new accidental!second!language grammarlib.ugent.be/fulltxt/rug01/001/891/626/rug01-001891626... ·...
TRANSCRIPT
ACCIDENTAL SECOND LANGUAGE GRAMMAR
ACQUISITION OF ENGLISH BY FLEMISH
CHILDREN THROUGH WATCHING SUBTITLED TELEVISION PROGRAMMES OR FILMS A Contrastive Survey of ASO, TSO and B-‐Stream
Master dissertation Supervisor: Prof. Dr. M. Van Herreweghe Elien Becelaere Master of Arts in Linguistics and Literature: Dutch – English 2011-‐2012
FACULTY OF ARTS AND PHILOSOPHY
2
PREFACE
First of all, I would like to thank all the students who were so kind to participate in the study.
Secondly, many thanks go out to Hendrik Provost and Pieter Blomme from the Don Bosco College
Zwijnaarde, to Tom Scheerens from Visitatie Mariakerke, to Christel Lekens from the Atheneum
Merelbeke, and to Franciska Gruwez from the Don Bosco Institute Sint-‐Denijs-‐Westrem, who made
this study possible in the first place by letting me conduct the research at their school.
Lastly, a special thanks goes out as well to the teachers who gave up some of their teaching time for
my research.
Many thanks go out as well to my supervisor, Prof. Dr. Mieke Van Herreweghe who supported and
guided me throughout the whole process of conducting my research and writing my thesis.
Finally, I would also like to thank my family and boyfriend for their continuous support during the
preparation and writing of my thesis.
3
TABLE OF CONTENTS
1 INTRODUCTION ......................................................................................................................... 5
2 LITERATURE SURVEY .................................................................................................................. 9
2.1 SECOND LANGUAGE ACQUISITION .................................................................................................. 9 2.2 THE INFLUENCE OF MULTIMEDIA AND SUBTITLES .............................................................................. 15 2.2.1 Multimedia ............................................................................................................................... 15 2.2.2 Subtitling ................................................................................................................................... 17 2.3 MULTIMEDIA LEARNING AND THE COGNITIVE LOAD THEORY ............................................................... 24 2.3.1 A cognitive theory of learning .................................................................................................. 25 2.3.2 Principles in multimedia learning ............................................................................................. 29
3 RESEARCH ............................................................................................................................... 34
3.1 EDUCATIONAL SYSTEM IN FLANDERS ............................................................................................ 34 3.2 PARTICIPANTS ......................................................................................................................... 37 3.3 DESIGN OF THE RESEARCH .......................................................................................................... 38 3.4 GRAMMATICAL FOCI ................................................................................................................. 41 3.5 PROCEDURE ............................................................................................................................ 42 3.5.1 BSO ........................................................................................................................................... 43 3.5.2 TSO ............................................................................................................................................ 44 3.5.3 ASO ........................................................................................................................................... 44 3.6 PROCESSING OF RESULTS ........................................................................................................... 45
4 RESULTS .................................................................................................................................. 47
4.1 OVERVIEW RESULTS SURVEY ....................................................................................................... 47 4.2 CONTRASTIVE ANALYSIS RESULTS TESTS WITH FRAGMENT AND EDUCATION TYPES ................................... 49 4.2.1 Contrastive analysis results in general ...................................................................................... 49 4.2.2 Contrastive analysis results grammatical foci ........................................................................... 52 4.2.3 Contrastive analysis results questions fragment and ‘new’ questions ..................................... 56 4.3 CONTRASTIVE ANALYSIS EXPOSURE TO ENGLISH AND RESULTS TESTS .................................................... 58 4.4 CONTRASTIVE ANALYSIS VIEWING HABITS AND RESULTS TESTS ............................................................ 61 4.5 CONTRASTIVE ANALYSIS RESULTS TESTS AND THE LANGUAGE PROFICIENCY IN ENGLISH ............................. 63
5 DISCUSSION ............................................................................................................................ 66
6 CONCLUSION .......................................................................................................................... 71
7 REFERENCES ............................................................................................................................ 73
8 APPENDIXES ............................................................................................................................ 76
8.1 APPENDIX 1 – THE SURVEY ........................................................................................................ 76 8.2 APPENDIX 2 – TRANSCRIPT OF FRAGMENT SHREK ........................................................................... 79 8.3 APPENDIX 3 – THE TEST ............................................................................................................. 86 8.4 APPENDIX 4: RESULTS ............................................................................................................... 90
4
8.4.1 Figure 1: Table Crosstabulation Education type with Contact with English. ............................ 90 8.4.2 Figure 2: Crosstabulation Education type with frequency of watching English television
programmes and movies. ......................................................................................................... 90 8.4.3 Figure 3: Crosstabulation Education type with the frequency of using Dutch subtitles. .......... 90 8.4.4 Figure 4: Crosstabulation Education type with the use of English. ........................................... 91 8.4.5 Figure 5: Boxplot graph correlation of the results in total with the three education types. .... 91 8.4.6 Figure 6: Boxplot graph of the correlation results in total with the three fragment types. ..... 92 8.4.7 Figure 7: Graph comparing the results on the first and second test from each education type. .......................................................................................................................................... 92 8.4.8 Figure 8: Graph showing in percentages the results on each grammatical aspect correlated to
the fragment types. .................................................................................................................. 93 8.4.9 Figure 9: Graph showing in percentages the results on each grammatical aspect correlated to
the education types. ................................................................................................................. 93 8.4.10 Figure 10: Graph showing the comparison of the results in total of both fragment questions
and non-‐fragment questions, in correlation to the three education types. ............................. 94 8.4.11 Figure 11: Boxplot graph that correlates the results on the first test to the extent of contact
with English. .............................................................................................................................. 95 8.4.12 Figure 12: Boxplot graph that correlates the results on the second test to the extent of
contact with English. ................................................................................................................. 96 8.4.13 Figure 13: Boxplot graph with the correlation of the results of the first test to the extent of
contact with English from participants from ASO. ................................................................... 97 8.4.14 Figure 14: Summary of the cases that are processed in the correlation of the results from the
first test to the types of television programmes and films watched most by all participants. . 97 8.4.15 Figure 15: Boxplot graph with the correlation of the results from the first test to the types of
television programmes and films watched most by all participants. ....................................... 98 8.4.16 Figure 16: Boxplot graph with the correlation of the results from the first test to the ability of
speaking English of all participants. .......................................................................................... 99 8.4.17 Figure 17: Boxplot graph with the correlation of the results from the second test to the ability
of speaking English of all participants. .................................................................................... 100
5
1 INTRODUCTION Upon talking about the rise of English to an old teacher of mine that teaches English in the first years
of secondary education, she spontaneously acknowledged how she had already perceived that her
students, before having received any instruction in English yet, had already acquired some English.
Surprisingly, she only referred to vocabulary acquisition and stated that she did not observe any, or
only very little, grammatical knowledge. Secondly, I had already observed myself that at the end of
secondary education, after having received eight years of French instruction and six to five years of
less intensive English instruction most Flemish students are far more fluent in English than they are in
French, despite the difference in length and intensity of instruction. Research by Housen, Janssens
and Pierrard (2001, in Kuppens, 2010:68) has also shown this discrepancy in language proficiency in
students from Flemish schools.
I found this very intriguing. It is already more or less common knowledge that the immense exposure
to English, mostly through popular media, influences our knowledge of English. Incidental acquisition
of English vocabulary has already been confirmed by research, but grammar acquisition has not.
However, it would be rather odd if incidental language acquisition was restricted to only vocabulary
acquisition. Especially when we consider the omnipresence of English in the Flemish society. In
contrast to French, English is the most extensively used language in most of the popular media, and
in other domains of society as well. Recent research by Kuppens (2010), Charlotte Lippens (2010) and
Caroline Lippens (2010) already showed the influence of different types of media such as video,
gaming, internet and music on the incidental acquisition of English by Flemish children, prior to
instruction. My own study (Becelaere, 2011) that was conducted for my bachelor paper, also
indicated that Flemish children, prior to English instruction, already mastered a level in English that
was higher than expected. In that research, the influence of subtitled television on the implicit
English grammar acquisition of children in their first year of secondary ASO-‐education was
investigated. The design of the research was threefold and was also used as a basis for the research
in this study. First, the participants were asked to fill in a questionnaire that enquired about their
background; age, gender, mother tongue language and the like. Secondly, they watched a fragment
after which they had to take a small test, the third part of the research. The test consisted of twenty
questions that were all intended to test the participants’ proficiency in four grammatical aspects of
English. The participants were presented with a Dutch sentence and three possible English
translations of which only one was correct. The participants were asked to indicate the correct
answer. Two groups were involved in the research, one group watched an English-‐spoken video-‐
6
fragment with Dutch subtitles and the other group watched an English-‐spoken fragment with English
subtitles. All participants were students in general secondary education (ASO). No significant
differences in performance on the test were found between the two groups, however. The study did
indicate that the participants had already implicitly acquired some knowledge of English, but it was
not able to explicitly indicate that implicit grammar acquisition had taken place. Implicit vocabulary
acquisition through watching subtitled television and films has already been confirmed through
other researches, but grammar acquisition has not. It seems highly unlikely, however, that no
grammar knowledge is acquired as well through the exposure to English. As will become clear in the
literary study, other researchers are also convinced that implicit grammar acquisition occurs as well.
In addition, there is the rise of multimedia technologies and other media such as video, computer,
gaming and other means, that also contributed to the development of English as a lingua franca.
Since their emergence, these new technologies have been recommended as tools in learning
environments. They are used more and more for educational purposes in different knowledge
domains, including langue learning. That is also why educational sciences, and more specifically
Mayer’s Cognitive theory of Multimedia Learning (CTML) (2009) and Sweller’s Cognitive Load theory
(Van Merriënboer and Sweller, 2005), are included in this dissertation. In addition to the insights that
these theories provide in the process of learning and the possible value of multimedia on that
process, it also caused us to include an extra type of video in this study. According to the CTML
(Mayer, 2009) subtitles would in fact obstruct the learning process.
Previous research (Charlotte Lippens, 2010:127-‐128) has shown that there is a difference between
students of the three main types in secondary education concerning the extent of acquisition of
English vocabulary prior to instruction and through media. However, this is only one study and it did
not focus on either subtitled video or on grammar acquisition. Therefore it was also decided to
include the two other main education types of secondary education in Flanders in the research as
well, thus students from AS0, BSO and TSO participated in the research. It would be interesting to
see if the results from this research would support the same conclusion or if different outcomes
would be obtained. In addition, the test is carried out a second time now, two weeks after the first
test was taken. This way we hope to ascertain whether the fragment has any enduring effects. We
also acknowledge that it is difficult to determine the influence of subtitled television on the
incidental grammar acquisition in English by means of only one fragment. Especially when it is very
likely that the participants have already acquired prior knowledge of English through the exposure to
popular English media, such as subtitled television programmes and films. For that reason, the test
consists of sentences used in the fragment and sentences that are not, that are new to the
7
participants. This way, we hope to observe the influence of the seen fragment.
In conclusion, the main research questions of this dissertation are:
(1) Have the participants incidentally acquired English grammar by watching the fragment? The
influence from the fragment will be determined in a threefold manner: by comparing the overall
results on the test, by comparing the results from the questions from which the sentences are
used in the fragment that is shown and from the questions that use sentences that are new to the
participants, and by comparing the results on each of the four grammatical aspects that were
focused on in the research: question words, adverbs ending in -‐ly, negation, and the correct
sentence order.
(2) Which of the three different types of fragment exerts the greatest influence on incidental
grammar acquisition of participants who have not yet received instruction in English; English-‐
spoken with English subtitles, English-‐spoken with Dutch subtitles, or English-‐spoken without
subtitles?
(3) Is there a substantial difference between the participants from ASO-‐, BSO-‐, and TSO-‐secondary
education in either the extent of influence on implicit grammar acquisition or in which subtitle-‐
type exerts the most influence on implicit grammar acquisition?
The subquestions in this dissertation are:
a) Is there a difference between the results achieved by participants who stated to have a lot of
contact with English and the results from participants who stated to have little contact with
English trough English-‐spoken media or other means? There will be a strong focus on the
influence through the exposure to subtitled television programmes and films.
b) Do participants who state to be proficient in English achieve better results in the two tests, than
participants who stated not to be proficient in English?
We will try to formulate an answer to these questions by means of the conducted research. It is
expected that indications for implicit grammar acquisition of English will become apparent in this
research. According to linguistics, the groups that saw the fragments with Dutch or English subtitles
are expected to achieve the best results. According to educational sciences, the group that saw the
fragment without subtitles is expected to achieve the best results. A second expectation is that
differences between the participants from the three education types will become clear, both in the
results from the tests as in the survey. Lastly, participants who stated to come into contact on a
regular basis with anglophone media, in particular with subtitled television programmes and films,
are expected to achieve higher results than participants who only occasionally come into contact
with anglophone media.
8
In this dissertation we start with an introduction on second language acquisition, after which the
influence of media on incidental language acquisition will be discussed, with a focus on the influence
of subtitles. This will be followed by insights from educational sciences, in particular Sweller’s
cognitive load theory and Mayer’s multimedia theory will be discussed in detail here. Then the
methodology of the research will be clarified, after which the results of the research will be
considered and analysed. Finally, we will be able to form a conclusion on the incidental acquisition of
English grammar and the role of watching subtitled television programmes and films in that process.
9
2 LITERATURE SURVEY In this section, we will discuss the theoretical aspects that provide the theoretical framework for our
research and that have initiated us to implement it. First, we will discuss the concept of second
language acquisition, relying more specifically on the views from Krashen (1987) and Rod (1994).
After that the influence of multimedia and subtitles will be considered and elaborated on, and to
conclude, we will discuss Mayer’s multimedia learning theory (2009), while also considering Sweller’s
cognitive load theory (Van Merriënboer and Sweller, 2005).
2.1 Second language acquisition
Since this research is concerned with second language acquisition (SLA), some elaboration on this
concept is needed. This theoretical framework is mainly based on the work of Krashen (1987), with
additional observations from Ellis (1994). The framework is limited to mostly these two researchers
because it is not the intention to elaborate too much on the theoretical aspects of SLA and because
we believe that this already represents SLA in a sufficient manner for this research. With second
language acquisition we mean the acquisition of a foreign language that is not spoken (frequently) in
the direct environment of the learner. The most conventional manner to come into contact with the
foreign language is in an educational setting. However, popular media such as television, internet
and radio are nowadays a widespread means as well for coming into contact with foreign languages,
as is also the case in this research. SLA does not necessarily refer to the second language that a
learner acquires, it can just as well be used to refer to the third or fourth language that the learner
acquires and that is a foreign language. Thus, in this research SLA is used to refer to the acquisition of
any foreign language that has no (official) social function in the learner’s community. We also make a
distinction between implicit and explicit language acquisition. Explicit language acquisition takes
place in a formal instructional setting, whereas implicit language acquisition occurs in an informal
non-‐instructional setting. Krashen refers to this distinction as respectively language learning and
language acquisition. Language learning implies a conscious process and conscious knowledge of the
foreign language (1987:10), the learning is explicit. Language acquisition is implicit or incidental, it
implies a ‘subconscious process’ (1987:10). The acquisition of a language is a slow process and occurs
only ‘when comprehension of real messages occurs’ (1987:67). No distinction is made here between
adults and children, adults acquire a language in the same mode as children (1987:10).
10
Krashen also believes that these two varieties of mastering languages both lead to a different type or
system of knowledge that are not complementary, they ‘coexist strictly independently’ (Van Lommel
et al., 2006:245). This is also known as the ‘non-‐interface position’ (Ellis, 1994:356). Ellis (1994) refers
to this distinction as ‘naturalistic and instructed second language acquisition’ (1994:12), but he
strongly objects to the idea that implicit language acquisition occurs subconsciously while explicit SLA
would occur consciously. He also objects to the idea that these lead to two types of knowledge that
are entirely separate from one another (1994:356). In this dissertation, we will not make a distinction
between language learning and language acquisition, these terms will be used interchangeably. To
distinguish between non-‐instructed and instructed language learning, we will refer to the terms
implicit and explicit second language acquisition or learning. Krashen (1987), however, appears to
have a valid point when distinguishing between these two modes of language learning by arguing
that implicit language acquisition occurs in a less conscious manner, thus implying a difference in the
language learning process in the two modes. Ellis (1994) does not explicitly agree on this point, but
he also does not exclude the option that the language learning process differs in implicit and explicit
SLA, it is ‘an open question’ (1994:12). The belief that language knowledge coming from implicit or
explicit language learning concerns two different types of knowledge that do not complement one
another and are entirely separate from one another is, however, harder to grasp. Especially since no
actual confirmation from any research has yet been given for this hypothesis.
In addition to this distinction between acquisition and learning, Krashen (1987) has developed four
other hypotheses: the natural order hypothesis, the monitor hypothesis, the input hypothesis and
the affective filter hypothesis (1987:9). These hypotheses are all closely interrelated and often
depend on one another. The natural order hypothesis implies that there is a ‘natural order’ in
acquiring the grammar of a second language that is similar to the grammar acquisition in the first
language. This order is different for each language, but develops in more or less the same manner for
the same language (1987:12-‐13). The most important value of this hypothesis lies for Krashen in the
implication that grammatical sequencing in language learning is abundant and should thus be
rejected (1987:14). Ellis (1994) disagrees on this point. He also believes that there is ‘a natural order
of acquisition’, which he also refers to as developmental sequences (1994:20). Nonetheless,
according to Ellis, there is evidence that suggests the usefulness of grammar instruction. But the
grammatical rules presented in the grammatical instruction need to be adjusted to the appropriate
developmental sequence in the acquisition order of the learners, in order for the learners to be able
to integrate these new rules in their ‘mental grammars’ (1994:22). Instruction can not push students
at the right developmental sequence, because it does not ‘enable learners to ‘beat’ a developmental
sequence’ (1994:22). Research by Macaro and Masterman (2006:319-‐320) supports this assumption.
11
The monitor hypothesis on the other hand, again makes a clear distinction between language
acquisition and learning, but in functions now. In this view, knowledge from language acquisition
initiates the communication in the foreign language and leads to fluency in the foreign language,
while the knowledge coming from language learning only functions as a monitor (1987:15). However,
the use of this monitor is restricted and only applies when there is sufficient ‘Time’, a ‘Focus on form’
and knowledge of the rule (1987:16). In this view, acquisition is once more seen as essential while
learning is considered to be only secondary in SLA (1987:20).
The input hypothesis focuses on how language is acquired. Krashen (1994) states that in order to
acquire a language, the learner needs to be provided with foreign language input that is always one
level or stage above his or her already acquired level. If the acquirer is at stage i, the preferred input
should be i + 1, input which the acquirer should understand. The acquirer is not initially focussed on
form but on meaning, form is acquired as a result (1987:20-‐21). The acquirer will understand the
meaning by not only relying on his or her already acquired knowledge of the language, but by also
relying on information from the context, or knowledge of the world (1987:21). Providing input that is
slightly higher than the competence of the acquirer is in Krashen’s view also the most beneficial
method to encourage speaking. ‘Speaking fluency’ can not develop from instruction, but needs to
develop over time and will only emerge when the acquirer is ready. Providing ‘comprehensible input’
will support this process (1987:22). The concept ‘comprehensible input’ needs some clarification.
This input is not expected to exactly correspond to the required i + 1, it does not need to be ‘finely-‐
tuned’ to the acquirer’s competence (1987:22). This condition is supported by the characteristics of
‘caretaker speech’ in first language acquisition by children (1987:22). This type of speech is not
specifically aimed at i + 1 and tends to adjust itself according to the progress of the child. In addition,
it is foremost intended to be comprehensible and it is consistent with the ‘“here and now” principle’
(1987:22-‐23). Comprehensible input as Krashen describes it is quite similar to caretaker speech.
Similar types of speech in SLA are ‘foreigner-‐talk’, ‘teacher-‐talk’ and ‘interlanguage-‐talk’ (speech of
other learners of the second language) (1987:24). These provide the acquirer with roughly tuned
input. An unexpected implication of this hypothesis is that for beginners, the classroom might be
very useful since it provides learners with the required level of input (1987:30). Although Krashen
does emphasis the importance of understanding here, he also states clearly that comprehension
alone ‘is not sufficient’ for acquiring a language (1987:66).
Ellis (1994) has some theoretical objections to the input hypothesis from which two will be
mentioned here. The first objection is that learning and comprehension are two different processes
that might not even be related to one another (Sharwood Smith, 1986 in Ellis, 1994:27). This view
12
does not object to the possible supportive position of comprehension in learning, but it does accord
less importance to its role in the language learning process. In addition, Tschirner (2001) also agrees
that comprehension does not necessarily lead to language learning. He reasons that when too much
focus is on comprehension, it will prevent the learner from actual learning because not enough
attention is given to linguistic information (2001:313). Secondly, Ellis points out the lack of ‘learner
output’. In the input hypothesis the learner is expected to acquire the language solely through
receiving input (1994:27). However, various researchers consider learner output to be an essential
aspect of language acquisition as well. Vanderplank explicitly states the importance of ‘productive
practice’ in the language learning process when he considers the influence of subtitled television on
language acquisition (1988:279). Swain’s ‘comprehensible output hypothesis’ (1985, in Ellis, 1994:27)
incorporates this notion that (pushed) learner output is essential. This hypothesis states that learners
will only acquire or learn to apply certain grammatical features through pushed output. Learners are
forced to pay attention to the correct form and will learn to appropriately produce the foreign
language through actual use and trial and error (Ellis, 1994:27, Krashen, 1998:175). Krashen strongly
disagrees with this view. He believes that a foreign language can be acquired up to a very high level
without the occurrence of actual language production in the learning process (1998:177). In addition,
Krashen states that forcing learners to produce output could actually harm the learning process,
because it could possibly provoke anxiety (1998:179), which negatively affects language acquisition
(Gardner, Tremblay and Masgoret, 1997:345). Learners should only produce output when they are
ready to produce output. Research by Nobuyoshi and Ellis (1993 in Ellis, 1994:283) indicated that
pushed output might not lead to language acquisition, but it could lead to an improvement of
student’s implementation of already known grammatical features (Ellis, 1994:283-‐284). We could
also discuss whether the input of subtitled television can be perceived as comprehensible input to
the learner. Krashen has not mentioned subtitled television, but according to him television in the
foreign language does not provide comprehensible input and consequently does not lead to
language acquisition (1987:63). The visual support from television is not enough to encourage
language acquisition. However, in subtitled television there is additional support from the subtitling
itself. This support might provide learners with the required comprehension, as some researchers
already stated. We will discuss this topic more fully when we discuss comprehension strategies and
familiarity in the next section.
