Elsevier

Economics Letters

Volume 121, Issue 3, December 2013, Pages 425-427
Economics Letters

Personal characteristics and lying: An experimental investigation

https://doi.org/10.1016/j.econlet.2013.09.005Get rights and content

Highlights

  • We explore personal characteristics and the decision to lie.

  • Sex, age, student debt, size of return and religious observance were insignificant.

  • Important individual variables were; major, religious importance and salience.

  • Relevant family traits were; divorced parents, single parent and economic status.

Abstract

We explore the relationship between personal characteristics and the decision to lie to an anonymous partner in a cheap talk environment. We find that sex, age, grade point average, student debt, size of return, socioeconomic status, and average time spent in religious observation are not related to the decision to lie. A subject’s major of study, the marital status of their parents, whether or not they were raised by a single parent, religious importance and whether or not the subjects came to collect their pay were important explanatory variables.

Introduction

A growing body of literature explores the likelihood of lying in response to economic incentives. Increasingly, this literature focuses on the role an individual’s sex plays in determining propensity to lie, often with conflicting results. For example, Dreber and Johannesson (2008), Erat and Gneezy (2011) and Friesen and Gangadharan (2012) find that males are more likely to lie than females, while Childs (2012a) and Gylfason et al. (2013) find no difference in lying by sex.

Childs (2012a) posits two potential explanations for the difference in findings. While Childs (2012a) and Dreber and Johannesson (2008) use essentially the same environment, the incentives for lying are much greater in Childs ($10 compared to $1.25) and thus it could be that females have a higher threshold cost of lying. Gylfason et al. (2013) test this hypothesis by using rewards similar in size to Dreber and Johannesson ($0.87) and find no difference in lying by sex despite the smaller incentive.

Childs (2012a) suggests cultural differences as an alternative explanation for the difference in findings. Dreber and Johannesson (2008) use Swedish subjects while Childs’ subjects were Canadian. Gylfason et al. (2013) argue that Icelandic subjects are not effectively different from Swedish subjects, thereby controlling for differences in culture. Gylfason et al. conjecture that differences in individual characteristics, beyond simply sex, can explain the differences between the findings of Dreber and Johannesson (2008) and Childs (2012a).

We explore the impact of personal characteristics on the propensity to lie for large ($10) and small ($2) returns. The data on large returns is drawn from the same experiment as Childs (2012a), the data on small returns is drawn from a subsequent experiment at the same university. We find that major, marital status of parents, whether or not the subject was raised by a single parent, religious importance, and whether or not the subject claimed their payment are all important explanatory variables. Sex, age, GPA, debt, size of return, socioeconomic status, and time spent in religious observance were not found to be statistically significant. Further, the level of incentive was not found to have a differential impact by sex.

Below is a description of the experimental design and survey, followed by results and conclusions.

Section snippets

Experimental design

The experiment is a modification of the one used in Childs (2012a), which in turn was based on Gneezy (2005). Individuals in different rooms are paired with a partner. Subjects in one room are senders and subjects in the other room are receivers. Senders see two payoffs and send a simple message advising receivers which payoff would yield a higher payoff for receivers. Receivers then choose a payoff. The payoffs were A = ($15, $5) or B = ($5, $15)1

Results

Fig. 1 shows the portion of subjects sending false messages by incentive level and sex. In neither treatment is the difference between males and females statistically significant3 (Low return p-value = 0.584, high return p-value = 0.996). We do see an increase in the portion of females sending false messages when the return increases, but this increase is not statistically significant (p-value = 0.362).

We cannot support Childs’ (2012a)

Conclusion and discussion

A number of researchers have found that males are more likely to lie in a simple sender–receiver game for a monetary reward than females (Dreber and Johannesson, 2008, Erat and Gneezy, 2011, Friesen and Gangadharan, 2012). Others (Childs, 2012a, Gylfason et al., 2013) have found no difference in lying by sex. Childs (2012a) suggests two potential explanations for the difference in findings; females may have a higher minimum threshold return to induce dishonesty and cultural differences between

References (8)

There are more references available in the full text version of this article.

Cited by (13)

  • On religion, lying, and social preferences

    2019, Economics Letters
    Citation Excerpt :

    Tan (2006) finds religiosity produces counterworking effects on social preferences. This study explores the relation between religiosity and lying and social preferences, while conducting it on a largely religious campus, mitigating the group membership effect (Childs, 2013). The rest of the paper proceeds as follows: Section 2 presents the experiment procedures, Section 3—the empirical results, and Section 4 concludes.

  • Gender differences in cheating: Loss vs. gain framing

    2018, Economics Letters
    Citation Excerpt :

    Our study is the first to test for gender differences in cheating using the die-paradigm and considering separately the Gain and Loss domains. Childs (2012, 2013), Cappelen et al. (2013) and Grolleau et al. (2016), among others, investigate gender differences in cheating using other tasks. In the die-paradigm, the experimental evidence when reports are associated with gains is mixed; e.g., Clot et al. (2014) find that females cheat more than males, while Conrads et al. (2017) find the opposite, and Muehlheusser et al. (2015) do not observe gender differences (see Abeler et al., 2016 for a recent meta-study on the die-paradigm).

  • Personality in Gneezy's cheap talk game: The interaction between Honesty-Humility and Extraversion in predicting deceptive behavior

    2016, Personality and Individual Differences
    Citation Excerpt :

    The theoretical assumption that individuals are entirely self-interested (Fehr & Gachter, 2000) is on the defensive, with evidence of people preferring not to maximize their own gains when able to do so (Charness & Gneezy, 2008). In recent economic studies, the focus on why some individuals prefer not to lie in response to economic incentives has been trained on their personal characteristics (Childs, 2012, 2013; Dreber & Johannesson, 2008; Gylfason, Arnardottir, & Kristinsson, 2013; Muehlheusser, Roider, & Wallmeier, 2015), as opposed to personality characteristics emphasized in personality research (Hilbig & Zettler, 2015). Personality can affect the decisions that individuals make (Ben-Ner, Kong, & Putterman, 2004; Boone et al., 1999; Cobb-Clark and Schurer, 2012; Lönnqvist, Verkasalo, & Walkowitz, 2011), and some trait variables have been shown to predict unethical behavior, such as the Locus of Control (Street and Street, 2006) and Honesty-Humility (Lee, Ashton, & Shin, 2005).

View all citing articles on Scopus
View full text