The following article guides you through a step-by-step process of displaying multiple data measurements, such as percent and values, simultaneously in one table or chart.
Let's use the ‘Distribution of answers' and 'Mean overall rating’ examples.
When analyzing the responses to rating-scale questions, researchers commonly record the results in a table, which contains the following rows (or columns):
‘Source’ ratings in percent (distribution of answers);
Top-N and Bottom-N (for example, Top-2 and Bottom-6 for a 10-point scale or Top-2 and Bottom-2 for a 5-point one);
Average (Mean) overall rating;
Number of respondents (sample size) and so on.
In general, this table can look like this:
Show rating-scale results in a table
Go to the Meta-Editor, find rating-scale variables related to different brands (step 1), and make sure the codes of their values are correct (step 2). If they are not, change them manually (step 3).
Select all the variables you need and group them. When variables are grouped, any edits you make to one variable will also happen to the rest. This can be an efficient way to make consistent edits, such as renaming or recoding variables.
Choose one variable from the group and set up the appropriate values into ‘Top-2’ and ‘Bottom-2’ values using the OR function. The same values will automatically appear in other variables in the group.
Again use the OR function to recode all the meaningful source values, from ‘1' to '10’, into the ‘Mean’ value. This will present the average overall rating. The final set of values looks like that for all grouped variables:
If the grouped variables have very long titles, it might be inconvenient to work with them later. You can gracefully shorten them by cutting the common part. To do it, select one variable and click the Trim icon above. In the open dialog box, highlight the unnecessary part (usually it is the common part and it is highlighted by default) and then click OK.
Go to the crosstab interface and choose a GRID-Report mode:
Find the variables you have adjusted, choose one, select only ‘Top-2’, ‘Bottom-2', and ‘Mean’ options and drag them to the Rows. The grid works in such a way that these three options will populate columns and the variable will populate a row. Select and add the other variables to the rows by selecting from the pop-up menu 'As Variable’.
Select only the ‘Mean’ column and choose a ‘Mean’ metric for it instead of ‘Vert%'. Also make sure that ‘Vert%’ is selected for the other columns: you need it to get the respective proportions to the total number of respondents.
Click the Calculate button above to get the results. And now you can switch to the Chart mode.
Show rating-scale results as charts
Adjust settings on the CHART tab:
Select ‘Vertical Bars Stacked’ as the chart type;
Set Series in Columns;
Switch on Data labels, select the ‘Show data markers’ check box;
Switch on Legend and align its position to Left-Middle.
Adjust settings on the SERIES tab:
Switch on ‘Autonaming’;
Go to the Y-axis tab to set the scale there from 0 to 100;
In the submenu of the ‘Mean’ series select Line as the chart type;
Adjust colors of series if you need to;
Now, you may move the mean to a separate group, for which a different y-axis will be set and the mean will look visually more expressive:
Here is the final chart view, thet can be exported to a dashboard:
