The Applied Behavioural Sciences in Commercial and Policy Practice

Submitted as a graduate literature review in applied consumer psychology and behavioural economics.

Abstract

This paper examines the practitioner-oriented behavioural science literature that has crystallised into a de facto canon for commercial, charitable, and governmental behaviour-change work over the past two-and-a-half decades. Using a single corpus of twenty texts identified by an industry-practitioner reading list, the review applies four complementary methodologies - narrative, systematic, scoping, and meta-analytic - each calibrated to the type of question it can legitimately answer. The narrative review traces the field's intellectual evolution from the heuristics-and-biases tradition through the institutionalisation of nudge units to its current critical-maturing phase. The systematic component codes twelve in-scope texts against a seven-category taxonomy of psychological mechanisms, finding that heuristics-and-biases dominate commercial application, social influence and identity dominate charitable application, and choice architecture dominates governmental application. The scoping review identifies five domain gaps in the corpus: methodological reflexivity, post-2020 algorithmic persuasion, cross-cultural generalisability, long-term effect decay, and ethical-philosophical engagement. The meta-analytic synthesis triangulates effect sizes from the primary research underlying the corpus against three recent meta-analyses (Hummel & Maedche, 2019; Mertens et al., 2022; DellaVigna & Linos, 2022) and a publication-bias correction (Maier et al., 2022), and concludes that the practitioner canon systematically overstates the magnitude of behavioural interventions by a factor of approximately five to six relative to corrected, in-field effect sizes. The paper closes with implications for practitioners who train from this canon: the books remain operationally useful but require pairing with methodological-critical literature to avoid the inferential errors that the canon's confident register tends to reproduce.

Keywords: behavioural science, applied psychology, nudging, replication crisis, choice architecture, persuasion, literature review

1. Introduction

The applied behavioural sciences have, since approximately 2000, generated a distinctive practitioner literature that occupies the conceptual space between peer-reviewed psychology and trade-press business writing. Authored predominantly by academics writing for general audiences (Kahneman, 2011; Thaler & Sunstein, 2008; Cialdini, 2016), by practitioners with academic training (Halpern, 2015; Ariely, 2008; Barden, 2013), and by senior commercial figures synthesising behavioural research for marketing application (Sutherland, 2019), the corpus now functions as the de facto curriculum for entry into the field. Boutique consultancies, government nudge units, and in-house behavioural science teams draw their working vocabulary, primary frameworks, and citation practices from a relatively small set of canonical texts.

This paper takes one such practitioner reading list - compiled for the use of staff at an Australian behavioural-marketing studio working across commercial ("Commerce"), charitable ("Cause"), and governmental ("Culture") engagements - as its operational corpus. The list comprises twenty book-length texts and three ongoing serial subscriptions. Its compiler explicitly describes the selection as "deliberately tight" and instrumental: each text is included for "specific work it does for the practice." The list therefore offers a useful methodological proxy for the working canon of contemporary applied behavioural science as it is read, taught, and operationalised.

A single-method review would be poorly suited to this corpus. A narrative review alone would describe but not measure. A systematic review alone would impose categorical rigour on a body of work that resists clean inclusion criteria. A scoping review alone would map without engaging the underlying evidence. A meta-analysis is technically impossible against trade-press books, since those books do not themselves contain the raw effect sizes that meta-analysis aggregates. Each method, used in isolation, would leave decisive questions about the literature unanswered.

This paper therefore deploys all four methods in sequence, each calibrated to the question it can legitimately answer. Section 2 specifies the methodology and acknowledges its limitations. Section 3 presents a narrative review of the field's emergence. Section 4 conducts a systematic coding of the in-scope corpus against a mechanism taxonomy. Section 5 scopes the boundaries of what the corpus does and does not cover. Section 6 synthesises effect sizes from the primary studies that the corpus relies on, in dialogue with three recent meta-analyses of the broader nudging literature. Sections 7 and 8 discuss implications and limitations.

The central finding, previewed here, is that the practitioner corpus is internally coherent and operationally useful but treats the underlying empirical science as more settled than current evidence supports - a gap that practitioners who train from the canon can close only by pairing it with the critical-methodological literature it largely omits.

