What the evidence says about communications and deep change

This distils two internal research briefs on how ontology-level change happens, produced through multi-perspective research with independent citation review (both graded B+ by that review — meaning the citations were checked, not that the claims are certain). Full sources below; honesty notes travel with the numbers.

Five findings

1. Information alone is the weakest lever we have measured. In the largest second-order meta-analysis of climate field interventions, education sits at d = 0.09 — the weakest category in its table — while appeals to values and responsibility reach d = 0.28 [1]. So messaging is not dead; arguing is. Honesty note: these studies measure behaviours (recycling, conservation), not worldviews — the extension to deep dispositions is our inference from convergence across literatures, not the paper’s claim. 2. Broadcast media reliably moves norms, salience and behaviour — not personal beliefs. The strongest field test of repeated narrative media is Paluck’s year-long radio experiment in Rwanda: perceived norms and behaviour moved; it “did little to change listeners’ personal beliefs” [2]. The Lancet’s review of mass-media campaigns attributes success to “concurrent availability of required services and products… and policies that support behaviour change” [3]. The truth campaign — advertising’s best “deep” case — accounted for roughly 22% of a four-year youth-smoking decline, with tax, law and trend doing the rest [4]. What campaigns demonstrably do well: correct false norms (most people underestimate how much others share intrinsic values [5]), name what was illegible, legitimise change after the ground has moved, and recruit people into the other zones. 3. The form of communication that does move beliefs is dialogic. Deep canvassing — ten-minute, non-judgmental, perspective-taking conversations — durably reduced prejudice in randomised field experiments, with effects lasting months [6]. It is communication behaving like relationship: one-to-one encounter, personal narrative, the listener’s own experience doing the work. The exception that proves where the mechanism lives. 4. Lived and built change converts attitudes at population scale — when it arrives with local legitimacy. Ireland’s workplace smoking ban: support among bar workers who themselves smoked rose from 39% to 67% within a year [7]. England’s 5p bag charge changed behaviour across every demographic within a month and increased support for further charges [8]. US same-sex-marriage legalisation roughly doubled the rate of decline in anti-gay bias — except where it arrived only by federal imposition, where bias increased [9]. The gate is ownership: structure converts when communities pass it, and can backfire when it is done to them. 5. Practice carries the only direct causal evidence on connection-level measures — and channel multiplication without variety is null. A 582-person randomised trial of prompted noticing of urban nature produced sustained gains [10]; the field’s meta-analysis puts sustained effects specifically with repeated practice [11]. Meanwhile the classic multi-channel community campaigns (COMMIT, Minnesota Heart Health) were null [12][13], and an overview of 25 systematic reviews found “no compelling evidence that multifaceted interventions are more effective than single-component interventions” [14]. What seems to work is coupling different kinds of work — story to practice to structure — rather than more channels of the same kind. We hold that as a corroborated hypothesis, not a result: nobody has randomised the combinations.

What we take from this

Don’t ask communications to perform the shift; ask it to do what it demonstrably does — name, normalise, legitimise, recruit — and end every story in something people can join or do. Put the conversion weight on practice, participation and locally owned structural change. And fund the coupling, because that is the part nobody’s single-channel budget covers.

Sources

1. Bergquist, Thiel, Goldberg & van der Linden, ‘Field interventions for climate change mitigation behaviors: a second-order meta-analysis’, PNAS 120(13), 2023
2. Paluck, ‘Reducing Intergroup Prejudice and Conflict Using the Media: A Field Experiment in Rwanda’, JPSP 96(3), 2009
3. Wakefield, Loken & Hornik, ‘Use of mass media campaigns to change health behaviour’, The Lancet 376, 2010
4. Farrelly et al., ‘Evidence of a dose–response relationship between “truth” antismoking ads and youth smoking prevalence’, AJPH 95(3), 2005
5. Common Cause Foundation, ‘Perceptions Matter’, 2016
6. Broockman & Kalla, ‘Durably reducing transphobia: A field experiment on door-to-door canvassing’, Science 352(6282), 2016
7. Before and after study of bar workers’ perceptions of smoke-free legislation, BMC Public Health 7:131, 2007
8. Thomas et al., ‘The English Plastic Bag Charge Changed Behavior and Increased Support for Other Charges’, Frontiers in Psychology 10, 2019
9. Ofosu et al., ‘Same-sex marriage legalization associated with reduced implicit and explicit antigay bias’, PNAS 116(18), 2019
10. McEwan et al., ‘A Smartphone App for Improving Mental Health through Connecting with Urban Nature’, IJERPH 16(18), 2019
11. Sheffield, Butler & Richardson, ‘Improving Nature Connectedness in Adults: A Meta-Analysis, Review and Agenda’, Sustainability 14(19), 2022
12. Community Intervention Trial for Smoking Cessation (COMMIT): I. cohort results, AJPH 85(2), 1995
13. Luepker et al., Minnesota Heart Health Program, AJPH 84(9), 1994
14. Squires et al., ‘Are multifaceted interventions more effective than single-component interventions?’, Implementation Science 9:152, 2014