SriPIIce Educntion lntf'rnationaf \0/. 17, Xo. 2,June 2006.jJjJ. 77-93

lntematiounl Cor111ril of.\ssocialitwr.. in Srimu t:ilumtion

Theoretical Concepts for Using Multimedia in Sience Education

GIRWIDZ, R., >-'RUBITZKO, T., *SCHAAL, S., '''*BOGNER, F. X., *Uniuersity of Edumtion. Ludwii(Sburg, Gt>mwny, **Univrrsity of Bayr('ll//i, ßa)'l'l'llth, Germany

rlwmtiCI' Multirotfing (u.)ing various kinds of representation), multimodality (addressing seueml SI'II.!Ps). and interactivity arp sfJerial Jeatures of multimnlia. We derived SI'7Jt'l'ftl guidelines 011 tht• basis of fJ.\yrlwfogicaf findings in ordf'l' to app~v muftinwdia, speäfically dfsignrd to assist effectivf fpam ing in tomfJlex fields of Jmowft>dge. I Ve f ocussed on f1romoting cogniliue.fli•xibility. buifding adt>quatl' men­talmodels. implrml'llting "~ituated lraming, "stnuturing lmowfrdge mul linliing fields of Jwowlnlge. IH lllf'll as adjusling rognitiue Ioad. In this cont1ibution, Wl' fn -esent an OIIPI1Jiew of theoretiail ron~idP­mtions and il/ustrate f>os~iblr implementations. Thr aim was to delinratr guidrlines for a thPOry-guid­ed droelopment of multimedia applimtions.

1\1-.r 1\'0/U>S: Multimrdia, ICT, scirncr edumtion, situated leaming.

Introduction

Modern media o ffer innovati,·e and spccific informatio n and communication opponunities that adctrcss multimoctali t~· ( i.e., the possibilit} to address sevcral scnscs), multicoding of infonnation (i.P., a ffering differem prcsemations and code sysLCms), and interacti,ity wilh a learncrl user. \"'h en focussing on learning proccsses, it should bc clear that thesc fcatures charactcrizc a rather superficial strucwre of communication and intcraction, and further aspects of learning have LO be considered. For example, multicoding might be hclpful in assisting lcarning by combining the potential of different code systems. However, combining repre­sentations may incrcase the densit:y o f info nnation and gene rate cognitive over­load.

Thereforc , funher aspects of learn ing should be considcred thaL extcnd beyond the supcrficial perceptual strucwre. This paper derivcs somc theore tically wcll-founded iclcas to cope ''~th scveral requirements from the psychology of learn­ing, as it is shown in th c oveniew of Figure I.

Multimedia in Science Learning The Superficial Structure of Information and Interaction:

M ultimodali ty iVI ulticoding lmeractivi ty

Deeper Structure of Learning: Cognitive Flexibility I Situated Learning Cognitive Load I Knowledge Structurc

\lental \Iodels

Figw'l' I. Theoretical C.onctpts for L~ing .\lultimedia in Scit'llrl' l~·dumtion

78 Cirwidz., R., Rubitz.ko, T, Schaal, S., and Bogner, F X.

In the following, we brie11y outline some unclerlying theories and subsequent­ly describe some specific cxamplcs \\~th the intent to bridge the gap betwecn the­ory and practice. The intcntion is to show how multimedia applications can be designed according to relevant thcoretical guidelines. A subsequent paperwill sug­gest a specific 6-lcsson unit based on the desCJibed theoretical guidelines (Girwidz, Rubitzko, Schaal, & Bogncr, prescnt issuc) .

Multimodality, Multicoding, and Interactivity

Firstly, we consicler the superficially noticcable aspect5 of interaction with mul­Limedia. According to Weidenmann (2002), multimodality, multicoding, and inte­ractivity describe spccial fcawres of information and communication struclllres, offering new ways for teaching and learning. Multimedia applications normally adclrcss auraland visual scnses. Inte ractivity is expecteclto have motivational efTects by crcating a new sense of responsibility for a lcarning process and by permiaing learncrs to play an active roJe.

Multimodality (Sounds and Pictures)

A multimcclia application prcscnting audio ancl visual information activatcs dif­ferent sensory systcms and provieles a more realistic ancl authentic approach. Maycr ( 1997) clescribecl the combined presentation of verbal ancl visual informa­tion as specifically helpful for inexperic nced lcarners. Mayer and Moreno ( 1998) oltllincd a split-attention-cffcct and bettcr learning results when ,·erbal and ,·isual aspect.s are combinecl, compareclw a writ.ten text. alone. The authors assumecl the importancc of two sepanne channels für proccssing visual and auclitory informa­tion in working mcmory. The combination of verbal and written components, thercfore, may Iead to a bcucr proccssing in limited working mcmory (Morcno & Mayer, 1999).

Multicodiug

\"'c idenmann (2002) highlightcd thc importance of different coding systems when using multimedia to promote lcarning. Any processing of information is based on code-spccific expressions, especially in tl1c early stages of a learning proccss. Aclditionally, physiological swdics providccl evidcnce that spccific cercbral cortcx areas are responsible for processing textual and visual information (Springer & Oeut~ch , 1998). Ditl"c rent codes ofl"cr difl"crent ways to communicate infonnation, but thcy require spccific skills. Scl111otz and Bannen (2003) distin­guishcd textual and pictorial represcntations, ancl statcd that cvcn diflcrent kinds of visualization engagc different knowlcdgc structures that art~ useful for specific applications. Howe,·er, illustrations embcddecl in texts do not ine\'itably rcsult in positive effccts. Only less giftecl lcarners scem to profit gencrally from picturcs in texL Gifted stuclents seem to be able to dcvelop an adcquatc memal moclcl with­out pictorial prcscntations (sec Schnotz & Bannen, 2003; tVIayer, 1997). Visualizations arc useful if they present f~1cts and concepts in task specific ways. Othcrwise, they interferc with other concepts and miglu cvcn causc difficulties (Schnotz & Bannen, 2003). Thercfore, a task specilic and appropriatc form of pre­sentation plars an imponant roJe in lcarni ng.

