needs met through computer game play among adolescents

11
Needs met through computer game play among adolescents John Colwell * Department of Psychology, University of Westminster, 309 Regent St., London W1B 2UW, United Kingdom Received 2 November 2006; received in revised form 14 June 2007; accepted 22 June 2007 Available online 14 August 2007 Abstract Despite much research on video/computer game play, especially the relationship between such play and aggression, little attention has been directed at needs met through play. The aim of the present study was to extend Selnow’s (1984) needs’ scale, and to identify individual differences in needs among adolescents. Focus groups were conducted, followed by a content analysis, and items corresponding to emerging themes were added to reasons identified in previous research for use in a quantitative survey. Four factors emerged from a factor analysis of the results; ‘companionship’, ‘prefer to friends’, ‘fun challenge’, and ‘stress relief’, and these, with the exception of ‘companionship’, were found to be significant predictors of play. Ó 2007 Elsevier Ltd. All rights reserved. Keywords: Computer games; Electronic friendship; Mood regulation 1. Introduction The main research focus in computer game research within psychology has been on investigat- ing possible links with aggression, with comparatively little attention paid to reasons for play. However, Selnow (1984) adapted items from a ‘catalogue of needs’ scale, developed by Greenberg 0191-8869/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.paid.2007.06.021 * Tel.: +44 (0) 207 911 5000; fax: +44 (0) 207 911 5074. E-mail address: [email protected] www.elsevier.com/locate/paid Personality and Individual Differences 43 (2007) 2072–2082

Upload: john-colwell

Post on 11-Sep-2016

228 views

Category:

Documents


9 download

TRANSCRIPT

www.elsevier.com/locate/paid

Personality and Individual Differences 43 (2007) 2072–2082

Needs met through computer game play among adolescents

John Colwell *

Department of Psychology, University of Westminster, 309 Regent St., London W1B 2UW, United Kingdom

Received 2 November 2006; received in revised form 14 June 2007; accepted 22 June 2007Available online 14 August 2007

Abstract

Despite much research on video/computer game play, especially the relationship between such play andaggression, little attention has been directed at needs met through play. The aim of the present study was toextend Selnow’s (1984) needs’ scale, and to identify individual differences in needs among adolescents.Focus groups were conducted, followed by a content analysis, and items corresponding to emerging themeswere added to reasons identified in previous research for use in a quantitative survey. Four factors emergedfrom a factor analysis of the results; ‘companionship’, ‘prefer to friends’, ‘fun challenge’, and ‘stress relief’,and these, with the exception of ‘companionship’, were found to be significant predictors of play.� 2007 Elsevier Ltd. All rights reserved.

Keywords: Computer games; Electronic friendship; Mood regulation

1. Introduction

The main research focus in computer game research within psychology has been on investigat-ing possible links with aggression, with comparatively little attention paid to reasons for play.However, Selnow (1984) adapted items from a ‘catalogue of needs’ scale, developed by Greenberg

0191-8869/$ - see front matter � 2007 Elsevier Ltd. All rights reserved.doi:10.1016/j.paid.2007.06.021

* Tel.: +44 (0) 207 911 5000; fax: +44 (0) 207 911 5074.E-mail address: [email protected]

J. Colwell / Personality and Individual Differences 43 (2007) 2072–2082 2073

(1974) for watching television, for use in a questionnaire survey among 10–14 year olds. Factoranalysis produced five factors: ‘computer games preferred to friends’, ‘learn about people’, ‘com-panionship’, ‘action’, and ‘solitude/escape’, and all were found to correlate highly with an index ofarcade play. Selnow argued that computer games were providing ‘electronic friendship’ – and thatthey may be a substitute for real friends, a fear echoed by other commentators (e.g. Miller, 1993).The same scale was used by Colwell, Grady, and Rhaiti (1995) on 11–17 year olds. Support wasobtained for the ‘electronic friendship’ hypothesis, with a significant correlation between the‘games preferred to friends’ factor and play, but only for boys.

