Make no mistake about it, interview are a form of psychometric test and should be treated as such.
For whatever reason, certain employers refuse to knowingly incorporate psychometric tests into their selection process. Some aren’t sold on their effectiveness, some don’t feel they know enough to use them correctly, and others just hate the word, psychometric, sounding too cold and clinical to be used at their organisations.
But what is a psychometric test really, and why are employee interviews treated differently to personality questionnaires, ability tests and situational judgement tests?
What makes a test psychometric?
A psychometric test is literally any form of assessment designed to measure a psychological construct. Competencies, values, knowledge, skills, abilities, character, motivations, interpersonal skills, leadership styles, virtually everything that interviews attempt to measure, are classic examples of psychological constructs. Psychological constructs exist only as latent traits, as opposed to physical traits such as height, weight, eye colour or hair colour, and can only be measured indirectly through the use of psychometric assessments.
Just because you aren’t handing the candidate a questionnaire, does not mean you aren’t using psychometrics, you are just giving the candidate questions verbally.
How the data is collected, whether it’s done verbally, in writing, from a computer or from multiple choice questions, does not dictate whether the assessment is “psychometric” or not, it’s what the assessment aspires to measure that dictates whether the assessment is psychometric.
Now with that in mind, let’s contrast how interviews are designed compared to other psychometric tests. To meet the quality control standards set by the British Psychological Society (BPS) and / or the American Psychological Association (APA), significant research efforts must be conducted on the assessment, which include the following requirements:
- Validity: The assessment must be demonstrably valid, with clear statistical evidence suggesting that the assessment measures the psychological construct(s) that it purports to measure.
- Reliability: The assessment must be sufficiently accurate, precise and replicable. Typically, for reliability statistics such as Cronbach’s alpha, a minimum reliability coefficient of .7 is required for use in practise.
- Fairness: The assessment must not show adverse impact against legally protected groups. Typically, standardised effect sizes are used to quantify group differences, with an effect size of .5 indicating significant adverse impact.
- Benchmarking: Norm groups are required, ensuring that candidate’s scores can be standardised and compared against a wider average for the relevant population. For example, if the candidate is a graduate, their score must be benchmarked against the graduate population.
In order to meet or exceed these set standards, research projects spanning several months, or even years must be undertaken, requiring thousands of participants and a team of psychologists.
Now consider how interviews are normally designed.
Almost always, interview questions are written by non-psychologists with almost no understanding of the psychological constructs they aim to measure. Often, questions aren’t written at all, and are instead created on the spot by the interviewer.
Even with the most structured, standardised interview, almost no research is undertaken to quantify the validity, reliability or fairness of the interview, leaving too many questions unanswered. Considering this lack of standardisation, it is no surprise that most unstructured interviews make for very ineffective selection tools. If interviews were treated like other psychometric tests, almost everyone’s interviews would fail to meet BPS and APA guidelines.
Now don’t get me wrong, a well-designed, well validated, structured interview can be a powerful predictor of future performance. However, interviews aren’t special, they are just another psychometric test that employing organisations can utilise to select and assess candidates. They have their benefits and disadvantages relative to other assessments available.
So there you have it, unequivocally, interviews are as much a psychometric test as any personality questionnaire, cognitive ability test or situational judgement test. All these assessments aim to do the same thing, measure key psychological constructs that should predict future job performance.
Too long have interviews been held in excessively high esteem and thus avoided the high standards imposed on other psychometric tests. Instead, interviews should be viewed the same way as any other psychometric test, and held to comparable standards of quality.