The use of personality testing to assess the quality of job applicants and employees has been an on-again, off-again practice -- with interest growing recently. But there was a reason the use declined once before.
No doubt as far back as the Stone Age, observers were struck by the performance differences among individuals: Some worked hard at everything and got a lot done, others did not.
The search to explain why was not very systematic until the field of personnel and industrial psychology began around World War I and dedicated itself to that task.
Employers typically described the most important factor leading to different performance outcomes as "character," but researchers would later begin to refer to it as personality -- a pattern of responses associated with an individual's disposition -- that persists across contexts and over time.
Systematic efforts to assess job applicants for hiring, and then internal candidates for promotion and advancement, using personality flourished in the 1950s. Companies such as Sears, Roebuck and Standard Oil of California pioneered these approaches. Some, like Procter and Gamble, used it to predict who would get to the top ranks of the company when hiring entry-level employees.
What did companies look at to predict who would succeed as a manager?
Rather than assessing work-related competencies, employers used interviews to assess character with questions about their family life, their extracurricular activities and their general life experiences.
General Electric became one of the leaders in what became known as "trait-rating" psychological scales for assessing individual candidates for managerial development. Personnel psychologists, working for employers and in universities, were actively involved in designing these tests.
By 1954, 63 percent of large corporations were using standardized tests of personality for hiring decisions, and twenty-five percent of the companies used them as part of their promotion process to assess potential for leadership positions.
But the psychologists studying employee selection began to have their doubts as to whether personality truly did predict much about future job performance.
By the early-1960s, the consensus among researchers was that personality was not a useful criterion for assessing individuals. It's not clear whether employers paid much attention to the researchers, at least initially. But within a decade or so, personality-based assessments had largely disappeared from the lists of "best practices" in human resources.
Given this, it was something of a surprise to outsiders ? at least to me ? to discover the resurgence of interest and research on personality as a predictor of job performance that took place in the 1990s.
In brief, the resurgence was based on new techniques, which concluded that the earlier studies discrediting personality were not as persuasive as they appeared at the time.
These reinterpretations were followed by a veritable explosion of new research showing relationships between personality and various aspects of job performance. (The aspect of personality most consistently associated with performance, by the way, is "conscientiousness.")
A host of new and existing vendors around the same time began to offer new personality tests that were said to be important predictors of a range of job outcomes. To what extent these new offerings were spawned by the new research isn't entirely clear.
They clearly tap into the long-standing employer belief in the role of "character" and may also have been pulled along by the popular business literature about the importance of organizational culture and having employees who "fit" with the values of the organization.
It was particularly interesting given this resurgence of research to read conclusions about the usefulness of personality measures in Personnel Psychology from a panel of five prominent personnel psychologists, all former editors of research journals where the new personality research had been published.
After reviewing all the research, they conclude that personality explains so little about actual job outcomes that we should think carefully about using it at all for employment decisions, according to "Reconsidering the Use of Personality Tests in Personnel Selection Contexts," authored by Frederick P. Morgeson, Michael A. Campion, Robert L. Diboye, John R. Hollenback, Kevin Murphy and Neal Schmitt.
The least valid of the personality measures are the ones most employers are likely to use: published tests that individual candidates complete themselves. In addition to their low overall predictive power, the other concern about these tests, expressed by some of the editors more strongly, is that it may be easy for individuals to "fake" their responses.
A cynic might look at this assessment, written in the fall of 2007, and observe that it is essentially the same as the assessments of personality research from the early 1960s: Does scientific knowledge progress at all?
If it wasn't useful in the 1960s and isn't useful now, why was there a resurgence of interest in it?
In fairness to the researchers, their interest is, at least in part, in understanding whether relationships exist at all and not just whether those relationships are big enough to be practical and useful.
It does seem, however, that an awful lot of research firepower was directed over this period at an old target that was not very promising to begin with, especially given all the new developments in the world of work that are under explored.
For practitioners, however, the conclusions from these editors are remarkably useful and unambiguous: Don't rely on personality to assess employees.