Scientists' identities shape engagement with environmental activism

The study examines whether and how scientists’ social identities shape their engagement in environmental activism. Against the backdrop of climate change and biodiversity loss, and a gap between scientific consensus and policy action, the paper situates growing scientist participation in social movements and the tension between traditional norms of scientific objectivity/neutrality and advocacy. Drawing on social identity theory and literature on collective action, the authors focus on potential conflicts and fit between scientist and activist identities. The key research questions test whether: (1) stronger scientist identification predicts greater climate activism; (2) stronger activist identification predicts activism; (3) perceiving science and activism as compatible predicts activism; and (4) scientists who strongly identify as scientists but perceive science–activism as incompatible are more likely to endorse techno-solutionism. The study also explores how identity content relates to support for technological fixes versus political/system change.

The paper reviews the rise of scientist activism (e.g., March for Science; groups such as Scientists for Extinction Rebellion and Scientist Rebellion) and debates over the appropriate role of scientists in advocacy. Traditional arguments emphasize maintaining objectivity and impartiality by separating science from advocacy to protect credibility, while others argue scientists have social and intellectual responsibilities to act given the robustness of climate science. Social identity theory and models of politicized identities are discussed as frameworks for understanding engagement, highlighting how group identification, norms, and values guide collective action. Prior surveys indicate barriers such as low perceived efficacy, uncertainty about effective actions, lack of activist networks, high workloads, and institutional constraints, and suggest a divergence between willingness to engage and actual participation. The review also introduces techno-solutionism—a belief in technical fixes for complex social problems—as a potentially attractive pathway for scientists that may sidestep political action.

Design and preregistration: Cross-sectional, mixed-methods study (quantitative survey with open-ended responses), preregistered on OSF. Data and analysis code are publicly available.

Sample and recruitment: 329 natural and social scientists from 41 countries (41.64% UK; 14.29% USA; 7.3% Germany; 4.56% Australia; 3.65% Ireland). Gender: 54.1% female, 40.7% male, 2.4% non-binary, 2.7% prefer not to say; age M=40.11, SD=12.03, range 22–77. About 53% reported membership in an environmental activist group. Disciplines included psychology (28%), biology (20%), earth science (13%), sociology (13%), engineering/technology (8%); roughly half natural sciences (n=159) and half social sciences (n=169). Recruitment (12/02/2022–01/10/2022) via Twitter and scientific societies/centers; participants were unpaid. Inclusion targeted scientists concerned about climate change (engaged or not in activism). Power analysis (targeting r≥0.15, 90% power, alpha .05) suggested N=374; finalized sample allowed detection of r≈0.19 at revised alpha.

Measures:

• Scientist identity strength: 8 items adapted from validated social identity scales (1–7 Likert; half reverse-scored), α=0.91; M=5.26 (SD≈1). Higher scores indicate stronger identification as a scientist.
• Environmental activist identity strength: 8 items (1–7; half reverse-scored), α=0.86; M=4.73 (SD=1.27).
• Scientist–activist compatibility (identity content): 4 items assessing perceived compatibility, objectivity, impartiality, and duty (two reverse-scored). 1–5 Likert; PCA supported a single index, α=0.76 (loadings 0.70–0.84); M=4.20, SD=0.70. Two additional items assessed reputational/credibility concerns, α=0.66; M=3.21, SD=0.92.
• Climate change risk perceptions: 2 items (1–5) on perceived impact to self and close others; M(self)=4.66 (SD=0.54); M(others)=4.73 (SD=0.51).
• Activism engagement: Adapted Activism Orientation Scale (20 behaviors; frequency 0–3 from never to often); α=0.91; M=1.45 (SD=0.56).
• Techno-solutionism: Single item “Inventing new technologies is the only way to successfully tackle climate change” (1–5); M=2.04 (SD=1.06). Two alternative “only way” items on political systems and human behavior included.
• Additional factors potentially affecting engagement: 12 items (1–5) including work, family, financial constraints, transport, visa/residency concerns, uncertainty about effectiveness and actions, lack of interest, fears of others’ views, lack of energy, COVID-19 concern, and not knowing activists. Items analyzed individually (α for composite=0.62; PCA did not yield reliable subscales).

Procedure: After consent, participants completed scientist and activist identity scales (counter-balanced), then scientist–activist compatibility, climate risk perceptions, activism engagement, solutions to climate change, additional engagement factors, activist group membership, and demographics. Within-scale item order was randomized.

Qualitative component: Open-ended questions asked for “3 things that prevent people like you from taking action” and “3 things that people like you gain from taking action.” Thematic analysis was conducted inductively following Braun and Clarke’s approach: coding, theme development, review/refinement, and naming; responses were organized and tracked in spreadsheets.

Analysis plan: For activism (normally distributed), standard multiple regression with stepwise blocks: (1) risk perceptions, scientist–activist compatibility, additional factors, age; (2) add activist identity; (3) retain significant predictors for a parsimonious model. For techno-solutionism (ordinal, positively skewed), cumulative link models (ordinal regression) with analogous steps. Bonferroni corrections were applied for multiple testing. Model diagnostics included VIFs (<3), bootstrap validation (10,000 iterations), and outlier checks (leverage, Cook’s distance, covariance ratios); retaining all data (approx. 2.7% flagged did not change conclusions). Group comparisons (activist group members vs non-members) used Welch’s t-tests and Wilcoxon rank-sum tests. PCA used oblimin rotation and parallel analysis.

