Leading by example from high-status individuals: exploring a crucial missing link in climate change mitigation

The study addresses how visible low-carbon behaviour from high-status individuals (politicians and celebrities) influences public willingness to adopt similar behaviours, perceptions of leaders, and broader climate-related attitudes. It responds to a longstanding challenge: large-scale voluntary behaviour change has been difficult to achieve, while governments often avoid unpopular demand-side measures, creating a governance trap where each side waits for the other to act. The authors propose that leaders’ personal actions can send powerful signals beyond words, drawing on embodied leadership and credibility enhancing displays (CREDs). They pre-registered hypotheses testing whether leaders who lead by example increase observers’ willingness to act (vs leaders not leading by example and vs a disembodied message), improve perceptions of leaders (commitment, effectiveness, warmth/competence, knowledge, approval, and possible reactance), and shift perspectives on climate change (leaders’ efficacy, others’ willingness, moral salience/responsibility, pro-environmental identity, policy support, concern/risk), with political orientation as a potential moderator.

The paper situates the work within debates about the role of individual behaviour in climate mitigation, noting tensions between systemic change and individual responsibility and the frequent media focus on leaders’ perceived hypocrisy. Prior research shows both positive and negative social evaluations of pro-environmental behaviour: do-gooder derogation and stigma can produce backlash, yet pro-environmental actors are often perceived as warm and competent, and opinion leaders can shift norms. Theoretical framing includes: (1) Embodied leadership—leaders’ physical actions convey meaning, direction, and commitment beyond discourse; and (2) Credibility Enhancing Displays (CREDs)—costly, effortful actions signal genuine belief and increase credibility. Past studies find advocates with lower-carbon lifestyles are seen as more credible, and ambassadors with personal investments in solutions increase uptake. The authors argue that high-status individuals, who tend to have large carbon footprints and outsized social influence, are especially consequential norm referents. They contrast embodied with disembodied framings of behaviour change and hypothesize that visible, sustained, multi-behaviour changes from leaders may be particularly influential and credible.

Design and preregistration: A pre-registered, between-subjects survey experiment tested a 2 (Leader Action: Leading by Example [LBE] vs Not Leading by Example [NLBE]) × 2 (Leader Type: Politician vs Celebrity) design, plus a control (Disembodied, no leader). OSF preregistration: https://doi.org/10.17605/OSF.IO/83UXA. Sample and recruitment: UK nationally representative sample by gender, age, and ethnicity recruited via Prolific in April 2021. Target n=1300 (≈260 per condition) based on power analysis (medium-to-small effects, d≈0.15, 95% power, α=0.05) and prior studies. Compensation £1.25 for ~10 minutes. After manipulation and attention checks, valid n=1267. Randomization checks showed no significant demographic imbalances across conditions. Procedure and materials: Implemented in Qualtrics. After consent, participants were instructed to imagine watching either an interview with a politician they might vote for or a celebrity they like/admire (leader conditions) or a TV report about climate change (Disembodied control). All leader vignettes mentioned the need for new technology, international agreements, and behaviour change. The interviewer then asked what behaviour changes are needed; specified were: flying less, eating less meat, driving electric cars, improving home efficiency, and active travel—said to typically halve a person’s footprint. The interviewer then asked if the leader had already adopted these behaviours for two years (Yes in LBE; No in NLBE). The control used similar content without a leader. Direct quotes were avoided to keep focus on actions. Items within question blocks were randomized to mitigate order effects; the overall question order was fixed to allow priming. Manipulation and attention checks were included. Measures: - Willingness To Act: 7 Likert items (−3 to +3) on willingness to make significant lifestyle changes; fly less; eat less meat; switch to an electric car; use public transport more; improve home energy efficiency; make sacrifices. - Perceptions of leaders (leader conditions only): • Leader’s Climate Commitment (via PCA separated into): Cares/Believes; Knowledgeable; Exaggerates. • Effectiveness (persuasiveness, dedication/effort to climate action). • Warmth/Competence (including trustworthiness, honesty, morality, shared values, inspirational). • Reactance (felt preached at, told what to do, manipulated). • Increased Approval (vote likelihood for politician; like/admire more for celebrity). - Perspectives on climate change: Leaders’ efficacy; Others’ Willingness To Act; Moral Salience/Responsibility; Pro-environmental Identity; Support for Climate Action (tech investment; strong international agreements); Concern/Risk Perception. - Appetite for Leadership: 6 items on expectations that leaders should act first and effects on own willingness; also perceived collective action. - Generalised Trust; demographics; perceived leader gender; survey topic identification. Analysis: Novel scales underwent factor/component analysis for reliability and dimensionality. Confirmatory and exploratory tests included two-way omnibus MANOVA (Leader Action, Leader Type; no control), follow-up ANOVAs with Bonferroni correction (p<0.003 for 15 tests), MANOVAs including the Disembodied control for relevant DVs, Tukey’s HSD for multiple comparisons, and tests of political orientation moderation on Willingness To Act. Ethics approval: Cardiff University (ref: EC.21.01.12.6239R2A).

