Rasch measurement using SPSS

[@xiexiao_orlando35]

Thank you Dr. Vahid for generously sharing your knowledge in this channel. I have been marathon-watching all your videos using SPSS, AMOS, WINSTEPS, FACETS and CITESPACE, basically a total of 22 hours in a few days. It is so wonderful to recap the CTT, SEM and RASCH in this way.

[@richardkirkden3121]

Great video, really clear! I had no trouble following the steps and running the Rasch model for a set of responses to the General Environmental Behaviour (GEB-50) scale. But is it possible for me to go further in SPSS and get ‘individual person scores’, i.e. to find where individual respondents are located on the GEB scale? I might be confused about what the Rasch model is for - I know it can be used to validate a psychological scale, but articles I have read seem to have gone on to estimate person scores in order to use the validated scale to quantify people’s abilities or values (in this case their attitudes about the environment).

[@abdullahalfarwan1458]

Thank you for this great explanation

[@shaliniverma6841]

thanks for this video sir I have tried many time but it shows the same messag Warning messages: 1: running command '"C:/PROGRA~1/R/R-33~1.0/bin/x64/R" CMD INSTALL -l "C:\Users\xyz\Documents\R\win-library\3.3" C:\Users\xyz\AppData\Local\Temp\RtmpOEyptg/downloaded_packages/polycor_0.7-10.tar.gz' had status 1 2: In install.packages("ltm") : installation of package ‘polycor’ had non-zero exit status please suggest what to do

[@rogerturingan6141]

great video sir... Thank u so much... May i ask if u have a video that uses rasch analysis in scale validation using jamovi?

[@VahidAryadoust]

JAMOVI has Rasch and IRT extensions but they did not work for me. I have asked the developers to see if they can resolve the problem.

[@roldancardona5335]

Perhaps you can as well share some videos on extended Rasch in SPSS

[@VahidAryadoust]

I will soon release more videos n Rasch. Please keep an eye out for them.

[@roldancardona5335]

@@VahidAryadoust Thanks

[@sb9190]

Please teach DIF calculated using SPSS

[@techplusexposing196]

how can I transform the dichotomous data of mine to likert scale using Rasch analysis?

[@suhaalmuhaissen2795]

Thank you for the video it is very useful. My problem is I need to conduct partial-credit Rasch-model for ordinal variables not dichotomous variables, how to do this?

[@VahidAryadoust]

Please watch the video on the validation of Likert scale in this channel.

[@mentalvalve]

Is the data available to be downloaded for study purpose?

[@luzberenicelopezhernandez1634]

Very useful

[@lethecat]

Thank you for this helpful sharing. May I ask a very basic question? In the Rasch model, is the ability in the item characteristic curves equivalent to the total score in the first graph? If it is, what is is the transformation to obtain the ability from the total score? If it is not, how do we get the ability scores then? Thanks again!

[@VahidAryadoust]

I suggest you read 'Best Test Design". You will find your answers there.

[@lethecat]

@@VahidAryadoust thank you very much for your response. I will read that.

[@lethecat]

@@VahidAryadoust Oops, I have SPSS28 on my computer, and it allows me to install the extension...

[@joannabarbararoman7355]

A question: is this method suitable for the Likert-style data? Let's say I have 16-item scale with possible responses 1-7 and I recode the first item into 7 different 0-1 variables (1r1, 1r2, 1r3, 1r4, 1r5, 1r6, 1r7), and do the same for the rest of the items. Should I do 16 separate analyses in SPSS by putting only variables for one item each time?

[@VahidAryadoust]

Since your items are dichotomous, the analysis can be run. If you are not concerned about multidimensionality, you can include all the items in the analysis. You can also watch my video on jMetrik (free IRT/Rasch software).

[@redlove2626]

the video is much appreciated.thanks a ton. Which is advisable or preferable-Rasch model using Winsteps or SPSS?

[@VahidAryadoust]

They use different estimation methods. Whichever you use, ensure to re-run the analysis several times and compare the results. This helps you to get an idea about parameter stability.

[@redlove2626]

@@VahidAryadoust is there something like winsteps is more advanced? Or SPSS would serve the same purpose? As I have SPSS but Winsteps I will have to buy .

[@VahidAryadoust]

@@redlove2626 I do not think so. But Winsteps provides you with more in-depth output.

[@redlove2626]

@@VahidAryadoust thanks for your reply. One last question please. Does the sample size make a difference in the results of rasch analysis. I have a pilot data of 125 responses and the main data of 700 responses. So I have done EFA on 125 responses. Now I want to apply rasch model to check reliability of variable having binary coded items. I am now confused whether it will be done on 125 or 700? Please clarify

[@VahidAryadoust]

@@redlove2626 yes, sample often makes a difference. I suggest you check the reliability of both the samples.

[@mehdimontazeri2792]

Thanks for the great video. I do all steps but when I click OK to run rasch, It doesn't working. There is also some differences. I use spss 27 & R 3.6. What should I do?

[@mehdimontazeri2792]

I also tried your versions but there was no defference. not working

[@VahidAryadoust]

@@mehdimontazeri2792 Sorry to hear that. Please try R 3.5.

[@Belphagor13]

Would t you just use RUMM instead?

[@bangcan391]

Say i did all the steps but when i tried to install the Rasch extension it says that i have not installed the R Plugin which i already did, it does not seem to say error when configurating the plugins, any tips?

