How to Analyse Constant Sum (Sum, Sum %, Average)?
Constant Sum allows respondents to assign a numeric value to each choice in your question (Constant Sum: Choice) OR to each choice in your statements (Constant Sum: Matrix), e.g.
1: Constant Sum: Choice - assign a numeric value to each choice in your question
2: Constant Sum: Matrix - assign a numeric value to each choice in your statements
This is one of the most complex analysis you can do in Glow as we have to analyse open ended numeric answers.
Constant Sum Analysis
Next, let's see you how to analyse:
Using these 3 Values:
Then show you how to Crosstab them
Constant Sum: Choice
Sum
Sum lets you sum the numeric answers given by respondents for each of your choices.
For example, in the Chart and Table below, you can see that the Sum of Brand A was 270 and we can also see n=4 respondents answered for Brand A (not shown in the chart and table are the individual responses for Brand A, which were 80 + 40 + 100 + 50)
Chart:
Table:
Sum %
Sum % divides the Sum value of each brand by the Total Sum value across all brands. For example, if the Total Sum is 1200 (270 + 80 + 70 + 35 + 100 + 115 + 120 + 75 + 45 + 90 + 200), then:
- Brand A captured 22.5% of total sum value (270 / 1200)
- Brand E captured 8.3% of total sum value (100 / 1200)
Chart
Table
Average
Average divides the Sum Value by the Choice Answered Value (NOT the Question Answered Value) to show you the average numeric value given by respondents for each choice. Example:
- 270 / n=4 respondents who saw Brand A = Average of 67.5%
- 75 / n=5 respondents who saw Brand L = Average answer of 15%
Chart
Table
Crosstab
Using any of the 3 Values above: Sum, Sum %, or Average, you can cross your answers across your demographic or segment filters by
- clicking the '+' button in your table columns where your filters are
- selecting the filters you need, e.g. Completes vs. Males vs. Females,
- click Add,
Et voila! You've just created a Constant Sum: Choice Crosstab.
Constant Sum: Matrix
For our Constant Sum: Matrix examples below, we'll use the Sum % as it's the most popular value for this question type, but you can just as easily use Sums or Averages for your analysis.
Sum %
Sum % divides the Sum value of each brand by the Total Sum value across all brands in each Statement. For example, if the Total Sum for each occassions is 1200, then:
- Brand A captured 16% of total sum value for Private Occassions (195 / 1200)
- Brand A captured 11% of total sum value for Work Occassions (130 / 1200)
- Brand A captured 22% of total sum value for Social Occassions (265 / 1200)
We've also used a stack bar chart.
Crosstab
To cross your answers across your demographic or segment filters:
- Change Rows to "Filters"
- Click the '+' button in your table rows where your filters are
- Select the filters you need, e.g. Completes vs. Males vs. Females, then click Add
- Use the Statement dropdown to switch between each occassion.
For example, below we can see that for Private Occasions:
- Brand X captured 19% of total sum value among Males (150 / 800)
- Brand X captured 5% of total sum value among Females (20 / 400)
Constant Sum analysis done! 🎉
You'll notice Constant Sum also functions as a normal choice or matrix question with count, row %, column %, total %. This is RARE to use for Constant Sum but perfectly possible! So if you want to use it this way, here's how: