Flying Under the Radar: How Frames Influence Public Officials’ Perceptions of Corruption

ABSTRACT: The study of the forces that lead citizens and public officials to tolerate corruption has attracted scholarly attention for decades. We seek to contribute to this literature by arguing that -since corruption is an interpersonal process- public officials’ perceptions of and dispositions toward it are influenced by how it is framed. To test this claim, we conduct an original experiment on a representative sample of civil servants working in a large urban municipal government in Mexico. We find that, even when evaluating clear examples of corruption, public officials are more likely to tolerate the illegal disregard for the bureaucratic procedure when doing so is framed not as a monetary exchange but as a way in which resources can be redistributed, institutions can be made more flexible, and organizations can be made more efficacious. Our findings highlight the importance of studying corruption as a social and interpersonal communication process.

CITATION: Zizumbo Colunga, D., Meza, O. Flying Under the Radar: How Frames Influence Public Officials’ Perceptions of Corruption. Political Behavior. 45, 995–1014 (2023). https://doi.org/10.1007/s11109-021-09745-3

Note: What follows is the authors' pre-print version. The final published version can be accessed [HERE]

Public officials play a central role in the way government resources are assigned, distributed, and administered. Consequently, they are at the core of the emergence, diffusion, and perpetuation of corruption. While research on the factors that determine voters’ perceptions and tolerance of corruption has advanced rapidly over the last twenty years (Agerberg, 2019; Batista Pereira, 2020; Carlin et al., 2015; Redlawsk & McCann, 2005; Winters & Weitz-Shapiro, 2013; Zechmeister & Zizumbo-Colunga, 2013) inquiry into the forces that drive individual public officials to see corruption as acceptable or justifiable behavior—to tolerate corruption—has advanced at a more gradual pace (Guo & Tu, 2017; Jackson & Smith, 1996; Peters & Welch, 1978). 

We seek to contribute to this literature by conceptualizing public officials’ perceptions of corruption as the result of an interpersonal process in which framing plays a central role. Specifically, we argue that public officials are more likely to tolerate the illegal subversion of bureaucratic procedure for private benefit—what we henceforth call bureaucratic misbehavior—when it is framed as a way to (a) reach Pareto-efficient outcomes, (b) redistribute centralized resources, (c) increase institutional flexibility, and (d) improve organizational effectiveness. Furthermore, we argue that bureaucratic misbehavior tends to be seen more favorably when quid pro quo offers are kept out of sight. 

To examine the impact of framing on civil servants’ attitudes, we analyze data from an original experiment conducted on a random sample of public officials working in a municipal government setting in Mexico. Following a procedure similar to the one used by other scholars (Guo & Tu, 2017; Jackson & Smith, 1996; Peters & Welch, 1978), we ask participants to read a series of scenarios involving bureaucratic misbehavior, and then, we question them about their perceptions of these behaviors. Yet, unlike previous studies, rather than comparing civil servants’ perceptions across different scenarios, we vary randomly the way bureaucratic misbehavior is framed within each scenario. This gives us the opportunity to identify the effect of frames while accounting for individual and scenario-level differences. We find that, even when evaluating clear examples of corruption, public officials’ perceptions are influenced by framing. We also find some evidence that frames promoting organizational benefits tend to have a more consistent effect on civil servants’ attitudes than pro-social frames and that, to some extent, making corrupt offers explicit mutes the effect of framing. Our findings highlight the importance of studying corruption as a social and interpersonal communication process.

In the following two sections, we contextualize our study in light of previous research on bureaucratic corruption and describe our argument. In the third section, we introduce the methodological design used to test our hypothesis. In the final two sections, we discuss our results and their theoretical and empirical implications.

Bureaucratic Corruption

Corruption drains trillions of dollars from the global economy every year (Dreher & Herzfeld, 2005; Gupta et al., 2016; Kaufmann, 2005; Mo, 2001) and severely distorts the implementation of the best-designed policies (Dal Bó & Rossi, 2007; Fredriksson & Svensson, 2003; Rose-Ackerman, 1975). While grand corruption tends to capture most of the public’s attention, most acts of corruption involve local and low-level officials subverting bureaucratic procedures illegally (Lambsdorff, 2006; Masters & Graycar, 2016). Thus, to fully understand corruption, it is critical to understand the factors that shape civil servants’ attitudes and willingness to misbehave.

The literature on bureaucratic corruption is wide and broad. Scholars who conceptualize corruption as a fruit of socialization have found important differences across countries and individuals. Gender (Dollar et al., 2001; Swamy et al., 2001), culture (Fisman & Miguel, 2007; Lee & Guven, 2013; Treisman, 2000), and education (Asongu & Nwachukwu, 2015; Glaeser & Saks, 2006; Truex, 2011) have all been found to play a significant role in civil servants’ willingness to misbehave. The literature that conceptualizes corruption as a rational choice (Becker & Stigler, 1974; Nas et al., 1986; Rose-Ackerman, 1999) notes that even the most virtuous individuals can misbehave given the right incentives. Thus, institutions (Lederman et al., 2005; Rose-Ackerman, 1999), public-sector wages (Di Tella & Schargrodsky, 2003; Van Rijckeghem & Weder, 2001), and auditing systems (Carranza, 2008; Rojas, 2020; Rothstein, 2011) can influence the expected utility of corruption and, in doing so, affect the prevalence of bureaucratic misbehavior.  

These theoretical outlooks on bureaucratic corruption disregard three key factors about this behavior. First, systemic low-level bureaucratic misbehavior—the type that ordinary people most likely confront—is often the result of public officials allowing people to gain access to services from which they are otherwise excluded (Marquette & Peiffer, 2018), or of citizens attempting to game the system to make their way through patchy, unfair, or deficient public services (Peeters et al., 2020). Second, previous views of corruption often disregard that, rather than acting in isolation, civil servants frequently reproduce behavioral patterns that are in line with societal and interpersonal norms (Jancsics, 2019); therefore, their decisions are embedded in a web of interpersonal relations and trust from which they cannot escape (Perez-Chiqués & Meza, forthcoming). Thus, civil servants often act in the service of the group, or kin, rather than exclusively to the service of their personal interests (Bukuluki, 2013). Finally, as some studies have noted, when interviewed after the fact, many corrupt public officials justify their acts as legitimate, or deny having knowledge that they were engaging in corruption (Anand et al., 2004; Ashforth & Anand, 2003). Thus, at least in some cases, the key informational presuppositions from which expected utility models are developed may not correspond to the starting point from which bureaucratic corruption is conducted. 

