Edited by:
Reviewed by:
*Correspondence:
This article was submitted to Cognitive Science, a section of the journal Frontiers in Psychology
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
This paper combines experimental and field data to examine how authorities with discretion over how rules are enforced penalize transgressors when the social context of the transgression elicits expectations of leniency. Specifically, we test how transgressors are punished when it is their birthday: a day that triggers expectations of lenient treatment. First, in three scenario studies we explore individuals’ intuitions about how they would behave and expect to be treated if they transgressed on their birthdays, as well as how they would imagine penalizing a birthday transgressor. Second, using more than 134,000 arrest records for drunk driving in Washington State, we establish that police officers penalize drivers more harshly when it is their birthday. Then, in a lab experiment in which we grant participants discretion over enforcing the rules of an essay-writing contest, we test psychological reactance toward transgressors who make their birthday salient, even subtly, as the mechanism behind this increased stringency. We rule out several alternative explanations for this effect, including public safety concerns, negative affect and overcompensation for bias. We conclude with a discussion of the theoretical and practical implications of our findings for the literatures on punishment, rule-breaking, and legal transgressions.
Scholars have theorized about when and how to punish individuals who transgress laws, rules, or regulations (
In this paper, we examine how transgressions are penalized when they occur in a social context that elicits expectations of leniency. Transgressions that occur in a social context in which the transgressor expects leniency put authorities with the discretion over punishing them into a difficult bind, needing to balance the motivation to meet the expectations elicited by the social context with the competing motivation to punish fairly and effectively. We argue that although transgressors may believe that transgressing in a social context that elicits expectations of leniency will lead to lighter penalties, this belief is misguided. Instead, we argue that—contrary to intuition—when authorities have a responsibility to enforce rules, but face a conflicting motivation to be lenient, they resolve this conflict in favor of harsher penalties rather than in favor of increased leniency.
Our research makes several theoretical contributions. First, we contribute to existing research on punishment by exploring how the social context of transgressions influences punishment decisions. Second, we extend our understanding of how individuals manage conflicting motivations when they have discretion over penalties. This is important because situational factors that trigger expectations of preferential treatment are pervasive, but many do not justify leniency in punishment decisions. Third, our work extends the literature on bias in punishment decisions by shifting the focus from discrimination on the basis of demographic characteristics such as race or gender to a focus on how people manage competing motivations to act. Ultimately, this work informs our understanding of the challenges in exercising discretion fairly and effectively (
While laws and regulations provide guidelines for how to punish transgressions, individuals (e.g., managers, judges, or police officers), and groups (e.g., panels, boards, or juries) typically have discretion regarding whether and how much to punish those who transgress. Discretion over arrests (
In the United States, the risk that discretion in punishment leads to unfair treatment of certain demographic groups has led to a number of high-profile initiatives to understand the extent and implications of these biases (e.g.,
One body of work that focuses directly on how we are motivated to punish is the literature on “just deserts” (
However, transgressions always occur in a broader social context. Elements of this broader social context may motivate expectations of leniency, but are arguably unrelated to the crime itself. Some elements of the social context create legitimate reasons to treat transgressors gently. For example, there is a norm of treating young people more leniently than adults when they transgress rules, as there are good arguments for why culpability is impaired before reaching maturity (
Other aspects of the social context that motivate people to treat transgressors more leniently are less legitimate. Social norms of deference to authority (
“Special days” such as birthdays also elicit expectations of preferential or favorable treatment that may extend to expectations of leniency in the context of transgressions. Birthdays are part of a larger class of days that have social or religious significance (e.g., Christmas, Yom Kippur) and are associated with strong norms of helping, kindness, and forgiveness. These days affect pro-social behavior such as charitable contributions (
It is not a big leap to suggest that the expectation that individuals should receive special treatment on their birthdays will even extend to expectations of leniency when important rules are broken. On December 2, 2012, the rapper known as “The Game” was pulled over by Los Angeles Police because the car he was driving had invalid license plates. A celebrity website recounted that although the car was unregistered, the police released him and did not tow his car “because it was his birthday.”