A later study by Colwell and Payne (2000) of 12–14 year olds also included Selnow’s scale. Aprincipal components analysis resulted in two factors. Factor 1, ‘companionship’, containeditems from Selnow’s ‘companionship’, ‘solitude’, ‘learn about people’, and ‘action’ factors –play seemed to be providing friendship and something to do whilst alone. Factor 2, ‘preferto friends’, mainly consisted of Selnow’s ‘computer games preferred to friends’ items. Both fac-tors correlated positively with play, but for boys only. A negative correlation between gameexposure and the number of good friends in class, again for boys, provided additional supportfor the ‘electronic friendship’ hypothesis. A replication in Japan among 12–13 year olds(Colwell & Kato, 2003) produced essentially the same factor structure. However, ‘companion-ship’ did not correlate with any of the play measures, and ‘prefer to friends’ correlated signif-icantly with frequency and duration of play, as well as with game exposure, for both boys andgirls.

Greenberg’s (1974) rationale was to base items on essays written by British schoolchildren onwhy they liked watching television, and adapted for use for video game play by Selnow (1984). Incontrast Kestenbaum and Weinstein (1985) adopted a Freudian conceptual framework, suggest-ing that games serve a homeostatic function, particularly for those who are experiencing conflictsassociated with entering adolescence. Supposed oedipal conflicts are likely to re-emerge duringthis period, and there may be limited opportunities for resolving resultant developmental stress.Therefore the aggressive content of games, instead of encouraging aggressive behaviour (presum-ably towards the father), may provide a fantasy arena for the safe expression of aggressive feel-ings. Also heavy players tended to play more when they felt tense, in order to relax, and thatthey were more likely to feel guilty after winning against their fathers.

McClure (1985) found that 14–18 year-olds liked games because they were ‘a challenge’ and‘fun’. However other reasons offered, and confirmed, were ‘good mood’ and ‘escape from pres-sure’, which bear some relation to the release of pent up energy identified by Kestenbaum andWeinstein (1985). Additional support for some of these reasons was obtained by Griffiths andHunt (1995), who also found that games were often chosen when ‘there was nothing else todo’, or ’because friends do’, and some players said that they found it difficult to stop playing.Wood, Gupta, Derevensky, and Griffiths (2004) found that enjoyment, excitement, relaxation,making friends, alleviating depression, and escape, differentiated between high and low frequencyplayers, but, as with the Griffiths and Hunt (1995) and McClure (1985) research, no rationale wasoffered for inclusion of these reasons.

The rationale for the present study was to identify reasons for play through focus groupdiscussions. This was followed by a questionnaire survey containing items to measure thosereasons, as well as reasons found in previous studies, which could then be subjected to factoranalysis.

2074 J. Colwell / Personality and Individual Differences 43 (2007) 2072–2082

2. Qualitative study

2.1. Method

2.1.1. ParticipantsFourteen group discussions were conducted in two primary (boys and girls 8–11 years old) and

three secondary schools (boys and girls 11–15 years old) in an outer London Borough.

2.1.2. Interview guideDiscussion centred around topics such as leisure activities, computer game play, game prefer-

ences, changes in play over time, and reasons for game play.

2.1.3. ProcedureChildren were recruited to take part in a discussion of ‘leisure time’ activities. Each focus group

contained five same sex children (as far as possible from the same friendship groupings to encour-age openness and honesty).

2.1.4. Data analysisThe data were subjected to a thematic analysis – the extraction of the common themes emerging

from the interviews.

2.2. Summary of results

2.2.1. Leisure activitiesAll children reported engagement in a wide range of activities, some of which centred on school,

for example soccer, and others which centred on home, and these included both more formalactivities such as instrument tuition, and informal activities such as watching television. Sponta-neous groupings of activities usually resulted in categories such as sports related or exercise, musiclessons and rehearsing, spending time with friends outside school, and a category sometimes giventhe varying descriptors of ‘screen’, ‘watching’, or ‘technology’. This included watching televisionand playing computer games, as well as (for some girls) chatting to friends on the telephone. Theywere always available when there was nothing else to do, or to relax when tired, or to obtain pri-vacy. For some younger boys, the attraction appeared to be the excitement of being involved in anaction game. As children grew older, and homework became more demanding, game play wassaid to be used as a kind of relaxing aside, or as an occasional escape, whilst appearing to be en-gaged in doing homework.

