BENEATH THE SURFACE OF MORAL HYPOCRISY
ABSTRACT
Though many studies have been published on the topic, moral hypocrisy remains an interesting but inconclusively demonstrated behavioral hypothesis that may prove beneficial in the study of ethical decision-making. Using randomized and biased computer programs as ethical assignment interventions, results are obtained that unambiguously demonstrate moral hypocrisy in test subjects, while shedding greater light on the underlying psychology of the subjects who engage in this behavioral pattern. A broader range of subjects expands the known generalizability of results obtained so far, while eliminating the chance that earlier results were an artifact of subject immaturity. Gender, socioeconomic background, and age failed to reach significance in explaining study results.
We live in a world in which unethical behavioral choices are displayed with such media prominence that one almost questions the assumption of an innate morality in our species. Open the newspaper or turn on the television news and you cannot miss stories about athletes caught up in illegal doping to gain competitive advantage, world prominence, and wealth; business leaders trading on insider information to the detriment of their shareholders; and governmental leaders gaining personal benefit through the allocation of governmental contracts. In short, we live in a world in which understanding moral behavior and the underlying motivation to overcome personal temptations in making ethical choices is critical.
Yet, it has been pointed out that moral standards do not function as fixed internal regulators of conduct. There are many psychological processes by which self-regulatory mechanisms can be disengaged (Bandura, 1986).
As Bandura has pointed out, these mechanisms of moral disengagement are not limited to extreme behaviors, such as those associated with terrorism or military interrogation, but rather, these mechanisms routinely operate in normal everyday behavioral choices. We witness moral disengagement anytime individuals perform activities which further their own interests but have negative consequences for others (Bandura, 2004). Among these mechanisms are advantageous comparison, euphemistic language, and displacement of responsibility. Studies show that people behave more cruelly when assault actions are given a sanitized label than when they are called aggression (Diener, Dineen, Endersen, Beaman, & Fraser, 1975). Similarly, when people are allowed to displace responsibility for their actions, they will behave in ways they normally would find abhorrent (Diener, 1977; Milgram, 1974).
More recently, moral hypocrisy has been used to explain unethical decision-making in everyday life (Batson, Kobrynowicz, Dinnerstein, Kampf, & Wilson, 1997; Batson, Thompson, & Chen, 2002; Batson, Thompson, Seuferling, Whitney, & Strongman, 1999). Moral hypocrisy exists when individuals seek to create the impression of being moral while not actually having to bear the costs of being so. On the surface, this has an intuitive appeal in that individuals avoid negative feedback to their self-image, while not having to suffer difficult consequences they would wish to avoid.
Evidence for Moral Hypocrisy
Through a series of eight studies, Batson and his colleagues introduced the concept of moral hypocrisy: the desire to appear moral without actually bearing the actual cost of being moral (Batson, Kobrynowicz, Dinnerstein, Kampf, & Wilson, 1997; Batson, Thompson, & Chen, 2002; Batson, Thompson, Seuferling, Whitney, & Strongman, 1999). In each of the experiments, the subject was faced with a simple task assignment decision, in which one individual (either the actual subject making the assignment decision, or a fictitious paired subject, would be assigned to a positive task, while the other would be assigned to a neutral task. In the positive task condition, the subject would have the opportunity to win a small cash prize (usually $30 in a raffle); in the neutral condition, the subject would engage in a “rather dull and boring†task and would not have the chance to win a prize.
In the first three experiments (Batson, et. al., 1997), the studies focused on exploring the relationship between moral reasoning and moral decision-making. The pattern of results led Batson and colleagues to speculate about a behavioral mechanism which they called moral hypocrisy. In the last four experiments (studies 2 and 3 from Batson, et. al., 1999; and the studies reported in Batson, et. al, 2002), the emphasis was on understanding variables which moderate moral hypocrisy, such as self awareness and the salience of a moral standard. However, Study 1 (Batson, et. al, 1999) and Study 2 (Batson, et. al, 2002) are the only studies that attempt to provide direct evidence for the existence of moral hypocrisy.