In this paper, we focus on the social context of birthdays as a situational cue that motivates leniency for two reasons. First, even though this norm should be irrelevant in punishment decisions, for someone with the responsibility to penalize transgressors, a birthday elicits a competing motivation to treat that particular individual with leniency. Second, birthdays are ideal for studying the effect of the social context in non-experimental field settings because they are randomly distributed in the population: in most contexts, authorities that apprehend transgressors do not and could not have known it was transgressor’s birthday before apprehending them. This means that birthday and non-birthday transgressors are randomly assigned to punishers, reducing endogeneity concerns about selection bias and causality when identifying variation in punishment decisions using data from the field.
If someone is caught transgressing a rule on their birthday, they have a choice between making that fact salient or not. There are several reasons why transgressors are likely to make salient a relevant fact that may motivate lenient treatment of them (“But it’s my birthday!”). Individuals tend to volunteer reasons for their misbehavior in order to save face and reduce embarrassment (
The person in a position of authority needs to manage competing motivations to be lenient and to punish in a fair and effective way. Several reasons suggest that these competing motivations might be resolved in favor of increased leniency. Of course, the authority might try to ignore the expectation of leniency and punish transgressions as if this expectation did not exist. However, as prior research demonstrates, individuals often proceed rather automatically to enact scripted cues to behave in certain ways, even when the cue is completely irrelevant to the behavior it mindlessly triggers (
Alternatively, it might be uncomfortable for an authority to behave counter to this expectation, particularly if they have the discretion over the transgressor’s punishment. Extensive research shows that people tend to behave consistently with what is perceived to be the normative expectation in the situation, particularly when those norms are made salient (
Such a prediction, however, ignores important psychological mechanisms involving the interaction between the transgressor and the authority. If the transgressor draws attention to his birthday, even subtly (as they are likely to, in an effort to capitalize on the expectation they have that this will lead to lenient treatment), this may shift the decision process of the authority. Although authorities may be indifferent or even positively inclined to treat a birthday transgressor with leniency, they may be particularly sensitive to perceptions that leniency is being solicited and react negatively to them. Consistent with the drive to punish in a fair and effective way, authorities may penalize transgressions more stringently when the social context cues expectations of leniency, because they will react negatively to any perception that their leniency is being solicited.
There is sound theoretical basis for such a prediction of stringency in psychological reactance theory (
Second, reactance effects increase when requests for help appear inappropriate or illegitimate (
Finally, when authorities have an obligation to penalize transgressions, they are motivated to ensure that the punishment is fair and appropriate.
We argue that when individuals perceive that someone is demanding something from them, whether the demand is explicit (actively soliciting leniency) or implicit (making salient an element of the social context that creates an expectation of leniency), they will experience the demand as a threat to their autonomy, and become less inclined to do it (
We now present six studies that test this argument using multiple methods and data sources. First, a series of scenario studies establishes that birthdays do represent a social context in which transgressors expect leniency, even though when individuals imagine themselves as authority figures with discretion over punishment decisions, they report higher levels of psychological reactance toward birthday offenders. Second, using 9 years of DUI (Driving Under the Influence) arrest records in the state of Washington (over 134,000 arrest records), we show that police officers punish marginal offenders more stringently on their birthdays than on other days. In a series of robustness checks, we show that it is unlikely that these results are explained by substantive differences in intoxication or public safety risk, but are instead likely based on the discretionary decisions of officers. Third, in a lab experiment in which we vary the birthday status of individuals who have transgressed rules, we demonstrate that individuals treat transgressors more stringently on their birthdays as a function of the psychological reactance triggered by the birthday status of the transgressor. A final study, using a similar experimental paradigm in the lab, rules out overcompensation for bias as the mechanism behind our effects.