2. Methodology

The corpus consists of twenty book-length texts and three serial sources as identified in the source reading list. For the analytic portions of this review, three texts are excluded as falling outside the behavioural-science substantive domain: Rumelt (2011) on strategy formation, Minto (1987) on structured business writing, and Zinsser (1976) on prose craft. Ogilvy (1983) is retained for the narrative review but excluded from the systematic coding because it pre-dates the modern behavioural-science synthesis and serves the corpus primarily as a stylistic referent. The Behavioural Insights Team's MINDSPACE and EAST frameworks are treated as a single source unit (Behavioural Insights Team, 2010, 2014). This produces a net systematic-review corpus of twelve texts.

Each method is operationalised as follows.

The narrative review (Section 3) follows the conventions described by Greenhalgh et al. (2018): a structured prose synthesis ordered by intellectual chronology, foregrounding the conceptual development of the field rather than the aggregation of effects. It does not apply formal inclusion criteria beyond the corpus boundary.

The systematic review (Section 4) follows the structured-coding logic of PRISMA-adjacent reviews adapted for a small-corpus theoretical synthesis (Page et al., 2021). The research question is pre-specified: Across the in-scope corpus, which psychological mechanisms are most consistently invoked to explain consumer and citizen behaviour change, and how do mechanism-attributions vary across the three application tiers (Commerce, Cause, Culture)? Inclusion criteria, coding scheme, and inter-rater reliability procedures are specified in Section 4.

The scoping review (Section 5) follows the framework of Arksey and O'Malley (2005) as refined by Levac et al. (2010), mapping the extent, range, and nature of what the corpus covers and identifying the substantive territory it leaves uncovered.

The meta-analytic synthesis (Section 6) departs from the conventional definition of meta-analysis (Glass, 1976) in an important respect. The trade-press books in the corpus do not themselves report extractable effect sizes; they cite and paraphrase primary studies that do. The synthesis therefore aggregates effect sizes from the primary studies most heavily cited within the corpus, supplemented by three recent meta-analyses of the broader nudging literature (Hummel & Maedche, 2019; Mertens et al., 2022; DellaVigna & Linos, 2022) and one publication-bias correction (Maier et al., 2022). This is a meta-analytic synthesis of the corpus's underlying evidence base, not of the corpus itself.

Three methodological limitations are flagged in advance. First, the operational corpus is one practitioner's list, not a systematically sampled population of behavioural-science texts; the findings should therefore be read as descriptive of contemporary practitioner training rather than of the academic field as a whole. Second, the systematic coding scheme is theoretical rather than empirically derived and inherits the biases of its source taxonomy. Third, the meta-analytic synthesis aggregates studies of varying methodological quality and population characteristics; the resulting figures are indicative rather than precise.

3. Narrative Review

The development of applied behavioural science as a coherent practitioner discipline can be parsed into five overlapping phases, each represented in the corpus.

Phase one: heuristics-and-biases foundations (1970s-1990s). The intellectual prehistory of the field is dominated by Kahneman and Tversky's research program on systematic deviations from rational-actor models of decision-making, published across more than two decades of journal articles culminating in the Nobel Memorial Prize in Economic Sciences (Kahneman, 2002) and synthesised for general readership in Kahneman (2011). The corpus also draws on Cialdini's (1984, 2007) parallel program on social influence, distilled from a year of participant observation across compliance professions including sales, fundraising, and cult recruitment. Both research programs were already mature when the practitioner literature began to crystallise, providing it with a stock of robust-seeming findings to draw upon.

Phase two: crystallisation and popular accessibility (2000-2008). The corpus represents this phase through Ariely (2008) and Thaler and Sunstein (2008). Both books translate the heuristics-and-biases and choice-architecture literatures for non-academic audiences. Nudge in particular reframes a body of academic findings as a normative-political programme - what Thaler and Sunstein term "libertarian paternalism" - that proved immediately compelling to policy actors. Within two years of the book's publication, the United Kingdom established the world's first dedicated government nudge unit (Halpern, 2015).

