Analyzing Other Data by Groups/Segments

You can create Groups/Segments and then analyze other data by these groups in order to see the data behind the segments (or in survey terms, to see how the groups answered other questions in the survey).

1. Create your Groups/Segments as described in Create Groups/Segments.
2. Create a new page and insert a table or chart. In the Report tree, hover and click the + > Table in Displayr. In Q, right-click in the Report tree and select Add Table.
3. You will be asked which data to show. Select the other data you are interested in that you want to correlate with the groups.  For example, if you are curious about the age brackets of the groups, you would select the Age data here.
4. Now that you have a new table or chart, you need to break it down by the groups:

In Q:

1. Click into the Inputs > DATA SOURCE > Columns dropdown in the Object Inspector (that currently shows "SUMMARY").
2. A list of all the data appears, and also at the bottom of the list, observe there is a new data item called "Groups/Segments date time".  Select this data.
3. The groups are now shown in the columns of your table/chart.

In Displayr:

1. In the object inspector , go to Data > Data Source > Columns (that currently shows 'SUMMARY').
2. A list of all the data appears, and also at the bottom of the list, observe there is a new data item called 'Groups/Segments date time'. Select this data.
3. The groups are now shown in the columns of your table/chart.

Please keep in mind that:

• The original visual display of Groups/Segments is actually a model that shows you the probability of people belonging to a group.  For example, a person responding to your survey may have a 25% chance of being in Segment 1, and a 75% chance of being in Segment 2, and the display accurately represents this.
• The "Groups/Segments date time" data (as shown in the screenshot above) - and any statistics shown with it - are categorical data, not probabilities.  This means that the person in the example will be included only in the column of Segment 2, and will not be included in Segment 1 at all.
• Although the categorical interpretation of the data is not as accurate as the probabilities, it is much easier to interpret - people find it easier to understand the percentages of people in categories, rather than understanding average probabilities or numeric correlations.  Just keep in mind that these breakdowns are not as accurate as the original model shown in the display.
• For more information on interpreting the data, please refer to http://mktresearch.org/wiki/Latent_Class_Analysis

Next

Displayr: How to Analyze Data by Groups/Segments