We ran three studies, in separate online samples, to explore individuals’ intuitions about whether they would expect to be treated leniently if they were pulled over for drunk on their birthday (Study 1a), whether they would volunteer that it was their birthday if they happened to be pulled over by a police officer on that day (Study 1b), and what they believe they would do themselves if they had discretion over penalizing a marginally drunk driver on their birthday (Study 1c). Together, our aim was to build a picture of what might occur in an actual interaction between a transgressor and an authority with discretion over penalizing the transgression on the transgressor’s birthday. These studies were conducted on Amazon Turk, an online labor market where ‘requesters’ can post short tasks for ‘workers’ to complete for a small fee. Studies have found that data collected through Amazon Turk are of comparable quality to data collected through more traditional methods (
The first solicited individuals’ intuitions about how they expect they would be treated if they made their birthday salient in the context of transgressing. We predicted that their intuition would be that they would be treated more leniently on their birthdays.
We paid 306 participants (60% male;
Participants were then asked to make a forced choice prediction about whether the officer would arrest them or release them with a warning.
No one failed the attention check in this study, and everyone completed the main outcome measures; thus, results are reported for the whole sample. There were significant differences by condition in terms of the proportion of respondents who believed they would be arrested, χ2(1,
This study solicited individuals’ intuitions about they would do if they were stopped for drinking and driving on their birthday, as well as reasons behind their choice.
We paid 112 participants (56% male;
We then asked them to indicate (on a 5-point scale) to what extent they agreed with four statements about why they might have made the choice they did: (1) It would result in the most lenient treatment from the officer; (2) It was the best excuse for my behavior; (3) It was the most appropriate choice to make; and (4) It would be the easiest thing to do.
We did not include an attention check in this study, so results are reported for the whole sample. Thirty-five of the respondents (31%) said that they would mention their birthday to the officer. Those who said they would mention their birthday to the officer reported significantly higher levels of agreement with the statements that doing so: (1) would result in more lenient treatment from the officer [
In our final scenario study, we asked participants to imagine themselves in the role of the police officer. We wanted to see if the leniency they predicted they would receive as the driver would translate when they imagined themselves in the role of the police officer. We also wanted to assess how individuals in the role of the authority reacted to different ways that drivers might make their birthday salient, as a preliminary test of Hypothesis 2.
We paid 273 participants (62% male;
[
[
[
[
Participants were then asked to make a forced choice prediction about whether they would arrest the driver or release them with a warning.
We also tested individuals’ psychological reactions to the scenarios. We used a 3-item measure of threat to freedom that has been used to study psychological reactance (
Four participants failed the attention check question in this study; results are reported for the remaining 269 participants. There were significant differences in whether participants reported they would arrest the driver, χ2(3,
The scenarios also elicited different levels of psychological reactance in the participants,
Participants’ negative affect also significantly differed by condition,
These results provide preliminary support for Hypothesis 2, that a transgressor’s birthday elicits negative psychological reactions among individuals with discretion over their punishment. In addition, the more obvious the effort to capitalize on the social expectation of leniency, the more negative the reaction.
Together, these results suggest three things. First, birthdays do represent a social context in which individuals expect to receive lenient treatment for their transgressions – even if the transgression is quite severe. Second, though still a minority, a substantial proportion of individuals claim they would mention it was their birthday to a police officer if they were pulled over for drunk driving on that day. Third, individuals imagining themselves in the role of a police officer believe they would only treat birthday offenders more stringently if the driver attempted to use that fact to solicit lenient treatment for his offense. Third, even though respondents reported they would only treat birthday offenders more stringently if they used the birthday to solicit leniency (this was also the only condition that elicited significantly more negative affect from the respondent), any mention of the transgressor’s birthday elicited psychological reactance. This last finding suggests that individuals with discretion over punishment may have more general psychological reactions to birthday transgressors.
In Study 2, we use a unique sample of field data to identify how individuals in positions of authority actually penalize transgressions in a social context that elicits expectations of leniency. Specifically, we study arrests involving suspicion of DUI of alcohol in the state of Washington, and test whether otherwise similar drivers are more likely to be arrested if it is their birthday, compared to those for whom it is not their birthday.
