Introduction
Instances of unethical behavior are widespread. A big scandal within the academic field was the discovery of academic fraud conducted by Dutch scientist Diederik Stapel. Stapel conducted a lot of popular and highly cited studies in the field of social psychology. However, in 2010 he was found guilty of fabricating and manipulating the data of over 50 of his studies (Bhattacharjee, 2013). A relatively more frequently occurring example of deception is that of false insurance claims. On a yearly basis, these claims cost insurers millions of euros. In the Netherlands, 10.001 cases of false insurance claims were confirmed in 2016, which on average would have cost Dutch insurers 8.000 euro per claim (“CBV Factsheet”, 2017). In the media, most attention is given to scandals like this. However, there are also a lot of examples of dishonest behavior that seem small, but happen all the time. Think of lying on your resume, buying clothes at a shop to return them after wearing them for a day (“wardrobing”), keeping excessive change money that was accidentally given to you, cheating on exams, illegally downloading movies and music and so on. People confess to tell one or two lies per day (DePaulo, Kashy, Kirkendol, Wyer, & Epstein, 1996). Taken together these small instances of dishonesty potentially have a big impact on society.
A lot of research has been done on unethical behavior by studying it in a laboratory setting. An established finding is that dishonesty is often not the result of a few bad individuals that cheat maximally, but rather a lot of people that cheat a little (Ayal & Gino, 2011; Fischbacher & Föllmi-Heusi, 2013; Mazar, Amir, & Ariely, 2008; Shalvi, Dana, Handgraaf, & De Dreu, 2011). In several studies conducted by Ariely and his colleagues they use the “Matrix task” to study dishonesty. Subjects receive a paper with 20 matrices filled with 12 three-digit numbers. Two of the numbers in each matrix add exactly up to 10. Subjects are instructed to find as many combinations as possible within limited time. For every combination found, they are financially rewarded. In these experiments subjects have the possibility to cheat, because the paper is shredded afterwards, so their stated performance cannot be checked. In this situation, almost all subjects are dishonest. However, subjects only overstate their performance by two matrices on average, instead of cheating to the maximum amount (Gino, Ayal, & Ariely, 2009, 2013; Mazar et al., 2008; Mead, Baumeister, Gino, Schweitzer, & Ariely, 2009). Mazar et al. (2008) state that this ‘partial dishonesty’ can be explained by the fact that people are tempted to cheat when this leads to personal benefits, but at the same time want to maintain a positive self-concept. They find a balance between these factors by cheating a little bit.
There are a lot of factors that influence where people draw the line regarding their unethical behavior. For example, the notion that the authority that people can steal from is wealthy makes it more likely that people will behave dishonestly. When people are in in the visible proximity of wealth (cash), they are more likely to cheat, compared to being in a setting of scarcity (Gino & Pierce, 2009). Vohs, Mead, & Goode (2006, 2008) found that people behave more self-focused and self-serving after being primed with visuals of money. It is therefore possible that, when people are being primed with visuals of the benefits that they can earn by being dishonest the tendency to cheat will increase. This has to the best of my knowledge not been researched so far. Therefore, this study will try to answer this question.
To look at the external validity of results found by laboratory studies, several field experiments on dishonesty have been conducted. After measuring dishonest behavior in the lab, participants of the study of Potters and Stoop (2016) unknowingly became part of a field experiment. In the lab, participants had – possibly dishonestly – earned a specific amount of money. This money would be paid on their bank account the next day. Potters and Stoop deliberately paid some participants too much and send them an e-mail asking to verify if payment was received successfully. Not replying to this e-mail and not mentioning the overpayment was marked as dishonest behavior. They found that behavior in the lab correlated to behavior in the field. The individuals who did not mention the overpayment, were the same individuals who were dishonest in the lab environment. Similar results were found by Dai, Galeotti and Villeval (2017). The current research will analyze if this is also the case for results found in an online experiment.
Knowing the person that you steal from or lie to decreases the extent of dishonest behavior displayed (Azar, Yosef, & Bar-Eli, 2013; van Swol, Malhotra, & Braun, 2012). Some researchers – especially students writing their thesis – use a convenience sample, since this is an easy, fast and cheap method to find respondents. A convenience sample often consists of people that the experimenter knows personally, such as family and friends. Using this kind of sample when studying dishonesty is a potential confounding factor. Participants that know the experimenter could act as if they are honest, because they do not want to deceive the experimenter. To my knowledge this possibly confounding factor has not been researched before in relation to dishonesty. Taking all discussed literature into account, this research will therefore try to answer the following research question:
‘Does being familiar with the experimenter influence the extent to which people display dishonest behavior in different types of experiments? Is this influenced by being primed with possible benefits?’
The research question will be answered by conducting an online and field experiment. Subjects are all members of a Dutch student association. Some of the subjects are familiar with the experimenter of this research whereas others are not. During the online experiment subjects must complete a lottery-variant of the matrix task used by Ariely and his colleagues. They are instructed to solve as many matrices as possible in two minutes. Afterwards they must record the number of matrices they have solved. It is not possible to check if this amount is answered truthfully, therefore subjects can cheat without getting caught. The number of matrices solved corresponds to the amount of lottery tickets earned. With these tickets, subjects take part in a lottery and have a chance of winning a Bol.com gift card worth €50. This prize incentivizes subjects to lie about the number of matrices solved. Half of the participants will be primed with a picture of the prize they can win on all pages of the online survey. Based on the research of Potters and Stoop (2016), subjects will unknowingly become part of a field experiment after finishing the online experiment. All subjects will receive an e-mail with the amount of lottery tickets earned. Subjects are asked to send an e-mail when the number is incorrect. However, the amount of tickets stated is always incorrect. Subjects that do not mention the mistake will be marked as dishonest.
The outcomes of this research are important for the interpretation of previous research on dishonesty conducted with a convenience sample and for comparable upcoming research, especially that conducted by students. If people that are familiar with the experimenter indeed behave differently than people unfamiliar with the experimenter, future research on dishonesty should be conducted with participants that are unfamiliar with the experimenter only. Furthermore, by doing both an online and field experiment, this research can add to the literature of the generalizability results found in an online experiment to the field. So far this has only been studied for lab experiments. Next to that, we can observe if the effects of being familiar with the experimenter hold both in an online and a field setting. Studying the effect of priming with possible benefits of being dishonest could give us insights on how to make dishonest behavior less likely to prevail. For example, in situations where dishonest behavior is likely to occur the focus should not be on the benefits that can be earned.