Reading the Mind in the Eyes test: Is it really valid?

Page 2 of 2 [ 18 posts ]  Go to page Previous  1, 2

Macaco
Emu Egg
Emu Egg

Joined: 20 May 2016
Age: 55
Gender: Male
Posts: 1
Location: Quebec

20 May 2016, 6:31 am

I have studied the RMET for the last 2 years. The test is not perfect, but almost.
We found that due to specific design properties, internal consistence measures are difficult to obtain (Alpha scores).
However, the test re-test procedures of the RMET are extremely accurate, thus confirming psychometric properties and high fidelity index (0.89).
We tested between different populations, we also compared results of many researches with-in and across-culture with some very interesting results.
Statistically the test poses some challenges, but it remains an excellent tool to explore human emotions.
Hope this helps.



Ettina
Veteran
Veteran

User avatar

Joined: 13 Jan 2011
Age: 34
Gender: Female
Posts: 3,971

20 May 2016, 9:45 am

What bothers me is that there are far better designed tests of emotion reading capability that are not getting nearly as much publicity as this test.

Personally, my favorite is the Emotion Multimorph Test. Basically, a computer gradually morphs a face from neutral to an emotion, and the subject presses a button when they think they know which emotion it's portraying, and then chooses out of four options - sadness, fear, anger and happiness.

The options are the same for each item, and use simple vocabulary that most people will understand. Plus, because it tests only four basic emotions, you can come up with emotion-specific scores - for example people with amygdala abnormalities tend to be especially poor at recognizing fear. If you know where a person is particularly weak, you know what to work on.

In contrast, the RMET uses complex vocabulary that many people - especially people with communication disabilities or poorer education - don't understand, it doesn't allow you to calculate emotion-specific scores, and it's pretty easy to guess the correct answer with only minimal emotion-reading ability. (If you can only tell positive vs negative, for example, you'll get 30% right on all the negative emotions in the multimorph test, but probably do pretty well on the RMET.) Plus, as people have noted, the pictures used in the RMET don't necessarily actually portray the emotions that are marked as 'correct'.