In all U.S. states, driving while intoxicated by alcohol (drunk driving) is prohibited and has a severe impact on public safety. Economists have estimated that intoxicated drivers create externalities of at least 30 cents per mile driven due to social welfare costs of traffic fatalities (
Driving under the influence laws are enforced by several police agencies in Washington State, including the Washington State Patrol, which is responsible for monitoring and enforcing the state’s highway systems, as well as local agencies, including municipal police, county sheriff’s offices, and Indian Nations agencies. In Washington State, DUI laws are primarily based on the driver’s blood alcohol level (BAC). Drivers whose BAC exceeds 0.08% are said to be in
When an officer suspects a driver of DUI, she typically administers a field sobriety test. Furthermore, the officer administers a mobile breath test (“breathalyzer”), which estimates the BAC of the driver. If the officer determines the driver to be intoxicated, the driver is placed under arrest and taken to a field station for a formal (and admissible in court) breath test. If the officer observes a mobile BAC greater than 0.08, the decision is straightforward. The driver is almost certainly in
Arresting the driver presents several potential costs for the officer. First, the arrest process takes the officer off the road for several hours and thereby precludes her from potentially arresting an even more highly intoxicated driver. Second, drivers who do not violate the
Our data include every DUI arrest in Washington State from 2001 to 2009. These data include the agency and identity of the arresting officer as well as the name, age, gender, and ethnicity of the driver. Also included are the date, time, and location of the arrest. The data also note the primary criminal charge, which allows us to exclude DUI arrests that are secondary to more severe crimes such as weapons violations, violent crimes, or outstanding arrest warrants. The data also identify the mobile BAC reading, when taken, as well as the court-admissible BAC reading from the police station. Since the data also identify the exact time of each test, we know the length of delay before the driver was given the court-admissible test. We present basic summary statistics in
Study 2: Descriptive statistics for DUI arrests.
All arrests |
Birthday arrests | Other arrests | ||
---|---|---|---|---|
Variable | Mean | Mean | Mean | |
Field BAC | 0.133 | 0.049 | 0.134 | 0.134 |
0.94 | 0.24 | 0.92 | 0.94 | |
Field BAC – Station BAC | -2.18 | 35.12 | -1.06 | -2.18 |
Minutes from field to station | 60.73 | 46.75 | 59.17 | 60.74 |
Birthday driver | 0.004 | 0.062 | 1 | 0 |
Driver age | 34.33 | 11.38 | 37.08 | 34.32 |
Female driver | 0.21 | 0.41 | 0.23 | 0.21 |
White driver | 0.83 | 0.38 | 0.82 | 0.83 |
Number of observations | 134,507 | 518 | 133,989 |
One weakness in our data is that we are unable to observe drivers stopped for suspicion of DUI but not arrested. Only drivers who were arrested appear in our data, creating potential survivor bias in any standard regression analysis. We will address this weakness by exploiting the discrete threshold at BAC = 0.08 in order to infer distributions of non-arrested drivers in the data. Another weakness is the relative rarity of birthdays, which represent 0.38% of all arrests vs. 0.27% (one out of 365.25) of all days. This rarity means that we must infer differences in birthday traffic stops from a substantially smaller sample than the total DUI database.
Using driver’s birthday as the context in which there is a social expectation of lenient treatment has several important characteristics from an identification perspective. First, the norm of preferential treatment on one’s birthday is universally known and widely observed. Second, birthdays are unobservable to an officer prior to a traffic stop and thus unlikely to create an unobservable selection bias in traffic stops. Third, birthdays are randomly distributed and uncorrelated with other factors that might affect officer leniency. This third point is critical for our decision to examine birthdays instead of other holidays such as Valentine’s Day, Mother’s Day, or Christmas, which may affect the officer. An officer showing leniency on Valentine’s Day, for example, may simply want to avoid a 2 h arrest that keeps him from dinner with a spouse, or he may be in a foul mood due to working on a holiday.
