The relationship between social class and unethical and prosocial (traffic) behavior: two naturalistic replication studies

ABSTRACT Whereas some previous studies have shown that upper-class individuals are more unethical and less prosocial than lower-class individuals, other studies reported contradictory results. The present research provides two replication attempts by examining whether upper-class drivers would be more likely to cut off a pedestrian at a marked crosswalk (Study 1) and less likely to let a pedestrian cross an unmarked pedestrian crossing (Study 2) than lower-class drivers. In both studies, the vehicle status as an index of the social class of the driver was not significantly related to whether the driver let the pedestrian cross the street. Overall, it appears that the associations between social class and unethical behavior and prosocial behavior are less robust and generalizable than initially thought.

individuals from lower-class backgrounds (reviewed below).Other studies, however, provided failed replications and either found that the rich and the poor do not differ in their unethical and prosocial behavior or that upper-class individuals are even less unethical and more prosocial than lower-class individuals.Given these contradictory findings, we deemed it important to provide further tests of the idea that upper-class individuals are more unethical and less prosocial than lower-class individuals.Concretely, in two naturalistic field studies, we examined whether individuals from different social classes would differ in their unethical and prosocial traffic behavior.

Social class and unethical behavior
As just noted, proponents of the idea that there are social class differences in unethical behavior argue that upper-class individuals more than lower-class individuals put their own interests ahead of the interests of others.In fact, research has shown that upper-class individuals tend to endorse the view that they are more deserving than others (Piff, 2014).In addition to these heightened feelings of psychological entitlement, upper-class individuals are more likely to report narcissistic personality tendencies (Greitemeyer & Sagioglou, 2016;Piff, 2014) and they hold more favorable attitudes toward greed (Piff et al., 2012) than lower-class individuals.
Based on the reasoning that upper-class more than lower-class individuals exhibit selffocused social-cognitive tendencies, Piff et al. (2012) hypothesized that upper-class individuals would be prone to engage in unethical behavior.In support of their hypothesis, upper-class individuals more than lower-class individuals drove recklessly, took goods from others for their own benefit, lied in negotiations, and cheated to increase their chances of winning a cash prize.Piff et al. (2012) research has garnered substantial attention from the popular press and has been cited over 1,200 times on Google Scholar (as of April, 2023).
Subsequent research successfully replicated the finding of a positive relationship between social class and unethical behavior (Boyd et al., 2015;Coughenour et al., 2020;Morling et al., 2014).Another study (Dubois et al., 2015) also found a positive relationship between social class and unethical behavior, but only when the behavior was self-beneficial.When unethical behavior was performed for the benefit of others, upper-class individuals were less likely to engage in unethical behavior than lowerclass individuals.
However, further research did not find any evidence for the claim that the rich are more unethical than the poor (Clerke et al., 2018;Gsottbauer et al., 2022;Jung et al., 2023;Trautmann et al., 2013), with upper-class individuals either being equally or even more ethical than lower-class individuals.
Given these inconclusive findings as to whether individuals from different social classes differ in their unethical behavior, we felt it important to provide another test of the idea that upper-class individuals are less ethical than lowerclass individuals.Study 1 of the present research was a direct replication of Piff et al. (2012) who found that individuals from upper-class backgrounds behave more unethically while driving than lower-class individuals.Concretely, individuals who drove upper-class cars were more likely to engage in unethical behavior by cutting off a pedestrian at a crosswalk than were individuals who drove lower-class cars.

