Plausible value PreviousNext
Mplus Discussion > Missing Data Modeling >
 Daniel K posted on Thursday, August 12, 2010 - 11:14 pm
Dear Dr. Muthen

What are plausible values of latent variables for? The values are simply factor scores? Then, in LGM, the values are individual intercept and slope values (like EB in HLM)?

Instead of latent means analysis, can we simply compare mean plausible values?

Thank you very much for your kind consideration.

Dan Kim
 Linda K. Muthen posted on Friday, August 13, 2010 - 9:45 am
Plausible values are factor scores. I would not use factor scores as their means are not the same as latent variable means in a latent variable model.
 Christoph Weber posted on Tuesday, August 16, 2011 - 2:55 am
Dear Dr. Muthen,

I am using complex data (PISA). There are 5 plausible values for test scores in math and reading. I am estimating a path model, where reading is an indep. var and math is the dep. var.

I am using the type = imputation command in conjunction with the cluster, weight and complex command.

1.) Is it necessary to use 25 Data sets (i.e.: Dat1 = mathpv1, readpv1; Dat2 = mathpv1, readpv2; ...; Dat6 = mathpv2, readpv1, ...) or would it also be correct to use just 5 data sets (Dat1 = mathpv1, readpv1; Dat2 = mathpv2, readpv2, ...)?

2.) What would be when I further use multiple imputation for the other variables (x1 - x10) of the model? Are 5 data sets (mathpv1, readpv1, x1imput1, x2imput1, ...) enough? Or do I also have to create data sets for each combination of imputed values?

Best regards and thanks

Christoph Weber
 Linda K. Muthen posted on Wednesday, August 17, 2011 - 1:08 pm
I think you need to direct this question to someone who is knowledgeable about the PISA data.
 Christoph Weber posted on Wednesday, August 17, 2011 - 2:48 pm
but is it correct to use type=imputation to analyse plausible values?
Christoph Weber
 Bengt O. Muthen posted on Wednesday, August 17, 2011 - 6:07 pm
Yes. And typically you would have imputed data sets where each includes all your latent variables (math and reading in your case). So you would have say 5 of those, not 5 for one of the latents and 5 for the other.
 Christoph Weber posted on Thursday, August 18, 2011 - 6:12 am
Thanks, I have further another question.
I'm running a path model with the data.

My dependent variable (y) is a binary var. The other variables are continuous.
y ON x1 x2 x3 x4 x5;

x1 ON x2 x3 x4 x5;

x2 ON x3 x4 x5;

x3 ON x4 x5;

I use MLR Estimation and request the stdyx solution.

How are the Coeff. standardizes?
How are these stand. coeff. interpreted?
Is R² something like Pseudo R²?

And I don't get exp(b) - coefficients?

Christoph Weber
 Bengt O. Muthen posted on Friday, August 19, 2011 - 6:31 pm
Standardization is as usual, described in the UG - with the exception that with a categorical DV, the residual variance is either 1 (probit link) or pi-square/3 (logit link).

R-square uses the usual formulas and these same residual variances, as advocated by McKelvey and Zavoina (1975) and also in books on categorical data.

You should get exp(b), but if you don't, define them in Model Constraint as New parameters.
 Christoph Weber posted on Saturday, August 20, 2011 - 11:37 am
Thanks for your help
Christoph Weber
 Artur Pokropek posted on Friday, February 03, 2012 - 2:07 am
I've got a little question. In DATA IMPUTATION command we could find option for rounding number of decimals for imputed continuous variables (ROUNDING=). But how to set the number of decimals for plausible values i.e. imputations for latent variable? When I put name of latent variable in rounding option (for instance ROUNDING = f1 (5); ) it does not work.
Thank You
 Linda K. Muthen posted on Friday, February 03, 2012 - 8:59 am
There is no such option for plausible values.
 Ksnow posted on Tuesday, July 03, 2012 - 10:42 pm
Dear Dr. Muthen ¡G
I am doing analysis about bifactor. I want to get the plausible value of bifactor model. Can Mplus do that? I have
tried to add the plausible value code in the bifactor model. However, it doesn't work? Am I wrong or the Mplus can not do that¡H
 Linda K. Muthen posted on Wednesday, July 04, 2012 - 11:02 am
Please send the output and license number to
 Jan Zirk posted on Tuesday, September 25, 2012 - 6:14 pm
Dear Bengt or Tihomir,

In Asparouhov & Muthen (2010; Plausible Values for Latent Variables Using Mplus) you mention the plausible values (pvs) but when they are extracted from a categorical variable or a latent factor MPlus gives us mean median, SD and CI values; did you mean in Table 3 mean or median pvs?
 Tihomir Asparouhov posted on Wednesday, September 26, 2012 - 5:46 pm
Table 3 is the mean pvs, but the median will give the same result because the posterior distribution for the plausible value is normal in that particular example.
 Jan Zirk posted on Thursday, September 27, 2012 - 1:37 am
Thank you very much.
 Maria Rasmusson posted on Friday, May 03, 2013 - 1:09 am
Dear Dr Muthén,

In one of my PhD studies I've made secondary analysis on Swedish PISA data. The SEM models were tested with Mplus using the maximum likelihood parameter estimator (MLR) with the two-level complex analysis type. PISA 2009 data was used.

