Storek, J.S.; (2011) The hubris and humility effect and the domain-masculine intelligence type: exploration of determinants of gender differences in self-estimation of ability. Doctoral thesis , UCL (University College London).

Preview

PDF 1331910.pdf Download (1MB)

Abstract

This thesis explores the potential determinants of gender differences in self-estimated intelligence. In particular, it addresses the determinants of gender differences in the ‘domain-masculine intelligence type’ that is expected to yield the most significant gender differences in the self-estimated intelligence model (SEI). Equally, it sets to confirm the occurrence of the ‘hubris-humility effect’ (HHE), i.e. male overestimation and female underestimation of cognitive abilities, specifically in the above intelligence type. The thesis contains eight chapters, ten correlational studies and five experimental studies. The thesis is divided in two sections. Section one contains the ten correlational studies and section two the five experimental studies. All studies are independent but related. Chapter one contains a review of the relevant literature. It is divided into three sub-sections: overview, intelligence and hubris-humility effect (HHE) and domain-masculine intelligence type (DMIQ): gender differences in self-estimated intelligence. Chapter two (Studies 1 and 2) introduces the domain-masculine intelligence type and demonstrates it is the most sensitive indicator of gender differences in the SEI model. HHE is shown to be the most pronounced and confined to occurring on DMIQ. Equally, gender is shown as the best predictor of DMIQ, over and above a number of other demographic variables. Chapter three (Studies 3 to 5) sets to validate the occurrence of HHE on DMIQ, while it introduces psychometric intelligence (‘g’) and implicit beliefs about intelligence as possible determinants of DMIQ. Studies 3 and 4 examine the role ‘g’, as measured by fluid (Gf) and crystallised (Gc) intelligence tests, play in DMIQ. Results confirm the occurrence of HHE on DMIQ and reveal significant gender differences in Gf and Gc, with medium and large effect sizes. Gender is shown to influence the relationship between ‘g’ and DMIQ. Contrary to prediction, a psychometric intelligence measure (Gf), and not gender, is the best predictor of DMIQ. Implicit beliefs about intelligence play no role in the prediction of DMIQ. Study 5 adds gender identity variables, i.e. masculinity and femininity, and self-construct measures, i.e. self-esteem and self-control, to Gf and Gc, as possible predictors of DMIQ. Results validate the existence of HHE on DMIQ and confirm gender as the best predictor DMIQ, over and above ‘g’, gender identity variables and self-construct measures. Chapter four (Studies 6 and 7) examines the role gender identity, i.e. masculinity and femininity, affect measures, i.e. positive and negative affect, and self-constructs, i.e. self-esteem and self-control, play as potential determinants of DMIQ. Both studies confirm the existence of HHE on DMIQ. Study 6 confirms gender as the best and only predictor of DMIQ. Study 7 affirms masculinity as the best predictor of the intelligence type, followed by gender. Chapter five (Studies 8 and 9) examines the role of culture in DMIQ and its impact on the existence of HHE on DMIQ. Gender identity variables are also included to validate the earlier findings and to explore the role masculinity plays as a predictor of DMIQ, in three distinct cultures. Study 8 was conducted in Czech Republic and Study 9 in Colombia and United Kingdom. Results confirm the occurrence of HHE on DMIQ in all three cultures, with medium effect size for the Czech sample and large effect sizes for the Colombian and British samples. Gender is shown to influence the relationship between gender identity variables and DMIQ. Contrary to prediction, masculinity and not gender, is the best predictor of DMIQ in the Czech Republic sample. In the Colombian sample, none of the entered variables significantly contributes to the prediction of DMIQ. In the British sample, gender is affirmed as the best predictor of DMIQ, followed by masculinity. The results suggest that culture influences the composition of DMIQ determinant(s). Chapter six (Study 10) explores the role of DMIQ in a precocious sample, i.e. members of Mensa UK. It also sets to validate the occurrence of HHE prevails on DMIQ in a population that is knowledgeable about intelligence as well as aware of its own intellectual superiority. Beliefs about intelligence and gender identity variables are also included to explore whether they will play a role in the prediction of the intelligence type. The results confirm the existence of HHE on DMIQ in this precocious population, providing additional evidence for the degree of embeddedness and impact of HHE on highly gifted individuals. Gender is confirmed as the only and best predictor of DMIQ. Chapter seven (Studies 11 to 15) contains five independent experimental studies. Study 14 was conducted with three independent samples to test three varying task-confidence conditions. The results of the three individual conditions are reported in the Appendix, while the combined total results are reported in Study 14. The five experiments consist of repeated measurement of DMIQ and a psychometric task (TCAP) that also includes task-success probability probes (TSP). Participants are asked to estimate DMIQ before and after the task. The task contains numerical, reasoning, and crystallised intelligence items as well as task-success or task-confidence probes. The number of the psychometric items and probes are manipulated per experiment to assess their impact on the results. As such, the task is expected to be gender-stereotype inducing. As in the correlational studies, HHE is predicted to occur in the pre- and posttask DMIQ conditions. Results of all five studies validate the existence of HHE on DMIQ1 and DMIQ2, with medium to very large effect sizes. Likewise, a significant decrease in the DMIQ estimates is observed in all five studies, with small to medium effect sizes. In addition, male advantage is confirmed on the psychometric task and the task-success probes. Gender differences in TCAP are observed in Studies 11, 12 and 15, with males correctly solving significantly more psychometric problems than females. Equally, gender differences in TSP occur in Studies 11, 12 and 13, with males providing significantly higher task-confidence answers than females. To validate the earlier results, gender is expected as the best predictor of DMIQ1 and DMIQ2. Results reveal that gender is the best predictor of DMIQ1 in three out of five studies and in two out of five studies in DMIQ2. Unexpectedly, task-success probes are twice the best predictor of DMIQ1 and three times the best predictor of DMIQ2. Moreover, gender influences the relationship between TPS and DMIQ1 and DMIQ2 in all five studies. Equally, gender influences the relationship between TCAP and DMIQ1 and DMIQ2, in all but one analysis. Surprisingly, the DMIQ1 and DMIQ2 estimates that are provided by participants in the three task-success probability groups, i.e. low, average and high, are startlingly accurate, with the exception of Study 14. That is, low DMIQ estimates are provided by participants with low task-success confidence, average estimates are provided by participants with average task-success confidence and the highest DMIQ estimates by individuals with highest task-success confidence. Results for TCAP are complex and less accurate. Yet, for both TSP and TCAP, males provide significantly higher DMIQ1 and DMIQ2 estimates than females, providing further evidence for the occurrence of male hubris in the self-estimation of ability process. Chapter eight presents a brief summary of results and conclusions of this research. Equally, limitations of this research are discussed and a number of future research recommendations provided. The appendix includes the three individual condition studies of Study 14; that is Studies 14A, 14B and 14C. The TCAP and TSP overviews for Studies 11 to 15 are also included. Finally, Study 16 that uses the combined sample made of the fifteen individual study samples (N = 2292) is integrated. Study 16 tests the main objectives of this thesis through previously used hypotheses and as such provides a summary overview of the results. All main objectives of this thesis are corroborated.

Download activity - last month

Export as JSONXMLCSV

Download activity - last 12 months

Export as CSVXMLJSON

Downloads by country - last 12 months

Export as XMLJSONCSV

Archive Staff Only

View Item