SWITCH ON THE RELATION BETWEEN COMMITMENT AND ESCALATION OF UNETHICAL BEHAVIOR
Master thesis, MScBA, specialization Human Resource Management University of Groningen, Faculty of Management and Organization
June 20, 2012
RICHARD ANDOH Studentnumber: 2057182
Wibenaheerd 56 9726 PC Groningen tel.: +31 50 363 8900 e-mail: r.s.andoh@student.rug.nl
Supervisor L.B. Mulder
Acknowledgment: Helpful feedback on earlier drafts of this thesis was given by students in my thesis group. More, I want to thank Laetitia Mulder for her insightful theoretical and practical advice and patience during the feedback rounds that were held.
CHANGING THE DOMAINS OF (IM)MORALITY: THE EFFECT OF A DOMAIN SWITCH ON THE RELATION BETWEEN COMMITMENT AND
ESCALATION OF UNETHICAL BEHAVIOR
ABSTRACT
This study attempts to uncover the factors that contribute to escalation of unethical behavior and examines how a domain switch influences the process. I combine research findings regarding the influence of commitment on escalation and available studies addressing moral compensation and moral licensing. Firstly, I hypothesizes that the degree of commitment to prior unethical behavior influences future unethical escalation. Secondly, I hypothesize that a domain switch moderates the relation between commitment and escalation of unethical behavior. The hypotheses were tested in an experimental computer study in which people played a game involving unethical options, commitment and domain switch were manipulated. Initially, the results showed no moderating effect of a domain switch. After further investigation a moderating effect was found, but only directly after the domain switch had occurred. Implications for the topics on: escalations of unethical behavior, commitment and for the concept of morality across domains are discussed.
Keywords: commitment; domain switch; escalation; unethical behavior;
CHANGING THE DOMAINS OF (IM)MORALITY: THE EFFECT OF A DOMAIN SWITCH ON THE RELATION BETWEEN COMMITMENT AND
ESCALATION OF UNETHICAL BEHAVIOR
Unethical behavior1 in organizations is a phenomenon that has received an increased amount of attention over the past few years. Unethical acts have shown to cause detrimental effects on individuals and society as a whole (Gino, Schweitzer, Mead
& Ariely, 2011). The illustrious case of the American Energy company Enron -filing the largest bankruptcy in the US history- initiated a domino-effect with major financial and social consequences (Ratanen, 2007). Historically, many similar incidents have occurred (e.g. WorldCom, Arthur Andersen and Tycon); in retrospect forcing people to ponder upon how this could have happened (Ratanen, 2007). The damaging and long term consequences of unethical behavior urge organizations and individuals to understand the conditions under which these behaviors manifest (Aquino, Freeman, Reed II, Lim &
Phelps, 2003,). Cases such as Enron and WorldCom provoke researchers and managers to question why these organizations and the people involved made unethical conduct prosper to such an extent. New insights are necessary in order to gain a greater understanding of unethical behavior in general, and more importantly; to prevent this from occurring in the future.
According to Jones (1991) behavior can be defined as unethical when it is harmful to others and is either legally or morally unaccepted. Based upon this definition, unethical behavior exists on a wide spectrum (e.g. cheating, lying, violation of rules and
1 Unethical and immoral are used interchangeably based upon the literature to which the reference is made, but both words entail the definition as formulated by Jones (1991)
other forms of unaccepted behavior). Substantial effort has been put into understanding why and how people behave unethically (e.g. Tenbrunsel & Smith-Cowe, 2008, Weaver
& Reynolds, 2006). Multiple studies and business cases have shown compelling evidence for the notion that unethical behaviour has a tendency to escalate (e.g. Enron, Martens, Kosloff & Jackson, 2010; Milgram, 1974). This thought is supported in the experiment by Martens et al. (2010); in which the results indicated that people who were given the task of killing five bugs, killed more bugs in their subsequent test than the people who killed one bug in their initial test.
