Scribd
Upload
15 views12 pages

Sem Stats

Uploaded by

averrosherbal.managermarketing
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as XLS, PDF, TXT or read online on Scribd
15 views12 pages

Sem Stats

Uploaded by

averrosherbal.managermarketing
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as XLS, PDF, TXT or read online on Scribd
You are on page 1/ 12

Chi-Square

Chi-Square difference test


Chi-Square ddl p
Model 0 30.2 33 60.72%
Model 1 38.4 35 31.81%
Chi-Square 8.2 2 1.66%

You just need to fill in B4, C4, B5, & C5 (ie, the cells on a white background)
Model 1 should be more constrained than model 2

Page 1
Chi-Square

Test 1.66%
model 0 is better. Risk: 5%
Model 0 is better. Risk: 10%

lls on a white background)

If Model 0 is better, the fit of Model 1 is significantly lowered by the additional constraints
If no model is better (test > 5% or 10%), you should usually use model 1

For an example, see http://www.geocities.com/joophox/publist/semfamre.pdf , p. 14

Page 2
Chi-Square

the additional constraints

mre.pdf , p. 14

Page 3
Rel & Val

lambda 1-lambda2
LV--> item 1 0.75 0.438
LV--> item 2 0.8 0.360
LV--> item 3 0.817 0.333
LV--> item 4 0.868 0.247
LV--> item 5 0.735 0.460
LV--> item 6 0.657 0.568
LV--> item 7
LV--> item 8
LV--> item 9
LV--> item 10
3.595 sum (lambda2)
2.405 sum (1-lambda)2
Just fill in this column 0.599 Rho vc
21.409 Sum lambda2
0.899 Joreskog rho

Page 4
Rel & Val

I've never seen any construct giving good results with 7 or more indicators, so I limited to 10 items

If you have less indicators than are present in the B column, just delete the values!

=AVE: average variance extracted (=d5/k): convergent validity if >0,5


(Check also that Lambda's c.r. are significative & that smc>0,5

Construct reliability

Ref. : Fornell, C. & D. F. Larcker (1981), Evaluating Structural Equations Models with
Unobservable Variables and Measurement Error, Journal of Marketing Research , 18, 39-
50.

Page 5
GammaHat & RMSEA

Chi-deux df n p
329 76 494 44

Enter your values here

Page 6
GammaHat & RMSEA

Fhat GammaHat Adjusted GammaHat


0.513 0.977 0.703
Steiger's 1980 GammaHat

###
###
###
RMSEA Lo 0.072 ###
RMSEA 0.082
RMSEA Hi 0.091 ###
###

Moschopoulos' (1984) approximation is used to compute the RMSEA confidence interval. ###
Se
Moschopoulos, Panagis G. (1984), "On a New Transformation to Normality," Communications in###
Stati
http://bama.ua.edu/cgi-bin/wa?A2=ind9505&L=semnet&P=R6799&I=1

Page 7
LV with 1 item

indicator standard-deviation lambda must be fixed at


0.980 0.931

You just need to fill in B3

Page 8
LV with 1 item

uniqueness fixed at
0.096

Page 9
Cronbach's alpha
number of items 5 0.833
average correlation between items 0.500

What if I remove 2 items with the "same" correlation with the others?
Alpha becomes ,,, 0.750
number of items becomes ,,, 3

What if I remove 1 item with the "same" correlation with the others?
Alpha becomes ,,, 0.800
number of items becomes ,,, 4

What if I add 1 item with the "same" correlation with the others?
Alpha becomes ,,, 0.857
number of items becomes ,,, 6

What if I add 2 items with the "same" correlation with the others?
Alpha becomes ,,, 0.875
number of items becomes ,,, 7
You just need to fill in B2 & B3
By Michaël Korchia (professor at KEDGE BS)

www.watoowatoo.net/mk
Thanks to all SEMnet contributors
(comments welcome)
Version 1.3.B, 08/06/21

You might also like