Multiple response sets

[baltzis]

Environment (for bugs)

• JASP version:
• OS name and version:
• Analysis:

  Environment (for enhancement requests)

• Enhancement: Descriptive statistics for multiple response sets.
• Purpose: Many surveys include multiple response questions that need to be calculated. I think it would be very useful to include some enhancement at least for descriptive statistics for this type of data.

[vankesteren]

Hi Baltzis,

Thanks for your suggestion! Could you provide an example dataset with such multiple response sets, and an example of what type of descriptives you would like to see? Then we can judge what preference to give to this enhancement request 👍

Erik-Jan

[baltzis]

Thank you for considering the suggestion. Actually, this is a common type of measurement in social sciences and I am providing an example in the attached file as well as the output from SPSS extracted using the menu "Analyze/Multiple response" and/or the menu "Analyze/Custom Tables" (after defining the multiple response sets):

In this example, a multiple response set measures a range of activities in a field (e.g. engagement with visual arts). The researcher needs to know the distribution of the studied population in a variety of specific activities (painting, sculpture, installations, etc.). The question includes the possibility to give multiple answers, because in real life most of the artists are engaged in more than one activities. In this example, the sub-questions are binary (Yes/No), but there could be a range of answers as well (e.g. in a set of questions forming a Likert scale). The multiple response set (in SPSS) calculates the frequencies and percentages of "Yes" (or of the specified answer) for both the total of answers and the total of cases.

However, another situation is when the frequencies and percentages of both answers ("Yes" and "No", or a range of answers if the question includes more categories) are required (see the last couple of tables in the output).

There are options, of course, to define or exclude an answer as missing value and decide whether you need it in the table or not.

Regards Alexandros multiple responses.zip

[vankesteren]

Thank you for the example files! If I understand correctly, this enhancement request is generally about custom tables where input variables can be grouped, whether it is multiple choice questions or other grouped responses.

I will discuss with the team what the desired priority is for such functionality, which determines in which timeframe you can expect to see it in JASP.

Erik-Jan

[baltzis]

Thank you very much for your answer. You understand correctly.

Another aspect of the same problem is to get descriptives for scores of the items included in a scale (not just percent of answers). In the example attached you can see the data from an actual survey concerning a scale to measure the importance assigned by visual artists to a variety of their working conditions (Cronbach's Alpha = .837). In the output file you can see descriptive statistics concerning the scores for the 17 items included in the scale. In SPSS while the percent of answers for grouped items are obtained through multiple response sets, descriptives for scores for grouped items are obtained through "Reports/Case summaries". However, I think the problem is the same: how to get results for grouped items, which is a common situation in social science surveys.

Regards Alexandros scale.zip

[baltzis]

Sorry, my mistake. I forgot that score means and standard deviations for each item are produced by reliability analysis. However, to get sums, counts and percents, the suggested enhancement is needed.

Regards Alexandros

[HoracioSD]

Hi, I´m an SPSS user but exploring alternatives to it. I would like to know if Jasp can deal with multiple response variables (used in surveys, in questions with many optional answers, with chekboxes). Can Jasp run frequencies managing this kind of variables?

Thanks in advance!

Horacio

[tomtomme]

@HoracioSD Sorry for the late response. This is not yet implemented.

[tomtomme]

@HoracioSD & @baltzis While this issue focuses on the descriptives of multiple response variables, it still seems to be a duplicate of the more broad IRT discussed here: https://github.com/jasp-stats/jasp-issues/issues/160 Thus I am closing this as a duplicate. Lets discuss there. Please reopen, if you disagree!

[shun2wang]

that's different things I think😉