The games industry in the United Kingdom was valued at almost £8B in the latest market valuation report (Purdie, 2024), with over £1.5B in mobile game sales in 2023. A 2024 survey of over 3600 UK participants aged 16+ found that 36% of female respondents (and 34% of respondents overall) had played mobile games, up from 20% in 2019 (Clement, 2025). Although the number of women playing video games continues to increase, their representation varies between different genres, platforms, and gaming activities (e.g., Le Ngoc, 2024). Previous research, as outlined below, has suggested multiple barriers to women’s play and participation in video game culture. However, there are few studies that provide evidence for these barriers in a broad sample of UK women who play mobile games.

A brief note on terminology: in relation to our study, we use “women” to mean people who identify as women. When discussing other authors’ work, we use “men,” “women,” and “people of other genders” to refer to those people who were designated that way in the context of their studies; similarly, we use “female” and “male” in accordance with how they are used by those other authors. Not all studies offer definitions of the terms used and may use them differently from each other.

Barriers to Play

Casual Vs. Core Games

One reason that women might be reluctant to consider themselves ‘gamers’ and fully participate in the gaming community is that the sorts of games that they are assumed to play, and that are marketed towards them, are viewed as less central, less serious, and less important than those chosen by and designed for men. There is a commonly drawn distinction within the gaming community between ‘causal games’ and ‘core games’ (or ‘games’ simpliciter) (Chess & Paul, 2019; Paaßen et al., 2017), along with a “subtle devaluation and blatant feminization (Vanderhoef, 2013, p. 4) of the former. Industry efforts to separate the two markets “have worked their way into the gaming subculture, furthering the stratification of gamers by gamers” (Hanford, 2018). The Global Gamer Study of 2023 found that women players are more likely to identify as casual gamers than men (44% vs. 28%, Le Ngoc, 2024, see e.g., Shaw, 2011), though there are some casual genres that seem to be played equally by men and women, such as puzzle games (Eklund, 2016).

How best to cash out the distinction is contested, but casual games are often free or cheaper than their ‘core’ counterparts, can be played in shorter bursts, are thought to be easier to learn, and can be (and often are designed for) lower-powered hardware such as mobile phones (Juul, 2012; Ramirez, 2015). It is commonly assumed that women play casual games rather than core games (Chess, 2016; Eklund, 2016; Juul, 2012; Paaßen et al., 2017), and the “perception of femininity in the casual games market guides the design of casual games” (Chess, 2016, p. 108). Those video games marketed towards women often involve pragmatic play (Chess, 2009) or emotional labour, constructing women “as being social- or care-centric players” (Chess, 2017, p. 96). Yet, these games are often devalued by the wider video game industry and gaming community. For instance, Chess (2017) notes of FarmVille, a popular casual game with social mechanics, that it:

[W]on “Game of the Year” in 2010 at the Game Developers Association Conference, yet the video game industry has been primarily indifferent or outright insulting regarding this feat… It would seem that just as affective [labour] is devalued in the real world, so is it devalued in the gaming world and as a form of play. (p. 118)

Attitudes: Violence, Invisibility and Hostility

As suggested above, women’s attitudes to gaming derive in part from their relation to their perception of games through advertising. Women are often an “invisible audience” for video games and not considered the primary market for core games. This invisibility affects how they view video games: women cite not knowing what games to try and feeling that “games are too violent” as explanations for why they do not play more games (Blackburn & Scharrer, 2019; Chess, 2020). Although most published games contain some kind of violence (APA Task Force on Violent Media, 2015), there are now a wide range of games and genres, and the discourse of games as primarily violent is slowly disappearing (Heritage & Humphreys, 2024). Chess (2017) suggests that women may not be aware of the range of games that are available: another impact of marketing strategy.

At the same time, decades of assumptions about gender and gaming (Eklund, 2016) have fed into additional barriers to women in the form of hostility from the gaming community. This includes the sexualisation of female video game characters to serve the presumed straight male gamer, limited representation of women in games and gaming culture, and harassment and sexist abuse (e.g., Bustos-Ortega et al., 2024; Fox & Tang, 2017; Nakandala et al., 2017). Vulnerability to this harassment may increase for women when moving from casual to core gaming spaces (Chess, 2020), creating an atmosphere that might make women feel like games are not “for them.”

