IMPS 2001 July 15 19 2001 Osaka Japan
- Slides: 38
IMPS 2001, July 15 -19, 2001 Osaka, Japan All about variable selection in factor analysis and structural equation modeling Yutaka Kano Osaka University School of Human Sciences 1
2 Today’s talk n n n Motivation for variable selection How SEFA (and SCo. FA) works Derivation of the statistics Theoretical property What does variable selection with model fit mean? Summary
3 Needs for variable selection n Variable selection in EFA is an important but time-consuming process n n n Composite scale construction Reliability analysis Variable selection in SEM should be less important but … n n Indicator selection Improvement of model fit
4 Recent literature n n n Little et. al. (1999). On selecting indicators for multivariate measurement and modeling with latent variables. Psychological Methods, 4, 192 -211. Fabrigar et. al. (1999). Evaluating the use of EFA in psychological research. Psychological Methods, 4, 272 -299. Kano et. al. (in press, 2000, 1994).
Procedures for variable selection in EFA n Usual procedure n n Magnitude of communalities Interpretability Towards simple structures Our approach n Model fit 5
6 Programs for variable selection in factor analysis n Exploratory analysis n n n SEFA(Stepwise variable selection in EFA) http: //koko 15. hus. osakau. ac. jp/~harada/sefa 2001/stepwise/ Confirmatory analysis n n SCo. FA(Stepwise Confirmatory FA) http: //koko 16. hus. osakau. ac. jp/~harada/scofa/input. html
7 Example_1 n n A questionnaire on perception on physical exercise n=653, p=15, one-factor model Data was collected by Dr Oka (Waseda U. ) Conclusion n Remove X 2, X 9, X 13, X 14
13 SCo. Fa: 24 Pschological variable
Original Model (p=24) 14
15 Theory of SEFA and SCo. FA n n n Obtain estimates for a current model Construct predicted chi-square for each one-variable-deleted model using the estimates, without tedious iterations Take a sort of LM approach
16 Known quantities and goal_1
17 Known quantities and goal_2
18 Basic idea We construct T 02’ as LM test
19 Final formula for T 2 Note: This is Browne’s (Browne 1982) statistic of goodness-of-fit using general estimates
21 Question 1 n n Can T 2 work even if X 1 is inconsistent? Estimate for Θ is biased.
23 Question 2 n n n Can SEFA identify an uncorrelated variable? Unfortunately, no We have developed a way of testing zero communality in SEFA (see Harada-Kano, IMPS)
24 Question 3 n What is the actual meaning of variable selection with model fit? n The following shows an illustrative example:
25 Answer 3_1: Example again n X 2, X 9, X 13, X 14 are to be removed
26 Answer 3_2: Example again n n Best fitted model with correlated errors SEFA conclusion: X 2, X 9, X 13, X 14 are to be removed
27 Answer 3_3: Example again n n Variables to be deleted are identified so as to break up the correlated errors Correlated errors may cause n Different interpretation of FA results n n n Common factors considered are not enough to explain correlations between observed variables Such variables are not good indicators (e. g. , in SEM) Inaccurate reliability estimates n n Green-Hershberger (2000), Raykov (2001) Kano-Azuma (2001, IMPS)
28 Question 4 n How one should do if SEFA or SCo. FA identifies a variable with large factor loading estimate as inconsistent?
29 Answer 4_1: Reliability n If one employs the alpha coefficient or (s)he has to delete it to have a good-fit model.
30 Answer 4_2: Reliability n If one employs (s)he can remain it, and compare reliability between models.
31 Answer 4_3: Example ρ' 0. 64 α 0. 74 Bad-fitted One-factor Model based ρ 0. 76
32 Answer 4_4: Example ρ' α 0. 64 0. 74 0. 63
33 Answer 4_5: Example ρ' α 0. 60 0. 78 0. 63
34 Summary_1 n n A new option for variable selection was introduced, which is based on model fit. You can easily access the programs on the internet n SEFA(Stepwise variable selection in EFA) n n http: //koko 15. hus. osakau. ac. jp/~harada/sefa 2001/stepwise/ SCo. FA(Stepwise Confirmatory FA) n http: //koko 16. hus. osakau. ac. jp/~harada/scofa/input. html
35 Summary_2 n n It enjoys preferable theoretical properties Testing null communality is important n n n Uncorrelated variables cannot be identified Variable selection with model fit can find out error correlations Traditional reliability coefficients based on a poor-fit model have serious bias
36 Summary_3 n n High communality variables can be inconsistent Whether such variables should be removed depends n Reliability has to be figured out using nonstandard factor model
37 References n n Harada, A. and Kano, Y. (2001) Variable selection and test of communality in EFA. IMPS 2001, Osaka Kano, Y. (in press). Variable selection for structural models. Journal of Statistical Inference and Planning. Kano, Y. and Harada, A. (2000). Stepwise variable selection in factor analysis. Psychometrika, 65, 7 -22. Kano, Y. and Ihara, M. (1994). Identification of inconsistent variates in factor analysis. Psychometrika, Vol. 59, 5 -20
Thank you for coming to Osaka and being at my talk n n 38 Tako. Yaki performance will start soon You can understand how octopus relates to Osaka, if you see and taste it