Aware of these limitations, an emerging strand of scholarship has started to view corruption as a social process that is justified, normalized, and socialized within bureaucratic organizations and that, at a certain point, can “fly under the radar” of bureaucrats (Anand et al., 2004; Arellano Gault, 2017; Ashforth & Anand, 2003; Pinto et al., 2008). While the literature interested in the normalization of corruption has acknowledged that public officials’ understanding of bureaucratic misbehavior “may be abetted by the malleability of symbolism in general and of language in particular” (Ashforth & Anand, 2003, p. 22), it has not yet empirically identified a specific mechanism capable of guiding the justification, normalization, and socialization of corruption. We seek to fill this gap in the literature by theorizing that bureaucratic misbehavior is the result of a relational and communicative process, one in which solicitors and organizations—those requesting that public officials misbehave—attempt to obfuscate their demands through the use of justificatory frames. In the short run, we argue, framing influences bureaucrats’ disposition to perceive, accept, and denounce misbehavior. Long term, however, it plays a central role in the process of corruption consolidation (Meza & Pérez‐Chiqués, 2020). 

Framing and Bureaucratic Corruption 

In the context of communication, framing refers to a strategy in which individuals or organizations—message sources—emphasize a specific angle of an issue to make their messages, arguments, or calls to action more persuasive to receptors (Chong & Druckman, 2007; Goffman, 1974; Iyengar, 1996). Research on framing has typically focused on the way politicians, media outlets, and opinion leaders use language to rally public support around themselves and their preferred policies (Ceresola, 2018; Chen & Zhang, 2016; K. M. McGraw, 1990; Zmolnig, 2018). However, framing theory is a potent paradigm that transcends top-down mass communication. Frames, for instance, are frequently used in salary negotiations, romantic exchanges, commercial deals, and financial transactions (Huang et al., 2021; Majer et al., 2020; A. P. McGraw & Tetlock, 2005). Furthermore, framing becomes particularly important when individuals attempt to communicate embarrassing, risky, or ethically questionable messages (Chen & Zhang, 2016; McLaren, 2015; Pinker, 2008). Frames protect remitments’ self-image, anchor negotiations, smoothen relational exchanges, and, when used effectively, can persuade others to agree to engage in risky behavior. Hung, Sena, Li, and Ozdemir (2021) found that, by framing their requests, participants in risky peer-to-peer lending platforms were able to persuade lenders to invest in them despite their bad credit histories. Yang, Vosgrau, and Loeswenstein (2013) found that investors are more willing to pay for risky assets when sellers frame them more formally. Schindler and Pfattheicher (2017) found that individuals are more likely to cheat at economic games when doing so is framed as a way to avoid losses. Because corrupt requests are simultaneously risky, illegal, and ethically questionable, we argue that framing plays a central role in the way they are communicated (Chen & Zhang, 2016; Zmolnig, 2018).

While the literature touching on the interlink between frames and corruption has focused on the way frames are used retrospectively (Anand et al., 2004; Arellano Gault, 2017; Ashforth & Anand, 2003; Kish-Gephart et al., 2010; Pinto et al., 2008; Zyglidopoulos et al., 2009), the literature on corruption normalization and consolidation provides the analytical groundwork to propose that frames are also used prospectively to persuade public officials to break the bureaucratic procedure to commit acts of corruption.  Such opportunities abound in organizational contexts where framing allows corrupt external actors to introduce acts of corruption in non-corrupt organizations, where example and peer pressure serve to consolidate and exploit bureaucratic misbehavior (Ashforth & Anand, 2003; Peeters et al., 2020; Pinto et al., 2008), and in contexts in which a culture of corruption is in the process of being socialized (Ashforth & Anand 2003). Based on this literature, we identify two classes of frames (pro-social and pro-institutional) that allow individuals to make their misbehavior approaches more acceptable. 

The first class of frames we investigate—which we call pro-social frames—relate to public officials’ tendency to justify acts of corruption as being so harmless compared to their social benefits that it is not reasonable to say that they were an act of corruption. Insofar as public officials attempt to maximize social utility, they may be persuaded by solicitors using these frames to perceive otherwise unacceptable actions as tolerable strategies to produce social welfare (Arellano Gault, 2017; Barr & Serra, 2009; Pinto et al., 2008; Schindler & Pfattheicher, 2017). Moreover, because public officials, like most individuals, use evaluations of harm as a signal to identify unethical behavior (Feinberg, 1987; Gray et al., 2014), they may disregard or comply with ethically questionable acts so long as that they do not harm others directly. There is evidence that corrupt politicians and bureaucrats use harmlessness or Pareto efficiency frames to rationalize their acts and seek redemption (Anand et al., 2004; Rabl & Kühlmann, 2009), and that citizens use this kind of justificatory frames to excuse their use of social influences to obtain undue public benefits (Arellano Gault 2018). Yet, so far, there is little evidence on the extent to which justifications appealing to the Pareto Efficiency of corruption persuade public officials to misbehave when used as a way to frame requests. 

The second type of pro-social framing that can be conducive to the justification of bureaucratic misbehavior emphasize the redistributive nature of corruption. Given the existence of powerful actors in and outside public administration, solicitors can frame corruption as a way in which civil servants contravene unfair laws to “reduce great prejudices toward weak persons” (Arellano Gault, 2017, p. 834). Fairness is a core dimension by which individuals form their evaluations of moral behavior (Kahneman et al., 1986), and, for centuries, appeals to redistribution have allowed bandits and criminals to transcend their status and become outlaw heroes (Hobsbawm, 2000; Seal, 2009). Thus, by highlighting social redistribution as a byproduct of their requests for bureaucratic misbehavior, solicitors may create a state of cognitive dissonance that can be resolved by normalizing and justifying corruption. In line with this logic, ethnographic and qualitative research has found corrupt politicians and bureaucrats to use a redistributive logic to justify and rationalize their misdeeds. Olivier de Sardan and others, for instance, have found that corrupt politicians in Africa often use redistributive and collectivist arguments to justify their actions (Blundo & Olivier de Sardan, 2006; Bukuluki, 2013; Olivier de Sardan, 1999), and both Li (2018) and Arellano Gault et al. (2019) have found that citizens pulling strings within public administration (i.e., exerting gianxi, palancas, or jeitnho) frequently use redistributive frames to justify their attempts to gain undue social influence in government. If rationalizations are indicators of the frames public officials consider legitimate, redistributive frames may play an important role in the process by which bureaucratic corruption becomes justified. 