Social class and prosocial behavior
Whereas upper-class individuals are assumed to prioritize their own interests, there is evidence that those lower in social class are more oriented to others in general and more sensitive to the needs of others in particular (Kraus et al., 2012;Piff et al., 2016).For example, those lower in social class pay more attention to other human beings (Dietze & Knowles, 2016), display superior performance in assessing other people's emotions (Kraus et al., 2010), are better at inferring emotional states of targets from images of their eyes (Dietze & Knowles, 2021), and report greater compassion for others than those higher in social class (Stellar et al., 2012).
Due to the heightened focus on the needs of other people, it has been hypothesized that lower-class individuals are particularly prosocial.In support for this hypothesis, Piff and colleagues (Piff et al., 2010) found that lower-class individuals are more charitable and generous toward others than are upper-class individuals.Subsequent research found that the negative relationship between social class and helping was evident not only in adults but also in children (Guinote et al., 2015;Miller et al., 2015).
Given these mixed findings, Study 2 of the present research aimed to shed more light on whether there are social class differences in prosociality.The design was similar to Study 1, but there was one crucial modification.As in Study 1, it was recorded whether a motorist let a pedestrian cross a street.Unlike in Study 1 where the pedestrian waited at a marked crosswalk, the pedestrian wanted to cross an unmarked pedestrian crossing (i.e., the motorists were not legally required to let the pedestrian cross the street).

The present research
Two studies provided a new test of the ideas that upper-class individuals behave more unethically and less prosocially than lower-class individuals.Our university requires no formal approval from an ethics committee if the research is in accordance with guidelines of the German Psychological Society.The data for both studies are openly accessible at: https://osf.io/7zuny/?view_only=fdb6fa741104427089885576fecde150.
The preregistration, including information on how we determined the sample size, can be found here: https://aspredicted.org/WPL_5PF.

Study 1
Study 1 is a direct replication of Piff et al. (2012) who found that upper-class individuals are more likely to drive through a marked crosswalk without letting a waiting pedestrian cross the street.

Method
Study 1 was carried out in a small city in Tyrol, Austria.A confederate was waiting at a marked pedestrian crosswalk.A coder who positioned him-or herself near the crosswalk in a car recorded whether an approaching vehicle would cut off the confederate.
As in Piff et al. (2012), observer ratings of the vehicle status were used to index the driver's social class.Based on the estimated vehicle value, make, age, and physical appearance, coders rated the perceived status of the vehicle using a scale from 1 (low status) to 5 (high status). 1 For example, the lowest score was given to a vehicle with current acquisition value less than 2,000 euros, low or medium status brand (e.g., Nissan, Ford), manufactured in the 90's or early 2000's, and the vehicle was in poor condition, whereas the highest value was given to a vehicle with current acquisition value more than 50,000 euros, high status brand (e.g., Mercedes, BMW, Ferrari), looks like it was bought in the last few years, and the vehicle was in great condition.
The procedure was also identical to Piff et al. (2012).Only vehicles approaching the crosswalk were coded if the confederate was the only pedestrian and there were no other vehicles in front of the vehicle.The confederates waited until a vehicle was about 15 meters away before entering the beginning of the crosswalk and facing the oncoming vehicle to signal their intent to cross.
Of 374 drivers, 72 (19.3%) failed to let the pedestrian cross the crosswalk.The coder also rated the perceived status of the vehicle (M = 2.61, SD = 0.85), as well as some control variables: the vehicle driver's sex (175 female, 199 male) and age (1 = 18-35 years (n = 117), 2 = 36-55 years (n = 200), 56 years and up (n = 57).In 265 cases the pedestrian was female, whereas in the remaining 109 cases the pedestrian was male.Piff et al. (2012) assessed the same control variables and used the same categories.

Results
To test whether upper-class drivers are more likely to cut off the waiting pedestrian, a binary logistic regression was performed on the data.Results showed that social class was not a significant predictor variable, b = .10,SE = .15,p = .527.Percentages of cars cutting off the pedestrian as a function of vehicle status are shown in Figure 1.This relationship remained almost unchanged when controlling for the driver's perceived age and sex and the pedestrian sex, b = .09,SE = .16,p = .580.None of the control variables received a significant regression weight, all ps > .347.

