To study the phonological processing abilities with non-word repetition task in children who stutter, we performed three non-word repetition tasks in two groups of children who stutter and who don't stutter. Responses from the nonword repetition task were scored into two different methods in below:

First, the per cent correct nonwords were scored based each nonword as a correct or error response.

Second, the overall per cent correct phonemes (the total number of phonemes that were correctly produced in each nonword divided by the total number of phonemes ∗ 100).

Each of these tasks has been made taking into account some of the factors affecting the non-word repetition performance. Each of these assignments includes sub-scales (one to five syllable). The aims of our study were:

1) To compare the performance of the two groups in three tasks to identify the task that distinguishes better between the two groups.

2) To identify items from each sub-scale for each task that could better differentiate between the two groups and put them in the form of a non-word repetition list.

I want to know which statistical methods I should use. Please guide me about the above issues.