We identify officer stringency in DUI enforcement by observing how often officers arrest
We first present birthday arrest frequency for marginal and
The goal of our analysis is to identify how the behavioral interaction between the transgressor (driver) and punisher (officer) are affected by the expectation of leniency associated with birthdays. Consequently, we are concerned that the increased likelihood of a discretionary arrest on birthdays might simply reflect fundamental differences between the characteristics of birthday and non-birthday drivers. Similarly, the police who arrest them or the conditions under which they are arrested might be different. To address this, we implement regression analysis that estimates a decrease in the likelihood of a birthday driver for all arrests above the
Because officers cannot observe on which side of the threshold a moderately intoxicated driver lies before arrest, the assignment near the threshold is random for those stopped for DUI. Since DUI stops are randomly assigned to either side of the threshold, our theoretical argument is that the behavioral interaction between the driver and officer is the mechanism driving any discrete increase in birthday probability at the
Our first specification uses logistic regression to estimate the probability that an arrest involves a driver birthday as a function of the
Our base model with no control variables is presented in column 1 of
Study 2: Regression models predicting birthday arrests.
(1) | (2) | (3) | (4) | |
---|---|---|---|---|
Logit | Logit | Logit | OLS | |
Driver sample: | All | All | All | All |
Dependent variable: | Birthday | Birthday | Birthday | Birthday |
-0.543∗ (0.265) | -0.536∗ (0.265) | -0.488† (0.259) | -0.0021 (0.0013) | |
BAC | 0.008 (0.013) | 0.009 (0.013) | 0.006 (0.013) | 0.00003 (0.00005) |
BAC2 | -0.00003 (0.0001) | -0.00004 (0.0001) | -6.3 |
-3.1 |
BAC3 | 5.3 |
7.5 |
-3.8 |
-4.7 |
BAC4 | -8.4 |
-1.1 |
2.8 |
8.4 |
Month/Day dummies | No | Yes | Yes | Yes |
Year dummies | No | Yes | Yes | Yes |
Age | No | No | -0.363† (0.217) | -0.028∗∗ (0.003) |
Age2 | No | No | 0.012† (0.006) | 0.001∗∗ (0.0001) |
Age3 | No | No | -0.0002† (0.0001) | -0.00002∗∗ (1.7 |
Age4 | No | No | 9.1 |
9.4 |
Male | No | No | -0.060 (0.112) | -0.00007 (0.0005) |
Driver ethnicity dummies | No | No | Yes | Yes |
County dummies | No | No | Yes | Yes |
Officer FE | No | No | No | Yes |
Pseudo R-squared | 0.0007 | 0.0370 | 0.0412 | 0.0451 |
Number of observations | 134,507 | 134,507 | 133,795 | 134,507 |
As a robustness test, in column 4 we report a linear probability model with officer fixed effects, since the rarity of birthday events does not allow the use of logit models with fixed effects. The linear fixed effect model produces a coefficient very similar to our marginal effects, but with reduced statistical significance (
We note that we cannot estimate the marginal effect at other points farther below the threshold due to our identification strategy. Furthermore, the low number of birthday arrests in our data makes more formal testing of regression discontinuity models difficult, so our results would be more compelling if we could triangulate them using additional data.
To address possible differences in our non-birthday and birthday samples, we also created a matched sample based on observable driver demographics, BAC, and stop characteristics. We implement a propensity score matching algorithm that chooses the ten nearest non-birthday neighbors for each birthday arrest, which reduces our sample to 5,117 arrests (some non-birthday arrests are neighbors for multiple birthday arrests). Using this matched sample, we repeat the
Our evidence of higher stringency toward birthday drivers supports Hypothesis 3a, but raises a number of alternative explanations. One natural concern with our identification strategy is that BAC readings may not accurately represent the public safety risks of birthday drivers, and that the increased stringency we observe represents a rational police response to expectations of future accidents or increasing intoxication. We systematically examine these alternative explanations by testing differences in arrested drivers’ characteristics.
