Born to Code: Does the Portrayal of Computer Scientists as Geniuses Undermine Adolescent Youths’ Motivational Beliefs?

The study investigates whether portraying computer scientists as inherently gifted “geniuses” undermines adolescents’ motivational beliefs in pSTEM. Drawing on situated expectancy-value theory, the author notes that popular media often depicts computer scientists as nerdy, brilliant White men, which may discourage interest, particularly among girls and historically marginalized groups. Prior correlational work links endorsement of genius stereotypes with lower pSTEM expectancy and value beliefs, but experimental evidence is lacking. The hypothesis is that exposure to media emphasizing genius stereotypes will decrease high school students’ pSTEM expectancy and value beliefs, with gender explored as a potential moderator.

The paper reviews evidence that pSTEM fields are stereotyped as requiring innate brilliance, contrasting with growth mindset perspectives. Studies show adolescents and young adults perceive success in STEM as dependent on brilliance and that such beliefs correlate with lower motivation among those who do not see themselves as naturally gifted. Media is identified as a conduit for transmitting these stereotypes. Prior experimental work with undergraduates found that exposure to nerdy stereotypes reduced women’s interest in computer science. Situated expectancy-value theory provides a framework for how sociocultural stereotypes shape expectancy (self-concept and expectations for success) and value (attainment and utility) beliefs. The paper positions the current experiment as filling a gap by testing genius stereotypes experimentally among high school students and assessing both expectancy and value outcomes.

Design: Between-subjects experimental design with pretest-posttest measures. Participants were randomly assigned to a genius-stereotype condition or a control condition. A priori power analysis indicated 212 participants were needed for within-group differences. Participants: 213 high school students from four Northern California high schools (of 226 who began; 13 did not complete). Assignment: control n=110 (51.6%), genius n=103 (48.4%). Gender: girls n=100 (46.9%), boys n=113 (53.1%). Grade levels: sophomores 47.4%, juniors 39.4%, seniors 12.7%, one first-year. Ethnic-racial background: Asian 47%, White 30%, Latina/o/x 12%, multiethnic 10%, Black 1, Native American 1, unreported 1. Setting/Procedure: After IRB approval, science teachers administered an online survey during class. Students (all enrolled in at least one science and one math course) provided demographics, completed pretest pSTEM expectancy-value scales, read a randomly assigned mock article, then completed posttest scales. The experiment occurred in the final ~10 minutes of a broader 20-minute survey session. Materials/Stimuli: Two fabricated high school newspaper-style articles titled “College Life: Meet Markus, a Computer Science Major,” matched in length and layout to a district online paper. Genius condition: portrayed Markus as naturally gifted, early coder, engaged in hacking, advanced STEM coursework; included quote asserting that success requires “real talent” and that some people are not meant for it, urging exclusive focus on advanced math and self-teaching coding. Control condition: portrayed Markus as discovering coding later, taking a broad array of high school courses (including art), and stating anyone can learn to code; directly refuted the need to be a “genius.” Measures: pSTEM expectancy-value beliefs (Eccles & Wigfield framework), administered pre- and post-article. Expectancy (10 items; self-concept and expectations for success; 5-point scale; pre α=0.92, post α=0.94). Value (4 items; attainment and utility value; 5-point scale; pre α=0.86, post α=0.85). Example items: expectancy—“How well do you expect to do in your pSTEM courses this year?”; value—“How important is it to you to do well in pSTEM courses?” Items averaged to form scales. Background variables: gender, ethnic/racial background, year in school, math grade. Analytic Approach: Bivariate correlations and t-tests assessed baseline equivalence (no significant pretest differences by condition). Two mixed-design repeated-measures ANCOVAs tested Condition (genius vs. control) × Time (pre vs. post) on value and expectancy outcomes separately, with gender as between-subjects factor and math grade as covariate. Assumptions were met (skewness/kurtosis within ±1; sphericity not violated).

• Baseline equivalence: No significant condition differences at pretest for pSTEM value, expectancy, math grade, gender, race/ethnicity, or year in school (all p > .05).
• Primary effect on Value beliefs: Significant Condition × Time interaction indicating a greater decline in value beliefs in the genius condition relative to control; F(1,210)=4.53, p=.03, ηp²=.021. Means: Genius pre M=3.41 (SD=0.81) → post M=3.19 (SD=0.83), Δ≈−0.22; Control pre M=3.40 (SD=0.78) → post M=3.33 (SD=0.68), Δ≈−0.07.
• Expectancy beliefs: No significant Condition × Time effect; F(1,210)=1.01, p ns, ηp²=.005.
• Gender moderation: Non-significant for both outcomes (Condition × Time × Gender, p=.840 for value).

Findings show that brief exposure to media portraying a computer science major as a genius reduced adolescents’ pSTEM value beliefs but did not alter expectancy beliefs. This supports situated expectancy-value theory by demonstrating that cultural stereotypes can depress the perceived value of pSTEM among youth. Results align with prior correlational research linking genius-stereotype endorsement to lower value and extend experimental work on nerd stereotypes affecting women’s interest. The lack of effect on expectancy may reflect that competence beliefs are more resistant to brief media exposure and may require sustained feedback or evaluation to shift. No gender moderation was observed, suggesting the genius stereotype’s devaluing effect may be broadly impactful, though individual differences (e.g., self-views of talent, stereotype endorsement) could shape susceptibility and warrant further study, particularly among underrepresented groups.

The study provides experimental evidence that genius-framed portrayals of computer scientists can diminish high school students’ valuation of pSTEM subjects. Educators and institutions should foster inclusive environments that avoid reinforcing genius stereotypes in classrooms, curricula, and media. Interventions aligned with growth mindset approaches may help mitigate stereotype effects. Future research should examine repeated or longitudinal media exposure, include varying character genders and identities, test identity and belonging outcomes alongside motivation, and explore moderators such as gender and race/ethnicity with larger, more diverse samples.

• Sample size and diversity were insufficient to test race/ethnicity as a moderator.
• The manipulation used only a male character; effects of protagonist gender were not examined.
• Brief, single-exposure manipulation limits conclusions about durability; longitudinal effects were not assessed.
• Outcomes focused on motivational beliefs; identity and belonging were not measured though they may also be affected.