Edit

Share via

Viva Glint aggregate groupings - trends and best practices

  • Article

Aggregate trending informs historical data over time. Trend points reflect matching items from one cycle to the cycle immediately preceding it. Changing items within an aggregate (group) interrupts trend.

Term Definition
Aggregate A grouping of two or more items that measure a concept within a survey
Aggregate score The average score of all the items divided by the number of item responses
Trend Any difference in the score of a single item from one cycle to the previous cycle
Aggregate trend Any difference in the combined score from one cycle to the previous cycle. Aggregate trend exists only when all items in two consecutive cycles are identical. 

Tip

It's best not to use aggregate groupings in surveys.

Aggregate grouping can lead to unreliable aggregate trend

Why not to use aggregate groupings in surveys:

  • Surveying frequently with individual questions produces more useful, focused results.
  • Measuring single items are better than multiple-item measures when it comes to utility and reliability. They can predict future behavioral outcomes, and they can do it consistently over time.
  • Grouping multiple items may limit the intervention opportunities for a single item within that group. An important item can become lost inside the group of items.
  • Action taking should only happen at the single item level. Aggregate scores aren't recommended for action taking decisions.

Caution

  • Trend is unreliable when it isn't measuring apples to apples. When there are differences in the items of an aggregate between the current cycle and the same aggregate in the previous cycle, trend is undependable.
  • Trend is shown when at least one item is identical in the same aggregate in two cycles. This means that unless all items are identical, the aggregate trend isn't measuring a consistent group of items and isn't as reliable as single item trend data.

Aggregate trend examples across 2 cycles

There's a lot to consider when comparing survey cycles with aggregate groupings.

Use case 1

Three items are part of the aggregate in both cycles, but because they aren't the same three items; no aggregate trend forms.

Chart showing that Cycle A and Cycle B don't share the same three items.

Use case 2

One item is removed from the aggregate in the second cycle. Cycle A contains three items, but Cycle B contains only two items. Aggregate Trend doesn’t exist because the aggregate isn't identical; there's no aggregate history. Individual items show trend but they don't reflect upon the aggregate as a whole unit.

Chart showing the 2 cycles share no history because the aggregate trend isn't identical.

Use case 3

A new question is added to the aggregate grouping. A new item is added to Cycle B so no aggregate trend is associated. There's no identical history.

Chart showing that no aggregate history is associated between the 2 cycles.

Aggregate trend across three or more cycles

Important

Trend always looks at the previous cycle.

Use case 4

Chart showing only the last 2 cycles are considered for aggregate trending.

Use case 5

Chart showing aggregate trend becomes more complicated across multiple cycles.