Personality assessment is the process of gathering information about an individual to make inferences about personal characteristics including thoughts, feelings, and behaviors. Raymond B. Cattell identified three primary sources of obtaining such personality information: life-data, information collected from objective records; test-data, information obtained in constructed situations where a person’s behavior can be observed and objectively scored; and questionnaire-data, or information from self-report questionnaires. Each type of data is used to make assessments of personality within contemporary industrial-organizational psychology (I/O psychology). Common forms of life-data might include information contained in a resume or an application blank and examinations of court, financial, or driving records in background checks. Test-data would include scores on personality-based dimensions derived from the assessment center method. However, by far the most common form of personality data in I/O psychology is questionnaire-data.
Self-report measures of personality can be divided into two broad categories: clinical and nonclinical. Self-report clinical measures, such as the Minnesota Multiphasic Personality Inventory (MMPI), have been used for some workers, such as airline pilots and police officers, to ensure that a potential employee does not suffer from an underlying psychological disorder. These clinical measures are generally given along with an interview (life-data) in the context of an individual assessment. Decisions to use these clinical evaluations for personnel selection, however, must be made carefully, because there is a notable possibility of violating the Americans With Disabilities Act.
Nonclinical self-report personality assessments, which are much more widely used than clinical assessments, are typically designed to assess personality traits. Personality traits are characteristic behaviors, thoughts, and feelings of an individual that tend to occur across diverse situations and are relatively stable over time. A trait that has been particularly important in the context of I/O psychology is conscientiousness, which is associated with a tendency to be organized, thorough, systematic and efficient. Assuming that these characteristics are desirable in an employee, a self-report questionnaire may be administered to make inferences about the conscientiousness of individuals within an applicant pool.
Content versus Empirical Scale Development
Although there are numerous approaches to constructing a personality assessment, two broad approaches can be identified. By far the more common approach to scale development is the content approach. In this approach items are written based on a theory of the construct the set of items is intended to measure. By way of example, an item such as I enjoy the company of others might be written for a sociability scale. Once written, the items are then typically empirically evaluated using principles of construct validation. As factor analysis is frequently used to evaluate items, this approach is also commonly referred to as the factor analytic approach.
As P. E. Meehl (1945) pointed out, however, interpreting an individual’s response to such an item requires certain assumptions. For example, it must be assumed that all respondents have interpreted the item in the same way, that people are aware of and can report their own behavior, and that people are willing to tell you about their behavior. Some personality measurement theorists felt these assumptions were untenable and suggested a different approach to personality test construction, the empirical keying approach. According to this approach, a personality item is useful to the extent that responses to it accurately differentiate two groups. For example, an item would be included on a depression scale if, and only if, depressed people responded to the item differently from nondepressed people. In the classic empirical keying approach the content of the item is irrelevant; whether the item appears theoretically related to the construct does not matter. Because the response to the item is considered to be the behavior of interest, interpretation of scores from an empirically keyed measure does not require the assumptions associated with the content approach. Although the empirical keying approach was the basis for such well-known measures as the MMPI and the California Psychological Inventory, the vast majority of personality assessments in use today are based on the content approach to scale development.
Normative Versus Ipsative Assessment
Most personality assessments given in I/O contexts provide normative scores. Normative scores result when the responses to one item are independent from responses to other items. The common Likert-type rating scale, in which the respondents use the scale to place themselves along the trait continuum as represented by a single item, will result in normative scale scores. Ipsative scores, in contrast, result from response formats in which respondents choose, rank order, or otherwise indicate preference among a set of statements presented in an item. The Myers-Briggs Type Indicator is a well-known measure that provides ipsative scores.
Normative and ipsative scores result in different inferences about a person’s trait standing. Normative scores allow for inferences regarding the amount of a trait that an individual possesses compared with other people. Ipsative scores, in contrast, support inferences about the amount of a trait possessed by the individual compared with the other traits assessed by the measure. Thus, a high score on a particular scale in an ipsative measure does not suggest that the respondent has a high standing on that trait, but suggests rather that the respondent has a higher standing on that trait than on any of the other traits assessed by the measure. Ipsative measures, therefore, are useful for identifying a person’s particular strengths and weaknesses (i.e., intraindividual differences) and may be particularly useful in vocational guidance contexts. In many I/O contexts, however, the explicit desire is to compare the scores from a set of people (i.e., interindividual differences), as in personnel selection. When comparing people is the goal, ipsative scores are inappropriate and such measures should not be used.
Origins of Personality Assessment in I/O Psychology
Applications of personality assessment within I/O psychology began as early as 1915 with the creation of the Division of Applied Psychology and the Bureau of Salesmanship Research at the Carnegie Institute of Technology. In addition to developing technologies for the selection of salesmen, this group of researchers also sought to develop measures of personality (or temperament/character as it was referred to at that time). Personality assessment gained further acceptance during World War I when United States military researchers developed the Woodworth Personal Data Sheet to identify individuals who might be susceptible to war neuroses. With the development of several multitrait assessment tools, the popularity of personality testing grew through the 1940s and 1950s. For example, a survey of more than 600 American companies conducted in 1953 indicated that nearly 40% of those companies used measures of personality or vocational interests in their selection systems.
Three factors led to a marked decline in the popularity of personality testing in applied contexts during the 1960s and 1970s. First, two influential literature reviews were published that suggested that there was little evidence for the criterion-related validities of personality measures for the prediction of job performance. Second, the trait-situation debate dominated personality psychology over this period of time. On the situationist side of the debate, led by Walter Mischel, it was argued that aspects of the situation, not personal characteristics, were the driving force behind behavior. Third, Title VII of the Civil Rights Act of 1964 brought increased legal responsibilities to the use of assessments in the context of personnel decisions. Based largely on these factors, many organizations decided to forgo personality assessments in their selection systems, opting to avoid possible legal issues resulting from the administration of these tests.
