01372 841813

Every organisation wants “smart” people, but what exactly do they need to be smart at?

On Intelligence

One of the key findings of occupational psychology research in the last fifty years has been that the single best predictor of performance at school and in work is that of general mental ability or intelligence. People blessed with higher levels of mental ability appear to benefit by having more successful careers and better quality of life than those whose intelligence is assessed as below average.

As clear as this finding is, it is not without flaws. From a social perspective, many people find it simply unacceptable that there might be an inbuilt inequality between people. The American Declaration of Independence dictum that “all men are created equal” suggests an admonition to a civilised society to put such differences aside and to create a level playing field where all have equal opportunity. And yet, in a society that also generally and pervasively adopts a “more is better” principle of life, it is hard to resist the temptation to look on those people with greater intelligence as more productive, more useful, and more valuable than their counterparts.

From a scientific perspective, the biggest problem is that of how to define intelligence. Almost everyone these days is familiar with the IQ test, which typically yields a score somewhere between 70 and 130. A few months ago UK newspaper headlines featured an 11 year old whose IQ score (162) was claimed to be higher than that of Einstein. But what exactly does an IQ test measure? When we look at the sort of questions that IQ tests ask, we find that they contain relatively simple problems that have a single correct answer that can be derived through a process of logical deduction.

See below for an example IQ test problem:


This type of question is far removed from the sort of problems that people have to deal with in their everyday lives. For example, a typical workplace problem might run from “how to complete a delivery schedule on time?” to “how to design a new organisational structure?” Nevertheless publishers of IQ tests make the claim that their tests measure, at least in part, a general intelligence factor called g that underpins and accounts for a large part of the effectiveness of the kind of mental reasoning that people use at work and in their home lives.

The evidence for the general intelligence factor g is derived out of the statistical analysis of scores from people who have been asked to take a variety of different IQ tests. For any particular person, the scores on one test are usually found to be pretty close to their scores on other tests. Note however, the g is inferred as opposed to being directly measurable. One cannot say for sure that a person’s score on a particular IQ test fully reflects their g. For that reason, great care has to be taken when making inferences about someone’s level of intelligence from their score on a particular IQ test. The comparison between Einstein and the 11 year old girl mentioned earlier is completely spurious. In any case, t is not known whether Einstein ever took an IQ test, and even if he had, it would certainly not have been the same as that taken by the 11 year old girl.

Today, the importance attached to being “smart” is reflected in the competency models, assessment methods and selection criteria for personnel hire of most organisations, and a great variety of tests for measuring intelligence and ability exist. Professional standards for the choice of such tests point to the need for organisations to establish a link between performance on the test and performance on the job in question in order to justify their use. Simply put, a person’s scores on the test should be related to measures of that person’s performance on the job such that the higher the test score, the better their performance on the job. In practice, such relationships are found to be rather weak and their measurement suffers from the problem that job performance itself is often difficult to measure since job performance is rarely down to a single individual’s efforts but instead reflects a myriad of contextual factors such as the overall business climate, the effect of local work conditions and impact of managers and colleagues, as well as a measure of luck.

Work psychologists are frequently called on to design, or choose, tests that will accurately predict work performance. To make such choices effectively, they need to analyse the factors that make people successful or unsuccessful in a specific role. Since correlation is not the same as causation (which is to say that, just because events are related this does not mean that one causes the other) the research design that would establish whether a particular test accurately predicts work performance needs to be longitudinal (i.e. to observe cause and effect over time) and to involve comparisons with a control group so that the actual cause of good job performance can be identified. But such research designs, which are the hallmark of good pharmaceutical drug-testing, are rarely used by work psychologists. Few organisations are willing to spend the time or money. By consequence, the use of many ability or IQ tests to predict who might perform well in a specific role is at best a matter of some speculative inference and at worst heavily misguided. It’s a bit like trying to guess peoples’ weights from the size of their feet. At the extremes one can judge accurately, for example, that someone with very small feet will weigh less than someone with very large ones. Bit in the middle of the range there are a lot of people for whom size of foot will not be a reliable indicator that they weigh more or less than a person with a size or two smaller or larger feet.

What, then, are organisations to do if they want to ensure that they hire people with the appropriate intellect to perform well in their allotted roles? Here are a few points to bear in mind:

  1. Appreciate that IQ and ability test performance are only a small part of intelligence. Be clear about the extent of the relationship between test and job performance.
  2. Distinguish between speed and power of thinking. IQ tests tend to require high speed on relatively simple problems. But powerful thinking involves the capacity to see problems as well as to solve them. So it is important to investigate candidates’ ability to appreciate the degree of complexity inherent in a work role.
  3. If tests are to be used, bear in mind that test-taking style can provide useful data – sometimes more useful than the test score itself. Consider for example, speed, accuracy and error-rate. If the job requires slow careful methodical analysis then high speed, error-prone test performance might be a negative indicator.
  4. Investigate through interview or other means a candidate’s ability to learn from experience. Consider whether the role in question requires the ability to think in new and different ways, or whether the person will be required to think in very much the same way as in previous roles.
  5. Analyse the type of decision-making required in the role, noting the distinction between simple, complicated and complex systems. Simple systems follow well-established cause and effect laws that are easy for people to identify and to use logic to solve. Complicated systems also follow well-established cause and effect laws, but at greater levels of depth and logical sophistication. Complex systems, which are composed of a great variety of interacting parts, are much more dynamic and unpredictable than simple or complicated systems. They may adhere to certain principles but they do not follow precise rules. Skill in logical analysis is less effective when making decisions with complex systems than experience, judgement, and mental flexibility.
  6. Use multiple data sources and multiple perspectives to build up a picture of a candidates’ problem solving capacity. Do not place sole reliance on any single measure.