Discussion
Cutting a pedestrian at a marked crosswalk is against the Austrian Vehicle Act and most drivers allowed the pedestrian to cross the street accordingly.Nonetheless, quite a number of drivers failed to give way to the pedestrian, clearly behaving unethically.In contrast to our hypothesis, however, the vehicle status as an index of the social class of the driver was not significantly related to whether the driver drove through the crosswalk without yielding to the waiting pedestrian.
Please note that we had adequate power (80%) to detect a significant relationship between social class and unethical driving behavior, based on the effect size reported in Piff et al. (2012) study (Lakens, 2022).It is also of note that the number of participants was almost 2.5 times larger than the sample size in the original Piff et al. (2012) study (374 vs. 152), following suggestions by Simonsohn (2015) that the replication sample size should be 2.5 times the original sample size.
Piff and colleagues (2012) found that upper-class individuals were more likely to engage in a variety of unethical behaviors (not just unethical driving) than lower-class individuals (see also Boyd et al., 2015;Coughenour et al., 2020;Morling et al., 2014).However, not only the present study, but also other investigations did not find a significant positive association between social class and the tendency toward unethical behavior (Clerke et al., 2018;Gsottbauer et al., 2022;Jung et al., 2023;Trautmann et al., 2013).As it is, it is unclear whether the rich are actually more unethical than the poor.
Study 2 aimed to further investigate the hypothesis that there are differences in traffic behavior between individuals from higher and lower social classes, but with a focus on prosocial behavior rather than unethical behavior.

Study 2
Study 2 also examined the relationship between social class and driving behavior.Like in Study 1, it was examined whether a driver would let a waiting pedestrian cross a street.In contrast to Study 1, however, there was no marked crosswalk.Letting a pedestrian cross an unmarked pedestrian crossing was therefore not a legal requirement, but a courtesy of the motorist.It was hypothesized that individuals who drive upper-class cars would behave less prosocially (i.e., would be less likely to let the waiting pedestrian cross the unmarked pedestrian crossing) than individuals who drive lower-class cars.
As there were relatively few cars with high status in Study 1, Study 2 was carried out in a larger city (with the expectation that relatively more higher income individuals live in a large city).Because of financial constraints, the confederate rather than independent observers recorded whether the driver stopped the vehicle and rated the vehicle's status.Finally, as there were no significant effects of the control variables in Study 1, they were not assessed in Study 2.

Method
Study 2 was carried out in Innsbruck, the capital of Tyrol, Austria.Similar to Study 1, a female pedestrian was waiting to cross a street.In contrast to Study 1, there was no marked crossing, meaning that the drivers were not legally obliged to let the pedestrian cross the pedestrian crossing.The pedestrian recorded whether an approaching vehicle let her cross the street (of 365 drivers, 176 [48.2%] did so) and rated the perceived status of the vehicle (M = 3.53, SD = 0.84).

Results
A binary logistic regression was performed to test whether the social classes differ in whether they let the pedestrian cross the pedestrian crossing.As in Study 1, social class was not a significant predictor variable, b = .02,SE = .13,p = .903.Percentages of cars letting the pedestrian cross the street as a function of vehicle status are shown in Figure 2.

Discussion
As in Study 1, the social classes were remarkably similar in their driving behavior.Notably, whereas failing to let the pedestrian cross the marked crosswalk was illegal in Study 1, there was no legal obligation to let the pedestrian cross the unmarked pedestrian crossing in Study 2.Not surprisingly, then, percentages of cars letting the pedestrian cross the street were higher in Study 1 than in Study 2. Nevertheless, it is noteworthy that the level of prosocial traffic behavior was high: although there was no legal requirement, about half of the drivers let the pedestrian cross the pedestrian crossing.Importantly, however, individuals who drove lower-class cars were not more likely to do so than individuals who drove upper-class cars.
In contrast to Study 1 that was a direct replication of a previous study (Piff et al., 2012), Study 2 examined the hitherto unstudied association between social class and prosocial traffic behavior.Regarding other forms of prosocial behavior, some previous work (Guinote et al., 2015;Miller et al., 2015;Piff et al., 2010) found that upper-class individuals are less prosocial than lower-class individuals, whereas other studies either found that upper-class individuals are more prosocial than lower-class individuals (Andreoni et al., 2021;Korndörfer et al., 2015;Liu & Hao, 2017;Wang et al., 2021) or that both do not differ in their prosociality (Callan et al., 2017;Gheorghiu et al., 2021;Greitemeyer & Sagioglou, 2018;Rao et al., 2022;Stamos et al., 2020;Van Doesum et al., 2017).The present findings are in line with the latter findings: upperclass individuals did not differ from lower-class individuals in their prosociality while driving.Overall, as with the relationship between social class and unethical behavior, the evidence is equivocal as to whether upper-class individuals are less prosocial than lower-class individuals.

