Girls’ Video Gaming Behaviour and Undergraduate Degree Selection:

A Secondary Data Analysis Approach

Girls’ uptake of physical science, technology, engineering and mathematics

(PSTEM) degrees continues to be poor. Identifying and targeting interventions

for girl groups that are likely to go into STEM degrees may be a possible

solution. This paper, using a self-determination theory and self-socialisation

framework, determines whether one girl group’s, “geek girls”, video gaming

behaviour is associated with their choice of undergraduate degree by using two

secondary datasets: a cross-sectional study of the Net Generation (n = 814) and

the Longitudinal Study of Young People in England (LSYPE) dataset (n = 7342).

Chi-square analysis shows that girls who were currently PSTEM degree were

more likely to be gamers and engage in multiplayer gamers. Further, using

logistic regressions, girls who were heavy gamers (>9 hrs/wk) at 13-14 years

were found to be more likely to pursue a PSTEM degree but this was influenced

by their socio-economic status. Similar associations with boys and PSTEM

degrees was not found or weak. Therefore, girls were self-socialising or self-

determining their identity groups through gaming. This research can provide the

basis for whether encouraging gaming in adolescent girls can help them onto

PSTEM pathways.

Keywords: STEM; gender studies; video games; degree; higher education;

longitudinal studies

1 Introduction

The challenge of encouraging adolescent girls to enter higher education for studying

science, technology, engineering and mathematics (STEM) subjects particularly the

physical STEM (PSTEM)1 subjects has plagued both Western societies and educators as

there has not been any significant increase in STEM participation rate for girls in the

1 The term STEM includes two main fields: physical sciences (e.g. physics, computer science

etc) and the biological sciences (such as medicine, veterinary sciences, zoology etc).

Physical STEM (PSTEM) and Biological STEM (BSTEM) are used to distinguish these two

main fields (see for example McPherson, Banchefsky, & Park, 2018).

last decade (Smith, 2011; WISE, 2015). There is a recognition that this issue is multi-

factored. Much of the research around girls and STEM subjects have focused mainly on

how their socio-economic and parental background (Archer et al., 2012b; Cherney &

Campbell, 2011; Rozek, Svoboda, Harackiewicz, Hulleman, & Hyde, 2017), their

attitudes to STEM subjects (Barkatsas, Kasimatis, & Gialamas, 2009; Simpson, Che, &

Bridges, 2016) and their STEM aspirations (Archer et al., 2013; Levine, Serio,

Radaram, Chaudhuri, & Talbert, 2015) can affect their engagement.

Recent research has started investigated how stereotypes affect the STEM

participation of girls (such as by Archer et al., 2012a; Cheryan, Siy, Vichayapai, Drury,

& Kim, 2011; Master, Cheryan, & Meltzoff, 2016; Starr, 2018). STEM stereotypes are

mainly associated with persons who are male, geniuses, wear glasses and play video

games and who are sometimes referred to as geeks or nerds (Cheryan et al., 2011; Starr,

2018). Understanding how stereotypes can affect engagement may enable educators to

create more effective interventions for their students. However, there is less

understanding of how girls use current stereotypes to legitimise their participation in

STEM studies. Some studies have shown that girls legitimise their STEM interests

through harnessing the genius stereotype by emphasising their academic achievements

(Archer et al., 2012a) or the appearance stereotype by appearing less feminine (Ong,

2005). There is less research, however, on how girls use the video gaming stereotype to

legitimise their participation in STEM studies, although there are number of studies

focusing on how video gaming can encourage girls into STEM studies (Appel, 2012;

Feng, Spence, & Pratt, 2007; Gnambs & Appel, 2017). Therefore, this research paper

extends this field by investigating whether a particular stereotype, that of the “geek girl”

gamer, is associated with engagement in STEM, particularly PSTEM, subjects at the

higher education level.

1.1 The Geek Girl Stereotype

A geek means someone who has expertise in a certain field usually to do with

technology such as computer geeks, trivia geeks, gamers and hackers (McArthur, 2008).

Males have stereotypically been associated with geek terminology and are popularised

in the media as being good at PSTEM subjects (such as physics, engineering etc) as well

as being particularly interested in video games. Take, for example, the popular sitcom,

‘The Big Bang Theory’, the male actors portray geeks who are scientists in the PSTEM

subjects and who are video gamers. The female actors who portray scientists are mainly

in the biological sciences and are not gamers. In this media portrayal of scientists, there

is a clear distinction of the roles of male and female scientists and their predilection to

gaming, to the point where video gaming seems to occupy the male actors’ lives. Salter

and Blodgett (2012) explain that video gaming is a hypermasculine sub-culture where

‘hardcore’ gaming such as network/multiplayer games are the norm. Whilst there has

been an increase in gaming amongst females, the perception of their gaming has mainly

been around social or console-based games, for example, CandyCrush (IAB, 2014;

Tomkinson & Harper, 2015). Even so, Padilla-Walker, Nelson, Carroll, and Jensen

(2010) noted that girls were generally less likely to spend time on video games as well

as play less violent games. For these reasons, girls are seen as casual gamers, that is, not

having a complete time and energy commitment to their gaming. The hypermasculine

sub-culture view these females gamers as not ‘true’ geeks and they are sometimes

referred to as “geek girls”, “girl geeks” or “fake geeks” (see Tomkinson & Harper, 2015

for a discussion on the terminology). It is these geek girls that this paper is interested in,

in determining whether they are legitimising their participation in STEM through

gaming.

1.2 Self-socialisation and self-determination theory

Whilst media can shape or influence the identity of the adolescent, the adolescent may

seek out media that fits their evolving identity (see Coyne, Padilla-Walker, & Howard,

2013). Therefore, girls who are undertaking or intend to undertake a PSTEM degree are

probably more likely to conform to the media’s and society’s construction of the geek

by engaging in prolong periods of gaming and in particular hardcore gaming (or vice

versa: girls who are gamers may feel the need to conform to studying PSTEM subjects).

This is what Arnett (1995) refers to self-socialisation by media. Arnett (1995) explains

that an adolescent forms their identity through partly trying to emulate persons or

conceptualisations in the media and therefore, contributes to “the formation of

occupational aspiration”. Girls intending to do a PSTEM degree may then engage in

more video gaming to feel part of the PSTEM community and provide legitimacy to

their intentions (or vice versa).

This conception also aligns with the work of Deci and Ryan (2014) on self-

determination theory (SDT) which suggests a person’s intrinsic motivation are

dependent on three psychological needs: relatedness, competence and supportive

autonomy. Relatedness refers to the feeling of being part of a community, whilst

competence indicates adolescents feeling capable and confident in achieving their goal

whilst supportive-autonomy indicates students feel in control of the decisions they make

and are supported in these decisions (Deci & Ryan, 2014; Kasser & Ryan, 1996). SDT,

therefore, suggests that if an adolescent has a particular goal, such as doing a PSTEM

degree, then they are more likely to be intrinsically motivated if they feel part of the

community such as the PSTEM and video gaming community, having feelings of

competence such as engaging with PSTEM subjects successfully and supportive-

autonomy such as having the volition to select PSTEM subjects

2 Hypotheses

Therefore, for girls who plan on undertaking PSTEM degrees, playing video games

particularly hardcore video games that are representative of a science geek, may be a

way of forming their identity (or vice versa). Boys, conversely, may not have the

pressure of conforming to the hardcore video gamer stereotype when seeking to do

PSTEM degrees (or vice versa), as this is a legitimate domain of the adolescent boy.

Recent studies suggest that there is some merit to this argument about video games and

STEM degrees, for example, Turner (2014) found in her longitudinal cohort study that

students who played video games were more likely to go on to do a STEM degree and

confirmed that boys were also more likely to do a STEM degree. Whilst Lantz (2015)

noted in her cross-sectional study of undergraduate students that just under half felt that

playing video games influenced their choice of STEM majors. However, both Turner

and Lantz did not draw any association between the gaming intensity (i.e. time spent on

gaming) and type of games by gender for the different types of STEM degree

(biological sciences vs physical sciences). The degree type is an important distinction to

make as girls often select biological STEM (BSTEM) degrees over PSTEM degrees

(WISE, 2015). Further, for girls planning on studying a PSTEM degree and hence

conforming to the gamer identity, then the expectation is that for girls, their time spent

on gaming should increase over the years. Therefore, the research hypotheses for this

paper are:

1. Girls who play hardcore video games (gamer type) are more likely to do

a PSTEM degree

2. Girl’s gaming intensity (hours spend on gaming) is positively associated

with their likelihood of pursuing a PSTEM degree

3. An increase in girl’s gaming intensity (hours spend on gaming) will be

positively associated with their likelihood of pursuing a PSTEM degree

3 Design and Overview of Studies 1 and 2

The three research hypotheses are answered using a secondary data analysis approach.

Secondary data analysis is not a statistical or analytical approach, but rather it is a

methodological approach. In a secondary data analysis approach, the research uses data

that is already in existence such as in a repository. Secondary data analysis offers a way

of investigating research questions using larger datasets than which would be possible

for the researcher otherwise, in terms of resources. Secondly, it allows research data to

be used parsimoniously. However, secondary data has the issue of not always having

the exact research data that is needed for answering research hypotheses and sometimes

proxies must be used. In this paper, two secondary data sources are used. The first

research hypothesis (Study 1) is answered using a cross-sectional survey from a UK

Economic Social Research Council (ESRC) funded project on the Net Generation (see

Hosein, Ramanau, & Jones, 2010; Jones & Hosein, 2010) during the first year of their

university life. The second and third hypotheses (Study 2) are answered using a

longitudinal cohort study, the LSYPE (Longitudinal Study of Young People in England)

(see Anders, 2012) which collected data on the same adolescents from when they were

13/14 years to 19/20 years.

Both datasets have data related to gaming and degree type. However, the gaming

data differs in both datasets. The Net Generation dataset has information on the game

type (hardcore and softcore gaming) and degree choice. On the other hand, the LSYPE

dataset has the number of hours played (gaming intensity) when the participants were

13/14 years and their eventual degree choice at 18 years but does not have information

on the particular game type. The LSYPE dataset, also unlike the Net Generation dataset

has the advantage of having data on those adolescents who did not go to university.

Using these two datasets thus allow us to answer the research hypotheses related to the

type (Study 1) and intensity (Study 2) of games and provide insights into how they are

associated with degree choice for girls.

The theoretical framework of self-socialisation by media is applied to Studies 1

and 2, and the SDT framework is only applied to Study 2. Study 1 did not have any

variables that would correspond to the SDT framework.

3.1 Study 1: Net Generation Dataset – Gamer Type and Degree

The Net Generation data set is a cross-sectional survey of UK first-year students and

their technology use which was taken during the autumn of 2008, that is, when these

students first started their degree programme. The Net Generation dataset surveyed

students from five different universities in a range of modules which was used as a

proxy for degree programmes. These modules included biosciences, veterinary sciences,

computer science, sociology, accounting and general science. This was an opportunity

sample rather than a representative sample.

3.1.1 Selection and Coding of Variables

This dataset had only data variables to support Arnett’s theory of self-socialisation

where the media was the type of games. The following variables were selected from the

Net Generation dataset. This dataset did not have any data relating to ethnicity.

Gender: The sample was limited to students who were under 20 years (n = 814) where

333 were male students, in order for it to be comparable to LSYPE dataset which had

information on students when they were circa 19 years old.

Degree Programme: Degree programmes were coded into three categories. STEM

degree programmes were split into BSTEM (which included veterinary sciences)

degrees and PSTEM degrees. The third category of degree was coded as Social Sciences

and Humanities.

Gamer Type: In the Net Generation dataset, students were asked the extent they played

three types of games: multiplayer games, web games and mobile/console games (where

1= very frequently to 5 = not at all). Recognising that this scale may have some issues

of validity, the data were recoded into a binary code for gamer type. If the student

indicated not at all (code 5) they were recoded as “0: not a gamer” and if they provided

any frequency (code 1 to 4), they were recoded as “1: a gamer” for that game type (i.e.

multiplayer, web or mobile/console games). The data for the three types of games were

combined to determine whether the student played any game (1: a gamer) or no games

(0: not a gamer).

3.2 Study 2: LSYPE Dataset – Gamer Intensity and Degree

LSYPE is a longitudinal survey which collected data in seven waves for adolescents

from the age of 13/14 to 19/20 years in England, which ran from 2004 to 2010. The

LSYPE dataset initially had 15770 students during Wave 1. By Wave 7, this had

dropped to 8323 of which 4116 were girls. This time period covers during the middle of

secondary school to the first couple of years of university. Anders (2012) indicates that

one drawback of the LSYPE is that there is an oversampling of adolescents entering

higher education but it is still able to provide insights into university access. To

minimise the effect of oversampling, the analysis used a weighted sample (LSYPE

variable: W7_lsype_wt_incskip).

3.2.1 Selection, Derivation and Coding of Variables

LSYPE variables were selected to approximate the concepts of autonomy, competence

and relatedness from SDT and Arnett’s self-socialisation in media in order to determine

how these constructs affected the choice of degree made (see Table 1 for a list of

variables). Further, where possible the variables in Table 1 were matched to the

suggested variables proposed by All, Nuñez Castellar, and Van Looy (2016) for when

researching gaming. Although All et al. (2016) suggested including the variables of age

and current ability, these were not included as all the adolescents were of similar age

and there was no variable for current ability. The main outcome variable was:

Degree Programmes: Degree programmes were similarly grouped as in Study 1: the

Net Generation dataset. An additional category of “No degree” was included for those

adolescents who did not enter higher education.

3.2.1.1 Competence: SDT

Adolescents’ feelings of competence were approximated using two variables, one on

their past performance and the other on their subject self-concept. Whilst past

performance is not a measure of feelings, past performance is known to be a driver for

feelings of competence and self-efficacy (Sitzmann & Yeo, 2013) and is hence used as a

proxy. Self-concept is a measure of feelings of competence and it represents an

individual’s composite view of their mastery in a particular area (Bong & Skaalvik,

2003). Hence, the variables used to represent competence are:

Past Performance: Students’ past performance was determined by a national

examination that occurs when adolescents were circa 10 years old (Key Stage 2, KS2)

in the areas of English, Mathematics and Science.

Self-Concept: Self-concept was determined by the sum of two questions on the extent

they like/dislike a subject (1 = “Don’t like it at all to 4 = “Like it a lot”) and how good

they thought they were in the subject (1= “Not good at all” to 4 = “Very good”), which

was only measured during Wave 1. The two items for each subject were found to be

unidimensional using a principal component analysis and had moderate reliabilities for

Mathematics (Cronbach α = 0.65); English (Cronbach α = 0.69); Science (Cronbach α =

0.75) and Information Communications and Technology, ICT (Cronbach α = 0.76).

