## 4.2 The Study: Deception Detection

### 4.2.1 Study Information

#### 4.2.1.1 Description

Laboratory studies have consistently shown that people make terrible lie detectors. It seems that liars in these settings do not provide reliable, observable indicators of their deception. But this could in part be due to the necessarily low stakes of lying in experimental studies. High-stakes lies, such as are common in criminal justice contexts, may elicit involuntary cues from liars that are impossible to replicate in an ethical laboratory study.

This is supported by research. Ecologically valid studies of high-stakes lies have found that there are reliable behavioural cues to deception. For example, ten Brinke & Porter (2012) examined videos of people pleading for the safe return of missing relatives. In some of those cases, the pleader was later revealed to have murdered the missing person. In this high-stakes scenario, dishonest murderers showed signs of emotional leakage, which could be used to reliably predict their dishonesty.

In addition to emotional leakage, cues of dishonesty can be found in observers’ reactions. Compared to truthful pleas for the return of a lost loved one, the pleas of deceptive murderers elicits stronger physiological arousal in observers. Some observers are aware of their heightened physiological arousal; however, this does not translate to accuracy in detecting deception. This suggests that people are (perhaps subconsciously) aware of something “off” about the deceptive pleader, but are unwilling or unable to label that individual a liar. As such, implicit cues of deception (such as heightened physiological arousal of observers) may be more useful for detecting deception than asking people to explicitly label someone as dishonest.

#### 4.2.2.2 Instruments and Measures

The pleader videos will be sampled from the same pleader videos used by ten Brinke & Porter (2012). Each participant will view one video, which will be sampled randomly from the collection of videos. Half the participants will be shown a video of a deceptive murderer, and half will be shown video of a truthful pleader.

The study protocol questions will be:

1. Did you recognize the person in the video?
2. Were you at all suspicious that the pleader in the video may have been lying?

#### 4.2.2.3 Analytic Strategy

A χ2 test of independence will be conducted to determine whether people who viewed honest pleaders were more likely than those who viewed a deceptive pleader to donate their participation incentive to charity. That is, the independent variable will be the honesty of the pleader (truthful or deceptive), and the dependent variable will be the participants choice of participation incentive (donation to charity or kept for self). Significance will be inferred at α = .05.

#### 4.2.2.4 Sample Size and Power Analysis

Research using people’s willingness to make a hypothetical donation showed that participants were more willing to donate to honest pleaders, with a standardized effect size of d = 0.47 (which is ρ = .23 in the metric of a correlation coefficient). This effect is related but distinct from the effect of interest to this study. Unlike that one, this relies on real-world donations. Additionally, the donations are being made to a charity and not being given directly to the pleader in the video. Additionally, this is only a single study and may therefore not be an accurate estimate of the population effect size. However, it is the closest facsimile of the effect of interest that we are aware of in published literature. As such, our power analysis was based on its effect size.

Assuming a population effect size of w = .23, 95% power will be obtained with N = 245.6 participants. In prior research, 6.4% of participants were suspicious that some of the pleaders were being dishonest. Because we want to be able to drop suspicious participants, we are aiming for a total sample of N = 263. This will give yield 95% power even after dropping 6.4% of participants from the analyses.

### References

Gunderson, C. A., Baker, A., Pence, A. D., & ten Brinke, L. (2021). Interpersonal consequences of deceptive expressions of sadness. Personality and Social Psychology Bulletin. https://doi.org/10.1177/01461672211059700
ten Brinke, L., & Porter, S. (2012). Cry me a river: Identifying the behavioral consequences of extremely high-stakes interpersonal deception. Law and Human Behavior, 36(6), 469–477. https://doi.org/10.1037/h0093929