I planning an experiment and need some feedback. I will test how well a test person react to a moving object. The object will be presented in three different forms. The interaction between the test person and the object will be via a computer “game”. I will not go into details but would just like to get general recommendation about the experimental design.

The object will be presented in 3 different forms each moving across the screen separately. The test person have to detect and respond to the moving objects (3 different forms that varies in detectability). I will score how successful the test person is to respond (success vs. no success).

Response variable, binomial: ‘success vs. no success’

Explanatory variable: ‘form of the object’ (3 object categories w/different difficulty levels)

Co-variable: ‘Age of the test person’, although I consider to have students with approximately the same age to elude this variable.

Co-factor: ‘Gender of the test person’

Random effect: 'Test person ID'

I would like to get recommendations about:

1.      How many survey participants (test persons) do I need?

Each test person will only be tested once (i.e. one trial). In one trial, I will repeat each object category a number of times in a random order.

2.      How many times should each object category be repeated?

The 3 object categories have different difficulty levels, by repeating them I will get a sufficient sample size. Although, the drawback by repeating them is that the test person will also improve his/her search image and become more effective. This could be controlled for by including a “time” or “stage” effect as a co-variable.

I would like to get the right balance between the number of survey participants and number of repeated moving objects per trial. 

Best regards,

Ronny

More Ronny Steen's questions See All
Similar questions and discussions