Phase three: institutionalisation (2008-2014). The corpus represents this phase through Halpern (2015) and the Behavioural Insights Team's published frameworks (Behavioural Insights Team, 2010, 2014). The institutional development of dedicated behavioural-science units within governments, public health bodies, and major corporations created a feedback loop: institutional demand legitimised the field, and the field's expanding institutional footprint legitimised the underlying science. Kahneman (2011) appeared at the midpoint of this phase and consolidated public familiarity with terms such as System 1, System 2, anchoring, and availability.

Phase four: domain diversification (2014-2019). As behavioural science became established, it diversified into specialised applications. The corpus represents this through Eyal (2014) on technology and habit formation, Barden (2013) on commercial marketing, Cialdini (2016) on antecedent influence ("pre-suasion"), and Sutherland (2019) on commercial application within the advertising industry. This phase also produced significant works on charitable application (Andresen, 2006, though predating it) and on behavioural science in user-experience design (Evans, 2018).

Phase five: critical maturation (2019-present). The most recent phase of the field, partially represented in the corpus, is marked by critical engagement with the limits and ethics of behavioural intervention. Sharot's (2017) neuroscience-grounded treatment of belief change, Chance's (2022) explicit grappling with the ethics of influence, and the broader post-replication-crisis re-evaluation of foundational findings (Open Science Collaboration, 2015; Camerer et al., 2018) characterise this phase. The corpus is somewhat thin here - a finding pursued in Section 5.

Across all five phases, the corpus exhibits a consistent register: confident, applied, sceptical of academic abstraction but reliant on academic credentialing for its authority. The inclusion of Ogilvy (1983) as a stylistic exemplar is telling. Ogilvy pre-dates modern behavioural science but exemplifies the posture the corpus aspires to: commercial seriousness, intellectual restraint, evidence-orientation. The corpus is, in this sense, as much a project of professional self-fashioning as it is of empirical synthesis.

4. Systematic Review

4.1 Research question and protocol

The systematic review is structured around the following research question: Across the in-scope corpus, which psychological mechanisms are most consistently invoked to explain consumer and citizen behaviour change, and how do mechanism-attributions vary across the three application tiers (Commerce, Cause, Culture)?

Inclusion criteria for the systematic coding required that a text (a) be present on the source reading list, (b) substantively engage with empirical behavioural science, and (c) propose mechanism-level explanations for behaviour. Twelve of the twenty texts met these criteria. Excluded texts (Rumelt, Minto, Zinsser, Ogilvy, and the four texts focused on strategy, communication craft, or pre-modern advertising) are reserved for the narrative and scoping treatments.

A seven-category coding scheme was applied, derived from a synthesis of the West, Sheeran and Michie (2019) and Michie et al. (2013) behaviour-change taxonomies, modified to capture the categories most commonly invoked in the practitioner literature:

• Heuristics and biases - anchoring, availability, framing, loss aversion, decoy effects

• Social influence - social proof, normative influence, authority, conformity

• Identity and self-concept - consistency, commitment, signalling, narrative-identity

• Affective processes - emotion, fear appeals, hope, hedonic forecasting

• Choice architecture - defaults, friction, salience, ordering effects

• Habit and reward - cues, variable reinforcement, trigger-action loops

• Trust and credibility - source effects, narrative, perceived expertise

Each of the twelve texts was coded for the prominence of each mechanism category, on a three-point scale: peripheral (mentioned but not central), substantial (treated as a significant explanatory category), and central (treated as a primary organising mechanism). Coding decisions used the texts' table-of-contents structure, chapter weighting, and citation patterns as evidence.

4.2 Findings

Table 1 summarises the mechanism-prominence coding across the twelve in-scope texts, grouped by application tier.

Three patterns emerge from the coding.

First, mechanism distribution differs systematically by application tier. Commerce-oriented texts cluster around heuristics-and-biases (Kahneman, Ariely, Barden) and choice architecture (Thaler & Sunstein, Evans), with habit/reward operating as a sub-specialism for technology and retention work (Eyal). Cause-oriented texts emphasise social influence, identity, and affective processes (Sharot, Chance), with trust elevated as a primary category - consistent with the sector's structural challenge of source credibility. Culture-oriented work, represented here only by Halpern, prioritises choice architecture above all other categories, reflecting the BIT's policy-design orientation.