Questions regarding “why” and “how” unethical behavior escalates, have received an increasing amount of attention over the past few years. Examination of literature shows an important finding: a single unethical act can influence the perception of the self to such an extent that behaving unethically can become the norm (Cialdini, 2007;
Martens et al., 2010). More, the degree to which escalation of unethical behavior becomes a norm to a person; is influenced by the situation and the individual characteristics (Wiltermuth, 2011). Monin & Jordan (2009) go on by asserting that there is a dynamic between the situation and the individual. More specifically, a person continuously uses behavior in prior situations as a source of feedback regarding: the right decision, attitudes values and beliefs; and uses that information to determine future behavior (Cialdini, 2007). While multiple researches have attempted explain why and how escalation might occur, there is little empirical support for these arguments. The goal of this research therefore is to contribute to a greater theoretical and empirical understanding of the dynamic conditions under which unethical escalation happens,
because the exact factors contributing to this phenomenon remain largely unanswered (Merrit et al., 2010).
The present research therefore aims to provide new insights in field of unethical conduct by investigating the conditions under which escalation of unethical behaviour emerges and continues. The reasoning for this is outlined below.
THEORETICAL FRAMEWORK
The effect of commitment on escalation
Multiple studies show that commitment is one of the main predictors of future behavioural responses (Burke & Reitzes, 1991; Staw, 1981). More specific, Staw (1981) argues that commitment ties an individual to a sequence of behaviours. In addition, Burke and Reitzes (1991) assert that commitment is the feeling that connects an individual to a set of self-meanings, consequently producing a sequence of activities. Thus people have the tendency to behave in accordance with their commitment to a certain activity. Hence, an employee who has invested a significant amount of time and effort into the act of stealing from the inventory will be more likely to continue in the future due to a strong link of the behaviour to the self image of this person.
Literature suggests two apparent reasons for why and how commitment “ties” a person and determines future behaviour. Firstly, people view that acting in a way that deviates from their previous behavior indicates inconsistency; which is naturally considered an undesirable personality trait (Merrit et al., 2010). Secondly, people are unwilling to change a course of action after serious investment in that course, because it forces a person to admit that a poor decision has been made (Festinger, 1957). This
unwillingness to change is referred to as the sunk cost effect or “a greater tendency to continue an endeavour once an investment in money, effort, or time has been made” (Arkes & Blumer, 1985, p. 2). A practical example is the resistance to terminate a business project which has required significant investment, because stopping the project after a great deal of money has been spent would imply a loss of the invested amount of money. In a more recent article Lunenburg (2010) uses the expression “throwing good money after bad” as a way to describe this process, which is defined as a continuation of a course set by a prior decision in “order to justify the original decision” (p. 1). This sense of being locked “into a course of action” in order to prevent the sense of discomfort or investment is referred to as entrapment (Fox & Staw, 1979; Staw, 1981, p. 578).
Hence, a higher sense of commitment increases the value of that course of action to a person and the likeliness that this person will continue this behaviour in the future.
Less apparent is the rationalization process through which commitment can also play a role in escalation of unethical behaviour. People have the tendency to rationalize why an act is acceptable and in the process convince themselves that their prior decision was no mistake (Fox & Staw, 1979). Thus, while a person rationally knows that stealing money is wrong, the same person can justify this immoral act afterwards by adding that it was because other people steal too. In a business setting, the danger is that an employee will start seeing unethical behaviour as acceptable business conduct by thinking: “that’s just the way we do things around here”.
Literature clearly gives rise to the notion that commitment foster escalation, which is referred to as “the increased likelihood and intensitiy” for future immoral behaviour (Tsang, 2002, p. 27). Despite the strong theoretical groundwork, the effects of
commitment on escalation of unethical behaviour lack empirical support. Only the results in the study by Martens et al. (2010) have shown that individuals with a higher degree of commitment (i.e. the ones who already killed five bugs), were more likely to escalate their previous immoral acts than the ones who were less committed (i.e. the ones who had only killed one bug). The researchers in this study concluded that people were more motivated to continue when they had killed five insects in order to rationalize that their pervious behavior was justified. Quitting or killing fewer bugs in a subsequent task would entail admitting that they were wrong. In support of this is the finding that people would only escalate when they were told that they were really killing insects. People who were told that the insects were not killed did not show the same degree of escalation. Thus in this experiment there was a clear case of increased commitment through: the need to be consistent, unwillingness to admit a wrong decision and a rationalization of the unethical acts. While the different processes contributing to a high commitment can be described separately, these processes appear to occur in combination instead of isolation in practice.
