2 Cognitive Dissonance
The Effect of Choice on Attitude Change within the Induced Compliance Paradigm
written by Benjamin Buttlar (revision), Maria Solovjeva (original draft), Adira Daniel (revision), and David C. Vaidis (revision)
2.1 The Classic
Paul had a big heart and a passion for life. He called himself a “foodie” and loved planning long dinners with friends, filled with laughter and lively political debates. One of his favorite topics was animal suffering. He spoke passionately about animal welfare and felt a deep connection with animals. He strongly believed in ethical animal farming and often said that respecting animals was part of what makes us human. I’ll never forget seeing him stand at the dinner table, glass in hand, speaking out against animal cruelty—while enjoying one of his favorite meals: a medium-rare steak. That was Paul—driven, sincere, and full of beautiful contradictions.
Cognitive dissonance theory is a foundational framework in social psychology and psychological science (Devine and Brodish 2003; Haggbloom et al. 2002). It explains a powerful mental process that helps us understand how people make sense of their thoughts, beliefs, and behaviors. This theory has established dissonance processes as a fundamental psychological mechanism and a key theoretical construct. Beyond demonstrating a basic psychological effect, the theory provides a comprehensive model for understanding how individuals construct and reshape their perceptions of reality.
The model proposed by Leon Festinger (1957) states people feel uncomfortable when they have two thoughts (or cognitions) that don’t fit together. These thoughts can be about the world, themselves, or their behavior. When two related thoughts contradict each other, we experience cognitive dissonance. For example, Paul who cares deeply about animals but eats meat may feel discomfort because their beliefs and actions don’t align. The thought “I care about animals” conflicts with “I eat meat, which causes animals to suffer” (Bastian and Loughnan 2017; Gradidge et al. 2021; Loughnan, Haslam, and Bastian 2010).
#definition Cognition
Cognitions can be thought of as bits of knowledge, such as opinions, attitudes, or beliefs about the world, about oneself or one’s behavior. If two cognitions are relevant to each other, but the opposite of one follows from the other, these cognitions are said to be inconsistent.
#definition Cognitive Dissonance
Cognitive dissonance describes the discomforting state that people experience when they hold cognitions that are relevant to each other but inconsistent with each other—meaning that one bit of knowledge suggests the opposite of another bit of knowledge.
#yourturn
Can you think of other real-world examples in which someone might experience dissonance?
People generally prefer consistency between their thoughts and actions as they navigate through life. When one holds inconsistent cognitions, this can lead to cognitive dissonance—a discomforting state that people try to reduce by using various strategies. Research, however, has mostly focused on one strategy to get rid of this uncomfortable feeling: modifying inconsistent cognitions, especially through attitude shifts that align people’s attitudes with a discrepant behavior (for recent reviews, see Buttlar, Pauer, and van Harreveld 2025; McGrath, A. 2017). In Paul’s case, he might reduce dissonance by convincing himself that “ethical meat” doesn’t really harm animals (e.g., Rothgerber, H. 2015). This shift in belief helps him feel better about eating meat, even though it conflicts with his values. A change in attitudes may thereby allow people to resolve the inconsistency between conflicting cognitions and the discomfort after engaging in counterattitudinal behavior.
#yourturn
Take one of your own real-world examples: How might people modify a discrepant cognition to resolve the dissonance?
The History in a Nutshell
The state of cognitive dissonance was originally demonstrated using a forced-compliance paradigm. In a classic experiment by Leon Festinger and James Carlsmith in 1959, participants were asked to perform a boring task. After the task, some participants were asked to lie to the next participant by saying that the task was interesting, while others were asked to tell the truth about the boring task. Some were paid $1 to lie, others $20. The idea was that lying creates dissonance—especially when the reward is small. Festinger and Carlsmith (1959) expected that participants should experience cognitive dissonance when they lie, because one cognition (“I said the task was interesting”) is inconsistent with another cognition (“The task was boring”). If you’re paid only $1, you don’t have a good reason to lie, so you might change your opinion to believe the task was more interesting than it really was. That way, your actions and beliefs match. The results showed that people paid just $1 were more likely to say the task was interesting, while those paid $20 did not change their attitudes. This surprising result challenged earlier theories like reinforcement theory (Skinner 1958), which said people change their behavior when they get bigger rewards.