We first address whether the mobile BAC of birthday drivers stopped by police accurately reflects their level of alcohol consumption, relative to other drivers. In this alternative explanation, marginally drunk birthday drivers have alcohol in their stomach that has not yet entered the bloodstream due to binge drinking or a stomach full of food, and officers arrest the driver because of their tacit knowledge that their BAC will continue to climb above the
We next address the alternative explanation that marginally drunk birthday drivers may be inherently more dangerous than their non-birthday counterparts and that stringency toward them is a rational public safety response. For this alternative explanation, we test whether birthday drivers are more likely to drink (and drive) later in the evening, making their arrest a pre-emptive strategy for law enforcement. The average time of arrest is also nearly identical between the two groups (10:54 p.m. vs. 10:59 p.m.,
We also examined whether arrested birthday drivers were more likely to have it be their first DUI arrest (in the state) compared to other drivers. To do so, we used only those drivers where no officer discretion was involved (thus eliminating any birthday bias), and found that although it was somewhat more likely that birthday drivers were being arrested for the first time compared to non-birthday drivers (91% vs. 89%), the difference was statistically indistinguishable (Fisher’s exact test,
Another alternative explanation is that officers might punish marginally drunk birthday drivers more stringently because they believe birthday drivers are more likely to be “scared straight” by the arrest. Although we cannot observe other confounds (such as differences in conviction and sentencing), we tested whether those whose first arrest was on their birthday were less likely to be arrested for a later DUI. For
Finally, we examine whether intoxicated birthday drivers were more likely to be involved in an accident compared to other drunk drivers. Arresting after an accident where the driver has a positive BAC involve no police discretion (hence we had excluded arrests involving accidents from our main sample). To test whether officers may be arresting birthday drivers at a higher rate because they have insider knowledge that they are more likely than other drivers to cause later accidents, we compare the ratio of birthday drivers among those arrested for DUI offenses ending in accidents (arrests excluded from our main sample) to the ratio of birthday drivers among discretionary DUI arrests. The percentage of drunk-driving accidents involving birthday drivers is 0.43%, compared to 0.41% for discretionary officer arrests (
This similarity in birthday rates for non-discretionary accident rates also casts doubt on an alternative explanation that officers give fewer breath tests to birthday drivers, and consequently might show more stringency toward those under 0.08 to compensate for this prior leniency. If that were the case, then we would expect a lower average rate of birthday drivers in discretionary tests (non-accidents) than in mandatory ones, which we do not. Together, these tests cast doubt on alternative explanations for police stringency toward birthday drivers, but of course cannot disprove them.
Of course, we cannot know whether drivers are actually soliciting leniency in their interactions with police officers when they are pulled over on their birthdays. However, if Study 1b is any indication, a substantial minority of the individuals (31% of the study sample) reported that they would mention their birthday to the police officer. Alternatively, contrary to Study 1c, which suggested that reactance in the condition in which participants noticed the birthday was equal to the control condition, officers in the field may react negatively to drunk driver even if their birthday isn’t mentioned. Whatever occurs between the officers and the drivers in the field, our analysis provides support for Hypothesis 3a: drivers who are at the margins of the legal limit for blood alcohol are more likely to be arrested on their birthday than on other days. This effect appears to be unrelated to the public safety risk of these drivers, their demographics, and the conditions under which they are arrested.
The data from Study 2 do not allow us to test whether psychological reactance explains the apparent stringency toward birthday drivers, neither can they reveal whether this is a more general behavioral response or whether it is idiosyncratic to the setting of drunk driving. We address these concerns by designing a laboratory experiment using a different type of transgression, which additionally allows us to test psychological reactance as our hypothesized mechanism (Hypothesis 3b).
The behavioral lab at a UK-based business school (43% male;
Participants were randomly assigned to one of three conditions. There were two birthday conditions: one in which the transgressor was using his birthday as a reason to solicit preferential treatment (
We informed participants that the lab was partnering with a nearby school specializing in English as a Second Language to evaluate a student essay-writing competition. Participants were all assigned to the role of “evaluator” and tasked to judge three of the essays competing for prizes. We also told participants that, because teachers typically know the students in their classes before grading any of their work, the students had written a short paragraph about themselves, which would be attached to each essay. We used actual example essays from the American College Testing writing assessment arguing in favor of extending high school by 1 year. We chose two essays that the assessment service used as examples of poorly written essays (that had scored 1 and 2 out of 5) and one example of a good essay (that had scored 5 out of 5).
