Combining multiple choice data

[Brenna Quinn]

Hello,

I am still with a data set from a survey I administered in Survey Monkey. I am attempting to see if I can "combine" in a way some variables.

School nurses were asked what kind of students they use pain assessment methods on. For example: "For which students do you use objective clinical assessment to assess pain?" Answers were mild special needs, moderate special needs, severe special needs, and profound special needs. The survey item item was select all that apply.

Because it was select all that apply when the data was loaded into STATA, 4 different variables were created. Each variable only counts how many respondents checked that certain option.

Example:
tab Mildstandardapproach

Mild standard approach | Freq. Percent Cum.
---------------------------------+-----------------------------------
Students with Mild Special Needs | 597 100.00 100.00
---------------------------------+-----------------------------------
Total | 597 100.00


Do you know if it is possible to combine the different variables? I am attempting to do so in order to make comparisons between the degree, license, and years of experience a nurse has with how they assess pain for different students.

brenna

[daniel klein]

The survey item item was select all that apply

From a methodological point of view, it is better to have explicit "yes/no" buttons for each of the items/categories for two major reasons.

1. Respondents get tired quickly, lose motivation and/or are "lazy". Empirical studies have shown that items presented later in a list get checked significantly less often.

2. Items that are not checked can mean they do not apply or the respondent skipped the item (which would best be represented by a missing value). If there is no explicit "yes/no" button ,there is no way of distinguishing between these two possibilities.

In a short list, where all the items all seem to capture one dimension (intensity of needs), these might not be a big problems. Anyway, if you plan on collecting data, think of this aspect when designing the questionnaire.

Do you know if it is possible to combine the different variables? I am attempting to do so in order to make comparisons between the degree, license, and years of experience a nurse has with how they assess pain for different students.

I think this question goes beyond the technical aspect/commands to combine variables. As mentioned above, it seems the items all measure the same dimension: the intensity of needs. You need to carefully about how exactly you want to combine the information. How would a combined variable look like? Would it record the highest value a respondent checked? The lowest? Something else?

Also, it might help if you tell us more about what you are trying to find out with your comparisons. In other words, what is/are your research question/s? What are the hypotheses you want to test?

Best
Daniel