#definition Forced-Compliance Paradigm
A very early paradigm of cognitive dissonance initially stemming from the persuasion field where participants are asked to perform a discrepant behavior (mainly a speech or essay), being more or less incentivized for doing it. Classic studies (e.g., Festinger and Carlsmith 1959) report counter-intuitive results showing more attitude change when paid a small amount of money in comparison to a bigger amount. This contradicted predominant theoretical frameworks at the time, such as the reinforcement theory (Skinner 1958), which would have predicted that people adjust their attitudes especially when they receive a high reward.
Festinger’s (1957) original formulation of cognitive dissonance was broad in scope. Researchers rapidly narrowed their focus to attitude change after cognition inconsistency in behavioral contexts, ultimately leading to a focus on developing experimental procedures (Vaidis and Gosling 2011; for more insights on the rich and evolving history of the dissonance theory see Aronson 1992; Harmon-Jones, E. and Harmon-Jones, C. 2007; Vaidis and Bran 2020). Early in the theory development, researchers such as Jack Brehm, Arthur Cohen, and Joel Cooper emphasized choice as a necessary condition for observing dissonance-induced attitude change. From Brehm and Cohen’s (1962) foundational theoretical revision to key experiments, choice emerged as a central moderator of dissonance effects, and a necessary point to distinguishing the dissonance effect from other theories. After that, the theory was challenged by competing explanations (e.g., self-perception theory, Bem 1967). This led to multiple debates that prompted refinements, introducing additional boundary conditions for dissonance. As a result, the paradigm evolved to highlight the individual responsibility for the counterattitudinal behavior, that was summarized in choice but also took various forms with similar constructs like volition, freedom, aversive consequences, and perceived responsibility as critical variables for producing attitude change (e.g., Linder, Cooper, and Jones 1967; Cooper and Fazio 1984). According to these perspectives, Paul would experience dissonance about eating meat only if (a) he feels like he had a choice, (b) he believes that meat consumption harms animals, and (c) he feels responsible for his action (e.g., not blaming others for the harm inflicted on animals). The recognition of these additional boundary conditions culminated in the development of the induced-compliance paradigm—a more nuanced iteration of the forced-compliance procedure. These developments represented a crucial consolidation phase for the theory.
#definition Induced-Compliance Paradigm
An evolution of the forced-compliance paradigm (Linder, Cooper, and Jones 1967) where participants have to perform a behavior that is inconsistent with their attitudes—typically writing or delivering a counter-attitudinal essay. In this version, all participants engage in the discrepant task, but they differ in the justification provided. Half are explicitly told they have a free choice in performing the task (freedom condition), while the other half are simply asked to do it, as in the original forced-compliance paradigm. The key difference lies in the source of justification: an external incentive in the traditional condition, versus perceived freedom of choice in the revised one. The classic results show a greater change in the attitude in the choice condition (high choice) compared to the control condition (low choice).
Although the induced-compliance paradigm led to several refinements of Festinger’s original theory (1957), most of these refinements were reconsidered later on. For example, Harmon-Jones, E. and Brehm, J. W. and Greenberg, J. and Simon, L. and Nelson, D. E. (1996) found that people can still change their attitudes even if they don’t feel responsible or don’t see negative outcomes. However, choice retained its central role in the induced-compliance paradigm as the key moderator variable for attitude change and this paradigm remains widely used today when people engage in dissonance studies (e.g., Cooper and Feldman 2019; Forstmann and Sagioglou 2020; Randles et al. 2015).
#yourturn
Take one of your own real-world examples: Imagine how the different refinements in regard to choice, aversive consequences, and responsibility of dissonance theory would apply to your example?
