Chapter 14, Second-Order, Dynamic, and Multitrait Multimethod Models Video Solutions, A Beginner's Guide to Structural Equation Modeling

Problem 1

The psychological research literature tends to suggest that drug use and depression are leading indicators of suicide among teenagers. (Note: Set variance of Suicide $=1$ for model identification purposes). Given the following data set information, create and run a LISRELSIMPLIS program to conduct a second-order factor analysis.
```
Observed Variables: drug1 drug2 drug3 drug4 depress1
depress2 depress3 depress4
Sample Size 200
Correlation Matrix
1.000
0.628 1.000
0.623 0.646 1.000
0.542 0.656 0.626 1.000
0.496 0.557 0.579 0.640 1.000
0.374 0.392 0.425 0.451 0.590 1.000
0.406 0.439 0.446 0.444 0.668 . 488 1.000
Standard Deviations 1.379 1.314 1.288 1.388 1.405 1.269
1.435 1.423
Latent Variables: drugs depress Suicide
```
The second-order factor model is diagrammed in Figure 14.4:
( FIGURE CAN'T COPY )

Nick Johnson

Numerade Educator

Problem 2

A sports physician was interested in studying heart rate and muscle fatigue of female soccer players. She collected data after three soccer games over a 3-week period. A dynamic factor model was used to determine if heart rate and muscle fatigue were stable across time for the 150 female soccer players.

Create a LISREL-SIMPLIS program to analyze and interpret the dynamic factor model. Include a diagram of the dynamic factor model. The data set information including observed variables, covariance matrix, sample size, and latent variables are provided below:
```
Observed Variables: HR1 MF1 HR2 MF2 HR3 MF3
Covariance Matrix
10.75
7.00 9.34
7.00 5.00 11.50
5.03 5.00 7.49 9.96
3.89 4.00 3.84 3.65 9.51
2.90 2.00 2.15 2.88 3.55 5.50
Sample Size: 150
Latent Variables: Time1 Time2 Time3
```

Check back soon!

Problem 3

Students provided ratings of their classroom behavior, motivation to achieve, and attitude toward learning. Teachers, likewise, provided ratings of student classroom behavior, perception of students' motivation to achieve, and attitude toward learning. Finally, other students or peers provided ratings on these three traits. The three ratings (student, teacher, and peer) on three traits (behavior, motivate, attitude) were analyzed in a SEM Multitrait Multimethod model. The Multitrait Multimethod Model is diagrammed in Figure 14.5:
( FIGURE CAN'T COPY )
a. Create and run a LISREL-SIMPLIS program to analyze the three sets of ratings on the three traits as a MTMM model. The observed variables, correlation matrix, sample size, and latent variables are:
```
Observed Variables: X1 X2 X3 X4 X5 X6 X7 X8 X9
Correlation Matrix
1.0
.31 . 38 1.0
.35 . 23 . 16 1.0
. 26 . 22 . 21 . 62 1.0
. 15 . .11 . 15 . 49 . . 62 1.0
```
```
.43 . .31 . . 24 . .61 . . 48 . .33 1.0
.40 . 35 . 19 . 49 . .45 . 32 . 74 1.0
.26 . 20 . 18 . . 43 - .41 . 33 . .52 . . 47 1.0
Sample Size: 300
Latent Variables: behavior motivate attitude student
teacher peer
```
b. Create and run a LISREL-SIMPLIS program to compute a CTCU and CU model using the data information from above. Compare the CTCU and CU model results to determine if a method effect exists. Also, compare the CTCU model with the MTMM model above, which provides clearer results?

Check back soon!