2.2.2. Amount of game playChildren had usually begun to play computer games on a sibling’s or friend’s machine, and this

had led to pressure on parents to buy equipment, usually in the form of a Christmas or birthdaypresent. Access was available at home to at least one game format for practically all children.However, there was evidence of clear gender and age differences in game play, with play amongyear 4–8 (up to 12 years old) children, especially boys, heavier. Less time was said to be spent ongame play by the time children reached year 9 (13–14 year olds), and especially year 10 (14–15

J. Colwell / Personality and Individual Differences 43 (2007) 2072–2082 2075

year olds). This age reduction in time spent in game play was due to other activities, such as home-work, or sports, which involved increased independence, and which provided opportunities forsocial interaction, and also paid employment.

2.2.3. Game preferencesChildren reported knowledge of, and experience of playing, a wide variety of games and game

types. Racing games (e.g. cars racing), puzzle games (brainteasers), role play games (the player cantake on another identity), shoot ‘em up games (shooting and killing using a variety of weapons)and beat ‘em up games (punching and/or kicking), sports games (e.g. golf simulations), and plat-form games (usually running and jumping onto platforms) were all mentioned. Girls expressed apreference for puzzle, role play, and platform games, whereas boys preferred games which in-volved action such as racing games, sports games, and beat and shoot ‘em ups.

2.2.4. Reasons for game play/experience of playingChildren were asked about both the positive and negative aspects of play – with the assumption

that those positive aspects would provide insight into motivations for play. Several distinct themesemerged. The first and most often mentioned dimensions were ‘for fun’, ‘enjoyment’, and ‘excite-ment’. This could be due to several game properties; for example the characters in the game, or thegraphics, or the opportunity to engage in fantasy play, to the extent sometimes of being able toengage in (imaginative) activities which were not possible in real life. This enjoyment could bethe challenge of playing with others, and thus the possibility of winning or being successful, thoughthis also raised one potential negative consequence – failure, which could be the cause of some frus-tration. There was some evidence too of game play being seen as something to do when a friend isnot available, or even to shut off the world. Alternatively the challenge may be that of a new game.

Also some games in particular had very straightforward goals of winning or losing which led tofeelings of success. Thus a degree of concentration was required – said to be a source of enjoy-ment, as were the feelings of control, mastery, and achievement when a required level of skillwas attained. The tension created by possible success or failure often led to strong arousal andexcitement. Also there was evidence that game play is regarded as credible social currency – com-puter game play was said to be cool.

Game play was also said to be used as a means of what may be termed mood alteration or ‘let-ting off steam’, following problems at school, or with friends (e.g. being left out), or with parents.This could involve for example ‘electronic aggression’ as a means of mood change (feeling better)after an occurrence of a negative nature, such a being bullied or a parental reprimand. Feelings ofanger, guilt, or frustration could be dissipated after some time spent in game play, and it was saidthat you could end up feeling much happier. It is as if troubles could be forgotten through absorp-tion in a game – participants were able to articulate this aspect of game play surprisingly well, andto resort to it when appropriate.

In contrast to the benefits of game play, very few negative aspects were mentioned. One con-cerned frustrations of various kinds; waiting for a game to be loaded, being called away in themidst of play, but especially losing at a certain level of difficulty in a hierarchical game, and havingto begin all over again. Another worry was the amount of time spent in game play, to the detri-ment of other activities. Indeed the compulsion to play was felt by some to almost become addic-tive, especially when trying to reach higher levels of a game.

2076 J. Colwell / Personality and Individual Differences 43 (2007) 2072–2082

2.3. Discussion

Most children appear to engage in a range of leisure activities, among which is a category re-ferred to as ‘screen’, ‘watching’, or ‘technology, which includes watching television and playingcomputer games. These were said to be available when seeking a relaxation activity, or when therewas nothing else to do, a reason chosen by participants in the paper by Griffiths and Hunt (1995),and included in Selnow’s (1984) needs’ scale, or when seeking privacy (also in Selnow’s scale).Prime reasons offered for game play were because they are ‘fun’ and ‘a challenge’ and this is con-sistent with findings reported by Griffiths and Hunt (1995) and by McClure (1985). An additionalreason because they are exciting, also reported by Griffiths and Hunt (1995), who also offered‘friends do’ as an option, chosen by 18% of participants. This social aspect of play ties in withthe current finding that play usually began with others, and social play can be an implicit partof the challenge created in a game, and that it is ‘cool’ to be seen as a player. This clearly relatesto the ‘companionship’ factor identified in earlier research (Colwell & Kato, 2003; Colwell & Pay-ne, 2000; Selnow, 1984). Games are also seen as an opportunity for fantasy and imaginative play–said to be why game play is fun.

