# Syntax generator for meta-analytic structural equation modeling with # the two-stage structural equation modeling # For more information, see http://courses.nus.edu.sg/course/psycwlm/internet/tssem.zip # Version 1.10 Date: 22 April, 2007 # This syntax is used to analyze correlation matrices # Comments. The program will ignore everything after it. # The syntax is quite similar to the LISREL style. # Analysis of correlation matrices, covariance matrices or variances AN=1 # AN=1: Analysis of correlation matrices # AN=2: Analysis of covariance matrices # AN=3: Testing the homogeneity of variances. # Please note that Stage 2 analysis will not be performed for AN=3 # NG is the number of groups or studies in MASEM. # The program expects there will be enough information for each group. NG=4 # NI is the number of variables. NI=5 # CM means covariance matrix. Covariance matrix is used here. # FU means full matrix. If the matrix is symmetric, SY can be used. # FI is the ASCII file for the data. For example, the data file is raw.dat. # Note that this line is directly copied to generate LISREL syntax. CM SY FI=CLUSTER.dat # PH is the file for the parameter estimates of the PHI matrix, i.e, the pooled correlation matrix. # Note that it CANNOT be used as input at the Stage 2 directly # because it duplicates output for several groups. # PH=tssem1.cor # EC is the file for the asymptotic covariance matrix of the parameter estimates. # Note that it CANNOT be used as the weight matrix at the Stage 2 directly # because it contains the elements of the standard deviations. # EC=tssem1.ack GP=1 NO=28 GP=2 NO=81 GP=3 NO=41 GP=4 NO=187 MV=1 # Variable 1 is missing # This is the Stage 2 model # Everything after "%" will be copied to generate the stage 2 model %LA %HC EI E C JP %SE %EI JP HC E C/ %MO NX=3 NY=2 PH=ST,FR GA=FU,FI BE=FU,FI PS=DI,FR %FR GA(1,1) GA(1,2) GA(1,3) GA(2,1) GA(2,2) GA(2,3)C %BE(2,1) %PD %OU ME=WL # End of input syntax