Analysis Multiple response question (dichotomies)

Multiple response refers to the situation where people are allowed to tick more than one answer option for a question.
Analyzing the answers given will be explained using the following steps:

• The question
• Coding in SPSS
• Creating a frequency table for multiple response items (method 1)
• The result of method 1
• Interpreting the output
• Creating a frequency table for multiple response items (method 2)
• The result of method 2
• Download the sample data file

The question

Coding in SPSS

Do you understand why we don't need missing values in this situation?

Creating a frequency table for multiple response items (method 1)

Before we can make a table we have to define the variable set first. In method 1 we use from the menu: Analyze > Multiple Response > Define Variable Sets... You choose the (dichotomous) variables that together make up the set. You also tell SPSS which of the two codes has to be used for counting. In our example that is the code 1 (=ticked).

Please pay attention to the note in the definition window. This definition can only be used for the Analyse > Multiple Response method.
When we apply method 2 (see below) we will have to define the Variable Set in a different way through the DATA menu.

Once the set has been defined it can be used to make a frequency table by:

The result of method 1:

We have edited the frequency table only a little. Here you see it together with some info about the cases.

Interpreting the output

• 48 respondents ticked at least one of the options. Two respondents either skipped the question or never use the internet.
• Those 48 respondents ticked 148 boxes in total. That is an average of slightly more than 3 per respondent. Apparently people use the internet at a lot of places.
• It is a little surpising that using the internet at friends/family has the highest score; 87.5% of the respondents does it.
  This option has a "market share" of 28.4%.

Creating a frequency table for multiple response items (method 2)

This time we use the Data menu to define the multiple response set.

Once we have defined the multiple response set we can use it.
Choose Analyze > Tables > Custom Tables:

Drag the set variable $V05 to the Rows position. We can specify the statistics and (sub)totals we want using the button shown under Define:

We have chosen for the following options:

The result of method 2:

Note that the totals that are reported here are different from the ones of method 1.
The total for Count says 48, which is the total number of respondents who ticked at least one option. The Column N% total is based on that base number, so it shows 100%. Hence these two totals are not equal to the sum of the values in the columns above them.
So unlike method 1 you can't see that there are a total of 148 boxes ticked by the respondents and that it adds to 308%, so to an average of about 3 ticks per respondent.

Note: Custom tables offers you a lot of options with respect to the content and the layout of the table, far more than method 1 does.
Experiment with it and find out what it is capable of.