What test instrument can quickly assess a candidate’s personality preferences; is cheap; available to almost everyone; marketed by dozens of vendors under a variety of names; and, is recommended “unreliable and untrustworthy” by most testing professionals? Yes, there are others, but I was referring to the DISC.
(Although this article focused on the DISC, you could apply its comments to any test, assessment, interview, exercise, role-play, and so forth used to separate qualified from unqualified applicants.)
Some Background
DISC development began in the early 1900s when the Army asked psychologist William Marston to investigate why different soldiers who received the same training behaved differently. He published a report about 10 years later entitled “Emotions of Normal People.”
As far as we know, Marston’s objective was to describe “mental energy”… not assess and classify people for a job. Shortly afterward, another psychologist used Marston’s theory to develop a pencil-and-paper test. It asked people to choose between pairs of adjectives (i.e., which adjective is most like you and which adjective is least like you); then it added items together and reported scores for dominance, extraversion, need for security, and need for structure. (Remember that Marston was NOT trying to hire the most qualified people for a job — just explain normal behavior.)
So far, so good.
After that, the DISC grew in popularity and became another chapter in misused pseudoscience. Vendors and users alike tested everything making wild predictions about future performance. I have not verified it, but there were even some rumors about separate DISC profiles developed for inanimate objects, barnyard livestock, and certain vegetables. What happened? In short, as one great philosopher said, “When the only tool you know how to use is hammer, every problem looks like a nail.” That is, to the uninformed DISC user, every hiring problem looked like a DISC profile.
Test Construction
There are four things that define a good hiring test (i.e., one with scores you can use and trust):
- It is well-constructed (e.g., all individual items supposed to measure expressiveness, actually measure expressiveness … and not dominance, security, or structure). This is termed inter-item reliability.
- It delivers the same results over time (e.g., for the same person, November 2008 scores are the same as November 2007). This is termed test-retest reliability.
- Scores predict performance (e.g., all high performers have similar profiles and all low performers have similar profiles). This is termed criterion validity.
- It incorporates a theory of job performance (e.g., you are trying to predict job performance, right?). This is termed knowing your butt from your elbow.
Just common sense, right?
Overall, the idea of checkingoff descriptive adjectives seems like a good way to demonstrate preferences, but this design has some serious limitations. Consider, for example, the following items taken from one version of the DISC: Fussy, Self-reliant, Persistent, Optimistic, God-fearing, Devout, and Moderate. Can you clearly identify to which of the four factors they relate? Neither can I. Fuzzy items lead to fuzzy results.
In another case, a DISC vendor proudly published on the Web a study it claimed proved the VALIDITY of its DISC (i.e., if you recall, validity proves test scores actually correlate with job performance). Unfortunately, the vendor must have been disoriented because he only reported inter-item reliability — not validity — leaving the reader wondering if the vendor either successfully completed a class in psychometrics or read the Standards for Educational and Psychological Testing (both are requirements for test professionals).
Personality Scores and Job Performance
Are scores on personality tests highly predictive of performance? There is a long history of research showing: maybe. That is, if the employee has all the other performance elements necessary to do the job, and the personality test is job-related, personality can make a difference. Self-descriptive test scores represent how the applicant wants to present themselves — it may not be reality. So, even if an applicant tells an interviewer he/she is organized, it’s no guarantee he or she will be good with details. It is just bad science to claim the DISC (or any other personality test) will accurately predict managerial performance, capability for organization, character, or personal responsibility.
What about the four DISC factors? Remember that Marston never intended the DISC be used for employee selection? Decades of personality research shows only factors consistently relate to job performance: being extraverted, not being neurotic, and being conscientious. Skills-based research, on the other hand, shows that intelligence is the greatest predictor of ability (e.g., people who are job-smart outperform those who are not). I don’t know about you, but it seems the DISC falls short on both. But, if you have been using the DISC as a hiring tool for some time, and rigorously monitor your results, you probably already know that.
What about the most-least scoring system? Well, that leaves something to be desired too. You see, if there are eight expressive adjectives and I check-off four … but you check-off a different set of four expressive adjectives … have we taken the same test?
The same goes for comparing expressive adjectives with ones for dominance, security, and structure. Ipsative scoring (as it is termed) tells us a great deal about individual preferences, but it is generally discouraged for hiring by test professionals.
So What Good Is the DISC?
If you need some help understanding broad differences between people, and you are willing to take a test never intended for serious business use, then the DISC is a fun tool. If you understand that when you describe yourself as primarily dominant (or any other factor), your scores will probably show you are dominant (amazing!); or, when you want to facilitate a quickie workshop in communication styles, the DISC is for you.
Always remember, the vast majority of personality tests were designed to evaluate differences between people — not evaluate people for jobs.