I had to run a Principal Component Analysis on one section of my survey to test validity and reliability; now I have to run a Cronbach's alpha on that information to check its validity and reliability, but am unsure how to do that in SPSS. Does anyone have a clue as to how to accomplish this?
I ran the PCA and determined that the results were statistically significant and likely factorizable, so now I have to re-run that data with Cronbach's alpha to check for validity and reliability and I am not sure how to do it in SPSS...if that clears up the initial question
Please can you tell, what was significant?
If you did a PCA and rotated so that your items clearly loaded on one component then you know which item loads/belongs to which sub-test (hopefully this reflects your a priori considerations). With this information you can run a reliability analysis in SPSS.
Analyze-->Scale-->Reliability Analysis. Model:Alpha
Now include all items of one factor/component/sub-test (not all items!). You will now get Cronbach's alpha.
Then repeat this procedure for each sub-test.
It is also advisable to check the correlation matrix, item-total correalation and alpha when item is deleted. This will give you additional information about the contribution of each item and how "good" it fits to the scale.
Ok, so the components that were loaded are the ones that I run the Cronbach's alpha on? You'll have to excuse me..I have literally had to teach myself statistics in the past 2 months and this is a completely new language to me. I think it makes sense...I will run it and see what my results are. Thank you!!!
Uh, 2 month and doing an PCA is hard I guess. ;-)
I think you mean the right thing, but components are the "factors" where your items load on. You have to run as many reliability analyses as components/factors/sub-tests were extracted/chosen.
So, how did you determine how many components/factors should be extracted?
For each reliability analysis, you chose the items that loaded on this component/factor.
Just let me know if I can help with your data.
SPSS did all of the analysis and left me with a mass quantity of tables that I am still trying to interpret. I have the table that gives me the five components that were extracted. I am not even sure if I need to do a reliability/validity test on this to be quite honest, but I wanted to do it for my own peace of mind.
In exploratory factorial analysis there are at least three different criterions to consider for the extraction of factors. First the Eigenwert or Kaiser-Criterion, I guess that is the one your five-factor solution is based on. But there were also other more elaborated criterions like the inspection of the scree plot or the parallel analysis.
Moreover, it is a good idea to check the necessary assumption before you run an analysis. On the output you will see the Kaiser-Meyer-Olkin-Measure (under section KMO and Barlett’s Test). The KMO should be over 0.70 (at least) and Barlett-Test should be significant. What scale have the items underlying your analysis? Did you measure the participants answers with likert scales o are the items dichotomized? What are the correlations between your items?
Further I guess it s a good idea to read some introduction literature on factor analysis before you try to publicize something. Just to understand the basics of this technique: maybe this short article (with explanations of an SPSS-output) may help you: http://ro.ecu.edu.au/cgi/viewcontent.cgi?article=1373&context=jephc
And in more detail: Schepers, J. Factorial Analysis. Or Fabrigar, L. Exploratory Factor Analysis
I hope these informations help a little.
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Statistical Software
- SPSS