Time and Leisure Equality

Another potential barrier to women playing video games is lack of time. Across various studies, women have been found to have less leisure time than men (e.g., Noonan & Glass, 2012; Yerkes et al., 2020a; Kan et al., 2021) and women’s leisure time is more likely to be interrupted (Craig & Mullan, 2013). A study of UK respondents from 2015 found that women have an average of 40 min less leisure time per day than men (Payne, 2017). Further, Yerkes et al. (2020b) found that women respondents felt more rushed than men and were less likely to use their free time to relax and recover.

These differences in leisure time have an impact on opportunities to game. In a study of the social networking game YoWorld, “not having enough time” was the main reason cited for women stopping playing the game (Bergstrom, 2019a). Work commitments and “life getting in the way” were the next most frequent responses, while cost, boredom and bullying were relatively rare reasons. In a study of American video game players, women had less free time, and played for shorter time periods than men (Winn & Heeter, 2009). Accordingly, lack of time may also act as a barrier to identifying as “gamers.” As Hanford (2018) writes:

The “gamer” identity is strongly wrapped up in the amount of work and investment one puts into gaming. The effort that gamers put into their play becomes a badge of commitment and their abilities merit their use of the term. This cycle of legitimacy and work has a multitude of effects, from the shaming of women at gaming conventions to the common critique of Anita Sarkeesian not playing enough games to critique them. (p. 150)

Of course, gender is not the only relevant factor in how we spend our leisure time; age plays an intersecting role. A Spanish study (Garrido et al., 2022) found that as people age, they spend less time in cultural leisure activities (film, theatre, reading books, and most games) and social leisure activities (talking on the phone, religious practices), and more time on passive leisure activities (watching TV or listening to the radio, napping, resting). However, in older adults, women were even less likely to participate in cultural leisure activities than men.

Chess (2017, 2020) argues that playing video games is an important issue for leisure equality:

Video games create an opportunity to enact leisure equality. It’s a space that has been largely dominated by masculinity until this point and it is a space where femininity gets largely ignored, harassed or mocked. A renewed focus on video games creates a new set of opportunities for thinking about leisure equality. (Chess, 2021)

Indeed, a study of 800 US and UK gamers found that 67% of women players felt that gaming was “a vital means to escape the daily grind,” though 42% felt guilty about taking time to play games (GameHouse, 2023). Surveys on this topic may be subject to reporting biases precisely because of the emotions involved. For example, underreporting of time spent playing the online multiplayer game Everquest II is higher for women compared to men, and for older players compared to younger, and this is attributed to differing social desirability of gaming (Kahn et al., 2014; Williams et al., 2009). Women were also more likely than men to report anticipating guilt about playing a violent video game (Hartmann et al., 2014, though this relates to guilt about the moral content of the video game rather than guilt about not doing some other activity). It seems plausible that the typical bias is to underreport rather than overreport barriers, so while it is important to be critical of the results, we assume it is more likely that they underestimate rather than overestimate the scale of the issues.

Current Study

The aim of the current study is to examine whether the barriers discussed above are evident in the attitudes and behaviours of women who play video games. We investigate how women living in the United Kingdom who play mobile games feel about gaming, and how this relates to their gaming behaviour. We present results from a new study of 1000 UK women and consider what gender theorists, game scholars, and the games industry can learn from this. Although this is mainly an exploratory study, we had three main research questions that informed the design.

The first research question relates to how women think about their gaming identity. Given the literature above, we expected that most women would identify as ‘casual’ gamers, if they identified as ‘gamers’ at all.

The second research question concerns women’s attitudes and emotions related to games. The studies above suggest that women are likely to feel guilty about taking leisure time to play, and to keep their gaming secret for fear of judgement. Experiencing these negative emotions may be inversely proportional to experiencing positive emotions like pride and anticipation of gaming.

The third research question is whether women’s attitudes are a barrier to their gaming behaviour. This might be evidenced, for example, if the likelihood of experiencing guilt or keeping their gaming secret was associated with indices of being excluded from gaming culture. Furthermore, the literature on leisure time might predict that women are less likely to play games if they feel guilty about spending time playing. Similarly, women may be motivated to play depending on their attitudes and emotions and depending on the perceived benefits. For example, if playing reduces stress, then they should choose to play when they are stressed.