3.2.1.2 Supportive-Autonomy: SDT

Adolescents’ feelings of supportive-autonomy were not directly measured in the

LSYPE dataset. However, socio-economic variables that are known to negatively affect

the extent that students are supported onto PSTEM degrees such as ethnicity, social

deprivation and gender (see for example Archer et al., 2012b; Parker et al., 2012 with

regards to access and subject choice) are used as proxies for measuring students’

perceived supportive-autonomy within their culture and society. These variables are:

IDACI Score: The income deprivation affecting children index (IDACI) score was used

to measure deprivation which is based on the postcode. It ranges from 0 to 1, where 0 =

not living in a deprived area to 1 = living in a very deprived area

Ethnicity: Ethnicity was recoded from the LSYPE data into four categories, where 1 =

Other (such as mixed, Chinese), 2 = African or Caribbean, 3 = South Asian (such as

Indian, Pakistani) and 4 = White.

3.2.1.3 Relatedness: SDT

Adolescents’ feelings of relatedness to the gaming community were approximated based

on the number of hours they spent on gaming (intensity) and the change in gaming

intensity over a period of two years. Gaming community in this context does not refer to

a specific gaming community but rather the feelings arising from their identity and

notion of being a gamer (Salter & Blodgett, 2012). The variables used are:

Gaming Intensity: Adolescents’ level of gaming was measured only in Waves 1 and 2

in the LSYPE dataset. A derived variable of gaming intensity was created to measure

adolescent’s gaming behaviour which represented the number of hours spent per week

gaming which included computer, video and smartphone games. For Wave 1, the

derived gaming intensity variable was computed as the product of LSYPE variable:

W1hcomG (Number of days/week spent playing computer or video games) and

W1HcomG2 (Number of hours per day spent playing computer or video games) to

provide the total number of gaming hours/ week. As the responses for W1hcomG1 was

a range, for example, 3-4 days, the lower bound of the range was used for the product

which provided a more conservative calculation for gaming intensity. Following the

creation of the gaming intensity variable, it was further categorised visually through

examining a histogram of female adolescents at the Wave 1 data for participants and

grouping the number of gaming hours/ week based on where they clustered (see Figure

A.1 in Appendices). The female adolescents’ histogram was used as the benchmark for

the gaming intensity. Adolescents were then grouped as being a Non-Gamer (0

hrs/week); Light Gamer (1 to 3 hrs/week); Moderate Gamer (4 to 8 hrs/week) and

Heavy Gamer (>9 hrs) using the histogram.

Change in Gaming Intensity: Wave 2 gaming intensity was calculated similarly to

Wave 1 Gaming intensity and the adolescents were categorised into the same groupings

of gaming intensity as Wave 1. A change in the gaming intensity categories between

Wave 1 and Wave 2 was calculated, to determine whether the students had increased a

category in their gaming intensity (coded 1), decreased their gaming intensity (coded -

1), or stayed the same (coded 0). This variable was referred to as Change in Gaming

Intensity.

Table 1: Dependent variables included for determining the degree outcomes based on

the description from All et al. (2016)

VARIABLES IN LSYPE

RELATION TO SDT

VARIABLES IN ALL ET AL. (2016)

DESCRIPTION

PAST PERFORMANCE

Competence Past Performance Prior academic achievement. Students’ scores at 10 years of age for English, Maths and Science (KS2)

SELF-CONCEPT Competence Motivation Motivation towards the learning content. Measured using their self-concept of English, Science, ICT and Maths

GENDER Autonomy Gender Gender (male/female). Male and female students analysed separately

ETHNICITY Autonomy Socio-economic Status

Four ethnic groups

IDACI SCORE Autonomy Socio-economic Status

Level of deprivation based on postcode

GAMING INTENSITY

Relatedness Game experience Hours spent playing games. Measured using gaming intensity categories

CHANGE IN GAMING INTENSITY

Relatedness Game experience Longitudinal change in gaming intensity

4 Analysis and Results

The analysis and results of the hypotheses are presented based on the studies they

related to, Hypothesis 1 for Study 1 and Hypotheses 2 and 3 for Study 2.

4.1 Study 1 and Hypothesis 1: Net Generation Dataset – Gamer Type and Degree

To test hypothesis 1, chi-square analysis was used to determine the likelihood of female

and male gamers doing a PSTEM degree depending on their game type preference. The

chi-square analysis found that female gamers were more likely to do a PSTEM degree

(χ2(2)=15.65, p<0.01; Cramer’s V = 0.18) regardless of their gamer type: multiplayer

gamers (χ2(2)=13.03, p<0.01; Cramer’s V = 0.16), web gamers (χ2(2)=17.23, p<0.01;

Cramer’s V = 0.1) and device gamers (χ2(2)=11.09, p<0.01; Cramer’s V = 0.15).

Female students who were doing a BSTEM degree were the least likely to be a gamer

(see Table 2). This may suggest those female students who have a predisposition for

games may be more likely to do a PSTEM degree (failing to reject Hypothesis 1).

Table 2: Female gamer types and their degrees in the Net Generation dataseta

Degree Gamer Multiplayer

Web Device Total

Female p<0.01 p<0.01 p<0.01 p<0.01 p<0.01

BSTEM 77 (66%) 30 (26%) 52 (45%) 67 (58%) 116Social Sciences and Humanities

280 (81%) 144 (41%) 214 (62%) 253 (73%) 347

PSTEM 18 (100%) 11 (61%) 16 (89%) 15 (83%) 18Total 375 (78%) 185 (38%) 282 (59%) 335 (70%) 481

Male p=0.16 p<0.01 p=0.12 p=0.29BSTEM 26 (87%) 11 (37%) 19 (63%) 25 (83%) 30Social Sciences and Humanities 224 (94%) 153 (64%) 167 (70%) 213 (89%) 239

PSTEM 62 (97%) 53 (83%) 52 (81%) 60 (94%) 64Total 312 (94%) 217 (65%) 238 (71%) 298 (89%) 333

a: Percentages are based on row totals

Male students, on the other hand, appear to have a preference for multiplayer

games if they were doing a PSTEM degree (χ2(2)=19.65, p<0.01; Cramer’s V = 0.24)

but otherwise had a similar distribution of game types across all degree programmes.

4.2 Study 2: LSYPE Dataset – Gamer Intensity and Degree

To determine whether gaming intensity and change in intensity was related to the type

of degree (hypotheses 2 and 3), a multinomial logistic regression was performed for

each gender where the type of degree was regressed against gaming intensity, change in

gaming intensity, past performance, IDACI score, self-concept and ethnicity. Key

participant characteristics for LSYPE are presented in Table 3 and Table 4. As the

variables in the LSYPE are weighted, in some analyses, these results may be different

from the totals because of rounding errors.