Ambiguity in Batson’s Results
The impetus for the current study largely rests on the results of (Batson et al., 1999)’s Study 1 and (Batson et al., 2002) Study 2. The first of these studies tested moral decision making by allowing the subject to assign herself to either a positive study condition (potential to win $30) or to the neutral condition (no money), and another subject (fictitious) to the remaining condition. The subject was told that most people consider the fairest way to assign tasks to be flipping a coin. In this way, Batson, et. al., made the moral standard of fairness salient. In this study, Batson et. al. provided subjects with a labeled coin (one side read “SELF to POS†the other side read “OTHER to POSâ€). Of the forty subjects in the experiment, 28 chose to flip the coin. Of these, four assigned the other individual to the positive study condition, while the remaining 24 assigned themselves to the positive study condition.
Following task assignment, subjects were asked to rate the morality of their task assignment, on a 9-point scale (1 being lowest, 9 being highest in moral behavior response). The mean response for the individuals who flipped the coin was 7.64. This high mean score for the morality of the assignment decision, in spite of the fact that task assignment significantly deviated (in favor of the study subject) from what chance would allow, was interpreted as evidence for moral hypocrisy behavior on the part of some subjects.
Two issues arise with this result. The first is that there is no way to examine the decisions of particular individuals because there is no way to know the result of the coin flip in any specific instance. Second, the mean score of 7.64 is not as unambiguous as it appears on the surface. With a coin flip, we expect each side to come up 50% of the time. Thus, of 28 individuals, we expect 14 of them (on average) to flip “SELF to POSâ€, in which case, the individual can assign themselves to the positive task and feel quite moral in the approach taken. Additionally, four individuals assigned the other subject to the positive condition, and we can expect them to feel quite moral about their decision as well. This leaves (on average) ten subjects who may have made a decision in favor of themselves when the coin came up “OTHER to POSâ€. If the 18 individuals who either assigned the other subject to the positive condition, or had “SELF to POS†each rated the morality of their own decision as a “9â€, then the remaining ten individuals would only need to average 5 out of 9 to create a mean score of 7.57, or a very close approximation of the 7.64 result. If the subjects who assigned themselves to the positive condition in spite of the result experienced such a drop off in mean morality of decision scores, then moral hypocrisy would be a less appropriate conclusion. If “SELF to POS†came up more than 50% of the time, then the strength of the moral hypocrisy assertion weakens.
Recognizing this, Batson, et al (2002) explore this question (Study 2) of whether earlier moral hypocrisy results are consistent with some subjects knowingly making the assignment in their own favor, regardless of the coin flip. In order to address this possibility, Batson and colleagues combined surreptitious observation of participants with a coin that was modified to make the outcome highly visible to the observer (each side of the coin had a different colored label.) Of 44 participants, 16 subjects flipped the coin and assigned themselves to the positive condition after either having the coin come up “other to positive†or “otherwise fiddle[d] the coin flip to assign self to positive.†On a 1 to 9 scale, the mean self-reported morality was 5.56 with a standard deviation of 2.37 (almost twice that of any other behavioral category outcome.