Using MultimPdia in ScinlCe Educalion 79

l11teractivity

Learning theories more and more emphasizc active learning ancl self-pacing in lcaming Situations. llc rcwith, adequatc ancl individual fccclback is considcrcd as a fundamental prc rcq uisite. Computers oiTer remarkable advantages for imeractiYe lcarn ing (Steppi, 1989). Real interaclivity requires that (i) lcarners can be creati,·e. and may choose and moclify contenL<~ by themseiYes; (ii) the program adapts ancl dynamically rcacts tO learners' actions; (iii) learners gain control o,·er the proccss­cs of lcarning; ancl (iv) assist.ance or guidance is g iYen on demancl by thc multi­lllCdia system. Interaction ancl communication might bccome even morc a ttractive by the use of light pens, voicc recognition , data g loves, and/ or eye camcras. I Iowcver, users havc to be familiar with thc information tcchno logies and be able to use them appropriately. New skills are required to copc with the ncw possibili­ties, ancl to organize and structure multiple sources of infonnation.

In summary, mullimedia systems oller new possibilitics fo r presenting know­lcclge and a bcu.c r accessibili ty to clata. They can combine information , makc it availablc in flexible a ncl int.e ractive ways, and help to rcalizc the spatial and tem­poral contiguity principle, meaning that corresponding tCXL'i and picturcs arc clo­scly posi tioncd tagether and presented simultaneously (rather than successi,·c ly).

T heoretical Guidelines and Concepts

These featurcs of' multimedia charactcrize superficial structures of a lcaming environment. To make them effectivc l'or learning, thcy should be part or a morc dc tailed conceptualizat ion. Mayer ( 1997, 2001) presentcd a thcory ancl empirical-1)' wcll-foundecl gu iclc lincs für multimcclia learning based on three thcorctical assumptions. First, thc human information processing systcm is based on two chan­nc ls fo r auditory and visual input, meaning that verbaland pictorial information is proccssed sepa ra tc ly. Sccond, each channcl has a limitcd capacit)'. Th ird, learning rcquires act.ive processing and happcns when attentio n Ieads to thc sclection ancl organizalion or pe rccivcd infonnation, so that it can be intcgrated cohere ntly in an cxisting knowlcdge structure. All thcsc processes are sensitive and can casily be upset by cognitive ovcrload. Maycr ancl his group prescnt a well-im·estigated cata­loguc of effects that should be considcrcd when designing multimedia applications f'or lcarning. Among thcm are:

(i) Multimedia d'f'ect: Learners perform better whcn information is prcscnted in worcls and pictures than in worcls a lone.

(ii) Modalit)' c fl'cct: Be tter transfer is expected when animalions are ofl(Ted tagether with narration rather than with wriuen tcxt.

(iii) Sparial contiguity eiTect: Lcarncrs perform bcucr when text is placcd ncar rathe r than l~tr frorn corresponding pictures.

(iv) Temporal contiguity e ffect: lt is bc tte r to prcscnt an imations and cnrre­sponding narration simultancously rathe r than succcssivcly.

(v) Cohercncc c l'f'cct: Irre levant words, picturcs, and sounds sho uld be excluded.

(vi) Redundanq• cfl'cct: There is bctt<:r transfer from an imation and narration than fro m animation, narration, and on-scrccn tcxt.

80 Ginuidz, R., Rubitzko, T, Sduwl, S., and ß ognPI; F X.

(vii)Signaling e ffect: Bc tle r transfc r is achievccl whe n narra tio n is signaled rather rhan non-signakd.

(Sec r-.'layer, 2002, also for thc pre-train ing effect and personalization princi­ple .)

These eJTecL<> wcre inYcstigated using short m ultimedia applications dcaling with causc-and-effert-chains and explaining how things \rork, r. g., how a bicycle pump o r car brakcs work, or how lightning sto rms dc,·elop. \1\'hcn using m ultime­clia fo r 111orc complcx cont.e nts ancl for tcaching units comprising more t.ha n onc lesson , further thcories a nd guidelincs become intc resting 10 assist a nd manage complcx systems and interrclations. Duc to space limi tations, only the basic idcas can be sketchcd he re . Rcfcrcnces to thc Iitera ture a re presem cd for thosc requi­ring morc cle tails. Spccific cxamples a rc prcsented in cve ry scgmcnt to illustrate how 10 apply thcOI")' a nd how to bridgc the gap bc t.wcen theory a ncl practicc.

Looking at complex lields or knowlcdge acquisition, it is intcnded to (i) Iüster a fl exible use of various kin cls of knowlcdgc represC'n tations, (ii) hclp to dcvelop aclcqua tc me ntal mode ls for scic nce phc no me na, (iii) assist in constructing a we il o rganizcd knowlcclgc struclllre to guarantcc access to knowleclgc tha t is necclccl for probkm soh-ing and applications. (i,·) usc rich contcxts and arn111ge situatcd learn­ing a nd anchorcd instruction, in o rcler 10 avoicl "in ert knowlcdgc," ancl, last but not least, (v) usc th c specia l bcnd its o r nlltlt"imedia , i.e., multimodality, n1111ticod­ing, and in tcracti\'il)', witho 111 procl ucing cognitive overloacl.

Fostering Cognitive Flexibility

"Cognitive rl cxibility" indJJdcs the abili1y to restrucwrc acguircd knowl cclgc according lO thc clcmands or a g ivcn Situation (Spiro & Jehng, 1990) . T hus, a knowlcdgc enscmblc can bc constructed ta ilored to thc nceds ofa problem-solving si tua tion , or to s11 pport lcarning ancl lin king or ncw conccpts (Spiro, Fcltovich, .J acobson , & Co11lson, 1992) . Cognitivc lkxibili ty he lps tu apply knowleclgc unclc r various conditions in an e lfec live way.

a) Cog11itive Flexibifity and Multiple R.epresel/lation

To fostc r cogJJitivc fkx ibility, knowkclge sho11ld genc rall)' bc consolidatc cl from d il'fc re n t conccptua l pc rspcctivcs. Knowlcdgc shoulcl Iw structurcd a ncl taught in diflcrcnl forms to bc func tional in m11ltiple situations. One assumption of th c cognitivc lkxibil iry thcoq• is thal spccific leam ing enviro111n cnts arc nccd­ccl. Facts sltould br prese ntecl a nd lcarnccl in many cl ifl"c re nt. ways, ancl knowlcdge sho uld bc intcgra lccl in a va riety of see na rios (Spiro, Coulson , Fe lwvich, & Anderso11. 198R; Spiro ct al., 1992). T his is panic11larly imponant for ill-struclllred kn owlcclgc domains with h igh across-case irregulari ties a ncl conceptual com plcxi­lics. Kozma (200:i), l(w instan rc, clescribccl a cliffe rc ncc be twcC' n cxpe n s a nd sl.ll­clenL~ (novices) rcgarcli ng 1lw JJsc o f llllllti p lc rcprcsenlatio ns. \l\1hile cxpc rts use them wi1 h purposc. novicc s1nden ts may fi1ce dil1iculties in con uccting m11ltiplc reprcscntations adcquatcly. T hcir Observations and a rgumc nts often Iimit lh cm­sehrc~ lo supe rlicial fcawrcs. ovices also tcnd to concc ntratc on a singlc dcscrip­tion, wh i k expcrts w;e vario us reprcsenta tions and scctn to a l 1 nnatc casi ly from one 10 a nother. In gcncral, nwn talmu hi-coding improvcs acccss to knowlcdge and prob lem-~olvi 11g 1 C'ch n iqucs.

