Path Analysis with SPSS/AMOS

advertisement
Path Analysis
SPSS/AMOS
Theory of Planned Behavior
Zero-Order Correlations
• PATH-INGRAM.sav data file from SPSS
data page.
Attitude SubNorm
Attitude
PBC
Intent
Behavior
1.000
.472
.665
.767
.525
SubNorm
.472
1.000
.505
.411
.379
PBC
.665
.505
1.000
.458
.496
Intent
.767
.411
.458
1.000
.503
Behavior
.525
.379
.496
.503
1.000
Predicting Behavior
Beta
t
Sig.
-1.089
.281
Intent
.350 2.894
.005
PBC
.336 2.781
.007
(Constant)
Predicting Intention
Beta
t
Sig.
2.137
.037
Attitude
.807 6.966
.000
SubNorm
.095
.348
PBC
-.126 -1.069
(Constant)
.946
.290
Path Diagram
AMOS Graphics
• Click Analyze, IBM SPSS AMOS. In the
AMOS window which will open click File,
New:
Draw That Path Diagram
• Click on the “Draw observed variables”
icon which I have circled on the image two
slides above.
• Move the cursor over into the drawing
space on the right.
• Hold down the left mouse button while you
move the cursor to draw a rectangle.
Release the mouse button.
Icons
Duplicate Icons
• Draw one rectangle.
• Now click the Duplicate Objects icon,
boxed in black on the slide above
• Point at that rectangle, hold down the left
mouse button while you move to the
desired location for the second rectangle,
and release the mouse button.
Altering/Moving Objects
• Change the Shape of Objects
– Click the icon
– Click the object and move the mouse.
• Move Objects
– Click the icon Click the object and move the
mouse
Set Object Properties
• Click on the “List variables in data set”
icon.
• Drag and drop variable names to the
boxes.
• To view/edit object properties, right-click
the object and select Object Properties
Draw Paths
• Click on the “Draw paths” icon.
• Draw a path from Attitude to Intent (hold
down the left mouse button at the point
you wish to start the path and then drag it
to the ending point and release the mouse
button).
• The borders of the objects being
connected will change color when
selected.
Draw Covariances
• Click on the “Draw Covariances” icon.
• Draw a covariance from SubNorm to
Attitude.
• Use the “Change the shape of objects”
tool to increase or decrease the arc of
these covariances.
Adding An Unique Variable
• Click on the “Add a unique variable to an
existing variable” icon.
• Move the cursor over the Intent variable
and click the left mouse button to add the
error variable.
• Right-click the error circle leading to Intent,
select Object Properties, and name the
variable “e1.”
Analysis Properties
•
Click the “Analysis properties” icon -to display the Analysis Properties window.
Select the Output tab and ask for the
output shown below.
Conduct the Analysis
• Finish drawing the path diagram
(illustrated earlier) and then
• Click on the “Calculate estimates” icon.
• In the “Save As” window browse to the
desired folder and give the file a name.
• Click Save.
One or More Variables
Not Named
• You may get this error even when every
variable in the model is named.
• In my experience, you might as well start
over from scratch at this point.
• Suggested curses can be found at
http://www.vnutz.com/curse_and_swear
OK, Stop Cussing
• In BlackBoard, go to Documents,
Structural Equation Modeling & Path
Analysis, Path Analysis Files.
• Download the files.
• Open Path-Ingram.sav in SPSS.
• Analyze, AMOS
• File, Open, Path-Ingram.amw
• Calculate Estimates
View the Output Path Diagram
• Click the icon outlined in red below.
• The one to the left will display the input
path diagram.
Standardize the Coefficients
• Click “Standardized
estimates.”
Export the Path Diagram
• Click the “Copy the path diagram to the
clipboard icon. Open a Word document or
photo editor and paste in the path
diagram.
View the Output Details
• Click the “View text” icon.
Export the Details
• The Copy to Clipboard icon (green dot,
above) can be used to copy the output to
another document via the clipboard.
2 Output
•
•
•
•
Chi-square = .847
Degrees of freedom = 2
Probability level = .655
The null here is that our model fits the data
just as well as a saturated model (one with
every variable connected to every other
variable).
R2
Variable
Intent
Behavior
Estimate
.600
.343
• These are for Intention predicted from
Attitude, Subjective Norms, and
Perceived Behavioral Control, and
• Behavior predicted from Intention and
Perceived Behavioral Control.
Standardized Direct Effects
Intent
Behavior
SubNorm
.095
.000
PBC
-.126
.336
Attitude
.807
.000
Intent
.000
.350
These are all shown in the path
diagram.
Standardized Indirect Effects
Intent
Behavior
SubNorm
.000
.033
PBC
.000
-.044
Attitude
.000
.282
These are products of
coefficients. For example,
Attitude to Behavior is .81(.35)
= .28.
Intent
.000
.000
Goodness of Fit Indices
• GFI = .994.
• This tells you what proportion of the
variance in the sample variancecovariance matrix is accounted for by the
model.
• This should exceed .9 for a good model.
• For the saturated model it will be a perfect
1.
The Normed Fit Index (NFI)
• NFI = .994. .9 or higher is good.
• Compares our model to the independence
model (a model with no paths or
covariances)
• The Comparative Fit Index (CFI) is similar,
and good with smaller samples.
• CFI = 1.000
Root Mean Square Error of Approximation
• Estimates lack of fit compared to the
saturated model.
• RMSEA of .05 or less indicates good fit,
and .08 or less adequate fit.
• RMSEA here is .000.
Download