This is also fleshed out in the experiment by Milgram (1974) in which people were asked to administer electric shocks to another participant with an increasing voltage, demonstrated that over sixty per cent of the research subjects shocked their participant- accomplice with a voltage that was reported to be deadly. Based upon the findings regarding escalation, the Milgram experiment -among other findings- also shows that people experience the same underlying causes for escalation of unethical behaviour as in the Martens et al. (2010) experiment. People in the Milgram experiment continued administering shocks, because they viewed themselves as being committed to the procedure by initially conducting a morally acceptable act (administering a minor shock).
Clearly people are not per se unwilling to stop the escalation process, but are forced by multiple psychological processes which take place in order to sustain an acceptable self- image. This stimulates behavioural escalation into an unethical deed. Resultantly these experiments and the available theories suggest that a high degree of commitment is a requisite for future escalation. I therefore pose the following:
Hypothesis 1: Individuals with a high degree of commitment to their initial behaviour will show a higher degree of escalation in their subsequent behaviour, compared to individuals with a lower degree of commitment to their initial immoral behaviour.
(IM)Morality across domains
Findings on the effects of commitment on behavioral escalation have shown that immoral escalation primarily occurs when individuals are highly committed to their behavior. This gives rise to the question whether people with a high degree of commitment always escalate or that there are conditions under which escalation is inhibited. A possible answer to this question can be found in the large amount of research supporting the argument that escalation only occurs within the same domain (e.g.
Martens et al., 2010; Milgram, 1974).
The available research on cross domain moral behavior has given rise to the idea that it is exactly the same domain behavior which allows escalation to emerge. A large field of research regarding cross domain moral behavior has shown that the moral self- image fluctuates and varies across different types of moral behavior (e.g. Monin &
Jordan, 2009). As a consequence of their immoral decisions in one domain, individuals perceive a decrease in their self-worth; and cope with this loss of worth by engaging in
moral behavior in another domain (Sachdeva, Iliev, & Medin, 2009). Thus, when behavior occurs across domains individuals have the tendency to compensate either for moral or immoral behavior.
Research on moral compensation and moral licensing supports this argument by asserting that individuals consider their previous immoral and moral actions to determine whether or not to engage in future unethical behavior in another domain (Miller &
Effron, 2011). With regard to this, Monin and Jordan (2011) argue that feeling moral in one domain can serve to compensate for feeling less adequate in other moral domains.
Consequently, the perception of being moral can permit people to conduct an immoral act or express an immoral opinion in another domain (Effron, Cameron & Monin, 2009;
Monin & Miller, 2001). In literature the moral self-image is even referred to as continuously changing; similar to the monetary value in a deposit account fluctuating based upon good deeds and bad deeds (Miller & Effron, 2010). In a nutshell, when initial behavior is immoral, people are likely to engage in moral behavior in another domain and vice versa (Effron, Cameron & Monin, 2009; Mazar and Zhong, 2010; Monin & Miller, 2001). This was confirmed in a study conducted by Mazar and Zhong (2010) in which participants were given the task to choose between environmentally friendly or conventional products. The results showed that participants who chose more green products (moral behavior) showed more immoral subsequent behavior when they were given the chance to cheat on an experimental assignment or to steal money from the experimenter. Thus, whereas escalation is stimulated by the need to be consistent and to justify prior decisions; licensing and compensation indicate that people can act inconsistently and justify this through a dynamic self-image.
Boiled down, on one hand there is a pattern of moral compensation and on the other hand there is a pattern of immoral escalation. This gives rise to the question: When do people compensate and when do they escalate? The findings show that this is determined by whether or not subsequent behavior is in the same domain as prior behavior. When behavior is in the same domain, commitment increases as a result of: the need to be consistent, unwillingness to admit a mistake and rationalization. At glance, when behavior is in a different domain individuals can unchain themselves from their commitment and self-meaning to a certain type of behavior (e.g. immoral); and compensate for this in another domain. The apparent difference between escalation and compensation is the occurrence of a domain switch. I therefore expect that individuals can unchain themselves from their commitment to their initial immoral behavior and will show compensation in subsequent behavior when the domains are switched. I therefore hypothesize the following:
Hypothesis 2: A domain switch moderates the relationship between commitment and immoral escalation by inhibiting future immoral escalation.