Using Multimr>dia in Science Eduwtion 81

The application in Figure 2 shows a ct~mera in a realistic arrangemcnt (pho­tography) combined with an abstract drawing. :VIanipulations of a realislic camera, simultaneouslr cause modifications on an abstract dra\\ing. For example, chanbring the aperturc modifies the bundle of light entering the camera. Thus. learners are assisted lO build connections between different represemations.

In general, linking textual, mathematical, and piclOiial representations may facilitate a proper handling of problems.

Figurt 2. A \"irtunl Cnmtra i11 Tlt'o RPpYPsmtntiolls.

b) Restruch~riug

:Vlastei~ing different symbol systems is one area of competence, the other is co~

ping with various dcscriptions in a single S}'Stem. For example, identical witing dia~ grams, as thosc prcsentecl in Figurc 3, can bc drawn so cli.fferently t.hat novices may not recognizc thcsc circu itries as idcntical. Computer animations can illustrate equivalence among thcse circuitries by showing the essential transformations step by step (see also l-Hirtcl, 1992).

rn Figurr 3. Sl'tlf'ral hmiii'S of n C.omputtr Animation. Subsrqumtly Tmnsfening Ont Ci1ruitr)' into. \nolhtr (lo

Show ThrirHquil'flll'llrf).

82 Ginoidz, R. , Rubitzlw, T, Srhaal, S., and Bogne1; F X.

c) SuppTantation

The so-called supplamalion principle may pro,ide substalllial suppon in rela­ling different operaüons lO each other (Salomon, 1979, 1994) . An example is detailed in Figurc 4, describing the bchaviour of an oscillator both in a rcpresen­tation similar to realit:y. and by the usc ofa corresponcling y(t)-diagrttm. T hc arrow connects the concrete and thc abstract reprcscntation. As the oscillator is mo\ing up ancl down. an arrow points to corrcsponding positions in the graphical descrip­tion of this process. Thus, thc concept or a line graph is illustrated.

yfm OIS

01

00!

001

01

ou t/a

l'igurr -1. /·iwnl' nf an Animation: Y(t}diagmm multltl' Comsponding Smwrio.

d) Comparing and Linking Different Representations

More giftcd lcarners can discover interclcpcnclencics on thcir own, when cor­responcling rcprcsentations are orfered simultancously. For cxample, the rompu­rer program "Atomos~ suppons the understancling of sub-atomic su·ucturcs in physics by clescr ibing c lcctron clensities within a hydrogen awm (Girwidz, Gößwein, & Stein rück, 2000), as indicatccl in Figure 5. DifTerclll kinds of diagrams

_ .... ::i;!:::····

i'lt;W<' 5. Srll'rtrd Oilfllay~ of Ihr Cnmfmli•rl'mgHun ":\tnmo\ ..

Using Multimedia in Science Edu.cation 83

and figures are put together. All of them desCiibc the same issue - the probability to find the e lectron at a certain position. Density of dots, intcnsity of colours or relief contours Iead to an intui tive understanding of high and low values.

e) Linking Knowledge by Using Hypermedia Systems

Spiro el. al. (1992) described "ill-structured domains" as knowledge domains with a complex structure that may mislead novices in particular. In orcler to con­nect multiple facts with various aspects in such a domain, a whole fielcl of know­ledge neecls to be worked through. Multimedia can offer support: A nonlinear medium like hypertext might be very weil suitecl for the kinds of 'lanclscape criss­crossings' recommenclecl by Cognitive Flexibilit)' Theory (Spiro et al., 1992). In hypermeclia systems, re lationships are represented by links interconnecting infor­mation nodes. These nocles a re compositions of textual or pictorial rcpresenta­tions, auclio-visual information in vicleos, animations, or interactive sirnulations. The information nocles can be organizecl sequentiaiiy, hierarchically, or arbitrarily. ßy linking multiple kinds of information, hypermeclia systems can promote a flex­ible access on knowleclge.

Construction of Adequate Mental Models

The term "mental models" points to analogaus cognitive rcpresentations of complex interdepenclencies within a knowledge domain. A classical example is the functioning of a steam engine or an e lectric buzzer (De I<Jeer & ßrown 1983) that is exemplifiecl in Figure 6. The design of multimedia applications can be based on the underlying theories, particularly when external pictorial representations are employed.

conductor

Figure 6. Eiearie ßur.41:r (from de 1\lter & Bmwn, 1983).

For research in teach ing ancllearning, mental models offer an attractive theo­rcr.ical backgrouncl. The)' are increasingly usecl for explanar.ions, but, unfortumuc­ly the tennisnot usecl consistently (Ballstaeclt, 1997). This paper fo llows thc initial cletinition of.Johnson-Laird (1980), Forbus and Gentner (1986), See! (1 986), and Weidenmann ( 1991). Mental models are analogous, pictorial representations enabling t.he brain to simulate complex syst.ems ancl to imagin e how they rnight work uncler different settings. Typical examples wirhin such a cont.exL are imagining astronomic processes, atmospheric circulation, or plant photoS)'nthesis. Mental models follow the assumption that human beings construct cognitive moclels of reality, reflecting aspect<; that are important for an individual. They give a reference

84 Cinvidz, R., Rubitzko, T, Schaal, S., and Bognt>r, F X.

framc for underst:anding new issues ancl pro,~cle a base for subsequent planning (Dutke, 1994).

Superficial configurations ancl cleeper structures shoulcl be distinguished. Funhermore, Einsiedler (1996), as weil as Schnotz, Bannen. and Seufert (2002), discriminatcd the medial rcprcsentation and i t~ corresponcling sensory perception on the one hand, from the fundamental subject structure, on thc other. Hencc, multimedia mercly can assist rhe construction of mental modcls and can only gi\'e hints in cxtcrnal representations, while mental models arc consu·ucted in the brain of an individual learncr. Seel ( 1986) aclded two funher aspects to bc consiclercd: the compalibility of external represcnt:ation to the internal mental represcmation, and its specific fit to a particular topic. Nevcrthelcss, cluc to its specific featurcs, multim eclia proviele the best basis for presenting information and supporling ade­quate mental models (lssing & Klimsa. 2002).