Second, the foundational texts function as cross-tier suppliers of mechanisms. Kahneman, the Cialdinis, and Thaler and Sunstein are cited across all three application tiers; the tier-specific texts inherit their mechanism vocabulary from these foundational works. This pattern is consistent with what Greenhalgh et al. (2018) describe as a "seed-and-branch" structure in mature applied literatures.

Third, identity emerges as the most under-recognised cross-cutting mechanism in the Commerce corpus. Although identity-based explanations dominate the Cause-tier texts (Sharot, Chance) and appear centrally in the most recent foundational work (Cialdini's Pre-Suasion, with its addition of Unity as the seventh principle of influence), the Commerce-tier texts treat identity peripherally or not at all. This is a substantively important finding: the commercial application of behavioural science, as represented by this corpus, may be systematically under-leveraging identity-based mechanisms relative to the available evidence base.

5. Scoping Review

The scoping review asks a different question: not what the corpus contains, but what it does not. Five substantive territories are under-represented or absent.

5.1 Methodological reflexivity and the replication crisis. The most consequential gap. Of the twelve in-scope texts, none engages substantively with the replication crisis that has reshaped large portions of the underlying literature (Open Science Collaboration, 2015; Camerer et al., 2018). Specific affected findings include several that the corpus relies on heavily: priming effects (Doyen et al., 2012; Harris et al., 2013, on the failure to replicate elderly-walking priming), money-priming (Rohrer, Pashler, & Harris, 2015), ego-depletion (Hagger et al., 2016), and the broader category of "social priming" that informs much of Cialdini's Pre-Suasion (2016). The reading list's compiler flags this gap in the marginalia ("Behavioural economics has its share of overclaiming and methodological controversy") but the books themselves remain confident-register popular syntheses. Additionally, the field has been shaped by documented research-integrity failures in directly relevant areas, including the Wansink retractions (Cornell University, 2018) and the data-fabrication findings affecting the 2012 signing-at-the-top dishonesty study co-authored by Ariely (Simonsohn, Simmons & Nelson, 2021), which raises a particular concern given Ariely's prominence in the corpus.

5.2 Algorithmic and platform-mediated persuasion. The corpus is largely pre-algorithmic. Eyal (2014) on technology-induced habit formation is the closest treatment, but contemporary developments - large-language-model-mediated persuasion (Salvi et al., 2024), recommender-system-shaped preference formation, and the dark-pattern literature on digitally manipulative design (Mathur et al., 2019; Gray et al., 2018) - are absent. For a practitioner working post-2024, this is a substantial blind spot.

5.3 Cross-cultural validity. Every in-scope text is anglophone-Western in both authorship and evidence base. The systematic limitations of WEIRD-sample psychology (Henrich, Heine & Norenzayan, 2010) - that findings from Western, Educated, Industrialised, Rich, and Democratic populations may not generalise - are not engaged. This is non-trivial: norm-based interventions in particular are known to be culturally sensitive (Schultz et al., 2007 versus the more variable replications in non-Western contexts), and the practitioner who applies the canon's defaults to non-Western audiences should be on notice.

5.4 Long-term effect decay. The corpus implicitly treats behavioural interventions as durable. The empirical literature increasingly suggests otherwise: Allcott and Rogers (2014) document the decay of normative-feedback effects on household energy use over time, and Frey and Rogers (2014) more broadly characterise nudges as often producing temporary rather than sustained effects. This temporal dimension is largely absent from the corpus.

5.5 Ethical and political-philosophical engagement. Chance (2022) is the principal exception, and Sutherland (2019) gestures at the question, but the corpus is overwhelmingly oriented toward operational effectiveness rather than normative critique. The substantial philosophical literature on the ethics of nudging (Hausman & Welch, 2010; Sunstein, 2014; Conly, 2013) is unrepresented. For practitioners working on Cause and Culture briefs in particular - where the ethics of changing audience behaviour are non-trivial - this is a significant lacuna.

