There are a lot of myths about recruitment and selection. The biggest myth is that it's an art, not a science. Here we present a rationale to determine exactly what to select for within the selection process - and hence we illustrate the scientific selection process.

What is performance?

Performance is when an employee does what they say they will.

An employee can be considered to perform if they meet their agreed objectives or if they do what it says in their job description. They will perform if they do what they agree to do when asked by their manager - and indeed, if they do what can reasonably be expected by their manager, without being asked.

It's that simple.

How can I tell if someone will perform?

This is where the science comes in.

Firstly, you have to ignore the dependence of an employee's performance on their manager, the job they do and the environment in which they do it. Manager, job and environment have a huge impact, but we can't consider those when hiring staff. We have to assume that all three will enable performance, not thwart it. We must select on the attributes of the person.

So, telling in advance who will perform involves the idea of prediction. The hiring manager is trying to predict who, from the candidates presented, will perform.

How does a manager predict who will perform?

To make a prediction, we need criteria - something to base the prediction on. Qualifications suggest the person has a good education but we would make a leap of faith if we said that because someone had a degree in this or that, they would perform. Likewise, experience tells us little. Generally, no written material about a candidate tells us anything about their future performance.

To predict, we have to test the person sitting before us.

So what tests do we administer to predict performance?

Well, that depends on what we expect the jobholder to do and to what level. We have to define what we mean by performance in each jobholder's case. Objectives, ad hoc requests and those things managers just expect employees to pick up will all be in the accountabilities of the job description. So ultimately performance is defined in a well-crafted job description.

So, tests must be designed to elicit evidence to show who will excel in each accountability in the job description.

So what do I look for?

That answer is again simple - take the job description, break it into the personal characteristics, competencies and behaviours that will secure performance - then test for presence of those. The premise is that a person with those personal characteristics, competencies and behaviours will perform - all other things (like management, job and environment) being equal. Selection is made on the test outcomes that evidence the personal characteristics, competencies and behaviours required. Base your prediction on that evidence.

So, selection seeks evidence of the right personal characteristics, competencies and behaviours.

But shouldn't I hire for attitude?

Let's kill that one stone dead. Those who say 'hire for attitude' mean 'hire for personality'.

A person's attitude depends on the environment in which they work. Their personality determines how they will react to a particular work environment so test for personality fit with that needed by the job. That's one of the personal characteristics. You can't test someone for an attitude towards a job, firm and manager that they have not yet joined.

So, finally, what do I select for?

Select for the personal characteristics, competencies and behaviours that are needed in order to excel in each job.

And developing those criteria takes skill, knowledge, time and energy.

But without those, you're guessing.