The challenge reason may be associated with other motivations for play. A computer gameusually offers the opportunity for success in some form, and the feeling of control or masterycould cause players to feel like ‘winners’, a finding in line with results in the study by Kesten-baum and Weinstein (1985). This suggests a boost in self-esteem, which may relate to Domi-nick’s (1984) assertion that boys who are low in self-esteem would be attracted to game play.However, computer games also appear to be being used as a means of altering mood state.A feeling of upset-ness or anger (a negative mood), due for example to being ‘told off’ by aparent or ‘let down’ by a friend, can be changed to a more positive mood through game play.Good mood was given as a description of play in the study by Griffiths and Hunt (1995). AlsoMcClure’s (1985) study found that games were an escape from pressure, and the research byKestenbaum and Weinstein (1985) suggests that it was a way of reducing tension and becomingrelaxed. However, in addition to these functions, the suggestion here is that children andadolescents may deliberately choose to play, in the knowledge that they will feel better as aresult. Additionally, some concern was expressed over the sheer amount of time spent on play,along with the worry that play could become addictive, a reason identified by participants in thestudy by Griffiths and Hunt (1995).

The ‘prefer to friends’ factor reported in earlier research (Colwell & Kato, 2003; Colwell &Payne, 2000) did not seem to appear as an emerging theme. However, computer games were saidto be an opportunity for ‘privacy’, or to ‘shut the outside world out’, and these comments may besymptomatic of that factor. It is also possible that articulation of this factor may be very difficult,particularly in a group situation

The emerging themes, unless already included in Selnow’s needs’ scale, were translated intoscale questions as follows:

Playing computer games helps me to relax.I play computer games because I like to feel in control.I play computer games because I can use my imagination.Playing computer games gives me a chance to compete with my friends.

J. Colwell / Personality and Individual Differences 43 (2007) 2072–2082 2077

I play computer games because I like a challenge.Playing computer games lets your anger out.I like computer games which are thrilling and exciting.I like computer games because they are fun.I play computer games when I feel frustrated.I often play computer games when I get left out by my friends.Playing computer games is the best way to block off the world around you.When I do well at computer games it makes me feel like a winner.It is cool to play computer games.Playing computer games gets everything off my shoulders.I play computer games because I get involved in them.Often when I play computer games I find it difficult to stop.I like playing computer games when I’ve had a bad day.After I’ve played a computer game I feel everything is all right.

3. Quantitative study

3.1. Method

3.1.1. SampleThe sample consisted of 482 school students (257 boys and 225 girls), with age range 11–15

years old (school years 6–10), drawn from two primary and two secondary schools in an outerLondon Borough. Ten questionnaires were discarded due to a large amount of missing data,and this left a usable sample of 472 respondents.

3.1.2. QuestionnaireRespondents were asked to indicate games systems used and then to complete items to measure

frequency of play, and the duration of each play session. This was followed by 32 five point items(strongly disagree, disagree, not sure, agree, strongly agree) to measure needs fulfilled throughplay, consisting of 18 items derived from the qualitative stage, 13 items used by Colwell and Payne(2000) to measure friendship needs and one item on guilt (Kestenbaum & Weinstein, 1985) ‘play-ing computer games sometimes makes me feel guilty’.

3.1.3. ProcedureThe questionnaire was completed during a form period.

3.2. Results

3.2.1. Effects of sex and ageMales played significantly more frequently than girls, t (439) = 6.86, p < .001, and they played

for a significantly longer time on each occasion, t (439) = 5.84, p < .001. There was also a signif-icant negative correlation between age and total weekly play (frequency X duration, r = �.16,p < .001), although this is clearly accounted for by frequency of play (r = �.25, p < .001).

2078 J. Colwell / Personality and Individual Differences 43 (2007) 2072–2082

3.2.2. Needs’ scaleSome items were badly skewed, and could not be satisfactorily transformed. Therefore, follow-

ing recommendations by Fabrigar, Wegener, MacCallum, and Strahan (1999), principal axis fac-toring was employed, with direct oblimin rotation (all prior research had used orthogonalrotations), resulting in seven factors with eigenvalues greater than one. However, the scree plotsuggested a three or four factor solution. Examination of residuals from requesting three factorsrevealed a large number in excess of .05, and indeed some in excess of .1, and therefore it seemedlikely that there was an additional factor (Tabachnick & Fidell, 2001). A satisfactory (and moremeaningful) residuals solution was produced by extracting four factors, explaining 46% of the var-iance (Table 1).

