How to Interpret Ability Test Results
Lead consultant at Test Partnership, Ben Schwencke, explains how to interpret ability test results and norm group benchmarking.
Transcript
Interpreting ability test results. Now when people are using ability tests for the first time, the results and the reports can seem a little bit confusing, mostly because the scores are almost always norm referenced as opposed to simply saying how many questions the candidate got correct, and there are at least two good reasons for this.
Firstly, say the campus gets 12 out of 15 on a numerical reasoning test. Is that good? Is that bad?
Unfortunately we can't tell without some kind of external benchmark.
Now imagine we're looking at graduates and the average graduate gets 14 out of 15.
It's a really easy test.
That would mean the count didn't do particularly well.
What if the average graduate gets 8 out of 15?
In which case the count did quite well.
Unfortunately without that external benchmark, it's not possible to say whether the candidates raw score, as we call it, is good, bad or indifferent.
The second reason is often, with more advanced assessments, the difficulty of the assessment and of the questions also impacts the particular score.
So with item Bank assessments where different candidates get different combinations of questions the camera gets 12 out of 15, maybe they were given particularly hard questions which means they get a higher score than a candidate that got 13 questions correct, but given easier questions.
As a result we simply can't rely on the candidates raw score, the number of questions they got correct, to inform selection decisions.
It isn't ethical, it isn't effective, as a result what we do is we use norm reference scores.
So when looking at the report of a candidate what you'll see is a percentile score is probably the most common way of this representing this information - what that means is if the candidate gets at the 50th percentile it means they've scored exactly average for their population whether it be graduates, professionals, managers, apprentices, it means they've scored exactly in the middle.
So if it's a fifteen question test and they get 12 out of 15 and the average graduate gets 12 out of 15, they're exactly 50th percentile.
Now if the candidate gets the 80th percentile they're now in the top 20 percent of performance for that population if they give a 20th percentile it means they're at the bottom 20 percent.
And there are other scores that we can provide for example stem schools which are very simple 1 to 10 score, with 5 and 6 being average.
We can get T scores, we can get Z scores, but they all represent the same piece of information which is how well the candidate did compared to that chosen norm group - the relevant population that they belong.
