Do Warnings of Response Verification Moderate the Relationship Between Personality and Cognitive Ability? Nicholas L. Vasilopoulos and Jeffrey M. Cucina The George Washington University Julia M. McElreath U. S. Customs and Border Protection This article presents the results of 2 studies that examine the effect of presenting a warning of response verification by others on the relationship between measures of personality and cognitive ability. In both a field study and a laboratory study, stronger correlations were generally found between measures of personality and cognitive ability when a warning of verification was present. The results from the studies also showed that the warning of verification was associated with slower item response latencies for certain personality scales, suggesting that exposure to the warning increased the complexity of the response decision. Results are discussed in terms of the effects of altering test instructions on the construct validity of personality measures used in applied settings. For more than a decade, personnel psychologists have expressed considerable interest in personality testing (Behling, 1998; Mount & Barrick, 1998). Much of the enthusiasm for personality testing can be attributed to meta-analytic results suggesting that meaning- ful relationships exist between measures of personality and indices of job performance (Barrick & Mount, 1991; Barrick, Mount, & Judge, 2001; Hurtz & Donovan, 2000; Salgado, 1997; Tett, Jack- son, & Rothstein, 1991). Also influential were the results from primary studies showing that many of the personality measures predictive of job performance are unrelated to measures of cogni- tive ability (Costa & McCrae, 1992b; McHenry, Hough, Toquam, Hanson, & Ashworth, 1990; Rosse, Miller, & Barnes, 1991). Collectively, these findings have been interpreted as evidence that personality dimensions such as conscientiousness can predict per- formance in domains that are not predicted by cognitive ability (Schmidt & Hunter, 1998). Despite this conclusion, some research- ers have questioned whether applicant faking compromises the utility of a personality test used for personnel selection (Snell, Sydell, & Lueke, 1999). One method used to address applicant faking is a score correc- tion on the basis of a social desirability scale embedded in the personality inventory. Score corrections represent a reactive ap- proach to faking because the inclusion of a social desirability scale is not expected to alter the item response process. Instead, correc- tions are made after the inventory is completed in an attempt to remove social desirability bias from an applicant’s score. The effectiveness of a score correction appears to depend on its pur- pose. Score corrections have little impact on the criterion-related and construct validity of personality measures (Barrick & Mount, 1996; Ellingson, Sackett, & Hough, 1999; Ellingson, Smith, & Sackett, 2001; Hough, 1998; Hough, Eaton, Dunnette, Kamp & McCloy, 1990; Ones & Viswesvaran, 1998a; Ones, Viswesvaran, & Reiss, 1996; Smith & Ellingson, 2002). However, score correc- tions do limit the number of fakers at the top of the score distri- bution. This reduces the chance that a faker is hired when selection decisions are made top-down (Christiansen, Goffin, Johnston, & Rothstein, 1994; Rosse, Stecher, Levin, & Miller, 1998; Stark, Chernyshenko, Chan, Lee, & Drasgow, 2001). Other methods used to address applicant faking represent pro- active approaches in that they are built into the personality inven- tory in an attempt to either dissuade applicants from faking or make it difficult for applicants to fake effectively. Examples of proactive approaches include forced-choice item formats (Jackson, Wroblewski, & Ashton, 2000) and warnings of response verifica- tion (Dwight & Donovan, 2003). Although proactive approaches seem reasonable, they involve a fundamental change that may alter the relationship between the personality measure and other mea- sures used in personnel selection. Accordingly, two studies were conducted to explore whether the relationship between scores on measures of personality and cognitive ability are moderated by a warning of response verification by others. The existence of moderators of the personality-cognitive ability relationship is suggested by the results of two meta-analytic re- views. Ackerman and Heggestad (1997) found significant corre- lations between general cognitive ability and the Big Five person- ality dimensions of Extroversion ( = .08), Neuroticism ( = -.15), and Openness to Experience ( = .33). The heterogeneous distribution of correlations for Conscientiousness, Extroversion, Neuroticism, and Openness to Experience (indicated by significant Q values; see Hedges & Olkin, 1985) suggests that the relationship between cognitive ability and these factors is moderated by a third variable. More recently, Cortina, Goldstein, Payne, Davison, and Gilliland (2000) reported a true-score correlation between cogni- tive ability and conscientiousness (.08). However, only 22% of the variance in the distribution of correlations was attributed to sam- pling error and range restriction, well below the 75% threshold for detecting moderators proposed by Schmidt and Hunter (1977). Nicholas L. Vasilopoulos and Jeffrey M. Cucina, The George Wash- ington University, Washington, District of Columbia; Julia M. McElreath, U.S. Customs and Border Protection, Washington, District of Columbia. We thank Rebecca Goldenberg and Adam Malamut for their assistance in collecting the data reported in Study 2. Correspondence concerning this article should be addressed to Nicholas L. Vasilopoulos, The George Washington University, Department of Or- ganizational Sciences and Communication, 2147 F Street Northwest, Washington, DC 20052-0001. E-mail: nlv@gwu.edu Journal of Applied Psychology Copyright 2005 by the American Psychological Association 2005, Vol. 90, No. 2, 306 –322 0021-9010/05/$12.00 DOI: 10.1037/0021-9010.90.2.306 306