Recognizing the Potential “Destructive” Power of Social Norms

In this post Cheyanne and Russell share an early finding from their research into social norms in anti-corruption programming. Evidence shows that efforts to combat a harmful practice by depicting it as widespread or frequent can backfire by unintentionally increasing the practice. This bodes ill for well-intentioned programming like “I Paid a Bribe” sites that depict corruption as widespread, or for efforts to have the issue raised more in the press.

In our effort to figure out if social norms are a key missing link in anti-corruption programming, we have been carefully mining social norms literature for the key points anti-corruption practitioners need to know.  We have much more to do, but felt one reasonably clear finding was worthy of sharing early.  It illuminates potential unintended consequences of some common forms of anti-corruption programming.

Saying how bad something is, can make it worse

Put simply, if a program conveys a practice like littering, drug use, or corruption as widespread in an effort to educate the public about its deleterious effects, the program may actually result in an increase of the negative behavior.

Evidence from a range of disciplines (e.g. psychology, sociology, marketing) that study social norms agrees on one finding: efforts to combat a harmful practice by depicting it as widespread or frequent can backfire by unintentionally increasing the practice.  This is an example of the “social proof” effect, and it derives from the tendency for human beings to imitate what they perceive is typical or common among their peers.  They do this either because doing so serves as a cognitive shortcut in unfamiliar situations, or because we feel validated or comfortable imitating the behaviors of those similar to us.

Psychology Professor Robert Cialdini has written most extensively on the subject, and describes in an article, “Within the statement ‘Many people are doing this undesirable thing’ lurks the powerful and undercutting normative message ‘Many people are doing it.’” If people are emotionally or cognitively motivated to do what they perceive as typical or common, increasing perceptions that a negative behavior is common can lead individuals to normalize, or worse, even increase that behavior!

This means that awareness-raising campaigns that are meant to inform or shock the public by how bad the situation is in the hopes of reducing the behavior can actually increase that behavior. This bodes ill for well-intentioned programming like “I Paid a Bribe” sites that depict corruption as widespread, or for efforts to have the issue raised more in the press. Headlines that scream about the abuse of power day after day could reinforce the perception that everyone is doing it and that such abuse is socially permissible, excusable, or unassailable.

This is really interesting, but is there evidence?

Yes but no!  Most of the evidence for social proof is derived from laboratory experiments (some about corruption) and field experiments on issues unrelated to corruption.  For instance, Robert Cialdini has shown that individuals are more likely to litter in an environment that is already littered – because the environment signals that littering is common – and may do so even with subtle reminders in their environment that littering is disapproved of. Marketing experiments have shown that signs at national parks designed to deter theft of firewood have increased theft if the signs depict stealing firewood as common, albeit wrong, whereas signs that depict firewood theft as disapproved of but less frequent lead to reduced theft. The same has been shown for a study examining fliers in hotel rooms that chastise residents for not reusing their towels as opposed to fliers indicating that most residents reuse their towels instead of immediately throwing them in the laundry. A field experiment on social proof tried to shift individuals’ energy consumption by providing them with information about town-wide average levels of energy consumption.  The researchers found that the information about mean levels of consumption “acted like a magnet” – people who consumed above the mean amount reduced their consumption, but people who consumed below the mean amount increased their consumption to match the average level of gas consumption in their community!

Is there evidence that social proof happens with corruption?

A few academics have examined the effects of perceptions on corruption and the results suggest the social proof effects hold.  These experiments have been conducted primarily in North America and Western and Southern Europe. One laboratory experiment elicited subjects’ perceptions about the frequency of corruption, and found that those who perceived corruption as common were more likely to behave in a “corrupt” way in an experimental game. When subjects were primed with information that the type of corruption relevant to the game was infrequent in real life, they behaved less “corruptly.” A slightly different field experiment engaged individuals from different geographic regions of Italy – one widely perceived as more corrupt than the other – in a series of corruption games.  Individuals from the more corrupt region were more likely to give and take bribes, and the author argues that perceptions that corruption is common fueled the willingness of individuals from the more “corrupt” region to engage in corruption in the game.

However, we have not been able to find research or evaluations that show this is happening on the ground in reaction to anti-corruption campaigns.  This is not the same as saying it does not happen.  Here we get caught by the dearth of robust/quality evaluations on anti-corruption programming.  Absent evaluations that compare levels of corruption after an intervention has taken place versus a situation with no intervention – and without any evidence from evaluations that look for unintended effects of programs – we cannot confirm that the social proof effect occurs for anti-corruption programming as it has in numerous laboratory and field experiments.

How transferable are the findings on social proof?

Given that our focus is on corruption in fragile contexts, and the bulk of evidence is either laboratory-based or conducted in the field in North America, we are concerned about the transferability of the theory.  Many different aspects of context matter, including the topic being addressed, culture, and state of security/fragility.   For instance, would the phenomena be different in more collective cultures rather than the largely individualistic cultures of North America?   Would higher degrees of insecurity make the impact of possible social sanction – some argue a key feature in maintaining compliance to a social norm – more powerful?  Are the drivers behind negative behaviors like binge drinking (from which much social proof evidence is derived) comparable to corruption, where there are winners as well as losers?

If social proof is applicable to corruption in fragile states, then what…

There are clearly still questions to be asked and research to be done. Regardless, the theory and experimental evidence is troubling in light of institutionalized anti-corruption practice.  What does the Corruption Perceptions Index indicate about and do for countries with high levels of corruption?  For those countries at the bottom of the list, it certainly indicates that there is a widespread (read common) corruption problem. In doing so, does the CPI cement perceptions among citizens that corruption in their context is normal, and maybe even unassailable?

If these perceptions can sustain or increase corruption, those who conduct research on corruption and who design anti-corruption programs need to become more cognizant of the potential unintended effects.  Given the transferability questions, at least programs should be monitoring for unintended negative effects and at best redeveloping their messages away from the amount or frequency of the problem.

Please send us your experience with corruption and social norms

We are eager to hear from anti-corruption practitioners if you have ever faced the problem of social proof in programming, and what you have done about it.  Has program monitoring or evaluation shown that corruption has unintentionally gone up in reaction to an awareness raising campaign or school outreach program?  If you have research, evaluations or personal experience that supports or contradicts these findings we would really like to hear from you.  Please be in touch.

Cheyanne Scharbatke-Church is Principal at Besa: Catalyzing Strategic Change, a social enterprise committed to catalysing significant change on strategic issues in places experiencing conflict and structural or overt physical violence. She has significant experience working on anti-corruption and state legitimacy in the DRC and Uganda.  As a Professor of Practice she teaches and consults on program design, monitoring, evaluation and learning. Cheyanne is also a regular author and co-curator of the blog series on corruption in fragile states.

Russell Hathaway is a Masters student at the Fletcher School of Tufts University studying conflict resolution and development economics. He received his BA from The University of Chicago, where he studied the linkages between post-conflict disarmament, demobilization, and reintegration (DDR) and transitional justice. Outside of the classroom, Russell has researched sexual violence in the Syrian Civil War at UN Headquarters, authoring a report on the subject for UN Women. He also served in a project support capacity at the international development organization Heartland Alliance and has traveled extensively in Turkey and Germany. Most recently, Russell served in the Bureau of Near Eastern Affairs and the political section of U.S. Embassy Berlin as a U.S. Foreign Service Intern with the State Department. Following graduate school, he will join the Foreign Service through the Pickering Foreign Affairs Fellowship.