Multimedia a11d Me11tal Models

Multimedia may combine se\'craltypes of prescntation and tlms a\·oid overem­phasising superlicial aspects or a speciiic represcntaLion. Figure 7 combines a real sctting (phoLOgraphy) with a graphical representation. The photography is (gra­clually) substitutcd by an abstract symbol set within a dynamic computer ,-isualiza­tion, illustrating the undcrlying process for the de,·elopment of clouds within a tcmpcrature in\'crsion scuing. Dillerent shacles of grcy in the right picturc inclicate the tcmperaLUrcs involved, and arrows show the flow ofair (Figurc 7). Such an ana­logous external reprcscntation may hclp in constructing an adequatc mental rnodcl.

Weidenmann ( 1991) specificd differcm t}"pcs of suppon für thc construction of mental modds when using illttstrations: (i) i\nivation: Pictures can ani'"'tte an already cxisting mental moclel and cstablish starting points. (ii) Construction: Illustrations can show llow singlc well-known componcnts are intcgnnecl within a st tpcrior strucwre, and hence assist in structuring and unif·ying knowledgc. (iii) Focus: Picturcs can emphasize spc-cial aspccts of an existing mental modcl and may acUust or elaboratc thcse aspccL<;. (iv) S11bstit ution; Visnal rcprescntations can il­lustrme complcx ancl clymunic aspects to show the intc-rplar of sening in moclcls. (Picturcs can also be lincd up to rield animations and show time eiependent a­spects.)

Besides picturcs or animmions. intcractivc simulations ma~' darifY important corrc-lations ancl relationships and assist in thc constn tction of aclequatc mental modcls.

U~ing Multimerlin in Science Edumtion 85

Situated Learning

Theories of situa ted lcarn ing considcr learning as eiependent o n activity, con­tcxt ancl cultural e rwironmcm (Lave 1988; Lave & \o\lenger, 1990). Conscquently, (i) knowlcdgc has tobe prcsemed in authentic contexts, under circumstances and in applications wherc this knowledgc normally is usccl , (ii ) learning nceds social interacLion and cooperation.

In the following, we consider the first aspect:

a) Anchored lnstruction

Anchorcd instruc tion (CTGV, 1993: ßransford, Shcrwoocl, Hassclbring, Kinzer, & \Villiams, 1990) implemcnts some thcorctical aspects of situated cognition theo­ll '· One important inte ntio n is to overcome "inen knowledgc," knowkdgc that can bc repeat ed in classroom tcsts, but canno t hc lp in sillrations whcre problcms nccd a solution. Initially, rcscarchcrs or the GTGV (Cogn itirc TechnoiO).,")' Group of Vanderbilt) f(>cused o n thc dcvelopmcnt of imeractivc viclcodiscs intcnclcd to sup­pon ancl stimulate swdcnts (and thcir tcachcrs) lO deal with complcx, realislic problems. Thosc videodiscs provided interesting, realistic .. anchors" as kick-o!E for tcaching and lcarning. Th cir narrative charactcr is intcncled to catch swdcnts' interest and motivatc them to investigatc th c problt:ms prcsentcd.

The corllcnt inm lvcd (i.c., facts, conccpt<;, thcoric~ and pr·inciplcs) must be mcaningful for an individual. Thus. knowlcdge gain is both a valuablc rcsult and simultanco11sl)' a tool to ropc with relevant qucstions. Spccilic ''anchors" may link knowledge w applications ancl offer a framework for integrating knowledge from different clomains. Anchoring knowledgc to realistic frames is thought LO support specific problem-sohiug strategies (Goldmann, Pctrosino, Sherwood, Gar-rison. Hickey, Bransforcl, & Pellcgrino, l 996).

b) Simulatiug aud Modelliug of Problems

A dccpcr understanding of rele,·<mt parameters in rcalistic seenarios may atisc from Simulations ancl programs which allow modelling. An example is the calcula-

Figurt' 8: .\/01Miing DiJff'rt' /1/ Stralegies of a \'iltua/ .\lammallo Suaessful~\' Sumitll' \\'inter. \ 'ruiah/es. ntrh a1, lnlllfalion, /Jody Si:.~•. and MofJifi(\' 1/mJP lo JJ,, Cholf'll.

86 Girwidz., R., Rubitz.ko, 1:, Schaal, S., and Bognn; F X.

tion of cnergy requirements for a room, where the sizc, thc kinds ofwalls, \\indows and insulations can be varied. Similarly, for interdisciplinary learning the energy balance of living beings is an intcresting topic. Surviving a winter neccls variables, such as insulation, ingestion , and mobilit:y to be taken into accoum. A specific application is shown in Figure 8 (,;nual mammal). Therefore. different habitals or hibernation strategics (such as winter-dormancy) can be varicd as well as the body size and the fur's attributcs. Learners can test their chosen settings ,;a this simula­tion and get instanl fcedback.

Structuring Knowledge

Oe .Jong ancl ~joo (1992) iclentiliccltwo imponant pans of learning processes: structuring knowledge and linking itto prior knowledge. Well structured and prop­c rly organizccl knowledgc is also im portant for problem-sol\'ing (Reif, 1981). A hicrarchical structure improvcs accessibility ancl key words can point to relentnt detai ls. Van I lem·elen ( 1991) auached special im portance to mediat i ng generat principlcs. Clark ( 1992) clistinguishecl venical linkagc (e.g., assigning a problem to a generat principle) and horizontal linkage (P.g., connecting a problem to simi lar knowledgc structures, such as analogies). Horizontal links are of special inte rest when u·ansfcr is requested. CharL'i, mind maps, and diagrams may illustnue cogni­tive Connections, help to analyze a knowledge domain, and impro,·e recognition and recall of specific lcarning mauer (Beisser, jonassen, &.: Grabowski 1994). Effective knowledge managemcnt involves a wcll-organizecl structurc of knowledge and incluclcs techniques lO refin e and cxtend dcclarative and procedunll know­ledge.

