Gender Differences in Personality: An Item-Level Analysis

Personality is a social and organizational construct with a substantial history and discourse. One particular area in personality that is of interest is gender differences in personality. Gender differences have been found on scales measuring various aspects of personality, such as narcissism (Grijalva et al., 2014). While there are differences present in personality data, there hasn’t been a consistent explanation for why this occurs. This research looked specifically at the construction of personality items to begin to understand the differences in personality by gender. While social roles and social context are mostly referenced to inform the response patterns of men and women, this research looked to gather greater insight into the impact of social roles and stereotypes on personality items and if the adjustment of the item content can result in the reduction of these differences. To examine these questions, two studies were conducted. The first study explored if certain personality items function differently between men and women. Furthermore, it looked to answer if it is possible to predict which items may show DIF (differential item functioning) by assessing the item for stereotype content. The second study focused on understanding personality item content and the impact of stereotypes through an experimental lens by manipulating personality items to reflect or remove stereotypes and if this would influence the endorsement of the item by gender. The results for the first study indicated that a small number of IPIP personality items showed DIF. Moreover, a majority of the items that did show DIF were coded to possess stereotype content. There was mixed support regarding endorsement by gender, with the clearest finding that agentic and competent items showing DIF were more likely to be endorsed by men compared to communal, not competent, or neutral items. The results for study 2 showed mixed results as well. The masculine and feminine-written items resulted in the most significant interactions between stereotype, item content, and gender, whereas, the items reflecting the other stereotypes (e.g., agentic, communal, warm, competent) showed fewer statistically significant interactions. The pattern of results in both studies present an opportunity to assess how we measure personality and construct personality items. Psychometric and practical implications are discussed