The questions and predictions mainly follow from the empirical observations in the literature above. However, although we are agnostic about the most suitable framework through which to view our results, the predictions may also be derived from various theories. For example, Chess (2017) invokes actor-network theory (Bijker et al., 1987; Latour, 2005) and notions of cultural assemblage (Taylor, 2009) to emphasise that there are complex relationships beyond those between an individual player and a game, including between other players, game makers, and wider cultural discourses. Given this, social role theory (Chappetta & Barth, 2022; Eagly, 1987) might predict that differences in leisure time and the under-representation of women in gaming in the past would lead to a stereotype that playing video games does not fit with a stereotypical feminine role. Similarly, Kelly et al. (2022) suggest that the stereotype content model (Fiske et al., 2002) would predict that gaming identity for women is related to notions of gaming competence. Women may attend to these stereotypes and group affiliations, developing gender schema that categorise games as ‘masculine’ (Beasley & Collins Standley, 2002; Bem, 1981). This may lead to them managing their social identity by avoiding association with the gaming community (Stone, 2019) or fearing being penalised for violating a perceived gender norm by playing games (Eareckson & Heilman, 2024), leading to negative emotions surrounding playing games.

Method

Participants and Procedure

Ethical review was conducted by Cardiff University. A well-known national survey company (Opinion Matters) was contracted to obtain a representative sample of 1000 women in the United Kingdom who play mobile games. A set of filtering questions confirmed that the respondents identified as women and had played a video game on their mobile device in the last 3 months. The sample was fairly balanced in terms of age and representative of UK regions (not significantly different from proportion of national populations, see Table 1). Participants answered a series of survey questions through an online interface.

Table 1 Demographic Information of Participants
Full size table

Materials

To ensure the questions and possible responses were relevant for the gaming industry, this survey was designed in partnership with a video games company, Undone Games. The survey included 14 questions of different kinds organised in several sections (see Table 2). The first section collected demographic information. The second section asked about the participant’s gaming habits, including the context in which they played and the type of games they played (multiple choice, multiple-answer questions, based on categories chosen by the gaming company to align with mobile market sectors). Participants were asked how much time they spent in an average day playing on mobiles, and an identical question about playing on other platforms. The third section asked about their gaming attitudes (Table 3), including how they would define their gaming style, and their level of agreement with various attitudinal statements about gaming, such as whether they felt guilt, shame or stress related to gaming. Although free text responses were available, these were used in less than 2% of cases, so these answers were ignored in the main analyses.

Table 2 Summary of Survey Materials
Full size table
Table 3 Summary of Main Gaming Attitude Questions (Numbers Represent Number of Respondents)
Full size table

Statistical Analysis

All analyses were carried out in R (R Core Team, 2021) using RStudio (RStudio Team, 2020). Likert scale responses were treated as ordinal variables. To increase statistical power for some analyses, Likert variables were divided into two binary categories ‘agree’ (“Somewhat agree,” “Strongly agree”) and ‘do not agree’ (all other choices). Each question that allowed multiple responses was split into multiple independent binary variables indicating whether the participant had selected the response or not. The final dataset included 116 variables for 1000 participants (see Table 2, discussed below).

Our exploratory analysis involved identifying factors that are associated with the key barriers outlined above, for example which factors predicted the level of agreement with experiencing guilt over playing games. However, looking at simple correlations (or categorical measures of association) between a target variable and over 100 other variables risks identifying false positive associations. While a (multinomial or ordinal) regression approach might evaluate the predictive strength of variables against each other, it would struggle to identify robust interactions with so many variables, and to avoid issues with multicollinearity.

Therefore, we use a ‘random forest’ method to identify the most important predictors. Random forest analysis is a machine learning method that can deal with correlated predictor variables and unbalanced datasets and identifies complex interactions in a tractable manner (see Hothorn et al., 2006; Strobl et al., 2008). The method is based on the calculation of thousands of decision trees. A decision tree is a way of explaining variation in a target variable by splitting the data into sub-groups according to predictor variables. The process identifies a series of yes-no questions to ask about the predictor variables to make a guess about the target variable. The structure is similar to a dialogue tree, or the boardgame ‘guess who.’ For example, imagine we had asked people to classify several games as ‘core’ or ‘casual.’ A decision tree predicting people’s answers might include a question “Does the game cost more than $10?”, and if the answer is yes, then there is a sub-question “Does it require high-spec hardware?” Answering ‘yes’ to both would predict it was more likely to be a core game, and ending up in one of the other branches might suggest it was less likely to be a core game.

