When configuring quotas for a survey, you can define targets in two ways:
As a percentage of total completes (e.g. 25% should be Gen Z)
Or with a minimum number of respondents (e.g. at least 50 must be Gen Z)
Quotas defined using a percentage of total completes are used for weighting, and those with minimum numbers of responses are used for boosting. You can combine these in any combination and his guide explains how these values interact to determine the number of responses to collect for each group.
1. The Higher of Percentage or Minimum is Used
For each quota line, we calculate:
The count implied by the % of total completes
The minimum_respondents, if defined
We then use whichever is larger.
This ensures small but important groups are adequately represented, while still allowing you to model your sample using proportions.
Example 1
Target: 500 completes
Quota line: LGBTQ+
total: 5% → 25 people
minimum_respondents: 50
RESULT: We use 50, because it’s higher than 25.
Example 2
Target: 1,000 completes
Quota line: Group A
total: 50% → 500 people
minimum_respondents: 300
RESULT: We use 500, because it’s higher than 300.
2. Remaining Quotas Are Recalculated Proportionally
Once any quota lines with a high minimum have been locked in, the remaining completes are distributed across the rest of the lines based on their percentage values, normalized to the leftover total.
Example 3
Target: 1,000 completes
Group A: total: 50%
Group B: total: 30%
Group C: total: 20%, minimum_respondents: 300
Calculation:
Group C would get 200 by percentage (20% of 1,000), but the minimum is 300 → we lock in 300
Remaining completes = 1,000 - 300 = 700
A + B’s combined % = 80%
A = 50/80 = 62.5% of 700 → 438
B = 30/80 = 37.5% of 700 → 262
RESULT:
Group A: 438
Group B: 262
Group C: 300
3. Rounding Adjustments
Suppose the final allocations don’t add up exactly to the target (due to rounding). In that case, the system will adjust by adding or removing a small number of completes from a flexible quota (one without a minimum).