The Science

The science of selection is an actuarial process. All selection methods make mistakes.

Research has shown that as many as 90% of candidates distort the truth to some extent in employment interviews.

Interviewers have unconscious biases about people that are not job-relevant.

The single most important attribute of any selection method is criterion validity.

Interviews have the greatest validity when they are highly structured and based on sound job analysis. Otherwise, their validity is relatively poor.

CVs have a poor record of predicting performance. They are unstandardised and they tell readers what the candidate wants them to know.

Experience is a relatively weak predictor of performance.

Educational qualifications and academic performance are not of themselves a substitute for ability.

One hundred years of published research has established a league table of assessment methods for selection and training.

No single assessment method provides perfect prediction of job performance.

Certain combinations of assessment methods provide the best prediction, but they also fall short of perfect prediction.

Confirmation bias often conceals selection errors and perpetuates poor practice.

Tests used in development should also be able to demonstrate a correlation with job outcomes if they are to be considered credible.

Tests and other assessment methods are not in competition to see which one is right. Rather, they are complementary.

‘How much’ validity matters. The greater the validity, the better the prediction.