METHOD
Design and participants
Participants were 268 students (172 male, 96 female; Mage=21, SDage= 2.05) of the Faculty for Business Economics at the University of Groningen, who enrolled for an experimental study. They were randomly assigned to one of four conditions of a 2 (Commitment: high vs low) x 2(Domain: switch vs no switch) between-subject design experiment.
Procedure
The experiment consisted of two different games: a cheating game, in which participants were assessed based upon the degree to which they lied; and a stealing game, in which the participants were assessed based upon the degree to which they stole. Some participants were assigned to both games and some were assigned to either the stealing or cheating game. The games were divided in two phases, which will be addressed in the commitment section.
Cheating game. In the cheating game all participants (player one) were told that
they were linked to another participant (player two) and that they could both earn money.
In reality every participant was player one. They were informed that one pair would be randomly selected at the end of the experiment by a computer to receive the prize money.
In addition, the participants were informed that they would play a random number of rounds and that every round would imply a decision regarding the division of money between player one and player two. In every round ten options regarding the division of money were given. The decisions ranged from A-J. Option A meant player one would get
€0 and player 2 would get €50, option B meant player one would get €5 and player 2 would get €45, option C meant player one would get €10 and player 2 would get €40, option D meant player one would get €15 and player 2 would get €35, option E meant player one would get €20 and player 2 would get €30, option F meant player one would get €30 and player 2 would get €25, option G meant player one would get €35 and player 2 would get €15, option H meant player one would get €40 and player 2 would get €10, option I meant player one would get €45 and player 2 would get €5 and option J meant player one would get €50 and player 2 would get €0. Player one was given the
assignment to send a message regarding the right decision to player two, after which player two would make a decision. Player one was informed that player two would receive no other information except for the information given by player one regarding the choice player two was supposed to make and the fee received at the end of each round.
Participants were thus informed that when player one decided to send player two the message to choose option J, this would result in the largest financial benefit for player one. The degree to which player one lied ranged from low to high, A being low and J being high. The decisions -ranging from A-J- were made over a course of several rounds.
Stealing game. In the stealing game all participants (player one) were told that
they were linked to another participant (player two) and that they could both earn money.
In reality every participant was player one. They were told that one pair would be randomly selected at the end of the experiment by a computer to receive the prize money.
Financially, the decisions in the stealing game were identical to the cheating game.
Hence, the division of money ranged From A-J, the only difference was that in this case the game was framed as “How much are you stealing from player two”. After the explanation, the participants were informed that they would play a random number of rounds and that every round would imply a decision regarding the amount of money player chose to steal.
Afterwards the participants were requested to answer a few questions about the experiment in order to check whether the manipulations were administered successfully.
After completing the experiment participants were debriefed and either received credits for a course or a fee of €7 for their participation.
Manipulation
Commitment. Every participant played the games in two phases. Commitment
was manipulated by changing the length of the first phase. In the high commitment condition this first phase consisted of twelve rounds whereas in the low commitment condition the first phase consisted of two rounds. In both the high and low commitment condition the second phase consisted of five rounds.
Domain switch. The domain switch was manipulated by the presence or the
absence of a change in the type of game that the participant was playing. There were two no switch conditions and two switch conditions: steal-steal, cheat-cheat, steal-cheat and cheat-steal. In the no switch condition the participant played one game during a number of rounds and were at the beginning of the second phase given the explicit message that they would continue playing the same game. This implied that participants who played the cheating game in the first phase, continued playing the same game in the second phase, the same goes for stealing. In the switch condition participants who played the cheating game in the first phase, changed to the stealing game in the second phase after they received the explicit message that they would play a different game. The same goes for when the player started with the stealing game.
Dependent measures
Escalation. In order to determine the degree of escalation, firstly the scores
ranging from A-J were recoded ranging from one to ten, A being one and J being ten.
Thus one implied not stealing/cheating and ten implied highest degree of stealing/cheating. Secondly, the mean initial behavior was calculated by adding the
scores of the first phase and by dividing this sum by the number of rounds played and the mean of subsequent behavior was measured by adding the scores in the second phase and by dividing them by the number of rounds played. Based upon these numbers, the extent to which participants escalated in the different conditions was measured by (Mean initial behavior) – (Mean subseqent behavior); 2 initial rounds in low commitment, 12 initial rounds in high commitment and 5 subsequent rounds in both low and high commitment.
