Tell me more ×
Cross Validated is a question and answer site for statisticians, data analysts, data miners and data visualization experts. It's 100% free, no registration required.
up vote 2 down vote favorite

I performed a CFA to reduce my items into factors. However, due to the violations of non-normality, the results were really weird, and transformations didn't help. There were some standard regression coefficients that were negative and ones that were over 1 despite the factors itself having good cronbach alphas.

  1. Where could I have gone wrong in my CFA? What's the meaning behind getting path coefficients over 1?
  2. Would it be suitable to even include such a bad-fit model into the write-up of my results?
share|improve this question
1  
Welcome to the site. The coefficients are called "standardized", not standard. Cronbach is a last name, and should be capitalized. If you have standardized coefficients greater than one, you probably also have variances below zero. The situation is referred to as "Heywood case" in factor analysis literature, so you might want to Google this phrase and see if something helps. Non-normality may or may not be a culprit. It would help if you demonstrated your results, including the non-normality diagnostics. – StasK Oct 15 '11 at 4:39
What's your sample size? How many factors and how many items per factor? Are factors correlated? What do you get when you run exploratory factor analysis with the same number of factors? – Jeromy Anglim Oct 16 '11 at 2:11
Related: stats.stackexchange.com/questions/4081/… – Henry Oct 16 '11 at 2:22

Know someone who can answer? Share a link to this question via email, Google+, Twitter, or Facebook.

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Browse other questions tagged or ask your own question.