The last hypothesis, the affective filter hypothesis, deals with factors that relate to the learners. Ellis
states that these are ‘general factors’ that include ‘individual differences’ (1994:35). This hypothesis
mainly states that affective factors such as motivation, self-‐confidence and anxiety also affect
language acquisition or implicit SLA. Motivation in SLA is usually ‘integrative’ according to Krashen
13
(1987:31). A positive attitude towards SLA, a high motivation, no anxiety and a high self-‐confidence
will lead to a low affective filter in acquirers. These learners ‘will be more open to input’ and acquire
language more easily and deeply (1987:31). However, input remains the primary source of language
acquisition, these variables only affect how the acquisition of the second language develops
(1987:32). Krashen states that this hypothesis mainly concerns language acquisition that occurs
subconsciously (1987:10), the acquirer is only aware of ‘using the language for communication’
(1987:10). In this case, motivation would seem to be aimed at comprehending what is being said in
order to be able to communicate. However, there are two objections here. First, Masgoret and
Gardner (2003) describe a student with integrative motivation as ‘one who is motivated to learn the
second language, has an openness to identification with the other language community, and has
favorable attitudes toward the learning situation’ (2003:128). This implies a more conscious state of
mind; the intention is to learn the second language, not only to understand it. This seems to conflict
with the opinion that the affective filter mostly influences incidental or subconscious SLA. Secondly,
if we agree with Ellis (1994:27) that comprehension does not necessarily lead to SLA, motivation
solely aimed at comprehension would scarcely influence language acquisition. The influence of these
factors is by no means denied, but it is questioned whether these factors really do affect implicit
language learning more than explicit language learning. In our view, the affective filter would apply in
both implicit and explicit language learning, presumably even stronger in explicit language learning.
In addition, Krashen (1987) refers to only a few factors whereas there are many more of these
factors recognised by other researchers such as Ellis (1994) and Gardner, Tremblay and Masgoret
(1997). These aspects all affect the language learning process and are each more learner-‐specific
characteristics. Aspects like age and gender will not be discussed here because they are not relevant
in this research. Motivation and attitudes are perceived as important aspects in language learning.
These two factors can be influenced greatly by emotional and cultural knowledge. Emotional and
cultural knowledge are very important in Tschirner’s view (2001). In the multicultural society we live
in, ‘intercultural understanding and appropriate cross-‐cultural behaviour are likely to become
essential skills for all humans’ (Tschirner, 2001:312). Cultural knowledge can be achieved and
deducted to a great extent from authentic video documents. Cultural knowledge can also help
students to achieve language knowledge. Emotional learning is mostly important because it can
improve the feeling of identification (Tschirner, 2001:316). Identification with the native speakers of
the foreign language is important to motivate students to acquire the language in question. This can
be enabled by the use of visual and auditory material that exercise great emotional influence.
Cultural knowledge can also play an important role in feelings of identification. Gardner, Tremblay
and Masgoret (1997:345) state that when a student feels positively connected or is interested in the
14
culture of the foreign language, the student will be more motivated to acquire that foreign language
than one that can not relate to the culture. The same applies to the identification that Tschirner
referred to (Tschirner, 2001:316). One who can identify himself with the culture or speakers from the
foreign language is more apt to learning the language itself. These aspects are personal, but do have
a great influence on the language learning process. Gardner and Lambert even state that a strong
motivation originates in this feeling of identification or desire to identify with members of ‘the new
linguistic community’ (1959:272, in Masgoret and Gardner, 2003:131). The same applies for the
differences between students in self-‐confidence. Gardner et al. say that: ‘self-‐confidence includes
two components [...]: anxiety as the affective aspect and self-‐evaluation of proficiency as the
cognitive component’ (Gardner et al., 1997:346). A self-‐confident person possesses a higher level of
independence and is better able to choose the learning strategies fit for him-‐ or herself and adjusting
them to his or her own needs. Language anxiety has proven to have a negative effect on the learning
of a foreign language. Language anxiety incorporates anxiety of tests and the probability of having a
negative evaluation, fear of communicating in a foreign language, and so on. Strongly related to self-‐
confidence, is the intelligence and language-‐aptitude of a student. Students with a higher level of
intelligence and language aptitude are better equipped to acquire a foreign language, and this
advantage may lead to a higher level of (self-‐) confidence in using and learning a foreign language.
Both the elements intelligence and language aptitude are interrelated to the last factor of motivation
and attitude. It is evident that a high level of self-‐confidence in language acquisition and an aptitude
for learning foreign languages will positively influence the motivation and attitudes towards language
acquisition. And a high motivation and good attitude will lead to better language acquisition; the
learner will try harder to master the language and to find the fit language learning strategies for him
or her (Gardner et al., 1997).
For our research, the natural order hypothesis, the input hypothesis and the affective-‐filter
hypothesis are most interesting. If grammatical sequencing is not needed in grammatical instruction
and learners receive enough comprehensible input, we could expect learners to acquire grammatical
knowledge of a language without having received formal instruction. And as will become clear in the
following section, we can infer that under the right circumstances subtitled television does provide
learners with comprehensible input. In addition to that, it is also assumed that subtitled television
has a positive influence on the affective factors, which would result in a low affective filter and more
actual learning.
15
2.2 The influence of multimedia and subtitles
The influence of multimedia and subtitles will first be considered from a more linguistic angle, after
which we will have a closer look into the Multimedia Learning theory from Mayer (2009) and the
Cognitive Load theory from Sweller (Van Merriënboer and Sweller, 2005). We will compare these two
views on the influence and possible use of (multi) media and already make some observations.
Furthermore, it should also be made clear what is signified with the terms media and multimedia in
this dissertation. The definition found in the Concise Oxford English Dictionary is used as a starting
point (2008). With media we mean ‘the main means of mass communication (especially television,
radio, and newspapers) regarded collectively’ (COED, 2008:886). Following this description, internet
can also be viewed as a form of media. Communication in this paper also includes music and
television programmes or films. Mostly, media in this sense will be referred to as popular media or
culture, including gaming as well. Multimedia, on the other hand, is a more technical and specific
term. According to the Concise Oxford English Dictionary, multimedia refers to ‘using more than one
medium of expression or communication’ (COED, 2008:938). This is also more or less what Mayer
implies when he refers to multimedia, a multimedia message or presentation includes ‘presentations
involving words and pictures’ (2009:5). It includes two different types of communication-‐media. In
this dissertation, what will be discussed is the influence of the medium television, more specifically
subtitled television programmes and films. A differentiation will be made between a presentation
with two communication modes, a pictorial and a spoken verbal mode, and the same presentation
with an extra written verbal mode in the form of subtitles, either in the same language as the spoken
verbal mode or in a different language.
2.2.1 Multimedia
Multimedia like television, video games and the internet nowadays provide an informal learning
environment. Children and adults incidentally acquire a language and cognitive skills by making use
of these technologies (Greenfield, 2009:71). In Flanders, English is the foreign language that is most
widely available in both a formal and an informal manner, it is omnipresent in society. Two examples
are the use of English as a language of communication in many work domains and the presence of
English in education, children receive formal instruction in English from secondary education on and
the use of English in higher education is greatly increasing. Next to that, English is also very present in
popular culture. It is the dominating language in music, in television programmes and films, on the
internet, and in gaming. In a research conducted by Caroline Lippens (2010) to investigate the extent
16
of English influence in the everyday lives of Flemish children before formal instruction, it already
became clear that children are mostly exposed to English ‘through popular culture, especially
through watching English television programmes, playing English games, and listening to English
music’ (Caroline Lippens, 2010:143). Other studies by Charlotte Lippens (2010) and Kuppens (2010)
support the hypothesis about the influence of these media on the incidental acquisition of English
from Flemish youngsters prior to formal instruction. According to Charlotte Lippens ‘it is reasonable
to state that the informants have a basic of English vocabulary, which differs in size for each
individual’ (2010:128). Similar effects have been noted in Finland, where the situation is comparable
to the Flemish situation. Björklund (2000, in Sjöholm, 2004:689) found similar results concerning
incidental acquisition of English by primary schools pupils before formal instruction in the language.
The pupils already acquired ‘a remarkable number of English words and phrases’ and the majority of
the pupils also obtained relatively good results on ‘an English listening-‐comprehension test’
(2004:689).
Seeing that exposure to multimedia already incidentally leads to language learning, it can be
suspected that media and multimedia could also be used successfully in formal language learning
environments. Although, it is very likely that these devices first need some alterations or
manipulations to be in accordance with the learning objectives. Tschirner (2001) acknowledges the
utility of multimedia, and more specifically of digital video, in language learning environments as
well, and he also recognises that multimedia can not be implemented as such in classrooms.
Tschirner (2001) points out that ‘[t]he use of new media in language classes will only improve
teaching and learning when it goes hand in hand with curricular and methodological innovation’
(2001:306). Tschirner agrees on this point with the ‘Learner-‐Centered Approach’ favoured by Mayer
(2009:13), and he proposes to look at SLA theories when implementing multimedia in language
learning environments to support language learning in the most effective manner possible
(Tschirner, 2001:306). In this belief, technology needs to be adapted to the study material and the
learners, not the other way around (2009:12). This notion will be elaborated on further on, when
discussing Mayer’s multimedia learning theory.
According to Tschirner (2001) and Sjöholm (2004), the exposure to English through popular media
and culture also plays a great role in the improvement of motivation in language learning. Following
Gardner et al. (1997), motivation can be defined as ‘the individual’s attitudes, desires, and effort to
learn the L2’ (1997:345). Sjöholm (2004) indicates that young Swedish learners of English in Finland
show a very positive attitude towards learning English. However, this positive attitude does not seem
‘to stem from classroom experiences, but rather from experiences from outside the school context’
17
(2004:688), presumably from English input through mass media (2004:692). Tschirner (2001)
attributes the improvement in motivation in part to another possible advantage of digital video.
Digital video, with or without subtitles, allows learners to come into contact or to gain insight into
the culture of the speakers of the foreign language. Learners can acquire ‘cultural knowledge’ that
can lead to identification with the foreign language speakers, an important factor in SLA (2001:312).
However, we do need to make some observations concerning our research. We are researching the
incidental language acquisition of children that are about twelve years old. Many of the television
programmes and films that are designed for these children, or younger children, take place in a
fantasy world. Popular films like Harry Potter, Shrek, Ice Age and so many others only display in part
the culture of the native speakers of English. Another problem is the notion of a native speaker or
culture of English. English is a lingua franca that has many native speakers all over the world. The two
most common varieties in Western Europe are British and American English. In education, British
English is mostly the norm but outside the classroom American English is often the dominant variaty.
Thus we can not always assign cultural value to a language, or to the television programmes and
films.
2.2.2 Subtitling
That (multi)media positively influences language acquisition to some extent is clearly obvious. In this
paper, however, the emphasis lays on the effects of subtitled television programmes or films in
informal learning environments on language acquisition. Subtitling is not a homogenous concept,
there are three different types of subtitling: standard subtitling, reversed subtitling and intralingual
subtitling or captioning (Koolstra & Beentjes, 1999:53). Standard subtitling is the most common type
with a foreign language soundtrack and native language subtitles. It is mostly used in small countries
as an adaptation method to make foreign language programmes and movies available and
understandable for the viewers in those countries. Subtitling is preferred to dubbing in small
countries because of its relatively low cost (Koolstra and Beentjes, 1999:52). Countries in Europe
where subtitling is very common are ‘Belgium, Denmark, Finland, Greece, Luxembourg, the
Netherlands, Portugal and Sweden’ (Koolstra, Peeters and Spinhof, 2002:325). Another type of
subtitling is reversed subtitling. Here the soundtrack is spoken in the native language and subtitles
are in the foreign language. Captioning or intralingual subtitling is more specifically for deaf or
hearing-‐impaired people as both the soundtrack and the subtitles are in the same language.
However, this type of subtitling is also used occasionally for language teaching purposes (Koolstra et
al., 1999:53). In this research we are concerned with possible differences in influence of SLA between
standard subtitling, intralingual subtitling, and no subtitling but foreign language soundtrack.
18
Digital video with standard subtitling is a very informative format, it presents the viewer with three
different types of information: ‘the visual image, the soundtrack in a foreign language; and the
subtitles in the native language’ (d’Ydewalle and De Bruyker, 2007:196). Digital video with
intralingual subtitling also offers three different media of information, but the verbal information is
all in the same language, the foreign language. Both types supply the viewers with a ‘very rich input
environment’ and an access to authentic language use of the foreign language (Tschirner, 2001:306).
Research by d’Ydewalle and Gielen (1992, in d’Ydewalle and De Bruyker, 2007:196) showed that
subtitles are read in a more or less compulsory way, paying attention to both the visual image and
the subtitles happens automatically. d’Ydewalle and De Bruyker also investigated the reading
behaviour when watching subtitled television and compared the reading behaviour in standard
subtitling with that in reversed subtitles. They found that processing standard subtitles involves
‘typical reading behaviour’ (2007:204), while reversed subtitling shows a less regular reading pattern
(207:203). However, even in the reversed subtitling-‐situation where the subtitles are actually
redundant for comprehension, the subtitles are still read and automatically processed to some
extent (2007:203). D’Ydewalle and De Bruyker do not know though to what extent this processing
took place and if this resulted in language acquisition, since they did not investigate this. They are
convinced, however, that language acquisition, mainly vocabulary acquisition, takes place when
watching subtitled television, both in the standard and reversed subtitling version (2007:204).
In another research by d’Ydewalle in cooperation with Van de Poel (1999), one objective was to test
the sensitive language acquisition hypothesis by comparing the results from younger and older
children with standard and reversed subtitling. The research showed clear vocabulary acquisition in
both groups, but it did not support the sensitive language acquisition hypothesis. Both the younger
and older children performed better in the standard subtitling mode, with the foreign language in
the soundtrack. However, the researchers did comment that previous studies had noted that adults,
on the other hand, showed better vocabulary acquisition in the reversed subtitling mode than in the
standard subtitling mode (Pavakanun and d’Ydewalle, 1992; d’Ydewalle and Pavakanun, 1996, 1997,
in 1999:241). This suggests that adults benefit more from watching subtitled television when the
foreign language is visually displayed in the subtitles, while children benefit more when the foreign
language is aurally present in the soundtrack and the mother tongue in the subtitles. D’Ydewalle and
Van de Poel gave two possible explanations here. Firstly, that the way in which children and adults
process information differs and secondly that younger children simply have not yet developed
‘sufficient reading abilities’ (1999: 241-‐242). Danan (2004) suggested that this difference in
performance between older and younger children or adults may indicate a difference in familiarity
with subtitled programmes (2004:73-‐74). When viewers of subtitled programmes are not familiar
19
with them, they have not yet developed ‘strategies to process subtitles efficiently and derive the
most benefits from them’ (Danan, 2004:74). The subtitles may lead to distraction instead of
comprehension and learning.
Research by Koolstra and Beentjes (1999) has shown implicitly that children in subtitling countries
already know how to handle subtitles from a young age on and are quickly familiar with the concept.
They conducted a research to investigate the effect of subtitles on children that had not yet received
instruction in the foreign language and on children that had already received instruction in the
foreign language. The children in the first group were on average between nine and ten years old,
the children from the second group were between eleven and twelve years old. The research was
conducted in the Netherlands, were subtitling is a common practice. The children viewed an English
programme with either standard subtitling, no subtitling or spoken in their native language. No
significant difference in vocabulary acquisition was found between the children that had already
received formal instruction and those that had not, but in both groups the strongest effects on
vocabulary acquisition was found in the standard subtitling type (1999:58). Koolstra and Beentjes
also conducted a word recognition test, in which children that viewed the subtitled programme
achieved again better results than those that viewed the non-‐subtitled programme. Thus, both the
younger and older children were already very familiar with standard subtitling. In addition, it was
suggested that the translation in the subtitles positively affects the word recognition because it
presents the participants with both visual and auditory input (1999:58). Thus, Koolstra and Beentjes
(1999) believe that standard subtitling would make it easer to recognise and process the foreign
language.
They also concluded from their word recognition test that although the subtitles are automatically
processed by viewers, the soundtrack is processed as well. The participants in their study, who were
familiar with subtitles, were able to switch effortlessly ‘between reading the subtitles and listening to
the English words spoken in the soundtrack’ (1999:58). Research by Bird and Williams (2002), Borrás
and Lafayette (1994) and Vanderplank (1988) pointed in the same direction. In all three studies,
intralingual subtitling was investigated and the participants were all learners of the foreign language
issued in the studies. Bird and Williams found indications that even though there was a strong
reliance on the visual support from the text, the sound was fully processed as well by the participants
(2002:517). Borrás and Lafayette (1994) came to the conclusion that simultaneous listening and
reading occurs, but that a certain amount of familiarity is beneficial or even necessary (1994:70). An
empirical research by Vanderplank (1988) supports this view as well, although in a more indirect
manner. The objective of the research was to have a better view on the attitudes of participants to
20
the use of intralingual subtitled television programmes as a supporting method for language
learning, and on how they experienced viewing intralingual subtitles in a foreign language (1988:273-‐
274). The participants indicated to notice inconsistencies between the soundtrack and the subtitles,
an inaccuracy that often occurs in subtitling. That implicitly indicates that the soundtrack is being
processed as well here.
As has already been stated above, Danan (2004) believes that familiarity to subtitles is ‘a key to
beneficial use of subtitled programs for language learning’ (2004:73). Familiarity with subtitles gives
the viewers time to adapt to them and to develop ‘strategies to process subtitles efficiently’
(2004:74). This belief is partly based on research by Vanderplank (1988), Koolstra and Beentjes
(1999) and Borrás and Lafayette (1994). Participants in Vanderplank’s (1988) research that were
unfamiliar with watching subtitled programmes indicated to have adjusted their way of viewing
television programmes and to have developed techniques to fully comprehend and take in the text
from the subtitles (1988:275). In addition, there was also one Danish participant who was familiar
with standard subtitles, and it was observed that this participant adapted more quickly to intralingual
subtitles (1988:276). Research by Koolstra and Beentjes (1999), indicated that older children showed
a higher extent of language acquisition than younger children (1999:58), although there was not a
significant difference in their research. The participants from the research lived in the Netherlands
and were thus accustomed to viewing subtitled English television programmes and films. Koolstra
and Beentjes ascribed the difference in language acquisition chiefly to the difference in frequency of
watching subtitled programmes at home, with the older children having had longer experience
viewing subtitled television at home (1999:59). Lastly, Borrás and Lafayette (1994) made an
interesting observation after working with participants who were unfamiliar with subtitles in their
research for the influence of subtitling on language acquisition. In concluding their research they
suggested that familiarity with subtitles might enhance ‘students’ attitudes toward and the effective
use of target language subtitles’ in language learning environments (1994:70).
Listening comprehension could in this perspective also be seen as a technique developed over time
by viewers of subtitled television. The visual input from the subtitles can support the auditory input,
and subsequently lead to a better understanding and listening comprehension. This listening
comprehension can improve or develop as well with increased familiarity (Danan, 2004:69-‐70).
Danan even describes intralingual subtitles as a ‘hearing-‐aid’ (2004:70). Borrás and Lafayette (1994)
did not describe or even mention listening comprehension this explicitly in their conclusion, but they
do agree that intralingual subtitles result in better comprehension. In addition they also emphasized
the importance of listening comprehension by concluding that this type of subtitling might help the
21
foreign language learner to ‘associate the aural and written forms of words more easily and quickly
than video without subtitles’ (1994:70). Bird and Williams (2002), however, also conclude from their
research the importance of the enhancement of listening comprehension in language learning when
viewing intralingual subtitled television, but they attribute this enhancement in comprehension also
to ‘faster decoding’ (2002:517). Through the visual presence or support of words, viewers are able to
make ‘a potentially more rapid judgement based on visual information’ (2002:517). Participants in
Vanderplank’s (1988) research reported that the intralingual subtitles made it easier or even enabled
them to follow and comprehend English television programmes, especially with ‘fast, authentic
speech and unfamiliar accents’ (1988:275). This textual support thus leads to a higher level of
understanding, but it also has another positive side-‐effect. Language learners appear to experience
less or no anxiety and to have more confidence due to the support through subtitles. Vanderplank
(1988) already observed that his participants were ‘relaxed and attentive’ (1988:275) while viewing
the programmes with intralingual subtitles in the research. Borrás and Lafayette (1994) perceived
better attitudes towards language learning in participants that had viewed the videos with subtitles
than in those that had not. They also believe that learners develop a higher ‘Perceived Self Efficacy’
(1994:70), a concept similar to self-‐confidence (Gardner et al., 1997:347). In addition, Borrás and
Lafayette also noticed a higher confidence and better performance when speaking the foreign
language in those participants who had watched the subtitled videos, than in those who had not.
(1994:69-‐71).
These studies also contradict Krashen (1987) when he says that television, or radio, is not an
adequate tool to acquire the beginnings of a language (1987:63). The studies above, especially those
on incidental language acquisition, have shown clearly that it is possible to acquire (the beginnings
of) a language on the basis of input from media, including subtitled television, without formal
instruction. Some nuances need to be made, however. First of all, Krashen speaks of television in the
foreign language without subtitling, in that sense the foreign language is much harder to
comprehend, even with the visual support. Secondly, Krashen has also based this assumption on his
own experience, he and his children watched some programmes with Spanish as the foreign
language for a couple of years, but did not acquire much Spanish aside from some very basic
vocabulary (1987:63). It can be expected, though, that aside from the television programmes, not
much other Spanish input was available. That is in firm contrast with the input that our participants
in Flanders receive, who are exposed to English to a large extent in multiple ways. As was perceived
by Kuppens (2010), this higher exposure does influence language acquisition in a positive way. The
participants who reported the highest exposure to English were also the participants who achieved
the highest results (2010:74). From the positive effects that watching subtitled television during SLA
22
has on affective factors such as self-‐confidence and motivation, it can be inferred that subtitled
television positively influences the low affective filter. It creates a ‘low-‐anxiety environment’ for the
learner (Danan, 2004:74), as both Danan (2004:74) and Vanderplank (1988:277) already indicated.