- Strong main effects: Leader Action significantly affected outcomes (omnibus MANOVA F(15,989)=107.9, p<0.001; Wilks’ λ=0.379; η²=0.621). Leader Type had a smaller multivariate effect (F(15,989)=5.89, p<0.001; λ=0.918; η²=0.082). No significant interaction. - Willingness To Act: LBE increased willingness vs NLBE (M=1.347 vs 1.139; F(1,1007)=8.94, p=0.003; d=0.19; η²=0.009). Compared to Disembodied, Disembodied > NLBE (p=0.004) and LBE > NLBE (p=0.007); Disembodied ≈ LBE (ns). Item-level: Disembodied and LBE > NLBE for willingness to make significant lifestyle changes, use public transport more, and make some sacrifices; LBE > NLBE for flying less. No significant differences for eating less meat, switching to an electric car, or improving home energy efficiency. - Perceptions of leaders (all p<0.001, large/medium effects, unless noted): • Cares/Believes: LBE (M=2.199, SD=0.762) > NLBE (M=−0.094, SD=1.196); F=1321.261; η²p≈0.568. • Warmth/Competence: LBE (M=1.248, SD=0.986) > NLBE (M=0.045, SD=1.167); F=314.011; η²≈0.238. • Effectiveness: LBE (M=0.860, SD=0.515) > NLBE (M=0.139, SD=0.701); F=347.184; η²≈0.257. • Knowledgeable: LBE (M=1.805, SD=1.022) > NLBE (M=1.195, SD=1.181); F=76.986; η²≈0.071. • Increased Approval: LBE (M=0.746, SD=0.796) > NLBE (M=−0.304, SD=0.986); F=349.584; η²≈0.258. Politician garnered higher approval than celebrity overall (M=0.332 vs 0.126; p=0.001). • Exaggerates: no significant LBE vs NLBE difference. • Reactance: NLBE produced higher reactance (M=0.284) than Disembodied (M=−0.718) and LBE (M=−0.495) (both p<0.001). LBE did not increase reactance. - Perspectives on climate change: Mostly null. No significant differences for Others’ Willingness To Act, Leaders’ Efficacy, Moral Salience/Responsibility, Support for Climate Action, Concern/Risk Perception. Pro-environmental Identity: Disembodied > NLBE (p=0.020); other comparisons ns. - Political orientation: No significant moderation of LBE effects on Willingness To Act, though descriptive patterns suggest LBE may boost willingness among the political right. - Appetite for leadership (descriptive across conditions): 86% agreed leaders should set an example by acting first; 79% disagreed that leaders’ personal behaviour is irrelevant; 77% agreed everyone should change at about the same time; 90% agreed biggest emitters should change most; 53% said they’d be more willing to change if leaders went first; 64% said they’d be more willing if most others were changing.

Findings support that visible, effortful low-carbon actions by high-status individuals function as credibility enhancing displays within an embodied leadership framework. Leading by example increased public willingness to adopt impactful behaviours relative to leaders who did not model such actions and substantially enhanced perceptions of leaders’ commitment, knowledge, trustworthiness, warmth/competence, effectiveness, and approval—key qualities for influence and policy leadership. Contrary to concerns about do-gooder derogation, leading by example did not trigger reactance or perceptions of exaggeration; rather, failing to lead by example increased reactance, suggesting that advocacy without congruent personal action can undermine public enthusiasm. While a “disembodied” message matched LBE on willingness, the real-world salience of leaders’ behaviour and frequent media scrutiny of hypocrisy make disembodied approaches difficult to sustain. Effects were generally similar for politicians and celebrities, though politicians saw slightly greater approval gains, implying potential to rebuild trust in political leaders through embodied, consistent actions. Although the net effect on willingness was small in this one-shot vignette, the authors argue real-world, repeated modelling by multiple leaders could amplify influence through social networks and norm change, potentially engaging constituencies (including the political right) less inclined toward climate action by aligning with values of self-regulation and personal responsibility.

The study provides experimental evidence that embodied leadership—leaders visibly adopting multiple high-impact low-carbon behaviours—can raise public willingness to act and markedly improve perceptions of leaders’ credibility, trustworthiness, competence, and effectiveness. These signals go beyond reducing a leader’s own emissions, potentially catalysing broader social change and helping to escape the governance trap wherein publics and governments wait on each other. The authors propose leading by example as a crucial missing link in climate mitigation and outline practical recommendations for leaders: frame personal action as part of a broader solution set; adopt behaviour suites that meaningfully cut footprints; communicate impact when asked (licence to preach); be consistent over time; and acknowledge others’ constraints. Future research should test real-world behavioural impacts, multiple leader exposures, diverse leader types and contexts, and differences between single vs multiple behaviours across cultures.

- Outcome measures emphasize willingness rather than observed behaviour; translation to actual behaviour change remains untested. - Vignette realism: sharp contrasts between LBE and NLBE may be less clear in practice; leaders might avoid discussing personal behaviour rather than explicitly admitting not acting. - Single-country sample (UK) limits generalisability; cultural variation in responses is unknown. - One-off fictional leader exposure likely underestimates cumulative or network effects of repeated, multi-leader modelling. - Limited leader types (politician, celebrity) and simplified stimuli; business, religious, local community leaders and more nuanced communication strategies were not examined. - Focus on a suite of behaviours; comparative effects of single vs multiple behaviours need study. - Political moderation analyses may have been underpowered after subgrouping.