The problem with references

14 May 2014

The problem with references as a way of selecting candidates for a job is they only work if the referee is a perceptive person who you trust. And therefore they don’t work at all.

When I’m recruiting someone, I’m normally trying to work out how good a lawyer they are (but you can substitute engineer, teacher, street-sweeper as appropriate).

A reference, essentially, is another person’s view of how good a lawyer they are. This could be extremely valuable. Recruitment processes are inevitably flawed, because they give a limited snapshot of a candidate and their capabilities. Someone who knows a candidate well has had a much better opportunity to assess them.

However, this only works if the person giving the reference is perceptive and trusted. Otherwise, I can’t tell whether they’re giving an honest reference or whether their honest opinion is worth anything. More than that, I need to know everyone’s referee well enough to compare references from different people (one person’s ‘not a complete idiot’ may mean much the same as another’s ‘brilliant legal mind’).

In a world where I knew everyone’s referees that might work. But that just isn’t possible. So I’d be left with information about some candidates but not others. And candidates who happened to know the same people I did would gain an advantage. Which is both unfair and likely to lead to discrimination.

References do work as a type of fraud detection. It’s worth checking that people really did attend the universities and do the jobs they say they did. But that’s really the limit of their usefulness (in my not so humble opinion).

blog comments powered by Disqus