The high degree of variance in these results indicates that at least a couple of subjects rated their morality unusually high, though clearly others rated their morality very low. The difference between this group and the subjects who assigned themselves to the positive condition without making any pretense of using ethical assignment proved statistically significant, though two concerns remain. First, “moral hypocrisy subjects†consisted of three different behavioral patterns. A comparison of morality rating between these three behaviors did not reach significance; though one must question whether this is a function of similar morality rating or rather is caused by insufficient statistical power. Given the large standard deviation, it would be difficult to exclude statistical power as an explanation for the non-significant result (Hoenig & Heisey, 2001; Lenth, 2001). Thus, it is impossible to determine whether high self-rated morality was associated with one behavioral type more than others. Second is the qualitative analysis of subject response patterns. though subjects falling under the “moral hypocrisy†label considered themselves more ethical than those that made no attempt to use the coin for task assignment, the mean response is only slightly above the scale midpoint. It seems reasonable to interpret that these subjects, taken as a whole, rated their behavior as neither highly ethical nor highly unethical. We would expect subjects who at least made the ethical attempt to view their own behavior more positively than those who ignored the ethical assignment procedure altogether. Should these results be interpreted as indicative that subjects felt moral without bearing the costs of being so (the definition of moral hypocrisy)? We view this as questionable without further evidence.
Prefrontal Cortex Maturity
There is one other reason to question the general applicability of findings to the broader population at large. The Batson, et. al. (1997, 1999, 2002) studies have used introductory psychology students, presumably 18-19 years old, on average. A recent longitudinal study of brain development using neuro-imaging to map the development of the brain in a sample of children from the age of 13 to 21 (images taken every two years) found that the prefrontal cortex matured last, reaching full maturity at approximately the age of 20 years (Gogtay et al., 2004). It has been observed that disorders associated with dysfunction in the prefrontal cortex and corticobasal ganglia-thalamocortical circuits result in behavioral changes such as impulsivity and apathy (DeLong, 2000). Additionally, subjects with lesions in this area experience severely compromised ability to reason and make rational decisions (Kandel, 2000).
Given that the ability to resist the impulse of personal temptation is governed by this late developing region of the brain, it cannot be ruled out that Batson, et al’s (2002) findings are a result of subject cognitive immaturity. The only way to rule this possibility out is to test Batson’s method on a group of subjects sufficiently beyond the developmental phase of the brain. Subjects 21 and older should provide an adequate test.
The two studies reported in this article were designed to test this result in a stronger manner with the hope of either confirming or disconfirming the moral hypocrisy hypothesis. More importantly, if the hypothesis were to be confirmed, the second experiment was designed to isolate the decision behavior of specific subjects so that we could further explore the psychological mechanism underpinning this behavior.
Each of the two studies replicates the Batson, et. al., 1999 Study 1, with an important difference: instead of a labeled coin, we allowed subjects to use a randomized computer program for assignment. In each experiment, subjects were shown how to use the computer program for task assignment, and walked through the process a few times. (This was largely a device to allow the subject to see each type of assignment pairing (Self to Pos/Other to Neutral; and Self to Neutral/Other to Pos) come up randomly.
In the first experiment, subjects were then seated at a work station with a program that truly was random in assignment. We did this to confirm that the random computer assignment would replicate the results from Batson, et. al.’s 1999 Study 1 with the labeled coin. It also provided a baseline for the comparison of the results from the second experiment.
In the second experiment, we replicate the first, except that the computer program at the subject’s work-station is biased to only respond “OTHER to POS/SELF to NEUTRALâ€. In this way, we introduced an intervention that the psyche might reasonably perceive as dubious in spite of contrary experience in training, and yet in which we knew the assignment outcome each individual received. The results of each experiment are reported and discussed in the remainder of this article.
Study 1: Randomized Computer Assignment
In Study 1, we replicate Batson, et. al.’s 1999 Study 1 with task assignment involving the use of a labeled coin. Though the primary purpose of this replication was to ensure that our random computer intervention did not unduly bias subject behavior, and thus interfere with the result, we also sought to broaden the generalizability of moral hypocrisy as a phenomenon. Thus, we used a larger sample that included a broader range of individuals, measured in terms of age and socioeconomic upbringing, while duplicating the use of mixed gender.
Method
Participants. We used 60 subjects in the first study. All participants were undergraduates at the University of South Florida, and received five points extra credit towards their course grade for participating. Of the 60 participants, 16 were male and 44 were female. The age of participants ranged from 21 to 45 (males M=25.12, females M=24.83).
