Creating a Correlation

In the following example, we are going to correlate a set of statements regarding different aspects of quality with a total quality score, to see which areas have the most impact on total satisfaction. In the example, the statements have been collected in a grid.

1. Add the Grid to a new aggregated table in rows, as shown belowhere.

Note that the header has the “collapsed” property set.



Figure 1 - Example of a statement grid added to a table

2. Add the correlation object to the column area by selecting it from the table objects area of the table toolbar and dragging it onto the column area.

The correlation header will indicate that no variable has yet been assigned to this header .



Figure 2 - The empty correlation header

3. To assign the variable that contains the question that you wish to correlate your individual categories to (in this case, a variable that contains a total quality score), drag the variable from the data source directly onto the correlation header item (the blue box). 

When the variable is dragged correctly onto the correlation header item, it will turn green to indicate that you may now drop the variable to assign it to the correlation header .



Figure 3 - The correlation header item ready for the variable

4. Refresh the table to return the correlation value of overall satisfaction for each of the statements in the rows .

In this example, the statement concerning Ownership (with a correlation value of 0.94) is most likely to impact overall satisfaction, whilst Content (with a correlation value of 0.19) is least likely to impact total satisfaction.

Figure 4 - The resulting correlation table