This has been great. As a review and a refresher it is fantastic. As an introduction for a person who is learning Rasch, it would be very good as well. I hope that other people take some time to watch this if they are getting into Rasch or using Winsteps. Much appreciated.
thanks for a really good video. I have 2 simple (I hope) questions: How do I tell the program to put item labels into a table (say a wrightmap). Also, I I want to run the analysis separately for males/females - Can I tell it to select only males or do I need to use separate data files? Many thanks.
Hi Valid. Thank you for taking the time to make this video. I am new to winsteps and am trying to teach myself. Can you do a video on test form equating and interpreting? Two versions of same test with common items? Actually, the hardest part is actual interpretation. In general in your videos, with your expertise, it would be great if you could add more commentary on interpretation, such as if this were an 'achievement test' you may expect this output...if it were a proficiency test it you expect xxx out put and so forth. Thank you again!
Can I ask for help? I'm processing our research data using Winsteps. However, the data doesn't analyze. Is their a needed format for excel? Please reply. Thank you!
Hello Vahid! Thanks for the AMAZING video. Can I ask you a question? How do I force the Winsteps to apply the Rating Scale Model? (where the distance between the Andrich categories thresholds are the same across all items) Because when I'm analyzing my data, the distance is varying according to items, as in a Partial Credit Model. Thanks in advance :D Pedro
Hi Pedro, The default in WINSTEPS is the Rating Scale Model. Please check the input file. If you have removed the semicolon in the following input line, the model estimated will be the Partial Credit Model. If you there is no semicolon, please add one and run the analysis. ; GROUPS = 0 ; Partial Credit model: in case items have different rating scales
Hi Vahid, You're right! Can I ask one more question? I'm a PhD student working with Rasch Measurement Model. When I use STATA 14 to work with the Rating Scale Model, STATA 14 shows an output table for the difficulty coefficients of each category threshold. I'm analyzing a Likert Scale, so STATA 14 displays the difficulty coefficients for the thresholds of Item 1 (Option0 vs Option1, Option1 vs Option2, Option2 vs Option3, Option3 vs Option4), Item 2, Item 3 etc. Now that I'm starting the analysis with Winsteps, it displays the Category Probability Curves and if I click on "Absolute x-axis", it shows the graph of the probabilities in relation to the "measures on latent variables" (ability coefficients) for each item. But do you know which output shows the actual values for the difficulty coefficients in Winsteps? I'd like to see the same table as I see in STATA 14. Is it possible? Best of all, Pedro
Pedro, If I understand your question correctly, what you need to look at in WINSTEPS is TABLE 3.2. There are two sub-tables there, the one on Top will provide category raw scores and the corresponding difficulty measures, fit etc. and the sub-table at the bottom will give you threshold difficulty measures (or the Tau parameters), e.g., the threshold between Option0 and Option1 , etc.
Hi Simon, This video is about fundamental considerations in Rasch modeling. For DIF analysis, please see my Channel at: ru-vid.com/show-UCfu2GCdjq50W-kL-cv3rcLw?view_as=subscriber
Hello Vahid. Thanks for your informative videos and great published papers about DIF. My question is that is the free version of Winsteps available for downloading? Because I am in Iran and due to US sanctions I can not buy it. The other question is that does the BIGSTEP version that is free, include the statistics for DIF analysis in a large scale? The last question, is there any other free program for DIF analysis in a large scale? if yes which one is preferable? thanks in advance.
I replied to you in an email. I suggest you use an R package. see: hansjoerg.me/2018/04/23/rasch-in-r-tutorial/ Winsteps has a free version called Ministeps, but I am not sure about Bigstep. Please get in touch with Dr Mike Linacre.