These five gaps do not invalidate the corpus. They identify the perimeter of what reading it will and will not equip a practitioner to do.

6. Meta-Analytic Synthesis

A genuine meta-analysis cannot be performed on the corpus directly, because the corpus consists of trade-press syntheses rather than primary empirical reports. What follows is therefore a meta-analytic synthesis of the underlying evidence base - the primary studies most heavily cited within the corpus, set in dialogue with three recent meta-analyses of the broader nudging and behavioural-intervention literature.

6.1 Effect sizes in foundational corpus-cited studies

Table 2 summarises effect sizes from the primary studies that the corpus most prominently relies on.

The studies in Table 2 are, in their original forms, methodologically serious and have shaped applied practice for two decades. Several have been subject to robust replication (Johnson and Goldstein's default-effect findings in particular have been re-confirmed across multiple natural experiments; Hallsworth et al.'s tax-letter findings are themselves a large-scale field experiment). Others have been more contested: the broader priming literature on which several of Cialdini's (2016) examples rely has shown systematic replication problems, as documented in Section 5.

6.2 Three meta-analyses of the broader literature

The corpus rarely cites the meta-analytic literature on behavioural interventions. Three meta-analyses and one publication-bias correction are particularly relevant.

Hummel and Maedche (2019) conducted a meta-analysis of 100 primary studies of nudging interventions across domains. They reported that approximately 62 per cent of nudge interventions produced statistically significant effects in the intended direction, with median effect sizes varying substantially by intervention type.

Mertens et al. (2022), published in Proceedings of the National Academy of Sciences, performed a more ambitious meta-analysis across 212 studies, reporting a pooled effect size of Cohen's d = 0.43. This figure was widely cited as evidence that nudging "works."

Maier et al. (2022), published in the same journal as a response, applied corrections for publication bias to the Mertens et al. dataset and reported that the bias-corrected effect size was substantially smaller - approximately d = 0.08, an order of magnitude below the headline figure.

DellaVigna and Linos (2022), published in Econometrica, compared two distinct populations of nudge studies: those published in academic journals and those conducted at scale by the United States Office of Evaluation Sciences and the United Kingdom's Behavioural Insights Team. Across 126 randomised controlled trials run by these government nudge units, they reported an average effect on take-up of approximately 1.4 percentage points, compared with approximately 8.7 percentage points in published academic studies. The discrepancy was largely attributable to publication selection: academic journals published the studies with larger effects, while the nudge units' full population of trials included many null and small-effect findings that never reached journals.

6.3 Synthesis

Taken together, these meta-analyses suggest two robust conclusions.

First, behavioural interventions reliably produce effects in the intended direction, but the effects are smaller than the trade-press literature typically conveys. The strongest interventions - defaults in particular - produce large, durable effects. The modal nudging intervention, however, produces a small effect.

Second, the gap between published effect sizes and in-field effect sizes is approximately five-to-six-fold, on the DellaVigna and Linos figures. The practitioner who internalises the academic-literature effect sizes (as the corpus encourages) will systematically over-predict the impact of their interventions in real-world deployment.

This is the central meta-analytic finding of the present review. The corpus is not wrong - the mechanisms it describes are real, and many of the foundational effects it cites are robust. But it is calibrated to a higher effect-size expectation than the bias-corrected evidence supports.

7. Discussion

The four methods, taken together, produce a coherent picture of the practitioner canon.

The narrative review establishes that the corpus represents a real intellectual lineage - five overlapping phases from the heuristics-and-biases programme through institutional adoption to the field's current critical maturation. The corpus is not arbitrary; it tracks the field's actual development.

The systematic review establishes that the canon has internal differentiation. The Commerce tier draws disproportionately on cognitive mechanisms; the Cause tier draws disproportionately on social and identity-based mechanisms; the Culture tier draws disproportionately on choice architecture. This is consistent with the structural differences between the three application contexts and suggests that the canon's tier-based organisation reflects substantive rather than merely organisational logic.

The scoping review establishes that the canon is bounded. It is strongest on operational mechanisms, weakest on methodological reflexivity and ethical-philosophical engagement, and substantially blind to algorithmic and cross-cultural developments. A practitioner trained solely on this corpus will be operationally competent but methodologically incurious, and may be unprepared for the next decade of practice.

The meta-analytic synthesis establishes that the canon, although accurate at the mechanism level, is systematically miscalibrated at the effect-size level. The interventions described in the corpus produce, on average, smaller effects in the field than the publications it cites would suggest.

The implication for practitioners is not to abandon the canon but to read it as a useful but incomplete training. Three pairings are recommended. The foundational texts should be paired with the replication-crisis literature (Open Science Collaboration, 2015; Camerer et al., 2018). The applied texts should be paired with the meta-analytic literature on real-world effect sizes (DellaVigna and Linos, 2022; Maier et al., 2022). The ethical territory should be supplemented with the political-philosophical literature on nudging (Hausman and Welch, 2010; Sunstein, 2014).

A second implication concerns the practitioner's epistemic posture. The corpus is written in a confident register - this is part of what makes it pedagogically effective. But the underlying evidence base supports a less confident register than the books exhibit. The reading list's compiler is correct to instruct practitioners to "resist the cult"; the present review provides the empirical basis for that resistance.

8. Limitations

This review has four limitations that warrant explicit acknowledgement.

First, the corpus is one practitioner's selection. A different selection - for instance, one biased toward academic rather than trade-press behavioural science - would yield a different review. The findings should be read as descriptive of one influential practitioner curriculum rather than as a comprehensive map of the field.

Second, the systematic coding scheme imposes categorical structure on texts that are themselves often categorically permissive. Different coding decisions would yield different prominence ratings. Inter-rater reliability checks, standard in formal systematic reviews, were not performed for this single-author work.

Third, the meta-analytic synthesis aggregates effect sizes from studies of varying methodological quality, population characteristics, and outcome measures. The figures presented should be read as indicative of magnitude rather than as precise meta-analytic estimates.

Fourth, the review is itself dated. The behavioural-science literature is evolving rapidly, particularly in algorithmic and ethical domains, and a review conducted in five years would identify a substantially different scoping perimeter.

9. Conclusion

The practitioner canon of applied behavioural science, as represented by this corpus, is a coherent, internally differentiated, and operationally useful body of work. It is also calibrated to a more confident register than the underlying empirical literature supports, and it leaves substantial gaps in methodological reflexivity, ethical engagement, algorithmic literacy, and cross-cultural application.

A practitioner who reads the canon and stops will be well-equipped for the work of a decade ago. A practitioner who reads the canon and pairs it with the critical-methodological literature, the meta-analytic corrections, and the ethical-philosophical commentary will be well-equipped for the work of the next decade. The canon is necessary; it is not sufficient.

The deeper implication is one the reading list's compiler already articulates: the behavioural sciences do not, by themselves, differentiate one practice from another. The application differentiates. The reading sharpens the application. It does not replace it.

References

Allcott, H., & Rogers, T. (2014). The short-run and long-run effects of behavioral interventions: Experimental evidence from energy conservation. American Economic Review, 104(10), 3003-3037.

Andresen, K. (2006). Robin Hood marketing: Stealing corporate savvy to sell just causes. Jossey-Bass.

Ariely, D. (2008). Predictably irrational: The hidden forces that shape our decisions. HarperCollins.

Ariely, D., Loewenstein, G., & Prelec, D. (2003). "Coherent arbitrariness": Stable demand curves without stable preferences. Quarterly Journal of Economics, 118(1), 73-106.

Arksey, H., & O'Malley, L. (2005). Scoping studies: Towards a methodological framework. International Journal of Social Research Methodology, 8(1), 19-32.

Barden, P. (2013). Decoded: The science behind why we buy. Wiley.

Behavioural Insights Team. (2010). MINDSPACE: Influencing behaviour through public policy. Cabinet Office, United Kingdom.

Behavioural Insights Team. (2014). EAST: Four simple ways to apply behavioural insights. Cabinet Office, United Kingdom.

