Data management/Multiple response questions

As already mentioned, multiple response questions can be incorrectly analysed and interpreted when the wrong denominator is used for calculating percentage responses.

For example, if it is desired to compute the percentage of the homesteads in which boys are used to make dairy products, the researcher has two choices for the denominator.

• It could be the total number of homesteads.
• It could be only those homesteads which are involved in the activity.

The two results have different interpretations.

Alternatively, the researcher may wish to compute the number of times that boys are used to make dairy products as a percentage of the sum of times that all family member categories are involved in the activity. This results in a different denominator and a different interpretation.