Procedure. Following Batson, et al., participation was by individual appointment. Upon arrival, participants were greeted by the experimenter and escorted into the research room. The experimenter explained that there were 2 participants in the study and that it was important that they not meet or see one another. (No other participant was actually involved.) The subject was then left alone with written instructions that indicated the experiment was an attempt to understand the effects of task assignments on feelings and responses of individuals:
We are currently examining the effects of task assignments on feelings and responses. Of course, any work-task may have positive, neutral, or negative consequences. In this study, we focus mainly on the effects of positive and neutral consequences. You and another person—who you will not know and will not meet—will be assigned to two different tasks. One of the work-tasks has positive consequences and the other has neutral consequences. You both will work on your tasks independently; you will not meet each other, or learn of each other’s identity.
Task consequences. In order to ensure a real and replicable result, it was important to have real task consequences. In the introduction, following the overview to the study, the following statements appeared for the subjects to read:
Work-Task 1: If you are assigned to Work-Task 1, the positive outcome task, you will be given a short series of five questions to answer. For each correct answer, you will be given 1 raffle ticket; the prize in the raffle is $100. Only participants in this study are eligible for the raffle so if you answer all the questions correctly, you stand a relatively good chance of winning.
Work-Task 2: If you are assigned the “neutral†consequences work-task (Work-Task 2), you will simply be asked to answer five short questions and be given feedback about whether your answers are right or wrong. You will not be given any raffle tickets. Also, you should know that most participants find Work-Task 2 rather dull and boring. If you are assigned this task, we hope you don’t find it too bad.
The chance to assign tasks using randomized computer assignment. Following the work-task descriptions, the instructions ask the subject to signal the experimenter for an introduction to the computer program available for task assignment. At this time, the subject was shown several times how to use the random assignment program on the computer, the program was reset, and the subject was given a chance to demonstrate their understanding. This also provided an opportunity for the subject to see that the program was random.
The subject was then left to read a section of the introduction entitled, “Procedures†in which the subject finds out that they are the member of their pairing that will be allowed to assign the work-tasks. They were also given the following statement about fairness in assignment of tasks:
The decision is entirely up to you. You can assign yourself and the other person any way you choose. The other person does not know and will not know that you are assigning the tasks; he or she will be told that the task assignment was made entirely by chance.
Most of the people that have participated in this study in the past felt that giving both you and the other person an equal chance at being assigned the positive task (Work-Task 1) is the fairest approach. For example, you might use a computer to make random assignments. For this reason, we have provided you with a computer that will assist you. The two of you will never meet; your anonymity is assured.
The first sentence in the second paragraph makes the moral standard of fairness salient to the subject. The statement connects fairness to the use of the random assignment program that the subject has been introduced to, and given ample opportunity to witness different assignment decisions coming up at random during separate practice and training trials.
Subjects were then allowed to progress forward in the task assignment materials to the task assignment decision. On this form, the first line reads, “Work-Task 1—the positive consequence work-task—will be done by ________.†The second line reads, “Work-Task 2—the neutral consequences work-task—will be done by _______.†The participant is asked to fill in “Self†on one of the lines, and “Other†on the remaining line. The task assignment form was then turned in to the experimenter. This task assignment is the major dependent variable in this study.
Perceptions of the morality of one’s decision. Following the assignment of tasks, the subject was left with a short series of questions related to their task assignment decision, while ostensibly the experimenter left to prepare the separate tasks and notify the other “experimenter†(the one that would be leading the other subject through parallel procedures, if the subject existed) which task they would be carrying out. Included were questions asking about how important the subject’s own welfare was to the decision, the welfare of the other subject, and the importance of being fair.
Additionally, the subject was asked whether or not they chose to use the computer in assigning the tasks, whether they thought the computer was fair, and whether they believed themselves to have been fair in assigning the tasks. Answers to this final question ranged from “In a very fair way†at one end of the scale, to “there is no fair or unfair way to assign tasks†in the middle, to “In a very unfair way†at the opposite end of the morality scale.