Hence, the escalation scores in the low commitment condition (Alpha = 0.87) and the high commitment condition (Alpha = 0.95) were based upon the difference between phase one (initial behavior) and phase two (subsequent behavior) which were ranged from 1 to 10 meaning a higher score indicated a higher degree of stealing/cheating. All items started with “choose the message you want to send to player two”.
Manipulation checks
I checked the commitment manipulation by the following four statements: “I supported the choices I had made before”, “there was nothing wrong with the choices I had made before”, “I felt personally responsible for the choices I had made before”, “I felt responsible for prior choices”, which were measured by a seven point answering scale (1= don’t agree, 7= highly agree). In addition, I checked the domain switch manipulation by the following five statements: "The decisions in the last five rounds were more or less the same as the decisions in the prior round(s)” (reverse recoded), "I experienced the last five rounds as a complete different situation than the prior rounds”,
“At a certain point there was a change in the rules of the experiment”, “the last five rounds comprised a different decision than the prior round(s)”, “there were a lot of
commonalities in the rules of the last five and the prior round(s)” (reverse recoded), which were measured by a seven point answering scale (1= don’t agree, 7= highly agree.).
RESULTS Manipulation checks
A 2(commitment) x 2(domain switch) ANOVA was performed on the commitment manipulation check.
Commitment check. There was no main effect of commitment on the commitment
check, F(1, 267)= 1.59, p= 0.21 >.05. Participants in the low commitment condition (M = 5.24, SD = 1.06) showed the same degree of commitment as the participants in the high commitment condition (M = 5.27, SD = 1.05). More, there was no main effect of commitment on the domain check F(1, 267)= 1.59, p= 0.21 >.05. Participants in the low commitment condition (M = 2.92, SD = 1.43) showed the same the degree of experienced domain switch as the participants in the high commitment condition (M = 2.72, SD = 1.35).
Domain switch check. Domain switch was positively related to the domain switch
check (r = .65, p < .05). Participants in the domain switch condition (M = 3.73, SD = 1.19) indicated a higher degree of experienced domain switch, than participants in the no domain switch condition (M = 1.92, SD = 0.90). In addition, domain switch was negatively related to the commitment check (r = -.17, p < .05). Participants in the no domain switch condition (M = 5.43, SD = 1.02) showed a higher degree of experienced commitment than the participants in the domain switch condition (M = 5.08, SD = 1.05).
The results of the analysis are shown in Table 1 and 2.
TABLE 1
Descriptive Statistics: mean, standard deviation and number
Mean Std. Deviation N
Commitment check Low commitment 2.92 1.43 136
High commitment 2.71 1.35 132
Total 2.82 1.39 268
NO domain switch 1.92 .90 135
Domain switch 3.73 1.19 133
Total 2.82 1.39 268
Domain check NO domain switch 5.43 1.02 135
Domain switch 5.08 1.05 133
Total 5.25 1.05 268
Low commitment 5.24 1.06 136
High commitment 5.27 1.05 132
Total 5.25 1.05 268
The effects of commitment and domain switch
I hypothesized that a high degree of commitment leads to more escalation than a low degree of commitment. I performed a 2(commitment) x 2(domain switch) ANOVA on the extent to which commitment leads to more escalation. There was no significant main effect of commitment, F(1, 267)= 0.01, p= 0.94 >.05 Participants in a low commitment condition (M= 0.15, SD= 1.76) showed the same degree of escalation as participants in a high commitment condition (M=0.17, SD= 1.50). Thus, no evidence was found for the hypothesis that higher commitment increased escalation.
TABLE 2 Correlations
1 2 3 4 5
1.Domain- checkmean
Pearson
Correlation 1 -.18** .65** -.08 -.03
Sig. (2-tailed) .00 .00 .21 .64
N 268 268 268 268 268
2.Commitment- checkmean
Pearson
Correlation -.18** 1.00 -.17** .01 .03
Sig. (2-tailed) .00 .00 .85 .69
N 268 268 268 268 268
3.Domainswitch Pearson
Correlation .65** -.17** 1.00 .01 -.05
Sig. (2-tailed) .00 .01 .91 .46
N 268 268 268 268 268
4.Commitment Pearson
Correlation -.08 .01 .01 1.00 .01
Sig. (2-tailed) .21 .85 .91 .94
N 268 268 268 268 268
5.Escalation Pearson
Correlation -.03 ,03 -.05 .01 1,00
Sig. (2-tailed) .64 .69 .46 .94
N 268 268 268 268 268
**. Correlation is significant at the 0.01 level (2-tailed).
More, I hypothesized that domain switch had a negative effect on the relation between commitment and escalation. A 2(commitment) x 2(domain switch) ANOVA was performed on how participants escalated in subsequent behavior. This showed no interaction effect of domain switch, F(1, 267)= 0.97, p= 0.33> 0.05. Hence, no support was found for a moderating effect of domain switch. See Table 3 for the analysis.