Secondly, the implication that watching subtitled television leads to a higher comprehension of the
foreign language can also indicate that subtitled television provides learners with comprehensible
input. However, as is already mentioned, familiarity with subtitles is an important factor here.
So far, it is quite clear that watching subtitled programmes can add value to the learning process. It is
not clear, however, which type of subtitling has the best effect. Danan (2004) suggests that
beginners should start with standard subtitling, but that more experienced learners benefit more
from intralingual subtitling (2004:75). There is a difference between the intentional and incidental
learning here. Research aimed at investigating intentional learning seems to favour intralingual
subtitling, while research on incidental learning seems to favour standard subtitling. Of course,
standard subtitling is the type encountered in incidental learning, but it would be interesting to see
whether intralingual subtitling would lead to more learning. Thus, research by Kuppens (2010),
Caroline Lippens (2010), Charlotte Lippens (2010) and Björklund (2000, in Sjöholm, 2004:689) has
shown the occurrence of incidental language acquisition through the influence of mass media,
including watching subtitled television and films, in children before formal instruction of the foreign
language in question, English. Caroline Lippens (2010) and Charlotte Lippens (2010) focused in their
research on the vocabulary knowledge of students. Vanderplank (1988) then, has shown in his
research that his participants, who were all language learners of the foreign language, all manifested
improved language acquisition after viewing intralingually subtitled television programmes. Two
other studies in intentional language learning environments where an improvement in language
acquisition and attitudes towards language learning were found, are those by Bird and Williams
(2002) and Borrás and Lafayette (1994). Both studies focused on intralingual subtitles. The only
research mentioned in this section that focused on the learning of vocabulary or grammar through
the use of subtitling as well, was that by Kuppens (2010). Kuppens investigated the extent of implicit
language acquisition in the different types of media. She compared the participants’ reported use of
those media to the results of her participants. She also specifically looked for the influence of
watching standard subtitled television and she sought for indications of language acquisition in
translation skills, the participants were asked to translate sentences from the foreign language into
the native language and vice versa. Especially translating from the native to the foreign language
involves grammatical knowledge as well. Watching subtitled television manifested significant effects
on the test, more than the other media researched (2010:78). The more a participant reported to
watch subtitled television, the higher the results were (2010:74-‐76).
23
d’Ydewalle and Van de Poel (1999) and Koolstra and Beentjes (1999) focused on the implicit
acquisition of vocabulary through the use of subtitles. D’Ydewalle and Van de Poel (1999) used
reversed and standard subtitling in their research, and showed that both types of subtitling led to
language acquisition, but that it appeared that older children and adults benefitted more from
reversed subtitling than from standard subtitling. Koolstra and Beentjes (1999) also found evidence
for vocabulary acquisition through subtitling, but they only used no subtitling or standard subtitling
in their research. d’Ydewalle and Van de Poel (1999), however, focussed on incidental grammar
acquisition as well in their research but the results from their syntax test were not significant to
manifest grammar acquisition of the foreign language (1999:234). This should not be too surprising,
though, since words are perceived to be ‘the easiest building blocks in acquiring a new language’
(1999:240), and are thus more easily to acquire. Secondly, grammar acquisition is also harder to
detect than vocabulary acquisition, especially in implicit language acquisition. Another explanation
comes from Tschirner (2001) who states that language learners would use ‘comprehension
strategies’ (2001:313) when listening to (authentic) foreign language use. Listeners focus primarily on
comprehending what is being said by relying on all clues available, also syntactic ones. However,
because the focus is very much on comprehending the meaning, less attention might be given to
learning language. Koolstra and Beentjes point out as well that understanding or determining the
meaning of words in an implicit or more natural manner is achieved by ‘making use of the semantic
and syntactical cues that the context provides’ (1999:51). They do believe language acquisition
follows from this, although they do not explicitly confirm or state that grammar acquisition takes
place as well, although they do imply it by referring to ‘syntactical cues’ as well (1999:51). There is,
however, not much support or evidence for incidental grammar acquisition through watching
subtitled television. Only Kuppens (2010) found some evidence of grammar acquisition, but that
acquisition is limited to those participants that very frequently view subtitled television. Van
Lommel, Laenen and d’Ydewalle (2006), like d’Ydewalle and Van de Poel (1999), did not find
significant support for grammar acquisition in both reversed and standard subtitling. Van Lommel et
al. did note a positive influence on the grammar test when the grammatical rules issued in the
fragment were presented before watching the subtitled fragment, but still the results were not
significant (2006:254). They argue that language acquisition is a gradual process in which vocabulary
acquisition should be obtained first before grammatical acquisition can occur (2006:255).
Thus, we can state with certitude that grammar acquisition through watching subtitled television is
harder to achieve than vocabulary acquisition. Nonetheless, grammar acquisition is expected to
occur. It is only harder to detect and since the viewers in our research acquire it implicitly, they are
not always able to explain why they use this or that particular form, or why that specific sentence
24
order is wrong. As Krashen (1987) puts it, they only have ‘a “feel” for correctness’ (1987:10). There is
no conscious awareness of the grammatical rules. Since the participants in our research have not yet
received formal instruction in the foreign language, English here, it will also be possible to ascertain
whether grammar acquisition has taken place already and to compare the results with the
participants’ reported frequency of watching English-‐spoken subtitled television. It is hard to specify
the extent of grammar acquisition of only one fragment, certainly because every participant has
already been exposed to English spoken media to a greater or lesser extent and is thus expected to
have already achieved some language knowledge. Finally, it is also very important to keep in mind
that viewing subtitled television is only a supporting device in the language acquisition process. This
device alone will not be sufficient to actually acquire a language. As Vanderplank states, reception of
language input is good, but ‘productive practice’ (1988:279) is also essential in the language learning
process.
2.3 Multimedia learning and the cognitive load theory
Now that we have obtained some more background in the linguistic view on second language
acquisition and the influence or positive effects from media and subtitled television on both
intentional and incidental SLA, we can acquire more insight into Mayer’s multimedia theory (2009)
and Sweller’s cognitive load theory (Van Merriënboer and Sweller, 2005). When discussing Mayer’s
Multimedia theory, it is important to note that this theory has not been tested to a great extent on
language learning yet. The topics used in the studies to test the learning effects were mostly derived
from the area of physics and mechanics, as Mayer (2009) has acknowledged as well: ‘the research
reported in this book focuses mainly on multimedia messages aimed at explaining how some systems
works, including mechanical, physical, and biological systems’ (2009:51). However, Mayer does not
limit his theory to just those fields of study, he always refers to the ‘multimedia instructional
message’ upon which his theory is applicable. That message is ‘a communication using words and
pictures that are intended to promote learning’ (2009:30). In a research by Mayer where multimedia
is tested in SLA, he finds that in language learning having multiple modes of representation, in both
the foreign and native language, is beneficial for learning as the word recognition is higher. However,
being able to choose the preferred mode of representation also leads to better language learning
(1998:33). Another remark is that in his theory, Mayer explicitly refers to intentional learning.
However, the processing of information and the limits to the amount of information in intentional
learning can be expected to be similar in incidental learning. Koolstra and Beentjes (1999) even
assume that with subtitled television, the intent and motivation to understand the foreign language
25
might even be higher in an incidental language learning environment where the viewers can choose
what they view themselves (1999:59). It can be argued that the same applies to intentional learning
environments.
2.3.1 A cognitive theory of learning
Mayer’s Multimedia Learning theory (2009) is basically built upon a cognitive theory of learning that
is based on ‘the human information-‐processing system’ (2009:61). This cognitive theory is based
upon three assumptions that are ‘derived from the learning sciences’ (2009:62). These three
assumptions are the Dual-‐Channel Assumption, the Limited-‐Capacity Assumption and the Active-‐
Processing Assumption (2009:64-‐67). With this cognitive theory of multimedia learning in mind,
Mayer researched the ways in which multimedia can positively affect the learning process, or which
methods do exactly the opposite. From this research he has derived twelve different principles for
designing instructional multimedia messages that encourage learning: Coherence, Signaling,
Redundancy, Spatial Contiguity, Temporal Contiguity, Segmenting, Pre-‐training, Modality,
Multimedia, Personalization, Voice and Image (2009:52-‐53). This theory is an example of the
previously mentioned ‘Learner-‐Centered approach’ (Mayer, 2009:13). These principles show how
instructional multimedia messages can and should be adjusted in favour of the learners. The learner
is central here, while the technology is subordinate and is adapted in function of the learner. The
cognitive theory and the three underlying assumptions will be discussed first, after which we will
discuss the principles that are useful for our research. We will already compare them to the
previously discussed findings from linguistics if necessary and discuss some dissimilarities.
A learner-‐centered approach also implies an understanding of how people learn and of how their
cognitive system works. That way learning environments, and more specifically instructional
multimedia messages in this case, can be designed to be ‘compatible with how people learn’
(2009:60). The cognitive model of multimedia learning is aimed at closely representing ‘the human
information-‐processing system’ (2009:61). First of all, a multimedia presentation implies that
different modes of information are presented in both a verbal or pictorial form. Verbal information
can include both spoken and written text, while pictorial information can include static graphics such
as pictures or dynamic graphics such as videos (2009:5). The cognitive theory of learning believes
that auditory and visual material are processed in different channels by humans, a visual and an
auditory channel (2009:9). At the basis of this model lies the dual-‐channel assumption, which will be
discussed later on. As can be seen in the image below, the model consists of three different
memories. First, there is the sensory memory, then the work memory and lastly there is the long-‐
26
term memory. These three memories represent more or less the three phases of successfully
processing information. The sensory memory receives the information first through the ears and
eyes, the auditory and visual receivers. The visual and auditory information is held for a very short
period in the respective segment of the sensory memory. Printed text will be held in the visual
sensory memory. From here, the visual and auditory information will be transmitted into the working
memory, where the actual processing of the information starts. Again, the information is held in this
storage for a short period and according to Mayer, it is here that knowledge is actively constructed.
As can be seen in the model, the visual and auditory information are transmitted into the verbal and
pictorial model in this memory. These are combined into one model, together with prior knowledge
that was already stored in the long-‐term memory. In contrast to the long-‐term memory, the storage-‐
space in the working memory is limited. Only small amounts of information can be processed at the
same time here. It is in the long-‐term memory where knowledge and information is held for a long
period of time. The presumptions of the limited capacity of the working memory and of the active
processing of information into knowledge are based on the assumptions of limited capacity and
active processing (Mayer, 2009:61-‐63).
1
First, the dual channel assumption (Mayer, 2009:64-‐66) states that humans have a different
information-‐processing channel for ‘visually represented material and auditorially represented
material’ (2009:64). It is a compromise between the presentation-‐mode approach and the sensory-‐
modality approach. The presentation-‐mode approach is closely related to Paivio’s dual-‐coding theory
(2009:65) and makes a distinction between the verbal and non-‐verbal presentation of information,
between words and illustrations and the like. The sensory-‐modality approach, however, distinguishes
information on whether it is initially processed visually or aurally, through the eyes or through the
ears. This approach closely relates to Baddeley’s model of working memory (2009:65). The
compromise here is that the sensory-‐modality approach is applied in the sensory memory to 1 McCormick, Bob. “Multimodal Instruction” Stuff4Educators.com. Stuff4Educatiors.com. Web. 3 May 2012. <http://stuff4educators.com/index.php?p=1_12_Multimodal-‐Instruction>
27
distinguish between the nature of the material, while the presentation mode is applied in the
working memory to make a distinction between models in the working memory that are constructed
on the basis of verbal or non-‐verbal information. Thus, information such as printed text, like
subtitles, is processed initially through the visual channel in the sensory memory, but in the working
memory it is processed as verbal information and constructed into the verbal model, possibly
together with spoken verbal information that is processed through the auditory channel.
Furthermore, depending on the cognitive effort that the learners dedicate to the learning activity,
information that was originally presented in one channel can also be represented in the other
channel (2009:66). A straightforward example from Mayer is that an experienced reader may
represent on-‐screen text such as subtitles that is initially processed in the visual channel, also in the
auditory channel by transforming the visual information into auditory information (2009:66). Mayer
calls this ‘cross-‐channel representations’ (2009:66).
Secondly, the limited-‐capacity assumption (2009:66-‐67) assumes that there is a limit to the amount
of information humans can process in the visual and auditory channels at a time. This implies that
not all information that is presented, is and can be processed by the memory, and that choices have
to be made concerning which information is processed or not and which information is used to
construct models in the working memory and which is not. Generally it is assumed that the average
memory span incorporates about ‘five to seven chunks’ of information (2009:67). A new approach to
this assumption is Sweller’s cognitive load theory (CLT) (Mayer, 2009:79-‐82). Sweller states that CLT
‘assumes a limited working memory that stores about seven elements but operates on just two or
four elements. It is able to deal with information for no more than twenty s [seconds] unless it is
refreshed by rehearsal’ (Van Merriënboer and Sweller, 2005:148). These limitations only apply to
novel information. In his theory, Sweller differentiates between three different types of cognitive
load: germane cognitive load, extraneous cognitive load and intrinsic cognitive load (Mayer, 2009:79-‐
81). Mayer defines germane cognitive load as ‘cognitive processing during learning that is aimed at
making sense of the essential material and that can be attributed to the learner’s level of motivation’
(2009:81). This type of cognitive load needs to be encouraged in the learning process. Extraneous
cognitive load is cognitive load that stems from the manner in which the information is presented to
the learner (free translation from Valcke, 2010:163). This is extra, unnecessary, cognitive load that
can be avoided by adapting the manner of presentation to the learner and learning material. Intrinsic
cognitive load then, is cognitive load that is inherent to the complexity of the information and is thus
invariable (Valcke, 2010:163). It can be managed, though, through limiting the information given at
one time. The complexity of the information depends on the ‘element interactivity’ of the
information or ‘the number of elements that must be processed simultaneously’ (Van Merriënboer
28
and Sweller, 2005:150). However, the issue of the limited capacity of the working memory is of
importance here. Germane cognitive load or ‘generative cognitive processing’ (2009:81) is the last
stage in processing, but if there is too much extraneous or intrinsic cognitive load, the learner will
not be able to reach the stage of germane cognitive load and will not actually learn (2009:81-‐82). In
conclusion, to obtain the best learning effects it is important to encourage germane load, by limiting
the intrinsic cognitive load and avoiding extraneous cognitive load.
Two related theories are the split-‐attention effect and the expertise reversal effect. The split-‐
attention effect occurs ‘when learners must split their attention between and mentally integrate
multiple sources of information’ (Yeung, Jin and Sweller, 1997:2). This can lead to an increased, and
unnecessary, cognitive load. This concept is also discussed in Greenfield as ‘divided attention’ in
multitasking (Greenfield, 2009:79). They noted that when multitasking was prerequisite ‘cognitive
processing was less mindful and more automatic’ (Foerde et al. in Greenfield, 2009:79). Moreno and
Mayer (1998) state that when split-‐attention occurs, the visual memory is overloaded and is thus
unable to process all information, it can ‘select fewer pieces of relevant information’ (1998:162). The
expertise reversal effect occurs when it appears that instructional methods work better for learners
with less expertise than for experienced learners. In some cases, the instructional methods could
even negatively effect the learner’s learning process (Van Merriënboer and Sweller, 2005:152). This
can be analogised to the concept of familiarity that was discussed earlier, although Danan (2004)
only discusses how familiarity can positively affect the learning process and increase the processing
of subtitles.
Finally, the active-‐processing assumption (2009:67-‐70) is in contrast with the more common
assumption that the human mind passively takes in information. Mayer believes that information is
processed and used to form a mental model or knowledge structure. Some of these structures
include ‘process, comparison, generalization, enumeration, and classification’ (2009:68). Three
crucial processes in active learning are: ‘selecting relevant material, organizing selected material, and
integrating selected material with existing knowledge’ (2009:70). Thus, learning is not just taking in
information but is actively processing information and forming knowledge structures with it, the
learner has an active role in the learning process. This implies that the design of the instructional
message should best be adjusted to the system of processing that is described here to facilitate the
processing of the information for the learners. Especially less experienced learners will benefit from
a design adapted to those needs. According to Mayer, the design of a multimedia message should
have ‘a coherent structure’ and the message should provide the learner with ‘guidance [...] on how
to build the structure’ (2009:69).
29
2.3.2 Principles in multimedia learning
Now that we have some more background on Mayer’s multimedia theory, we can have a closer look
into some of Mayer’s principles to improve multimedia design. All principles will be discussed, but
not all principles will be discussed to the same extent. Some principles are already more or less
inherent aspects of subtitled television. These principles include the multimedia principle, the spatial
contiguity principle, the temporal contiguity principle, the personalization principle and the voice
principle. Those principles will only be discussed briefly. More attention will be given to those
principles that contradict the positive influence of subtitles or those principles that when they are
applied to the medium, could lead to better results with subtitled television, especially in a learning
environment. Contradicting principles are the coherence principle, the redundancy principle and the
modality, while the principles that could possible enhance the effect of this medium in a formal
learning environment are the segmenting principle, the signaling principle and the pre-‐training
principle. These principles will be discussed last.
First, we will discuss the principles that support certain aspects that are inherent to the medium of
subtitled television. First of all, there is the multimedia principle. This principle basically states that a
multimedia message containing both words and pictures will enhance learning (2009:223). A
multimedia message consisting of both words and pictures provides the learner with both verbal and
non-‐verbal information about the same material. These two types of information present the same
information in a different manner and can thus complement each other, which could improve
learning (2009:228). In short, a multimedia message ‘allows learners to hold corresponding verbal
and pictorial representations in working memory at the same time, thus increasing the chances that
learners will be able to build mental connections between them’ (2009:229). Secondly, according to
the spatial contiguity principle learning improves ‘when corresponding words and pictures are
presented near rather than far from each on the page or screen’ (2009:135). This principle actually
supports the multimedia principle in that it encourages learners to make connections between the
visual and verbal information by presenting the two types of information in such close proximity and
thus it also positively influences the following knowledge structures (2009:141). This in turn reduces
the extraneous cognitive load of the multimedia message and could enable or encourage generative
cognitive processing (2009:142). The temporal contiguity principle complements the previous
principle in that it states that verbal and pictorial information should also be presented
simultaneously (2009:153). Again, this encourages learners to make connections between the two
different modes (2009:158-‐159). Lastly, the personalization principle and the voice principle have
30
both not yet been sufficiently confirmed through research, but they are theoretically very convincing
and they will be researched in the near future (2009:254-‐255). The personalization principle states
that learning is enhanced in multimedia when the verbal information is in ‘a conversational style
rather than formal style’ (2009:242). The voice principle is related to this principle, by assuming that
‘people learn more deeply when the words in a multimedia message are spoken by a friendly human
voice than by a machine voice’ (2009:255). Both principles are based on the assumption that
personalization and a human voice would unconsciously evoke a social response in learners. This
would lead to a higher motivation and could consequently encourage generative cognitive
processing (2009:248-‐249).
Subtitled television is in accordance with all these principles, they actually favour the use of subtitles.
Vanderplank (1988) even provides additional evidence for the personalisation principle. He points
out that television provides the viewers with more or less authentic language, and viewers also
perceive it as such (1988:278). This is consistent with what most television programmes and movies
are trying to achieve in their viewers: a feeling of reality. That includes real speech with various
accents and of course people, with a natural human voice. Tschirner (2001) provides us with a
different view on the personalisation principle that concentrates more specifically on the positive
effect of personalisation on language learning. He states that ‘Humans automatically assign feelings
to voices, faces, and interactions between people, and they perceive these feelings emotionally’
(2001:317). According to Tschirner this personalisation may strengthen the feeling of identification
with the culture of the speakers of the foreign language, which in turn could contribute to the
language acquisition of the learners (2001:317). In this view, personalisation could increase learner’s
motivation and thus positively influence the generative cognitive processing.
Secondly, we will discuss the principles that conflict with the use of subtitles. The coherence principle
states that the exclusion of extraneous material in the multimedia presentation will lead to better
learning (2009:89). This principle is threefold. It focuses on redundant material in the form of either
‘interesting but irrelevant words and pictures’, ‘interesting but irrelevant sound and music’ and
‘unneeded words and pictures’ (2009:89). There is only no clarity yet on irrelevant sound and music
because of conflicting studies, but on the whole learning improved when redundant information is
omitted. The redundancy principle builds on the coherency principle, but it is more specific and
states that non-‐verbal pictorial and auditory information will result in better learning than non-‐verbal
pictorial information and verbal information that is presented twice as both printed text and spoken
text (2009:118). The additional printed text is here seen as redundant information, it would lead to
extraneous cognitive load. People only have limited capacity to process information and especially
31
the visual channel could be overloaded with information because both the pictorial information and
the printed text are processed here (2009:123). However, this can not be compared to the split-‐
attention effect. In the split-‐attention effect the conflicting means of information are both necessary
for the learning process, they present different information (Yeung, Yin and Sweller, 1997:2-‐3).
Research by Mayer confirmed this principle. However, learners in his research were all native
speakers of the language that was spoken and written, and this was not tested on language learning
(Mayer, 2009:126-‐127). The modality principle resembles the redundancy principle in that it also
prefers spoken words above printed words, also when the two modes are only in combination with
pictorial information and not with each other and pictorial information (2009:200). This principle is
built on the dual channel hypothesis and the limited capacity assumption. Information that is
presented in pictorial form and as printed text will only enter the visual channel, leaving the auditory
channel unexploited. That could lead to an overload of information in the visual channel because of
the limited capacity for processing. Secondly, this manner of presentation could also lead to split
attention, since all information is presented in the same mode (2009:206-‐208). Research supported
this principle.
In conclusion, these three principles do not favour subtitled media and regard it as a cause of
cognitive overload because it would present the learner with redundant information. However, these
three principles have not been tested on language learning yet. Subtitling in SLA can rather be
viewed as supporting since it can be used as a comprehension strategy. Subtitles support the
auditory information. Nonetheless, according to the coherence principle television programmes and
films will always present the viewer with redundant information. Typical aspects of television are
corresponding music and the complex manner of telling. The first aspect is in this view simply
redundant, while the second aspect is not conform to the principle of coherence. It becomes clear
how certain aspects of multimedia that are seen as advantages in linguistics are seen differently in
educational sciences. The rich input environment that is praised several times in linguistics, also by
Koolstra and Beentjes (1999:54) and Tschirner (2001:306), is rather seen as a disadvantage in
educational sciences. According to these views, the fragment shown without subtitles would
influence language acquisition the most.