Debriefing. When participants finished this final questionnaire, the experimenter returned and debriefed the subject fully, which included revealing the true nature of the experiment. Participants were not asked to answer any questions (as either work-task would have required), subjects were allowed to complete five raffle tickets, as would have been available had they participated successfully in the positive consequences work-task. Following the completion of data collection, the raffle was held and the $100 awarded to the winner.
Results
Task assignment. Of the 60 subjects participating in study 1, 24 reported using the computer for task assignment, and 36 did not. Of the 36 who did not use the random computer program, 30 (.833) assigned themselves to the positive task, and 6 (.167) assigned the other subject to the positive task. Of the 24 who reported using the random computer program, 19 (.791) assigned themselves to the positive task, and only 5 (.208) assigned the other subject to the positive task. This pattern differs significantly from random (Χ2 = 8.167, significance p<.004).
Predicting who would use the computer for task assignment. One question before us is whether any of the control variables successfully predicted which subjects would choose to use the random computer program for task assignment. Along with age, gender, and socioeconomic upbringing, we asked the subjects five questions following their task assignment. They were:
How important was your own welfare to your task assignment?
How important was the welfare of the other subject in your task assignment?
How important was being fair to both participants to your task assignment?
How important was it to treat both participants equally in deciding task assignments?
How important was the concern for giving both subjects equal opportunity to be assigned to each task?
Each question was answered using a five item scale that ranged from “Very Important†(= 5) to “Very Unimportant†(= 1). The midpoint response was, “Neither Important nor Unimportant†(= 3).
Of the demographic variables (i.e., age , gender and socioeconomic upbringing) only gender (2-tailed t = -2.605) (p = .012) came close to reaching a level of significance in explaining variance in the decision to use the computer program. Of the five response question variables, only the answer to question 4 proved significant (t = 4.829) (p = .000). Thus, we conclude that none of the demographic variable had explanatory usefulness for this question, and only the subject’s interest in treating both subjects fairly predicted the use of the computer for task assignment.
Predicting who would assign the other subject to the positive condition. We used the same set of variables in an attempt to identify any predictable bias in which subjects would choose to assign the positive task to the other (unknown) subject. None of the demographic variables approached significance, and only the answer to question 2 (welfare of the other subject) came close to significance (2-tailed t = 1.634) (p = .109).
Perceived morality of the assignment. When subjects were asked to rate the morality of their task assignment on a 5-point scale (1 = in a very fair way; 5 = in a very unfair way; 3 = there is no fair or unfair way to assign tasks), we obtain confirmation of Batson, et. al.’s (1999) moral hypocrisy finding. When we compared the subjects who had selected themselves without using the computer (M = 3.55), to those subjects who selected themselves with the computer but reported that they perceived the computer to be unfair (M = 1.40), the difference is statistically significant (2-tailed t = 5.569) (p = .000). Thus, subjects who used the computer but were able to convince themselves the computer was unfair—in spite of contrary evidence through initial training—were also able to perceive their own choice in a much more moral light: Moral hypocrisy.
We gain further insight into how strong this phenomenon is when we compare the subjects who chose themselves despite using the computer and reported perceiving it to be unfair (M = 1.40), with the group of subjects who chose the other subject with or without using the computer (M = 2.09)—arguably the most noble assignment decision. In this case, there is no statistically significant difference between the two groups (2-tailed t = 1.377) (p = .190). Yet, also notice that the group engaging moral hypocrisy in their self-evaluation actually have a (non-significant) favorable shift in the sample mean with very small sample numbers (N = 11 and 5 respectively), thus with a larger sample, it is possible this effect would become significant. This should receive further study.
