Mathematical Model of Person Judgment, Part 1 - Manuscript

We present the first part of a fully parametrized mathematical model of person judgment, in an attempt to streamline and better organize theory in this research area. The model integrates several key concepts from Kenny’s (1994) Weighted Average Model and Brunswik’s (1956) lens model with one another. It incorporates time as a continuous variable and also accounts for target-effects, situation-effects, target by situation interaction-effects, randomness overall and between-target differences in randomness. All of these can influence the average level of a substance variable that may be observed by perceivers („cues“) and thus inform their judgments of targets. We discuss the model’s implications with regard to reliability, inter-rater agreement, and predictive validity. We also explain how the model connects to – and sometimes helps specify – several important concepts from the person judgment literature.

Mathematical Model of Person Judgment: Part 1 (Cue Emergence)

We present the first part of a fully parametrized mathematical model of person judgment, in an attempt to streamline and better organize theory in this research area. The model integrates several key concepts from Kenny’s (1994) Weighted Average Model and Brunswik’s (1956) lens model with one another. It incorporates time as a continuous variable and also accounts for target-effects, situation-effects, target by situation interaction-effects, randomness overall and between-target differences in randomness. All of these can influence the average level of a substance variable that may be observed by perceivers („cues“) and thus inform their judgments of targets. We discuss the model’s implications with regard to reliability, inter-rater agreement, and predictive validity. We also explain how the model connects to – and sometimes helps specify – several important concepts from the person judgment literature.

How information processing style shapes people's algorithm adoption

Across four studies I tested why people are averse to relying on algorithmic judgments in person judgment tasks (e. g., student admissions), and examined how such aversions can be attenuated. I proposed that people tend to focus more on case-specific information (vs. general propositions) in person-judgment tasks, and that algorithms (vs. human experts) are believed to be skilled at addressing general propositions (vs. case-specific information). Thus, I posited that in person-judgment tasks, people would be less averse to relying on algorithmic judgments when they focus more on general propositions (vs. case-specific information). By varying the perceived importance of case-specific information and general propositions, the research provides support for these hypotheses. In addition, the results reveal the mechanism underlying algorithm aversion in person judgments and provide a cost-effective way to increase consumers' algorithm adoption.

There Is No Primacy Effect in Interpersonal Perception: A Series of Preregistered Analyses Using Judgments of Actual Behavior

First impressions are commonly assumed to be particularly important: Information about a person that we obtain early on may shape our overall impression of that person more strongly than information obtained later. In contrast to previous research, the present series of preregistered analyses uses actual person judgment data to investigate this so-called primacy effect: Perceivers ( N = 1,395) judged the videotaped behavior of target persons ( N = 200) in 10 different situations. Separate subsamples of about 200 perceivers each were used in moving from exploratory to increasingly confirmatory analyses. Contrary to our expectations, no primacy effect was found. Instead, judgments of the targets in later situations were more strongly associated with overall impressions, indicating an acquaintance effect. Relying on early information seems unreasonable when more comprehensive information is readily available. Early information may, however, affect perceivers’ behavioral reactions to the targets and thus their future interactions, if such interactions are possible.

From photograph to face-to-face: Brief interactions change person and personality judgments

Research into ‘first impressions’ frequently uses photographs of faces as representations of unknown others. This is the case for legitimate concerns about standardisation, sample size and experimental control. However, there is little work investigating the robustness of these first impressions through first interactions. A sample of 97 pairs of stranger participants (N= 194) completed personality measures (of Big Five and Triarchic Model of Psychopathy), made personality and social judgments of a photograph of the face of their to-be partner, then engaged in five minutes (maximum) of unstructured interaction with their partner and then made their judgments again. The behaviour of the participants in the interaction was coded using 76 criteria. Generally, before and after judgments were correlated, but significantly different at Time 2. Personality judgment accuracy at Time 1 was poor overall but at Time 2 participants showed self-other agreement on Neuroticism, Extraversion and psychopathic Boldness. At Time 1 participant ratings of confidence were more similar to negative valence but at Time 2 confidence was a positive attribute. Coded behaviours related to ‘engagement’ were those that influenced the person judgments the most, and these were related to Extraversion, Agreeableness and Boldness of participants. Overall, the results of this study show that first personality and person judgments change from photographs to face-to-face interaction. Person judgment research should be aware of the extent to which judgments of photographs relate to first interaction.

How many targets do you need for a person judgment study?

Person judgment studies require adequate sampling of both the targets of the judgments (K) and those who judge them (N). Despite person judgment studies being appraised in terms of both the size of N and K, there has been little research looking for the ‘right’ number of targets. The current study uses two datasets to investigate the effect of increasing K. Sample One studies judges’ (N=193) ratings of the threat posed by point light targets (K=23) and Sample Two investigates the accuracy of (N=792) judges’ Big Five judgments from targets’ (K=50) self-selected photographs. Accuracy values were created from randomly selected subsets of targets of increasing number (k=3, k=4, etc). Our results show that accuracy values stabilise at k≈10. We find evidence that effect sizes of sample performance increase with greater K. The results encourage, but do not mandate, as large as K as possible.