Quota cells and the one that never fills
Every study has a cell that stalls at sixty percent. Which cell it is was decided before the study went live.
Outline. Each heading below carries one line saying what the section has to establish. Expand each into two to four paragraphs, put at least one number or worked example in the piece that nobody else could publish, then delete this note.
The last cell is never the one you expected, and by the time it stalls the options are all bad. The decision that caused it was made when the quota grid was drawn.
Interlocked and nested are not the same grid
Show the same targets as independent margins versus a full interlock, and the reach each implies.
Rarity multiplies, and it multiplies faster than people estimate
Worked example: three conditions that are each common producing a cell that is not.
The stall shows up at seventy percent, not at the start
Explain why the easy cells fill first and mask the problem for the first week of field.
Four ways out, ranked by what they cost the data
Loosen the cell, extend field, blend a second source, or re-weight. Say which one you would defend to a client.
Per-cell fill reporting, and why most buyers never see it
A quota report that only shows the total is hiding the exact number that decides the timeline.
-
Healthcare sample and the cost of verification
Self-reported clinical roles are wrong often enough that the whole category is priced around finding out.
9 min read -
Writing a screener people cannot game
Most screeners announce what they are looking for in the first two questions, and then the researcher is surprised that everyone qualifies.
6 min read -
Your B2B incidence rate is probably wrong
Two people can compute incidence from the same fieldwork and disagree by a factor of two, both correctly. The disagreement is almost never about arithmetic.
5 min read
