Speaker
Abstract
Background. General Factor of Personality (GFP) is a higher-order factor consistently found in personality inventories, explaining correlations between all personality traits in the socially desirable direction. The long-standing controversy about GFP is whether it is a real thing (individual differences can really be reduced to a single continuum from "bad personality" to "good personality"), or it is an artefact of response biases, most notably socially desirable responding.
I argue that this controversy cannot be resolved while using normative response formats easily lending themselves to socially desirable responding. Instead, researchers could measure personality with ipsative (or relative to self) response formats that force respondents to choose among statements that describe their personality better or worse. As it is no longer possible to endorse all desirable statements, ipsative formats prevent socially desirable responding and thus are better suited to study GFP. The use of Thurstonian Item Response Model (TIRT, Brown & Maydeu-Olivares, 2011) further ensures that the scale scores extracted from such questionnaires are free from ipsative constraints.
Method. This research investigated the construct validity of GFP in two studies, using the same personality inventory - the Customer Contact Styles Questionnaire (CCSQ published by SHL) but different external measures for construct validation. The CCSQ consists of 128 items arranged in 32 blocks of 4 items, where participants must rate each statement using a 5-point Likert scale and also select one statement which describes them most and one that describes them least. Therefore, CCSQ combines both ipsative and normative response formats and allows comparison between them.
Study 1 recruited N=279 undergraduate psychology students to complete CCSQ under a research condition. GFP extracted from both response types was validated using the Balanced Inventory of Desirable Responding (BIDR), the Situational Test of Emotional Management (STEM), and the Geneva Emotion Recognition Test (GERT).
Study 2 recruited N=219 call centre employees as part of a validation study. GFP extracted from both response formats was validated using the incentive bonus that the employees received according to their performance.
Results. In Study 1, GFP accounted for 24.9% and 13.6% of trait variance in normative and ipsative data, respectively. Normative GFP correlated with self-reported BIDR scales socially deceptive enhancement (r=.29) and impression management (r=.19) only. Ipsative GFP correlated with BIDR impression management (r=.31), GERT and STEM (r=.24 and r=0.20).
In Study 2, GFP accounted for 32.9% and 12.6% trait variance in normative and ipsative data, respectively. Both normative and ipsative GFPs correlated with the incentive bonus (r=.30 and r=.29, respectively).
However, format-specific GFPs did not correlate with each other in Study 1 (r=.01) and correlated only moderately in Study 2 (r=.20), suggesting distinct constructs.
Discussion. Past research has related GFP to social effectiveness by finding large overlaps with the measures of trait emotional intelligence and assessment centre ratings. Using a personality inventory that combines both the normative and ipsative formats, this research found that GFP demonstrated attributes of both – a substantive trait and a method-related artefact.
| Oral presentation | What ipsative measures can tell us about the General Factor of Personality |
|---|---|
| Author | Anna Brown |
| Affiliation | University of Kent |
| Keywords | GFP, ipsative, social desirability |