Personality testing was given new life in applied contexts during the 1980s and early 1990s. Personality theorists finally found their voice in the trait-situation debate, effectively arguing that personal characteristics can predict behavior. The heart of the argument was the principle of aggregation, which suggests personality generally does not predict single instances of behavior well, but it does predict lawful patterns in behavior across diverse situations. But the biggest boon to personality assessment in I/O contexts was the emergence of the Big Five and subsequent meta-analyses demonstrating the criterion-related validities of some of these broad traits.
Criterion-Related Validity and Utility of Personality Assessment
One reason for the lack of strong criterion-related validity findings for personality assessments in the 1950s was the broadside approach taken by researchers. This tendency to correlate every available personality test score with all available performance measures was said to have resulted in large numbers of small criterion-related validity coefficients, many of which would have been expected, on the basis of theory, to be small. With the emergence of the Big Five trait taxonomy in the 1980s, conceptual links between the traits and the criterion variables could be drawn. A result of this better predictor-criterion linkage was stronger evidence for the criterion-related validities of personality assessments. To date, numerous meta-analyses on the relationships between personality test scores and measures of work performance have resulted in positive findings. The strongest findings have been associated with the conscientiousness trait, which seems to be associated with most job-related criteria (i.e., performance, training, attendance, etc.) across almost all jobs. However, the criterion-related validities remain modest, even after the corrections typically employed in meta-analytic procedures. For example, one of the most widely cited meta-analyses reported corrected criterion-related validities for conscientiousness in the range of .20 to .22 across performance criteria and occupational groups.
Many personality assessments frequently used in I/O settings were not created explicitly for applied use. That is, the questionnaires were created to provide a general assessment of personality; and as such, the items in these measures tend to be very general and do not typically convey information about any specific situational context. When responding to such a contextual items, respondents may consider their behaviors across a wide range of social situations, such as at home with family, at a gathering with friends, at a public event, or at work. Research has found, however, that when the item content was contextualized in a work setting, for example by adding the phrase at work to the end of each item, the criterion-related validity of the test was higher than when a contextual items were used. Thus, by including work-based situational cues within personality items, the criterion-related validity of personality scores can be enhanced.
Despite the improved validity associated with the contextualization of personality items, the criterion-related validity of personality assessments is clearly lower than that of many other available selection tools, such as ability tests, assessment centers, and work samples. Despite the lower criterion-related validities, personality assessment can still be of value in selection contexts. First, the correlations between personality test scores and scores from cognitive ability tests tend to be small, suggesting that personality tests can improve prediction of performance above and beyond cognitive ability test scores. Second, personality test scores tend not to show the large mean differences between racial groups that are found with cognitive ability tests. Third, these tests can often be administered quickly and typically are relatively inexpensive.
Impression Management and Faking
A major issue facing the application of personality assessment is the possibility of impression management, which is also known as socially desirable responding or faking. Impression management occurs when an individual changes a response to a personality item to create a positive impression. Consider a situation in which a person would, under normal circumstances, respond to the item I am a hard worker with a response of neutral on a five point Likert-type scale. If that same person were presented with the same item when applying for a job and responded with agree completely to increase the chances of being hired, then the individual would be engaging in impression management.
The precise effects of faking on the criterion-related validity of personality measures is still being debated, but it would appear that the effect is rather small. However, impression management does appear to negatively influence the quality of selection decisions. Although this may seem contradictory, it must be recognized that the validity coefficient takes into account the full range of personality test scores, whereas selection is generally concerned only with scores over a certain portion, usually the high end, of that distribution. In a top-down selection context, the quality of selection decisions appears to be negatively affected by the fact that a number of low per-forming people will rise to the top of the personality test distribution, increasing their chances of being selected. Researchers are currently examining the precise impact of faking on selection and are working on ways to deal with faking to maintain the usefulness of personality assessments.
Summary
Personality assessment is the process of gathering information about a person to make an inference about the individual’s characteristic ways of behaving. Although there are numerous methods for assessing personality, the most common form of assessment in I/O psychology is the self-report questionnaire. Meta-analyses have shown that these self-report measures can provide information that is valid for predicting various organizational outcomes. Further, that criterion validity may be enhanced by writing items that are contextualized in workplace settings. Finally, although personality assessments can provide useful information for making personnel decisions, intentional response distortion on the part of the respondent may lessen the usefulness of those scores in applicant contexts.
References:
- Barrick, M. R., & Mount, M. K. (1991). The Big Five personality dimensions and job performance: A meta-analysis. Personnel Psychology, 44, 1-26.
- Guion, R. M., & Gottier, R. F. (1965). Validity of personality measures in personnel selection. Personnel Psychology, 18, 135-164.
- Kanfer, R., Ackerman, P. L., Murtha, T., & Goff, M. (1995). Personality and intelligence in industrial organizational psychology. In D. H. Saklofske & M. Zeidner (Eds.), International handbook of personality and intelligence (pp. 577-602). New York: Plenum.
- Meehl, P. E. (1945). The dynamics of structured personality tests. Journal of Clinical Psychology, 1, 296-303.
- Mueller-Hanson, R., Heggestad, E. D., & Thornton, G. C., III. (2003). Faking and selection: Considering the use of personality from a select-in and a select-out perspective. Journal of Applied Psychology, 88, 348-355.
- Schmit, M. J., Ryan, A. M., Stierwalt, S. L., & Powell, A. B. (1995). Frame-of-reference effects on personality scale scores and criterion-related validity. Journal of Applied Psychology, 80, 607-620.