Factor 1 (29.5% variance, alpha .77) included items (e.g. ‘like being with another person’,‘games are good company’) from the ‘companionship’ factor identified by Colwell and Payne(2000), and so that label was retained. The second factor (7.4% variance, alpha .75) consistedof items which reflected a preference for computer games compared to friends (e.g. ‘more excitingthan being with people’, ‘better than being with people’), and which corresponded closely withthose obtained by Colwell and Payne (2000). Therefore the label ‘prefer to friends’ was retained.Six items (e.g. fun, challenge, excitement, feeling cool) loaded on factor 3 (5.5% variance, alpha.79), which seemed to reflect elements of a positive challenge. The label ‘fun challenge’ was there-fore applied. Factor 4 (3.6% variance, alpha .77) consisted of five items (e.g. frustration, having abad day, letting anger out, shutting the world out). Taken together, these suggest using computergame play as a means of alleviating a negative emotional state. This factor was labeled ‘stressrelief’. All factors correlated positively with each other (Table 2), but the correlation between‘fun challenge and ‘prefer to friends’ factors was weaker. Regression scores for each factor wereretained for further analyses.

Table 1Factor analysis of the needs scale items

Item Factor 1 Factor 2 Factor 3 Factor 4

When I play computer games it is almost like being with another person .82Playing computer games helps me to forget I’m alone .59Computer games are good company to me .57Playing computer games is almost like being with a friend .57Playing computer games is more exciting than being with people .88Playing computer games is better than being with people .63It’s more fun to be playing computer games than playing with friends .61I like computer games which are thrilling and exciting .75I like computer games because they are fun .62I play computer games because I like a challenge .60Playing computer games makes me feel that I am part of the action .44I play computer games because I get involved in them .43It is cool to play computer games .42I play computer games when I feel frustrated .78Playing computer games lets your anger out .60I like playing computer games when I’ve had a bad day .58Playing computer games gets everything off my shoulders .47Playing computer games is the best way to block off the world .45

Table 2Factor correlation matrix

Companionship Prefer to friends Fun challenge Stress relief

Companionship .35 .37 .58Prefer to friends .22 .36Fun challenge .45

J. Colwell / Personality and Individual Differences 43 (2007) 2072–2082 2079

Correlations between the needs’ factors and frequency, duration, and total weekly play areshown in Table 3. All needs’ factors correlated positively and significantly with each of the gameplaying measures, suggesting that each factor may constitute a reason for playing computergames. However, the correlations for boys and girls separately show a differing pattern of results.The pattern for boys is similar to the overall pattern, whereas correlations for girls tend to be low-er for the ‘prefer to friends’ factor.

3.2.3. Multiple regression analysesStandard linear regression procedures were conducted in order to investigate how much vari-

ance was explained in each of the play measures (frequency, duration, and total weekly play).The independent variables entered in the equations were sex, age, and the needs’ factors. How-ever, age was excluded when the DV was duration of play, since the zero order correlation be-tween the variables was not significant. The results are shown in Table 4.

Five variables, sex, age, ‘fun challenge’, ‘stress relief’, and ‘prefer to friends’, were significantpredictors of frequency of play. The same variables, with the exception of age, were significantpredictors of total weekly play, and, with the exception of age and ‘stress relief, of duration ofplay.

3.3. Discussion

There are clear links between these results and past research. The ‘fun challenge’ factor corre-sponds to the two dimensions,’ for fun’, and ‘for a challenge’, found in the research by Griffithsand Hunt (1995), and also includes excitement, identified by Wood et al. (2004). However, theywere suggested as separate reasons, whereas in the present research they are part of the same fac-tor. The challenge and fun dimensions were also identified by McClure (1985), and the challenge

Table 3Zero order correlations between needs’ factors and game play measures

Whole sample Boys Girls

Frequency Duration Totalplay

Frequency Duration Totalplay

Frequency Duration Totalplay

Companionship .29*** .25*** .32*** .27*** .29** .34*** .24* .14 .21**

Prefer to friends .29*** . 29*** .34*** .29*** .30*** .38*** .17* .13 .21**

Fun challenge .38*** .34*** .43*** .31*** .28*** .36*** .31*** .28*** .34***

Stress relief .36*** .30*** .39*** .32*** .30*** .39*** .30*** .21** .29***

* p < .05.** p < .01.*** p < .001.

