Personality and Individual Differences
Emotional-Intelligence Claims Laboratory
Two fictional measures sold as emotional intelligence. One overlaps a personality questionnaire; the other overlaps a reasoning test. So the question is what either of them adds.
Simulated — fictional measures, no published test reproduced
Learning objective
By the end you should be able to say why "it correlates with the outcome" is a weaker claim than "it adds something we did not already have".
About 20 minutes. Nothing you do here is saved or sent anywhere.
What these measures are
All six are fictional descriptions of method types — a questionnaire, tasks with scored answers, ratings by other people. No published emotional-intelligence measure is reproduced, quoted, paraphrased or scored here, and no number on this page is an estimate of any real quantity. The tool takes no position on whether emotional intelligence exists. It is about what a correlation matrix does and does not license somebody to advertise.
How to use this
- Commit to an answer to the question below.
- Read the three bars.
- Slide the one control to the right and watch them swap over.
- Move to step 2 and ask whether either measure adds anything.
- Judge the five marketing claims against what you found.
First, a judgement
A fictional research group administers three measures, each described in the literature as measuring emotional intelligence: a questionnaire about how you usually handle feelings, a task whose answers are scored against a key, and an emotion-recognition accuracy test. They also collect ratings of social effectiveness by colleagues — the outcome each of the three measures is meant to predict, not a fourth measure of the same thing.
The laboratory
Six fictional measures: three carry the emotional-intelligence label, one is the colleague-rated outcome the label is sold on predicting, and two are the ordinary measures any claim of added value has to beat. Every correlation below is what the model implies directly, worked out from the measures' shared loadings rather than sampled from simulated people.
Need a hand?
Nothing here changes the controls for you, and none of it says what you are going to find.
Key terms
- Convergent validity
- Two measures of the same construct correlating, as they should if they are both measuring it.
- Method variance
- Shared variance arising from how measures were collected rather than from what they measure. Two self-report scales share a method whatever their content.
- Incremental validity
- What a measure adds to prediction once the established predictors are already in the model.
- Construct proliferation
- Introducing a new construct that substantially overlaps ones already in use.
Six measures, one matrix
The sliders do not change anyone's scores. They change what the measures are assumed to be made of — and every correlation on screen follows from that.
What each measure overlaps with
Every measure against every other
Does it add anything?
| Predictors | Variance accounted for |
|---|
What this evidence supports
Challenge — five things somebody wants to say
Each of these is the kind of sentence that appears in a brochure, a press release or a training-day slide. Judge each one against the evidence as you currently have the laboratory set. There is no stored answer key: change the settings and some of these verdicts change with them.
What this demonstrates
The self-report questionnaire overlaps a personality questionnaire heavily; the emotion task overlaps a reasoning test. Both are sold under one label, and they barely agree with each other. That is why the question worth asking is never whether a measure correlates with the outcome, but whether it adds anything once the ordinary measures are already in.
For teaching elsewhere: take this activity as one self-contained block of HTML, on the clipboard or as a file. Either way it is styled so that it will not disturb the page you put it into.