Thursday, 22 February 2018


We all like to understand ourselves a bit better and I was recently invited to take a psychometric test which looks at my Verbal Analysis Aptitude and Numerical Analysis Aptitude.

Both tests are 24 minutes against the clock and designed with a count-down timer to create a degree of pressure and stress to see how you respond.

Let’s examine a couple of really scenarios

Scenario 1 - There are 40 questions and Sam answers 20 of them, all correctly, in the time available. Score: Of the ones Sam answered Sam got them 100% right, but Sam only answers half the questions so Sam scores 50% Interpretation: Sam is slow, but correct.

Scenario 2- There are 40 questions and Alex answers 40 of them, getting only half of them right, in the time available. Score: Alex also scores 50%, but using a very different approach. Interpretation: Alex is fast, but not correct.

The scores are the same, but have very different interpretation which may be gathers be assessing the time spend on reading and responding to each question. Presupposing accurate timing of each was recorded.

Let’s examine different context and strategies with different implications

Strategy A – The aim is to get the most right, so if you are running out of time you are better guessing (and having a 20% chance of getting it right) rather than let the clock run-out and not score.

Strategy B – The aim is to get the least wrong, so if you are running out of time you are better spending time getting it right than being panicked or guessing wrong.

Real life implications

If you were in a business/culture/context where speed and getting most right made money or saved lives then you may consider Strategy A. Perhaps in this context Alex may be the better candidate.

If you were in a business/culture/context where accuracy getting is wrong cost lives or money then you may consider Strategy B. Perhaps in this context Sam may be the better candidate.

This was not an unusual psychometric test, but the implications of how the algorithms are interpreted and the context in which they are delivered produce very, very different outcomes.

What is interesting in this context is the GDPR Implications.

Under General Data Protection Regulation [GDPR] Sam or Alex are entitled to know how the process works. The have rights related to automated decision making including profiling. This type of test is specifically mentioned: “a recruitment aptitude test which uses pre-programmed algorithms and criteria.”


Because this type of processing is considered to be high-risk the GDPR requires you to carry out a Data Protection Impact Assessment (DPIA) to show that you have identified and assessed what those risks are and how you will address them.

GDPR also:
1.      requires you to give individuals specific information about the processing;
2.      obliges you to take steps to prevent errors, bias and discrimination; and
3.      gives individuals rights to challenge and request a review of the decision.

These provisions are designed to increase individuals’ understanding of how you might be using their personal data.

You must:
·         provide meaningful information about the logic involved in the decision-making process, as well as the significance and the envisaged consequences for the individual;
·         use appropriate mathematical or statistical procedures;
·         ensure that individuals can:
a)       obtain human intervention;
b)       express their point of view; and
c)       obtain an explanation of the decision and challenge it;
·         put appropriate technical and organisational measures in place, so that you can correct inaccuracies and minimise the risk of errors;
·         secure personal data in a way that is proportionate to the risk to the interests and rights of the individual, and that prevents discriminatory effects.

If you are an HR professional or expert in psychometric testing I would be very interested in your feedback, experience and suggestions.

No comments:

Post a Comment