Table 3: Ethnicity, Degree Type and Gaming Intensity Groups for LSYPE Sample

(weighted values)a

Female Male TotalEthnicityOther (Mixed, Chinese, etc.) 132 (4%) 130 (4%) 262 (4%)African or Caribbean 120 (3%) 87 (2%) 207 (3%)Asian (Indian, Pakistani etc.) 243 (7%) 233 (6%) 476 (6%)White 3229 (87%) 3164 (88%) 6393 (87%)DegreeBiological, Medical and Veterinary Sciences 399 (11%) 199 (5%) 597 (8%)PSTEM 135 (4%) 325 (9%) 460 (6%)Social Sciences and Humanities 1054 (28%) 740 (20%) 1794 (24%)No degree 2136 (57%) 2350 (65%) 4486 (61%)Wave 1 Gamer Intensity (Groups)Non-Gamer 1431 (38%) 548 (15%) 1978 (27%)Light Gamer 1617 (43%) 1376 (38%) 2993 (41%)Moderate Gamer 379 (10%) 710 (20%) 1088 (15%)Heavy Gamer 297 (8%) 980 (27%) 1277 (17%)Wave 2 Gamer Intensity (Groups)Non-Gamer 1963 (53%) 679 (19%) 2642 (36%)Light Gamer 1223 (33%) 1386 (38%) 2609 (36%)Moderate Gamer 285 (8%) 568 (16%) 853 (12%)Heavy Gamer 253 (7%) 980 (27%) 1234 (17%)N 3724 3613 7337

a: Percentages are based on column totals

4.2.1 Hypothesis 2: Gaming intensity and PSTEM degree

The overall multinomial regression for both boys χ2(48)=1365.82, p<0.01; Cox and

Snell’s pseudo R2 = 0.31) and girls (1366.50, p<0.01; Cox and Snell’s pseudo R2 =

0.31) were significant (see Table 5 and Table 6). Further, the Gamer Intensity categories

were also significant for both boys (χ2(9)=25.18, p<0.01) and girls (χ2(9)=26.41,

p<0.01). The results indicate those female students who were heavy gamers (> 9 hours)

were more likely to do a PSTEM degree (fail to reject Hypothesis 2). In particular,

female non-gamers in comparison to female heavy gamers were more likely to do a

BSTEM degree (2.5 times), Social Sciences (3.1 times) and No degree (4.1 times) than

a PSTEM degree. However, the male non-gamer in comparison to male heavy gamers

were more likely to do a Social Sciences degree over a PSTEM (1.8 times) but had a

statistically similar likelihood of doing a BSTEM or No degree.

Table 4: IDACI social deprivation score, self-concept and past performance of the

LSYPE sample (weighted values)

Female Male TotalM SD M SD M SD

IDACI 0.20 0.17 0.20 0.17 0.20 0.17KS2 Scores (Max 36)English 27.3 4.1 26.2 4.4 26.8 4.3Maths 26.5 4.7 27.2 4.9 26.9 4.8Science 28.4 3.6 28.7 3.7 28.5 3.6Self-Concept Scores (Max 8)English 6.2 1.3 5.9 1.3 6.1 1.3Maths 5.7 1.4 6.1 1.3 5.9 1.4Science 5.8 1.5 6.3 1.3 6.0 1.4ICT 5.9 1.6 6.3 1.4 6.2 1.5Game Intensity (hrs/wk) 3.9 5.8Wave 1 2.2 3.7 5.6 7.0 3.7 6.2Wave 2 1.8 3.9 5.7 7.4 0.20 0.17

In general, for both boys and girls, the results indicate in line with other studies

that ethnicity, deprivation, subject self-concept and past performance are major

contributors to the choice of degree. In particular, girls who had higher levels of social

deprivation (IDACI score) were significantly more likely to do a Social Sciences degree

(4.4 times) or No degree (58.3 times) than a PSTEM degree but this did not affect their

choice of a BSTEM degree over a PSTEM degree. For boys in high deprivation areas, a

similar pattern follows, with the likelihood of doing a Social Sciences degree, 3.1 times

and No degree 23.6 times. Further, the results suggest that girls from an Asian and

Other ethnicity versus White students were more likely to do a PSTEM degree than No

degree (13.8 and 3.9 times respectively). Mathematics self-concept appeared to be a

strong predictor for the selection of a PSTEM degree together with past performance in

mathematics. Girls and boys with a high mathematics self-concept were both more

likely to do a PSTEM degree than any other degree.

Table 5: Multinomial logistic regression predicting the type of degree for female

students based on gamer intensity (weighted values)BIOLOGICAL SCIENCES VS PSTEM 

SOCIAL SCIENCES VS PSTEM 

NO DEGREE VS PSTEM 

B SE OR B SE OR B SE ORETHNICITY**      OTHER (MIXED, CHINESE) -0.67 0.44 0.51 -0.56 0.38 0.57 -1.36 0.40 0.26**AFRICAN OR CARIBBEAN 1.52 1.04 4.59 1.24 1.03 3.45 -0.72 1.04 0.49ASIAN (INDIAN, PAKISTANI) -0.23 0.35 0.79 -0.48 0.33 0.62 -2.62 0.36 0.07**WHITE ref cat ref cat ref catIDACI** 0.86 0.83 2.35 1.48 0.78 4.38+ 4.07 0.77 58.32**KS2 SCORES      ENGLISH** 0.06 0.06 1.06 0.05 0.05 1.05 -0.10 0.05 0.91+MATHS** -0.05 0.05 0.95 -0.09 0.04 0.91* -0.17 0.04 0.85**SCIENCE* -0.07 0.06 0.93 -0.05 0.06 0.95 -0.16 0.06 0.86**SELF-CONCEPT      ENGLISH SC** -0.07 0.09 0.94 0.10 0.08 1.11 -0.09 0.08 0.92MATHS SC** -0.23 0.09 0.80* -0.32 0.08 0.72** -0.26 0.08 0.77**SCIENCE SC** 0.11 0.08 1.12 -0.12 0.07 0.89 -0.16 0.07 0.85*

ICT SC -0.05 0.07 0.95 0.01 0.07 1.01 -0.06 0.07 0.95GAMER TYPE*      NON-GAMER (0HRS) 0.93 0.39 2.54* 1.14 0.36 3.13** 1.41 0.36 4.11**LIGHT GAMER (1-3 HRS) 0.50 0.33 1.65 0.93 0.30 2.54** 0.90 0.30 2.46**MODERATE GAMER (4-8 HRS) 0.94 0.45 2.55* 1.34 0.42 3.83** 1.34 0.42 3.81**HEAVY GAMER (>9 HRS)

ref cat ref cat ref cat

CHANGE IN GAMER INTENSITY

     

DECREASE -0.14 0.37 0.87 -0.33 0.35 0.72 -0.24 0.35 0.79SAME -0.22 0.33 0.80 -0.33 0.31 0.72 -0.47 0.31 0.63INCREASE ref cat ref cat ref cat

+: p<0.1; *: p<0.05; **: p<0.01

Table 6: Multinomial logistic regression predicting the type of degree for male students

based on gamer intensity (weighted values)

BIOLOGICAL SCIENCES VS PSTEM 

SOCIAL SCIENCES VS PSTEM 

NO DEGREE VS PSTEM 

B SE OR B SE OR B SE ORETHNICITY**      OTHER (MIXED, CHINESE) 1.09 0.42 2.97** 0.10 0.39 1.10 -0.56 0.38 0.57AFRICAN OR CARIBBEAN 0.46 0.67 1.59 0.52 0.53 1.68 -1.30 0.54 0.27*ASIAN (INDIAN, PAKISTANI) 0.09 0.32 1.09 -0.34 0.26 0.71 -2.18 0.27 0.11**WHITE ref cat ref cat ref catIDACI** 0.91 0.68 2.47 1.13 0.53 3.09* 3.16 0.51 23.64**KS2 SCORES      ENGLISH** -0.02 0.05 0.98 0.03 0.04 1.03 -0.16 0.03 0.85**MATHS** 0.10 0.04 1.11* -0.01 0.03 0.99 -0.10 0.03 0.90**SCIENCE**

