²Õ´¦æ¬°¬ã¨s¤u§@§{
SEM ¸É¥R¸ê®Æ (SEM demos for OBRM Workshop)
l 2011
SEM Workshop Information
1. Simplis »yªk¤¤ªºÆ[¹îÅܶµ¦WºÙªº^¤å¤j¤p¼g»Ýn»PpsfÀɮפ¤§¹¥þ¬Û¦P¡A§_«h·|µLªk°õ¦æ»yªkÀÉ¡C¨Ò¦ppsfÀɤ¤¤§Æ[¹îÅܶµ¦W¬°self2¡ASimplis»yªk¤¤¼gSELF2=SE¡A«hµLªk°õ¦æ¸Ó»yªk¡A¥B·|¥X²{SELF2¬°¤£¥i¿ëÃѤ§Åܶµ¦WºÙ¦¹¿ù»~°T®§¡C
2. LISREL©xºô©Ò´£¨Ñ¤§°Ï§O®Ä«×°Ñ¦Ò¸ê®Æ(The assessment of
the validity of constructs)
l * ¥»¯¸LISREL¬ÛÃöµe±»P°Ó¼Ð¡AÄÝSSI©Ò¦³¡AAmos¬ÛÃöµe±»P°Ó¼Ð¡AÄÝSPSS©Ò¦³ *
l
½d¨ÒÀÉ®× (Demo files)
?
Amos ¾Þ§@¥Ü½d§ë¼v¤ù¤U¸ü(Handout for Amos demo)
?
Depress.zip ÀɮפU¸ü (Depress.zip used in the demo)
? Depress.zip ÀÉ®×»¡©ú: Depress15.sav ¬°SPSS dataÀɮסA¥]§tSelf-esteem 5 ÃD¡ADepress 4 ÃD¡AImpulsive 3 ÃD¡C self, depres, impuls ¬°«ez¤Tºc©À¦UÃD¥Ø¤§¥§¡È¡C¦¹Àɮ׬°LISRELºô¯¸½d¨Ò¸ê®ÆÀÉ¡A¥H¤U¥Ü½d¬Ò±Ä¥Î¦¹Àɮ׶i¦æ¡C
? Depress.zip contains 4 LISREL syntax files (.lpj) and 1 SPSS data file ( Depress15.sav has 5 Self-esteem items, 4 Depress items, 3 Impulsive items. self, depres, and impuls are the means of the above 3 constructs). This data file is an example provided by the LISREL website.
l
SEM ¾Þ§@¥Ü½d¼v¤ù <§Q¥Î¥k¤W¨¤ ¹Ï¥Ü¥i¥þ¿Ã¹õ¼½©ñ,¶È¨Ñºô¤WÆ[¬Ý,¤£´£¨Ñ¤U¸ü>
SEM demos (Video clips, click the to switch to the full-screen mode)
You may need to install powercam player to
watch the demo
l ¡@
?
LISREL
n
Data Management
Converting SPSS data with missing values into PRELIS system file (.psf) or LISREL data system file (.dsf)
J
Listwise deletion method
J
Multiple imputation method
n
CFAs & Full Model (Based on
the LISREL example from the SSI
website)
J
Creating depress15.dsf for LISREL analyses
J
Scaling latent variables by fixing PH
J
Scaling latent variables by fixing LX
J
Discriminant validity between SE and IMP
J
Full Model (DPàSE & DPàIMP)
J
LISREL and SIMPLIS syntax files of the above models
? Amos
n Observed Variable Path Analysis
n
Confirmatory Factor Analysis
n Testing mediation effect with bootstrapping
n
J
Bootstrapping in Amos
(Reference: Cheung,
G. W., & Lau, R. S. (2008). Testing mediation and suppression effects of
latent variables: Bootstrapping with structural equation models. Organizational
Research Methods, 11, 296-325.)
¡@
¡@
¡@