a) Notatio11s

Mind maps and conccpt maps are strucwred clisplays of kcy tcrms. including also tcxt-picture combimuions. Conccpt maps represcnt a knowleclge domain by using nodes (usually specilic kcy worcls or centrat statcmenL~) ancllines to indicatc Connections. So-called ·'reference maps" aim to depict knowlcdge strucwres and oller an appropriate f'ramc for access to informalion. Charts in panicular accetllu­atc a hierarchical structurc and show YCrtical ancl horizontal arrangements. Thus, these featurcs proviele a useful framework for transfonning knowledgc into a visu­al representation, which can casil)' be communicated. Visualized knowledge stru­ctures may assist various instructional aims (Ballstaedt, 1997), for instancc, to nwdi­aLe statcments, to mcmorize knowledge and/ or to ofTer threads for cxplorations. I lowcver. a simple transfcr from extcrnal to internal representations can not be assumcd (Einsiedler, 1996). Prescnting knowlcclge structurcs is also insuflicicnt for J onasscn and Wang (I 99~), who call l'or active processing and working with thcm.

b) Charts, Maps, a11d Comfmters

Structured network prescntations arc ofspecial intercst in modern multimedia applirations. ThC)' meet the demands of' modern thcorics of mental represcnta­tions and at, the samc time, suit currcnt programming tcchniques. Programmcd moch tles con tain dctails, horizontal and ve rLical Connections arc coverccl b)' "links.'' Espccially h)'penext can replicatc thc semamic structure of a knowledge domain. 'oclcs rcprescnt terms and links indicate logical connections. Hypermedia appli­

cations can Iw secn as a refincment ol' hypcrt ext, integrating pictures, graphs and

Using Multimedia in SciencP Eclu.cation 87

animatcd visuals. They can also expand on dcmand, to show more detailed stru­ctures. An examplc is shown in Figure 9, categorising different sources of e lcctric energy and illustrating each type by a picture.

Ngwl' 9: Diffmnt So111rPs of l>ilxtrir f:'nngy. Siudmts May i\tll!lJI' /o CorreSJJ()IIding Pagt•s by Clid1ing on thP 'Ji•nns.

Flcxibility, adaptabil ity, and nctworking are spccial advantages of modern media in supporting thc structuring of knowlcdge. Drawing rnaps with modern computcr programs is vcry simple. Evcn begin ncrs can apply them and tailor learn­ing paths specifically lO their individual dcmands and intercst<> (Girwidz & Krahmer, 2002).

c) Advauce Organizer

Learning with hypermcdia oftcn Iacks appropriatc scaffolding and a process­orielllccl guidancc. Advance organizcrs may bc realizccl as a refcrcnce map ancl can assist goal-di rectcd learn i ng. At thc beginn i ng of' a leaming un it, such a framcwork can also help to understand the overall context. Furthermore, structurecl maps may avoid the so-called "lost in hypcrspace" synclromc.

Considering Cognitive Load

a) Limitation of Worlling Memory

ßaddclcy ( 1992) dcscribcd three subcomponents of' working memory: (i) The centrat exccutivc, which controls aucntion ancl integrales information from two subordinated systems, (ii) thc visual-spatial skctch pad which proccsses imagcs ancl (i ii) thc phonological loop, which eieals with acoustic in/()l'lnation . T he capacit.y of individual working memory scems to bc st.rictly li mited to less than seven so-called chunks (groupcd and orgtmized cntities of knowledgc learncd in the past). Howevcr, 1he info rmat.ion quanti ty grouped within a particttla r eh unk secms l.o be ncarly unlimited (Mi lle r, 195ß; ß addclcy, 1990).

88 Ginuidz, R. , Rubit:dw, T., Scltaal, S., and Bognn: F .'<.

b) Determiuauts for Cognitive Load

"Cogniti\'C Ioad theory" (Sweller, 1994) focuscs on the limi tations of working memory as an imponant f~lctor für lcarning (Chandler & s,,·cller, 1991). Cnder adversc cirnunstanccs, pcrceiving ancl processing information may rcquirc: c\·cn more cogniti\'c capacitics than undcrstanding the material iL~elf. For cxamplc, stuclying equations with unfamiliar notalions causcs a hcavy cognitivc Ioad due to

unknown expressions, which nccd intcrpreting (Lcung, Lo"·· & Swellcr. 1997). Any confrontation with an unl~uniliar coclc S)'!>tem is \ "Cl)' likcly to proclucc a hca'y strain on mental rcsources (Sccl & \\'inn, 1997) .

Aclclitionally, int e raction bctwecn learncrs and the operation of a computcr program comributc to cognitivc Ioad. Thus, coopcratin: cl iscon:ry lcarning with intcractive animations, for cxample, may produce high cognitivc Ioad, bccausc of thc need f"or sinlllltancously coordinating intcractions wi th pccrs (Schnotz, Höckhcle r, & Grzonclzicl, 1999) . Abo sul~ject mauer can o\'erstrain cogni tin~ rcsourccs if wo many e lcmenL~ have to be processcd in working memory. In ordcr to rcduce cogn iti\'c Ioad, singu lar clement~ should bc combincd w form mea­IJingrul units bcför(' further working. Marcus, Coopcr, and Swcllcr ( 1996) put this approach into concretc terms and cxplaincd that, for cxamplc, cliagrams can pro­viele such schemas for functional clepenclcncics in mathcmatics and scicncc.

The use of dill"erent sensor~· modalities normally reduccs cognitin· Ioad (Tindall-Forcl, Chandler, & Swcllcr, 1997), unlcss thc prescntation itself causcs a Ioad, q~., for nct-working auditory and visual information (Jeung, Chandler, & Swellcr, 1997). Supplemcntary visual information in teXL'\ may producc o,·erloacl br absorbing too many cogniti,-e rcsourccs of rhc visual information processing chan­nel, whereas additional acoustic information in computcr basccl emironmcnts may reduce cognitive Ioad (Kalyuga, Chancl ler, & Swellcr, 1998). Thc use of colours for guiding purposes ancl for indicating dctails in picwres ma~- be helpful. Additional written text, bowcvcr, may incrcase cognitive Ioad, if lcarncrs hm·e to switch bctween processing pictorial and textual information (Kalyuga. Chandler. & Swellcr, 1999).

Figure 10 illustrates a bird's strategy in cold cmironmenLS to minimiLc thc cnergy loss in their (un-insulatcd) legs. Thc animation is supplcmentcd by appro­priatc verbal explanations. Furthermore. dctails that arc not necessary für expla­nations arc r;tcled out. The temperaturc is displayccl ncxt to the veins in ordcr to avoicl a split-attention-effcct. Appropriatc colours inclicatc cold and warm areas ancl provick a \'isual imprcssion of tempcr<Jturc distribution.

38' c 34 c

24 c 21 c 15 c ' 13 c

a~·c 5-0 c

Figw1• 10: Hrdurtion ofCognitil•t•l.oad 1!_1 .\'amtlirm and Cdour Coding.

