We explore a little about the science and technology that goes on behind the scenes at the Compatibility Test.
The compatibility test uses images for a number of reasons. The first is that it's a universal language and the test is taken in a number of countries around the world. Instead of translating the text into lots of languages, the test works for everyone. Secondly, the images tend to create a more instinctive reaction where text tends to be a more considered response. During the development of the test, we found that people react quite quickly to one image or another and select the result. With text, users clearly have to read the text first but then tend to consider the response.
Most of the stastics show that the initial iconic memory test produces a more realistic reaction. People who consider the text tend to choose answers that are closer what people think they should answer.
The algorithm doesn't just compare the number of images that people have chosen the same. There is a correlation between the images chosen and the match because, of course, if everyone chose exactly the same images they would have a pretty close match. However, certain images counteract other images so even choosing one different pair can have a significant change in the match percentage.
Behind the scenes, each image is related to personality traits and these are actually what is matched. This was also done so that in the future, we can modify the images but the underlying elements of what is being tested will remain the same.
The use of images of shapes and patterns rather than actual objects tended to produce better results. It's possible that depending on further testing we will include both types of images but using images of 'real' things tends to provoke a more obvious reaction.