Lastly, the three principles that could possibly enhance the positive influence from watching
subtitled television when implemented will be discussed. When these are implemented though, it is
no longer possible to speak of incidental SLA, but only of formal learning. Two principles have
actually already been more or less tested in research by Borrás and Lafayette (1994) and Van Lommel
et al. (2006). The segmenting principle states that learning improves when the multimedia message
32
is presented ‘in user-‐paced segments rather than as a continuous unit’ (2009:175). In segmenting, a
complex learning unit is divided into smaller parts ‘that are presented sequentially with pacing under
the learner’s control’ (2009:176). Thus the ‘element interactivity’ (Van Merriënboer and Sweller,
2005:150) is reduced, and consequently the intrinsic cognitive load should decrease as well, which
would encourage deep learning. Research supports this finding, but also points to some restrictions
of this principle. Segmenting seems to show the strongest effects with complex learning material
that is presented in a rapid tempo, and with learners that are unfamiliar with the learning material
(2009:185). Research by Borrás and Lafayette (1994) on the influence of subtitles has shown that the
students’ comprehension and production of the foreign language improved when they could control
the pace of the subtitles. According to the second principle that concerns pre-‐training, learning
would be deeper when learners have already acquired some prior knowledge such as names and
characteristics on the subject that is taught (2009:189). This principle is specifically aimed at learners
who are not familiar with the learning material, and it consequently also shows the best results when
used with this group. Already having prior knowledge on the learning material decreases the amount
of processing, especially that from intrinsic cognitive load. Less effort needs to be made to
understand the material, which means that more ‘cognitive resources’ can be used to construct a
knowledge structure or can in other words be used for generative cognitive processing (2009:193).
Similarities can be observed here with Krashen’s input hypothesis, more specifically with his
assumption that comprehensible input approximates i + 1. Input can not be too difficult or of a too
high level for the learner to understand, in order for learning to still occur. The learning material is
adjusted to the learners’ prior knowledge or level, to make it comprehensible. Research by Van
Lommel et al. (2006) seems to confirm this principle. Van Lommel et al. researched the effect of
subtitles on grammar acquisition. For that purpose they also tested the use of grammar lessons
before viewing a subtitled fragment. They noted a significant improvement of grammar acquisition in
the groups that had received instruction of some grammatical rules prior to viewing the subtitled
fragment (2006:254). They even noted that the grammatical rules were acquired ‘at a level allowing
also their application on new cases’ (2006:254). That implies that the participants understood the
learning material and were able to transfer their newly constructed knowledge onto new situations
or cases (2009:19). The last principle, the signaling principle, states that learning is fostered when the
organization of the learning material is highlighted (2009:109). Signaling gives the learner some
guidance about which information to process and on how to organize it in a knowledge structure, it
encourages deep learning (1999:110-‐112).
In conclusion, subtitled television is conform with five of the principles that Mayer suggests in his
cognitive theory of multimedia learning and it differs with three other principles. According to these
33
last three principles (the coherence, redundancy and modality principle) television without subtitles
would support language learning the most, since subtitles are perceived to be redundant
information. However, Plass et al., including Mayer, state explicitly that vocabulary acquisition is
obtained when learners ‘can establish a direct connection between a word in their native language,
the corresponding picture of an object or action, and its foreign equivalent’ (Plass et al., 1998:26).
Thus, we can infer that the multimedia-‐design should differ in a language learning environment to
obtain optimal learning results, although this subject is not discussed deeply further on. De
Westelinck, Valcke, De Craene and Kirschner (2004) argue as well that multimedia design should be
adjusted to other study domains. In their research they found that the implementation of some of
the multimedia principles did not lead to the expected better learning results in the domain of
educational sciences. They argue that especially when pictorial information is concerned, multimedia
principles can not just be implemented. Learners do not always acquire sufficient prior knowledge of
the iconic system or pictorial representations used in a specific knowledge domain. The presentation
of that iconic system could consequently lead to an increased cognitive load for students instead of a
decreased cognitive load (2004:569). Thus, the cognitive theory of multimedia learning is not fully
applicable to SLA. However, it provides us with some interesting insights in how learning works. In
particular when we also regard the comment where Mayer and others (1998) state that in a
language learning environment, the learner benefits from an extra, elsewhere seen as redundant,
mode of information such as subtitles. Accordingly, much like in the linguistic view, learners would
benefit most from standard subtitled television, as the translation presented in this type enables
learners to establish more and deeper connections (1998:26). It could also be argued that the
translation leads to more understanding and consequently decreases cognitive load. We can infer
that the cognitive multimedia theory thus also favours standard subtitling in a language learning
environment, in contrast with what the principles in Mayer’s book about multimedia learning state
(2009), and which would seem to prefer no subtitles at all.
34
3 RESEARCH In this section, the design and course of the research will be commented on, as well as on the
participants and on how the results of the research were processed. But first, more explanation will
be given on the educational system in Flanders in order to have a better understanding of why
students from the first year of three different types of secondary education in Flanders were chosen
as participants, and of how these three different types of secondary education actually differ from
one another.
3.1 Educational System in Flanders
Belgium is a trilingual country; Dutch (Flemish), French and German are the official languages. As is
explained in ‘Education in Flanders’ (De Ro, 2008:6), every language group has its own community.
Those communities govern the organisation of education in their district (2008:6) and are very
independent as to how they organise their system of education. The research is conducted in
Flanders, which implies that the Flemish Community governs the educational system. Mainstream
education system in Flanders is organised into three major stages of education: Nursery and Primary
education, Secondary education, and Higher education (2008:25-‐34). The traditional trajectory for
children in Flanders is to start in nursery at the age of 2.5 years and to proceed to primary education
at age six. If primary education passes without any problems, children enter secondary education at
age twelve and receive a certificate for the previous stage. Primary education is the same for every
child; there are no extra levels or options. In secondary education, however, there are. This can also
be deducted from the image below.
First of all, there is a difference between the A-‐stream en the B-‐stream. The A-‐stream is oriented at
those children who have successfully passed Primary education and obtained the aforementioned
certificate. The first two years of education in the A-‐stream are very much the same for every
student, it is supposed to give the students a broad knowledge-‐basis. The children in B-‐stream have
either not yet obtained the certificate from Primary education, or they have chosen intentionally for
this type to prepare for the following vocational education. After the first year in the B-‐stream the
first group can either start in the first year of the A-‐stream or they can choose to stay in the same
type of education and prepare for vocational education. That brings us to the second differentiation
in mainstream secondary education. After the first two years, students are able to choose between
four different types of education: General secondary education or ASO, Technical secondary
education or TSO, secondary Arts education or KSO, and Vocational secondary education or BSO.
35
Students coming from the A-‐stream can choose from the four types, students coming from the B-‐
stream and that chose to continue in the B-‐stream, can only choose vocational education. Those four
types all offer different perspectives. ASO offers a ‘broad general education’ (2008: 31) and most
students from this type will proceed to Higher education. It is the most popular type of secondary
education, as it counts 39.9 % of the pupils in Flanders. TSO puts more emphasis ‘on general and
technical/theoretical subjects’ (2008:31). Graduated students from this type can already exercise an
occupation or they can proceed to higher education. This type includes 31.3 % of the students. KSO
offers, next to a general education, ‘active arts practice’ (2008:31). After passing this type, students
can either exercise an occupation or proceed to higher education. Only 1.9 % pupils study this type of
education. BSO offers ‘a practice-‐oriented type of education’ (2008:31). Students can exercise an
occupation after passing this type of education. BSO counts 26.9 % pupils. Although this paper only
focusses on first year students of secondary education, it is important that these different types are
explained. When children start their first year of secondary education, it is common for most of them
to have already chosen which type of education they will study in the following years. The type of
education that is followed depends greatly on which school is enrolled in in the first year of
secondary education, whether it is oriented at ASO-‐, BSO-‐ or TSO. Depending on which school is
chosen, the students will already have some introductory courses in their first year curriculum,
preparing them for one of the different types of secondary education. It is possible to change from
education type during the first two years, but it is rather uncommon. That is why we can actually
already speak of ASO, TSO and BSO from the first year of secondary education on. KSO is the only
type that is effectively chosen after the second A-‐stream year in secondary education.
The students of each type of education have a more or less clear profile that reflects their choice and
the motives behind that choice. Their choice in secondary education also determines to some extent
how these students will proceed after finishing secondary education. ASO-‐students are assumed to
proceed to higher education, while TSO-‐students and especially BSO-‐students are expected to either
study an extra year to perfect their training or to immediately start working in the sector of the
occupation that they have chosen to become proficient in during secondary education. This
difference in profile and in type of education also implies that students from different types need
different teaching approaches. Research by Charlotte Lippens (2010:127) already showed a
discrepancy in incidental acquisition in English vocabulary between students from ASO, TSO and BSO,
with the latter clearly showing a lesser degree of incidental acquisition. It would be interesting to
compare these three types of education, and see whether there is a substantial difference in
incidental grammar acquisition and in the amount of contact with the English language through
popular media, especially through subtitled television.
36
Image from Education in Flanders (De Ro, 2008:25)
Academic bachelor
Bachelor following bachelor
Master following master
Doctor (Universities only)
1ste graad
2de graad
3de graad deeltijds
4de graad
years
21
21
43 General Art Technical Vocational
General Art Technical Vocational65
765
765
765
7
Vocational
ELEM
ENTA
RY E
DU
CAT
ION
SEC
ON
DA
RY E
DU
CAT
ION
(2)
TER
TIA
RY E
DU
CAT
ION
Pre-vocational
BASP
ECIA
L ED
UC
ATIO
N
COLLEGES OF HIGHEREDUCATION ONLY
UNIVERSITIES AND COLLEGES OFHIGHER EDUCATION (ASSOCIATION)
654321
321
Professional bachelor
43
43
5
Master
3.1 Structure of education
43
PRIMARY EDUCATION
NURSERY EDUCATION
1st stage
2st stage
3st stage
4th stage
PART-TIME
years
years
37
3.2 Participants
For the research we needed to find students that were old enough to watch English-‐spoken
television with subtitles, but that had not yet been taught in English. Furthermore, I also wanted to
compare students from ASO, BSO and TSO. That is why it was decided to conduct the research with
first year students of secondary education from ASO, TSO and BSO-‐level. These are, normally,
children of about twelve to thirteen years old. BSO-‐students do not yet receive instruction in English
until after their first two years of secondary education2. Some students from TSO and ASO, however,
receive instruction in English from the first year of secondary education on, but there are also
schools that only start with English courses in the second year of secondary education. We searched
for participants in schools situated in or around the city Ghent. The participants from ASO are
students from the Don Bosco college Zwijnaarde in Ghent. The TSO-‐participants study at the Don
Bosco Sint-‐Denijs-‐Westrem institute in Ghent. Lastly, the BSO-‐participants are students from two
schools in the area surrounding Ghent, Visitatie Mariakerke and Atheneum Merelbeke.
In total 124 students participate in the test, from which seven students are omitted: five students
because they indicate to speak English at home, two because they could only participate in the first
part of the test and were absent during the second part, and one student because she has a minor
form of autism. That leaves us with 116 effective respondents. From those students, 46 are female
and 70 are male. In ASO male participants are more numerous than female ones, but only slightly,
whereas the participants from TSO are almost uniformly male, with only one girl in this group. BSO is
the only group where there are more female than male participants. The majority of the participants
come from ASO. We have 59 students from ASO, 31 from TSO and 26 from BSO. All participants
together cover nine class-‐groups, three from each education type. Nine different test-‐groups were
needed for the implementation of the research, and working with class-‐groups was the most
straightforward system for both the schools and me. All students are born between the years 1997
and 2000. The most common birth year in ASO is 1999, whereas the most common birth year in TSO
and BSO is 1998. In total 102 participants speak Dutch as their mother tongue, and 107 students
speak Dutch at home. It has to be mentioned, though, that a small number of students who indicated
Dutch as their mother tongue did not list Dutch as a language spoken at home. It can be assumed
that (most of) those students do speak Dutch at home, but that they forgot to list it as such. Only six
students did not list Dutch as their mother tongue or as a language spoken at home. In total, eleven
2 Vlaams Ministerie van Onderwijs en Vorming. “Overzicht van de eindtermen en ontwikkelingsdoelen van de eerste graad” Vlaams Ministerie van Onderwijs en Vorming. Web. 28 April 2012. <http://www.ond.vlaanderen.be/curriculum/secundair-‐onderwijs/eerste-‐graad/index.htm>
38
languages other than Dutch are spoken by some participants either at home or as a mother tongue:
French, Turkish, Bulgarian, Slovak, Spanish, Russian, Albanian, Portuguese, Arabic, Moroccan and an
Indian language. The highest percentage of students speaking Dutch as their mother tongue can be
found in ASO, 98.3 %. In BSO 73.1 % of the participants speak Dutch as a mother tongue, in TSO 77.4
% of the participants do so. The percentage for Dutch as a language spoken at home stays the same
in ASO, but it is higher in both BSO and TSO, respectively 80.8 % and 87.1 %. The entire research was
conducted in the period between 13/02/2012 and 12/03/2012. More specific information on the
participation will be given in 3.5.
3.3 Design of the research
For the design of the research, my bachelor paper (Becelaere, 2011) served as the starting point. The
same threefold structure as in that research was kept: first the survey, then a fragment of a film or
television programme and lastly a test that is chiefly based on the fragment. The three parts of the
research will be discussed separately below. Some changes with respect to my bachelor paper
research design were made, though. A retake of the test was added, two weeks after the first part of
the research had taken place. Thus, it was possible to see if the presumed learning impact of the
fragment was still effective in time. Changes were also made to both the survey and the test, those
will be explained further on. Another substantial change was that instead of working with only two
different subtitle-‐types and only two ASO class-‐groups, there were now three different subtitle-‐types
and class-‐groups from ASO, TSO and BSO. The spoken language of the fragment remained English.
Thus, the research is now implemented with nine class-‐groups, from which one of each education
type has seen either a fragment subtitled in English, subtitled in Dutch or without subtitles, three
class-‐groups for each subtitle-‐type. Lastly, it is also important to note that the whole research is
carried out anonymously. No names are asked from the participants, only their class-‐groups and
class-‐numbers are asked to be able to distinguish the students from one another. The research itself
is entirely conducted in Dutch. English is only used when it was necessary, for instance in the tests.
The main objective of the survey (Appendix 1) is to obtain some more information on the students’
background, such as their sex, date of birth, if they had ever been taught English and the like. In
contrast with the survey in my bachelor paper, there is no focus anymore on what or on which
channel the students watch when watching television. The questions are of a more general kind and
enquire for instance about their preferences regarding subtitles and the language spoken in
television programmes and films, which type they usually watch and so on. The focus is more on
practical information. But, as was mentioned before, research indicates that all types of popular
39
media influence implicit learning of English, so the participants’ habits concerning music, the internet
and gaming are asked about as well. To be complete, the students are also asked whether there are
any other ways through which they come into contact with English. Some options like family and
friends are provided, but there is also some free space to write down other, and perhaps less
obvious, possibilities. Lastly, the participants are also asked what they think about their own English
and if they frequently use English words and expressions in their everyday, mostly Dutch, language.
To have a more extended view on their usage of English, the participants are also asked which words
or expressions they use. Some examples of English words and expressions are given here, since the
participants mostly use these phrases and words almost without realising it and might thus not be
able to come up with most of them. For this list, my own use of English words and expressions when
speaking Dutch was looked into, and more attention was giving to the use of English in Dutch wen
listening to conversations, to advertising and the like.
After the students have filled in this survey, they are asked to look attentively at the fragment shown
to them. In each school, the fragment has been shown by means of the television set or screen with
a multimedia projector and video player available. As mentioned earlier, there are three classes from
each of the three education types participating in the research. Each class from each education type
has seen the same fragment but with a different subtitle-‐type: English-‐spoken with English subtitles,
English-‐spoken with Dutch subtitles and English-‐spoken with no subtitles. The fragment shown in the
research originates from the fourth Shrek movie, Shrek Forever After (2010) [Transcript can be found
in Appendix 2]. The first fifteen minutes of the film are shown, chapters one to three are fully shown,
and a part of chapter four. This movie is chosen because there is a lot of interaction in this film.
Secondly, since it is a movie designed for children, the language is not too difficult to understand, nor
is the storyline too complicated to follow. That last argument is also partially why the first fifteen
minutes of this particular movie are chosen. Everything that happened in the previous three movies
is narrated in short with strong visual support, as well as the beginning of the new story of this
particular movie. That way, and by not starting in the middle of the story, it was hoped that the
students could follow the story more easily and would not (yet) lose interest in the movie from the
very beginning on because they could not keep up with what was going on in the movie. It has never
been necessary to explain who Shrek is and what the original story is about. The only concept or
word that sometimes needed some explanation was ogre, supposedly because of the difference in
pronunciation in English and Dutch.
After that, the test followed [Appendix 3]. This is the part in which most changes are made in
comparison with the previous research. Last year, it was observed that the questions seemed to be
40
too easy for most students. That is why an attempt was made to make the questions more difficult
this time. But at the same time, the questions could not be made too difficult since now our group of
participants is more differentiated. Thus, the test is made more difficult, but only slightly. To make
the test more diverse and difficult, a second and more difficult part is added to the test. Secondly, as
in the previous research, examples from the fragment the students have seen are used in the test,
combined with sentences that did not occur in the fragment. However, this time those other
sentences are examples from actual language use. Sentences are used from the novels Sense and
Sensibility by Jane Austen3 and The Picture of Dorian Gray by Oscar Wilde4, from the news sites from
NY Times5 and BBC News6, and from Stephen Fry’s blog The New Adventures of Mr Stephen Fry7. In
addition, this mixture of new and familiar sentences also enables us to possibly observe actual
influence from the fragment. A last change in comparison with the previous test is that all twenty
questions are mingled. This time, the questions are not kept together per grammatical element that
is being researched. Thus there is less continuity in the test.
The design of the first part of the test in our research is for the most part based on the test from my
bachelor paper. It consists of twenty multiple-‐choice questions in which the students are given a
Dutch sentence and three English sentences as possible translations. They have to indicate which
English sentence they believe is the correct English translation from the Dutch sentence. There is also
a fourth option (d) ‘Geen idee’, which denotes ‘no idea’ or ‘I don’t know’. The participants are told to
mark this option only when they really have no idea which is the correct answer, not even a
preference for one of the other three options. Thirteen sentences are sentences that are utilised in
3 2 Austen, Jane. Sense and Sensibility. Ed. Stephen Arkin (Wordsworth). Hertfordshire: Wordsworth Editions Limited, 2007. Print. Question 4: 2007:24 / Chapter 7, p 24. Question 20: 2007:12 / Chapter 3, p 12. 4 3 Wilde, Oscar. The Picture of Dorian Gray. Ed. John M. L. Drew (Wordsworth). Hertfordshire: Wordsworth Editions Limited, 2001. Question 26: 2001:156 / Chapter 17, p 156. Question 27: 2001:40 / Chapter 4:40. 5 4 The New York Times Online. <http://www.nytimes.com/?adxnnl=1&adxnnlx=1335611004-‐ZxZQA8tZoQgF31ToKZp0vg>. Question 7: Lyal, Sarah. “417,6. Million Bites Later...” The New York Times. 09 Feb. 2012. p. 1. Web: 28 April 2012. <http://www.nytimes.com/2012/02/10/world/europe/charlie-‐bit-‐my-‐finger-‐video-‐lifts-‐family-‐to-‐fame.html?pagewanted=1&_r=1&ref=todayspaper>. Question 9: Anderson, Sam. “The World of Charles Dickens, Complete with Pizza Hut.” The New York Times. 07 Feb. 2012. p. 1. Web: 28 April 2012. <http://www.nytimes.com/2012/02/12/magazine/dickens-‐world.html>. Question 28: Tabuchi, Hiroki. “A Confused Nuclear Cleanup.” The New York Times. 10 Feb. 2012. p. 1. Web: 28 April 2012. <http://www.nytimes.com/2012/02/11/business/global/after-‐fukushima-‐disaster-‐a-‐confused-‐effort-‐at-‐cleanup.html?_r=1&hp=&adxnnl=1&adxnnlx=1328958157-‐3ppkkMBg7yGzGn4nKS9sxQ>. 6 5 BBC. <http://www.bbc.co.uk/worldservice/]. Question 11: “Learning English: ‘used to’ / ‘get used to’.” BBC. Web: 28 April 2012. <http://www.bbc.co.uk/worldservice/learningenglish/grammar/learnit/learnitv93.shtml>. Question 16: Moore, Alan. “Viewpoint: V for Vendetta and the rise of Anonymous.” BBC, 10 Feb. 2012. Web: 28 April 2012. <http://www.bbc.com/news/technology-‐16968689>. Question 19: “Barack Obama: ‘I deserve a Second Term.’” BBC. 6 Feb. 2012. Web: 28 April 2012. <http://www.bbc.co.uk/news/world-‐us-‐canada-‐16913174>. 7 6 The New Adventures of Mr Stephen Fry. <http://www.stephenfry.com/>. Question 29: Fry, Stephen. “A Modest Proposal” Samfry. 19 Dec. 2011, p. 1. Web: 28 April 2012. <http://www.stephenfry.com/2011/12/19/a-‐modest-‐proposal/>. Question 30: Fry, Stephen. “The BBC and the future of broadcasting.” Samfry. 18 June 2008. Web: 28 April 2012. <http://www.stephenfry.com/2008/06/18/the-‐bbc-‐and-‐the-‐future-‐of-‐broadcasting/>.
41
the fragment the students have seen, sentences number: 1, 2, 3, 5, 6, 8, 10, 12, 13, 14, 15, 17, 18.
The other seven sentences are examples from actual language use, sentences number: 4, 7, 9, 11, 16,
19, 20. The questions will be discussed in more detail in the following section, 2.3. The second part of
the test is a classic examining technique at school. The students are given an English sentence in a
mixed sentence order, and they have to try to form a correct English sentence from it. The proper
punctuation is added as well in the jumbled up sentences, not only to give an indication as to which
type of sentence is expected but also to avoid that the punctuation would be neglected or forgotten.
The first five sentences in this part are examples from the fragment: 21, 22, 23, 24, 25; the last five
are examples from actual language-‐use: 26, 27, 28, 29, 30.