Using i\llullimNiia in Stience Erluration 89

"Cognitive Ioad" is adclit ionally innucnccd by a learncr's expcrtise. So, for novices any new information of a spccific knowledgc domain should be pre-orga­nizcd beforc imple mc ntation (Tuovincn & Swcller, 1999) . Howcver, if, for example, for cxperts a diagram is clear without textual cxplanatio n , redundant information in additionaltext would un ncccssarily increasc their cognitivc Ioad (Kalyuga PI al., 1998) .

Self-dirccted lcarning 111tl)' also causc cognitive ovcrload, if learncrs havc to

clccide which information is needccl next, or whcrc to find required information or how relevant spccific infonnation can be. Addit.ionally, t.cchnical proble ms or problcms with the uscr intcrface 111<1)' a ppcar.

c) Conb·ol lnformation Flow i11 Order to Adapt Cognitive Load

Swellcr (2002, 1994) suggcsted a classification or lcarning matcrials by taking into accOtnH whethcr within a working me mory informa tion may be processecl ste p by stcp or simultan cously. l'urthcnnovc, cognitivc Ioad strongly de pcncls 0 11 con­tcnt aspects. Thercfore, wc can o nly suggcst fcw gcncral principles to control the information now a nd t.o ac!just cognitive Ioad. Thr<'c approachcs a rc : (i) Allow users to co11 tro l the progress of work ancl admit adaptation to thdr individual requircments; ( ii) arrange information according to the principle or temporaland local contiguity, mcan ing that info rmation sh01Jid bc presen tt:d whcn and wherc it is ncccled. fo'or instan cc, text and pictures bclonging LOge ther shou ld not bc sepa­rattel (Mayt'r, 200 I); ( iii) thc "single concept principlc" that points to one single matter of f~1ct, tcrm, or conccpt. Diffe re nt aspcct~ can t.he bc trca ted in seq uc nce stcp by Slcp. In panicular, this can bc hc lpful fo r teach ing basic principles. Examples clcaling with fundamental principlcs conccrning waves are shown in Figure ll. On thc left hand a set of physica l tcrms is shown . Clicking on a tcrm starts playiug a short video clip to illttstrate th e unde rlying principle.

-w.;.. ..... - h-- .,..,.. Reflecllon 1 2 clrC\IIar wavee

Rellectlon 2 Spoclal w • .,.. ••• subjects

Sup•poeltlon o~:blellllt

StancUnewa ... Obetacle

Twom.dfa 1 Two medi•

Twomedla2 Dlep., .. on

Water••w• tnto Doppler .rr.et

/•lgttll' 11: tl l.üt of tlfi!ilimlion.v 01'1tling with I l'miP Pltnumtnw (tl) ond rm ürmtjJIP Joron tlrrmluatrrl Phtlllllllflton: lntrr-jJelll'lmliiiK Circular \\(n!f.5 / PrinrijJII' of Umlistw11t'rl Sttj'"fJOlilioll {B).

A Brief, Simplified Summary and Outlook

Usc multimodalit y, multicoding, and interact.ivity to fostcr deepcr lcarn ing, and

• assist cognitiw flcxibilit y by e nabling thc fl exible use or vario us reprcscnta­tions, change the superfic ial appearancc o f ol~jcct~. usc thc supplantation principlc to cstablish cognitivc COnnectio ns, ancllink d c tails using hypertcxt I hyperme dia

90 Cirwidz, R., Rubitzko, T, Sdwal, S., rmd Bognn; F. X

• hdp to build mental mocle ls, use multiple representations to activate exist­ing concepts, to construct and integrale components, to focus on special aspect~, or to illustrate the itlle rplay of settings

• use muhimedia to implement icleas of situatecl learning and anchored instruction , build Connections to authentic seuings, ancluse imeresting and challenging simulations and moclelling systems

• structure fi e lds of knowledge by using hypertext as ,,·eil as more ,; ·ual dis­plays, likc mind maps, concept maps, rcference maps or chans that modern media can make available in flexible ways

• be a'rare of the neccssity to avoid cogniti,·e o\·erload, implement features to control thc tlow of information, proviele plain user interfaces, and ofT e r fur­ther guiclance, perhaps even by using a workbook.

On the basis of the theoretical assumptions described, a specific multimedia programme ancl learning emironment was clesigned. Figures 7, 8, 10 are taken from these applicarions. See Girwiclz, Bogner, Schaal ancl Rubitzko (2006, present issuc).

References

BADOELEY, A. D. (1990). Human 1111'11101)': TheOI)' and pmrtire. Hillsclale, NJ: La\\Tence Erlbaum.

ßADOEI.EY, A. D. ( 1992). Working memory. Srimre, 255, 556-559.

ß.\LLST.\EDT, S. P. ( 1997). 1\'issensvmnilllung. [Transfer of knowledge] Wein heim: Psrchologie \'erlags Union.

ßEISSER, K L, j 01\ASSEN, D. H., &: GRABOWSKI, ß. L (1994). Using ancl sclecting graphic techn iques to acquirc structural knowleclge. Performance lmjm>vement Quarterly, 7(-1). 20-38.

ßRr\NSFORD,J D. , SHER\\"000, R. D. , HASSEI.BRINC, T S., KINZER, Cll., K., &: Wll.l.L\.\IS, S. M. (1990). Anchorcd instruc tion: vVII)' we neecl it and how technology can help. In D .. ix and R. Spiro (Eds.), Cognition, education and multimedia. (pp. 11 5-141). I!illsdale, [\{): Erlbaum Associates.

CIIA:-JDI.ER, P., & S\1-ELLER, J ( 1991). Cognitive Ioad theory and the format of instruction. Cognition and Tnstruction, 8(-1), 293-332.

CI.:\RK, R. E. ( 1 992) . Facilitating Domain-Ceneral Problem Sohing: Computers, Cognitive Processes and Insu·uction. In E. Oe Corte, M. C. Linn , H. Mandl ancl L. Verschaffel (Eds.), Computn~Based Learning Enuironments and Problem Solving ( ATO ASJ Series. Series F: Computer and Systems Sciences; 8-l) (pp. 265-285). Berlin: Springer.

CT GV ( 1993). Anchored insu-tiCtion ancl situatecl cognition re,·isited. Educational Terhnology, 33(3), 52-70.

