RMSEA PreviousNext
Mplus Discussion > Exploratory Factor Analysis >
 Cesar Daniel Costa Ball posted on Thursday, March 28, 2013 - 9:01 am
Hello, I ran an EFA and RMSEA = .061 but not as interprtar following "Probability RMSEA <= .05 p= 0.081"
 Bengt O. Muthen posted on Thursday, March 28, 2013 - 3:52 pm
The value 0.05 is often used as a limit that a good model should not exceed. This output says that with your estimate of 0.061, the probability is only 0.081 that this estimate is less than 0.05 and you have a good enough model in this sense.
 Cesar Daniel Costa Ball posted on Thursday, March 28, 2013 - 10:19 pm
Thank you for your prompt response.
I understand how to interpret the value of RMSEA. Does the probability associated with RMSEA is recommended to be less than p = .05?
 Linda K. Muthen posted on Friday, March 29, 2013 - 10:10 am
You want the probability of RMSEA being less than .05 to be high.
 Cesar Daniel Costa Ball posted on Friday, March 29, 2013 - 6:41 pm
When Chi square is higher than 0.05, and CFI / TLI are higher than 0.95 and RMASEA has a value less than 0.05, usually associated to the probability value is higher 0.05 RMASEA. Am I correct? Am I playing well the value and probability of RAMSEA?
The value must be less than 0.05 and also see that the probability is higher than 0.05, to accept the null hypothesis. Is this correct?
 Bengt O. Muthen posted on Saturday, March 30, 2013 - 6:33 am
Note that the chi square p-value should be lower than 0.05, not higher than 0.05.

I think you would want the probability that RMSEA is lower than 0.05 to be higher than 0.05, but see the original articles on RMSEA.
 Cesar Daniel Costa Ball posted on Saturday, March 30, 2013 - 11:39 am
Thanks for your help
 Suyin Chang posted on Monday, February 24, 2014 - 5:59 pm
Dear Mr. Muthen,

I found this thread by accident, but thought I should drop this clarifying question regarding your reply above.

Shouldn't the chi-square of the model fit be non-significant, i.e. with a p-value greater than 0.05? The one that should have a p-value lower than 0.05, i.e. significant, is not the chi-square of the baseline model?

 Bengt O. Muthen posted on Tuesday, February 25, 2014 - 11:47 am
I think there was some confusion in this interchange talking about (1) chi-square test of overall model fit and (2) p-value for RMSEA being less than 0.05.

Regarding (1), you want a p-value of at least 0.05 for a model to not be rejected.

Regarding (2), you want a rather high p-value.
 Suyin Chang posted on Tuesday, February 25, 2014 - 12:39 pm
Thanks Dr Muthen, this was precisely the carification I was thinking of. Quite helpful.
Back to top
Add Your Message Here
Username: Posting Information:
This is a private posting area. Only registered users and moderators may post messages here.
Options: Enable HTML code in message
Automatically activate URLs in message