First, I know this question is outside this project. But, I believe I will reach a larger audience through your outlet.
I used two instruments for collection of primary data. My questionnaire consists of 8 dichotomously scored items of 0 & 1 (sought to determine exposure to violence), while my teacher behaviour inventory consists of 5 items on a 5-point scale of 0-4 (sought to determine behavioural outcome). The instruments were administered only once.
Facial and content validity were conducted. Principal component factor analysis was conducted. And, then reliability test was run. The items sought to be validated are those that survived both factor analysis and reliability testing. I am trapped on validity testing.
Please, how do I run and report the result of data validity test using SPSS?
I need your helpful suggestions to scale through. Thank you.
I'm not sure I got your question, but first things first, I believe the order in which you describe the validity assessment (PCA > reliability) should rather be the other way around (reliability > PCA). Normally, I would carry out a reliability analysis on all items followed by a PCA on the whole set (8 dichotomous variables and the remaining 5 items) to see factor loading on the first (two) component(s). Then one can decide which variables to include and then carry out the cronbach's alpha reliability analysis. Given all looking adequate, one can then aggregate the scores of the remaining variables to create a single scale variable. Of course, this latter bit I mentioned is theoretical and, depending on your specific question and scales, might or might not be possible to do.
Theoretically, you can compare the value of the rxy table with an r-product mean in SPSS to assess validity (i.e., checking for bivariate correlation). Also, is the dichotomous assessment capturing a spectrum / same dimension ?
But then again, I am not sure I understood what exactly you might be looking for in terms of an answer. There is a plethora of validity indices than can be tested for. Thus it would be helpful if you could describe what you mean by "validity testing".
You can follow this thread for more info https://www.researchgate.net/post/how_to_validate_a_questionnaire_that_is_not_comprised_of_likert_scale
Thank you Andrei.
You got my question right. Thank you for the lucid explanation and the recommended link. They are helpful.
- Badges
- Science topic
- Similar topics
- Instrumentation Engineering
- Instruments