However, this single tree is just one of several possible effective trees (e.g., perhaps it’s better to ask about hardware before cost, or to adjust the price from $10 to $40, or to ask about game length rather than cost), so many trees are generated using random sub-samples of data and predictor variables. This ‘forest’ of trees is evaluated to give a “variable importance” score to each predictor variable for a given dependent variable. We used an importance measure based on permutation (Strobl et al., 2007): The values of a given variable are randomly permuted and the prediction accuracy of each tree is re-calculated. The variable importance measure is the average difference in prediction accuracy between the unpermuted data and the permuted data for each tree. If the variable is critical to accurate prediction, then there should be a large difference and the importance scores will be high. One can then compare the importance scores of different independent variables to see which is contributing more to prediction (the relative ranking of the importance scores is relevant, while the absolute values of the importance measure depend on the overall accuracy of the tree and various other factors). That is, if variable A has a higher importance score than variable B, it suggests that knowing about A provides more effective information for predicting the target variable (it’s more important to ask about A than about B). Essentially, the random forest process ranks a set of predictor variables according to how well they explain the variation in a target variable. This guided the investigation of specific associations in the data.

For each target variable, there were 111 possible non-demographic predictor variables. When applying random forests analysis, variables that had logical implication relationships with the target variable were removed from the predictor list. For example, when predicting a respondent’s level of agreement with the statement “I feel guilty about taking time to play games,” we omitted the predictors “I feel guilty if I spend too much time playing games,” and “I feel guilty about taking leisure time.” See the supporting materials for more details.

To estimate the strength of association between two specific variables, we use a range of methods including Fisher’s exact test (more accurate and fewer assumptions than the Chi-square test) for two binary categories, and ordinal regression with a cumulative distribution function (using the R package brms; Bürkner, 2021) for ordinal categorical variables. The final dataset included 116 variables (five demographic, 111 non-demographic) for 1000 participants. We divided the results into thematic groupings, including guilt, fear of judgement, time, and stress/boredom.

Results

Core and Casual Identities

We begin with a descriptive account of gamer identity. Participants were asked to pick labels that best described their gaming style, allowing them to choose multiple options from any of “Casual,” “Social,” “Competitive,” “Hardcore,” or “N/A.” “Casual” was chosen by 74.8%, and exclusively by 62.5%. The next most chosen label was “Social” (24%, 12.5% exclusively) followed by “Competitive” (10%, 3.9% exclusively) and “Hardcore” (4.7%, 2.3% exclusively). “N/A” was chosen by 3.7%. In terms of intersections (Fig. 1), “Casual” is occasionally alongside “Social” (9.7%) or “Competitive (4.6%), but rarely with “Hardcore” (1.4%).

Fig. 1
figure 1

‘Upset plot’ of set intersections between different gamer identity labels. The set intersection guide indicate which choices were made by each column of results. The insert on the left shows the overall frequency of a label being selected

Full size image

Puzzle games were the most popular genre, played by 77.5% of women, followed by similar ‘casual’ genres like word games (47.7%) and logic games (30.2%), and traditional games like card games (37.7%) and board games (23.4%). Genres such as real time strategy (RTS), shooters, driving, fighting, sports, and stealth were played by less than 10% of women (it should be noted that these questions are about mobile game genres, and women may be playing different genres on PCs or consoles). To identify clusters, we performed a principal component analysis (similar to Eklund, 2016, see supporting materials at https://osf.io/myeb8/?view_only=b6edbaaed1e0418daad4aeb2756cb871). However, the genres of games played were not significant predictors of our main variables of interest in the analyses below.

Guilt

Out of the 1000 respondents, almost a third (28.8%) agreed with the statement that they felt guilty about taking time to play games. In terms of demographics, older players were less likely to feel guilt than their younger counterparts (31% of women aged 16–24 reported feeling guilty compared to 24% of those aged 55 or older). An ordinal regression predicting participant’s agreement with feeling guilty by their age category revealed a significant linear term (β = -0.59, 95% CI = [-0.87, -0.30]), indicating a general decline of guilt in older age categories (which may be related to more time free from work or childcare, for example). There was some regional variation in the proportion feeling guilty (from 21% in the West Midlands to 42% in London), but exploratory analyses did not find any strong associations between region and behaviour or attitudes (see supporting information).