Discussion. This first study replicated and expanded upon Batson, et. al.’s Study 1 (1999), in which they used a labeled coin as a tool to facilitate random assignment of tasks by participants. The purpose served by this replication, first and foremost, was to establish that using the computer program, in place of the labeled coin, and would not bias the study results. This was necessary, as in the second study, we use a manipulation of this intervention to further explore the phenomenon of moral hypocrisy. We conclude that the replication for this purpose was successful.
The computer intervention also expands support for the Batson, et. al. (1999—Study 1) conclusion, in that in this study, it was reasonable to measure perceived fairness of the computer assignment. This allowed us to identify subjects who a) reportedly used the computer for task assignment; b) assigned themselves to the positive task; and c) reported the computer assignment to be unfair. It seems improbable that subjects would exhibit this response pattern when the computer had selected them for the positive task.
We also expanded upon the original study by using a much larger sample comprised of a more diverse subject pool. In our study, we expanded the age range (21 to 45; M = 26.65) and included subjects from socioeconomic upbringings that ranged from (self-reported) lower-middle class, to upper class (54.3% were closely grouped around lower middle to middle class with the remaining 45.7% distributed as follows: 8.4% self-reported themselves as poor; 8.4% reported themselves as wealthy, and 28.9% reported themselves in a range from middle class to upper class.) We found no statistically significant effects related to age, gender, or socioeconomic upbringing as predictors of task assignment or predilection for engaging in moral hypocrisy reasoning. Though this deserves further study in the future, it considerably broadens the generality of moral hypocrisy as a psychological influence on ethical decision-making.
Study 2: Biased Computer Condition
The problem with the above study is similar to the problem with the 1999 Batson, et. al. Study 1 and to some degree, the 2002 Batson, et. al, study 2: though we can conclude moral hypocrisy appears to be happening, we cannot explore the results at the individual level. The random computer program, as with the labeled coin, creates an expectation that half of the subjects who use the computer (or the coin) will be told to select themselves. We would expect those subjects to not only select themselves, but also feel quite good about doing so. An advantage of our randomized computer assignment over the labeled coin is that we were able to engage perception of fairness of the instrument. This led to a stronger conclusion of moral hypocrisy.
It would be helpful to be able to analyze subjects while knowing exactly what assignment came up on the coin (or computer). Thus, in Study 2, we replicate the method of Study 1, with the exception that after training the subject to use the randomized computer program, we placed the subject at a work station with an identical looking program interface, but one that was biased to only select the other subject. (Note, the program could not be used repeatedly without the experimenter resetting it, thus “fiddling†with the result was not possible.) In this way, the study allows us to explore moral hypocrisy both quantitatively and qualitatively by virtue of the fact we know that each subject who uses the computer is being told to assign the other subject to the positive study condition.
Method
Participants. Participants for Study 2 were 20 undergraduate students (11 women, 9 men), ranging in age from 19 to 40 (M = 26.30). Thus, the sample closely parallels the sample collected in Study 1. As with the first study, subjects received 5 points extra credit for participation.
Procedure. The procedure for Study 2 was exactly the same as in Study 1, except that the computer program at the subject’s work station was biased to always select the other subject for the positive task condition.
Results
Task assignment. Of the 20 subjects in Study 2, 15 subjects (.75) assigned themselves to the positive task condition, while 5 (.25) assigned the other subject to the positive task. Only 7 subjects (.35) reported using the computer to assign tasks, while 13 subjects (.65) made assignments without using the program. Of the 7 subjects who used the computer, only 2 (.29) assigned the other subject to the positive condition while 5 (.71) assigned themselves to the positive condition in spite of receiving a computer response indicating the other subject should be assigned to the positive task. Of the 13 subjects who made task assignments without the program, 10 (.77) assigned the positive task to themselves, and 3 (.23) assigned the positive task to the other subject.
