Should we blame psychometric tests for Paul Flowers’ mismanagement of the Co-operative Bank?

Posted 29 January 2014 by
John Hackston, Head of R&D at OPP
Cooperative Bank

The news headlines today seem to make uncomfortable reading for test publishers like OPP. Apparently, failed Co-op Chairman Paul Flowers “aced” psychometric tests during recruitment, thereby pipping the more experienced and skilled candidates at the post. Given the disastrous results of Paul Flowers’ leadership of the bank, surely this means that psychometric tests are useless and should be thrown out?

Well, not quite. Let’s just look at things in a little more detail. All of the coverage has been generated by the evidence given to the Treasury select committee by Rodney Baker-Bates, former deputy chairman of the Co-op Bank and, before that, Paul Flowers’ rival for the chairman role. If we look at what Baker-Bates actually said, a much more nuanced picture begins to emerge:

  1. Asked by MPs if he knew why Flowers had beaten him to the job, Baker-Bates said: “I was told afterwards he did very well on the psychometric tests.” This implies that the tests were the main or only factor in making a decision about who to appoint. As test publishers, we know that this is not best practice. Each assessment tool has its particular strengths and weaknesses, and a well-rounded selection process, especially for more high-stakes roles, should use a range of different methods. We’d strongly recommend that these methods include assessments of a candidate’s competence in core areas of the job – in this case, a proven track record in the banking sector might have been useful. Blaming the psychometric test looks suspiciously like a diversionary tactic to draw attention away from the many flaws in this selection process.

  2. In fact, we know that the Co-op selection process also involved an interview, so why was it implied that the decision was based solely on test results? It might be because blaming the tests means that no-one needs to take personal responsibility for Baker-Bates’ lack of success; “it’s not my fault, blame the tests”. This is a lot less personal, and a lot easier to say, than for example “In my opinion you didn’t come over well when I interviewed you”.

  3. The interview included in the selection process, on the basis of Baker-Bates’ account, left a lot to be desired. All the evidence suggests that interviews that work best in predicting later job success ask questions that are a) set out in advance and based on the competencies required for the job, and b) ask interviewees to give examples from their past experience related to these competencies. However, the interview for Chairman of the Co-operative Bank sounds rather more like an informal test of the applicant’s knowledge of the co-operative movement. This was likely to favour internal candidates like Paul Flowers, and was arguably not particularly relevant for the role. When someone doesn’t meet the basic core skills and professional competencies required for a particular role, the results of personality and ability assessments are potentially irrelevant. Why not get the basics right first?!

  4. Things are looking a little shaky, but there is more. Baker-Bates mentioned being told that Flowers did well on the tests, going on to say “I have to say the psychometric testing did surprise me. Paul was a clear winner. When I looked at the scores I was concerned about a couple.” Showing someone’s test results to a third party without good cause and prior consent is directly against everything that OPP teaches about the ethics of using psychometric tests. In some circumstances, it could be illegal. This does make us worry about how well the whole selection process was run.

What conclusions can we draw from this sorry affair? I would suggest that you follow the advice in our white paper on selection: first, establish what skills, attributes, behaviours, knowledge and experience are needed for the role, and then choose tools that demonstrably measure these factors and that are reliable, fair and cost-effective. These will often include ability tests and personality questionnaires, but could also include structured interviews, work sample tests or simulations, assessment centre exercises and of course questions about qualifications and past experience. Extensive research evidence shows that tests are one of the best single predictors of later job success – but nothing is perfect. Sometimes a candidate who would have done well in the job comes out poorly on a selection test; sometimes a candidate who will be a poor performer in the job does well on the test. This happens less often with tests than with interviews and most other selection methods, but it does happen. This is one reason why a good selection process should always include a range of different assessments.

So are psychometric tests to blame for a £1.5 billion deficit and the ruin of the Co-operative Bank? It is interesting that all the headlines mention psychometric tests, but none mention the interview – carried out in person by three members of the board of directors who would make the decision about who to appoint. Which do you think would have the biggest influence on their decision, the interview or the tests? I know what I think...

Posted in