Figure 2 shows key explanatory variables identified using the random forest approach. Guilt was strongly associated with keeping gaming a secret from friends and family – women who kept gaming a secret were nearly three times more likely to feel guilt than women who did not (63% vs. 23%, Fisher’s p < .001, see Fig. 3). There were also relationships to feeling excluded from gaming culture: Participants were more likely to feel guilty if they thought that gaming is mostly a male past-time (47% vs. 23%, Fisher’s p < .001), would be embarrassed to call themselves a gamer (44% vs. 22%, Fisher’s p < .001), or thought that people who play games on mobile or tablet are not gamers (44% vs. 24%, Fisher’s p < .001). Finally, there were correlations with barriers related to game marketing: Women were more likely to feel guilty if they didn’t know which games to try (41% vs. 23%, Fisher’s p < .001), or thought that games were too violent (44% vs. 23%, Fisher’s p < .001). Other factors had considerably lower importance scores, suggesting they were less important predictors.

Fig. 2
figure 2

Variable importance for factors predicting whether a respondent felty guilty about taking the time to play games. Higher scores indicate greater predictive power (not the direction of correlation). Only the relative distance between importance scores is meaningful, not the absolute importance score. The final row shows a condensed distribution for the 99 other less important variables

Full size image
Fig. 3
figure 3

The percentage of women feeling guilty about taking time to play games according to their responses to the key predictors identified by the random forest model. Percentages indicate the percentage within the subgroup agreeing or disagreeing with the question

Full size image

To test whether these key predictors contribute independently to predicting guilt (versus just being slightly different measures of the same thing), they were entered together as independent variables in an ordinal regression. All remained significant predictors suggesting that they were at least somewhat independent contributors (overall accuracy = 43%, correlation between model prediction and independent variable, tau = .43, z = 16.8, p < .001), except for thinking that people who play on mobiles or tablets are not gamers (see supporting materials at https://osf.io/myeb8/?view_only=b6edbaaed1e0418daad4aeb2756cb871). This latter variable may score highly in the random forest because of its predictive power in interaction with other variables.

The factors associated with guilt combine cumulatively. For example, Fig. 4 shows a decision tree based on just three variables. Only 17% of respondents feel guilty if they disagreed that gaming is a male pastime and would not be embarrassed to call themselves a gamer. On the other hand, 63% of respondents felt guilty if they did not disagree that gaming was a male pastime and strongly agreed that they did not know which games to try. The regression model predicted that a woman who strongly agreed with each of the six most important predictor statements above would be 97% likely to feel guilty, compared to 2% likely if they strongly disagreed with each statement.

Fig. 4
figure 4

Example of a decision tree with two levels using the most important variables identified by the random forest analysis to predict responses to whether participants feel guilty about taking the time to play games. Colours indicate responses to each question. ‘n’ represents the number of participants in each partition

Full size image

Secrets

Almost 1 in 6 respondents (16%) kept gaming a secret for fear of judgement. Due to the strong relationship with guilt identified above, the random forest analysis excluded questions related to guilt. The results revealed that keeping gaming a secret was associated with feeling anxiety or depression when scrolling social media: 31% of women who felt anxious or depressed kept gaming a secret, compared to 12% of women who did not, χ2(1) = 43.78, p < .001.

Motivation also played a role in secret-keeping. Women who played games to challenge themselves were half as likely to keep it a secret, χ2(1) = 13.81, p < .001. However, women who felt very proud of their achievements in games were more likely to keep their gaming a secret than those who did not feel as proud, 22% vs. 14%, χ2(1) = 5.14, p = .023.

Time

The median time playing games was 1–2 h per day and 4.6% of respondents play on their mobile for 5+ hours (Table 4). There is a positive correlation between time spent on mobile games and time spent playing games on other platforms, such as PC or console (Fisher’s test p < .001). There was also a positive correlation between time spent playing mobile games and time spent scrolling social media (Fisher’s test p < .001). There was no difference in time spent gaming between people who keep gaming a secret and those who do not (Fisher’s test p = .898). Likewise, there was no difference in time spent gaming based on how guilty respondents feel (Fisher’s test p = .951).