We provided participants with a scoring sheet and contest rules, which included a rule forbidding essays over 500 words from being eligible for prizes. The rule read
The birthday manipulation was included at the end of the personal statement attached to Essay #3 (the essay assessed by the American College Testing service as the one of the highest quality), which was always positioned last in the package. The handwritten personal statement also included a message that either mentioned the essay-writer’s birthday [
Participants were instructed to grade the essay’s unique ideas (10 points), persuasiveness (10 points), language quality (10 points), and grammar, spelling and punctuation (10 points). These points were summed to create a total score. We used participants’ scores as a manipulation check to confirm that the essay containing the birthday manipulation was evaluated as the “best” essay among the three, and thus the most likely to be nominated for a prize
We measured participants’ psychological reactance to each of the essay writers’ personal statements using the same 3-item measure of threat to freedom used in Study 1c (
The word counts of each essay were handwritten on each of the essays and circled. At 513 words, Essay #3 violated the 500-word limit rule by 13 words. Neither of the other two essays violated the word limit. The dependent variable of interest was whether participants treated Essay #3 with increased stringency by not nominating it for the prize even though it was the best of the three essays.
Five participants failed to complete all relevant measures, and were excluded from the analysis. Results are reported for the remaining 157 participants.
A repeated measures ANOVA with final score as the within-subjects factor confirmed the ranking of the essays provided by the American College Testing service. The essay scores significantly differed from each other,
We next established that participants’ psychological reactance to the third essay was affected by the condition to which they were assigned,
Hypothesis 3b predicted that an authority figure’s increased stringency (in this case, penalizing those who violated competition rules) as a function of targets’ birthdays is driven by psychological reactance. We used Preacher and Hayes’ PROCESS macro (
We ran three models for all the relevant comparisons, each using 5,000 bootstrap samples. We included the participants’ reactance toward the first and second essays as covariates in the analyses, as the reactance measures were significantly correlated with each other (between Essay 1 and Essay 2,
Study 3: Model summary information comparing indirect effects of birthday and control conditions on stringency via psychological reactance.
Consequent |
||||||
---|---|---|---|---|---|---|
M (psychological reactance) |
Y (stringency in punishment) | |||||
Antecedent | Coefficient | Coefficient | ||||
M (Reactance to Essay #3) | 0.42 | 0.19 | 0.028 | |||
Reactance to Essay #1 | 0.28 | 0.14 | 0.044 | -0.68 | 0.34 | 0.048 |
Reactance to Essay #2 | 0.39 | 0.12 | 0.002 | 0.19 | 0.30 | 0.519 |
X (Birthday-Soliciting-Leniency) | 0.72 | 0.18 | <0.001 | -0.11 | 0.45 | 0.813 |
AB (Effect of X on Y via M) | 0.31 | 0.17 | 95% CI: 0.035 to 0.708 | |||
X (Birthday-Mentioned) | 0.34 | 0.18 | 0.057 | 0.09 | 0.42 | 0.835 |
AB (Effect of X on Y via M) | 0.15 | 0.10 | 95% CI: 0.010 to 0.424 | |||
X (Birthday-Soliciting-Leniency) | 0.38 | 0.18 | 0.034 | -0.19 | 0.42 | 0.646 |
AB (Effect of X on Y via M) | 0.16 | 0.12 | 95% CI: 0.004 to 0.498 |
In a substantially different paradigm and using a different type of transgression, Study 3 shows that psychological reactance to transgressors on their birthdays drives authority figures’ increased stringency toward them. It is interesting to note the subtlety of the birthday manipulations in this study: even in the birthday-soliciting-leniency condition, the student did not use his birthday as an excuse for violating the rules of the essay-writing contest, but merely said they deserved the prize because it was their birthday. These subtle manipulations help strengthen our argument that merely making the social context of a birthday salient increases how stringently a transgressor will be treated by an authority figure with the discretion to do so. Additional tests confirmed that psychological reactance – the subjective perception that one’s freedom is threatened – functions as a mechanism behind this effect. In contrast, we did not find empirical support for negative affect as a mechanism in this study.