DE j<>:-JG, T , & NJOO, M. (1992). Learning and lnstruction with Computer Simulations: Learning Processes lnvolved. In E. De Corte, .\1. C. Linn, H . .Vlandl and L. \'erschaffe! (Eds.), Computer-Based Learning Environments and Problem Solving (NATO ASJ Scries. Serics F: Computer and Systems Sciences; 84) (pp. 41 1-427) . Bcrlin: Springer.

DE Kt.EER, J , &: BRO\\~, J S. {1983). Assumplions and ambiguistics in mechanistic

Using Multimedia in Srimce Education 91

mental models. In D. Centner and A. L. Stevens (Eds.) , J\1/m /almode/s (pp. 155-190). 1-Iillsdale, N. J.: Erlbau m.

Ou·nu:, S. (1994). MP111rtle J\1/odelle: Kon.\·trulitP des Wissens und \lr>t:stehens. [Mmlal Nlodels: Worliing ModPl about Knowledge and Undet:standing] KognitionS/J!>)'clwloKisthP Grundlagen Jiir diP Sojtware-EtgonomiP. Cötti ngen, Stuttgart: Verlag für angc­wandte Psychologie.

EI NSIEDLER, W. (1996). Wissensstrukturicrung im Unterricht. [Structuring Knowledge in lessonsl Zeitschriftfür Piidagogili, 42(2), 167-192.

FORIIUS, K. D., & GENTNER, D. (1986). Learn ing p h ysical doma ins: Toward a lheo­retical framework. In R. S. Michalski, J. C. Carbonell ancl T. .M. Mitchell (Ecls.), Machine leaminK, an artijicial inlelligenrP aj)jJroach (Vol. 2) (pp. 311-348). Los Altos: Morgan &wfma nn Publishers.

CIR\\'II)Z, R., C öß\\'EI:-1, 0., & STEI:--JRÜCK, H.-P. (2000). Atomphysik a m Computer.[Atom Physics with Computers] Plt)'sik in unst'l1'1' ZPit, 31(4), 165-167.

CIRWm:t., R., & KRAliML::R, P. (2002). Lernpfade durch das WvVVI! m it Mapping­Programmen [Lcarning Paths throug h the WWW with Mapping Programs] . Unterricht Ph)•sili, 13(69), 11 -13.

COL.OMA~:-1, s. R. , PETR0$11'/0 , A. J., SIIER\\'OOD, R. D., G.-\RRIS0:-1, S., HICKEY, D., ßRA:-\SFORD,j. D., & PELLEGRINO, j. W. ( 1996). Anchoring science instruction in multimedia learning e nvironments. In S. Vosniadou, E. Dc Corte, R. Glaser and H. Mandl (Ecls.) , lnfern.ationalfJersjm:tives on the desigu o.f f('{:/t:nology-sufJjJorterl /earningenvimnmmts (pp . 257-284) . Mahwah , Nj: Lawrcnce Erl baum Associat:es.

Hi\RTEL, H. (1992). Neue Ansätze zur Da rstellung und Be handlung von Grundbegriffen und Grundgrößen der Elektrizit~itslehre [ ew Wars for Presenting and Treating Basic Concepts and Base ltems of Elecuicityj. In K. Oette und P . .J. Pahl ( H rsg.) , MultimPrlia, \/pm etzu.ng und Software Jii r die Lehre (S. 423- 428) . Berl in: Springer.

ISSING, L. J., & Ku MSA, P. (2002) . Multimedia- Eine Chance f'i'n· Information u nd Lernen [Multimedia- A C hance für Informalion and Learning]. In L..J. Issing und P. Klimsa (ITrsg.}, Information und Lernen mit J\ilullimedia. (S. 1-2) Weinheim: Psychologie Verlags Union.

J Et:NC:, H.:J., CHA!':DLER, P. , & SWELLER, .J. ( 1997). The role o f visual indicators in dual sensory mode instruction. Erlurafional P~Jclwlogy, 17(3), 329-343.

Jol JNSON-LAIRD, P. N. ( 1980) . Mental Mode ls in Cognitive Scie nce. CoKnitive SciencP, 4, 7 1-115.

J ONASSEN, 0., & W ANC, S. (1993). Acquiring structura1 knowledge from semantical­ly structured hypertext. Journal of ComjJuter-Based I nstrurtion, 20( 1 ), l-8.

KALYUC.\, S., CHA.\IDLER, P., & $\\'ELLER, .J. ( 1998) . Le,·els of expenise and insu·ur­tional design. H tmla 11 Faclors, 40( 1 ), 1-17.

KALYUCA, S., CI-IANDLER, P., & SWELLER, .J. ( 1999). Managing split-al tention and red u ndancy in rnultimedia instruction. AjJjJlied-Cognitivt'-Psydwlogy, 13(4), 35 1-37 1.

Koz~t.\, R. (2003). The material features of multiple represcnta tions and thcir cog­nitive and social a ffo rdances for science understandi ng. Leaming and fmtmction, 13, 205-226.

92 Ginoidr., R., Rubit::.ko. T , Sc/wal. S. , mul Bognn; F. X.

L\\"E, j. (1988). Cognition in Pmctire: Mind, mathematirs. and rulture in eve1)'rla)' life. Cambridge, CUP.

L WE, .J., & WEi':GER, E. ( 1990). Situated Learning: LegitimalP Pe1ijJhera/ Participation. Cambridge, UK: Cambridge Unive rsity Press.

Lt::UN<:, M., Low. R., & Sm : LL.ER, .J. ( 1997). Learning from equatio ns o r \\'Ords. Inslructional Science, 25( 1), 37-70.

M.\RCL'S. N .. CoOPER, .\1.. & SWEU.ER.J. (1996). Unclerstanding instructions.Jouma/ of Educational Psyrltology, 88( 1), 49-63.

M:\YER, R. E. ( 1997). Multimedia learning: are wc asking the rig bt question? E<lucational PsJchologist, 32, 1-19.

MAYER, R. E. (2001). Multimedia Lmrning. NcwYork: Cambridge Uni,·e rSi[)· Press.

MAYER, R. E. (2002) l\lulrimedia Learning . The Psydwlog;' of l .earning and Motivation, 41,85-1 39.

MAYER, R. E., & MORE '0, R. (1998). Asplit-atte ntio n efTect in rnultime dia learning: E\iclence for dual processing sy tems in working m e mory. Journal of Educational Psycholo[!J~ 90(2), 312-320.

MJLLF.R, C. A. ( 1956) . Tbc magic numbe r seve n, plus o r minus [Wo: Some Iimits on our capacit}' für proccssing info rmation. Psychological Review, 63, 81-97.