Table 4 Distribution of Average Time Spent Playing Mobile Games Every Day
Full size table

What attitudes toward games do predict the amount of time spent playing mobile games? The random forest approach suggests several associations (Fig. 5). First, perhaps unsurprisingly, women who agree they do not play enough games to consider themselves to be a gamer play less than those who disagree (38% of women who agree play for more than an hour a day, compared to 53% of women who do not agree, Fisher’s p < .001). While this may be an accurate reflection of gamer values, we note that 28% of women who play for more than 5 h per day still don’t deem themselves to be playing enough to consider themselves a gamer. Similarly, 65% of women who play mobile games for more than 5 h a day still considered themselves ‘casual,’ while only 14% considered themselves ‘hardcore.’ The numbers are even more extreme for players who spend more than 5 h a day playing PC/console games (75%, 12%). This may reflect unrealistic perceptions of gamer behaviour, or that the identity tags of ‘gamer’ and ‘hardcore’ are perceived to be about the type of game one plays, or it may indicate a general aversion to ‘gamer’ identities.

Fig. 5
figure 5

Responses to attitudes to gaming according to the average number of hours of mobile games played daily

Full size image

The relation between time spent gaming and the willingness to pay for a mobile game is expected: Both are measures of investment in the hobby, and people may be more willing to play higher-quality games for longer. Interestingly, women tended to play more if they were proud of their achievements (50% of women who were proud played for more than an hour a day, compared to 37% of women who were not proud, Fisher’s p < .001). In contrast, women tended to play less if they agreed that gaming was mostly a male pastime (Fisher’s p = .022) or think that games were too violent (Fisher’s p = .031), though these effects were slightly weaker. Having friendly competition with friends was not a predictor on its own (Fisher’s p = .125) but had a moderate importance score in the random forest analysis because it interacted with other variables. For example, for people who have friendly competition and strongly disagreed that games are too violent, 66% played for more than an hour a day compared to 43% of those who did neither (Fisher’s p < .001).

Stress and Boredom

Respondents were asked whether they played when they were stressed or bored, and whether playing alleviates these negative emotions, which allowed us to investigate the motivation for, and benefits of, gaming. For example, Fig. 6 showed that the results for boredom are straightforward: Those who reported playing when bored were likely to agree that playing lowers boredom (73% agree vs. 27% do not agree), and those who reported that playing makes them feel less bored were more likely to report playing when bored (82% vs. 18%). This pattern is consistent with a general motivation to seek out games to help alleviate boredom.

Fig. 6
figure 6

Cross-section of respondents according to various attitudes to stress and boredom, showing an asymmetry for attitudes to stress

Full size image

However, there was an asymmetry for stress: most women who play when they are stressed feel less stressed when they play (77% vs. 23%), but under half of women who report that games make them feel less stressed report that they play when they are stressed (42% vs. 58%). That is, although most women feel less stressed when they play, they won’t necessarily seek out games when they feel stressed.

The difference in results for boredom versus stress may be due to the context of the two emotions, and to (potentially gendered) assumptions about permissible remedies. When players are bored, they may have little else to do (hence the boredom) and so have time to play games. In contrast, when participants are stressed, they may have other demands on their time and thus not see play as a viable (or justifiable) option, even though it might make them feel better. This experience may be exacerbated by the fact that the types of video games being advertised to women are often considered unimportant and unserious, as discussed above. This suggests that there is a missed opportunity to market games to women as a legitimate form of stress relief.

We note that there is a similar asymmetry between agreeing that playing alleviates loneliness and whether they play with others: Only 16% of women who reported that playing makes them less lonely play with friends or strangers, and most of those who play socially say that playing does not make them less lonely (63%). However, relatively few women overall reported that playing alleviated loneliness (20.6%), and more work is needed to understand how women use games to engage with others or achieve the feeling of engaging with others (see e.g., Ballard & Spencer, 2023; Barr & Copeland-Stewart, 2022).

Discussion

Taken as a whole, we propose that the results suggest that many UK women feel excluded from video game culture, and that this exclusion can be a barrier to playing mobile games, or being open about one’s play. In other words, many of the results can be explained by women’s sense that games and gaming culture are not for them. For example, as noted above, almost a third of women (29%) feel guilty about taking time to play. This is in keeping with previous research, which has found many gender disparities in leisure time, with women being almost twice as likely to feel guilty about taking ‘me time’ than men (GameHouse, 2023). However, our study found that women who look forward to playing are more likely to feel guilty than those who do not (34% vs. 25%).