3.4 Grammatical Foci
As was already mentioned in the introduction, the grammatical points of interest in this research are:
(1) Question words
(2) Adverbs ending in -‐ly
(3) Negation
(4) Word order in English sentences
The test is a mixture of questions concerning one of these grammatical aspects. Of course, these
elements are all interconnected with each other and are not independent units. Especially the word
order in English sentences can actually apply to every question in this test, but it is not always the
focus. These grammatical foci have been chosen because of their elementary role in language.
Especially the use of question words, of negation and of the right word order in English sentences are
basic aspects in the introduction to English during initial English instruction. Adverbs ending in -‐ly,
however, are of a somewhat more advanced level of grammar.
Questions 5, 7, 8, 15 and 27 focus on the use of question words. The question words used are where,
what, who and why. Only sentences 7 and 27 are examples of actual language use that are not used
in the fragment. In questions 1, 4, 9 and 17 the emphasis is placed on adverbs ending in -‐ly.
Questions 4 and 9 are examples of actual language use, questions 1 and 17 are taken from the
fragment. Question 1, ‘They lived happily ever after’ (option (c)), is actually quite remarkable. Besides
the fact that this is a fixed expression in English, this sentence also occurs at least five times during
the short fragment. That is why this sentence was chosen to be the first question of the test. It was
supposed to be an easy start that would give the students some confidence or a good feeling when
starting the test. However, there are only very few students who answered this question correctly.
42
Questions that focus on the usage of negation then, are questions 3, 10, 13, 16, 19, 21, 29 and 30.
The focus is more specifically on the place of not and never in the sentence and on the use of don’t.
Questions 3, 10, 13 and 21 are examples from the fragment, while questions 16, 19, 29 and 30 are
examples of actual language use.
Lastly, the questions that focus on word order are most numerous and include questions 2, 6, 11, 12,
14, 18, 20, 22, 23, 24, 25, 26 and 28. A special focus is also laid on the construction ‘to be used to’.
This expression occurs several times in the fragment, and thus it is expected that the participants
would observe how to use this expression correctly in the sentence. This expression occurs in
questions 2, 6 and 11. Lastly, in the second part of the test, the emphasis is to a large extent on the
correct word order in the sentence, not in the least because the participants have to form a correct
sentence themselves. In order not to make the test too difficult, the questions that focus on word
order are unmarked sentences, meaning simple sentences or independent clauses. There are no
sentences here that include inversion.
3.5 Procedure
As stated above, carrying out the research took almost a month. During that time, nine class-‐groups
participated twice in the research. Overall, every class-‐group participated very well and most were
even quite enthusiastic. The first part of the research lasted about forty-‐five minutes, more or less
corresponding to one lesson. In most of the class-‐groups that was sufficient time for everyone to
finish the first threefold part of the research. The second part of the research only took fifteen
minutes and took place during a lesson. All students had sufficient time to finish the test. After this
second part, there was an opportunity to ask some but not all class-‐groups whether they find the
second test to be more or less difficult than the first one. Further explanation will be given when the
more distinct observations are discussed. A teacher was almost always present in the classroom
while the research was carried out. But even without the presence of a teacher, the class-‐groups
participated very well. All groups were also very interested in their results on the two tests. That is
why afterwards the results of the test were passed on to the schools, so that they could tell the
students how they did. There were some slight differences between the participants with reference
to the three education types and to the types of video that were watched. For this reason, each
education type will be discussed separately.
43
3.5.1 BSO
The first class-‐groups to participate in the study were two BSO-‐class-‐groups from Visitatie
Mariakerke. One group saw the fragment with English subtitles, the other with Dutch subtitles. Both
groups cooperated very well and were also very attentive while watching the fragment. From what
could be observed, most students in both class-‐groups seemed to understand what was going on in
the fragment. This could also be observed from the fact that there was laughter at the right
moments. Of course, this could also be explained by the strong visual support present in the
fragment. These two class-‐groups found the second test as difficult or easy as the first one, there was
no difference to them. Nonetheless, one remarkable event occurred with the group that saw the
fragment with Dutch subtitles: they did not laugh at one particular moment in the fragment that is
supposed to be funny and where the English joke ‘Better out than in’ is made. The group that
watched the fragment with English subtitles did laugh with that joke, just like the group that watched
the fragment without subtitles. The same pattern was observed in the other groups from the other
two education-‐types. This moment, however, is designed to be funny, and the combination of the
visual and auditory support should be enough to understand the joke, like the other class-‐groups
have done. It was observed that the Dutch translation in the subtitles is very poor and does not
reflect the intended humour. On the contrary, the original expression and the consequential humour
of the moment were lost. The joke at that moment is about the three ogre-‐babies burping and
farting after having eaten, with the mother responding to that with: ‘Better out than in’. The Dutch
translation in the subtitles however is: ‘Geef de boer een stoel’ [Give the farmer a chair]. It is an old
Dutch saying which is not commonly used anymore, or at least not in Flanders. The meaning of the
joke was thus lost. It appears that the participants that watched the fragment with Dutch subtitles
rely more on the subtitles than on what is being said in English. That way, the joke could be easily
lost on the students. The third class-‐group from BSO was from Atheneum Merelbeke and watched
the fragment without subtitles. This class-‐group was least enthusiastic from all groups. They
participated very well and seemed concentrated while making the test after the fragment, but they
were not very interested in the fragment. After a few minutes, most members of the group lost
interest in the movie and started whispering to each other and scrabbling on their papers or hands.
Only three or four pupils watched the whole fragment intently. Possibly, this loss of interest can be
explained by the absence of the support from subtitles. This class-‐group also showed clearly that
they found the test fairly difficult. The three BSO-‐class-‐groups were the only groups where the
questions and possible English answers from the test were read out to the students.
44
3.5.2 TSO
The TSO-‐students all came from the same school, the Don Bosco Sint-‐Denijs-‐Westrem institute. The
majority of these students were male. All three groups were enthusiastic and seemed very
concentrated while making the test. The first group watched the fragment with English subtitles.
They laughed at the appropriate moments and seemed to understand what was going on in the
fragment quite well. This last impression was only strengthened by a remark of one of the students.
At one point in the fragment, Shrek really disappoints Fiona by implying that his life was better
before he rescued her from the dragon. After that one of the students commented by saying: ‘Game
over’, after which most other students started laughing. It shows that most of the students in this
group really understood what was going on at that particular moment. The second group watched
the fragment with Dutch subtitles. They were concentrated during the test and while watching the
video, but they did lose interest in the fragment in the end. Some of the students showed quite
clearly that they found the test difficult. In both class-‐groups, the majority of the participants found
making the test a second time easier than the first time. The third class-‐group watched the fragment
without subtitles. They seemed to understand what went on in the film, and laughed at the
appropriate times. At the end some students lost interest in the fragment. In contrast with the two
previous groups, this group found making the test a second time more difficult than the first time.
3.5.3 ASO
All ASO-‐students came from the same school, the Don Bosco college Zwijnaarde. All three groups
participated very well and were very attentive. The first group watched the fragment with English
subtitles. They seemed to understand quite well what went on in the fragment and laughed when
appropriate. After the first test, some of the students were interested in the intention behind the
research. After a short explanation, one of them stated very clearly that she did not believe that
watching English-‐spoken television programmes and film had any influence on her knowledge of the
language. She also claimed to find the test, and in particular the last part of it, very difficult. The
second class-‐group watched the fragment with Dutch subtitles. This group was a very quiet one. They
also did not laugh very much during the fragment. It was observed that these students seemed to
rely very much on the subtitles: the students in the back-‐row were consistently shifting to be able to
get a better view on the subtitles, although they could see what actually went on, on the screen very
well. The third group was a very enthusiastic class-‐group. They were also enthusiastic about seeing
the fragment completely in English, without any Dutch in it. They seemed to understand most of the
fragment and laughed at the right times. After the second part of the research, there was a
45
possibility to ask for their opinion about a possible difference in difficulty of the test between the
first and second time. The meanings were divided; some participants found it easier while others
found it more difficult the second time. There was no consensus.
3.6 Processing of Results
The results are processed in a statistical manner. The programme SPSS has been used for the
statistical computing. The statistical processing of the results consists of three steps. First, not all
questions, or more specifically not all the answers, could be introduced directly into SPPS and
needed to be edited. Secondly, all the results needed to be introduced into the programme but not
all results are valid. Finally, the results are processed in SPSS and could be analysed. This final step
will be discussed in the next section, 2. Results.
This first step mainly concerns questions from the survey. Answers to questions from the actual test
are simply introduced as right, wrong or invalid into SPSS. One part from the survey is adjusted a
little before entering the results on to the programme. It concerns questions seven to thirteen. The
purpose of these questions is to obtain some insight into the contact of the participants with English
through popular media or other possible ways. Specifically, the questions enquire about the
students’ use of music, gaming and internet, and if they use or come into contact with those media in
which language they mostly come into contact with them. The last question enquires about possible
ways to come in contact with English, other than media, television or films. Those questions are
taken together as an indicator of contact with English. This indicator has a four-‐point scale: 0= no, 1=
seldom, 2= sometimes, 3= often and 4= very much. Every point stands for either one of the three
media that are used in English or for another possible way of coming into contact with English.
In the second step more adjustments are made to the results, which asks for some more explanation.
It is obvious that questions that are answered unclearly, e.p. when two answers are given, they are
not taken into account in the processing of the results. Similar adjustments due to errors or
absentmindedness of the participants concern questions seven to thirteen, where the students’ use
of popular (English) media is looked into. The media specifically referred to are gaming, internet and
music. Lastly, question thirteen enquired whether there are any other ways the participants come
into contact with the English language. Some students, such as participant 68, filled in English music
or games here. Obviously, such answers are disregarded to avoid having double, invalid, results. The
same occurs with the questions that ask for the mother tongue, languages spoken at home and
possible other languages that are spoken. Some students, like participant fifteen, have listed their
46
mother tongue and his or her languages spoken at home again as an answer to the question about
other languages that the students might master. Furthermore, when students, such as participant 93,
answered one of the three questions of languages by saying they speak ‘a little French’ or a little of
any other language, that language was not taken into account. As a last remark, it should also be
mentioned that students who indicated to speak English at home or as their mother tongue, are
removed from the study.
The second part of the test also needs some clarification. The students have to form a sentence from
the jumbled up parts in the question. A first remark is that spelling errors are not taken into account,
solely correct word order counts. Inversion, on the other hand, is considered to be wrong. Inversion
is not a conventional use of language and mostly involves a marked meaning. The purpose of the
test, however, is to obtain conventional language use. For that reason, inverted sentences are not
assessed as being correct. There are only two exceptions here and those are questions 21 and 27.
Question 27 is a question taken from the fragment, it is highly informal but it is strongly emphasized
in the fragment. It has two valid answers: ‘Didn’t you share with the croissants?’ and ‘You didn’t
share with the croissants?’. This last option is not the most conventional option, but this is the option
used in the fragment. The inversion in the first option is naturally accepted, because in yes-‐/no-‐
question-‐sentences inversion is the more conventional word order. Question 27 is ‘Where did you
come across her?’. Here, inversion is required. Lastly, errors that are possibly due to inattentiveness
but could also be made intently are also considered wrong. Some examples of this are:
‘The dragon goes the bridge’ (participant 33)
‘He is very fascinating he’ (participant 6)
‘We are not talking about ideal society’ (participant 81)
It is not sure whether the participants had the correct answer in their mind when writing down the
answer or whether they assumed the answer given to be correct. Either way, these sentences are
assessed as being wrong.
47
4 RESULTS First, some general results will be presented that are relevant to our research questions, those topics
include the extent of contact with English and the amount of watching television, with a focus on
English spoken television programmes and films, and the use of subtitles. After that some contrastive
analyses will be made that should enable us to formulate an answer to our research questions. Not
all information from the survey is processed in the study, we focused on what is most relevant to our
research.
4.1 Overview results survey
As was already stated above, 116 students have participated in the study: 59 students from AS0, 26
from BSO and 31 from TSO. The participants are all between eleven and fourteen years old when the
tests are taken. The results in this section are all based upon what the participants have reported in
the survey. First, it is clear from the results that most participants are exposed to English to a great
extent. We could infer from the results that respectively 36 and 40 participants in the research come
into contact with English on a very regular basis, these are the students that respectively report a
high (‘often’) to very high (‘very much’) contact with English. They constitute 65.5 % of our
participants.
Figure 1: Table Crosstabulation Education type with Contact with English.
Educationtype * contactEnglish Crosstabulation Count
contactEnglish
Total no seldom sometimes often very much
Educationtype BSO 1 7 6 11 1 26
ASO 1 3 11 17 27 59
TSO 1 0 10 8 12 31
Total 3 10 27 36 40 116
A one-‐way ANOVA revealed that, concerning the contact with English, there is a highly significant
difference between the pupils in the different education types. A post-‐hoc Scheffe test shows that
the significant differences can only be found between participants from ASO and BSO (p ≤ 0.001) on
the one hand, and TSO and BSO (p ≤ 0.05) on the other. As can be perceived from the table above,
participants from BSO appear to be less exposed to English than participants from ASO and TSO. In
BSO there is a higher number of participants who indicate to have ‘no’ to ‘seldom’ contact with
English, than in ASO and TSO.
48
The general television viewing habits of all participants are in comparison very similar, the vast
majority regularly watches television programmes and films. However, when we look at the reported
viewing of specifically English spoken television programmes and films we find some slight
differences, but a one-‐way ANOVA showed no significant differences (p > 0.05). In comparison,
students from ASO and TSO report a higher frequency of viewing English spoken television
programmes than students from BSO. Overall, it is striking that respectively only two and six
participants report to not or only seldom watch English spoken television.
Figure 2: Crosstabulation Education type with frequency of watching English television programmes and movies. Educationtype * watchengprogmovies Crosstabulation
Count
watchengprogmovies
Total yes sometimes seldom no
Educationtype BSO 10 14 1 1 26
ASO 35 21 2 1 59
TSO 17 11 3 0 31
Total 62 46 6 2 116
When asked about whether the participants always view the English spoken programmes and movies
with Dutch subtitles, participants from TSO report the lowest use of Dutch subtitles of all three
education types, only half of the students always watches English spoken television with Dutch
subtitles. BSO students report the highest use of Dutch subtitles, and ASO students show a clearly
lower use of subtitles than BSO students but the majority always watches English spoken television
with Dutch subtitles. A one-‐way ANOVA revealed a significant difference concerning the use of Dutch
subtitles. The post-‐hoc Scheffe test shows that the only significant difference in use is found between
pupils from TSO and BSO (p ≤ 0.05).
Figure 3: Crosstabulation Education type with the frequency of using Dutch subtitles. Educationtype * Dutchsubtitles Crosstabulation
Count
Dutchsubtitles
Total yes not always no
Educationtype BSO 20 6 0 26
ASO 34 25 0 59
TSO 15 14 2 31
Total 69 45 2 116
49
Lastly, the participants are also asked to indicate whether they sometimes speak English. Participants
from ASO report the highest use of English, while participants from BSO report the lowest use.
However, when we look at all participants together, we see that a slightly higher number of
participants reports not to use English. There are also seven participants, two from BSO and five from
ASO, who did not answer this question correctly (= the missing value “99.00).
Figure 4: Crosstabulation Education type with the use of English Educationtype * sometimesspeakEnglish Crosstabulation
Count
sometimesspeakEnglish
Total yes no 99,00
Educationtype BSO 8 16 2 26
ASO 31 23 5 59
TSO 14 17 0 31
Total 53 56 7 116
4.2 Contrastive analysis results tests with fragment and education types
4.2.1 Contrastive analysis results in general
First we will correlate the results from the tests with the education type of the participants. When
we look at the results in general we have a total of 60 questions, every test consists of 30 questions.
Participants from BSO answered on average 30.39 out of 60 questions correctly, the minimum is
18.00 and the maximum is 48.00 (s.d.= 8.91). Participants from ASO answered on average 41.14
questions correctly, the minimum is 19.00 and the maximum is 56.00 (s.d.= 8.62). Participants from
TSO answered on average correctly on 36.97 questions, with a minimum of 16.00 and a maximum of
54.00 (s.d.= 9.70). The one-‐way ANOVA and post-‐hoc Scheffe test show a highly significant difference
between ASO and BSO (p ≤ 0.001), and between TSO and BSO (p ≤ 0.05). This is also clearly
discernible from the boxplot graph below. Similar results are found when we compare the results
from both tests separately.
50
Figure 5: Boxplot graph correlation of the results in total with the three education types.
Secondly, we will correlate the results from the test with the fragment types. With fragment types
we mean the three different types of fragments that are shown in the research itself. All fragment
types are spoken in English, but one includes Dutch subtitles (English-‐Dutch), one includes English
subtitles (English-‐English), and the last one has no subtitles (English). When we correlate the results
from the tests with the fragment types, we obtain a different outcome. First of all, the participants
are organized in a different manner. No distinction is made between the three education types, but
rather between the three subtitle types. The group that saw the fragment with Dutch subtitles
consists of 43 participants, the group that saw the fragment with English subtitles consists of 39
participants, and the last group that saw the fragment without subtitles consists of 34 participants.
The results from the two tests separately and taken together show similar results. That is why only
the results from the tests in general will be discussed. The results are more or less similar. The results
in general, 60 questions in total, show that the group that saw the fragment with English subtitles
answered on average 38.49 questions correctly, with a minimum of 18.00 and a maximum of 56.00
(s.d.= 10.40). The group that saw the fragment with Dutch subtitles obtained on average a result of
36.15, with a minimum result of 16.00 and a maximum result of 53.00 (s.d.= 10.12). The group that
saw the fragment without subtitles answered on average 38.18 out of 60 questions correctly, the
51
minimum is 19.00 and the maximum is 54.00 (s.d.= 8.97). The boxplot graph below also shows clearly
that the results from the participants from the three subtitle types only differ slightly. Indeed, a one-‐
way ANOVA pointed out that no significant differences are found between the three fragment types
(p > 0.05). There are also no significant differences between the results from the different fragment
types when correlated to the education types (p > 0.05).
Figure 6: Boxplot graph of the correlation results in total with the three fragment types.
When we compare the results from the first and the second test, we see that the results in the
second test are slightly higher. However, the difference is minimal as we can see when we correlate
the results to the education type. In the first test, students from ASO answered on average 20.12 out
of 30 questions correctly, with a minimum of 5.00 and a maximum of 28.00 (s.d.= 4.87). The mean
results in the second test are 21.02, with a minimum result of 12.00 and a maximum result of 29.00
(s.d.= 4.10). Students from BSO answered on average 14.73 questions correctly in the first test, the
minimum is 7.00 and the maximum is 24.00 (s.d.= 4.56). In the second test, they answered on
average correctly on 15.65 questions, with a minimum result of 9.00 and a maximum result of 26.00
(s.d.= 4.74). Pupils from TSO answered on average 18.19 out of 30 questions correctly in the first
52
test, the minimum is 8.00 and the maximum is 26.00 (s.d.= 4.81). In the second test, they answered
on average correctly on 18.77 questions, with a minimum result of 8.00 and a maximum result of
28.00 (s.d.= 5.28). These slightly higher results in the second test can also be perceived when we look
at the graph below that compares the results from the pupils from every education type in the first
and second test. However, these differences are only minimal.
Figure 7: Graph comparing the results on the first and second test from each education type.
4.2.2 Contrastive analysis results grammatical foci
We also focused on four grammatical aspects in this study: question words, adverbs ending in -‐ly,
negation, and sentence structure. The tasks in the test are not equally divided for every grammatical
aspect. There are five sentences concerned with question words, four with adverbs ending in -‐ly,
eight with negation, and thirteen with sentence structure. When we correlate the results for every
grammatical aspect that is studied in the research with the fragment types, we see that generally
there are no significant differences in the results from the three fragment types (p > 0.05). When we
correlate the results for every grammatical aspect that is studied in the research with the education
type, we do see significant mean differences between the three types (p ≤ 0.05). However, what is
recurrent in both the education types and in the fragment types are the results for the grammatical
aspects in general. Both question words and sentence structure show the best results, followed by
negation and adverbs. Especially the grammatical aspect adverbs ending in -‐ly has very low results in
general.
!"#$%&
'(#!'&!)#!*&
!+#,+&
'!#('&!)#$$&
(&
+&
!(&
!+&
'(&
'+&
-./& 0./& 1./&
!"#$%&#'
()$*+,-.'&/0"#'
1-20+34#-.'!"#$%&#'543#&'+.)'6"*-.)'7"#&'
1234&!& 1234&'&
53
Thus, as can also be perceived from the graph below, the results from participants from every
fragment type are quite similar. However, the results of the grammatical aspect adverbs endings in –
ly appear to show more differentiation, especially in the first test. Participants from the fragment
type with English subtitles answered on average 1.67 out of 4 questions correctly in the first test,
with a minimum result of 0.00 and a maximum result of 4.00 (s.d.= 0.97). Participants from the
fragment type with Dutch subtitles answered on average 0.92 questions correctly, with a minimum
of 0.00 and a maximum of 3.00 (s.d.= 0.84). Participants that saw the fragment without subtitles
answered on average 1.18 questions correctly, with a minimum result of 0.00 and a maximum result
of 4.00 (s.d.= 1.11). The one-‐way ANOVA indeed shows a significant difference here (p ≤ 0.01). The
post-‐hoc Scheffe test shows that the significant difference is found between the results from the
fragment type with English subtitles and the fragment type with Dutch subtitles (p ≤ 0.01). In the
second test, the results are again more similar. Participants from the fragment type with English
subtitles answered on average correctly on 1.21 out of 4 questions, with a minimum of 0.00 and a
maximum of 4.00 (s.d.= 1.13). Students from the fragment type with Dutch subtitles answered on
average 1.10 questions correctly, with a minimum result of 0.00 and a maximum result of 3.00 (s.d.=
0.94). Participants from the fragment type without subtitles answered on average correctly on 0.97
questions, the minimum is 0.00 and the maximum is 3.00 (s.d.= 0.90). The one-‐way ANOVA did not
show any significant differences in the second test (p > 0.05). This is conform with the results from
the other three grammatical aspects in this study that did not show any significant differences when
correlated to the fragment types, both in the first and second test.
54
Figure 8: Graph showing in percentages the results on each grammatical aspect correlated to the fragment types.
