lisrel850结构方程分析软件

Notes on the LISREL syntax files for Multilevel SEM examples. MSEM1A.LS8: 1 Factor CFA Model ------------------------------ The data used for this example were obtained from the data library of the Multilevel Project at the University of London. A single factor CFA model is fitted to the data (factor loadings are constrained to be equal for the between and within groups). The differences between this file and the usual 2-group LISREL syntax files are: 1) The addition of the $CLUSTER command following the RA command. 2) The addition of the $PREDICT command. Gender is used as a predictor (fixed part of the model). 3) Sample sizes are set to zero so that the program determines the sample sizes. 4) Raw data are provided in the form of a Prelis System File (PSF). Note: Observed variable labels are extracted from the PSF. 5) The MI keyword is used on the DA command line since the data contain missing values. Note: The MI keyword is not necessary if a PSF with a predefined global missing value is used. A global missing value can be assigned by using the Define Variables option from the Data menu of the PSF window. MSEM2A.LS8: 2 Factor CFA model, Gender as fixed effect ------------------------------------------------------ The data used for this example were obtained from the data library of the Multilevel Project at the University of London. Language and Socio vary between schools, but not within schools. A two factor CFA model is fitted to the data. The differences between this file and the usual 2-group LISREL syntax files are: 1) The addition of the $CLUSTER command following the RA command. 2) The addition of the $PREDICT command. Gender is used as a predictor (fixed part of the model). 3) Sample sizes are set to zero so that the program determines the sample sizes. 4) Raw data are provided in the form of a Prelis System File (PSF). Note: Observed variable labels are extracted from the PSF. 5) The MI option on the DA command line is used since the data contain missing values. Note: The MI keyword is not necessary if a PSF with a predefined global missing value is used. A global missing value can be assigned by using the Define Variables option from the Data menu of the PSF window. MSEM3A.LS8: MIMIC model, 2 school variables ------------------------------------------- This example analyzes data downloaded from the MPLUS website: www.statmodel.com. The variables x3 and x4 vary across, but not within schools. This dataset has no missing values. The differences between this file and the usual 2-group LISREL syntax files are: 1) The addition of the command $CLUSTER following the RA command. 2) Sample sizes are set to zero so that the program determines the sample sizes. 4) Raw data are provided in the form of a Prelis System File (PSF). MSEM4A.LS8: MIMIC model based on pooled data -------------------------------------------- Run MSEM4A.LS8 to compare the estimates for the pooled data with the within schools estimates when fitting a multilevel model (see MSEM3A.LS8 or MSEM3B.SPL). This example also illustrates how raw data can be read from a Prelis System File (PSF). MSEM5A.LS8: MIMIC model, Language and Socio are school variables ---------------------------------------------------------------- Language and Socio vary between schools, but not within schools. The differences between this file and the usual 2-group LISREL syntax files are: 1) The addition of the $CLUSTER command following the RA command. 2) Sample sizes are set to zero so that the program determines the sample sizes. 4) Raw data are provided in the form of a Prelis System File (PSF). Note: Observed variable labels are extracted from the PSF. 4) The MI keyword is used on the DA command line since the data contain missing values. Note: The MI keyword is not necessary if a PSF with a predefined global missing value is used. A global missing value can be assigned by using the Define Variables option from the Data menu of the PSF window. MSEM6A.LS8: 2 Factor CFA, School and Grade as fixed effects ----------------------------------------------------------- Data used in the example are from a S.A. Schools Project. See mlsem.doc in the documents folder on the LISREL 8.50 for Windows CD and the online Help file. This example illustrates confirmatory factor analysis with 2 latent variables and School and Grade as fixed effects. The differences between this file and the standard 2-group LISREL syntax files are: 1) The addition of the $CLUSTER command following the RA command. 2) The addition of the $PREDICT command. Grade is used as a predictor (fixed part of the model). 3) Sample sizes are set to zero so that the program determines the sample sizes. 4) Raw data are provided in the form of a Prelis System File (PSF). Note: Observed variable labels are extracted from the PSF. 5) The MI keyword is used on the DA command line since the data contain missing values. Note: The MI keyword is not necessary if a PSF with a predefined global missing value is used. A global missing value can be assigned by using the Define Variables option from the Data menu of the PSF window. MSEM7A.LS8: Model MSEM6B.LS8, using ASCII data ---------------------------------------------- This example uses the raw data file SA_SCHOOLS.RAW. Data from S.A. Schools Project. This example illustrates confirmatory factor analysis with 2 latent variables and Grade as fixed effect. The differences between this file and the standard 2-group LISREL syntax files are: 1) The addition of the $CLUSTER command following the RA command. 2) The addition of the $PREDICT command. Grade is used as a predictor (fixed part of the model). 3) Sample sizes are set to zero so that the program determines the sample sizes. 4) Raw data are input in the form of an ASCII file. 5) The MI keyword is used on the DA command line since the data contain missing values. MSEM8A.LS8: Model with endogenous and exogeneous latent variables and fixed --------------------------------------------------------------------------- x-variables ----------- Data from S.A. Schools Project. Note that the dataset SASCHOOLS.PSF used in this example is a subset of the data contained in SA_SCHOOLS.PSF obtained via LISTWISE deletion. LISTWISE deletion reduces this dataset by approximately 50%. Language and Socio vary across schools, but not within schools. In this example we treat Language, Socio, Mothedu and Fathedu as fixed exogenous observed variables. There are 7 Y and 8 X-variables. The differences between this file and the standard 2-group LISREL syntax files are: 1) The addition of the $CLUSTER command following the RA command. 2) The addition of the $PREDICT command. Grade is used as a predictor (fixed part of the model). 3) Sample sizes are set to zero so that the program determines the sample sizes. 4) Raw data are provided in the form of a Prelis System File (PSF). Note: Observed variable labels are extracted from the PSF. 5) The MI keyword is used on the DA command line since the data contain missing values. Note: The MI keyword is not necessary if a PSF with a predefined global missing value is used. A global missing value can be assigned by using the Define Variables option from the Data menu of the PSF window. MSEM9A.LS8: Model with endogenous and exogeneous latent variables and Language ------------------------------------------------------------------------------ as fixed effect --------------- Data from S.A. Schools project. Language and Socio vary across schools, but not within schools. In this example we treat Language, Socio