Those less familiar with games and gaming culture were also more likely to feel guilty: women were twice as likely to feel guilty if they did not know which games to try (41% vs. 23%) or if they thought games were too violent (29% vs. 10%). This aligns with barriers identified by Chess (2017): That women may not be aware of the diversity of games available, because there is a limited range marketed to them (thus the former), and yet when they think of ‘games’ simpliciter, they may not have casual games in mind (thus the latter).

Familiarity with gaming culture was not the only relevant factor in feeling guilty about gameplay; guilt also seemed to be related to whether women identified with gaming culture. 60% of women surveyed believed that they do not play enough games to consider themselves a ‘gamer.’ Women were twice as likely to feel guilty if they thought that gaming was mainly a male pastime (46% vs. 23%) or if they said they would be embarrassed to call themselves a gamer (55% vs. 35%). In contrast, women who felt that occasionally playing any game made you a gamer were 42% less likely to feel guilty. These patterns are suggestive of women feeling that games and gaming are not for them.

Another barrier to gaming for women is fear of judgement. For example, 16% of respondents keep gaming a secret from friends and family for fear of judgement. Furthermore, feeling and expressing pride in gaming as a woman is a difficult balancing act: Women who felt very proud of their achievements were actually slightly more likely to keep it a secret than those who did not feel as proud (22% vs. 14%). In contrast, women who were motivated to play games to challenge themselves were half as likely to keep it a secret (11% vs. 20%). Again, these patterns are suggestive of women feeling excluded from games and gaming culture: that playing games is not the sort of thing of which one should feel proud of or identify with.

This study showed that significant numbers of women feel guilty about playing video games, and ashamed about what others think of them taking time to play games. However, to our surprise, feeling guilty or ashamed did not correlate with how much time women spent playing games. We expected a significant proportion of women to report feelings of guilt around playing video games and taking leisure time more generally, and this proved to be the case. But while we anticipated feelings of guilt and shame would have a negative impact on the amount of time women spent playing games, we did not find such a connection. Those who feel guilty or ashamed do not play less, but they feel worse.

Limitations and Future Research Directions

This study only considered responses from people identifying as women. While the results are valid and revealing, they do not demonstrate that women feel more guilty or ashamed about playing mobile games than men, or people of other genders. Other studies suggest this is likely, such as the GameHouse study which found that women were more likely to feel guilty while playing mobile or other video games compared to men (42% vs. 30%; GameHouse, 2023). Future research might consider gendered patterns in attitudes towards playing games and the differences in causes and associations. It might also consider intersectional factors beyond gender and age, such as socio-economic status, dis/ability, caring duties, and so on, as well as comparing the current results to results from other countries or non-mobile gaming platforms. Furthermore, future work could consider the role of feminist identity (McCullough et al., 2020) and perception of others within the gaming community on one’s own self-perception (Yao & Rhodes, 2023), as well as directly investigating why women choose to quit or avoid games (Bergstrom, 2019b).

In relation to methodology, the survey asked participants to rate their agreement with statements about feeling guilt and shame. This follows the procedure in the Guilt Inventory (Jones et al., 2000), which also provides a set of statements (e.g., “Guilt and remorse have been a part of my life for as long as I can recall”), and asks participants to indicate their agreement on a 5-point Likert scale from “strongly agree” to “strongly disagree.” However, measuring and distinguishing between these feelings is complex (Giner-Sorolla & McGee, 2020). Several scales have been developed, such as the Test of Self-Conscious Affect (TOSCA; Tangney & Dearing, 2002) and the Guilt and Shame Proneness scale (GASP; Cohen et al., 2011), but there were several drawbacks to using these for the current study. For example, TOSCA is based on hypothetical scenarios and measures ‘constructive guilt’ around personal failure (e.g., guilt in relation to interpersonal conflict or committing crimes), while we were more interested in maladaptive feelings for consistently failing to adhere to a gendered expectation. Indeed, Giner-Sorolla et al. (2011, p. 445) suggest that TOSCA “may not measure the experience of guilt, but rather motivation to make amends for personal wrongdoing,” which does not fit our aim. Also, we were interested in chronic guilt specifically in relation to video game behaviour, rather than general proneness to guilt, and TOSCA measures correlate poorly with the Guilt Inventory measures (Benetti-McQuoid & Bursik, 2005). Furthermore, while there are specific scales developed to measure guilt around trauma (Günlü, 2025) or sexuality (Mosher, 2013), there are few that investigate guilt in relation to gendered expectations around leisure time. Future studies might develop the scales above and pair them with measures of gender role (see e.g., Benetti-McQuoid & Bursik, 2005; Eterović et al., 2022).