MORENO, R .. & Nl-\YER, R. E. (1999). Cog n itive principles o f multimeclia learning: Tb c rolc of modali[)' and contiguity effects. Journal of Educational Psychology, 91, 1-11.

Mous.w1, S.-Y., LO\\', R. , & SwELLFR, J. ( 1995). Rcclucing cognitive Ioad by mixing auditory and ,·isual prcsel1laLion modes . .Journal of Edurational Ps;·clwlog)'. 87(2), 31 9-334.

REIF, F. ( 1981). Tcacbing problem sol\'ing : A scie ntific a pproacb. Physics Tt'(lcher, 19. 310-316.

S.\LQ)J();-.;, C. { 1979). l nteraclion of mNI ia, cognition and learning. San Francisco: .Josscy-ßass.

SALOMOi\, C. ( 199-l) . lnferact ion of media, cognilion and learning. Hillsdale. New

J ersey: Erlbaum.

Sc:HNO'J'Z, W., & ßANNERT, M. (2003). Constructio n ancl ime rfe rence in learning fro m multiple re presentation. Learning rmd lnstruction, 13, 141- 156.

SCII:>:OTZ, v\'., BöCKJIELER, J.. & GRZO:>:DZlEL, I l. ( 1999) . Individual and co-o pera tive lcarning with ime ractiYe a nimated pictures. EumfJean Journal of Ps_)'cholog)' of Eduration, 1-1(2). 245-265.

ScHi':OTZ, W., 8 .\:>::--:ERT, .\IL. & SEUFI:: RT, T. (2002). Towards an lntegrati\'e \'icw of Text and Pierure Comprehcnsion: Visualization Eflect'i on the Construction of Me ntal Mode ls. In:J. Otero.J. A. Lcon and A. C. Graesser (Ecls.), Tlte psyclwlo­g)' of scimce Ir' XI l'Om.jJreltension ( pp. 385-4 16). llillsdale, 1\'j: Lawrc nce Erlbaum Associates.

SEEL, N. l\-1. (1986). Wissenserwe rb durch Mcclieu und "mentale Mode lle " [Ko nwl edge Acquisition by Means of tVlcclia a nd "\lental t\ lodcls'']. Unterrich tswissenschaft, -1. 38-1-401.

SEE I., N. M., & \Vt:--iN, \V. 0. ( 1997). Resea rch on media and learning: Distributed

Using Multimedia in Scienre Erluralion 93

cognilion and scmiotics. In R. 0. Tennyson ancl F. Schott (Eds.), lnsll'lftlional design: fnlemalionaljJPrsjJPrtiues, Vol. I: Theory, rcscarch, and models, (pp. 293-326). Mahwah , r-{J: Lawrence Erlbaum 1\ssociates.

SPIRO, R.J., COULSON, R. L., f ELTO\"ICH, P . .J. , & ANDERSOt'\, 0. K. {1988). Cognitive Flexibilit)' Theory: Ach·anccd Knowlcdgc Acqu isition in III-Stru ctu rcd Oomains. In V. Pa tel (Ed.), Tenth Annual Confcrence of the Cognilivc Scic ncc Society (pp. 375-383). Hillsdale, J J: Lawrcncc Eribaum Ass.

SPIRO, R. J., & JEtiNC, .J. ( 1990). Cognitivc flcxibil ity and hypertext: Theory and technology ror thc non-linear and multid intcnsional traversal or com plex sub­ject mauer. ltt 0. Nix ancl R. Spiro (Ecls.), Cognition, Erlucation, anrl i\llttliimnlia. (pp. 163-205). I lillsdale, NJ: Erlbaum.

SPIRO, R. J., fEL:I"O\'ICII , P . .J., jAC:OßSON, M.J. , & CüULSON, R. L. ( 1992). Cognitive Flexibility, Constructivism, and Hypertext: Random Access Instruction for Aclvanced Knowl-cclgc Acquisition in 111-Stntcturcd Oomains. In: T. Ouffy and D.Jonassen (Eds.), Construrliuism anrltlw Terlmolog)' of Inslruction: i\ ConTJersalion. (57-75). Hillsdalc N . .J.: Lawrence Erlbaum.

SPRI:'\CER, S. P. , & DEliTSCII, C. (1998). Linllf'S- rrrhles Grlt.im (üji - Righl ßmin]. Heidelberg: Spekt rum.

STEPP I, H. ( 1989). CtrJ: ComfmleJ~ßased 11-rtining. Plmmug; Design wul l~·utwirlllung

intera/(/ ivl'r l~t'm progmmme ( ComjJuler-ßasnl 1i-aini·ng. Plan n ing, Design aud DroelojJnwnl of l n/emtlivt' Vaming Pmgrams/. Stut:tgart: Klett.

SWELLER, .J. (199Ll) . Cognitive Ioad theory, lcarning diflicull)· and instructional dcsign. Learning anrl l nslmrlion, 4, 295-3 12.

SWELLER. J. (2002). Visualisation and lnstructional Design. In R. Ploetzner (Ed.}, Proceedill[fl of tlte fnlemational Wod,shop on Dynamir Visua/izalions and Learning. (pp. 150 1-15 10) Ti"rbingen: Knowleclge Media Research Center.

T!NDAI.I .-Fotm, S., Cl 11\NDI.ER, P., & S\\'EI.I .ER,J . ( 1997). \Vhen two scnsory modcs arc beuer than one . .Jou.J'JIId ofExjHtriuumtal Psytlwlof!,y: AfJjJlied, 3(4), 257-287.

TuovJ:\Et'\, .J.-E., & S\\' I·:I.I.ER, .J . (1999). A comparison of cogni tive Ioad associatcd with discovcry learning and workcd cxamples . .Jonmal ofEducalional P~yrholoro·.

91(2), 334-34 1.

II EUYELE:-< \ '.\:-.!, A. ( 1991). Learning to think likc a physicist: A review or rcscarch­bascd instructional strategies. A merica n.Jouma/ of Pltysirs, 59(19), 89 1-897.

WEIDENXtAN:-1, B. ( 1991). Lf!rnen mit ßildmedim: Psyrhologischt• 1111rf didahlisrlll' Cnmdlagm. Wtinhcim: ßeltz.

WEIDEN.\IANN, n. (2002). M ul ticodieru ng und Multi modalitüt im Lern prozcss [Mu ltiroding ancl Mtthitnoclality in Learning Proccsses] . In L..J. lssing, und P. Kli111sa (II rsg.) , lnj(mn"lion und Lnnl'll 'lllil Mu./ti.nlf'(/ia. (S. 45-62) Wein heim: Psycholog-ie VerlagstIn ion.