TABLE 3
ANOVA Results For The Main And Interaction Effects
Squares df Mean Square F Sig.
Corrected Model 4.12a 3 1.37 .51 .66
Intercept 6.79 1 6.79 2.52 .11
Commitment .02 1 .02 .01 .94
Domain switch 1.55 1 1.55 .58 .45
Commitment * Domain switch 2.62 1 2.62 .97 .33
Error 710.36 264 2.69
Total 721.25 268
Corrected Total 714.48 267
a. R Squared = ,006 (
A moderating effect was found when the first round after the switch was used in calculating the degree of escalation. Thus escalation was calculated by (Mean of the first phase) – (score in first round). Again a 2 (commitment) x 2 (domain switch) ANOVA was performed. This showed an interaction effect of domain switch, F(1, 267)= 6.29, p <
0.05. Participants in the no domain switch condition (M = -0.33, SD = 1.67) showed a lower degree of escalation than the participants in the domain switch condition (M = 0.10, SD = 2.60). See Table 4 for the analysis.
TABLE 4
ANOVA Results For Main And Interaction Effects
Source
Type III Sum
of Squares df Mean Square F Sig.
Corrected Model 46.79a 3 15.60 3.36 .02
Intercept 3.87 1 3.87 .83 .36
DOMAINSWITCH 12.49 1 12.50 2.69 .10
COMMITMENT 5.72 1 5.72 1.23 .27
DOMAINSWITCH *
COMMITMENT 29.20 1 29.20 6.29 .01
Error 1226.45 264 4.65
Total 1277.20 268
Corrected Total 1273.24 267
a. R Squared = ,037 (Adjusted R Squared = ,026)
DISCUSSION
The experimental study tested how commitment and a change in domains influenced the extent to which participants escalated in their immoral behavior. Results of the study show that high commitment participants and low commitment participants showed no difference in their escalating behavior. Thus, commitment had no influence on the escalating behavior of people. On the question whether the domain switch influenced the relation between commitment and more than one round was accounted for. Whether they played in a cheating game, stealing game; or played both for a defined number of consecutive rounds made no difference in their escalating behavior. The remarkable finding was made however, that there was an effect of domain switch on the relation between commitment and escalation, when only the first round in phase one was used to determine the degree of escalation. Thus, there was no influence of a domain switch on the relation between commitment and escalation when all rounds in phase two were
accounted for, but there was an effect when the first round was used to test the influence of a domain switch on the relation between commitment and escalation.
One possible explanation for why no main effect of commitment was found probably stems from the unsuccessful manipulation of commitment. As the manipulation checks show, people who had played the game for a longer time did not identify more with their previous behavior than people who played the game for a shorter time. It is well established that the degree of commitment depends on the amount of time and effort an individual puts into a task or activity (Staw, 1981; Tsang, 2002), but the exact point at which a person makes the decision to commit to a course of action remains highly subjective. Still, the question remains as to why the commitment manipulation failed despite the fact that the high commitment group played the game for a much longer amount of time than the low commitment group. There are two apparent explanations as to why the participants in the high and low commitment situation did not experience a difference. Firstly, people’s behavior in real life and the way they perceive their investment to something in real life differs from an experimental setting. The participant may have felt no commitment, because they apparently did not link their decisions to their self-image, possibly as a result of the artificial setting. Playing a game with another person and cheating while seeing that person in real life, will probably lead to a stronger emotional response than when this person is associated with a computer. Perhaps showing the other participant and the reactions as a result of the game can increased the felt commitment, due to higher emotional responses. Secondly, the number of rounds simply might not have been enough to activate a sense of commitment.