Perceived morality of the assignment. When we examined the self-rated morality of the subject’s task assignment decision (using the same 5-point scale as in Study 1), we obtain similar results: when we compare the group that assigned the positive task to themselves without using the program (M = 3.80), to the moral hypocrisy group (the group that reported using the computer but assigned themselves to the positive task anyway) (M = 1.60), we obtain a significant difference (2-tailed t = 4.407) (p = .001), with the moral hypocrisy subjects rating themselves significantly more moral than those subjects that assigned themselves the positive task without the computer.
When we compared this group (used computer, assigned self, perceived computer unfair) (M = 1.60) with the subjects who assigned the positive task to the other subject (M = 2.20), the difference is not significant (2-tailed t = 1.095) (p = .305). However, notice that again, the moral hypocrisy engaging group posts a (non-significant) higher mean score than the group that assigned the other individual to the positive task.
Taking a qualitative look at the result. In this study, due to the nature of the intervention, we know exactly what response each subject who used the computer received for task assignment. This allows us to take an individual level look at the dynamics underpinning moral hypocrisy (see table 1).
Table 1: Qualitative Look at Self-Reported Morality of Task Assignment
Very Fair Mostly Fair Not Fair or Unfair Mostly Unfair Very Unfair
Assigned self to positive task, used computer, perceived computer unfair. 3 1 1
Assigned other to positive task 1 1 2
Assigned self to positive task without computer 1 2 5 2
As can be seen in Table 1, 3 of the 5 subjects who demonstrated “moral hypocrisy†actually evaluated their task assignment decision as “Very Fairâ€, even though the computer program had told them to assign the other subject to the positive task. Of the remaining subjects in this group, 1 felt s/he had been “Mostly Fair†and 1 felt s/he had been neither fair nor unfair. None of the subjects who exhibited moral hypocrisy believed their decision to be either “Mostly Unfair†or “Very Unfairâ€.
Subjects who assigned the other subject to the positive task also did not report feeling that their task assignment was “Mostly Unfair†or “Very Unfairâ€, but 2 of the 4 reported themselves as “Not Fair, Not Unfairâ€, which is somewhat surprising, considering they had sacrificed a chance at the $100 raffle. We cannot explain this result.
In contrast with the other two groups, subjects who assigned themselves to the positive task without using the computer for task assignment generally acknowledged that their decision was either “Mostly Unfair†or “Very Unfairâ€, though 1 subject did report him/herself as “Mostly Fair†in this assignment choice. This is strong evidence about the power of moral hypocrisy to influence the self-perceptions of those individuals that are prone to it.
Discussion.
In this second study, we found a pattern of results very similar to both our first study, and to Batson, et. al.’s (1999) Study 1 and stronger support for moral hypocrisy than Batson, et. al.’s Study 2. By manipulating the computer program to select only the other subject for the positive task, we were able to isolate subjects who had specifically gone against the computer selection while using the program. This resulted in a pool of subjects who would either a) recognize that they had been unfair to the other subject, or b) engage in moral hypocrisy. In this experiment, the subjects engaged in explicit moral hypocrisy in order to deem their own actions as fair and just by denying the fairness of the computer assignment, in direct contrast to their training experience with the program.
General Discussion
In an attempt to further our understanding of moral hypocrisy in individuals, we developed two experiments. The first was designed to replicate Batson, et. al.’s 1999 Study 1, to extend the range of generalizability in the population, and to confirm that our computer program intervention would not substantially bias the results obtained. With a subject pool of sixty individuals, we reaffirmed that gender does not appear to play an explanatory role in moral hypocrisy. This is consistent with findings from Batson, et. al. 1999, and Batson, et. al., 2002. Additionally, our results indicate that socioeconomic upbringing and age also do not significantly explain moral hypocrisy results. It would perhaps be interesting in future work to explore whether or not socio-cultural differences, or educational differences play influence the outcomes. In our studies, as well as those by Batson and colleagues, all of the subjects have been working towards college degrees, and thus are more educated than average. It might be that subjects of different educational backgrounds would exhibit more or less of this behavior.
