Social Barriers to STEM: A controlled study, with elimination of demand characteristics
Prior research has suggested that social factors hinder the entry of women into STEM (Science, Technology, Engineering, and Mathematics) fields. However, it was noted that most of these studies asked very gender-specific questions, which could produce high demand characteristics and thus bias responses. Moreover, some studies lack a sufficient control. Consequently, an attempt to gain a better, more accurate understanding of social factors and STEM is what led to the creation of this study.
Aim: To understand the extent to which social factors impede women’s progression in STEM and, by testing for specific factors, to identify which factors contribute the most to such possible differences.
Method: A questionnaire was administered to 217 high school students, most of them studying the IB (International Baccalaureate) program, worldwide through online groups (e.g. Facebook). Their names, ages, geographical origin and involvement in STEM were recorded, after which they were asked to rate the impact of 7 different social factors on promoting interest in STEM on a scale of 1 to 5 (1= “Highly discouraged”, 5= “Highly encouraged”). The seven factors were:
1 - Influence of parents
2 - Influence of teachers
3 - Influence of classmates
4 - Influence of role models
5 - Influence of clubs and organizations
6 - Influence of availability of resources
7 - Influence of ease of access to resources
Lastly, as a filler, they were asked about any other factors that may have impacted their decision regarding pursuing STEM.
Results: Differences between ratings for each category were measured using a T-test at the level p=0.05. Among men (n=50) and women (n=87) who were pursuing STEM, mean ratings for all 7 factors were higher among men; However, only 3 of the 7 factors (classmates, role models, and ease of access to resources) showed statistically significant differences (p=0.017214, p=0.010231, and p=0.03985 respectively) between men and women.
Among men (n=20) and women (n=55) not pursuing STEM, there were no significant differences in social factors affecting their decision. In 3 of the 7 factors, women rated themselves to be more encouraged to enter STEM as compared to men, in 1 factor the mean ratings were equal, and for the other 3 factors, men were more encouraged. Thus, being a 50/50 split, there were truly no real differences found.
Conclusion: Findings suggest that among all social factors, it is primarily a subset (specifically: that statistically hinder women’s progression in STEM. It further reveals that although women entering STEM face social issues, women who do not enter STEM largely do so out of choice - approximately at the same level as why men choose to not enter STEM. These results are relevant to understanding exactly which social issues must be targeted the most to truly improve the outlook of women entering STEM fields.
Findings are consistent with prior surveys conducted on women’s opinions regarding barriers impacting their performance in STEM; an informal survey conducted by the same author in 2018 portrayed “lack of role models” as one of the highest factors hindering women entering STEM fields. This raises the reliability of our work. It is also consistent with prior research.
The cross-cultural, etic nature of the survey renders findings more generalizable than if confined to a specific geographical group. Nevertheless, it can be argued that cultural differences may have created variability in the responses, thus acting as a confounding factor. This could perhaps be eradicated by utilising a larger sample, or confirming the possible influence of culture through a second analysis of region of origin and STEM ratings. Although such an analysis has not been conducted as of yet, it will indeed be an insightful tool to obtain a more comprehensive, multivariable picture of the story.
The primary limitation to be weary of in this study is the relatively smaller sample size (n=217) as compared to previous research (n=492). This could have perhaps yielded from relatively lower advertising, or lower response-rate to the advertising. Thus, although a seemingly consistent trend is seen in findings, they cannot be extrapolated to the entire human population with certainty: more longitudinal research with larger sample sizes would require to be conducted in order to increase confidence in the results.
Moreover, as these social barriers to STEM were self-reported, again it is possible that participant bias (inherent or externally acquired), or the social desirability effect, or even social conformity effect (for those who may have guessed the aims of the study), could influence respondents’ answers in order to depict such a trend. Although the primary goal of this study was to eliminate such a factor in the first place (which supposedly we rightly did), we cannot assume that 100% of the participants had zero demand characteristics - particularly due to the overt and globally discussed nature of the issue.
The methodology could have further provided more detailed understanding by including more “social barriers”, or subdividing the existing 7 barriers into more specific ones. The reason behind why this was not done by the researchers was for the sake of convenience, as respondents tend to avoid long questionnaires. However, replications of this study with minor (or major) tweaks could refine our knowledge of which specific factors influence women’s entry into STEM.
Some of the news is actually good news: It shows that progress is being made on gender attitudes towards STEM: That parents, teachers, and organizations have nearly eliminated themselves as contributors to the gender bias, and that financial aid does not discriminate against women within STEM. Of course, this progress must continue, and strive to maintain minimal differences in such social factors on girls in STEM.
What it does reveal, however, is that classmates, role models, and resources are still factors to be tackled. This has face validity: It is more difficult to control the behaviors of young classmates as to make them aware of the consequences of their actions (and to not perpetuate gender traditions) as opposed to controlling them in informed adults and mentors. Similarly, one cannot “create” role models out of the blue, and a historically fewer number of recognized women in STEM makes it difficult for today’s women to have an ideal for inspiration. Lastly, ease of access to resources may differ due to several reasons, many which link to other points: A girl may lack access to help in her studies because her father is working and her mother did not study STEM (tying back to the ‘role model’ idea), she may lack resources because clubs and organizations have limited equipment that is ‘hoarded by the boys’ (linking to ‘clubs and organizations’), etc.
To resolve this, education regarding gender biases and discrimination must be perpetuated at a young age, so as to help students conduct gender-sensitive actions as opposed to amplifying existing stereotypes.
Furthermore, female role models must be made as publicly visible as possible. Such actions have already been instilled by the author within her own educational institution, UWCSEA Dover, by instilling posters of various female scientists and mathematicians all over the primary school building - to embed in young girls’ minds that they, too, can achieve in STEM. Expanding such a simple yet effective initiative in other schools worldwide could have, potentially, a dramatic interest in confidence among women to study STEM.
Lastly, to improve access to resources is a trickier job that can be done, on a broader scale, through collaboration between public policy makers, non-profit organizations educational institutions, and legal guardians. Governments and organizations must ensure that equal girls have access to resources as boys, schools should monitor and ensure this equal distribution, while parents must also be wary of unfair access to resources for their female children.
Nevertheless, we cannot deny that the explanation cannot be purely social: As prior data have suggested, including my previous work regarding empathy quotient differences, the gap could simply arise due to a lack of preference for STEM fields.