I have designed a questionnaire to test which of three motivational constructs can best be used to describe learners motivation. 25 statements were designed to test the three constructs (each question relating to one of the three), they were randomly ordered and participants were asked to respond to each statement by ticking in the appropriate place on one of 6 points along a scale from strongly agree to strongly disagree. I had planned to use exploratory factor analysis to analyse the data, although now I wonder whether a different method of analysis might be better? I was told I would need about 300 responses for factor analysis to be effective/appropriate, and although I hope not to get much fewer than that - I am doing a comparative study with two categories of students, so expect to collect somewhere between 125-140 from each of the two categories. Could anyone advice me what methods of analysis might be more suitable?
$\begingroup$
$\endgroup$
1
1 Answer
$\begingroup$
$\endgroup$
If you know which question relates to which construct, you should use confirmatory rather than exploratory factor analysis. If you have two categories of students, you should use multiple group analysis, and test several levels of invariance. If you have Likert scales, you would need to either use polychoric correlations (within each group) or explicit ordinal modeling (which is doable with IRT graded response models if you are better familiar with them than say with GLLAMM framework).
I had planned to use exploratory factor analysis to analyse the dataYou hardly so far described what you want, what is the aim. Only the title, as currently, drops tiny hint. So, compare or factor? $\endgroup$