A second alternative explanation that could drive our results is overcompensation for bias. When attention is drawn to factors that may bias an individual’s evaluation of a target, a typical response is to try to correct for that possibility by adjusting the judgment away from the direction of the bias (
The behavioral lab at a UK-based business school (31% male;
This experiment used a manipulation which simply mentioned the essay-writer’s birthday without actively soliciting leniency [
We ran a logistic regression with stringency as the dependent variable, and birthday condition, bias condition, and their interaction as independent variables. The coefficients for the bias salience condition (
Our evidence from the field and the lab was consistent with the predictions of our theory. When confronted with a social expectation of lenient treatment, individuals with the authority and discretion to punish them treat transgressors more stringently rather than more leniently. In our studies, we used the transgressor’s birthday as a social context that leads to an expectation of lenient treatment, as it has many attractive characteristics that allow us to test this phenomenon in the field. Increased stringency for birthday transgressors, which we identified both in the field and in the lab, runs counter to what individuals believe happens to transgressors on their birthdays. We find that this effect is driven by psychological reactance toward the transgressor. Moreover, psychological reactance increases as the salience of the transgressor’s birthday increases (as, we assume, is the perception that the target is using his birthday to actively solicit lenient treatment).
Our results contribute to several literatures. First, our research contributes to a broader literature on punishment from psychological perspectives (
Second, this paper contributes to our knowledge of how individuals behave when a context elicits two different motivations with conflicting behavior prescriptions. The large body of work in both psychology (e.g.,
Third, these findings extend our understanding of psychological reactance among individuals with the discretion over penalizing transgressions. Most of the literature on reactance has focused on refusals to help others (
Ultimately, these findings help us understand how discretion is exercised in the field, thus deepening our knowledge of how discrimination operates. Work on discrimination has focused almost exclusively on demographic characteristics such as age, race, ethnicity, and gender (
Our research also has important practical implications, both for alleged transgressors as well as those with discretion over punishing them—from managers and teachers to judges and jury members. Transgressors need to be aware that their intuitions about avoiding punishment by making leniency norms salient may backfire, resulting in harsher penalties than if they refrained from making the norm salient. In other words, transgressors may benefit from avoiding any perception that they are trying to capitalize on contextual factors that would suggest more lenient treatment. On the other hand, authorities with the responsibility to punish should be aware that in the face of conflicting motivations, they may make decisions that undermine the fairness, and ultimately the effectiveness, of their sanctions.
Compared to others responsible for punishing transgressions, law enforcement officials may be particularly likely to react negatively to perceptions that offenders are soliciting lenient treatment. They are accustomed to excuses and pleas for leniency from those they penalizing, to the extent that such pleas can become tiresome and prompt cynicism (
Our research also has important practical implications for managers, who are commonly given broad discretion to punish employees through oral reprimands, work suspension, or, in extreme cases, termination (
Our findings suggest that when authority figures have discretion over punishment decisions, making an expectation that a transgressor will be treated leniently salient leads to a negative psychological reaction, leading individuals with the authority to punish to do so more harshly. This might lead to the conclusion that discretion is overrated or overused. Yet, we do not want to suggest that our findings provide an argument against discretion, only a fair warning about some of its additional problematic qualities. Many dysfunctional consequences result when discretion is unavailable, such as under mandatory punishment guidelines (e.g., “three strikes” laws). These consequences include higher rates of violence and murders among repeat offenders and against witnesses of repeat offenses (
The authors contributed equally to this work. LP collected, analyzed, and wrote up the field data. LP and CM conceived the experiments. CM ran the experiments, analyzed and wrote up the lab data. LP and CM discussed the results and implications and commented on the manuscript at all stages, and wrote the paper jointly.
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
We thank Nicholas P. Lovrich and the Division of Government Studies and Services at Washington State University for the opportunity to work with the Washington State Patrol on this project. Brady Horn and Dick Doane provided considerable help in data provision. We also thank Francesca Gino, and participants at the 2012 Academy of Management Meeting in Boston as well as the Behavioral Ethics Conference at the University of Central Florida for their insightful comments on presentations based on earlier drafts of this paper. Finally, we thank the Edmond J. Safra Center for Ethics, London Business School and the Olin Business School at the Washington University in St. Louis for their financial support of this research.
We use this terminology to indicate that such drivers are at the