マルチグーループSEMをMplusで行う。データはHolzingerSwineford1939を用いる。Mplus用のデータへの変換は下部参照のこと。
今回は配置不変モデルのみ。
Grant-White学校のSEMのダイアグラム。
コード
TITLE: Multiple Group SEM using HolzingerSwineford1939 Data DATA: FILE = "HolzingerSwineford1939.dat"; LISTWISE = ON; VARIABLE: NAMES = id sex ageyr agemo school grade x1 x2 x3 x4 x5 x6 x7 x8 x9; USEVARIABLES = x1 x2 x3 x4 x5 x6 x7 x8 x9; GROUPING = school (1=Grant-White, 2=Pasteur); MISSING=.; Analysis: TYPE = GENERAL; ESTIMATOR = ML; ITERATIONS = 200; MODEL = NOMEANSTRUCTURE; INFORMATION = EXPECTED; MODEL: visual by x1 x2 x3; textual by x4 x5 x6; speed by x7 x8 x9; OUTPUT: SAMPSTAT STDYX;
結果
Group GRANT-WHITE VISUAL BY X1 0.727 0.061 11.957 0.000 X2 0.466 0.066 7.080 0.000 X3 0.651 0.061 10.598 0.000 TEXTUAL BY X4 0.857 0.030 28.579 0.000 X5 0.857 0.030 28.614 0.000 X6 0.795 0.034 23.414 0.000 SPEED BY X7 0.665 0.057 11.710 0.000 X8 0.793 0.052 15.298 0.000 X9 0.700 0.055 12.706 0.000 TEXTUAL WITH VISUAL 0.540 0.086 6.317 0.000 SPEED WITH VISUAL 0.536 0.093 5.733 0.000 TEXTUAL 0.345 0.091 3.781 0.000
Group PASTEUR VISUAL BY X1 0.771 0.063 12.153 0.000 X2 0.432 0.063 6.893 0.000 X3 0.600 0.061 9.901 0.000 TEXTUAL BY X4 0.823 0.031 26.218 0.000 X5 0.824 0.031 26.256 0.000 X6 0.860 0.029 29.410 0.000 SPEED BY X7 0.514 0.059 8.710 0.000 X8 0.679 0.063 10.726 0.000 X9 0.577 0.060 9.558 0.000 TEXTUAL WITH VISUAL 0.484 0.087 5.600 0.000 SPEED WITH VISUAL 0.340 0.114 2.994 0.003 TEXTUAL 0.333 0.100 3.342 0.001
データセット
Rで実行する。
library(lavaan) data(HolzingerSwineford1939) df1 <- HolzingerSwineford1939
Mplusへの書き出し1
library(MplusAutomation) variable.names(df1)
Mplus用のデータに変換
prepareMplusData(df1, filename="HolzingerSwineford1939.dat", overwrite=T)