2.2 The Aftermath
In recent years, psychologists have started re-examining classic theories using more rigorous methods, partly in response to the replication crisis in psychology. One method used to systematically test key findings in psychology is the Registered Replication Reports (RRR), which use pre-registered, standardized methodologies to systematically test key findings, reducing biases and ensuring that results are replicable across diverse samples. Specifically, researchers in multiple labs decide in advance how to run and analyze a study, using the same script and materials. By having multiple independent labs follow the same procedure to test whether research findings hold up, RRRs reduce systematic error and human tendencies that may distort results or conclusions of a study. For such a replication effort on cognitive dissonance, the induced-compliance paradigm offers a great foundation due to its pivotal role in the history of the theory.
The multi-laboratory RRR by Vaidis, Sleegers, and colleagues in 2024 (Vaidis et al. 2024) built upon a study conducted by Croyle and Cooper (1983, experiment 1) to replicate dissonance-induced attitude change in the induced-compliance paradigm. In the replication study, relying on 19 countries and almost 5,000 participants, students were informed about potential policy changes and then asked to write an essay contradicting their supporting opinion, with the potential to be applied under different conditions. They were then divided into three groups. The first group (high-choice inconsistency) wrote essays to support tuition fee increases while they had the opportunity to refuse to do it—they were reminded to participate voluntarily and consciously agreed to do so. The second group (low-choice inconsistency) was put in a situation where they were not explicitly free to refuse—they were simply told to write an essay to support an increase in fees. Thus, the first two comparisons served to assess the effect of choice. A third control group received all the similar information (potential change in the policies) but wrote an essay on a neutral topic with high freedom of choice (high-choice neutral). The control serves as a way to assess the effect of the inconsistent content of the essay.
According to cognitive dissonance theory, participants in the high-choice condition should have experienced more psychological discomfort and, consequently, a greater shift in their attitudes toward the position they argued for. However, Vaidis et al. (2024) found that both high- and low-choice participants in the inconsistency conditions changed their attitudes in line with their behavior (a within shift about 1 point on a 9 point scale), but the difference between these two groups was similar (all Cohen’s ds < 0.05, all ps > .15, with an overall power of 95% for the four primary analyses). These findings challenge that perceived freedom of choice is necessary for cognitive dissonance to occur—which is considered a central tenet of the theory.
Crucially, the effect of inconsistency on attitude change in the RRR was found across labs, consistently showing that writing a counterattitudinal essay produces more change than the neutral essay. This raises concern about the exact process behind these effects and opens the door for alternative explanations, like self-persuasion. Thus, the authors performed a limited number of exploratory analyses to disambiguate these results. Based on two specific items assessing the cognitive dissonance state—discomfort and conflict—that were collected during the study, they compared the reported affect of participants in the counterattitudinal essay conditions to the ones in the neutral essay condition. The results showed more discomfort and more conflict in the inconsistent conditions than in the neutral one (respectively ds > .22 and ds > 0.58, with all ps < .001). While having to keep in mind their exploratory purpose, these results light that cognitive dissonance may be a relevant process to take account of the attitude change results.
#definition Paradigm
Within Kuhn’s (1962) tradition, a paradigm transcends mere methodological procedure, constituting instead a foundational scientific framework that combines theoretical principles, methodological standards, and empirical expectations. This comprehensive system guides scientific practice by providing explanatory models and predicting experimental outcomes. The paradigm’s coherence depends on the alignment of these elements—when empirical results contradict theoretical predictions or methodological applications fail to produce expected findings, the entire paradigm faces fundamental challenges.
The results from the RRR suggest that the induced-compliance paradigm can still produce attitude change, but a central boundary condition for these effects that has been considered necessary for decades does not seem to hold up to scrutiny in a current replication. Which thus raises concerns about the induced-compliance paradigm: If freedom of choice were not a necessary precondition, then the paradigm would be based on flawed assumptions. This raises concerns that previous studies may have overestimated the role of choice, possibly due to false positives from small samples or selective reporting.
