Psyc451
Exam#1
Lab#5
Name: ____________________
Path Analysis
Dataset  clinical_pathdata451.sav
Walk-through:
Friend Social
Support
Stress
Family Social
Support
Depression
State Anxiety
Trait
Anxiety
1. Obtain the full path model
Please note: the “error” values aren’t obtained until you do the model comparison in Step 3
a. Using Stress as the criterion obtain the following:
R² for the model = _______
Variables
Friend Social
Support
error associated with Stress = _________
Family Social
Support
Trait
Anxiety
β (r)
b. Using State Anxiety as the criterion obtain the following:
R² for the model = _______
Variables
Friend Social
Support
error associated with State Anxiety = _________
Family Social
Support
Trait
Anxiety
β (r)
c.
Using Depression as the criterion obtain the following:
R² for the model = _______
Variables
Friend Social
Support
error associated with Depression = _________
Family Social
Support
β (r)
d. Conclusions:
Identify direct effects upon depression in the model
Identify indirect effects upon depression in the model
Trait
Anxiety
Stress
State
Anxiety
Friend Social
Support
Stress
Family Social
Support
Depression
State Anxiety
Trait
Anxiety
2. Obtain the hypothesized (reduced) path model
Please note: the “error” values aren’t obtained until you do the model comparison in Step 3
e. Using Stress as the criterion obtain the following:
R² for the model = _______
Variables
error associated with Stress = _________
Trait
Anxiety
β (r)
f.
Using State Anxiety as the criterion obtain the following:
R² for the model = _______
Variables
Friend Social
Support
error associated with State Anxiety = _________
Family Social
Support
β (r)
g. Using Depression as the criterion obtain the following:
R² for the model = _______
Variables
Trait
Anxiety
error associated with Depression = _________
Stress
State
Anxiety
β (r)
3. Compare the two models
a. N = ___________
d = __________
b. Fit of full model = _________ Fit of reduced model = __________
Q = ______ W = ______
c. Conclusion: Does removing the paths reduce the fit of the path model?
p = _______
Friend Social
Support
Stress
Depression
State Anxiety
Trait
Anxiety
4. Obtain the “trimmed” path model (drop all non-significant paths from the full model)
Please note: the “error” values aren’t obtained until you do the model comparison in Step 5
h. Using Stress as the criterion obtain the following:
R² for the model = _______
Variables
error associated with Stress = _________
Trait
Anxiety
β (r)
i.
Using State Anxiety as the criterion obtain the following:
R² for the model = _______
Variables
Friend Social
Support
error associated with State Anxiety = _________
Trait
Anxiety
β (r)
j.
Using Depression as the criterion obtain the following:
R² for the model = _______
Variables
Trait
Anxiety
error associated with Depression = _________
Stress
Friend Social
Support
β (r)
5. Compare the two models
a. N = ___________
d = __________
b. Fit of full model = _________ Fit of reduced model = __________
Q = ______ W = ______
c. Conclusion: Does removing the paths reduce the fit of the path model?
p = _______
Your Turn #1
use the data set  talent_pathdata451.sav
2. Working with the following variables, construct a causal model of college performance (first year college gpa -the outcome variable) with no more than 4 predictive layers – Socio-economic status, locus of control, self
concept, motivation, reading, writing, math, gender, high school program. Locus of control, self concept and SES
were measured in high school, reading, writing & math were measured when applying to college, motivation and
gender were measured during the beginning of the first year of college. Remember to pay attention to both the
measured and the manifested “when” for each variable!!
a.
b.
c.
d.
e.
f.
Draw the full model with all the arrows.
Draw a hypothesized model – dropping those paths that you think are non contributing
Compute the full model and draw it with all the path coefficients
Compute the hypothesized mode and draw it with the hypothesized path coefficients
Test the hypothesized model vs. the full model
Compute the “trimmed” model (dropping all non-significant paths from the full model) and draw it with al the
included path coefficients.
g. Test the trimmed model vs. the full model.
About the “Your Turn,” your Laboratory Project and the Conference
You may use the same criterion/predictors you used last week, change some, or
change all of them -- your choice. Remember, a “good story” isn’t just “a set of all
significant” predictors. Rather, it is a combination of things that do and don’t
correlate/have multivariate contributions, so we can tell an interesting story about how
these variables relate to the criterion.


Have a mix of “kinds of predictors” (demographics, etc.)
Have a mix of significant and nonsignificant multivariate contributors (6-8 predictors is plenty!!!)
You may try multiple criterion variables in any data set, but you must present below 3
analyses using data from at least 2 different data sets!
3. Construct a causal model with no more than 4 predictive layers. Remember to pay attention to both the measured
and the manifested “when” for each variable!!
h.
i.
j.
k.
l.
m.
Draw the full model with all the arrows.
Draw a hypothesized model – dropping those paths that you think are non contributing
Compute the full model and draw it with all the path coefficients
Compute the hypothesized mode and draw it with the hypothesized path coefficients
Test the hypothesized model vs. the full model
Compute the “trimmed” model (dropping all non-significant paths from the full model) and draw it with al the
included path coefficients.
n. Test the trimmed model vs. the full model.