However, between the results from participants from every education type there are various
significant differences. Concerning the question words in the first test, participants from BSO
answered on average 3.00 questions out of 5 correctly, the minimum is 0.00 and the maximum is
5.00 (s.d.= 1.27). Participants from ASO answered on average correctly on 3.63 questions, with a
minimum of 0.00 and a maximum of 5.00 (s.d.= 1.05). Students from TSO answered on average 3.58
questions correctly in the first test, the minimum is 0.00 and the maximum is 5.00 (s.d.= 1.09). The
ANOVA test revealed a significant difference (p ≤ 0.05), but when we look at the post-‐hoc Scheffe
test we only see two near significant differences (p = 0.06) between ASO and BSO on the one hand,
and TSO and BSO on the other. The results were more similar in the second test. Pupils from BSO
answered on average correctly on 3.15 out of 5 questions, with a minimum of 1.00 and a maximum
of 5.00 (s.d.= 1.05). Participants from ASO answered on average 3.68 questions correctly, the
minimum is 0.00 and the maximum is 5.00 (s.d.= 1.04). Pupils from TSO answered on average
correctly on 3.45 questions, with a minimum result of 1.00 and a maximum result of 5.00 (s.d.= 1.09).
A one-‐way ANOVA showed no significant differences in the results from the second test (p > 0.05).
The results for adverbs ending in –ly are relatively similar to each other in both the first and second
test, a one-‐way ANOVA shows no significant differences (p > 0.05).
!"!#
!"!#
!"!#
!"!#
!"$#
!"$#
!"$#
!"$#!#
!#
!#
!#
%#
&%#
'%#
(%#
)%#
*%#
+%#
,%#
-%#
.%#
&%%#
/012345#64782# 98:17;2# <1=>345# ?15@15A1#?@70A@071#
!"#$%&#'()
"*+",&-."/'
0*-11-2+-%'3#4"+&#'
!"#$%&#'0*-11-2+-%'-#4"+&#'5'6*-.1",&'784"'
55
Concerning the questions on negation, participants from ASO answered on average 5.31 out of 8
questions correctly in the first test, the minimum is 0.00 and the maximum is 8.00 (s.d.= 1.64).
Students from BSO answered on average correctly on 3.12, with a minimum of 0.00 and a maximum
of 7.00 (s.d.= 1.73). Participants from TSO answered on average 4.07 questions correctly, the
minimum is 0.00 and the maximum is 7.00 (s.d.= 1.98). The post-‐hoc Scheffe test shows significant
differences between the results of participants from ASO and BSO (p ≤ 0.001), and from ASO and TSO
(p ≤ 0.01). In the second test, participants from BSO answered on average 3.46 out of 8 questions
correctly, with a minimum of 0.00 and a maximum of 7.00 (s.d.= 2.28). Students from ASO answered
on average correctly on 5.56 questions in the second test, the minimum is 1.00 and the maximum is
8.00 (s.d.= 1.51). Students from TSO answered on average correctly on 4.61 questions, with a
minimum result of 1.00 and a maximum result of 8.00 (s.d.= 2.11). The one-‐way ANOVA revealed a
significant difference between the results. The post-‐hoc Scheffe test shows that the only significant
difference is found between the results of participants from ASO and BSO (p ≤ 0.001). Concerning the
questions on sentence structure, pupils from BSO obtained on average a result of 7.62 on 13
questions in the first test, with a minimum of 3.00 and a maximum of 13.00 (s.d.= 2.45). Students
from ASO answered on average correctly on 9.73 questions, the minimum is 3.00 and the maximum
is 13.00 (s.d.= 2.46). Participants from TSO answered on average 9.39 questions correctly, with a
minimum of 4.00 and a maximum of 13.00 (s.d.= 2.49). The one-‐way ANOVA revealed significant
results (p ≤ 0.05). From the post-‐hoc Scheffe test, we could see that the differences are found
between the results of participants from BSO and ASO (p ≤ 0.01) and from TSO and BSO (p ≤ 0.05). In
the second test, participants from BSO answered on average correctly on 8.04 out of 13 questions,
the minimum is 4.00 and the maximum is 13.00 (s.d.= 2.24). Pupils from ASO answered on average
10.58 questions from the second test correctly, with a minimum of 5.00 and a maximum of 13.00
(s.d.= 1.83). Participants from TSO answered on average correctly on 9.71 questions, with a minimum
of 4.00 and a maximum of 13.00 (s.d.= 2.44). The post-‐hoc Scheffe test showed significant
differences between the results of participants from ASO and BSO (p ≤ 0.001), and from TSO and BSO
(p ≤ 0.05).
56
Figure 9: Graph showing in percentages the results on each grammatical aspect correlated to the education types.
4.2.3 Contrastive analysis results questions fragment and ‘new’ questions
In the test eighteen questions used sentences from the fragment, whereas twelve questions used
unfamiliar sentences. It is remarkable that much better results were achieved in the fragment
questions, than in the non-‐fragment questions. The results were correlated with both the fragment
and education types. In both the correlation with the fragment types and the education types, we
looked at the results in total, from both tests. That implies that the fragment questions now amount
to 36 questions, and that the non-‐fragment questions amount to 24 questions. When correlated to
the three fragment types, the mean results for the fragment questions vary between 23.41 and
24.58, whereas the mean results for the non-‐fragment question vary between 12.74 and 13.91. A
one-‐way ANOVA showed no significant differences here (p > 0.05). The difference in results between
non-‐fragment questions and fragment questions is, however, striking. The correlation of the results
with the education types again shows different results than the correlation with the fragment types.
When we look at the results on the fragment questions, we see that on a total of 36 questions
participants from BSO answered on average 19.00 questions correctly, with a minimum result of
11.00 and a maximum result of 31.00 (s.d.= 5.91). Pupils from TSO answered on average 24.55
questions correctly, the minimum result is 11.00 and the maximum result is 34.00 (s.d.= 6.54).
Participants from ASO answered on average correctly on 26.27 questions, with a minimum of 11.00
and a maximum of 35.00 (s.d.= 5.67). When we look at the post-‐hoc Scheffe test we find a significant
!"#$
!"#$
!"#$
!"#$
%"#$
%"#$
%"#$
%"#$
&"#$
&"#$
&"#$
&"#$
'$
('$
)'$
*'$
+'$
,'$
-'$
.'$
/'$
0'$
(''$
1234567$869:4$ %:;39<4$ =3>?567$ "37@37A3$"@92A@293$
!"#$%&#'()
"*+",&-."/'
0*-11-2+-%'3#4"+&#'
!"#$%&#'0*-11-2+-%'3#4"+&#'5'67$+-28,'9:4"'
57
difference between the results from ASO and BSO (p ≤ 0.001), and from TSO and BSO (p ≤ 0.01).
When we look at the results from the non-‐fragment questions we see that participants from ASO
answered on average 14.86 out of 24 questions correctly, with a minimum result of 8.00 and a
maximum result of 22.00 (s.d.= 3.58). Students from BSO answered on average correctly on 11.39
questions, the minimum is 5.00 and the maximum is 19.00 (s.d.= 3.34). Participants from TSO
answered on average 12.42 questions correctly, with a minimum of 5.00 and a maximum of 22.00
(s.d.= 3.68). The one-‐way ANOVA revealed a significant difference (p ≤ 0.05), but when we look at the
post-‐hoc Scheffe test we notice that the same significant differences from in the fragment questions
are found in the non-‐fragment questions between the different education. In the non-‐fragment
questions the significant differences are found between BSO and ASO (p ≤ 0.001), and between TSO
and ASO (p ≤ 0.01). And again, there is a noticeable difference between the results from the
fragment questions and the non-‐fragment questions. Because of the difference in the number of
questions, the results are converted to percentages so that we can have a better and more realistic
view on the differences. Pupils from BSO answered on average correctly on 52.77 % of the fragment
questions, and on 47.44 % of the non-‐fragment questions. Participants from ASO answered on
average 72.98 % of the fragment-‐questions correctly, and 61.93 % of the non-‐fragment questions.
Students from TSO answered on average correctly on 68.19 % of the fragment questions, and to
51.75 % of the non-‐fragment questions.
Figure 10: Graph showing the comparison of the results in total of both fragment questions and non-‐fragment questions,
in correlation to the three education types.
!"#$$%
$"#&"% '(#)&%
*$#**%
')#&+%!)#$!%
,%),%",%+,%*,%!,%',%$,%(,%&,%
),,%
-./% 0./% 1./%
!"#$%&#'()
"*+",&-."/'
01$+-23,'&45"#'
6375-*8#3,'9*-.7",&'-,1':3,;9*-.7",&'<$"#23,#'
23456789% :;8<23456789%
58
4.3 Contrastive analysis exposure to English and results tests
When we correlated the results from the test with the level of contact with English, we did not yet
make a distinction between the education types. The level of contact with English was correlated
with the results from both the first and second test. In Participants that reported to have very limited
(‘no’) contact with English answered 9.67 out of 30 questions correctly in the first test, with a
minimum of 5.00 and a maximum of 15.00 (s.d.= 5.03). Pupils that reported a high (‘often’) level of
contact with English answered on average 19.61 questions correctly, the minimum result is 11.00
and the maximum result is 28.00 (s.d.= 5.46). Students that reported a very high (‘very much’)
exposure to English answered on average correctly on 19.53 questions, with a minimum result of
8.00 and a maximum result of 27.00 (s.d.= 4.75). The one-‐way ANOVA revealed a significant
difference. When we looked at the post-‐hoc Scheffe test, we found significant differences between
the lowest and the two highest levels of contact with English (p ≤ 0.05). This difference in results can
also be perceived from the boxplot graph below.
Figure 11: Boxplot graph that correlates the results on the first test to the extent of contact with English.
59
The results from the second test are less straightforward. Participants that reported a very limited
exposure to English (‘no’) answered on average 12.33 questions correctly, the minimum is 10.00 and
the maximum is 14.00 (s.d.= 2.08). Participants that reported to be exposed to English only seldom,
answered on average correctly on 16.30 questions, the minimum result is 10.00 and the maximum
result is 23.00 (s.d.= 4.88). Students that only sometimes come into contact with English on average
answer 17.37 questions correctly, with a minimum result of 10.00 and a maximum result of 26.00
(s.d.= 4.53). Participants that reported a high (‘often’) exposure to English answered on average
correctly on 20.31 questions, with a minimum of 9.00 and a maximum of 28.00 (s.d.= 5.09). Students
that have a very high contact with English (‘very much’) answered on average 20.73 out of 30
questions correctly, with a minimum result of 8.00 and a maximum result of 29.00 (s.d.= 4.49). The
one-‐way ANOVA revealed a significant difference (p ≤ 0.01), but the post-‐hoc Scheffe test is not
conclusive and does not show where the significant differences are to be found. The boxplot graph
again shows the differences in results between the different levels of exposure.
Figure 12: Boxplot graph that correlates the results on the second test to the extent of contact with English.
60
The results of the tests separately and the level of contact with English were also correlated to the
education types. In the first test, participants from ASO that reported a limited exposure to English
(‘seldom’) answered on average correctly on 19.33 out of 30 questions, with a minimum result of
17.00 and a maximum result of 23.00 (s.d.= 3.22). Students that reported an average exposure to
English (‘sometimes’) answered on average 18.00 questions correctly, the minimum is 11.00 and the
maximum is 25.00 (s.d.= 4.07). Participants that reported a high exposure to English (‘often’)
answered on average 21.94 questions correctly, with a minimum result of 11.00 and a maximum
result of 28.00 (s.d.= 5.24). Students that reported a very high exposure to English (‘very much’)
answered on average 20.48 questions correctly, the minimum result is 9.00 and the maximum result
is 27.00 (s.d.= 3.98). A one-‐way ANOVA revealed a significant difference (p ≤ 0.01). No post-‐hoc
Scheffe tests are available though, because the group that reported to have very limited (‘no’)
contact with English is too small for a post-‐hoc Scheffe test. The results for each group in the second
test are similar to the ones in the first, and the one-‐way ANOVA test also revealed a significant
difference in the second test (p ≤ 0.01).
Figure 13: Boxplot graph with the correlation of the results of the first test to the extent of contact with English from participants from ASO.
61
4.4 Contrastive analysis viewing habits and results tests
The results from the question concerning which subtitles-‐types participants use the most, are
compared here to the results from the tests. These are correlated to the education and fragment
types. The one-‐way ANOVA revealed no significant differences when we correlated the results with
the education types (p > 0.05). As can be seen in the case processing summary of the first test below,
the pupils filled in many options. One participant did not answer the question correctly (missing
value “99.00”).
Figure 14: Summary of the cases that are processed in the correlation of the results from the first test to the types of
television programmes and films watched most by all participants.
Case Processing Summary
whichtypemost
Cases
Valid Missing
N Percent N
totaal1 Dutch 24 100,0% 0
English-Dutch 60 100,0% 0
English 2 100,0% 0
English-English 1 100,0% 0
other 2 100,0% 0
English-Dutch+other 2 100,0% 0
Dutch+English-Dutch 16 100,0% 0
Dutch+other 2 100,0% 0
Dutch+English-Dutch+other 3 100,0% 0
English+other 1 100,0% 0
Dutch+English+other 1 100,0% 0
English-Dutch+English-
English
1 100,0% 0
99,00 1 100,0% 0
The results from all participants that reported to watch mostly to the types Dutch, English-‐Dutch,
English, Other, and Dutch + English-‐Dutch will be stated here. These are the types that are most
relevant to our study and also include more than one participant. The results from the first and
second test are again similar, thus only the first test will be discussed here. Participants that mostly
watch Dutch television programmes have a mean result of 15.96, with a minimum result of 5.00 and
a maximum result of 27.00 (s.d.= 4.97). Participants that stated to mainly watch English-‐spoken
programmes with Dutch subtitles answered on average 19.42 out of 30 questions correctly, the
62
minimum is 8.00 and the maximum is 28.00 (s.d.= 4.97). Students that reported to watch English
programmes without subtitles the most answered on average correctly on 19.00 questions, with a
minimum result of 15.00 and a maximum result of 23.00 (s.d.= 5.66). Participants that stated to
mainly watch other types than the ones mentioned in the question obtained a mean result of 13.00,
with a minimum of 8.00 and a maximum of 18.00 (s.d.= 7.07). Participants that stated to mostly
watch a combination of Dutch-‐spoken television programmes and English-‐spoken television
programmes with Dutch subtitles achieved a mean result of 20.56, the minimum result is 11.00 and
the maximum result is 27.00 (s.d.= 4.75). The one-‐way ANOVA revealed a significant difference in
both the first and second test (p ≤ 0.05). However, no post hoc Scheffe tests are available because
there are too many options and some of them include only one participant, as can also be seen in the
case processing summary of the first test. The differences in results can also be seen on the boxplot
graph below.
Figure 15: Boxplot graph with the correlation of the results from the first test to the types of television programmes and
films watched most by all participants.
63
4.5 Contrastive analysis results tests and the language proficiency in English
The correlation between the results of the tests and the results of the question whether students
believe they are able to speak English is also very interesting. No distinction is made between
education and fragment types, all participants are taken together. First of all, the answers in the test
are to be interpreted as such: ‘very well’ means a good speaking ability of English, ‘yes’ means a
minor ability in English and ‘no’ means no ability in English. In total seven participants answered this
question incorrectly (missing value: “99.00”). In the first test, participants that reported a good ability
of English answered on average 24.83 out of 30 questions correctly, the minimum result is 22.00 and
the maximum result is 27.00 (s.d.= 2.23). Pupils that stated to be able to speak English answered on
average correctly on 19.49 questions, with a minimum result of 7.00 and a maximum result of 28.00
(s.d.= 5.09). Students that reported to not speak English answered on average 14.89 questions
correctly, the minimum result is 5.00 and the maximum result is 22.00 (s.d.= 3.86). The one-‐way
ANOVA showed significant differences (p ≤ 0.001). From the post-‐hoc Scheffe test, we can see that
these differences are found between the results from participants that stated to speak very well
English and those that stated not to speak English (p ≤ 0.001), and between the participants that
stated to speak English and those that stated not to speak English (p ≤ 0.001). The divergence in
results can also be perceived from the boxplot graph below.
64
Figure 16: Boxplot graph with the correlation of the results from the first test to the ability of speaking English of all
participants.
Similar results are found in the second test. Students that stated to be able to speak English very well
answered on average correctly on 26.83 questions, the minimum is 26.00 and the maximum is 28.00
(s.d.= 0.98). Participants that reported to be able to speak English (‘yes’) answered on average 20.52
questions correctly, with a minimum result of 8.00 and a maximum result of 29.00 (s.d.= 4.57).
Participants that stated not to speak English answered on average correctly on 15.46 questions, the
minimum result is 9.00 and the maximum result is 23.00 (s.d.= 3.73). The one-‐way ANOVA again
revealed a significant difference (p ≤ 0.001). The post-‐hoc Scheffe test showed that there are
significant differences between the group that reported to speak English very well and the group that
reported not to speak English (p ≤ 0.001), between the group that reported to speak English very well
and the group that reported to be able to speak English (p ≤ 0.01), and between the group that
reported to be able to speak English and the group that reported not to speak English (p ≤ 0,001).
65
Figure 17: Boxplot graph with the correlation of the results from the second test to the ability of speaking English of all
participants.
66
5 DISCUSSION First of all, the aim of this study was to answer our research questions. We will focus on indications
for implicit grammar acquisition first and then on the differences in results between the three
fragment types and the three education types. Afterwards we will discuss the subquestions and
focus on the influence of frequent contact with English on the results in this test, after which we will
shortly discuss whether there is an interrelation between the self-‐reported ability of speaking English
from the students and their results in the tests. Secondly, this study again confirms the immense
exposure to English outside the classroom that Flemish children experience, especially through
popular media. Many participants indicated to come into contact with English on a very regular basis
through the internet, gaming and music but also by watching television. We focussed in particular on
the medium of television, and it appears from our research that English spoken television
programmes or films are frequently used by the majority of our participants. This in in agreement
with what research by Kuppens (2010), Charlotte Lippens (2010) and Caroline Lippens (2010) already
indicated.
Our first research question is concerned with the question whether incidental grammar acquisition
can be observed from the results of the research. This question is threefold and will be determined
by comparing the overall results of the test, by comparing the results from the questions from which
the sentences are used in the fragment that is shown and from the questions that use sentences that
are new to the participants, and by comparing the results on each of the four grammatical aspects
that are focused on in the research: question words, adverbs ending in -‐ly, negation, and the correct
sentence order. We will already partially answer the third research question as well here, which
concentrates on whether there is a difference in results between the students from TSO, BSO and
ASO. We do believe that our research provides indications for grammar acquisition. We do have to
note that we also believe that our participants already achieved a certain level of language
acquisition, including grammar acquisition, which also shows in our study and makes it harder for us
to observe any direct influence of the seen fragment. But we do believe that our research suggests
direct influence of the fragment as well. First, when we look at the results in general from the test we
see that especially the participants from TSO and ASO achieved results above chance-‐level. Pupils
from BSO obtained significantly lower results. On average students from BSO just passed the tests.
Especially for students from ASO and TSO, these results suggest that grammar acquisition has taken
place, although we can not discern here if these results are due to the influence of the seen fragment
or if they are due to previously acquired grammatical knowledge. As for a possible difference
67
between the first and the second test, we observed that the results for the first and second test are
very much alike. The results for the second test were slightly higher, but the difference is very
minimal and we can not conclude anything from this. From these results we can also not conclude
with certitude whether the effect of watching the fragment is enduring or not, but it is possible.
When we look at the results of the different grammar aspects, we see that in general participants
achieve the best results on the grammatical aspects question words and sentence structure,
followed by negation and adverbs ending in -‐ly. The questions that focus on adverbs ending in -‐ly
have the lowest scores, participants from al three education types were not able to answer half of
these questions correctly. The scores on question words and sentence structure are however high,
students appear to master these two aspects of grammar already quite well. Again, students from
TSO and ASO achieve in general significantly better results than pupils from BSO. Only in the aspect
adverbs ending on -‐ly do the pupils of TSO achieve a result that is more similar to that from the
pupils from BSO. A remarkable occurrence in the aspect adverbs is the first question in the test
where the correct sentence is: ‘They lived happily ever after’. This is a widely spread sentence, that is
used up to five times during the seen fragment, sometimes even with emphasis. It was expected that
this question would be easy and thus it was chosen as the first question in the test to give the
students a good feeling and confidence when starting the test. Yet, there are very few students who
chose the right option here, the answer given most is the option ‘They lived happy ever after’.
Presumably, the word happy is heard more often than happily, which could be a possible explanation
why many students rather chose the option with happy, a word they recognise, instead of the
correct option with the less familiar word happily. This difference in results between the four
grammatical aspects could also point to a difference in sequence in the natural order of grammar
acquisition of these grammatical aspects, that is discussed by both Krashen (1987:12-‐13) and Ellis
(1994:20). Perhaps, adverbs ending in -‐ly, and to a lesser extent negation, are acquired in a later
sequence of the development of grammar than question words and sentence structure. But overall,
there are indications that all participants have already acquired some level in the English grammar,
especially concerning question words and sentence structure. Finally, we look at the last possible
indicator of grammar acquisition: the comparison between the results from sentences used in the
fragment and sentences that are not. We find some remarkable results here. First of all, we find that
participants from ASO and TSO again achieve significantly higher results than pupils from BSO.
Secondly, we find that the results from fragment sentences are remarkably higher than those from
the non-‐fragments sentences, especially in the results from pupils from ASO and TSO. We believe
that this indicates that there is a direct influence of the fragment that the participants have seen, and
that these results suggest incidental grammar acquisition.
68
When we correlate the results of the tests with the fragment types, we observe no significant
differences in the results between the participants from each fragment type, the results are in
general very similar. As we said before, we also believe there are indications that the participants
already acquired a certain level in English. We believe that it is possible that this previous knowledge
could explain why there are no significant differences in influence on language acquisition between
the three fragment types used in this study. Another (partial) explanation could possibly be the
familiarity to subtitles, a factor that is emphasised by both Danan (2004:73) and Vanderplank
(1988:273-‐274). It is obvious from the survey that the participants are already very familiar with
watching subtitled television, especially English-‐spoken television with Dutch subtitling. Danan states
that once viewers are familiar with standard subtitling, they can also easily adapt themselves to
intralingual subtitling, which is very well suited for more ‘advanced students’ (2004:75). But this
familiarity to subtitles does not explain why participants that have seen the fragment without
subtitles obtain similar results as the participants from the other fragment types.