As an alternative to appealing to pro-social frames, solicitors may use institutional justifications as a strategy to obfuscate their requests. When civil servants enter public service, they are socialized into the norms, values, and traditions of government agencies. Throughout this process, individuals maintain their individuality but also become accustomed to using an institutional logic to understand the events that occur within the organization (Arellano Gault, 2017; Ashforth & Anand, 2003; DeHart-Davis, 2007; Pinto et al., 2008). Thus, rather than being driven by considerations of justice and social welfare, individuals become motivated by outcomes that allow the organization—and them as agents—to achieve their goals more effectively. 

The institutional justifications for corruption can be clustered into two groups. The first relates to arguments highlighting the capacity of corruption to allow organizations to bypass inflexible institutions. The actions of public agencies are centered on objectives and goals that can be obstructed by procedures and red tape. Thus, civil servants can come to see bureaucratic procedures as an opponent rather than an ally. To the extent that individuals seeking to justify corruption appeal to this logic, they can frame acts of corruption as the only viable path for organizations to achieve their goals. Consistent with this view, qualitative evidence has found flexibility to be a central logic behind the rationalization of misbehavior (Arellano Gault, 2017; Belle & Cantarelli, 2017; DeHart-Davis, 2007), and quantitative research has consistently found a strong association between the prevalence of red tape and corruption (Duvanova, 2014; Guriev, 2004). If these findings are indicative of the underlying logic that pushes civil servants to misbehave, we should observe that frames appealing to flexibility shape the extent to which public officials see corrupt acts as justifiable. 

The second group of institutional justifications that can be used to frame misbehavior conceptualize corruption as a way to promote organizational efficiency, that is, as a shortcut that allows bureaucrats and citizens to accomplish the same goals that they could achieve through long and involved bureaucratic procedures. When corruption becomes associated with organizational efficiency, citizens following the norms are seen as naïve or foolish, and public officials following protocol are regarded as nuisances. Thus, by appealing to public officials’ desire to be a facilitator of group objectives (Ashforth & Anand, 2003), solicitors requesting that public officials misbehave may make their corrupt approaches less evident by using efficiency frames. Although efficiency and flexibility frames may seem similar, they are distinct in that the former frame corruption as a way to bypass inflexible institutional rules, and the latter justify misbehavior as a way to facilitate the achievement of organizational objectives. 

In the next section, we test the capacity of frames to influence the way public officials see corruption. Before moving forward, however, it is important to note that the use of frames to persuade public officials to misbehave contrasts with a direct strategy that seeks to persuade public officials through the use of quid pro quo offers. Given the contraposed nature of both strategies, it is critical to evaluate the degree to which presenting offers can potentially spoil the capacity of frames to exert their persuasive effect.

Methods

To examine the capacity of frames to influence public officials’ views of bureaucratic misbehavior, we conducted an experiment with bureaucrats working in a large provincial urban municipality in Mexico. Although Mexico is one of the countries where bureaucratic corruption is most prevalent (Zizumbo-Colunga & Amador, 2018), the state in which our municipality is located has a below-average level of corruption within the country (INEGI, 2020). Thus, the context under study is similar to other large provincial-urban governments in Latin America.

To represent accurately the public officials working in the municipal government, we conducted the sampling from the payroll registered by the municipality’s Human Resource Department (HR). We stratified public officials by heads of department, mid-level, and operatives. While all heads of department were invited to participate in the study, mid-level officials and operatives were invited to participate with a probability proportional to the size of the department in which they worked. This strategy allows us to recover representative population estimates with a confidence level of 95 percent and an error margin of m=0.05. There were so few area directors that, instead of drawing a random sample, we interviewed all of them.

Two days before our team visited the municipality, we sent a letter endorsed by the government’s transparency unit to all the heads of department, informing them that enumerators would visit their unit to conduct a short survey. On the day of the interview, enumerators visited each unit, reserved a room in which participants could answer the survey quietly and without being interrupted, and invited the selected civil servants to participate in the study in groups of 8. Substitutions were allowed only if the initially sampled respondent was sick or absent during the days of fieldwork, which occurred in less than 10% of the cases. When such was the case, public officials were substituted by a colleague working in their same department, at their same hierarchical level, and, when possible, of their same gender.

Enumerators assigned seats to participants, being careful to separate them sufficiently so that they could not talk to each other or see each other’s answers. Then, they informed subjects that the survey was strictly voluntary and anonymous, that they could leave any question unanswered, and that nobody would be informed if they decided not to participate. Participants who did not give their informed consent (<1%) were thanked for their time and let out of the laboratory without our informing their superiors of their refusal. Those who consented continued to the first section of the study. Figure 1 shows the flow of the survey. 

In line with previous studies (e.g., Guo & Tu, 2017; Jackson & Smith, 1996; Peters & Welch, 1978), after being asked a series of demographic questions, civil servants were asked to evaluate eight different vignettes describing acts of bureaucratic misbehavior in four areas of public administration, front desk, archiving, budgeting, and purchases.  We selected these four areas because they are the focus of Mexico’s Federal Law of Transparency and Access to Public Information, the main legal instrument in the Mexican government’s fight against corruption (LGTAIP, 2015). We included two vignettes per area, one describing bureaucratic misbehavior in abstract terms, and one describing this type of action with an example (concrete). To reduce anchoring and memory effects, abstract and concrete vignettes about the same area were presented in separate sections of the questionnaire (see Figure 1).

Using such a diverse set of vignettes allows us to evaluate the effects of frames across a broad set of contexts and—within the limitations inherent to multiple-measures designs—to increase our effective number of observations.  We designed all vignettes to portray clear examples of bureaucratic misbehavior. All the examples described actions sanctioned legally by the state’s penal code. According to the state’s penal code, public officials engaging in the acts described in the vignettes risk being subject to administrative sanctions, bans from public service, and jail time (Guanajuato’s Penal Code, 2001). We chose to work with clearly illegal acts because, although doing so is likely to trigger extreme responses and high levels of social desirability, it removes the ambiguity between acts of bureaucratic misbehavior and innocent acts of bureaucratic discretion—a necessary precondition to study public officials’ perceptions of corruption.

Table 1 shows two of the eight vignettes to which each respondent reacted. The text in brackets corresponds to the sections manipulated experimentally. The wording of each vignette can be found in online Appendix A. 

