snowRMM returns an error after updating to 5.0.9

[graduallywaiting]

Hello,

I did an Partial Credit Model using the snowRMM ver 4.4.1 which returned results with no errors. However, upon updating the module to 5.0.9, this error is now appearing.

Did i do something wrong here?

[graduallywaiting]

I'm trying to do a partial credit model, however, it shows an RSM error.

[seol]

hey

you might check the 'Step' option before running analysis.

Cheers
Seol

[graduallywaiting]

Hello Dr. Seol,

I have already indicated the steps which is 4 (5 categories - 1). I intend to do the PCM since the number of categories in my dataset are not equal.

I was wondering why is it returning an error that RSM cannot be computed? even though i selected the PCM as the type of model.

Here is the error that I'm receiving:

Code: Select all

Error in datprep_RSM(X, W, sum0): RSM can not be computed since number of categories are not the same for each item!


private$.run()
private$.prepareRsmPlot(data)
eRm::RSM(data)
likLR(X, W, mpoints, groupvec, model, st.err = se, sum0, etaStart)
datprep_RSM(X, W, sum0)
stop("RSM can not be computed since number of categories are not the same for each item!\n")

Here is the commands that i entered:

Code: Select all

snowRMM::rasch(
    data = data,
    vars = vars(VNOS.1, VNOS.2, VNOS.3, VNOS.4, VNOS.5, VNOS.6, VNOS.7, VNOS.8, VNOS.9, VNOS.10, VNOS.11, VNOS.12, VNOS.13, VNOS.14, VNOS.15, VNOS.17, VNOS.18, VNOS.19, VNOS.20, VNOS.21, VNOS.22, VNOS.23, VNOS.24, VNOS.25, VNOS.26, VNOS.27, VNOS.28, VNOS.29),
    step = 4,
    type = "PCM",
    imean = FALSE,
    imeasure = FALSE,
    wrightmap = FALSE,
    aic = TRUE,
    bic = TRUE,
    caic = TRUE)

It seems to me the jamovi is trying to do an RSM even though i am actually asking for a PCM.

[seol]

Hi graduallywaiting

Could you send me an OMV file that produces errors at snow@cau.ac.kr? I will take a look at.

Cheers
Seol

[graduallywaiting]

Hello Dr. Seol,

I sent the file to your email using this email:graduallywaiting@protonmail.com

Thank you in advance.

Best regards,
graduallywaiting

[seol]

Hi graduallywaiting

Thanks for your email. Recently, I added eRm R package to run analysis. To run analysis,

1. the bottom category should be coded as 0.
2. The minimum and maximum values of a category must be the same across all items.

These are the requirements of eRm R package to run analysis, Otherwise error message will be happened.

I sent OMV file that is started with 0 to you and please check the OMV file.

One solution is to change the minimum or maximum value of items that are not identical, just one or two.

Cheers
Seol

[seol]

Hi graduallywaiting

I fixed a bug and implemented LR test, Martin-Loef Test, and Wald test in Rasch model within snowRMM module(ver.5.2.1) Update it from jamovi library now.

Cheers
Seol

[graduallywaiting]

Hello again Dr. Seol,

Thanks for your help. I was able to work out the problems in that i had in the previous data set. Also thank you adding more useful tests for the rasch models.

Regarding the Martin Loef's Test, I assume this test is for testing the unidimensionality assumption of the rasch model right?
I tried running it in some of my dataset, and it always gives a not significant result, i.e. the model is unidimensional. This result is consistent even with those models that is deemed to be have multiple dimensions based on theoretical basis and from a factor analysis perspective.

I have a data set from a questionnaire that was deemed to have at least two dimensions based on theoretical grounding, and as EFA, PCA and CFA shown to have multiple dimensions.

But when I do the Martin Loef Test it always shows that model is unidimensional.

My question now, is it appropriate to use the Martin Loef test for this case? Or is there any assumptions that are not met for this test?

Thanks again

[seol]

Hi graduallywaiting

This is a statistical issues not model issues. As far as i tested statistical values using R and snowRMM module for Martin Loeaf test, the results are same shown below.
Cheers
Seol