Obtaining additional decimal places PreviousNext
Mplus Discussion > Confirmatory Factor Analysis >
 Gareth posted on Monday, November 08, 2010 - 3:24 am
Is it possible to request more than three decimal places for the parameter estimates? I have a very small effect which is significant but estimated as 0.000. I probably need the fourth decimal place.

Thank you in advance.
 Linda K. Muthen posted on Monday, November 08, 2010 - 6:11 am
You can save the results using the RESULTS option of the SAVEDATA command. Results are saved with more decimals.
 Anonymous posted on Wednesday, November 07, 2012 - 6:58 am
I have been trying to obtain extra decimal places for my parameter estimates as follows:

format is f10.5;
file is data.dat;
estimates are estimates.dat;

Although the data is saved to 5 decimal places the parameter estimates are only saved to 3 decimal places. Is there a way round this?
 Linda K. Muthen posted on Wednesday, November 07, 2012 - 7:10 am
Use the RESULTS option not the ESTIMATES option.
 Simone Croft posted on Wednesday, February 18, 2015 - 3:54 am

I've been following the instructions above to get an extra decimal place in my results. I have added the following command:


I'm not getting any error messages regarding this line of code, but I'm also unable to find the actual dat file with the results. I've tried a couple of locations, but can't find the file. Do you have any ideas why this might be happening?

Thanks in advance,
 Linda K. Muthen posted on Wednesday, February 18, 2015 - 6:36 am
Look at the end of the output. Does it say it was saved? If so, copy the name of the file and search for it on your computer. Otherwise, send the output and your license number to support@statmodel.com.
 Ivan Jacob Agaloos Pesigan posted on Saturday, January 19, 2019 - 3:54 am
Is it possible to save confidence intervals for instance from CINTERVAL (BOOTSTRAP); in the same way you save results using SAVEDATA: RESULTS ARE: "results.dat";?
 Linda K. Muthen posted on Saturday, January 19, 2019 - 1:46 pm
No, this is not possible.
 Siddhartha Baviskar posted on Wednesday, February 13, 2019 - 7:25 am

I have been trying to obtain all results from my CFA with only 2 decimal places. I have tried the following:

FORMAT IS f10.2;
RESULTS ARE "cfa.dat";

I can open the cfa.dat file in Mplus, but all I can see is an array of numbers.

How can I obtain the equivalent of the output file with all results only up to 2 decimal places?

Thank you.
 Linda K. Muthen posted on Wednesday, February 13, 2019 - 5:55 pm
This option is not available in Mplus. The FORMAT option is for writing the data set not the results.
 Kathy Vu posted on Thursday, June 06, 2019 - 9:06 pm
Hi Drs. Muthen & Muthen,

I am also having difficulties with the RESULTS option of the SAVEDATA command. I am testing configural invariance and need the 4th decimal of model fit indices. I've tried:

results are decimals.txt;

results are decimals.dat;

results are decimals;

I keep receiving a file with numbers like below:
0.31391304E+01 0.33695652E+01 0.37889175E+01 0.26284862E+01 0.26746950E+01 0.27836504E+01 0.31234322E+01 0.29956522E+01 0.29565217E+01 0.29260869E+01

I am not sure what I am doing incorrectly. Any advice would be appreciated!
 Linda K. Muthen posted on Friday, June 07, 2019 - 12:11 pm
Go to the output where you saved the results. You will see what is saved. The order of the parameters is taken from the numbers in TECH1.
 Allison posted on Tuesday, January 21, 2020 - 8:07 am

I am similarly trying to extract additional decimal points for results from a BCH comparison in multilevel LPA. I have the following commands included at the end of my syntax:

OUTPUT: Tech7; Tech11; Tech14;

SAVEDATA: FILE=Test1.dat; SAVE=CPROB; RESULTS are decimals.dat;

I received an output file similar to the comment above from Kathy Vu that is difficult to match up to the BCH results, particularly since the BCH results don't seem to be appearing in the Test1.dat file. Is there a different way to export this information so I can obtain four decimal points for each BCH comparison?
 Tihomir Asparouhov posted on Wednesday, January 22, 2020 - 2:27 pm
It is possible to do this using the manual BCH method, see Section 3

You will need the command
SAVEDATA: RESULTS are decimals.dat;
in the final estimation step where all the BCH parameters are actual model parameters.
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