After reading each vignette, civil servants were asked to answer four questions on a scale from 1 to 7. First, they were asked to what extent they thought that, given the state of things, acquiescing to the request described in the vignette was justified (Justified). Then, they were asked about the extent to which they thought that—if found in a similar situation—they would personally agree to the request (Misbehave).  Afterward, participants were asked about the extent to which they thought the acts described could be construed as corruption (Corrupt), and finally, they were asked about the degree to which they considered that—if they witnessed a scenario like the one described in the vignette—they would report it to the authorities (Report).  We asked the questions in this order to reduce as much as possible the social desirability bias associated with asking individuals about their willingness to report corrupt acts just before asking about their attitudes toward them.

To test the capacity of framing to shape public officials’ attitudes, we randomly varied the justificatory frames organizations and solicitors use within each vignette. Doing so allows us to account for all confounding factors across individuals and scenarios and to control for the endogenous relation between public officials’ attitudes toward corruption and their willingness to accept frames (Anand et al., 2004; Gannett & Rector, 2015). It is important to acknowledge, however, that our design is not equipped to distinguish framing from informational effects. A more detailed design would be necessary to evaluate whether the effect of frames is driven by a shift in the way public officials view corruption or by the introduction of additional considerations. The results of our experiment reflect the net effect of the justificatory frames including both mechanisms. 

Each vignette was assigned to include no frame at all (Control), or a phrase framing the act as (a) an attempt to improve policy outcomes without harming anyone (Pareto Efficiency), (b) an attempt to redistribute resources (Redistribution), (c) an attempt to surpass institutional rigidity (Flexibility), or (d) as an attempt to improve institutional efficiency (Efficiency). Additionally, to test the extent to which explicit quid pro quo offers clarify bureaucratic misbehavior as corruption (Estrada-Gallego, 2009; Peters & Welch, 1978)—and potentially spoil the capacity of frames to influence public officials—each vignette was randomized into a secondary treatment: the mention of (a) no quid pro quo offer (Control),(b) an economic offer (Money), or (c) an offer to reciprocate with favors (Favor). In sum, each vignette could include (or not include) one of the frames and one of the pieces of quid pro quo offers mentioned before. The text in curly brackets in Table 1 shows an example of the Money treatment. A full description of each vignette can be found in online Appendix A.  

Results

In total, 630 of the 750 individuals invited agreed to participate in our survey (85%), of which 61.77% were male and 38.23% were female. On average, participants were 39.3 years of age, had about 14.35 years of schooling, and earned about 864.5 USD a month (17,290 pesos). While, to our knowledge, there is no comparable benchmark against which to compare our sample, in online Appendix C, we show that the demographic composition of our sample is similar to the demographics of the overall public official population in Mexico as proxied by INEGI.  

With respect to participants’ reactions to our scenarios, we found public officials to display low levels of tolerance for corruption. Even when we did not mention monetary or favor offers, scenarios were seen—on a 0 to 100 scale—as highly corrupt (90.8), unjustifiable (7.9), and worthy of being reported (87.1). These reactions are consistent with the relatively low levels of corruption reported traditionally in the municipality under study,  with previous studies finding that public officials are often clear about the acts that constitute corruption (Atkinson & Mancuso, 1985), and with the success of our scenarios in conveying unambiguous examples of bureaucratic misbehavior.  Meanwhile, it is important to acknowledge that, despite our efforts to provide civil servants with a private space to answer our questions honestly, these low levels of tolerance for corruption are also consistent with participants emitting socially desirable answers. Yet, as we discuss later, to the extent that social desirability biases our results, it does so in a conservative direction. 

Within the attitudinal bounds imposed by clear acts of corruption, to what extent do frames shape civil servants’ attitudes toward bureaucratic misbehavior? To investigate this question, we pooled participants’ responses across scenarios and fitted four distinct OLS models,  one for each dependent variable. Equation 1 describes the models estimated.

In the equation, the rating given by each individual (ϕ) is a function of the way in which the solicitor frames the request (as seeking Pareto-Efficiency, Redistribution, Flexibility, Efficiency, None), the type of offer suggested in the scenario (Monetary, Favors, None), the uniqueness of each scenario (θ), the uniqueness of each individual (δ), and some residual random error (e).  

Table 2 displays the effect of each condition within each of the two manipulations (Framing and Offer) on the four different outcomes measured. Contrary to previous studies finding individuals’ perceptions of explicit corruption to be stable across subpopulations and contexts, we find the ratings expressed by the subjects in our study to be systematically influenced by the frames used to describe bureaucratic misbehavior. Yet, we find only moderate evidence that frames influence civil servants’ behavioral dispositions directly. 

In general, as Table 2 shows, we find that frames that highlight the capacity of bureaucratic misbehavior to redistribute resources and increase institutional flexibility, and highlight the capacity of misbehavior to strengthen bureaucratic efficiency attenuate civil servants’ negative views of corruption (Columns 1 and 3). However, we find no evidence that Pareto efficiency frames significantly influenced public officials’ attitudes. 

The fact that—within pro-social frames—the effect of redistributive frames dominates over the effect of Pareto efficiency frames highlights a disconnect between normative legal theory and bureaucrats’ perceptions. While normative legal theory tends to place a strong link between perceptions of crime and wrongdoing (Feinberg, 1987; Gray et al., 2014), we find that public officials tend to put more weight on justice and fairness than on harm.

As Column 2 in Table 3 shows, we also find that institutional frames have a stronger effect than pro-social frames on public officials’ willingness to misbehave. This result is consistent with the idea that public officials absorb a bureaucratic mentality that leads them to display a higher sensibility to improve or navigate institutional pathologies than to increase social goods (DeHart-Davis, 2007). If such is the case, solicitors cognizant of the bureaucratic barriers to which civil servants are exposed may be more effective in persuading public officials to misbehave than citizens appealing to social fairness.  

In addition to these two patterns of results, it is important to note that neither pro-social nor institutional frames seem to influence citizens’ disposition to report corruption. This result suggests that, perhaps due to a generalized sense of duty promoted by transparency campaigns, public officials tend to hold more stable attitudes toward reporting than toward misbehaving. It is important, however, to interpret these null findings with caution since the high levels of social desirability associated with answering questions about corruption could bias our results in favor of the null hypothesis (see online Appendix F). Thus, our results should be understood as a conservative surrogate of potentially larger effects and not as a strong demonstration of the inflexibility of public officials’ attitudes.

