1993 Fiscal Year Final Research Report Summary
Development & research in covariance structure analysis.
| Project/Area Number |
04610088
|
|---|
| Research Category |
Grant-in-Aid for General Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Research Field |
Psychology
|
|---|
| Research Institution | National Center for University Entrance Examination |
|---|
Principal Investigator |
ISIZUKA Tomoichi National Center for University Entrance Examination, Associtae Professor., 研究開発部, 助教授 (00168238)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
TUTIDA Syouji University of Meiji, Department of letter, Associate professor., 文学部, 助教授 (90197707)
TOYODA Hideki National Center for University Entrance Examination, Assistant professor., 研究開発部, 助手 (60217578)
|
|---|
| Project Period (FY) |
1992 – 1993
|
| Keywords | Civaruabce Structure Analysis / identification rule / simultaneous equation model / generalized least squares / Reticular Aciton Model / Confirmatory Factor Analysis / noniterative estimation procedure / Asymptotic Covariance of Estimator |
|---|
| Research Abstract |
An alternative sufficient condition for checking identifiability of the simultaneous equation model is proposed using the notation fo the RAM (Reticular Aciton Model ; McArdle & McDonald, 1984). The new rule can be more widely applied for checking identifiability of the simultaneous equation model compared with other rules. Using the notation of the RAM, an estimation procedure for the simultaneous equation model based on GLS (generalized least squares) is also proposed.
When there are fixed parameters and linearly constrained parmeters in the model , the estimated which ontain the covariances between observable endogenous variables and residual variables can be obtained from explicit formulas of matrix. The consistency and the asymptotic covariance of this estimator are shown.
By using the data form kluegel, Singleton & Starnes (1989), the procedure is empirically compared to ML (maximum likelihood) and GLS (Browne, 1974).
The purpose of second study is to propose a noniterative estimation procedure for confirmatory factor analysis base on the instrumental variable method. It has the advantage compared with previous methods that linear equality constraints between unique and common factor variances are available.
The procedure is empirically compared to other methods. The conclusion, based on the data used in this study, is that the method described seems to work well.
|