General discussion
Research to date has provided mixed evidence on whether the social classes differ in their unethical and prosocial behavior.Inconsistent with the claim that upper-class individuals are more unethical and less prosocial than lower-class individuals, the present studies showed that the propensity to engage in unethical and prosocial driving behavior was unrelated to the driver's social class.
A limitation of both studies is that -as Piff and colleagues (2012)-we used the vehicle status as an indicator of the driver's social class.The value of a car may misrepresent the driver's actual social class if, for example, lower class individuals spend most of their money on a visible good (such as a fancy car) due to conspicuous consumer motives.It is also noteworthy that the present studies remain mute in terms of cause and effect, because no experimental design was employed where participants are randomly assigned to either a high or low social class condition.
On the other hand, the design of both studies has notable strengths.Both unethical and prosocial behavior can be considered sensitive issues so that participants may be untruthful in their reporting (when self-report measures are used) or their behavior in a laboratory study differs from their real-life behavior because participants are aware that their behavior is being monitored.As people's natural driving behavior was recorded, social desirability biases and demand effects were avoided.
Furthermore, our studies are useful additions to previous lab-based experimental work that addressed whether a manipulation of the perceived social class has a causal impact on a person's unethical/prosocial behavior but where the generalization to the 'real world' is unknown.No single methodological approach can provide unequivocal answers to the effects of social class on unethical/prosocial behavior.Hence, it is useful to combine lab-based experimental studies where causality can be determined with naturalistic field studies that have high ecological validity.
Proponents of the idea that upper-class individuals are more unethical and less prosocial than lower-class individuals assume that these differences occur (in part) because the latter are less selfish and more focused on other people.That a lack of socio-economic resources promotes an increased sense of other-orientation has been documented in a number of studies (for a review, Piff et al., 2016).However, a recent large-scale study (Weick et al., 2022) failed to find a negative relationship between social class and communal attitudes.Further empirical work would be welcome that examines not only whether there are social class differences in unethical and prosocial behavior, but also to what extent upper-class individuals differ from lower-class individuals in terms of the assumed underlying mechanisms (i.e., their self-and other-centered social-cognitive tendencies).Please also note that individuals who perceive their social class as relatively low do not behave less aggressively compared to those who perceive their social class as relatively high.In fact, they tend to behave more aggressively (e.g., Chen et al., 2018;Greitemeyer & Sagioglou, 2017, 2019;Peng et al., 2021).
It is important to keep in mind that the present studies focused on interactions between strangers.While upper-class and lower-class drivers did not differ in their unethicality and their prosociality toward an anonymous pedestrian, upperclass individuals might be more unethical and less prosocial than lower-class individuals in the context of ongoing relationships.Future research may examine how the rich and the poor behave when they have repeated interactions within existing relationships (cf.Hooker & Algoe, 2022).
To conclude, research to date has not provided conclusive evidence as to whether upper-class individuals are more unethical and less prosocial than lower-class individuals, with some research showing a positive relationship between social class and unethical behavior and a negative relationship between social class and prosocial behavior, other research showing the exact opposite, and further studies that found that social class was not related to unethicality and prosociality.Consistent with the latter studies, the present two studies suggest that the rich and the poor are similar in their unethical and prosocial driving behavior.

Figure 1 .
Figure 1.Percentage of cars cutting off the confederate at the marked crosswalk as a function of the vehicle status (study 1).

Figure 2 .
Figure 2. Percentage of cars letting the confederate cross the unmarked pedestrian crossing as a function of the vehicle status (Study 2).