Website Worth

Total Pageviews

Sunday

Factor analysis in Stata: PCF vs PF

1) use PCF when: you intent to develop one concept. If after rotation there are more than one factor, take the first one only.
  • Variance is considered 1 for each item.
  • Total variance is equal to number of items
2) use PF when: you guess there are more than one concept/dimension in the set of items
  • Concerns with shared variance (R2 of each item regressed on the rest), not the total variance
  • Total shared variance is always smaller than number of items