Camerer, C. F., Dreber, A., Holzmeister, F., Ho, T. H., Huber, J., Johannesson, M., Kirchler, M., Nave, G., Nosek, B. A., Pfeiffer, T., Altmejd, A., Buttrick, N., Chan, T., Chen, Y., Forsell, E., Gampa, A., Heikensten, E., Hummer, L., Imai, T., Isaksson, S., Manfredi, D., Rose, J., Wagenmakers, E.-J., & Wu, H. (2018). Evaluating the replicability of social science experiments in Nature and Science between 2010 and 2015. Nature Human Behaviour, 2(9), 637-644.

Chance, Z. (2022). Influence is your superpower: The science of winning hearts, sparking change, and making good things happen. Random House.

Cialdini, R. B. (2007). Influence: The psychology of persuasion (Rev. ed.). Harper Business. (Original work published 1984)

Cialdini, R. B. (2016). Pre-suasion: A revolutionary way to influence and persuade. Simon & Schuster.

Conly, S. (2013). Against autonomy: Justifying coercive paternalism. Cambridge University Press.

Cornell University. (2018). Statement of Cornell University Provost Michael I. Kotlikoff. (Statement on the Wansink investigation.)

DellaVigna, S., & Linos, E. (2022). RCTs to scale: Comprehensive evidence from two nudge units. Econometrica, 90(1), 81-116.

Doyen, S., Klein, O., Pichon, C.-L., & Cleeremans, A. (2012). Behavioral priming: It's all in the mind, but whose mind? PLOS ONE, 7(1), e29081.

Evans, D. (2018). Bottlenecks: Aligning UX design with user psychology. Apress.

Eyal, N. (2014). Hooked: How to build habit-forming products. Portfolio.

Frey, E., & Rogers, T. (2014). Persistence: How treatment effects persist after interventions stop. Policy Insights from the Behavioral and Brain Sciences, 1(1), 172-179.

Glass, G. V. (1976). Primary, secondary, and meta-analysis of research. Educational Researcher, 5(10), 3-8.

Goldstein, N. J., Cialdini, R. B., & Griskevicius, V. (2008). A room with a viewpoint: Using social norms to motivate environmental conservation in hotels. Journal of Consumer Research, 35(3), 472-482.

Gray, C. M., Kou, Y., Battles, B., Hoggatt, J., & Toombs, A. L. (2018). The dark (patterns) side of UX design. Proceedings of the 2018 CHI Conference on Human Factors in Computing Systems, 1-14.

Greenhalgh, T., Thorne, S., & Malterud, K. (2018). Time to challenge the spurious hierarchy of systematic over narrative reviews? European Journal of Clinical Investigation, 48(6), e12931.

Hagger, M. S., Chatzisarantis, N. L. D., Alberts, H., Anggono, C. O., Batailler, C., Birt, A. R., et al. (2016). A multilab preregistered replication of the ego-depletion effect. Perspectives on Psychological Science, 11(4), 546-573.

Hallsworth, M., List, J. A., Metcalfe, R. D., & Vlaev, I. (2017). The behavioralist as tax collector: Using natural field experiments to enhance tax compliance. Journal of Public Economics, 148, 14-31.

Halpern, D. (2015). Inside the nudge unit: How small changes can make a big difference. WH Allen.

Harris, C. R., Coburn, N., Rohrer, D., & Pashler, H. (2013). Two failures to replicate high-performance-goal priming effects. PLOS ONE, 8(8), e72467.

Hausman, D. M., & Welch, B. (2010). Debate: To nudge or not to nudge. Journal of Political Philosophy, 18(1), 123-136.

Henrich, J., Heine, S. J., & Norenzayan, A. (2010). The weirdest people in the world? Behavioral and Brain Sciences, 33(2-3), 61-83.

Hummel, D., & Maedche, A. (2019). How effective is nudging? A quantitative review on the effect sizes and limits of empirical nudging studies. Journal of Behavioral and Experimental Economics, 80, 47-58.

Johnson, E. J., & Goldstein, D. (2003). Do defaults save lives? Science, 302(5649), 1338-1339.

Kahneman, D. (2002). Maps of bounded rationality. Nobel Memorial Lecture.

