|
|
|
||||
|
|
|||||
Table 4.1, p. 47.
Estimating the intercept only model using langpost as the dependent variable, schoolnr as level 2 (group level) and pupilnr as level 1 (individual level).
Title: Random intercept model with no covariate
Data:
File is mlbook1.dat ;
Variable:
Names are
schoolnr pupilnr iq_verb iq_perf sex minority repeatgr aritpret classnr
aritpost langpret langpost ses denomina schoolse satiprin natitest
meetings currmeet mixedgra percmino aritdiff homework classsiz groupsiz;
Missing are all (-9999) ;
usevar langpost schoolnr;
cluster = schoolnr;
Analysis:
Type = twolevel random ;
estimator = ml;
model:
%within%
langpost;
%between%
langpost;
Loglikelihood
H0 Value -8126.609
H1 Value -8126.609
Information Criteria
Number of Free Parameters 3
Akaike (AIC) 16259.219
Bayesian (BIC) 16276.424
Sample-Size Adjusted BIC 16266.892
(n* = (n + 2) / 24)
RMSEA (Root Mean Square Error Of Approximation)
Estimate 0.000
SRMR (Standardized Root Mean Square Residual)
Value for Between 0.000
Value for Within 0.000
MODEL RESULTS
Estimates S.E. Est./S.E.
Within Level
Variances
LANGPOST 64.562 1.973 32.731
Between Level
Means
LANGPOST 40.363 0.428 94.405
Variances
LANGPOST 19.471 3.111 6.259
Table 4.2, p. 49.
The model includes only the predictor gndc_verb.
Title: Random intercept model with a covariate;
Data:
File is h:\xiao\mlbook1.dat ;
Variable:
Names are
schoolnr pupilnr iq_verb iq_perf sex minority repeatgr aritpret classnr
aritpost langpret langpost ses denomina schoolse satiprin natitest
meetings currmeet mixedgra percmino aritdiff homework classsiz groupsiz;
Missing are all (-9999) ;
within = iq_verb;
usevar langpost schoolnr iq_verb;
cluster = schoolnr;
centering = grandmean(iq_verb);
Analysis:
Type = twolevel random;
estimator = ml;
model:
%within%
langpost on iq_verb;
%between%
langpost;
Loglikelihood
H0 Value -7625.887
H1 Value -7625.886
Information Criteria
Number of Free Parameters 4
Akaike (AIC) 15259.773
Bayesian (BIC) 15282.713
Sample-Size Adjusted BIC 15270.004
(n* = (n + 2) / 24)
RMSEA (Root Mean Square Error Of Approximation)
Estimate 0.000
SRMR (Standardized Root Mean Square Residual)
Value for Between 0.000
Value for Within 0.000
MODEL RESULTS
Estimates S.E. Est./S.E.
Within Level
LANGPOST ON
IQ_VERB 2.488 0.071 35.275
Residual Variances
LANGPOST 42.225 1.288 32.774
Between Level
Means
LANGPOST 40.609 0.308 132.005
Variances
LANGPOST 9.515 1.577 6.032
Table 4.3 on page 51, ordinary least squares regression
Title: OLS regression;
Data:
File is mlbook1.dat ;
Variable:
Names are
schoolnr pupilnr iq_verb iq_perf sex minority repeatgr aritpret classnr
aritpost langpret langpost ses denomina schoolse satiprin natitest
meetings currmeet mixedgra percmino aritdiff homework classsiz groupsiz;
Missing are all (-9999) ;
usevar langpost iq_verb;
centering = grandmean(iq_verb);
Analysis:
Type = meanstructure;
model:
langpost on iq_verb;
MODEL RESULTS
Estimates S.E. Est./S.E.
LANGPOST ON
IQ_VERB 2.654 0.072 36.797
Intercepts
LANGPOST 40.935 0.149 274.397
Residual Variances
LANGPOST 50.897 1.505 33.816
Table 4.4 on page 55, within- and between-group regression
Title:
iq_verbc is grandmean centered version of iq_verb;
iqbar is the group-mean centered version of iq_verbc;
Data:
File is mlbook1_a.dat ;
Variable:
Names are
schoolnr pupilnr iq_verb iq_perf sex minority repeatgr aritpret classnr
aritpost langpret langpost ses denomina schoolses satiprin natitest
meetings currmeet mixedgra percmino aritdiff homework classsiz groupsiz
iq_verbc iqbar;
Missing are all (-9999) ;
usevar langpost iq_verbc iqbar;
within = iq_verbc;
between iqbar;
cluster = schoolnr;
Analysis:
Type = twolevel random;
estimator = ml;
model:
%within%
langpost on iq_verbc;
%between%
langpost on iqbar;
Loglikelihood
H0 Value -7613.774
H1 Value -7613.774
Information Criteria
Number of Free Parameters 5
Akaike (AIC) 15237.548
Bayesian (BIC) 15266.223
Sample-Size Adjusted BIC 15250.338
(n* = (n + 2) / 24)
RMSEA (Root Mean Square Error Of Approximation)
Estimate 0.000
SRMR (Standardized Root Mean Square Residual)
Value for Between 0.000
Value for Within 0.000
MODEL RESULTS
Estimates S.E. Est./S.E.
Within Level
LANGPOST ON
IQ_VERBC 2.414 0.072 33.691
Residual Variances
LANGPOST 42.149 1.284 32.826
Between Level
LANGPOST ON
IQBAR 1.593 0.313 5.089
Intercepts
LANGPOST 40.741 0.285 143.038
Residual Variances
LANGPOST 7.743 1.313 5.897
Table 4.5, on page 64. Currently Mplus 3 does not allow three-level modeling yet.
UCLA Researchers are invited to our Statistical Consulting Services
We recommend others to our list of Other Resources for Statistical Computing Help
These pages are Copyrighted (c) by UCLA Academic Technology Services