Analysis Multiple response question (categories)
Multiple response refers to the situation when people are allowed to tick
more than one answer option for a question.
Analyzing the answers given will be explained using the following steps:
People are allowed to tick a maximum of two options.
Hence there are two variables needed to process the answers.
The coding below is designed to do just that.
Note: An alternative way of coding multiple response
questions is to use dichotomies. How that is done is explained on the page: Defining Multiple Response Sets.
The key information we want to have is how many times each of the aspects for
fruit juice was chosen by the respondents.
We can view that info by making two frequency tables (for "important aspect 1"
and for "important aspect 2") and manually adding the frequencies for each
aspect.
For example price is mentioned 34 times in aspect 1 and 24 times in aspect 2.
Hence in total the aspect price is mentioned 55 times.
But of course we don't have to do these basic things
manually when we have a sophisticated statistical package
like SPSS.
Step 1 is to specify which variables have to be combined into a set; we use
the DATA menu for this.
Choose Multiple Response Sets
and complete the dialog window as shown on the
right.
The Categories
range from 1 through 5.
Note: The same dialog box can be reached
through:


Once the set has been created we can use it to make a frequency table for the
set.
Choose Analyze > Tables
> Custom Tables and drag the set variable $V05 to the
Rows position.
We can specify
the statistics and (sub)totals we want using the
two buttons shown under Define:
We have chosen for the following options: 

The result is:
 90 respondents have mentioned at least one aspect
 55 respondents indicated that price was an important aspect.
That is 61% of all people who responded and it is 31% of all the
answers given.
 In total the 90 respondents have ticked 55+52+34+26+10=177
aspects. So almost everyone did indeed tick two options as
requested.
There is an alternative approach
possible by using from the menu:
Analyze > Multiple
Response.
Again we have to define the variable set first. This is
done in a similar way as shown above.
But in this approach the definition of the set is
not saved in the data file. So this description can
only be used during an analysis run. Once you close
SPSS this info will be lost.
Once the Variable Set has been defined the menu
options Multiple
Response > Frequencies...
and Multiple
Response > Crosstabs...
become available. So now you can ask for a frequency
table for the group you just made.
The result looks like this:


Using this approach the information we obtain is presented slightly differently.
But from the output we infer:
 90 out of 100 respondents answered this question
 Those 90 people ticked a total of 177 boxes, which
is almost 2 boxes per respondent.
 Again we have two sets of percentages. One column
uses the total number of responses (177) as base value
for the percentages and the other column uses the number
of cases (90) as base.
61% of the respondents considered price to be an
important aspect. That is 31% of all the answers.
