I have self perception ratings for N= 90 students who are deaf. They rated how well they felt individuals could understand thier speech, and how well they felt they could understand spoken sentences. We then completed speech intellibity measures of their spoken speech (using % words correctly understood) and how well they could understand spoken sentences (percent of words the deaf individuals understood).
Now if I want to see how self perception ratings (likert scores) compare to percent correct scores, what is recomended?
I converted all scores to z scores and did correlations. is this correct?
It depends on whether peception ratings were done with Likert type items or proper Likert scales (the latter consisting of more than one item). If you have just one Likert type item, that has an ordinal scale and you need a correlation for ordinal data, such as Spearman' s r or Kendall's tau. If you have a proper Likert scale, you can choose to regard that as metric scale and use Pearson's r as correlation coefficient.
As to z transformation: With ordinal data (single Item) you can't do that. With metric data, you do not have to do that. z transformation -just as any other linear transformation- will not change the correlation.
- Badges
- Science topic
- Similar topics
- Correction