Hello from the Other Side: Can We Perceive Others’ Darkness? Observers’ Accuracy of the Dark Triad

Abstract

The current study builds on research concerning self-other agreement, specifically the Trait-Reputation-Identity (TRI) Model, and combines this theoretical approach and recent advances from multitrait-multimethod research, specifically the CTC(M-1) and the latent difference model. This combination was applied to avoid statistical problems associated with latent trait models of the type previously used in TRI Model research. A further aim was to fully exploit the TRI Model’s potential by simultaneously modeling more than one trait in one psychometric model. This makes it possible to explore whether raters use similar information to assess different traits or whether this information is observer-specific, as well as whether the traits themselves can shape an individual's Reputation. All analyses were based on a data set capturing the Dark Triad. As the Dark Triad has only rarely been examined from different rater perspectives before, the study also provides new insights into this network of maladaptive traits. A sample of 290 students, their best friends, and one parent were asked to fill out the NPI, Mach IV, and SRP-III. The results suggest that accuracy is high only for narcissism. Moreover, a dark halo could be observed among each of the other-raters for Machiavellianism and psychopathy; target’s standings on these two traits were also underestimated by the other-raters. The study’s combined use of the TRI Model and modern structural equation methods highlights the usefulness of the TRI Model and adds to the debate about the specific nature of Dark Triad traits.

Notes

1 Raters who are structurally different differ with respect to their perspective on and available information about the trait. Psychometrically, different raters (here targets, friends, and parents) do not belong to the same population of raters. In this case, it is most meaningful to contrast methods with one another, choose one method as the reference method (Eid et. al, 2008), and model either a correlated trait-correlated method minus 1 (CTC(M-1)) model (Eid et al., 2003) or a Latent Difference (LD) model (Pohl, Steyer, & Kraus, 2008).