When I performed the analysis I was not aware of the possibility to use all 5 plausible values offered for each student. Instead I used one of the plausible values and tested the models with each PV at a time and compared the five outputs (that did not differ much).

However, one of the anonymous reviewers of my manuscript informed me about the possibility to perform the testing of the CFA models and estimating the parameters using all the five PVs in Mplus in order to get correct standard errors (since group-differences are in focus).

How is this done? Should I use type=imputation as referred to above by Christoph? What will the input instruction look like when it is supposed to call for five different data files? Is it all done in one analysis?

Thanks in advance!
Best regards,
Maria Rasmusson
 Bengt O. Muthen posted on Friday, May 03, 2013 - 8:23 am
Yes, use TYPE=IMPUTATION and you will get averaged estimates and correct SEs and chi-2 using all the imputations. See UG ex 11.8, part 2. You find the data, including the "implist" file on our website under the User's Guide examples.
 Maria Rasmusson posted on Tuesday, May 07, 2013 - 2:27 am
Tusen tack för hjälpen!
Thanks a lot. It seems to work fine!

Best regards,
 Thomas Rodebaugh posted on Wednesday, June 26, 2013 - 10:32 am
hi there,

we're scratching our heads over this output message we've been getting:

Unknown option:

are there circumstances under which even proper use of the plausible command would produce this error? or perhaps somehow we are using it incorrectly?

a slightly shortened version of our input file is below, in case that helps. any insight much appreciated! we've tried a number of permutations of the below, but keep getting the same message. we're using mplus 7.

data: file is informant2.dat;
variable: names are [quite a few];
usevariables are avfmps12 avfmps16 avfmps19 avfmps24 avfmps30
fmps12 fmps16 fmps19 fmps24 fmps30;
missing are avfmps12 avfmps16 avfmps19 avfmps24 avfmps30
fmps12 fmps16 fmps19 fmps24 fmps30 (-99);
categorical are fmps12 fmps16 fmps19 fmps24 fmps30;
analysis: estimator = BAYES;
model: [is a complicated factor structure]
PLAUSIBLE = hplau.dat;
SAVE = hplau2.dat;
output: TECH1 TECH8;
 Bengt O. Muthen posted on Wednesday, June 26, 2013 - 3:24 pm
PLAUSIBLE= is not an option in the DATA IMPUTATION command. This was changed in Version 7 to


in the SAVEDATA command. So the change is

FILE IS hplau.dat;

where 10 is the number of imputations. See the Version 7 User's Guide.
 Thomas Rodebaugh posted on Monday, July 01, 2013 - 2:21 pm
thanks, bengt! it works fine now. that will teach us to use version 7 while looking at the version 6 manual!
 Hsien-Yuan Hsu posted on Tuesday, July 09, 2013 - 9:17 pm
Dear Dr. Muthen,

I have a quick question regarding Ex 11.7.

Factor scores for the latent variables f1, f2, and f3 would be obtained and can be used in secondary analysis.

My question is that:
Do I need to include important background variables (pointing to indicators) for better estimated factor scores? Do you have any reference?

I appreciate your reply in advance!

 Linda K. Muthen posted on Wednesday, July 10, 2013 - 12:28 pm
If the factors are well-measured by several good factor indicators, you don't need to include background variables. See the special issues of JEBS and JEM on NAEP.
 Dan Cloney posted on Monday, October 06, 2014 - 5:45 am

I have been provided two datasets that include information on the same participants.

The first is a dataset with observations on a number of variables that include some missing values.

The second is a set of 5 plausible values for a proposed new latent indicator. There are plausible values for a subset of the participants observed in the first dataset.

Can you suggest the best approach to incorporate these into a single imputation analysis?

Should I estimate an arbitrary model with the first dataset, then manually add the PVs to the saved imputed datasets, and then estimate the final model?
 Bengt O. Muthen posted on Monday, October 06, 2014 - 10:47 am
I don't know what to recommend here. It sounds like the first data set has more subjects than the second, but the second has more variables - the latent variable's plausible values. Not sure what the final modeling would be.
 Dan Cloney posted on Monday, October 06, 2014 - 2:52 pm
Hi Bengt,

Thank you for your response.

You are right, the first data set contains more subjects than the second. e.g., u11, u12, u13...u33 and 3000 subjects long, with some missing data.

The second data set contains fewer variables that the first: 5 PVs for one latent variable. e.g., f1_pv1...f1_pv5 and 2000 subjects long.

The final model is intended to be a growth model, that includes f1 as a (time-invariant) covariate.

Does that extra information give you any ideas?
 Bengt O. Muthen posted on Monday, October 06, 2014 - 4:35 pm
I assume that the two datasets have some observed variables in common. And that you don't have the observed indicators of the f1 factor in either dataset.
 Dan Cloney posted on Monday, October 06, 2014 - 5:02 pm
That's right, the observed indicators of f1 do not exist in either data set.

The only element common to both data sets are the subject and cluster IDs.
 Bengt O. Muthen posted on Monday, October 06, 2014 - 5:48 pm
Then I think all you can do is analyze the n=2000 subjects in common to the two data sets, merging the PV data sets with the n=2000 subset of the other data set to get those observed variables. Then use Type=Imputation data input.
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