[@bangcan391]

Nevermind, i find the fix both SPSS and the R plugin need to be installed in the same storage or drive, cheers

[@VahidAryadoust]

I cannot figure out what might have wrong. Some tips: ensure that you have installed the right versions or R and SPSS as in the video. also, sometimes restarting your machine can help!

[@marcelocoppi4254]

Hello, I'm trying to instal it on version 27 of SPSS, using R36, however, after doing everything when i click ok to do the analysis, nothing happens. Do you know what could be happening? TKS

[@VahidAryadoust]

Hi, have you tries to install R35? I am not sure if SPSS 27 is compatible with these; perhaps ask IBM SPSS.

[@marcelocoppi4254]

@@VahidAryadoust I'll try that...tks!

[@sebastiann7279]

Is the SPSSINC RASCH working if my SPSS has R 3.6. .... I installed it but is not working at all

[@VahidAryadoust]

Try R 3.5.

[@amysugiati1972]

do you have any suggest for me about, how is measuring scor exam using r program ?

[@matthewcourtney8683]

My team and I built the following R program online to do Rasch analysis: autopsych.shinyapps.io/version_1_0_0/ with exposition of app available here: journals.plos.org/plosone/article?id=10.1371/journal.pone.0257682

[@deanehrlich]

Can R 3.5 be installed on a Mac?

[@VahidAryadoust]

According to the following link, R 3.5 can be installed on MAC, too. I have not test it, but I hope it will work for you. www.r-statistics.com/2018/04/r-3-5-0-is-released-major-release-with-many-new-features/

[@markianbonganay4059]

do you have link also for the version 26 of spss? cuz I cant use the r35 cuz my spss is still version 25

[@VahidAryadoust]

Use R3.3 with version 25. Trial 26 version: www.ibm.com/support/pages/downloading-ibm-spss-statistics-26

[@matthewcourtney8683]

Have a go at dichotomous and partial credit Rasch analysis here: autopsych.shinyapps.io/version_1_0_0/ with exposition of web app here: journals.plos.org/plosone/article?id=10.1371/journal.pone.0257682

[@Penantang007]

Why when I run it, the error mesagges appear: Error # 1. Comman Name: SPSSINC, give me solution, please..thank's

[@VahidAryadoust]

I have not seen this error. I suggest you reinstall the R package. It might help.

[@kamilj.3893]

Does it work for non-binary data?

[@hkarimi7358]

Is it possible to do the Mixed Rasch Model in spss?

[@VahidAryadoust]

I do not think so. The micros for such an analysis are not available yet, from what I know.

[@matthewcourtney8683]

By mixed do you mean inclusive of both dichotomous and partial credit items? Have a go at dichotomous and partial credit Rasch analysis here: autopsych.shinyapps.io/version_1_0_0/ with exposition of web app here: journals.plos.org/plosone/article?id=10.1371/journal.pone.0257682

[@soehartosoeharto8471]

Can we run CFA using SPSS?

[@VahidAryadoust]

You should have AMOS for that. Previously, AMOS was available as an option in SPSS, but they have now separated them.

[@matthewcourtney8683]

Look up how to use the lava an package in R.

[@roldancardona5335]

STATS IRM /VARIABLES ITEMS=Item16 Item17 Item18 Item19 Item20 Item21 Item22 Item23 Item24 Item25 Item26 Item27 Item28 Item29 Item30 /OPTIONS MISSING = ALLAVAILABLE EXECUTE=TRUE /PRINT SUMMARY ITEMFIT /PLOT /SAVE. Error in solve.default(object$hessian) : system is computationally singular: reciprocal condition number = 1.409e-40 In addition: Warning message: In tpm(data = dta, na.action = missing) : Hessian matrix at convergence is not positive definite; unstable solution. What does this mean?

[@joannabarbararoman7355]

I see "Error in 1:ncol(out) : argument of length 0" and then there are no colourful graphs :((

[@VahidAryadoust]

It seems the R packages has not been installed properly. I suggest you follow the steps once again to ensure the package is linked with SPSS properly.

[@joannabarbararoman7355]

@@VahidAryadoust I am sure everything was installed properly. I am using SPSS 26 and R 3.5.3. I just noticed this error occurs only if I click "descriptives" as an output option. If I untick it, then I have all the graphs... strange

[@VahidAryadoust]

@@joannabarbararoman7355 I cannot figure out the cause of the problem, but for the time being, maybe run the analysis with the descriptive box unclick :-)

[@shyamacharya167]

Could you please show me how dichotomous and polyomas items together?

[@VahidAryadoust]

please see my videos on using jMetrik and Winsteps. You can use a partial credit model for hybrid items.

[@shyamacharya167]

@@VahidAryadoust thanks. I went through it. Could you please provide such your or any other video about using partial credit model by using R? Your videos are really wonderful.

[@VahidAryadoust]

@@shyamacharya167 sure, that is on my to-do list.

[@shyamacharya167]

@@VahidAryadoust Thank you.

[@matthewcourtney8683]

Have a go at dichotomous and partial credit Rasch analysis here: autopsych.shinyapps.io/version_1_0_0/ with exposition of web app here: journals.plos.org/plosone/article?id=10.1371/journal.pone.0257682