Table 4Multiple regression analyses: predicting frequency, duration, and total weekly play

Predictor Frequency of play Duration of play Total weekly playBeta unique variance Beta unique variance Beta unique variance

Sex �.20*** 3.3% �.15** 1.9% �.20*** .3.5%Age �.19*** 3.0% �.06Fun challenge .19*** 2.3% .20*** 2.7% .23*** 3.4%Stress relief .16** 1.0% .09 .14* 0.8%Companionship �.07 .00 �.03Prefer to friends .13** 1.2% .15** 1.7% .17*** 2.3%

R = .50 R2 = .24 R = .42 R2 = .17 R = .53 R2 = .27

* p < .05.** p < .01.*** p < .001.

2080 J. Colwell / Personality and Individual Differences 43 (2007) 2072–2082

dimension by Kestenbaum and Weinstein (1985), to be reasons for play, further supporting theirimportance. The results of the current research provide support for these reasons, but they alsopoint to other dimensions such as being cool, part of the action, and being involved, which arealso included in the factor. The McClure (1985) and Kestenbaum and Weinstein (1985) studiesalso found that computer games were played because they were an escape from pressure, andWood et al. (2004) that they were an escape, reasons which bear some similarity to the ‘stressrelief’ factor. However, evidence from the current research also suggests that play (frequency)can be used, perhaps deliberately, as a means of altering a negative mood to a more positiveone. This mechanism would seem to fit with Zillman’s (1988) concept of mood managementthrough media use, in which attempts are made to self-regulate emotions (e.g. Hoeksma,Oosterlam, & Schipper (2004)). Play may therefore be used as part of a ’balancing’ or ’homeo-static’ process in relation to mood. These two factors, ‘’stress relief’ and ‘fun challenge’ clearlyrelate to findings from previous research, but greater understanding is now provided of theirstructure and function. Both factors were predictors of frequency of play, but only ‘fun challenge’predicted duration of play.

The ‘prefer to friends’ factor predicted frequency and duration of play, and builds on previousresearch (Colwell & Kato, 2003; Colwell & Payne, 2000). The pattern of zero order correlationsbetween the factor and play measures was similar to that found with Japanese adolescents, in therelationship obtained for both male and female students, but it also bore some similarity to thatfound with the British sample, with correlations generally higher for boys than girls. The ‘com-panionship’ factor was similar to previous findings (Colwell & Kato, 2003; Colwell & Payne,2000), but again zero order correlations with play measures were somewhat different, in thatpositive correlations obtained for both male and female students. Colwell and Payne (2000) foundsignificant correlations for boys only and none were obtained by Colwell and Kato (2003) for theirJapanese sample. However, in regression equations (not conducted in previous research) ‘compan-ionship’ was not found to be a significant predictor of play.

The pattern of correlations between the factors would seem to make conceptual sense. Thestrong correlations between ‘stress relief’, ‘fun challenge’, and ‘companionship’ suggest that play-ing games which are exciting, a challenge, and which are fun, etc., and which have elements ofalmost like being with a friend, are all likely to contribute to altering mood state. Thus the ‘stress

J. Colwell / Personality and Individual Differences 43 (2007) 2072–2082 2081

relief’ factor appears to function as a way of coping with negative experiences, with game playfulfilling a mood regulation function. Those who play computer games primarily because theyare fun and a challenge are probably playing for more social reasons – play is ‘cool’, and, as such,an expression of group norms, and indeed part of the challenge may actually be competing withfriends. Therefore this kind of need satisfaction, ‘fun challenge,’ was found (and could be ex-pected) to be more weakly linked to play preferred to being with friends. In contrast, thosewho score high on the ‘prefer to friends’ factor may prefer game play (to real friends), and thepositive correlation with the ‘companionship’ factor is not surprising. Games fulfill needs morenormally found in friendship, and the strong correlation with ‘stress relief’ shows that such playcan also fulfill a homeostatic function. Therefore some support would appear to be provided forSelnow’s (1984) worry that, for some, game play may be providing ‘electronic friendship’.