We believe that the results of Study 1 sufficiently replicate and strengthen prior indications of moral hypocrisy in simple decision-making circumstances (Batson et al., 1997; Batson et al., 1999). Of the 60 subjects in the experiment, 36 chose to make assignments without the aid of the computer, while 24 reported using the computer. Similar to Batson and colleagues, of the 36 who did not use random assignment, 30 assigned themselves to the positive task, while only 6 assigned the other subject to the positive outcome task. Of the 24 subjects who used the computer, 19 of them still assigned themselves to the positive condition, while only 5 assigned the other individual to the positive condition. Of the 24 subjects who used the computer for assignment, 11 reported believing the computer to be unfair, in direct contradiction to their experience with the computer in training. The subjects who reported believing the computer unfair, also rated the morality of their assignment decision significantly higher than subjects who did not use the computer.
Based upon these results, it seems clear that using the computer program as an intervention in place of the labeled coin from Batson, et. al.’s 1999 Study 1, did not interfere with the experimental result. Thus, we concluded that it was also reasonable to proceed with the follow-up experiment in which we manipulated the computer program condition such that all subjects would obtain the result “OTHER to POS/ SELF to NEUTRAL). (Remember, all subjects went through the same training regardless of experimental condition.)
In this second experiment, we obtained a similar pattern of results. Of the 20 subjects, 13 chose to make assignments without the computer, while 7 reported using the computer. Of these 7 subjects, 5 assigned themselves to the positive outcome condition, in direct contradiction to the computer’s output. All 5 reported believing the computer to be unfair (again, in direct contradiction to their training experience). These subjects also rated the morality of their task assignment decision significantly higher than subjects who did not use the computer program.
There are three important outcomes from this second experiment. First, it provides unambiguous support for the moral hypocrisy hypothesis proposed by Batson and colleagues. Second, it allows us to start understanding this behavior proportionally within populations (on a small, speculative scale), given that 5 of 20 subjects (.25) clearly exhibited this behavior. On the positive side, it was only 25% of the subjects; on the negative side, it is concerning that 25% of these subjects were able to bias their perception of the world to this degree over such a small prize.
The third important outcome is it allows us to more clearly examine the psychological process taking place within the individual that is subject to moral hypocrisy patterns of reasoning. This manipulation allowed us to explore the more interesting issue of how subjects rationalize their decision patterns.
Several mechanisms have been proposed to explain moral disengagement, or behavior-standard discrepancy. Among the simplest is ordinary self-deception (Mele, 1987) in which the individual simply uses situational ambiguity as a basis for ex post interpretation of results in one’s own favor. The computer program intervention allowed for a rather extreme form of this behavior, in that the subject was given a chance to “prove to themselves†that the program randomly produced both sets of results, during their initial training. Thus, subjects should have believed in the randomness of the assignment decision, yet in both cases—when the computer assignment really was random, and when it was not random—similar proportions of subjects still were able to convince themselves that the computer had not been fair. This was then used to rationalize the assignment they preferred: themselves to the positive outcome condition.
The result could also be loosely interpreted as “displacing and diffusing responsibility†(Bandura, 1991), whereby the subject places responsibility for the outcome on the “unfair computer†and feels justified in correcting this perceived injustice, thus feeling somewhat self-righteous in the process, though this would be an admitted stretch.
Regardless of mechanism, and this is certainly an area that deserves further study, it appears that some proportion of the general population is vulnerable to moral hypocrisy when their self-interest is sufficiently engaged. This would be a particularly troublesome quality to find in leaders with significant decision-making power, such as politicians and organizational leaders. It would be beneficial if future studies pursue questions related to promotion and leadership acquisition. Specifically, are individuals who obtain leadership roles and positions of power, more/less/or equally susceptible to moral hypocrisy. These, for the time being, will remain questions for future study.