Practice Implications

The analyses of our survey of UK women found striking evidence of barriers to playing mobile video games. The results suggest that women tend to spend less time playing if they feel they don’t fit into gaming culture, for example if they believe that gaming is a male pastime, or that they don’t play enough games to be a gamer, are embarrassed to call themselves a gamer, or think video games are too violent. In contrast, women spend more time playing if they are proud of their gaming achievements. This suggests that guilt and fear of judgement are symptoms; while they are clearly negatively impacting women gamers, they are not directly barriers to play, and not the root of the problem.

Efforts to dismantle barriers are complicated by the interrelations between them. For example, recently, in response to their aforementioned survey, GameHouse launched a UK marketing campaign to encourage women to take more “me time” to play games. Despite the campaign’s positive message around leisure time, many of the games published by GameHouse focus on women in gendered service roles (cooking, caring for children or animals, planning a wedding) with time management mechanics, effectively recreating the pressure to be productive within leisure time activities (see e.g., Harrington & Dawson, 1995; Henderson, 1991; Ganzon, 2019). This suggests that understanding the relationships between barriers is complicated, and in reducing one barrier, it is easy to bolster another.

Removing barriers to (open) play for women may require deeper changes such as reducing leisure inequality – for instance, Chess (2021) encourages gamers to:

Find more people in your lives that aren’t getting enough leisure and play and help them to play more… because it will make us better humans and better feminists. And if more of us play more games we will make games better.

However, there are interim measures we might take that still acknowledge the many pressures on our time. Challenging the perception of casual games as less serious, lower quality, and less valuable than their ‘core’ counterparts is one such measure. For instance, Juul (2010) found that many players of casual games look for challenging experiences, but tend to play for shorter periods of time. Chess (2017) calls games with this feature ‘time positive’ (p. 52), allowing players to play for shorter intervals and to easily recommence play after a break. Accordingly, we echo Keogh’s (2016) argument that we should “constructively re-conceptualize causal games as those that are more flexible with the player’s time,” delineating them not on the basis of their hardware or skill demands, quality, or seriousness, but rather by how easily they can be “incorporated into the player’s everyday life” rather than existing “as an activity apart from it” (pp. 34–36).

The games industry, gaming scholars, and participants in gaming culture can also help by broadening the perception of who plays video games, who games are for, and what types of games are available. As discussed above, our study found that most women feel less stressed when they play, but won’t necessarily seek out games when they feel stressed. The need for games that alleviate stress is recognised by game makers, for instance a designer interviewed by Chess (2017) noted that “putting in mechanics that deliberately create stress or introduce the possibility of failure tend to not be as popular” with women who play mobile games as a way to relax or “decompress from the stresses of their lives outside of games” (p. 45–46). But participants’ responses in our study (both related to stress and unfamiliarity with the range of games available) suggest that marketers are not conveying the variety and potential benefits of mobile games. Chess (2017) notes that game companies are “well-known for being tight-lipped about actual and intended demographics” (p. 40), but our results suggest that the current marketing of, and messaging around, games is leaving out a significant proportion of possible gamers. Furthermore, since we found that pride in achievements is associated with more time playing, it may also help to normalize the celebration of gaming achievements for women.

There is danger in being too prescriptive about how we should design games for women, as too often such prescriptions reinforce spurious intuitions about gender difference (e.g., Ray, 2004). However, our results are grist to the mill of those arguing for more inclusive game design in general (Chess, 2017; GDI, 2024; Lazzaro, 2008).

Conclusion

This study explored attitudes of women who play mobile games in the United Kingdom. It found that, despite being an important part of their lives, many women experience negative emotions related to games. These emotions are associated with gendered barriers to play, such as not feeling like a core part of the gaming community. Understanding the barriers to gaming in more detail is important for overcoming them, to increase the quantity and quality of women’s gaming experience.