Surveys with Multiple Answer Questions

This tutorial explains how to transform tables that include multiple answer questions into a table where the multiple answer questions are transformed in such a way that they can be analyzed.

Dataset for the Multiple Answer Questions

The data are the results of a survey in which 15 consumers were asked to evaluate a variety of strawberries using 7 attributes. Consumers were asked to check off the attributes that best described the strawberry variety.

Goal of the Multiple Answer Questions feature

The raw output of the survey contains replies which are entered one after the other and separated by a comma in the Excel cells (see table below). With this formatting, data analysis and statistical techniques requiring a structured layout cannot be applied.

The objective here is to dissociate each attribute and to define for each consumer whether it has been checked or not. The new data format will allow us to perform statistical analyses, such as CLUSCATA analysis.

Setting up the Multiple Answer Questions dialog box with XLSTAT

• Open XLSTAT

• Select the Preparing data menu and then the Multiple Answer Questions feature.

• Select column B containing all the attributes in the Data field.

• Choose the separator corresponding to the comma in the Separator field.

• Check the Column Labels field since the dataset contains column headers.

• Launch the analysis by clicking OK.

Displaying the results of the Multiple Answer Questions tool

The output table shows the presence or absence of each attribute for each consumer.

The columns corresponding to the multiple-answer questions have been replaced by as many columns as there are answers in the input columns.
In each column, we find Yes/No answers depending on whether the respondents have checked or not the item.
Using this new table format, it may be interesting to classify subjects using CLUSCATA.