Finally, we find quid pro quo offers to play a central role in civil servants’ perception of what constitutes an act of corruption. When an act of bureaucratic misbehavior is framed as motivated by favor or monetary offers, it is more clearly seen as corrupt and unjustified. Across the board, offers seem to have a stronger and more consistent effect on public officials’ attitudes than frames. This pattern of results may emerge from corruption campaigns highlighting bribery and extorsion as paradigmatic exemplars of corruption, or from public officials’ disposition to prioritize reporting crimes with a stronger sanction. While all the acts in our task are illegal, Mexican law stipulates that when corruption involves a transaction of money or favors, it merits a more severe penalty. The figures in Table 3 illustrate the effects of framing (upper row) and quid pro quo offers (lower row) on public officials’ views of corrupt approaches. The dashed horizontal lines represent the control condition. 

In general, participants showed a limited willingness to embrace justificatory frames. However, like the panels in the top row show, within the bounds of the study, redistributive and institutional frames shaped civil servants’ attitudes. In total, framing increases corruption tolerance by between 2 and 3 percentage points, which, given the low levels of tolerance displayed in our study, translates into between 40 and 70% in respondents’ acceptance of corrupt acts. With respect to the use of quid pro quo offers, our results are consistent with a tradeoff hypothesis. While solicitors can strive to persuade public officials by offering enticements, they do so at the risk of making the unethical nature of their request more salient. Despite involving an explicit transgression of the law, a request to misbehave is perceived to be between 3 and 4.6 percentage points less corrupt and more justifiable and between 2 and 3 percentage points less reportable when it does not include an offer. But is this effect independent, or do offers spoil the effect of framing? To explore this question, we specify a model in which we interact both treatments (Table 5).

As Table 4 shows, we found some evidence consistent with the spoiling effect of quid pro quo offers. In general, when solicitors mention an exchange of favors, individuals become alert, and the effect of framing disappears. We found only one instance in which the effect of framing is potentiated by offers. When monetary offers are involved, efficiency frames seem critical for solicitors’ capacity to avoid detection. When they are not, frames seem redundant.   This result suggests that different offers trigger distinct considerations in the minds of individuals. Yet, more research is necessary to understand the varied moderating effects that different types of offers can generate.

Conclusion

Civil servants play a critical role in the emergence and maintenance of, and in the fight against, grand and petty corruption. Unsurprisingly, a small but growing body of literature has examined the forces that lead civil servants to tolerate and justify corruption (Guo & Tu, 2017; Jackson & Smith, 1996; Mancuso, 1993; Peters & Welch, 1978). In this paper, we have sought to contribute to this literature by conceptualizing corruption as an interpersonal phenomenon in which the framing of requests, offers, and exchanges matters. We have argued that frames that emphasize the capacity of bureaucrats to improve social and institutional outcomes through the illegal disregard of the bureaucratic procedure influence the way in which civil servants see corruption. 

To test these ideas, we conducted an original experiment on a representative sample of civil servants working in a large Mexican municipality like many in Latin America. We found civil servants to perceive bureaucratic misbehavior as less corrupt when it was framed to highlight its capacity to redistribute resources, improve institutional efficiency, and increase institutional flexibility. Further, consistent with previous scholarship, we found that pro-institutional frames have a more consistent effect than pro-social frames. Finally, we found that explicit quid pro quo offers contravene a framing strategy. Not only do they clarify bureaucratic misbehavior as corruption, but, in most cases, they spoil the effect of framing. 

Three specific results are noteworthy. First, even though previous studies have found that corrupt public officials tend to rationalize and frame their misbehavior in light of its harmlessness (Gannett & Rector, 2015; Rabl & Kühlmann, 2009), we found no evidence that civil servants themselves are persuaded by these types of frames. This result suggests that corrupt public officials may be unaware of the ineffectiveness of their strategy, that citizens and officials are affected differently by framing, that public officials anticipate that these frames will be used to hide misbehavior, or that civil servants in developed and developing democracies form their ethical evaluations differently. Future studies should compare the effect of frames on citizens and public officials, to explore these possibilities more deeply. 

Second, institutional frames seem to be at least as effective as redistributive frames in obfuscating the link between bureaucratic misbehavior and corruption. This result suggests that, in addition to assessing the ethical nature of their actions, civil servants evaluate the degree to which their behavior is harmful, fair, and just (Dungan et al., 2014; Kahneman et al., 1986) and consider the degree to which it improves the functioning of government. Interventions seeking to instill organizational loyalty among public officials should not lose sight of the capacity of corrupt solicitors to use this value to frame bureaucratic misbehavior. 

The third result to highlight is that, as previous correlational studies have noted (Guo & Tu, 2017; Jackson & Smith, 1996; Peters & Welch, 1978), portraying bureaucratic misbehavior as a favor or economic exchange clarifies these actions as corrupt, unjustifiable, denounceable, and un-endorsable. This finding is important for two reasons. On the one hand, not all acts of corruption involve a direct and explicit monetary exchange. Many involve breaking the law to benefit others without the explicit intention of having the favor returned. To the extent that these types of actions are more likely to pass unnoticed by civil servants, they may become more likely to be normalized and socialized. Interventions to reduce corruption in developing countries would do well to train civil servants to recognize corruption even in the absence of monetary and favor quid pro quo offers. On the other hand, this result is noteworthy because it suggests the existence of a tradeoff between entering into a process of economic bargaining (Atanasov, 2011; Estrada-Gallego, 2009; Shleifer & Vishny, 1993) and a persuasive communication process. It seems that solicitors and corrupt organizations (Arellano Gault, 2017; Pinto et al., 2008) may need to abandon this path as soon as they seek to enter an economic bargaining process. 

While the experimental framework used here allows us to account for important internal validity threats, it is critical to acknowledge that it also has limitations. On the one hand, our study was focused on four specific examples of social and institutional frames. Yet, many other frames (e.g., impunity, social acceptability, confidentiality) may also influence the attitudes of public officials. We hope to inspire others to theorize how other frames may also allow corruption to disseminate and consolidate in bureaucratic organizations. On the other hand, by including justification frames, we have simultaneously introduced additional information that could be novel to some of the study participants. Although we think that it is unlikely that our participants have never heard of the framings introduced here, our empirical strategy cannot disentangle information and framing effects. Future studies should develop a more nuanced empirical strategy to fulfill this objective. 

