Good day all!
I have run bivariate correlations to see if there is a connection between two scales, I ran Pearson's Correlation and have -.511 correlation between two scales (is this the correct test for Likert scale data)?
Second question: if I want to test if patterns exist between scale data and demographics, what is the best test for this? I have created dummy codes for each of my demographics (racial identity, gender identity, and level of education).
Thank you very much for your assistance.
JDM
When using Likert-scored items to create a scale, there are some preliminary steps you should take. First, use coefficient alpha to ensure that the items all measure the same the same thing (you want to get an alpha of at least .7). Next, do an exploratory factor analysis to determine whether the items all measure a single underlying concept (I recommend using Principal Axis for the extraction with an oblique rotation.)
Thank you Dr. Morgan.
Likert Scale 1 came with Cronbach's Alpha at.940.
Likert Scale 2 came with Cronbach's Alpha at.946.
When you recommend an exploratory factor, does this mean to place each of the questions scores from Likert Scale1 (7 questions) into the variables? I found a tutorial for SPSS, and adding each of the 7 questions into the variables, I tried the rotation, and it responded: Only one component was extracted. The solution cannot be rotated. Did I not run the test correctly?
Thank you.
With such high values on alpha, you can be assured that your variables all measure the same thing, so you don't need to do any factor analyses.
- Badges
- Science topic
- Similar topics
- Social Psychology
- Attitude