After this study the question therefore remains, what do we know about commitment? There is compelling evidence in literature supporting the assumption that commitment influences the degree of escalation. Martens et al. (2009) & Staw (1981) found that commitment can influence the self-image to such an extent that people continue their immoral behavior. Using these findings as groundwork, this paper examined individuals’ reactions to both different types of initial and subsequent immoral behavior and a degree of commitment, either high or low. While this paper finds no support for the available theory, it does provide researchers with a more skeptic view on whether and when commitment occurs and actually influences future escalation. This provides an interesting point of entry for future research.
In addition the data showed that a domain switch did not have an effect on the relation between commitment and escalation after more than one subsequent round, but did have an effect after just one round. One major contribution of this paper therefore, is to show how and when a domain change might predict future escalation. This paper advances and touches upon the frontiers of moral behavior and the effects of a domain switch on future moral behavior by indicating that a domain switch can disinhibit escalation, contrary to what was found in literature. Multiple studies show that people engage in moral behavior in another domain after they have been immoral in a previous domain (Effron, Cameron & Monin, 2009; Mazar and Zhong, 2010; Monin & Miller, 2001). Likewise I expected that a domain switch would lead to compensation, but the results show the opposite, it resulted in more escalation. How do these findings reconcile with what was found in literature? In line with what was shown by Martens et al. (2010), people who exhibited immoral behavior in the first phase, probably escalated because
they did not want to admit that their prior behavior was wrong. Thus, somehow despite the change in domains people were unable to unchain themselves from the commitment to their prior immoral behavior. Especially the idea of changing to another game might have further stimulated the rationalization processes in order to justify prior actions. Thus instead of being liberating, changing the domain could have led to more entrapment due to the need to have more reasons as to why prior behavior was acceptable. An employee who has been stealing from the manager, might therefore continue immoral behavior when given the chance to cheat on a test, because of the need rationalize the immoral norm and to prevent a sense of wrongdoing. A possible explanation for the reason why an effect was found after one round, but not after five can be because people initially –as a result of the message regarding a change of the game- are aware that they are playing a different game and therefore change their behavior consciously, but after a few rounds this awareness might fade away unconsciously. In essence two explanations for the finding regarding a domain switch can be distilled: a domain switch or the experience of playing a different game might be affected by the number of rounds played and a domain change stimulates escalation by forcing people to further rationalize their behavior in another domain.
These findings urge further continuation of research in the field of cross domain moral behavior. This experiment was limited, because participants were only examined based upon two domains (stealing & cheating) and based upon one switch; whereas multiple domain switches during the experiment could have given more information regarding possible patterns. Additionally, the results show that stealing and cheating were seen as different behaviors. These behaviors however, might not have been seen as being
different enough, meaning that the people somehow still viewed them as being part of the same domain as the number of rounds progressed. An explanation can be found in the set-up of the experiment, which in essence came down to pressing a button and acquiring a fictional amount of money. Thus perhaps the behaviors itself were seen as different, but due to the set-up of the game both were seen as being the same; a computer simulation.
Future studies should therefore not only use different domains, but should also make the required acts per domain differently. Meaning that in the cheating game people would push the button, but in the stealing game they would have to use arrow keys for example to increase the perceived difference. In addition, other domains (e.g. discrimination, fraud), might have yielded different results. This was not tested for, because the conscious choice was made to choose stealing and cheating due to prior successful experimental studies with both games.
Practically, the implications for the results are that if people act in an immoral way, attempting to change this behavior by switching domains, will not per se result in a decrease of their immoral behavior, but can even lead to more immoral behavior. Thus, when an employer finds out that an employee is stealing inventory one can expect that if this employee gets the opportunity to cheat on for e.g. a quality audit, there is a high probability that this will happen. In conclusion, when immoral behavior in a different domain occurs, this same employee will show the same/worse behavior. More, this study gives managers new insights into possible factors contributing to immoral escalation and unethical conduct in general. If employees can be locked into a course of action and throw good money after bad, managers should realize that when an employee is highly committed to a project and refuses to terminate it -despite strong evidence that this
project unethical- it is not because of the incompetence of the employees, but possibly because of the unwillingness to admit their mistakes. What is important -from a managerial and organization perspective- is the fact that a minor immoral deed can escalate into major future unethical behavior in different domains if left unnoticed, leading to what Martens et al. (2010) describe as an immoral norm.