Despite these limitations, we believe that our results highlight the importance of framing as a fundamental part of the communicative process by which corruption disseminates. It highlights key implications since anti-corruption strategies have long been aimed at fighting impunity, promoting self-restraint, and incentivizing whistleblowing. These strategies assume an organizational environment in which cooperation is clear and well defined. In several contexts, however, recognizing corruption is difficult since civil servants perceive the violation of the bureaucratic procedure for private gain as a normal way of attaining institutional objectives or as a way to sustain the public interest. This study demonstrates that communication frames constitute an important mechanism through which corruption becomes socialized and therefore consolidated. Consequentially, further interventions are needed to train public officials to identify and deal with corruption frames. Some interventions may involve training civil servants to see through frames and justifications. Others may involve socializing the importance of procedure beyond a legalistic duty, reducing red tape and institutional inflexibility, and providing citizens and civil servants with mechanisms to request or offer help within the legal system.

REFERENCES

Agerberg, M. (2019). The Curse of Knowledge? Education, Corruption, and Politics. Political Behavior, 41(2), 369–399. 

Anand, V., Ashforth, B. E., & Joshi, M. (2004). Business as Usual: The Acceptance and Perpetuation of Corruption in Organizations. Academy of Management Perspectives, 18(2), 39–53. 

Arellano Gault, D. (2017). Corruption as an Organizational Process: Understanding the Logic of The Denormalization of Corruption. Contaduría y Administración, 62(3), 827–842. 

Ashforth, B. E., & Anand, V. (2003). The Normalization of Corruption in Organizations. Research in Organizational Behavior, 25, 1–52.

Asongu, S., & Nwachukwu, J. (2015). The Incremental Effect of Education on Corruption: Evidence of Synergy from Lifelong Learning. Economics Bulletin, 35(4), 2288–2308.

Atanasov, P. D. (2011). Bribery as Negotiation: A Decision-Making Perspective (SSRN Scholarly Paper No. 1603362). Social Science Research Network. https://papers.ssrn.com/abstract=1603362

Atkinson, M. M., & Mancuso, M. (1985). Do We Need a Code of Conduct for Politicians? The Search for an Elite Political Culture of Corruption in Canada. Canadian Journal of Political Science / Revue Canadienne de Science Politique, 18(3), 459–480.

Barr, A., & Serra, D. (2009). The Effects of Externalities and Framing on Bribery in a Petty Corruption Experiment. Experimental Economics, 12(4), 488–503.

Batista Pereira, F. (2020). Do Female Politicians Face Stronger Backlash for Corruption Allegations? Evidence from Survey-Experiments in Brazil and Mexico. Political Behavior. 

Becker, G. S., & Stigler, G. J. (1974). Law Enforcement, Malfeasance, and Compensation of Enforcers. The Journal of Legal Studies, 3(1), 1–18. 

Belle, N., & Cantarelli, P. (2017). What Causes Unethical Behavior? A Meta-Analysis to Set an Agenda for Public Administration Research. Public Administration Review, 77(3), 327–339.

Berinsky, A. J., & Kinder, D. R. (2006). Making Sense of Issues Through Media Frames: Understanding the Kosovo Crisis. The Journal of Politics, 68(3), 640–656. 

Blundo, G., & Olivier de Sardan, J.-P. (2006). Everyday Corruption and the State: Citizens and Public Officials in Africa. Zed Books.

Bukuluki, P. (2013). “When I Steal, it is for the Benefit of Me and You”: Is Collectivism Engendering Corruption in Uganda? International Letters of Social and Humanistic Sciences, 5, 27–44.

Ley general de transparencia y acceso a la información pública, Pub. L. No. DOF 20-05-2021, 1 (2015). http://www.diputados.gob.mx/LeyesBiblio/pdf/LGTAIP_200521.pdf

Carlin, R. E., Love, G. J., & Martínez-Gallardo, C. (2015). Cushioning the Fall: Scandals, Economic Conditions, and Executive Approval. Political Behavior, 37(1), 109–130. 

Carranza, R. (2008). Plunder and Pain: Should Transitional Justice Engage with Corruption and Economic Crimes? International Journal of Transitional Justice, 2(3), 310–330. 

Ceresola, R. G. (2018). The U.S. Government’s Framing of Corruption: A Content Analysis of Public Integrity Section Reports, 1978–2013. Crime, Law and Social Change, 71(1), 47–65.

Chen, M., & Zhang, C. (2016). Framing Corruption in the Chinese Government: A Comparison of Frames Between Media, Government, and Netizens. International Journal of Communication, 10(0), 20.

Chong, D., & Druckman, J. N. (2007). Framing Theory. Annual Review of Political Science, 10(1), 103–126.

Dal Bó, E., & Rossi, M. A. (2007). Corruption and Inefficiency: Theory and Evidence from Electric Utilities. Journal of Public Economics, 91(5), 939–962. 

David, A. G., Jesús, H., Daniel, Á., & Emilio, Z. (2019). Estudio semiológico de la palanca en México: Un primer acercamiento empírico. In A. G. David & C. S. Manlio F. (Eds.), La resbaladilla de la corrupción: Estudios sobre los procesos sociales y organizacionales de la corrupción colusiva en el sector público (p. 264). CIDE.

DeHart-Davis, L. (2007). The Unbureaucratic Personality. Public Administration Review, 67(5), 892–903.

Di Tella, R., & Schargrodsky, E. (2003). The Role of Wages and Auditing during a Crackdown on Corruption in the City of Buenos Aires. The Journal of Law and Economics, 46(1), 269–292.

Dollar, D., Fisman, R., & Gatti, R. (2001). Are Women Really the “Fairer” Sex? Corruption and Women in Government. Journal of Economic Behavior & Organization, 46(4), 423–429. 

Dreher, A., & Herzfeld, T. (2005). The Economic Costs of Corruption: A Survey and New Evidence (SSRN Scholarly Paper No. 734184). Social Science Research Network. 

Dungan, J., Waytz, A., & Young, L. (2014). Corruption in the Context of Moral Tradeoffs. Journal of Interdisciplinary Economics, 26(1–2), 97–118. 

Duvanova, D. (2014). Economic Regulations, Red Tape, and Bureaucratic Corruption in Post-Communist Economies. World Development, 59, 298–312. 

Estrada-Gallego, F. (2009). Comunicación indirecta en situaciones de soborno y amenaza. Sociedad y economía, 16(1), 67–85.

Feinberg, J. (1987). Harm to Others (Vol. 1). Oxford University Press.