-0.16 0.06 0.85** -0.15 0.040.86*

* -0.17 0.04 0.85**SELF-CONCEPT      ENGLISH SC**

0.18 0.08 1.20* 0.18 0.061.19*

* -0.04 0.06 0.96MATHS SC**

-0.23 0.09 0.79** -0.31 0.070.73*

* -0.35 0.06 0.70**SCIENCE SC**

-0.11 0.08 0.90 -0.29 0.060.75*

* -0.33 0.06 0.72**ICT SC*

-0.17 0.07 0.84* -0.18 0.060.84*

* -0.16 0.06 0.85**GAMER INTENSITY**

     

NON-GAMER 0.18 0.37 1.20 0.56 0.28 1.75* 0.27 0.27 1.31

(0HRS)LIGHT GAMER (1-3 HRS) 0.07 0.24 1.07 0.27 0.19 1.31 -0.18 0.18 0.84MODERATE GAMER (4-8 HRS) 0.18 0.28 1.19 0.53 0.21 1.70* -0.07 0.20 0.94HEAVY GAMER (>9 HRS)

ref cat ref cat ref cat

CHANGE IN GAMER INTENSITY

     

DECREASE 0.14 0.28 1.14 0.30 0.21 1.34 0.18 0.20 1.19SAME 0.16 0.24 1.18 0.31 0.18 1.36+ 0.06 0.18 1.06INCREASE ref cat ref cat ref cat

+: p<0.1; *: p<0.05; **: p<0.01

4.2.2 Hypothesis 3: Increase in gaming intensity and PSTEM degree

The change in gaming intensity for boys and girls from Wave 1 to Wave 2 were also

examined to determine whether gaming intensity remained stable. The change in

gaming intensity for girls as they moved from Waves 1 to 2 did not affect their degree

choices (Table 5 and Table 6). For all girls, their gaming intensity fell from Wave 1 to

Wave 2 (rejecting Hypothesis 3), with over half of girls choosing not to play games and

increase of 15% (see Table B.1 in the Appendices). For boys, if their gaming intensity

stayed the same instead of increasing from Waves 1 to 2, they were more likely to do a

Social Sciences degree over a PSTEM degree (see Table B.2 in the Appendices).

Examining the change of gaming intensity from Wave 1 and Wave 2, both boys and

girls appeared to decrease their gaming by the same amount (around 32%, see Table B.3

and Table B.4 in the Appendices). However, this decrease in gaming varied depending

on the gamer. Almost three-quarters of girls who were moderate and heavy gamers

decreased their gaming intensity compared to only about half of the boys. Further, girls

who were heavy gamers and who eventually did a PSTEM degree had the lowest

decrease in gaming intensity. Interestingly, girls were more likely to remain as non-

gamers over this period in comparison to boys (73% vs 41%).

5 Discussion

This research aimed to determine, particularly for girls, whether there was a relationship

between being a geek girl gamer (one stereotype of the PSTEM student) and the choice

of a degree using the Net Generation and the LSYPE datasets. Firstly, the datasets

indicate that there is a low uptake of PSTEM degrees particularly for girls across both

samples of the Net Generation and LSYPE dataset. Further, the analysis indicates that

both female and male gamers were more likely to do a PSTEM degree. Both female and

male students who were doing or went on to do a PSTEM degree were more likely to

play multiplayer games and be heavy gamers (Hypotheses 1 and 2). However, this

relationship appears to be less strong for boys, as boys generally appear to play games

almost to the same extent regardless of their degree choice.

These findings appear to initially confirm both Arnett’s theory of self-

socialisation and SDT that girls would more likely engage with media that is a

representation of the PSTEM identity of being a geek. However, the LSYPE results

indicated that generally girls’ association with games dropped from Wave 1 (13/14

years) to Wave 2 (14/15 years) even for those who eventually go onto to do a PSTEM

degree but this was more variable in boys (therefore Hypothesis 3 is rejected). Based on

both Arnett’s theory and SDT, PSTEM girls’ gaming intensity should have increased

from Waves 1 to 2 as they engaged with the media that can shape their identity. There

are two possible explanations for this drop. Firstly, at the age of 13 to 14 years (during

Wave 1), in the British school system, students select the subjects that they would like

to study for the next two years. It is during this time students make their first decision

on whether they would like to study PSTEM subjects. For girls, who selected the

PSTEM subjects and were trying to fulfil society’s perception of the geek, perhaps they

felt they no longer had to play video games as they had now legitimised their status by

selecting PSTEM subjects. If this was true, there should be only a drop-in gaming

intensity for PSTEM girls only. However, the drop is seen across all girls, albeit slightly

more in PSTEM girls. Therefore, an alternative explanation is that girls are perhaps

more conscientious about their studies and are likely to drop any extraneous activities,

unlike boys. However, Rogers and Hallam (2010) did not note many differences in the

studying approaches between boys and girls who were nearing their General Certificate

of Secondary Education (GCSE) examinations (circa 16 years) but noted that girls were

more likely to spend time considering their work, such as reading over materials. Böö

(2014) found however that academic performance was negatively correlated with game

intensity and it may be as girls are generally more conscientious (Chamorro-Premuzic

& Furnham, 2009), they may have chosen to reduce their game time. These results

suggest more research is needed to understand how girls and boys prepare for

examinations, particularly as they may be disadvantaging themselves if one group

concentrates on examination outcomes beyond other activities that may make them a

well-rounded individual. It is unfortunate the LSYPE dataset did not measure the

gaming intensity beyond Wave 2, in order to determine whether the gaming intensity

increased after examinations.

The variables related to competence and autonomy were large influences in the

choice of a PSTEM degree. With regards to competence, students’ mathematics self-

concept affected the choice of degree for both boys and girls. Further, autonomy

associated variables, such as ethnicity and deprivation appeared to affect those students

who did not do a degree more strongly than other students. This confirms other studies

that students’ higher education outcomes are affected by the lack of autonomy

associated with students’ life circumstances. However, most interestingly, the analysis

indicates that the variable of relatedness (i.e. gaming intensity) was also strongly

associated with PSTEM degree. Researchers have previously demonstrated that gaming

had ill-effects on students’ risk behaviour (Padilla-Walker et al., 2010) and performance

outcomes but there were mixed results depending on gender particularly when prior

achievement was controlled (Böö, 2014; Burgess, Stermer, & Burgess, 2012; Walsh,

Fielder, Carey, & Carey, 2013). Further, there is some research that suggests there may

be merits to video gaming as it can make one more computer literate (Appel, 2012;

Gnambs & Appel, 2017). This research found that gaming intensity or gamer type

affected girls’ future educational outcomes positively in that they were more likely to go

on to do a PSTEM degree.

This may suggest, that heavy video gaming or multiplayer games that are

consumed by girls, should not be considered detrimental but rather be encouraged as it

can signpost an educator or a parent to direct girls to possible future PSTEM higher

education pathways. However, this approach raises the issues around whether the use of

stereotyping may be an appropriate and inclusive pedagogical approach as it may

further alienate girls who do not conform to the geek girl stereotype but who wish to go

into PSTEM degrees. It may also further stereotype or socialise girls into the idea that

girls who go onto to do PSTEM degrees have to be gamers. Hence a balanced but

cautious approach needs to be taken that inspires those girls who are already gamers

without alienating those who are not.

Perhaps an alternative approach may be that educators who want to encourage

girls into PSTEM degrees should implement more gamification in their teaching such as

multiplayer games. This approach can enable girls to embrace gaming positively and

encourage pathways to pursue PSTEM degrees. As gamification is on the increase

across the whole curriculum, its impact on girls and PSTEM will have to be monitored

(see for example Albuquerque, Bittencourt, Coelho, & Silva, 2017).

