Page 1 of 1

Missing values in snowIRT Distractor Analysis

Posted: Wed Nov 18, 2026 7:22 pm
by Belzile
Hi all,
I have a question about missing values in snowIRT Distractor Analysis. Apologies if this has already been addressed elsewhere. I looked through the forum but may have missed it.

My items contain five response options (A–E, coded 0–4), with 99 used for non-responses. If 99 is defined as missing in jamovi, Distractor Analysis will not run. If it is not defined as missing, the analysis runs, but “No Response” is treated as another response category and receives its own discrimination statistics.

I don't recall encountering this issue with previous versions of snowIRT, although I may be mistaken.

Is there a way to exclude non-responses from the statistical calculations while still obtaining the complete distractor table for each item (frequencies/proportions and discrimination indices for A–E)?

One reason I find this function particularly useful is that it provides the discrimination statistics for each distractor, not only the overall item statistics available through Reliability Analysis.

Thanks! :slightly_smiling_face:

Re: Missing values in snowIRT Distractor Analysis

Posted: Thu Nov 19, 2026 12:22 am
by seol
Hi Belzile

Thank you for using snowIRT and for reporting the issue with missing responses.

Distractor Analysis in snowIRT 5.3.0 now handles missing values in both the results and the plots.

- Missing responses. Values defined as missing in jamovi, such as 99, are excluded from each item’s response counts and proportions. The analysis can now run with missing responses.
- Updated plots. The Distractor plot and empirical item characteristic curve have been rebuilt using ggplot2. Both use respondents with complete scores across the selected items. The Distractor plot uses the same score groups and response proportions as the results tables.

Version 5.3.0 is expected to be available in the jamovi library next week. Thank you again for taking the time to share your feedback.

Best Regards
Seol