Fisman, R., & Miguel, E. (2007). Corruption, Norms, and Legal Enforcement: Evidence from Diplomatic Parking Tickets. Journal of Political Economy, 115(6), 1020–1048.

Fredriksson, P. G., & Svensson, J. (2003). Political instability, corruption and policy formation: The case of environmental policy. Journal of Public Economics, 87(7), 1383–1405. 

Gannett, A., & Rector, C. (2015). The Rationalization of Political Corruption. Public Integrity, 17(2), 165–175. 

Glaeser, E. L., & Saks, R. E. (2006). Corruption in America. Journal of Public Economics, 90(6), 1053–1072.

Goffman, E. (1974). Frame Analysis: An Essay on the Organization of Experience (Later Reprint edition). Northeastern University Press.

Gray, K., Schein, C., & Ward, A. F. (2014). The myth of harmless wrongs in moral cognition: Automatic dyadic completion from sin to suffering. Journal of Experimental Psychology. General, 143(4), 1600–1615. 

Guo, X., & Tu, W. (2017). Corruption Tolerance and its Influencing Factors—The Case of China’s Civil Servants. Journal of Chinese Governance, 2(3), 307–328. 

Gupta, S., Ogada, C., & Akitoby, B. (2016). Corruption: Costs and Mitigating Strategies (Working Paper SDN/16/06; IMF Staff Discussion Note). International Monetary Fund.

Guriev, S. (2004). Red Tape and Corruption. Journal of Development Economics, 73(2), 489–504. 

Código Penal del Estado de Guanajuato, Pub. L. No. P.O. 05-07-2018, 1 (2001). https://poderjudicial-gto.gob.mx/modules.php?name=Transparencia&file=index&func=contenido&cat=1&scat=9

Hobsbawm, E. (2000). Bandits. The New Press.

Huang, J., Sena, V., Li, J., & Ozdemir, S. (2021). Message Framing in P2P Lending Relationships. Journal of Business Research, 122, 761–773. 

INEGI. (2020). Encuesta Nacional de Calidad e Impacto Gubernamental (ENCIG) 2019: Principales Resultados (p. 20) [Data set]. https://www.inegi.org.mx/contenidos/programas/encig/2019/doc/encig2019_principales_resultados.pdf

Iyengar, S. (1996). Framing Responsibility for Political Issues. The ANNALS of the American Academy of Political and Social Science, 546(1), 59–70. 

Jackson, M., & Smith, R. (1996). Inside Moves and Outside Views: An Australian Case Study of Elite and Public Perceptions of Political Corruption. Governance, 9(1), 23–42. 

Jancsics, D. (2019). Corruption as Resource Transfer: An Interdisciplinary Synthesis. Public Administration Review, 79(4), 523–537. 

Kahneman, D., Knetsch, J. L., & Thaler, R. H. (1986). Fairness and The Assumptions of Economics. Journal of Business, S285–S300.

Kaufmann, D. (2005). Myths and Realities of Governance and Corruption. In A. Lopez-Claros & K. M. E. Schaw, Global Competitiveness Report 2005-06 (pp. 81–98). Pelgrave.

Kish-Gephart, J., Harrison, D., & Klebe Treviño, L. (2010). Bad Apples, Bad Cases, and Bad Barrels: Meta-Analytic Evidence About Sources of Unethical Decisions at Work. The Journal of Applied Psychology, 95, 1–31. 

Lambsdorff, J. G. (2006). Causes and Consequences of Corruption: What Do We Know From a Cross-Section of Countries. In S. Rose-Ackerman, International Handbook on the Economics of Corruption (pp. xiv–xxxviii). Edward Elgar Publishing.

Lederman, D., Loayza, N. V., & Soares, R. R. (2005). Accountability and Corruption: Political Institutions Matter. Economics and Politics, 17(1), 1–35. 

Lee, W.-S., & Guven, C. (2013). Engaging in Corruption: The Influence of Cultural Values and Contagion Effects at the Microlevel. Journal of Economic Psychology, 39, 287–300.

Li, L. (2018). The Moral Economy of Guanxi and The Market of Corruption: Networks, Brokers and Corruption in China’s Courts. International Political Science Review, 39(5), 634–646. 

Majer, J. M., Trötschel, R., Galinsky, A. D., & Loschelder, D. D. (2020). Open to Offers, but Resisting Requests: How The Framing of Anchors Affects Motivation and Negotiated Outcomes. Journal of Personality and Social Psychology, 119(3), 582–599. 

Mancuso, M. (1993). The Ethical Attitudes of British MPs: A Typology. Parliamentary Affairs, 46(2), 179–192.

Marquette, H., & Peiffer, C. (2018). Grappling With The “Real Politics” of Systemic Corruption: Theoretical Debates Versus “Real-World” Functions. Governance, 31(3), 499–514. 

Masters, A. B., & Graycar, A. (2016). Making Corruption Disappear in Local Government. Public Integrity, 18(1), 42–58.

McGraw, A. P., & Tetlock, P. E. (2005). Taboo Trade-Offs, Relational Framing, and the Acceptability of Exchanges. Journal of Consumer Psychology, 15(1), 2–15. 

McGraw, K. M. (1990). Avoiding Blame: An Experimental Investigation of Political Excuses and Justifications. British Journal of Political Science, 20(1), 119–131.

McLaren, R. M. (2015). Relational Framing Theory. In The International Encyclopedia of Interpersonal Communication (pp. 1–5). American Cancer Society. 

Meza, O., & Pérez‐Chiqués, E. (2020). Corruption Consolidation in Local Governments: A Grounded Analytical Framework. Public Administration, online first. 

Mo, P. H. (2001). Corruption and Economic Growth. Journal of Comparative Economics, 29(1), 66–79.

Montgomery, J. M., Nyhan, B., & Torres, M. (2018). How Conditioning on Posttreatment Variables Can Ruin Your Experiment and What to Do About It. American Journal of Political Science, 62(3), 760–775. 

Nas, T. F., Price, A. C., & Weber, C. T. (1986). A Policy-Oriented Theory of Corruption. The American Political Science Review, 80(1), 107–119.

Olivier de Sardan, J.-P. (1999). A Moral Economy of Corruption in Africa? The Journal of Modern African Studies, 37(1), 25–52.

Peeters, R., Gofen, A., & Meza, O. (2020). Gaming the System: Responses to Dissatisfaction with Public Services Beyond Exit and Voice. Public Administration, 98(4), 824–839.