Kahneman, D. (2011). Thinking, fast and slow. Farrar, Straus and Giroux.

Kahneman, D., & Tversky, A. (1979). Prospect theory: An analysis of decision under risk. Econometrica, 47(2), 263-291.

Levac, D., Colquhoun, H., & O'Brien, K. K. (2010). Scoping studies: Advancing the methodology. Implementation Science, 5, 69.

Maier, M., Bartoš, F., Stanley, T. D., Shanks, D. R., Harris, A. J. L., & Wagenmakers, E.-J. (2022). No evidence for nudging after adjusting for publication bias. Proceedings of the National Academy of Sciences, 119(31), e2200300119.

Mathur, A., Acar, G., Friedman, M. J., Lucherini, E., Mayer, J., Chetty, M., & Narayanan, A. (2019). Dark patterns at scale: Findings from a crawl of 11K shopping websites. Proceedings of the ACM on Human-Computer Interaction, 3(CSCW), 1-32.

Mertens, S., Herberz, M., Hahnel, U. J. J., & Brosch, T. (2022). The effectiveness of nudging: A meta-analysis of choice architecture interventions across behavioral domains. Proceedings of the National Academy of Sciences, 119(1), e2107346118.

Michie, S., Richardson, M., Johnston, M., Abraham, C., Francis, J., Hardeman, W., Eccles, M. P., Cane, J., & Wood, C. E. (2013). The Behavior Change Technique Taxonomy (v1) of 93 hierarchically clustered techniques. Annals of Behavioral Medicine, 46(1), 81-95.

Minto, B. (1987). The pyramid principle: Logic in writing and thinking. Pearson.

Ogilvy, D. (1983). Ogilvy on advertising. Crown.

Open Science Collaboration. (2015). Estimating the reproducibility of psychological science. Science, 349(6251), aac4716.

Page, M. J., McKenzie, J. E., Bossuyt, P. M., Boutron, I., Hoffmann, T. C., Mulrow, C. D., et al. (2021). The PRISMA 2020 statement: An updated guideline for reporting systematic reviews. BMJ, 372, n71.

Rohrer, D., Pashler, H., & Harris, C. R. (2015). Do subtle reminders of money change people's political views? Journal of Experimental Psychology: General, 144(4), e73-e85.

Rumelt, R. (2011). Good strategy/bad strategy: The difference and why it matters. Crown Business.

Salvi, F., Ribeiro, M. H., Gallotti, R., & West, R. (2024). On the conversational persuasiveness of large language models: A randomized controlled trial. arXiv preprint.

Schultz, P. W., Nolan, J. M., Cialdini, R. B., Goldstein, N. J., & Griskevicius, V. (2007). The constructive, destructive, and reconstructive power of social norms. Psychological Science, 18(5), 429-434.

Sharot, T. (2017). The influential mind: What the brain reveals about our power to change others. Henry Holt.

Simonsohn, U., Simmons, J. P., & Nelson, L. D. (2021). Evidence of fraud in an influential field experiment about dishonesty. Data Colada, 98.

Sunstein, C. R. (2014). Why nudge? The politics of libertarian paternalism. Yale University Press.

Sutherland, R. (2019). Alchemy: The surprising power of ideas that don't make sense. WH Allen.

Thaler, R. H., & Benartzi, S. (2004). Save More Tomorrow: Using behavioral economics to increase employee saving. Journal of Political Economy, 112(S1), S164-S187.

Thaler, R. H., & Sunstein, C. R. (2008). Nudge: Improving decisions about health, wealth, and happiness. Yale University Press.

Tversky, A., & Kahneman, D. (1991). Loss aversion in riskless choice: A reference-dependent model. Quarterly Journal of Economics, 106(4), 1039-1061.

West, R., Sheeran, P., & Michie, S. (2019). Behaviour change theories and techniques. In Behavioural insights and public policy: Lessons from around the world (pp. 84-110). OECD Publishing.

Zinsser, W. (1976). On writing well. Harper & Row.

Word count: approximately 5,200 words (main text, excluding abstract, tables, and references).