5.1 Limitations, implications and directions for future research

Whilst there are some limitations of this research, it opens up discussions for future

research as well as methodological issues. Firstly, regardless of whether girls self-

socialise or had a pre-disposition to gaming, there exists a relationship between girls’

gaming behaviour and that of doing a PSTEM degree. However, the question is how

educators can use this information for increasing girls’ participation in PSTEM subject

in an inclusive way. If we take the stance of self-socialisation, girls who do have a pre-

disposition towards gaming and think of themselves of geek girls, they probably could

be identified early by teachers/ parents and be encouraged to explore a PSTEM degree

pathway by connecting their gaming interest to their future employability such as

through invited talks from gaming experts or use of gaming in PSTEM subjects (see for

example Denner, Werner, & Ortiz, 2012). However, it is also important for girls who do

not want to engage with the geek culture to see more alternative female and male role

models highlighted during their education in schools and widening participation

contexts (such as science museums) to ensure that they feel less like an impostor when

they do not conform to the geek girl identity of a PSTEM person (Cheryan et al., 2011).

As there is still a limited number of female PSTEM role models that go against these

traditional stereotypes of geek girls, then in the short-term it is possible that educators

will have to encourage those girls who are gamers to pursue PSTEM degrees first whilst

working on a longer-term strategy of changing how girls perceive a PSTEM pathway.

Secondly, in our study, gaming intensity in the LSYPE study was only measured

during Waves 1 and 2, and therefore, it could not be determined whether students’

gaming intensity changed after their examination periods for students with different

degree programmes. Triangulating with the Net Generation data in which most students

were gamers, and with almost all students within the PSTEM degrees being a gamer,

the analysis suggests at some point, students, particularly girls, begin to re-engage with

games possibly after the pressure point of examinations. Future research should focus

on what types of activities both genders tend to terminate near examination periods,

when they begin to re-engage with the activities and the reasons why they began to re-

engage, as knowing these time points can affect their engagement with any planned

intervention activities, such as science camps or gamification.

Thirdly, the results indicate that whilst Arnett’s self-socialisation theory and

SDT can explain girls’ gaming behaviour, it is possible that girls may be engaging in

gaming for other reasons. Perhaps, girls who play certain types of games may be more

pre-disposed to studying degrees with problem-solving elements such as PSTEM

degree. For example, Adachi and Willoughby (2013) found that students (both boys and

girls) who play strategic games such as multiplayer games had better problem-solving

skills at secondary school level which indirectly led to better academic performance.

This may thus set a path for students to choose those subjects that they are interested in.

Further, Appel (2012) found that students who played games were more computer

literate which may predisposed students to computer-related degrees such as PSTEM. It

is likely that pre-disposition, self-socialisation and feelings of relatedness when it comes

to studying a PSTEM degree are all inter-related, and future research may want to

consider the relative importance of these in the choice of degree.

Finally, this research also raises a methodological issue. In the use of

longitudinal cohort studies, the research demonstrated that girls had changed their

gaming behaviour from Wave 1 to Wave 2 to the extent that their gaming intensity in

Wave 2 did not have any relationship to their eventual degree. It is uncertain whether

this change in behaviour is country-specific, that is dependent on the education system

or whether it is a behavioural change that occurs in girls regardless of country. Gnambs

and Appel (2017) noted in their national study of 14 to 15-year-old German adolescents

that around 28% were non-gamers which was the same average that was found for 13

to 17-year-olds in an American study (Lenhart, Smith, Anderson, Duggan, & Perrin,

2015). These figures are similar to LSYPE Wave 1 data but not for Wave 2 (27% vs

36%). If students’ examination periods cause these changes in behaviours, this will have

implications for cross-sectional surveys within any context where adolescents are facing

a major national external pressure such as examinations, as a survey of observed

behaviours can be affected. These observed behaviours, therefore, will not be localised

randomised effects but will occur across the whole sample. Therefore, it is important for

cross-sectional surveys to triangulate or cross-corroborate their variables with

longitudinal cohort studies to determine if they are representative of students over a

longer period when drawing implications for their study.

6 Conclusion

Two secondary large-scale datasets (one cross-sectional and the other longitudinal) have

confirmed that students who engage with video games, were heavy gamers or played

multiplayer games were more likely to study a PSTEM degree. These associations were

stronger in girls than in boys and therefore has implications for how we may want to

engage girls in PSTEM studies. Further, the research has raised methodological issues

of conducting cross-sectional surveys where national factors, such as national

examinations, can affect the observed behaviour from one year to the next for students.

7 References

Adachi, P. J. C., & Willoughby, T. (2013). More than just fun and games: The longitudinal relationships between strategic video games, self-reported problem solving skills, and academic grades. Journal of Youth and Adolescence, 42(7), 1041-1052. doi:10.1007/s10964-013-9913-9

Albuquerque, J., Bittencourt, I. I., Coelho, J. A. P. M., & Silva, A. P. (2017). Does gender stereotype threat in gamified educational environments cause anxiety? An experimental study. Computers & Education, 115(Supplement C), 161-170. doi:https://doi.org/10.1016/j.compedu.2017.08.005

All, A., Nuñez Castellar, E. P., & Van Looy, J. (2016). Assessing the effectiveness of digital game-based learning: Best practices. Computers & Education, 92–93(1), 90-103. doi:http://dx.doi.org/10.1016/j.compedu.2015.10.007

Anders, J. (2012). Using the longitudinal study of young people in England for research into higher education access. Retrieved from https://ideas.repec.org/p/qss/dqsswp/1213.html

Appel, M. (2012). Are heavy users of computer games and social media more computer literate? Computers & Education, 59(4), 1339-1349. doi:http://dx.doi.org/10.1016/j.compedu.2012.06.004

Archer, L., DeWitt, J., Osborne, J., Dillon, J., Willis, B., & Wong, B. (2012a). “Balancing acts'': Elementary school girls' negotiations of femininity, achievement, and science. Science Education, 96(6), 967-989. doi:doi:10.1002/sce.21031

Archer, L., DeWitt, J., Osborne, J., Dillon, J., Willis, B., & Wong, B. (2012b). Science aspirations, capital, and family habitus: How families shape children's engagement and identification with science. American Educational Research Journal, 49(5), 881-908.

Archer, L., DeWitt, J., Osborne, J., Dillon, J., Willis, B., & Wong, B. (2013). ‘Not girly, not sexy, not glamorous’: Primary school girls’ and parents’ constructions of science aspirations. Pedagogy, Culture & Society, 21(1), 171-194. doi:10.1080/14681366.2012.748676

Arnett, J. J. (1995). Adolescents' uses of media for self-socialization. Journal of Youth and Adolescence, 24(5), 519-533.

Barkatsas, A., Kasimatis, K., & Gialamas, V. (2009). Learning secondary mathematics with technology: Exploring the complex interrelationship between students’ attitudes, engagement, gender and achievement. Computers and Education, 52(3), 562-570.

Bong, M., & Skaalvik, E. M. (2003). Academic self-concept and self-efficacy: How different are they really? Educational Psychology Review, 15(1), 1-40. doi:10.1023/a:1021302408382

Böö, R. (2014). Video game playing, academic performance, educational activity, and motivation among secondary school students. Örebro University

Burgess, S. R., Stermer, S. P., & Burgess, M. C. R. (2012). Video game playing and academic performance in college students. College Student Journal, 46(2), 376-387.