#yourturn
Why might Kuhn’s definition of paradigm lead scientists to reconsider the induced-compliance paradigm in light of these RRR findings? And if a paradigm fails, do you think that automatically means the underlying theory is also wrong? Why or why not?
Reaction to the Failed RRR of Choice as a Boundary Condition for Attitude Change
Recent comments on this replication have argued that choice may nonetheless be a central boundary condition (Harmon-Jones, E. and Harmon-Jones, C. 2024; Pauer, Linne, and Erb 2024). For instance, a re-examination of the RRR by Vaidis et al. (2024) shows that the effect of counterattitudinal behavior on attitudes was moderated by the perceived freedom of choice on self-reported manipulation checks. That is, across the inconsistent high- and low-choice conditions, especially people who perceived higher (vs. lower) choice adapted their attitudes in line with their counterattitudinal behavior (Pauer, Linne, and Erb 2024). This might suggest that the subjective perception of choice, rather than actual experimental manipulation, may be a critical driver of dissonance effects. Another explanation for this replication failure might be that the induced-compliance paradigm may not be appropriate in a modern world, in which people, for instance, have to regularly “choose” to accept cookies to avoid effort (Harmon-Jones, E. and Harmon-Jones, C. 2024; Pauer, Linne, and Erb 2024). In a response to these commentaries, Sleegers et al. (2024) emphasized that the offered explanations were only post hoc analyses or explanations, not firm expectations stemming from the theory. They reminded the field that, regardless of the theoretical discomfort, the results remain what they are—emerging from a more structured and robust study, with more participants in a single experiment than combined totals of studies over several decades—the induced-compliance paradigm as defined has not been able to generate the predicted attitude changes in the RRR.
2.3 Conclusion
What Do These Findings Mean for Future Research?
Ultimately, this replication effort does not invalidate cognitive dissonance theory but rather highlights the need for refinement in its procedures. The induced-compliance paradigm in particular has long been a cornerstone of dissonance research, yet these findings suggest it may be flawed or at least in need of revision. If the expected effect of choice on dissonance is not as reliable as previously assumed, this casts doubt on the paradigm’s ability to isolate cognitive dissonance effects from other psychological mechanisms. Many of the classic studies using this paradigm may need re-evaluation, and researchers must consider which of the refinements to Festinger’s original theory (1957) are really required and whether alternative explanations, such as the self-perception (Bem 1967) may account for the observed effects. The RRR provided some potential explanations through their exploratory data, but there is a clear need for further research. Moving forward, cognitive dissonance research must embrace more rigorous methodologies to determine when and how dissonance processes truly operate (Vaidis and Bran 2019).
Moving forward, we propose three key considerations for future research. First, as a main lesson from these replication efforts, prioritizing pre-registered, large-scale replications is essential to distinguish robust findings from false positives (Sleegers et al. 2024). Second, echoing other calls (Gradidge et al. 2021; Vaidis and Bran 2019), researchers should incorporate direct measures of the subjective experience of dissonance, such as self-reported tension, arousal, or neural activity, to investigate the process of cognitive dissonance beyond its effects on attitudes. Third, regarding the induction of cognitive dissonance specifically, future research should explore, evaluate, and refine alternative paradigms that do not depend exclusively on choice-induced dissonance. Promising examples include newer approaches like induced hypocrisy (for a meta-analysis, see Priolo et al. 2019), and the meat paradox framework (Bastian and Loughnan 2017; Gradidge et al. 2021) — the latter of which has even recently received empirical validation through registered replication (Jacobs et al. 2024). Alternatively, maintaining the plausibility of the induced-compliance paradigm will require both substantial adaptation to contemporary contexts (e.g., Schwarz and Strack 2014) and, more crucially, rigorous validation through preregistered studies to evaluate these adaptations. These steps will lead to a better understanding of cognitive dissonance and its underlying processes.