This paper urges future research by managers, organizations and academics regarding the factors contributing to the escalation of unethical behavior. Additional investigation is necessary in order to make sure that cases such as Enron remain stories of the past; and can give managers and organizations the tools to nourish moral behavior
within individuals, organizations and society as a whole.
REFERENCES
Aquino, K., Freeman, D., Reed Ii, A., Lim, V. & Phelps, W. (2009). Moral Identity And The Self-Regulation Of (Im)Moral Behavior. Journal Of Personality And Social Psychology, 97, 123-141.
Arkes, H. R. & Blumer, C. (1985). The Psychology Of Sunk Cost. Organizational Behavior And Human Decision Processes 35, 124-140
Burke, P. J. & Reitzes, D. C. (1991). An Identity Theory Approach To Commitment.
Social Psychology Quarterly, 54, 239-25.
Cialdini, R. (2007). Influence: The Psychology Of Persuasion. New York: Harper Collins.
Festinger, L. (1957). A Theory Of Cognitive Dissonance. Stanford University Press Staw, B. M. Rationality And Justification In Organizational Life. In B. Staw & L.
Cummings (Eds.), Research In Organizational Behavior. Greenwich, Conn: Jai Press, 1979. Vol. 2.
Gino, F., Schweitzer, M., Mead, N., & Ariely, D. (2011). Unable To Resist Temptation:
How Self-Control Depletion Promotes Unethical Behavior. Organizational Behavior And Human Decision Processes, 115(2), 191-203.
Jones, T. M. (1991). Ethical Decision Making By Individuals In Organizations: An Issue- Contingent Model. Academy Of Management Review, 16: 366-395.
Lunenburg, F.C. (2010). Escalation Of Commitment: Patterns Of Retrospective Rationality. International Journal Of Management, Business, And Administration, 13.
Martens, A., Kosloff, S, & Jackson, Le. (2010). Evidence That Initial Obedient Killing Fuels Subsequent Volitional Killing Beyond The Effects Of Practice. Social Psychology And Personality Science, 1, 268-273.
Mazar, N. & Zhong, C. B. (2010). Do Green Products Make Us Better People?
Psychological Science, 21, 494–498.
Merritt, A. C., Effron, D. A. & Monin, B. (2010), Moral Self-Licensing: When Being Good Frees Us To Be Bad. Social And Personality Psychology Compass, 4, 344–
357.
Milgram, S. (1974). Obedience To Authority: An Experimental View. New York: Harper And Row.
Miller, D. T., & Effron, D. A. (2010). Psychological License: When It Is Needed And How It Functions. In M. P. Zanna And J. Olson (Ed..), Advances In Experimental Social Psychology, 43, (Pp. 117-158). San Diego, Ca: Academic Press
Monin, B., & Jordan, A. H. (2009). Dynamic Moral Identity: A Social Psychological Perspective. In D. Narvaez & D. Lapsley (Ed.), Moral Self, Identity And Character:
Prospects For A New Field Of Study. New York: Cambridge University Press.
Monin, B., & Miller, D.T. (2001). Moral Credentials And The Expression Of Prejudice.
Journal Of Personality And Social Psychology, 81: 33-43
Rantanen, M. (2007). Reasons Of Systemic Collapse In Enron In: Raimo P. Hämäläinen And Esa Saarinen (Ed.). Systems Intelligence In Leadership And Everyday Life.
Systems Analysis Laboratory, Helsinki University Of Technology, Espoo.
Sachdeva, S., Iliev, R., Medin, D. L. (2009). Sinning Saints And Saintly Sinners: The Paradox Of Moral Self-Regulation. Psychological Science, 20, 523-528.
Staw, B. M. (1981). The Escalation Of Commitment To A Course Of Action. Academy Of Management Review, 4, 577-587.
Tenbrunsel, A. E. & Smith-Crowe, K. (2008). Ethical Decision Making: Where We' Ve Been And Where We're Going. Academy Of Management Annals, 2, 545-607
Trevino, L.K., Weaver, G.R., & Reynolds, S.J. (2006). Behavioral Ethics In Organizations: A Review. Journal Of Management, 32, 951–990.
Wiltermuth, S. S. (2011). Cheating More When The Spoils Are Split. Organizational Behavior And Human Decision Processes, 115(2), 157-168.
