Personality inventories aren't good enough

The use of personality inventories in high-risk decision-making (such as employee selection) has always been controversial. In the article by Guion & Guttier (1965), there are reservations related to theories describing connections between personality dimensions and jobs, the quality of personality measurement, and especially - the stability and strength of relationships between personality dimensions and outcomes in the workplace.

 

 

 

In the 1990s, three changes occurred that brought personality assessment in the workplace and employee selection in particular back to center stage:

 1. Growing acceptance of the Big Five model (BIG FIVE) as a comprehensive model of normal personality;

 2. Several meta-analyses showed stable (albeit weak) connections between personality and outcomes in the workplace;

 3. Development in personality measurement tools (especially of the Big Five). As a result, the use of personality tests in employee selection processes (and in general) began to gain renewed momentum.

But the increasing use of personality tests in organizational contexts is not without criticism. The main criticism is that in these tests, people tend to describe their personality in a socially desirable way when they believe that the test results will be used for important decision-making. Studies have shown that examinees can indeed bias their results in personality tests, and that these biases affect the validity of personality tests and the decisions made based on them. It's important to note that there are other studies that have shown that tools can be developed to prevent cheating on personality tests.

 

 

The approach of organizational-occupational psychologists toward personality tests has changed greatly since the 1950s. Studies from the 1990s show evidence for the validity of at least some of the Big Five dimensions for job performance, albeit with relatively low levels of validity. It seems that psychologists' approach has changed significantly, even though the claim regarding the instability of relationships between personality dimensions and workplace outcomes has not yet been refuted. Validity levels found in studies stand at 0.06, and only one dimension among the Big Five (conscientiousness) was found to have a higher level of validity, but still a level categorized as low (0.12). However, at least some of the sources of instability found are related to the quality of the studies conducted and not to the quality of the tests. As better measurements for the Big Five develop, the theoretical likelihood increases that psychologists will be able to choose better tools for personality assessment. Nevertheless, there is a high chance that Guion and Guttier's (1965) claim regarding the validity and usefulness of personality assessment tools is still relevant today.

 

 

There are three main reasons for concern regarding the use of personality tests in the context of employee selection or career counseling for the same reasons.

Currently, there is no clear and convincing theory describing how personality dimensions should be related (or not related) to performance in specific jobs. Several attempts have been made to present a theory linking broad personality dimensions to various jobs.

Studies have shown that personality dimensions may be relevant for predicting contextual performance versus task performance, but this relationship does not help match personality traits to different jobs. Since personality dimensions do not show the same high and consistent predictive validity as personality tests, it may be necessary to examine which personality traits are relevant for which jobs.

The relationships of ability tests to job performance are monotonic relationships, meaning they have a uniform nature (as ability increases, job performance increases), but personality tests do not show such a pattern. For example, a manager with an unpleasant character is likely to have a low level of performance; as the manager's pleasantness increases, so is likely to increase their level of performance. But this relationship does not continue monotonically. From a certain level of pleasantness, it is likely that the quality of the manager's performance will decrease, as it will be difficult for them to perform "unpleasant" tasks, such as negative feedback, bad news, etc.

Beyond the complexity of the relationship presented above, the relationship between personality and performance may vary between roles, organizations, and situations. For example, when there are clear behavioral norms, it is likely that each individual's personality will have less impact on performance than in situations where there are few cues about behavior appropriate to the situation.

Countless attempts have been made to develop a method for determining personality traits that contribute to the quality of performance in different jobs. Tett & Burnett (2003) proposed a complex model in which situational factors influence tasks, social level, and organizational level in the job, and can amplify or reduce the impact of different personality traits. That is, it may not be the job itself that determines which personality traits contribute to good job performance, but other factors.

Studies have found that organizations typically use personality tests that are not designed to measure the Big Five. Tests such as 16PF and CPI are tests with a good psychometric level, but the results obtained from these tests do not align with the Big Five dimensions, and the relationships between the dimensions obtained from these tests are not found to have significant relationships with job performance. The MMPI, which is a questionnaire originally designed to diagnose psychopathologies, is also found to be in extensive use, but evidence for its predictive validity is weak. The MBTI questionnaire is very popular, but it is unclear what the test measures, and there is no evidence of links to various organizational behaviors.

Integrity tests have the best criterion validity and are the most useful in employee selection processes, but these are usually a group of measures that "work" well, but no one is clear why. Integrity tests with good predictive validity are usually found to have substantial or complete overlap with the Big Five dimensions.

In summary, there is a large gap between the research literature on the predictive validity of personality tests for job performance and their common use in practice, which places little emphasis on the validity of personality tests in the context of employee selection.

 

 

There are several disputes regarding the use of ability tests in employee selection, but ability tests are perceived by most organizational-occupational psychologists as valid, accurate tests with a reasonable level of predictive validity for job performance.

There are three main differences between ability tests and personality tests, which make it relatively easy to predict job performance using ability tests, but more difficult to predict it using personality tests.

The key to understanding the structure of cognitive abilities is the common finding that results of different ability tests are positively correlated with each other, a phenomenon called "positive manifold." This finding has three important implications:

a. The specific content of the ability test is not important. Two ability tests that have no common content will still have a high positive correlation.

b. It is easy to build a reliable ability test. Positive manifold means that different items will have a high positive correlation with each other, which makes it relatively easy to build reliable tests.

c. Since most job tasks require information processing, similar to ability tests, it is likely that any reliable ability test will correlate with measures of job performance.

In contrast, the structure of personality does not easily allow for building a reliable and valid test for predicting job performance. The Big Five have weak relationships with each other, so one personality measure has a weak or zero correlation with another. That is, the content of the test is very significant in building the test. In addition, a test measuring the "wrong" personality dimension will show very low or zero validity with job performance. Therefore, if the job requires social extroversion, and the test measures niceness and emotional stability, one might (mistakenly) conclude that personality tests have no validity for predicting job performance.

Ability tests typically ask a person to do something, while personality tests typically ask a person to describe themselves. Self-report tests of this type are subject to much criticism regarding the usefulness and accuracy of the information provided. In contrast, ability tests are more relevant to employee selection. The task required of the examinees is similar to what the test is designed to measure. To succeed in the test, the examinee is required to solve problems, manipulate information, complete tasks, and so on. Success in work also requires many of these things. In contrast, there are very significant differences between the task required in completing a personality questionnaire (telling about yourself) and the tasks required in the world of work.

A second and important difference between ability tests and personality tests stems from the research tradition accompanying each type. Ability tests are usually based on a psychometric tradition, while personality tests were often built from clinical experiences, where the emphasis is on understanding the person and not comparing examinees.

Cognitive ability tests are usually subject to rigorous criticism by potential users. First, the tests are used for high-risk decision-making (such as employee selection and university admission). Second, the results of the tests may have implications of an ethnic and community nature, such as the chance of being accepted for a job. Because of this, ability tests are at the heart of claims of discrimination in organizations, academic acceptance, and more.

Personality tests are usually not under criticism regarding their reliability and validity. They are usually not used as tools for high-risk decision-making, the likelihood that they will lead to ethnic or community discrimination is lower than ability tests, and therefore the psychometric sophistication required of them is lower.

 

Personality tests do not have a high level of predictive validity for job performance as ability tests do, and it is not likely that personality tests can present such a level. The problem is not necessarily the personality tests themselves, but the personality dimensions measured, the measurement strategies chosen, and the level of accountability associated with these tests.