David McClelland and Competencies

David McClelland is often cited as the source or founder of the modern competency movement for his 1973 paper, Testing for competence rather than for intelligence.

In his paper, he argues that aptitude and intelligence tests are not all that valid. For example, many of Binet's original tests were based on exercises that teachers used in French schools, thus it is no surprise that they correlate highly with grades in schools. And when he researched the manual of the “Differential Aptitude Test” of the Psychological Corporation, almost every coefficient involved predicting grades in courses. He continues that researchers have had great difficulty in demonstrating that grades are related to any other behaviors of importance, except for doing well on aptitude tests. Thus, while graduating from high school and college opens up higher level jobs, the research has generally shown that students who did poor in school (as long as they passed) did just as well in life as the top students.

School grades seem to have no real power in predicting competence in real life outcomes, aside from the advantage that credentials convey on the individual concerned. The exception are special test scores, such as motor ability for vehicle operators or typing tests for clerks, in other words criterion sampling.

For example, to be a policeman in Boston, you have to take a three hour intelligence test that involves words that will never be used on the job. Rather, as McClelland argues, you should examine some of the real words that good policemen use on the job and then test for them (that is, if you think vocabulary is important for the job), and not what some supervisor or higher-up thinks they should use on the job.

McClelland also argues that tests should be designed to reflect changes in what people have learned. He writes that it is difficult, if not impossible to find a characteristic that cannot be modified by training and/or experience. For example, most intelligent test makers keep their scoring a secret so people will not learn how to do better the next time they test. If people did better on them then that means that they are not really measuring innate aptitude that are stable from one administration to the next. Faking a high score is impossible if you are performing a real criterion behavior, such as reading or driving a car. Faking becomes possible the more indirect the connection is between the test behavior and the criterion behavior.

At this point he goes on to what most competencies should try to measure — clusters of life outcomes. McClelland says that if you move towards criterion based job analysis, there is the danger that the tests will become extremely specific to the criterion involved. Thus, one could end up with hundreds or even thousands of specific tests for each job. It might be more useful to assess competencies that are more useful in “clusters of life outcomes.” This could include occupational, leadership, and interpersonal skills.

McClelland is mostly interested in the testing aspect. In addition, his paper is one of those more broad, easy to read papers as it is based on some of his lectures, thus most readers will find it quite easy to read. After reading it I could see why he would be considered one of the founders of the present competency movement. First, he moves away from general aptitude testing because it has no real impact. Rather, he drives towards criterion referenced testing. Now I'm not sure if a lot of competency-driven organizations have really thought about this, but once you have identified the competencies, you have to be able to test for them.

Secondly, he stays away from simply making them task-based as you would probably wind up with more tests than what would be practicable. I think what really makes competencies valuable to organizations is that rather than trying to identify an individual to fit each and every job, you find individuals that fit your organization (clusters of life).

For example, in a food manufacturing organization we had people who moved from the production line, to distribution operations, to supervisors, to supply chain analysts at our main headquarters. In an era where good people are getting harder to find, the organization that finds the best means to identify such people gives it an advantage over its competitors. Thus, perhaps this sudden interest in competencies.

Note that McClelland did cluster personality or traits into competencies, rather than separate them into attributes. He did not believe in the saying “once a bigot, always a bigot.” Rather he wrote that there is no solid evidence that this trait of any other trait cannot be changed. Thus, if you cannot find the people with all the competencies you need, you can always train or develop them (of course some competencies are a lot more easy to train or develop than others). In addition, you can grow the people you presently have.

Third, competencies are identified by what superior performers do, not what you or someone else think they should be doing. You cannot rely on what the job description says they should be doing or the list of competencies you see posted on the web because every organization is different. You have to talk to the customers, peers, coworkers, and managers to identify who the superiors performers are.

I think this is why competencies can be so hard to define and understand. We want a clear picture of what a competency looks like — sort of a like a task that has a definite beginning and end. But to see a competency, you have to actually look at a superior performer. Imagine if you never seen or tasted hot fudge before and I tried to explain it to you:

me: “It taste sweet.”
you: “Like sugar?”
me: “Well, it also has a nutty, buttery taste.”
you: “Hmm, what does it look like?”
me: “Runny poop.”
you: “Yeeeew!!!!!”

When looking at the superior performer you have to ask yourself what makes her better than her peers? Why do the highest spending customers come to her and not Joe? Why do her peers come to her when they have tough question rather than to Sally who is a lot closer? Why does her manager give her a superior performance review every time? Why has she excelled at the last four job she has moved on to? Thus, a competency may start looking like a combination of things — personality, motive, skills, and knowledge.

For example, Sally might know the computer applications better than anyone else in the department. However, when asked questions by her peers, she mildly scolds them for not knowing the answer in the first place and then acts as if she is doing them a huge favor when she shows them how to do it.

While the superior performer may not know how to do what is asked, she is not afraid to dive in, check the help menus, twiddle around a few minutes until they get the results they were looking for and in addition, have fun doing it.

Task modeling would look at it from Sally's view — the beginning to end of each logical and necessary action that results in a major accomplishment. However, why is Sally not considered the superior performer?

Competency modeling looks at it from the superior performer's view — familiar with the applications (especially the help functions), not afraid to experiment, loves working with and helping others. Of course you are going to have to refine it more than this.

Thus, when it comes to hiring, do you hire experts in the technical applications, or those who are are at least familiar with it, love to experiment, and enjoy helping others?

Next Steps



McClelland, D.C. (1973). Testing for competence rather than for intelligence. American Psychologist, 28, 1-14.