Chamorro-Premuzic, T., & Furnham, A. (2009). Mainly openness: The relationship between the big five personality traits and learning approaches. Learning and Individual Differences, 19(4), 524-529. doi:https://doi.org/10.1016/j.lindif.2009.06.004

Cherney, I. D., & Campbell, K. L. (2011). A league of their own: Do single-sex schools increase girls’ participation in the physical sciences? Sex Roles, 65(9), 712. doi:10.1007/s11199-011-0013-6

Cheryan, S., Siy, J. O., Vichayapai, M., Drury, B. J., & Kim, S. (2011). Do female and male role models who embody STEM stereotypes hinder women’s anticipated success in STEM? Social Psychological and Personality Science, 2(6), 656-664. doi:10.1177/1948550611405218

Coyne, S. M., Padilla-Walker, L. M., & Howard, E. (2013). Emerging in a digital world: A decade review of media use, effects, and gratifications in emerging adulthood. Emerging Adulthood, 1(2), 125-137. doi:10.1177/2167696813479782

Deci, E. L., & Ryan, R. M. (2014). Autonomy and need satisfaction in close relationships: Relationships motivation theory. In N. Weinstein (Ed.), Human motivation and interpersonal relationships: Theory, research, and applications (pp. 53-73). Dordrecht: Springer Netherlands.

Denner, J., Werner, L., & Ortiz, E. (2012). Computer games created by middle school girls: Can they be used to measure understanding of computer science concepts? Computers & Education, 58(1), 240-249. doi:http://dx.doi.org/10.1016/j.compedu.2011.08.006

Feng, J., Spence, I., & Pratt, J. (2007). Playing an action video game reduces gender differences in spatial cognition. Psychological Science, 18(10), 850-855. doi:10.1111/j.1467-9280.2007.01990.x

Gnambs, T., & Appel, M. (2017). Is computer gaming associated with cognitive abilities? A population study among German adolescents. Intelligence, 61(Supplement C), 19-28. doi:https://doi.org/10.1016/j.intell.2016.12.004

Hosein, A., Ramanau, R., & Jones, C. (2010). Learning and living technologies: A longitudinal study of first-year students' frequency and competence in the use of ict. Learning, Media and Technology, 35(4), 403-418. doi:10.1080/17439884.2010.529913

IAB. (2014). Gaming revolution. Retrieved from https://iabuk.net/research/library/gaming-revolution

Jones, C., & Hosein, A. (2010). Profiling university students' use of technology: Where is the net generation divide? The International Journal of Technology Knowledge and Society, 6(3), 43-58.

Kasser, T., & Ryan, R. M. (1996). Further examining the American dream: Differential correlates of intrinsic and extrinsic goals. Personality and Social Psychology Bulletin, 22(3), 280-287. doi:10.1177/0146167296223006

Lantz, C. E. (2015). Women, gaming and STEM majors: Interest and motivation. (10799743 Ed.D.), University of Southern California, Ann Arbor. Retrieved from https://search.proquest.com/docview/2067420960?accountid=17256 ProQuest Dissertations & Theses Global database.

Lenhart, A., Smith, A., Anderson, M., Duggan, M., & Perrin, A. (2015). Teens, technology and friendships. Retrieved from http://www.pewinternet.org/2015/08/06/teens-technology-and-friendships/

Levine, M., Serio, N., Radaram, B., Chaudhuri, S., & Talbert, W. (2015). Addressing the STEM gender gap by designing and implementing an educational outreach chemistry camp for middle school girls. Journal of Chemical Education, 92(10), 1639-1644.

Master, A., Cheryan, S., & Meltzoff, A. N. (2016). Computing whether she belongs: Stereotypes undermine girls’ interest and sense of belonging in computer science. Journal of Educational Psychology, 108(3), 424-437. doi:10.1037/edu0000061

McArthur, J. A. (2008). Digital subculture: A geek meaning of style. Journal of Communication Inquiry, 33(1), 58-70. doi:10.1177/0196859908325676

McPherson, E., Banchefsky, S., & Park, B. (2018). Using social psychological theory to understand choice of a pstem academic major. Educational Psychology, 1-22. doi:10.1080/01443410.2018.1489526

Ong, M. (2005). Body projects of young women of color in physics: Intersections of gender, race, and science. Social Problems, 52(4), 593-617. doi:10.1525/sp.2005.52.4.593

Padilla-Walker, L. M., Nelson, L. J., Carroll, J. S., & Jensen, A. C. (2010). More than a just a game: Video game and internet use during emerging adulthood. Journal of Youth and Adolescence, 39(2), 103-113. doi:10.1007/s10964-008-9390-8

Parker, P. D., Schoon, I., Tsai, Y.-M., Nagy, G., Trautwein, U., & Eccles, J. S. (2012). Achievement, agency, gender, and socioeconomic background as predictors of postschool choices: A multicontext study. Developmental Psychology, 48(6), 1629-1642. doi:10.1037/a0029167

10.1037/a0029167.supp (Supplemental)Rogers, L., & Hallam, S. (2010). Gender differences in perceptions of studying for the

gcse. International Journal of Inclusive Education, 14(8), 795-811. doi:10.1080/13603110902721654

Rozek, C. S., Svoboda, R. C., Harackiewicz, J. M., Hulleman, C. S., & Hyde, J. S. (2017). Utility-value intervention with parents increases students’ STEM preparation and career pursuit. Proceedings of the National Academy of Sciences, 114(5), 909-914. doi:10.1073/pnas.1607386114

Salter, A., & Blodgett, B. (2012). Hypermasculinity & dickwolves: The contentious role of women in the new gaming public. Journal of Broadcasting & Electronic Media, 56(3), 401-416. doi:10.1080/08838151.2012.705199

Simpson, A., Che, S. M., & Bridges, W. C. (2016). Girls’ and boys’ academic self-concept in science in single-sex and coeducational classes. International Journal of Science and Mathematics Education, 14(8), 1407-1418. doi:10.1007/s10763-015-9676-8

Sitzmann, T., & Yeo, G. (2013). A meta-analytic investigation of the within-person self-efficacy domain: Is self-efficacy a product of past performance or a driver of future performance? Personnel Psychology, 66(3), 531-568. doi:10.1111/peps.12035

Smith, E. (2011). Women into science and engineering? Gendered participation in higher education STEM subjects. British Educational Research Journal, 37(6), 993-1014. doi:10.1080/01411926.2010.515019

Starr, C. R. (2018). “I’m not a science nerd!”:STEM stereotypes, identity, and motivation among undergraduate women. Psychology of Women Quarterly, 0(0), 0361684318793848. doi:10.1177/0361684318793848

Tomkinson, S., & Harper, T. (2015). The position of women in video game culture: Perez and day's twitter incident. Continuum, 29(4), 617-634. doi:10.1080/10304312.2015.1025362

Turner, A. J. (2014). Play to pay?: Adolescent video game play & STEM choice. In L. Robinson, S. R. Cotten, & J. Schulz (Eds.), Communication and information technologies annual (Vol. 8, pp. 55-71): Emerald Group Publishing Limited.