From the results in general, the results on the four different grammar aspects and the difference in
the results between fragment and non-‐fragment questions, we do believe that implicit grammar
acquisition of English can be observed here. This would contradict Krashen (1987:63), who stated not
to believe that people learn from video’s, and also Van Lommel et al. (2006) who stated that
vocabulary needs to be acquired first (2006:255). This indications of implicit grammar acquisition and
the fact that no difference was found between the effectiveness from the different fragment types,
also contradict Mayer’s CTML, and more specifically the redundancy (2009:118) and modality
principle (2009:200). In addition, the translation could possibly also enhance making connections
between the different mental representations in the learning process and consequently encourage
‘generative cognitive processing’ (Mayer, 2009:81) and thus learning, as Mayer has already
suggested himself (Plass et al., 1998:26). From this perspective, CTML can still be applied to language
learning with subtitled television if some adjustments are made. It would appear that implementing
subtitles, both standard and intralingual subtitles, in a video spoken in a foreign language improves
the comprehensibility and enhances language learning, even when there is no explicit intention to
learn the foreign language. Danan (2004:71) could thus be correct when she suggests that subtitled
television might provide the viewer with comprehensible input.
Concerning the three education types, significant differences were mostly found between the
education types ASO and BSO on the one hand, and TSO and BSO on the other hand. In most cases,
the pupils from BSO achieved significantly lower results than the students from ASO and TSO. From
the results of the survey, we see that pupils from BSO reported a lower contact with English than
69
students from ASO and TSO, but the results for watching English spoken television were fairly similar.
It seems that participants from BSO are slightly less exposed to English through other ways than the
medium television. During the research, it was already clear that students from ASO were most
interested during the fragment, while the pupils from BSO and TSO lost their interest in the fragment
most easily. Yet, pupils from TSO achieved higher results that are more similar to the results from the
participants from ASO, than pupils from BSO. We do not have a solid explanation for this
discrepancy, the only difference between the three education types is that students from BSO have a
slightly lower contact with English. If this were the cause for the difference in results, it would seem
that the extent of contact with the foreign language is highly important and that already a small
difference can influence the extent of acquisition greatly. However, additional support such as the
implementation of pre-‐training and signaling, according to Mayer’s multimedia principles (2009),
could perhaps enhance language acquisition here. These results also confirm Charlotte Lippens’
(2010:127) study that observed the same discrepancy.
In our first subquestion the aim was to find out whether there is a correlation between the
participants’ contact with English, especially through subtitled television, and their results on the
test, whether the extent of exposure to English correlates with the extent of language acquisition.
We found significant results here. First, we will discuss the correlation between the extent of contact
with English and the results. We saw that in general pupils who reported a high to very high exposure
to English subsequently also achieved higher results than those pupils who reported a lower contact
with English, with the differences even being significant between pupils who reported the highest
exposure and those who reported the lowest exposure. This is consistent with Kuppens’ (2010:74-‐76)
study that also only showed significant differences between those participants who reported the
highest contact with English through popular media and those participants who reported the lowest
contact with English through popular media. The results of the participants who stated a high
exposure and those who stated a very high exposure to English are similar. The results indicate that
only a high to very high exposure to English leads to more extensive language acquisition.
The correlation between contact with English and the level of the achieved results is clear, but now
we will focus on the possible correlation between the results on the test and which subtitle type the
participants reported to watch the most. We already saw that the viewing habits concerning
watching English-‐spoken television programmes and films are quite similar across the participants
from the three education types. In addition, the majority stated to always watch English-‐spoken
television with Dutch subtitles, with the pupils from BSO reporting the highest use. However, it
appears that watching English spoken television without subtitles also occurs quite frequently, with
70
the exception of students from BSO who favour Dutch subtitles. This is quite surprising, since
watching English spoken television without Dutch subtitles already requires a certain knowledge of
English, knowledge that is not acquired through instruction. The results from the question as to
which subtitle type is watched most are less straightforward. There is a significant difference in the
results but the post-‐hoc Scheffe test is inconclusive. We do, however, note that there is a difference
in results between in particular participants who mostly watch Dutch television programmes, and
students who mostly watch either English-‐spoken programmes with Dutch subtitles or a combination
of Dutch-‐spoken television programmes and English-‐spoken programmes with Dutch subtitles. The
mean results of the last two groups are clearly higher than the results of the first group. We do not
known, however, if these are significant results. But these results do suggest that a higher exposure
to English leads to higher results, this is in accord to the correlation between the extent of contact
with English and the level of results. This is also in agreement with the results in Kuppens’ study
(2010), the more a participant reported to watch subtitled television, the higher the results were
(2010:74-‐76).
Finally, our second sub-‐question was concerned with the possible correlation between the students’
reported ability of speaking English and their results on the test. In the results we saw that students
who reported to be able to speak English very well had higher results than those who stated to just
be able to speak English, and than those who stated not to be able to speak English. Consequently,
those that reported not to speak English had the lowest scores, and answered barely half of the
questions correctly. In both tests, these differences between the participants that stated not to be
able to speak English and both the participants that stated to be able to speak English and to speak it
very well were significant ones. This indicates not only that there is a correlation between the
students’ ability to speak English and their results on the test, but most importantly, it indicates that
the participants are well able to assess their ability of English.
71
6 CONCLUSION The main objective of our research was to find indications of implicit English grammar acquisition
through watching subtitled television in children that have not yet received instruction in English.
Which type of subtitling was the most effective was also researched, the participants were shown a
fragment with either Dutch subtitles, English subtitles or without subtitles. In addition, we also
wanted to see if there are differences in acquisition between students from the three main
secondary education types in Flanders; TSO, ASO and BSO. 116 participants in their first year of
secondary education in ASO, BSO or TSO participated in the research. We do believe to have found
indications of implicit grammar acquisition through viewing subtitled television programmes or films
in our research, both from the one fragment shown to the participants but also from the English-‐
spoken media the participants have already been exposed to prior to this research. We did not find
any significant differences in influence between the three types shown to the participants. We
believe that this could be explained by the prior knowledge of English most students have already
acquired trough the exposure to English-‐spoken media. In addition, all students are already familiar
with the medium subtitled television, they have already developed adequate viewing techniques.
Secondly, we have also confirmed that there is a discrepancy between the three different education
types discussed here, in general participants from ASO achieved the best results while participants
from BSO achieved the lowest results. We do not have a clear-‐cut explanation for this, although it
appears from the survey that these students are a little less exposed to English than both
participants from ASO and TSO. We believe that this could imply that language learning needs to be
adjusted to the participants of each secondary education type to enhance and encourage learning.
Thirdly, the research showed that there is a correlation between the extent of contact with English
and the extent of language acquisition, the higher the extent of exposure the higher the extent of
language acquisition. Lastly, the research also indicates that the pupils are very well aware of their
capabilities in English, of their language proficiency.
Although this research was successful, we do believe that improvements to the study can be made,
four changes in particular. First of all, including a larger number of participants could provide us with
more conclusive results. Secondly, the research could also be improved by adding a pre-‐test as well.
This way we would be able to have a view on the prior knowledge of the participants and to compare
it to the results after viewing the fragment. This would enable us to make a better assessment of the
influence of the seen fragment. Thirdly, more conclusive results could also be obtained by
conducting the study over a longer period of time, by adding more fragments followed by tests and
thus creating more opportunities for research, to have a longitudinal research. Finally, it would also
72
be very interesting to add a tested questionnaire to the research. That way the results would not
only be more objective, but it is also a good way to also enquire about the participants’ motivation,
whether they identify with the speakers of the foreign language, to their attitudes towards language
learning and the foreign language, and so on.
In addition, we also believe that more research is needed, both in the domain of implicit grammar
acquisition as in explicit language learning through media such as subtitled television. First of all,
more research into incidental grammar acquisition is necessary to confirm with more certitude that
in addition to implicitly acquiring vocabulary, grammar is acquired implicitly as well through watching
subtitled television. It would also be interesting to see whether this incidental language acquisition is
already present in younger children as well. Lastly, we also believe that more extensive research on
the use of the cognitive theory of multimedia learning (Mayer, 2009) in language learning would lead
to valuable insights. We believe in the value of CTML, but the theory needs some adjustments when
applied to language learning. We believe that with more research subtitled television could be
implemented successfully in language learning at schools. When adapted to the learners and when it
is used intently for language learning, watching subtitled television could be an effective tool in
language learning. Although we do have to keep in mind that the participants in this study have
received an immense input of the foreign language, English here, through a variety of media. That is
probably an important factor as to why students graduating from secondary school are more
proficient in English than French, students come into contact with English on a much larger scale
than with French, and especially outside the classroom as well.
73
7 REFERENCES -‐ Becelaere, Elien. Accidental Second Language Grammar Acquisition of English through Watching
Subtitles Television Programmes or Films. Unpublished dissertation: Universiteit Gent, 2011.
Print.
-‐ Bird, Stephen A., and John N. Williams. “The effect of bimodal input on implicit and explicit
memory: An investigation into the benefits of within-‐language subtitling.” Applied
Psycholinguistics 23.4 (2002): 509-‐533. Print.
-‐ Borrás, Isabel, and Robert C. Lafayette. “Effects of Multimedia Courseware Subtitling on the
Speaking Performance of College Students of French.” The Modern Language Journal 78.1
(1994): 61-‐75. Print.
-‐ Concise Oxford English Dictionary. 11th ed. 2008. Print.
-‐ Danan, Martine. “Captioning and Subtitling: Undervalued Language Learning Strategies” Meta:
Journal des traducteurs / Meta: Translators’ Journal 49.1 (2004): 67-‐77. Print.
-‐ De Ro, Jo (ed.). Education in Flanders. A broad view of the Flemish educational landscape. Haletra:
Houthalen-‐Helchteren, 2008. Print.
-‐ De Westelinck, Katrien, Martin Valcke, Brigitte De Craene, and Paul Kirschner. “Multimedia learning
in social sciences: limitations of external graphical representations” Computers in Human
Behavior 21 (2005): 555-‐573. Print.
-‐ d’Ydewalle, Géry, and Marijke Van de Poel. “Incidental Foreign-‐Language Acquisition by Children
Watching Subtitled Television Programs” Journal of Psycholinguistic Research 28.3 (1999): 227-‐
244. Print.
-‐ d’Ydewalle, Géry, and Wim De Bruycker “Eye Movements of Children and Adults While Reading
Television Subtitles” European Psychologist 12.3 (2007): 196-‐205. Print.
-‐ Ellis, Rod. The Study of Second Language Acquisition. Oxford: Oxford University Press, 1994. Print.
-‐ Gardner, R. C., Paul F. Tremblay, and Anne-‐Marie Masgoret. “Towards a Full Model of Second
Language Learning: An Empirical Investigation” The Modern Language Journal 81.3 (1997): 344-‐
362. Print.
-‐ Greenfield, Patricia M. “Technology and Informal Education: What Is Taught, What Is Learned.”
Science 323 (2009): 69-‐71. Print.
-‐ Koolstra, Cees M., and Johannes W.J. Beentjes. “Children’s Vocabulary Acquisition in a Foreign
Language Through Watching Subtitled Television Programs at Home” Educational Technology
Research and Development 47.1 (1999): 51-‐60. Print.
-‐ Koolstra, Cees M., Allerd L. Peeters and Herman Spinhof. “The Pros and Cons of Dubbing and
Subtitling” European Journal of Communication 17.3 (2002): 325-‐354. Print.
74
-‐ Krashen, Stephen D. Principles and Practice in Second Language Acquisition. Hertfordshire:
Prentice-‐Hall International (UK) Ltd., 1987. Print.
-‐ Krashen, Stephen. “Comprehensible output?” System 26 (1998): 175-‐182. Print.
-‐ Kuppens, An H. “Incidental Foreign Language Acquisition from Media Exposure.” Learning, Media
and Technology 35.1 (2010): 65-‐85. Print.
-‐ Lippens, Caroline. The incidental acquisition of English before the input of formal instruction: a
contrastive analysis of TSO, B-‐stream & ASO in Flanders. Unpublished dissertation (masterproef):
Universiteit Gent, 2010. Print.
-‐ Lippens, Charlotte. The English productive vocabulary of secondary school children in Flanders prior
to English instruction. Unpublished dissertation (masterproef): Universiteit Gent, 2010. Print.
-‐ Macaro, Ernesto, and Liz Masterman. “Does intensive explicit grammar instruction make all the
difference?” Language Teaching Research 10.3 (2006): 297-‐327. Print.
-‐ Masgoret, A.-‐M., and R. C. Gardner. “Attitudes, Motivation, and Second Language Learning: A
Meta-‐Analysis of Studies Conducted by Gardner and Associates” Language Learning 53.1 (2003):
123-‐163. Print.
-‐ Mayer, Richard E. Multimedia Learning. 2009. 2nd ed. New York: Cambridge University Press, 2010.
Print.
-‐ Moreno, Roxana, and Richard E. Mayer. “Verbal Redundancy in Multimedia Learning: When
Reading Helps Listening” Journal of Educational Psychology 94.1 (2002): 156-‐163. Print.
-‐ Plass, Jan L., Detlev Leutner, Dorothy M. Chun, and Richard E. Mayer. “Supporting Visual and Verbal
Learning Preferences in a Second-‐Language Multimedia Learning Environment” Journal of
Educational Psychology 90.1 (1998): 25-‐36. Print.
-‐ Plass, Jan L., Dorothy M. Chun, Richard E. Mayer, and Detlev Leutner. “Cognitive Load in reading a
foreign language text with multimedia aids and the influence of verbal and spatial abilities”
Computers in Human Behavior 19 (2003): 221-‐243. Print.
-‐ Sjöholm, Kaj. “The complexity of the learning and teaching of EFL among Swedish-‐minority students
in bilingual Finland” Journal of Curriculum Studies 36.6 (2004): 685-‐696. Print.
-‐ Shrek Forever After. The Final Chapter. Dir. Mike Mitchell. Dreamworks Animation SKG. 2010. DVD.
-‐ Tschirner, Erwin. “Language Acquisition in the Classroom: The Role of Digital Video” Computer
Assisted Language Learning 14.3-‐4 (2001): 305-‐319. Print.
-‐ Valcke, Martin. Onderwijskunde als ontwerpwetenshap. Een inleiding vor ontwikkelaars van
instructie en voor toekomstige leerkrachten. Gent: Academia Press, 2010. Print.
-‐ Vanderplank, Robert. “The value of teletext sub-‐titles in language learning” ELT Journal 42.4 (1988):
272-‐281. Print.
75
-‐ Van Lommel, Sven, Annouschka Laenen, and Géry d’Ydewalle. “Foreign-‐grammar acquisition while
watching subtitled television programmes” British Journal of Educational Psychology 76 (2006):
243-‐258. Print.
-‐ Van Merriënboer, Jeroen J. G., and John Sweller. “Cognitive Load Theory and Complex Learning:
Recent Developments and Future Directions” Educational Psychology Review 17.2 (2005): 147-‐
177. Print.
-‐ Yeung, Alexander Seeshing, Putai Jin, and John Sweller. “Cognitive Load and Learner Expertise:
Split-‐Attention and Redundancy Effects in Reading with Explanatory Notes” Contemporary
Educational Pscychology 23 (1997): 1-‐21. Print.
76
8 APPENDIXES
8.1 Appendix 1 – The Survey
!"#$%&'
‣!"#$%&!'''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''‣!"#$%()**+,&!'''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''
‣!-+.//,0+1$0)*&!''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''
‣!2/+1+,0$$#30$#+(&!'''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''
‣!4$$#34$#+(!5+%6,/7+(!08)9%&!'''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''‣!:(1+,+!0$#+(!19+!;+!%6,++70&!'''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''‣!<=+,3>(!?+#7+!0$#+(!8+.!;+!$#!#+%!5+8$1@!''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''‣!A+(!;+!//90!$#!($$,!++(!0$#+(7$*6!/B!1+,5+#9;7+!5+?++%0@!'''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''
‣!C/!;$D!?+#7+!0$$#30$#+(!8+.!;+!1$$,!5+#++,1@!'''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''
("%)$**&$+*,$-*.,'
E' -+%#$F80&!
!!2$(
!!G,/)?
H' "9;7!;+!=$$7!($$,!0+#+=9%9+!+(!I#*%@!
!!J$
!!K++(
L' :#%!;+!($$,!0+#+=9%9+!/B!($$,!I#*%!79;70D!79;7!;+!1$(!//7!($$,!M(5+#%0$#95+!6,/5,$**$N%!/B!I#*%@!
!!J$
!!:B!+(!0/+
!!C+#1+(
!!K++(
O' :#%!;+!($$,!M(5+#%0$#95+!6,/5,$**$N%!/B!I#*%!79;70D!79;7!;+!1$(!*+0!K+1+,#$(1%0$#95+!
/(1+,P0+#9(5@!!
!!J$
!!K9+0!$#P;1
!!K++(
Q'R$$,!79;7!;+!8+0!*++%0+!($$,&!
!!K+1+,#$(1%0$#95+!6,/5,$**$N%!+(!I#*%
!!M(5+#%0$#95+!6,/5,$**$N%!+(!I#*%!*+0!K+1+,#$(1%0$#95+!/(1+,P0+#9(5
!!M(5+#%0$#95+!6,/5,$**$N%!+(!I#*%!S/(1+,!/(1+,P0+#9(5
!!M(5+#%0$#95+!6,/5,$**$N%!+(!I#*%!*+0!M(5+#%0$#95+!/(1+,P0+#9(5
!!:(1+,+&!''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''
!"#$%&'%()*&&+% ,"-$
1
77
!" #$$%&'()*&+,-(&*./*&/(&0(+&).(12+3
&&#(4(%)$542+$).6&6(2-%7*(5
&&856()2+$).6&6(2-%7*(5&9(+&#(4(%)$542+$).6(&754(%:+().56
&&856()2+$).6&6(2-%7*(5&9(+&856()2+$).6(&754(%:+().56
&&856()2+$).6&6(2-%7*(5&;754(%&754(%:+().56
&&<54(%(=&""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""
>" [email protected]+(%&/(&A$$*&5$$%&9@;.(*3&
&&B$
&&<1&(5&+7(
&&#((5
C" #$$%&'()*(&9@;.(*&)@.2+(%&/(&0(+&9((2+=
&&#(4(%)$542+$).6(&9@;.(*
&&D%$52+$).6(&9@;.(*
&&856()2+$).6(&9@;.(*
&&<54(%(=&""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""
E" F@%1&/(&A$$*&7-&0(+&.5+(%5(+3&
&&B$
&&#((5
GH"&I79&/(&4$$%&A$$*&.5&J75+$J+&9(+&$54(%(&+$)(5&4$5&0(+&#(4(%)$5423
&&K7&/$L&'()*(3&"""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""
&&#((5
GG"&F-(()&/(&A$$*&2-())(+/(2&7-&4(&M..L&N)$,&F+$:75L&$54(%(&J7527)(2&71&7-&0(+&.5+(%5(+3
&&B$
&&<1&(5&+7(
&&#((5
GO"&P5&'()*(&+$$)&2-(()&/(&4$53
&&#(4(%)$542
&&856()2
&&D%$52
&&<54(%(=&"""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""
GQ"&I79&/(&576&7-&$54(%(&9$5.(%(5&.5&J75+$J+&9(+&0(+&856()23
&&R.$&856()2+$).6(&A%.(54(5
&&S-&%(.2
&&R.$&/(&1$9.).(
&&<54(%(=&"""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""
&&#((5
!"#$%&'%()*&&+% ,"-$
2
78
!"#$%&'($)*$+,'$)*-*./$(,0$)*$,.$1'2*.3$4,'$+*530,,'6
$$7,8$-**5$29*($-*./3
$$7,
$$:**'
!;#$%&'($)*$+,'$)*-*./$(,0$)*$,.$1'2*.3$4,'$3<5*4*'6
$$7,8$-**5$29*($-*./3
$$7,
$$:**'
!=#$><5**4$)*$-*./$,/$*'$09*$1'2*.36
$$7,
$$:**'
!?#$@*A5B&4$)*$&'$)*$(,2(,2*.&)43$0,,.2*A5B&4$C*0$+5&*'(*'$9/$/,C&.&*$1'2*.3*$D995(*'$9/$-&''*0)*3$
-9,.3$!"#$%#&'(!")'#*+%!"),-.!#/.+01#)2341#5,&"61#7'#8*"+691#:..#%';#0+6.<1#$,::#%';1#-",001#(;(.!"###6
$$%,,4
$$E/$*'$09*
$$:**'
!F#$G5B&3$,,'$D*.4*$D995(*'$)*$-*./$5*2*.C,H2$2*A5B&40$&'$)*$(,2(,2*.&)43$0,,.2*A5B&4#
$$IJ$CK$29(
$$L$.9+*$K9B
$$:9$D,K
$$:&M*
$$ED*39C*
$$N*.,O
$$P*,.
$$:QRS$T:*+*5$D,'',$.93*$K9BU
$$VWW$TV*30$W5&*'(3$W95*+*5U
$$QJ9$M,5*3
$$N&2J0
$$>C&.*
$$@99($.BM4
$$1')9K
$$>9$TDJ,0U
$$XB0*
$$>**$K9B$.,0*5
$$Y&33$K9B
$$XJ&..
$$PB(*
$$Z,<<K
$$L$4'9D
$$E'(*5*[$#############################################
$%&'"()"*+,((-" .%/'
3
79
8.2 Appendix 2 – Transcript of Fragment Shrek
K: King
Q: Queen
R: Rumpelstiltskin
P: Pinocchio
S: Shrek
F: Fiona
G: Guide
D: Donkey
P: Puss
BP: Butterpants kid
Father: father BP
GM: Gingerbread Man
MM: Muffin Man
MW: Man Woman
BBW: Big Bad Wolf
Rumpelstiltskin: Once upon a time, a long time ago, a king and a queen had a beautiful daughter,
named Fiona. But, she was possessed by a terrible curse: by day a lovely princess, by night a hideous
ogre. Only true love’s kiss could lift her curse. So Fiona waited in a tower, guarded by a dragon, until
the day when her true love would arrive. But as the days turned into years, the king and queen were
forced to resort to more desperate measures.
K: Eew.
Driver: Whoa, there.
K: I don’t know about this, Lillian. Fairy godmother said only true love’s kiss could break Fiona’s
curse.
Q: I don’t trust that woman, Harold. This may be our last hope. Besides, he does come highly
recommended by King Midas
K: But to put our daughter’s life in the hands of this ... Person? He’s devious, he’s deceitful, he’s...
