Research on political participation almost unanimously assumes that social pressure by neighbors induces collective behavior. Yet most experimental studies focus on individually based forms of political and civic behavior, such as voting and recycling, in Western industrialized societies. The paper tests the effect of neighborly social pressure on collective action in Tunisia. In a field experiment, I manipulate whether neighbors or community outsiders invite citizens to contribute to a public good (i.e., trash collection). I run the experiment in three neighborhoods of varying socioeconomic composition in Tunis (n = 1199). I do not find evidence to suggest that neighborly social pressure encourages participation in neighborhood cleanups, with low participation rates both for the neighbor and outsider contact conditions. While the effect of social pressure does not significantly vary across communities, overall participation rates do. Residents of the poor neighborhood are most likely to respond in a socially desirable way when asked about their intentions but least likely to participate. The paper also discusses some limitations of the study and outlines avenues for future research.
I do not find evidence to suggest that neighborly social pressure encourages participation in neighborhood cleanups, with low participation rates both for the neighbor and outsider contact conditions.
very low attendance and unverified treatment salience limit the randomized comparison's ability to establish a substantive null
While the effect of social pressure does not significantly vary across communities, overall participation rates do.
limited moderation power and neighborhood attendance differences reported only at p<0.1 constrain the three-community comparison
Residents of the poor neighborhood are most likely to respond in a socially desirable way when asked about their intentions but least likely to participate.
the intention–attendance comparison cannot distinguish social desirability from an intention–behavior gap
In particular, I find suggestive evidence that communities may face coordination problems.
descriptive beliefs about neighbors do not experimentally identify a coordination mechanism
campaigns that solely aim to increase the general awareness of environmental pollution may fail to reach active engagement in environmental protection and collective action
awareness was measured rather than manipulated in a study of recruiter identity
Derived from the full evaluation — not a separate score.
Strengths
Households were assigned within neighborhoods to neighbor or outsider recruitment. The paper identifies preregistered hypotheses, adjusts for reported covariate imbalances, and discloses the added reminder and intention call.
↳ Introduction; Recruitment of Participants; Experimental Analysis
The study moves from individually based behaviors such as voting and recycling to a time-intensive neighborhood public good in Tunisia. Comparing poor, mixed, and wealthy neighborhoods adds contextual variation.
↳ Introduction; Study Design
The literature review includes contradictory socioeconomic findings. The results explain how low participation, possible priming, and limited recruiter familiarity constrain interpretation.
↳ Introduction, references 17–18; Statistical Results, limitations paragraphs
Limitations
Only 34 recruited respondents attended, approximately 3% of the sample. Recruiter familiarity was measured only among attendees, leaving the overall treatment contrast unverified and small-effect conclusions uncertain.
↳ Statistical Results
Participants were encouraged to recruit others within the same neighborhoods, without a reported contamination analysis. For stated intentions, 252 of 1,199 respondents were unreachable and treatment-specific reachability is not reported in the supplied text.
↳ Recruitment of Participants; Statistical Results
The intention–attendance gap does not distinguish social desirability from other explanations. Awareness was measured rather than manipulated, so the experiment does not establish the effectiveness of awareness-only campaigns.
↳ Abstract; Discussion and Conclusion
The randomized recruitment comparison supplies new evidence about collective action in Tunisia, with transparent disclosure of the reminder-call protocol change. However, approximately 3% attendance and uncertain treatment salience make the null less informative about modest effects or neighborhood moderation. Encouraged neighborhood referrals and unanalysed phone nonresponse add uncertainty without demonstrating failure of random assignment. The discussion meaningfully qualifies the headline result, but its social-desirability and awareness-policy interpretations exceed the design. Municipal relevance is concrete, while implementation guidance remains preliminary.
Nabu’s assessment, alongside the field’s view.
Are you an author of this paper?
Sound3.2
Confidence mediumThe study extends neighbor-versus-outsider canvassing research to time-intensive collective action in three Tunis neighborhoods. Low attendance and uncertain treatment salience limit the evidentiary advance from the null and moderation results.
Within-neighborhood household randomization, preregistered hypotheses, covariate adjustment, and disclosed protocol changes support the design. Unaddressed interference and intention-outcome nonresponse are priced into the score, alongside uncertainty about treatment salience and small-effect sensitivity.
The paper distinguishes intentions from attendance and phrases its headline as absence of evidence. The social-desirability assertion and awareness-policy inference imply conclusions that the design does not identify.
The literature discussion engages competing socioeconomic findings and connects design limitations to cautious interpretation of the null. Broader claims about norms and policy keep the score below the next anchor band.
Caveats4 of 4 checks
3 reported values do not reconcile, but the key claims do not rest on them.
Ethics approval, written consent, preregistration, and a protocol deviation are declared. The referral-count sentence repeats La Marsa, leaving one neighborhood label unresolved; the signed-rank terminology noted in the reviews is not sufficient to establish a result-changing calculation error.
Ethics approval: declared · Preregistration: declared
36 references in manuscript 29 of 31 checkable references found in an index 5 are books, websites or datasets — counted, but not index-checkable
References not verified automatically are resolved manually within 24 hours.
No retraction or expression of concern found in Crossref (includes Retraction Watch).
Sources: Crossref (incl. Retraction Watch) ✓
Where this paper’s evidence sits on the path from initial observation to real-world use.
Recruitment, reminders, and observed attendance were tested in an operational cleanup rather than a laboratory setting. Low uptake and the absence of an effective tested alternative prevent direct implementation guidance.
AI-generated, human-governed. Something look off? Contact us to request a review.
Keep reading
Other papers Nabu has evaluated blind, starting with the same field.
Akihiro Nishi, Hirokazu Shirado, David G. Rand, et al. · 2015 · Economic and Business Sciences
wealth visibility facilitates the downstream consequences of initial inequality—in initially more unequal situations, wealth visibility leads to greater inequality than when wealth is invisible
Well-supportedGermani, Spitale · 2026 · Philosophy, ethics and religion
prompting should therefore be understood as an epistemic and normative act
SupportedHo, Liu, Tran, et al. · 2026 · Computer and information sciences
After one recursive step, increasing exposure from 0% to 100% makes recommendation ratings more concentrated, independently generated reviews of the same paper about 11% more similar, and reviews across different papers about 5% less diverse.
Supported