How to analyze multiple responses in SPSS
Generating New Variables in SPSS: The Multiple Response Command
The table below shows part of a Health Issues survey questionnaire.
Question:
Thinking about health related matters, did any of the following happen to you in the last 1 Month?
Response | ||
Response | Yes | No |
I was ill enough to go to the doctor | ||
I sought counseling for mental problems | ||
I had problems with infertility | ||
I suffered from a drinking problem | ||
I used illegal drugs | ||
My child had to go to hospital | ||
My partner had to go to hospital | ||
A close friend died | ||
My child suffered from drug or alcohol problems | ||
The answers to such a set of questions are regarded as multiple responses, since the answer to each does not preclude an answer for the others; the responses are not mutually exclusive.
The answer to each of these questions is regarded, for the purpose of SPSS coding and data entry, as. a separate variable. That is, a column is set up for each of the items to which a respondent can answer Yes or No. In this instance, therefore, there are 9 columns of data; one containing either a 1 (= Yes) or 2 (= No) for each case according to whether they were ill enough to go to the doctor, another column containing 1 or 2 for each case indicating whether they had sought counseling for mental problems, and so on.
To code this in SPSS we use the code hlth1, hlth2, hlth3…hlth9
After coding and typing the respective labels, SPSS will look like the figure below:
Figure 1 - Variable View
The data view will look like the figure 2 below after data entry
Figure 2 - data view for multiple responses
With the data entered in this way, if we wanted to see the number of Yes responses for each of these variables we would have to generate 9 separate frequency tables and note the number of Yes responses in each.
An alternative, which also allows us to do further analysis, is to use the Multiple Response command. The Multiple Response command allows us to analyze a number of separate variables at the same time, and is best used in situations where the responses to a number of separate variables that have a similar coding scheme all ‘point to’ a single underlying variable.
In this example, we can consider each of the items in the question as all pointing to the state of a respondent’s health. They are particular operationalizations, each of which captures just one dimension of this complex variable. It is therefore interesting to summarize the responses to these items at once, and to be able to use the pattern of responses across these items in further analysis with other variables, which is exactly what the Multiple Response command allows us to do.
To use the Multiple Response command we initially have to set up a Multiple Response
Set. This procedure instructs SPSS to group together the responses across a range of variables.
Before doing this it is important to have noted the coding scheme for the items that will make up the Multiple Response Set. In this instance, the coding is 1 - 9 for the various health conditions. We note this because we need to tell SPSS which value (or range of values) is of interest to us. Here we are interested in all the Yes responses to each item.
Figure 3 - Define the Multiple Response variable sets
Transfer the Multiple variables to the right
Figure 4 - defining the variable categories
Since we have 2 responses (Yes, 1 and No, 2) we select categories 1 to 2.
Now that we have defined the multiple responses set we can analyze it. The ‘new’ variable that we have just created and called hlth_con does not appear on the Data Editor window with the existing variables, but is stored in SPSS’s memory, and is accessed only through the
Analyze/Multiple Response/Frequencies command. It won’t appear in the normal dialog boxes we are familiar with which present the variables in the data file in a source variable list.
It is also not saved with the data file and will disappear when the file is closed, so it is a good idea to perform all the analysis you plan to undertake using the multiple response set before finishing your current SPSS session.
The simplest analysis we can undertake on a multiple response set is to run a frequency (Figure 5) on the new variable (Though you can also perform crosstabs
To analyze multiple frequencies
Figure 5 - Analyze Frequencies
The SPSS multiple frequencies command. Then Press OK