Perez-Chiqués, E., & Meza, O. (forthcoming). Trust-based Corruption Networks: A Comparative Analysis of Two Municipal Governments. Governance: An International Journal of Policy, Administration and Institutions.

Peters, J. G., & Welch, S. (1978). Political Corruption in America: A Search for Definitions and a Theory, or If Political Corruption Is in the Mainstream of American Politics Why Is It Not in the Mainstream of American Politics Research?*. American Political Science Review, 72(3), 974–984. 

Pinker, S. (2008). The Stuff of Thought: Language as a Window into Human Nature (Reprint edition). Penguin Books.

Pinto, J., Leana, C. R., & Pil, F. K. (2008). Corrupt Organizations or Organizations of Corrupt Individuals? Two Types of Organization-Level Corruption. Academy of Management Review, 33(3), 685–709. 

Rabl, T., & Kühlmann, T. (2009). Why or Why Not? Rationalizing Corruption in Organizations. Cross Cultural Management: An International Journal, 16(3), 268–286. 

Redlawsk, D. P., & McCann, J. A. (2005). Popular Interpretation of “Corruption” and their Partisan Consequences. Political Behavior, 27(3), 261–283.

Rojas, M. (2020). Corruption and Weak Institutions. In M. Rojas (Ed.), Well-Being in Latin America: Drivers and Policies (pp. 155–164). Springer International Publishing. 

Rose-Ackerman, S. (1975). The Economics of Corruption. Journal of Public Economics, 4(2), 187–203. 

Rose-Ackerman, S. (1999). Corruption and Government: Causes, Consequences, and Reform. Cambridge University Press.

Rothstein, B. (2011). Anti-Corruption: The Indirect ‘Big bang’ Approach. Review of International Political Economy, 18(2), 228–250. 

Schindler, S., & Pfattheicher, S. (2017). The Frame of the Game: Loss-Framing Increases Dishonest Behavior. Journal of Experimental Social Psychology, 69, 172–177. 

Seal, G. (2009). The Robin Hood Principle: Folklore, History, and the Social Bandit. Journal of Folklore Research, 46(1), 67–89.

Shleifer, A., & Vishny, R. W. (1993). Corruption. The Quarterly Journal of Economics, 108(3), 599–617.

Swamy, A., Knack, S., Lee, Y., & Azfar, O. (2001). Gender and Corruption. Journal of Development Economics, 64(1), 25–55.

Treisman, D. (2000). The Causes of Corruption: A Cross-National Study. Journal of Public Economics, 76(3), 399–457. 

Truex, R. (2011). Corruption, Attitudes, and Education: Survey Evidence from Nepal. World Development, 39(7), 1133–1142. 

Tversky, A., & Kahneman, D. (1981). The framing of decisions and the psychology of choice. Science, 211(4481), 453–458. 

Van Rijckeghem, C., & Weder, B. (2001). Bureaucratic corruption and the rate of temptation: Do wages in the civil service affect corruption and by how much? Journal of Development Economics, 65(2), 307–331. 

Winters, M. S., & Weitz-Shapiro, R. (2013). Lacking Information or Condoning Corruption: When Do Voters Support Corrupt Politicians? Comparative Politics, 45(4), 418–436. 

Yang, Y., Vosgerau, J., & Loewenstein, G. (2013). Framing Influences Willingness to Pay but Not Willingness to Accept. Journal of Marketing Research, 50(6), 725–738. 

Zechmeister, E. J., & Zizumbo-Colunga, D. (2013). The Varying Political Toll of Concerns About Corruption in Good Versus Bad Economic Times. Comparative Political Studies, 46(10), 1190–1218.

Zizumbo-Colunga, D., & Amador, B. (2018). Corrupción en México. In V. Romero, P. Paras, G. Pizzolito, & Z. Elizabeth, Cultura Política de la Democracia en  México y en las Américas, 2016/2017: Un Estudio Comparado sobre Democracia y Gobernabilidad (pp. 93–112). Vanderbilt University.

Zmolnig, R. (2018). Framing Corruption: How Language Affects Norms. Crime, Law and Social Change, 70(2), 179–195. 

Zyglidopoulos, S. C., Fleming, P. J., & Rothenberg, S. (2009). Rationalization, Overcompensation and the Escalation of Corruption in Organizations. Journal of Business Ethics, 84(1), 65–73.

Endnotes

[1] We understand tolerance for corruption as a continuum that goes from accepting, participating in, and promoting corruption to condemning, combating, and denouncing this behavior.

[2] The retrospective use of frames refers to the case when a person chooses and uses a frame to justify or rationalize an act of corruption once it has been committed.

[3] We included one abstract vignette and one concrete vignette per area.

[4] See online Appendix E for an analysis of the effective number of observations.

[5] We only asked respondents about their willingness to misbehave in concrete vignettes. After the study, the researchers aided the transparency unit of the municipality in its efforts to implement anti-corruption measures.

[6] See online appendix A for the precise wording of the exact question wording.

[7] There is no academic consensus on the appropriateness of holding information constant across frames. While some have designed experiments to isolate framing from information (Berinsky & Kinder, 2006; Tversky & Kahneman, 1981), others have taken a less stringent strategy (Chen & Zhang, 2016; Iyengar, 1996) perhaps due to the fact that accounting for informational differences across frames risks blocking an important mechanism by which media and relational frames exert their effect and, ultimately, introducing post treatment bias (Montgomery et al., 2018).

[8] We used randomization without replacement to reduce the treatment repetition (see online Appendix B).

[9] Replication material can be found at: https://doi.org/10.7910/DVN/0XM0MU

[10] The municipal government studied is in one of the five states with the lowest levels of perceived corruption to be the lowest (INEGI, 2020).

[11] We found little difference in the proportion of participants who justify at least one vignette in our sample and the proportion of Mexicans who justify bribes as measured by LAPOP (see online Appendix D).

[12] While we find little evidence of social desirability unbalances across conditions (online Appendix F) we do find that social desirability moderates the effect of framing toward the null hypothesis.

[13] Our results are robust to an ordered logistic specification that relaxes some of OLS’ assumptions about the dependent variable (see online Appendix J).

[14] In these models, individuals who hold their answer constant (strainghtline) do not contribute any information to the estimates. However, we find strainghtlining to be uncorrelated with demographics, and to moderate our effects toward the null hypothesis (online Appendix H). In addition, we found our results to be robust to clustering the standard errors at the individual level (See online Appendix I).

[15] As online Appendix G shows, we have limited statistical power to identify interaction effects, thus, our null findings should be interpreted carefully