I know that exploratory factor analysis needs some restrictions for statistical identification, such as fixing the variances of factors at 1 and fixing the correlation between factors at 0. Apart from these, are there other restrictions imposed on the example below, so that we only need to estimate 11 parameters?
PS. The output is quoted from the Stata manual (Example 5 for factor). From the manual and some other textbooks, I didn't find the answer, so I am here for seeking your help. Thank you very much.
Code:
. webuse bg2, clear . factor bg2cost1-bg2cost6, factors(2) ml Factor analysis/correlation Number of obs = 568 Method: maximum likelihood Retained factors = 2 Rotation: (unrotated) Number of params = 11 Schwarz's BIC = 83.4482 Log likelihood = -6.842448 (Akaike's) AIC = 35.6849 -------------------------------------------------------------------------- Factor | Eigenvalue Difference Proportion Cumulative -------------+------------------------------------------------------------ Factor1 | 1.02766 0.28115 0.5792 0.5792 Factor2 | 0.74651 . 0.4208 1.0000 -------------------------------------------------------------------------- LR test: independent vs. saturated: chi2(15) = 269.07 Prob>chi2 = 0.0000 LR test: 2 factors vs. saturated: chi2(4) = 13.58 Prob>chi2 = 0.0087 Factor loadings (pattern matrix) and unique variances ------------------------------------------------- Variable | Factor1 Factor2 | Uniqueness -------------+--------------------+-------------- bg2cost1 | -0.1371 0.4235 | 0.8018 bg2cost2 | 0.4140 0.1994 | 0.7888 bg2cost3 | 0.6199 0.3692 | 0.4794 bg2cost4 | 0.3577 0.0909 | 0.8638 bg2cost5 | -0.3752 0.4355 | 0.6695 bg2cost6 | -0.4295 0.4395 | 0.6224 -------------------------------------------------
0 Response to Number of parameters in exploratory factor analysis
Post a Comment