Using Blatant Extreme Responding for Detecting Faking in High-stakes Selection: Construct validity, relationship with general mental ability, and subgroup differences
Autor: | Erica Hauck, Jeff A. Weekley, Nicolas Roulin, Julia Levashina |
---|---|
Rok vydání: | 2014 |
Předmět: |
Strategy and Management
media_common.quotation_subject Ethnic group Personnel selection Construct validity Context (language use) Cognition General Business Management and Accounting Management of Technology and Innovation Scale (social sciences) Personality Big Five personality traits Psychology Social psychology General Psychology Applied Psychology media_common |
Zdroj: | International Journal of Selection and Assessment. 22:371-383 |
ISSN: | 0965-075X |
DOI: | 10.1111/ijsa.12084 |
Popis: | Although there has been a steady growth in research and use of self-report measures of personality in the last 20 years, faking in personality testing remains as a major concern. Blatant extreme responding (BER), which includes endorsing desirable extreme responses (i.e., 1 and 5 s), has recently been identified as a potential faking detection technique. In a large-scale (N = 358,033), high-stakes selection context, we investigate the construct validity of BER, the extent to which BER relates to general mental ability, and the extent to which BER differs across jobs, gender, and ethnic groups. We find that BER reflects applicant faking by showing that BER relates to a more established measure of faking, an unlikely virtue (UV) scale, and that applicants score higher than incumbents on BER. BER is (slightly) positively related to general mental ability whereas UV is negatively related to it. Applicants for managerial positions score slightly higher on BER than applicants for nonmanagerial positions. In addition, there was no gender or racial differences on BER. The implications of these findings for detecting faking in personnel selection are delineated. |
Databáze: | OpenAIRE |
Externí odkaz: | |
Nepřihlášeným uživatelům se plný text nezobrazuje | K zobrazení výsledku je třeba se přihlásit. |