Dear Vahid! Rasch analysis and WINSTEPS are new to me. Your step-by-step illustration helps me a lot. I am still struggling with the analysis. My questions are: 1) For item selection, apart from the infit outfit parameters, are there other parameters that are complementary? Like 'Measures', it checks the spread of the items. 2) How about the 'discrimination index'? I found from the manual that the assumption of Rasch model put all items discrimination index equals to 1. However, I found some papers still reporting discrimination index for the elimination of items with low discrimination power. I eventually found the button for discrimination on Winstep. But the output files displayed the label of the question numbers on that column instead of the discrimination index. I got stuck there. Thank you for your reply in advance.
Hi; these are very interesting questions, but it is not easy to answer them in short. Please watch the other videos in the same playlist for answers: ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-JVIa8jszCDQ.html
Just an update: Our comprehensive review of Rasch measurement in language assessment has been published: journals.sagepub.com/eprint/YXCQB3NYCICYVFQ4TSCG/full
assalamualaikum sir, I want ask to you how we make illustration of infit and outfit measure per item, and illustration of average score per measure (chart graph)? I am very difficult to make it.
Hi, I'm Shirleen here. Can I use the Winsteps in measuring the construct validity of the questionnaire? Pls advise me for what are the outputs that I should look at ? Thanks
Shirleen, Rasch measurement (performed via, e.g., WINSTEPS) is a very useful technique for validating of questionnaires. As to what WINSTEPS output to look at, it depends on the "claims" that you make about your instrument. Here is an example how Rasch measurement can be used to validate questionnaires: www.tandfonline.com/doi/abs/10.1080/09588221.2014.1000931 Here is another short paper which shows how Rasch measurement can be used in validation: www.rasch.org/rmt/rmt231f.htm
I want to thank you for uploading this video. It has been very informative, especially in integrating the “text book” definitions in explaining the Rasch Model Analysis. Thanks to you I will be able to apply this software to validate and analysie my thesis multiple choice questionnaire.
@ Norah Albadran: steps to convert excel and other spread sheets into WINSTEPS's input file are explained by Dr Linacre in this link: www.winsteps.com/winman/excelsssmenu.htm
I would like to find resources to prepare data file to analyze a three facet 30 students 6 raters and 5 dimensions of a rubric which I am considering as five items. I would appreciate your guidance.
Item mean (logits) is anchored to zero (is always zero). See this from TABLE 10.1: click "Output Tables , then --> 10.ITEM column. You can find the person mean from Table 6: Click "Output Tables , then --> 6. PERSON (row) scroll down to find the person mean at the bottom of the table.
Here are 2 papers that discuss it at some depth: www.tandfonline.com/doi/abs/10.1080/09588221.2019.1704788 arts.unimelb.edu.au/__data/assets/pdf_file/0008/3241619/8_2_S3_Zhu-and-Aryadoust.pdf
Thank prof, i and my group(30 people) have bought winsteps and facets softwares because of your video. But i hope you can make video about some details related this software, 1. Using excle input file for winsteps software and prepapring data especiallah for DIF 2. Preparing data for facets softwares, some of us will get trouble because of coding data in facets i think. I hope you always healthy and happy there, thanks for your sharing knowledge, i become your fans since you write about Scientometric analysis in frontier journal. Hihihi😄
Hi, there are various ways to test unidimensionality. If you are using WINSTEPS, you can run a principal component analysis of residuals (PCAR), which gives you a good indication the extent to which the test is unidimensional. I will make a video soon about unidimensionality, local independence, and fit. Stay tuned in!
You can also find some relevant information about unidimensionality int these papers: tandfonline.com/doi/full/10.1080/15434303.2011.628632 and tandfonline.com/doi/full/10.1080/15434303.2015.1133626
Hi Valid, try to import excel but recieve this error ; WININPUT 4.6.0 at excelmod line: 22070 Error: 1AD ActiveX component can't create object, do you know why? Ensure only have 1 tab.
Ignatius, you seem to have some missing components. To solve this, Winsteps website suggests: 1. Download www.winsteps.com/a/Winsteps-system-files.exe 2. Right click on the downloaded file and "Run as Administrator" 3. Necessary system support routines are installed and registered. 4. Winsteps and Facets should now function correctly.