He’s...
R: Rumpelstiltskin!
Mrs Highness!
Q: How do you do.
R: Down Fifi, get down!
As you can see, everything’s in order.
K: So you’ll put an end to our daughter’s curse.
R: And, in return, you sign the kingdom of Far Far Away over to me.
K: Lillian, this is madness.
Q: What choice do we have, Harold. Fiona has been locked away in that tower far too lang.
R: It’s not like she’s eh getting any younger.
K: But to sign over our entire kingdom?
R: Well, if your kingdom is worth more to you than your daughter
K: Nothing is worth more to us than our daughter.
80
R: Jump Fifi, jump. Just sign it and all your problems will disappear.
Guard: Your Highness! The princess! She’s been saved.
R: Huh. Who saved her?
R: No one would have guessed that an ogre named Shrek, whose roar was feared throughout the
land, would save the beautiful princess Fiona. True love’s kiss led to marriage and ogre babies. The
kingdom of Far Far Away was finally at peace. Oh, goody for them! And they lived happily ever after.
P: Sir? You gonna have to pay for that.
R: Maybe we can make a deal for it, little boy?
P: Oh, I’m not a real boy.
R: Do you wanna be?
P: Nobody needs your deals anymore Grumpel Stinkypants.
R: I wish that ogre was never born!
Kids: Wake up, daddy, wake up!
F: Good morning.
S: Good morning to you.
F: Better out than in.
S: Hehehe. That’s my line.
Did my little Fergus make a.. Woooah! Big grownup ogre stink...
Ooh, that’s diabolical.
G: And on your left, the loveable lug that showed us you don’t have to change your undies to change
the world. I wonder what Shrek’s up to in there.
S: Get in there.. Get.. Impossible to put on!
F: Okay, the dragon goes under the bridge, through the loop and finally, into the castle.
F & S: Pfffff.
S: Hmm.
D: Play date! Winter, spring, summer, autumn, all you gotta do is call!
Puss: ...and Shrek kissed the princess. She turned into a beautiful ogre. And they lived
D: ..happily..
F: ...ever..
S: ..after.
F: Look! A shooting star!
S: So, what did you wish for?
81
F: That every day could be like this one.
S: Ooh, come here, you.
P: (singing) One love, one heart.
Let’s get together and feel alright..
Kids: Morning, daddy.
F: Morning.
Better out than in.
S: Did my little Fergus make a..
Cute. Real cute.
G: This loveable lug taught us you don’t have to change your undies to change the world!
D: Play date!
F: Shrek! The outhouse is clogged up!
P: She turned into a beautiful ogre and they lived..
D: ..happily..
F: ..ever..
S: ..after.
Kids: Daddy, get up!
F: Morning.
Better out than in.
G: This loveable ogre...
D: Play date!
S: Noo..
D: You’ve got a friend
F: Outhouse again!
S: Come on! Oooh..
G: Undies!
F: Outhouse!
Kids: Get up! Get up!
P: And they lived..
D: ..happily..
F: ..ever..
S: ..after.
82
S: Rooaar [whispered]
-‐Flying in the air on the dragon-‐
[Laughter and music]
D: Nice landing baby. Hey, now remember, don’t eat the valet.
P: Happy birthday niños! Vamos a la fiesta!
Men: Hey, Shrek, Shrek!
Man 1: Ooh Mr Shrek would you sign our pitchforks?
Man 2: And our torches!
Man 3: Ooh man, you used to be so fierce.
Man 1: Yeah, when you were a real ogre!
S: A real ogre?
[Birthdaysong]
D: Come-‐on Shrek, it’s a singalong. You’ve got to sing along.
S: No, thanks.
D: Please? I’ll be your best friend.
S: Why does being your best friend entail me doing everything I don’t wanna do?
Please Felicia, not in Daddy’s ear.
Father: Excuse me Mr. Shrek. Could you do that ogre-‐roar of yours for my son? He’s a big, big fan.
BP: Do the roar.
S: You know, I’d rather not. It’s my kids birthday party and..
BP: Do the roar.
F: Euh, honey? Why don’t you go check on the cake.
S: Sure.
F: And don’t forget the candles.
MM: Hold still.
GM: Thanks for the pants, muffin man! I always wanted chaps! Yee-‐haw! Giddy up!
MM: Ah, monsieur Shrek! Your cake. Voila!
S: Oh, what is that supposed to be?
GM: That’s Sprinkels, the ogre.
MW: Isn’t he cute, he looks just like you.
D: But happy. It’s a party Shrek.
You gotta cheer up.
S: I’m in a great mood actually.
83
D: Ooh, I’m gonna lick me a rainbow.
S: Donkey!
Father: As long as you’re not doing anything, how about one of those famous Shrek-‐roars.
BP: Do the roar.
S: Let me set you straight Butter Pants. An ogre only roars when he’s angry. You don’t wanna see me
angry, do you?
BP: Do it.
S: Hold it together, just hold it together..
BP: Daddy, he’s getting away. Do something.
F: Oh, good. What happened to the cake?!
S: Trust me, it’s an improvement.
Q: Ugh. You licked it.
S: Noo..
Q: Just because you’re an ogre, doesn’t mean you have to eat like one.
BBW: Looks like you forgot the candles.
F: Okay, just watch the cake, I’ll go get them.
S: Watch the cake? Where’s the cake?!
Pig 1: We ate the cake.
S: What!? No, don’t cry..
F: Hey! I believe you promised my son a roar.
BP: Do the roar.
S: Roar.
BP: I don’t like it.
S: Pigs, we need another cake.
Pig 1: But we ate the others cakes.
Father: Ooh, come-‐on man, one roar.
D: Hey everybody! Shrek’s gonna do his famous ogre roar!
S: Not now, donkey.
S: Pigs, are there any cupcakes?
Pig 2: We ate them too.
Pig 1: They have lollypops.
Pig 3: No, I ate them.
Pig 2: You didn’t share?!
Pig 3: You didn’t share with the croissants!
S: Everything’s gonna be okay.
84
F: Shrek, what’s going on?
D: Come-‐on Shrek! Your fans are waiting!
BP: Do the roar.
F: We need the cake!
Everybody: Give us a roar! Cake! Give us a roar! Cake! Give us a roar!
S: ROOOAAAR
[Cheering]
BP: I love you, daddy.
Puss: Everybody, I have found another cake.
F: Shrek, are you okay?
Everybody: Shrek, Shrek, Shrek, ..
S: Aaargh!
[shocked silence]
F: Unbelievable!
S: Tell me about it, those villagers are..
F: I’m not talking about the villagers Shrek, I’m talking about you. Is this really how you want to
remember the kids’ first birthday?
S: Great! So this is all my fault?!
F: Yes. But.. You know what, let’s talk about this after the party, at home.
S: You mean that roadside attraction we live in? Step right up, see the dancing ogre. Don’t worry, he
won’t bite! I used to be an ogre, now I’m just a jolly green joke.
F: Okay. Okay. Maybe you’re not the ogre you used to be, but maybe that’s not such a bad thing.
S: Argh, I wouldn’t expect you to understand. It’s not like you’re a real ogre. You spent half your life
in a palace.
F: And the other half locked away in a tower.
S: Look, all I want is for things to go back to the way they used to be. Back when villagers were afraid
of me and I could take a mud bath in peace. When I could do what I wanted, when I wanted to do it.
Back when the world made sense!
F: You mean back before you rescued me from the Dragon’s Keep?
S: Exactly!
F: Shrek, you have three beautiful children. A wife who loves you. Friends who adore you. You have
everything. Why is it the only person who can’t see that, is you?
S: That’s just great.
85
S: If she thinks I’m gonna slink back there and apologise, she’s got another thing coming. She’s not
the boss of me. I’m an ogre and I’m not gonna apologise for acting like one.
R: Ooh, help me, help, please. Someone, anyone. Ooh, help me, please help. Please. I’m stuck. Oh,
help, ooh please help. Someone, anyone, oooh, help me! The pain. I can see a bright light. A tunnel..
Grandma, is that you?
S: Yeah, it’s me. Granny.
R: An ogre! Oh, please Mr. ogre, please don’t eat me!
S: I’m not gonna eat you.
R: Eh, but but, you are an ogre? Aren’t you?
S: Yeah, well I.. I used to be.
Move out or get crushed.
R: So, you’re not gonna eat me?
S: No thanks, I already got a big bowl of curly toed weirdo for breakfast
R: Hey hey, wait up! What’s your rush, where you going?
S: Nowhere.
R: What a coincidence, I’m just heading that way myself. But seriously, let me give you a ride. I insist.
Come-‐on, it’s the least I can do, after all you’ve done for me.
I got a hot rat cooking.
R: All right! Can I interest you in a mudslide? Slug and tonic? A liquid libation to ease that frustration?
Eyeball-‐tini?
S: Well, maybe just one.
86
8.3 Appendix 3 – The Test
!"#$%&&'%()*+)%,#'')'%-"#./%,#-'0
!" #$%&$$'($)%)*+%&,)+%$)%+$&-../+"
!" #$%&'()*%+'$!,,&'%*%-'!.%-"
/" #$%&'()*%+'$!,,)0%11'%*%-'!.%-"
2" #$%&'()*%+'$!,,)(&'%*%-'!.%-"
+" 3%%0')+%%"
0" 1.%234%$$)%*+$5"
!" 4'51%+'67'/%'!0'78-%"
/" 4'67'/%'51%+'!0'78-%"
2" 4'51%+'!0'78-%'67'/%"
+" 3%%0')+%%"
6" 1.%7$585*-2%(/$%75*-2%)/$89%:,5*&("
!" 4'6-516'6$!6'97:!0'076;'<!-7(+"
/" 4'+70=6'6-516'6$!6'97:!0;'<!-7(+"
2" 4'6-516'076'6$!6'97:!0;'<!-7(+"
+" 3%%0')+%%"
;" #$%&$.$)%7$$&%*<%$&.,,5"
!" #$%&'16-708(&'-%1%:/(%+'%!2$'76$%-"
/" #$%&'16-708%-'-%1%:/(%+'%!2$'76$%-"
2" #$%&'16-708'-%1%:/(%+'%!2$'76$%-"
+" 3%%0')+%%"
=">,8%?$@%A$%+$2$)48B
!" >$!6'+)+'9)1$'&75'?7-@
/" A)+'9)1$'&75'?7-'9$!6@
2" >$!6'+)+'&75'9)1$'?7-@
+" 3%%0')+%%"
C" D$%2,4%,&EA(%2*$48"
!" B75'67'/%'51%+'C%-2%
/" B75'51%+'67'/%'C%-2%
2" B75'51%+'C%-2%'67'/%"
+" 3%%0')+%%"
F">,,5*G%($H$%4<$I/J$.$%7/($*B
!" >$&'6$)1',!-D25(!-'*)+%7@
/" >$&'*)+%7'6$)1',!-D25(!-@
2" #$)1',!-D25(!-'*)+%7'9$&@
+" 3%%0')+%%"
E(!1F'""""""""""""""""""""""""""""""""""""""""""""""""""""""""""' E(!10-"F'"""""""""""""""""""""""""""""""""""""""""""""""""""
Name of school Date 1
87
!"#$%&'(&)*&+),)-*.&/)%&.)&%$$*%0
!" #$!%&$!''()(*&%+&%$(&,!-(.
/" #$!%&%+&%$(&,!-(&$!''()(*.
," 0$(&,!-(&$!''()(*&%+&1$!%.
*" 2(()&3*(("
1" 233)(&*445&(%)*5&6$$*&74-%/))3"
!" 45(67%$3)8&9:(;;(*&'+1(6<=;&;3-(&9!1*=9%"
/" 45(67%$3)8&9:(;;(*&'+1(6<=;;7&;3-(&9!1*=9%"
," 45(67%$3)8&9:(;;(*&'+1(6&;3-(&9!1*=9%"
*" 2(()&3*(("
>?"&8)&+$$%&/)&6')%&49)%)60
!" @+=A6(&8+))!&(!%&:(&)+%.
/" @+=A6(&8+))!&)+%&(!%&:(.
," @+=A6(&)+%&8+))!&(!%&:(.
*" 2(()&3*(("
::"&;5&54<7%&$3=>.&7))3&.-*)&/$5)?-9"
!" B&=9(*&%+&/=7&6(!;;7&(C'()935(&:!-(D='"
/" B&=9(*&6(!;;7&(C'()935(&:!-(D='&%+&/=7"
," B&%+&/=7&=9(*&6(!;;7&(C'()935(&:!-(D='"
*" 2(()&3*(("
:@"&;5&7$.&$3&))6&+*4%)&54/&<-*3A&%4).&B)'*.4)(&$3(&46%,'>%"
!" B&!;6(!*7&8+%&!&/38&/+1;&+<&,=6;7&%+(*&1(36*+(9&<+6&/6(!-<!9%"
/" B&8+%&!;6(!*7&!&/38&/+1;&+<&,=6;7&%+(*&1(36*+(9&<+6&/6(!-<!9%"
," B&!;6(!*7&!&/38&/+1;&+<&,=6;7&%+(*&1(36*+(9&<+6&/6(!-<!9%&8+%"
*" 2(()&3*((
:C"&;5&7),&7)%&6')%&+*$$+"
!" B&;3-(&3%&)+%"
/" B&)+%&;3-(&3%"
," B&*+)A%&;3-(&3%"
*" 2(()&3*(("
:D"&#)&7),,)6&.)&<$5)&49+)+)%)6"
!" #(&%$(&,!-(&!%("
/" #(&!%(&%$(&,!-("
," E%(&%$(&,!-(&1("
*" 2(()&3*(("
F;!9G&""""""""""""""""""""""""""""""""""""""""""""""""""""""""""& F;!9)6"G&"""""""""""""""""""""""""""""""""""""""""""""""""""
Name of school Date 2
88
!"#$%&'$('))'$*++(,
!" #$%&$'(&)!*'+,
-" .'(&)!*'+&/$%,
0" #$%&)!*'+&$'(,
+" 1''2&3+''"
!-#$./00&1'$1''23'4$1++4$4//&3$5'1#
!" 4%5'&6$%)7)&6%&!/!8&2'*'("
-" 4%5'&6$%)7)&2'*'(&6%&!/!8"&
0" 4%5'&6$%)7)&6%&2'*'(&!/!8"
+" 1''2&3+''"
9:"&6+($6+($75+8$5+2$'&4)'9&:;$5''($&4$<(')'#$
!" ;!(&<!(&!/!8&/!)&=2!>&!7&?'!0'"
-" ;!(&;!(&@/!8&/!)&5%)7&=2!>&!7&?'!0'"
0" ;!(&;!(&@/!8&/!)&=2!>>8&!7&?'!0'"
+" 1''2&3+''"
!=#$>'3$&2$''4$?''23@$.*(';#$
!" A7&3)&!&?!(78B&4$('C"
-" A7&!&?!(78&3)B&4$('C"
0" A)&!&?!(78&37B&4$('C"
+" 1''2&3+''"
!A#$B;$C'4$4/1$4&'3$;9++(#
!" A&!5&2%7&+%2'"
-" A&2%7&!5&+%2'"
0" A&!5&+%2'&2%7"
+" 1''2&3+''"
DE#$B;$*/F$4F$+9$<+4$*'0#
!" A&!>('!+8&$35&>%*'"
-" A&>%*'&$35&!>('!+8"
0" A&$35&!>('!+8&>%*'"
+" 1''2&3+''"
D>!)E&""""""""""""""""""""""""""""""""""""""""""""""""""""""""""& D>!)2("E&"""""""""""""""""""""""""""""""""""""""""""""""""""
Name of school Date 3
89
!""#$%%&$'(%)%$*+&,
!"#$%&%'()$*$+&),$),-$./0&112')1$*$304$*$5$*$1,2/-
#############################################################################################################################################################################
!!#$4'%-/$),-$6/&%7-$*$),-$%/270'$*$#$*$70-1
#############################################################################################################################################################################
!8#$92%'-11$*$),&1$*$#$*$&1
#############################################################################################################################################################################
!:#$,2;-$*$),/--$6-24<=4>$.,&>%/-'$*$#$*$304
#############################################################################################################################################################################
!?#$@41)$>&A-$*$>00A1$*$#$*$,-$*$304
#############################################################################################################################################################################
!B#$;-/3$*$&1$*$=21.&'2<'7$*$#$*$,-
#############################################################################################################################################################################
!C#$%&%$*$+,-/-$*$.09-$*$304$*$2./011$*$5$*$,-/
#############################################################################################################################################################################
!D#$2/-$*$2>>$292)-4/1$*$#$*$+-
#############################################################################################################################################################################
!E#$'0)$*$+-$*$)2>A&'7$*$#$*$2604)$2'$&%-2>$10.&-)3$*$2/-
#############################################################################################################################################################################
8F#$'-;-/$*$),-3$*$.2>>-%$*$#$*$@41)$)->-;&1&0'1$*$+-/-
#############################################################################################################################################################################
-%)"&#.$/((0$1233+%$4%)%5%0#+&'6
G>21H$##########################################################$ G>21'/#H$###################################################
Name of school Date 4
90
8.4 Appendix 4: Results
8.4.1 Figure 1: Table Crosstabulation Education type with Contact with English.
8.4.2 Figure 2: Crosstabulation Education type with frequency of watching English television
programmes and movies.
Educationtype * watchengprogmovies Crosstabulation
Count
watchengprogmovies
Total yes sometimes seldom no
Educationtype BSO 10 14 1 1 26
ASO 35 21 2 1 59
TSO 17 11 3 0 31
Total 62 46 6 2 116
8.4.3 Figure 3: Crosstabulation Education type with the frequency of using Dutch subtitles.
Educationtype * Dutchsubtitles Crosstabulation
Count
Dutchsubtitles
Total yes not always no
Educationtype BSO 20 6 0 26
ASO 34 25 0 59
TSO 15 14 2 31
Total 69 45 2 116
Educationtype * contactEnglish Crosstabulation
Count
contactEnglish Total
no seldom sometimes often very much
Educationtype
BSO 1 7 6 11 1 26
ASO 1 3 11 17 27 59
TSO 1 0 10 8 12 31
Total 3 10 27 36 40 116
91
8.4.4 Figure 4: Crosstabulation Education type with the use of English.
Educationtype * sometimesspeakEnglish Crosstabulation
Count
sometimesspeakEnglish
Total yes no 99,00
Educationtype BSO 8 16 2 26
ASO 31 23 5 59
TSO 14 17 0 31
Total 53 56 7 116
8.4.5 Figure 5: Boxplot graph correlation of the results in total with the three education types.
92
8.4.6 Figure 6: Boxplot graph of the correlation results in total with the three fragment types.
8.4.7 Figure 7: Graph comparing the results on the first and second test from each education type.
!"#$%&
'(#!'&!)#!*&
!+#,+&
'!#('&!)#$$&
(&
+&
!(&
!+&
'(&
'+&
-./& 0./& 1./&
!"#$%&#'
()$*+,-.'&/0"#'
1-20+34#-.'!"#$%&#'543#&'+.)'6"*-.)'7"#&'
1234&!& 1234&'&
93
8.4.8 Figure 8: Graph showing in percentages the results on each grammatical aspect correlated to
the fragment types.
8.4.9 Figure 9: Graph showing in percentages the results on each grammatical aspect correlated to
the education types.
!"!#
!"!#
!"!#
!"!#
!"$#
!"$#
!"$#
!"$#!#
!#
!#
!#
%#
&%#
'%#
(%#
)%#
*%#
+%#
,%#
-%#
.%#
&%%#
/012345#64782# 98:17;2# <1=>345# ?15@15A1#?@70A@071#
!"#$%&#'()
"*+",&-."/'
0*-11-2+-%'3#4"+&#'
!"#$%&#'0*-11-2+-%'-#4"+&#'5'6*-.1",&'784"'
!"#$
!"#$
!"#$
!"#$
%"#$
%"#$
%"#$
%"#$
&"#$
&"#$
&"#$
&"#$
'$
('$
)'$
*'$
+'$
,'$
-'$
.'$
/'$
0'$
(''$
1234567$869:4$ %:;39<4$ =3>?567$ "37@37A3$"@92A@293$
!"#$%&#'()
"*+",&-."/'
0*-11-2+-%'3#4"+&#'
!"#$%&#'0*-11-2+-%'3#4"+&#'5'67$+-28,'9:4"'
94
8.4.10 Figure 10: Graph showing the comparison of the results in total of both fragment questions
and non-‐fragment questions, in correlation to the three education types.
!"#$$%
$"#&"% '(#)&%
*$#**%
')#&+%!)#$!%
,%),%",%+,%*,%!,%',%$,%(,%&,%
),,%
-./% 0./% 1./%
!"#$%&#'()
"*+",&-."/'
01$+-23,'&45"#'
6375-*8#3,'9*-.7",&'-,1':3,;9*-.7",&'<$"#23,#'
23456789% :;8<23456789%
95
8.4.11 Figure 11: Boxplot graph that correlates the results on the first test to the extent of contact
with English.
96
8.4.12 Figure 12: Boxplot graph that correlates the results on the second test to the extent of
contact with English.
97
8.4.13 Figure 13: Boxplot graph with the correlation of the results of the first test to the extent of
contact with English from participants from ASO.
8.4.14 Figure 14: Summary of the cases that are processed in the correlation of the results from the
first test to the types of television programmes and films watched most by all participants.
Case Processing Summary
whichtypemost
Cases
Valid Missing
N Percent N
totaal1 Dutch 24 100,0% 0
English-Dutch 60 100,0% 0
English 2 100,0% 0
English-English 1 100,0% 0
other 2 100,0% 0
English-Dutch+other 2 100,0% 0
Dutch+English-Dutch 16 100,0% 0
Dutch+other 2 100,0% 0
Dutch+English-Dutch+other 3 100,0% 0
98
English+other 1 100,0% 0
Dutch+English+other 1 100,0% 0
English-Dutch+English-
English
1 100,0% 0
99,00 1 100,0% 0 8.4.15 Figure 15: Boxplot graph with the correlation of the results from the first test to the types of
television programmes and films watched most by all participants.
99
8.4.16 Figure 16: Boxplot graph with the correlation of the results from the first test to the ability of
speaking English of all participants.
100
8.4.17 Figure 17: Boxplot graph with the correlation of the results from the second test to the ability
of speaking English of all participants.