• Scientist identity strength was not significantly correlated with activism engagement overall: r(327)=0.08, p=0.17 (H1 not supported). Perceived reputational/credibility harm was also not significantly correlated: r(327)=-0.09, p=0.09.
• Components of scientist–activist compatibilism (objectivity and impartiality not compromised by activism, duty of stewardship, and ability to be both scientist and activist) each positively correlated with engagement, r=0.25 to 0.36 (all p<0.001). PCA yielded a composite compatibilism index (α=0.76) strongly associated with activism engagement, r(327)=0.42, p<0.001. Mean compatibilism was high (M=4.2, SD=0.70; 5.5% disagreed; 1.82% neutral).
• Final regression model for activism engagement: R²=0.52, F(4,324)=90.13, p<0.001. Significant predictors: activist identity (largest unique variance; H2 supported), scientist–activist compatibilism, age (positive), and interest in activism (lack of interest negative). All retained significance after Bonferroni correction; no multicollinearity (VIFs<3); bootstrap CIs aligned with original estimates. Compatibilism and scientist identity strength were orthogonal (r=0.02, p=0.70); no interaction between scientist identity strength and compatibilism on engagement.
• Additional factors: Uncertainty about action effectiveness, uncertainty about which actions to take, and lacking personal connections to activists were negatively correlated with engagement. Family commitments and COVID-19 impacts showed positive correlations with engagement. Financial and work commitments, transport access, and visa/residency concerns showed weak, non-significant positive correlations.
• Techno-solutionism: Most disagreed or were unsure (74% disagree; 15% unsure; 11% endorse). Scientist identity strength was not associated with techno-solutionism (χ²(1,N=329)=2.67, OR=1.02 [1.00,1.05], p=0.10). Higher scientist–activist compatibility predicted lower odds of endorsing techno-solutionism (χ²(1,N=329)=36.76, OR=0.80 [0.74,0.86], p<0.001) and higher odds of endorsing political system change as the sole solution (χ²(1,N=329)=11.12, OR=1.13 [1.05,1.22], p<0.001).
• Group comparisons: Non-group members were younger, less interested and engaged in activism, identified less as activists, were more uncertain about action effectiveness, perceived lower science–activism compatibility, reported greater concern about others’ opinions, and were relatively more supportive of techno-solutionism (though on average still disagreed). Scientist identity strength did not differ between groups. Among activist group members only, scientist identity strength correlated with activism (r(172)=0.25, p<0.001) and had a significant independent association in regression; however, adding activist identity rendered scientist identity non-significant, underscoring activist identity’s primacy.
• Thematic analysis revealed five constructions: (1) traditional views prioritizing objectivity/integrity and seeing activism as risky; (2) activism as professional/ethical obligation; (3) managing the reputation of the ‘scientist’ and perceived objectivity; (4) supporting activists rather than participating directly; and (5) a techno-solutionist stance emphasizing scientific/technological work over activism.

Findings show that the content of scientist identity—particularly beliefs about compatibility between scientific roles and activism, stewardship obligations, and maintenance of objectivity and impartiality—better explains activism engagement than merely the strength of identifying as a scientist. Activist identity remains the strongest predictor, but compatibilism contributes substantial unique variance and is largely independent of scientist identity strength, indicating that how scientists construe their roles and norms is pivotal. Qualitative analyses illuminate the nuanced ways scientists manage tensions between credibility and advocacy, ranging from adherence to traditional norms to framing activism as part of professional duty, adopting supportive roles, or favoring technological solutions. The negative association between compatibilism and techno-solutionism suggests that construing science and activism as incompatible may channel preferences toward technological fixes and away from political or collective action. Age-related effects may reflect career security or lifecycle dynamics enabling older scientists to engage more. The role of perceived efficacy and activist networks aligns with broader literature, emphasizing the importance of social context and collective/personal efficacy in facilitating activism.

Scientist identity content—especially beliefs about science–activism compatibility, stewardship, and preservation of objectivity—plays a key role in explaining scientists’ engagement in environmental activism, beyond the influence of activist identity and other factors. Compatibilist identity content is associated with greater activism and reduced endorsement of techno-solutionism, while activist identity remains the strongest single predictor of engagement. The study contributes a nuanced understanding of how scientists construct and integrate professional and activist roles, informing strategies to support effective scientist engagement on climate issues. Future work should track how identity content evolves with engagement, broaden samples across disciplines and regions, and examine causal pathways via longitudinal and qualitative designs.

• Sampling and generalizability: Opportunity sampling and overrepresentation of scientists with activist affiliations and of the Global North (particularly the UK) limit generalizability; some fields may be overrepresented. Political ideology in the sample skewed liberal.
• Design: Cross-sectional and correlational; causal inferences cannot be made. Timing and mode of recruitment (Twitter, societies) may bias participation.
• Measures: Some constructs measured with brief/new items (e.g., single-item techno-solutionism; two-item reputation concern). Additional factors did not form a reliable composite and were analyzed individually.
• Reputational concerns: Though not statistically related to engagement here, qualitative data suggest context-sensitive perceptions of credibility that warrant further study.
• Qualitative scope: Open responses indicate additional unmeasured influences; richer qualitative/ethnographic work is needed.