Walsh, J. L., Fielder, R. L., Carey, K. B., & Carey, M. P. (2013). Female college students’ media use and academic outcomes: Results from a longitudinal cohort study. Emerging Adulthood. doi:10.1177/2167696813479780

WISE. (2015). Women in science, technology, engineering and mathematics: From classroom to boardroom. Retrieved from Bradford, UK: https://www.wisecampaign.org.uk/uploads/wise/files/WISE_UK_Statistics_2014.pdf

8 Appendices

Appendix A. Figures

0 1 2 3 4 5 6 7 8 9 10 12 15 20 25 30 35 400

200

400

600

800

1000

1200

1400

1600

Girls' Gaming Hours/ Week

Freq

uenc

y of

Girl

s

Figure A.1: Histogram of girls’ gaming hours/week for Wave 1 in the LSYPE

Appendix B. Tables

Table B.1: LSYPE Female gaming behaviour in Wave 1 and Wave 2 (weighted values)a

FEMALE BIOLOGICAL SCIENCES

PSTEM SOCIAL SCIENCES AND HUMANITIES

NO DEGREE

TOTAL

WAVE 1 (13-14 YEARS) (P<0.01)B NON-GAMER (0 HOURS)

149 (37%) 36 (26%) 368 (35%) 878 (41%) 1431 (38%)

LIGHT GAMER (1-3 HRS) 174 (44%) 68 (50%) 501 (48%) 875 (41%) 1618 (43%)

MODERATE GAMER (4-8 HRS)

41 (10%) 11 (8%) 114 (11%) 213 (10%) 379 (10%)

HEAVY GAMER (>9 HRS) 35 (9%) 21 (15%) 70 (7%) 171 (8%) 297 (8%)

TOTAL 399 136 1053 2137 3725

WAVE 2 (14-15 YEARS) P=0.70B

NON-GAMER (0 HOURS)

216 (54%) 67 (50%) 540 (51%) 1140 (53%) 1963 (53%)

LIGHT GAMER (1-3 HRS)

130 (33%) 49 (36%) 367 (35%) 677 (32%) 1223 (33%)

MODERATE GAMER (4-8 HRS)

32 (8%) 9 (7%) 78 (7%) 166 (8%) 285 (8%)

HEAVY GAMER (>9 HRS)

21 (5%) 10 (7%) 69 (7%) 153 (7%) 253 (7%)

TOTAL 399 135 1054 2136 3724

a: Percentages based on column totals; B: χ2 probability for degree by gamer

Table A.2: LSYPE male gaming behaviour in Wave 1 and Wave 2 (weighted values)a

MALE BIOLOGICAL SCIENCES

PSTEM SOCIAL SCIENCES AND HUMANITIES

NO DEGREE

TOTAL

WAVE 1 (13-14 YEARS)(P<0.01)B NON-GAMER (0 HOURS)

22 (11%) 30 (9%) 95 (13%) 401 (17%) 548 (15%)

LIGHT GAMER (1-3 HRS) 83 (42%) 133 (41%) 301 (41%) 859 (37%) 1376 (38%)

MODERATE GAMER (4-8 HRS)

43 (22%) 67 (21%) 183 (25%) 416 (18%) 709 (20%)

HEAVY GAMER (>9 HRS) 51 (26%) 95 (29%) 161 (22%) 674 (29%) 981 (27%)

TOTAL 199 325 740 2350 3614

WAVE 2 (14-15 YEARS) (P<0.01)B

NON-GAMER (0 HOURS)

29 (15%) 47 (14%) 135 (18%) 468 (20%) 679 (19%)

LIGHT GAMER (1-3 HRS)

86 (43%) 124 (38%) 311 (42%) 864 (37%) 1385 (38%)

MODERATE GAMER (4-8 HRS)

32 (16%) 65 (20%) 124 (17%) 347 (15%) 568 (16%)

HEAVY GAMER (>9 HRS)

51 (26%) 89 (27%) 170 (23%) 671 (29%) 981 (27%)

TOTAL 198 325 740 2350 3613

a: Percentages based on column totals; B: χ2 probability for degree by gamer

Table B.3: The change in girls' gaming intensity from Waves 1 to 2 for the LSYPE

(weighted values)a

Biological Sciences PSTEM

Social Sciences and Humanities No degree Total

Non-gamerDecrease 0 (0%) 0 (0%) 0 (0%) 0 (0%) 0 (0%)Same 116 (78%) 28 (78%) 277 (75%) 623 (71%) 1044 (73%)Increase 33 (22%) 8 (22%) 92 (25%) 254 (29%) 387 (27%)Sub-Total 149 36 369 877 1431

Light GamerDecrease 72 (41%) 29 (43%) 205 (41%) 386 (44%) 692 (43%)Same 80 (46%) 33 (49%) 229 (46%) 370 (42%) 712 (44%)Increase 22 (13%) 5 (7%) 67 (13%) 119 (14%) 213 (13%)Sub-Total 174 67 501 875 1617

Moderate GamerDecrease 33 (80%) 8 (80%) 80 (70%) 153 (71%) 274 (72%)Same 4 (10%) 1 (10%) 19 (17%) 37 (17%) 61 (16%)Increase 4 (10%) 1 (10%) 15 (13%) 24 (11%) 44 (12%)Sub-Total 41 10 114 214 379

Heavy GamerDecrease 28 (80%) 14 (67%) 56 (80%) 130 (76%) 228 (77%)Same 7 (20%) 7 (33%) 14 (20%) 40 (24%) 68 (23%)Increase 0 (0%) 0 (0%) 0 (0%) 0 (0%) 0 (0%)Sub-Total 35 21 70 170 296

All GirlsDecrease 133 (33%) 51 (38%) 341 (32%) 669 (31%) 1194 (32%)Same 207 (52%) 69 (51%) 539 (51%) 1070 (50%) 1885 (51%)Increase 59 (15%) 14 (10%) 174 (17%) 397 (19%) 644 (17%)Total 399 134 1054 2136 3723

a: Percentages are based on column sub-totals

Table B.4: The change in boys' gaming intensity from Waves 1 to 2 (weighted values)a

Biological Sciences PSTEM

Social Sciences

and Humanities No degree Total

Non-gamerDecrease 0 (0%) 0 (0%) 0 (0%) 0 (0%) 0 (0%)Same 11 (50%) 14 (47%) 49 (52%) 148 (37%) 222 (41%)Increase 11 (50%) 16 (53%) 46 (48%) 253 (63%) 326 (59%)Sub-Total 22 30 95 401 548

Light GamerDecrease 12 (14%) 25 (19%) 51 (17%) 197 (23%) 285 (21%)Same 49 (59%) 67 (50%) 163 (54%) 396 (46%) 675 (49%)Increase 22 (27%) 42 (31%) 86 (29%) 266 (31%) 416 (30%)Sub-Total 83 134 300 859 1376

Moderate GamerDecrease 21 (48%) 30 (45%) 103 (56%) 215 (52%) 369 (52%)Same 7 (16%) 12 (18%) 39 (21%) 76 (18%) 134 (19%)Increase 16 (36%) 25 (37%) 41 (22%) 126 (30%) 208 (29%)Sub-Total 44 67 183 417 711

Heavy GamerDecrease 29 (57%) 47 (49%) 82 (51%) 345 (51%) 503 (51%)Same 22 (43%) 48 (51%) 79 (49%) 329 (49%) 478 (49%)Increase 0 (0%) 0 (0%) 0 (0%) 0 (0%) 0 (0%)Sub-Total 51 95 161 674 981

All BoysDecrease 62 (31%) 102 (31%) 236 (32%) 757 (32%) 1157 (32%)Same 89 (45%) 141 (43%) 330 (45%) 949 (40%) 1509 (42%)Increase 49 (25%) 83 (25%) 173 (23%) 645 (27%) 950 (26%)Total 200 326 739 2351 3616

a: Percentages are based on column sub-totals