Most conceptual work on computer game play within psychology has been directed at under-standing the effects of play, particularly of aggressive games, on behaviour. Competing theorieshave emerged. Catharsis theory (Feshbach & Singer, 1971) suggests that aggressive play will pro-vide a socially acceptable way of releasing aggressive tendencies within the context of a fantasyworld, thus making real aggressive behaviour less likely. Social learning theory (Bandura, 1986)suggests the opposite, as does the general aggression model (Anderson & Bushman, 2002), basedin part on social learning principles. The current research did not attempt to provide a test of thesetheories. However, the ‘stress relief’ factor would seem to bear some relation to the concept ofcatharsis, derived from Freudian theory, as is Kestenbaum and Weinstein’s (1985) notion of ten-sion release. Playing aggressive games was said to be a means of ’letting off steam’, and there wasno mention of feeling enhanced aggressive tendencies after such play, or evidence of a belief thatsuch play could increase aggressive behaviour.

Unlike much previous research, this study has begun with accounts of reasons for play, thusproviding a clear rationale for inclusion of specific dimensions in a subsequent questionnaire sur-vey. Factor analysis of the data resulted in four positively correlated factors. ‘Companionship’and ‘prefer to friends’ factors confirm, and therefore provide support for, the results of earlier re-search. Also ‘fun challenge’ and ‘stress relief’ factors are related to needs identified in previousresearch, but this research has provided more information on both the factor structure and therelationship between them. All factors, with the exception of ‘companionship’ were found to besignificant predictors of play.

Acknowledgements

Thanks to research assistants Mariana Bayley and David Sladen.

References

Anderson, C. A., & Bushman, B. J. (2002). Human aggression. Annual Review of Psychology, 53, 27–51.Bandura, A. (1986). Social foundations of thought and action: A social cognition theory. Englewood Cliffs, NJ:

Prentice Hall.Colwell, J., Grady, C., & Rhaiti, S. (1995). Computer games, self-esteem, and gratification of needs in adolescents.

Journal of Community and Applied Social Psychology, 5, 195–206.

2082 J. Colwell / Personality and Individual Differences 43 (2007) 2072–2082

Colwell, J., & Kato, M. (2003). Investigation of the relation between social isolation, self-esteem, aggression, andcomputer game play in Japanese adolescents. Asian Journal of Social Psychology, 6, 149–158.

Colwell, J., & Payne, J. (2000). Negative correlates of computer game play among adolescents. British Journal of

Psychology, 91, 295–310.Dominick, J. R. (1984). Video games, television violence, and aggression in teenagers. Journal of Communication, 34,

136–147.Fabrigar, L. R., Wegener, D. T., MacCallum, R. C., & Strahan, E. J. (1999). Evaluating the use of exploratory factor

analysis in psychological research. Psychological Methods, 3, 272–299.Feshbach, S., & Singer, R. D. (1971). Television and aggression: An experimental field study. San Francisco, CA:

Jossey-Bass.Greenberg, B. (1974). Gratifications of television viewing and their correlates for British children. In J. Blumler & E.

Katz (Eds.), The uses of mass communication (pp. 71–92). Beverley Hills, CA: Sage.Griffiths, M., & Hunt, N. (1995). Computer game playing in adolescence. Prevalence and demographic indicators.

Journal of Community and Applied Social Psychology, 5, 189–193.Hoeksma, J. B., Oosterlam, J., & Schipper, E. M. (2004). Emotion regulation and the dynamics of feelings: A

conceptual and methodological framework. Child Development, 75(2), 354–360.Kestenbaum, G. I., & Weinstein, L. (1985). Personality, psychopathology, and developmental issues in male adolescent

game use. Journal of the American Academy of Child Psychiatry, 24(3), 329–337.McClure, R. (1985). Age and computer game playing. Perceptual and motor skills, 61(1), 285–286.Miller, J. (1993). Computer games. Information Sheet. The Professional Association of Teachers.Selnow, G. W. (1984). Playing computer games: The electronic friend. Journal of Communication, 34, 148–156.Tabachnick, B. G., & Fidell, L. S. (2001). Using multivariate statistics. Allyn & Bacon.Wood, R. T. A., Gupta, R., Derevensky, J. L., & Griffiths, M. (2004). Video game playing and gambling in adolescents:

Common risk factors. Journal of Child and Adolescent Abuse, 14(1), 77–100.Zillman, D. (1988). Mood management through communication choices. American Behavioral Scientist, 31, 327–340.