Queendom Tools

Kish Grid Selection Generator

Kish Grid Respondent Selection

For field survey coordinators running face-to-face household surveys who must pick one adult per household without carrying a printed Kish grid booklet, this generates the correct table row and the selected member on the spot, so within-household selection stays random and sampling bias is reduced. This tool generates a Kish grid selection table for random within-household respondent selection to reduce sampling bias.

The Kish grid is a systematic method to select a single respondent from multiple eligible members within a household, ensuring each member has a known, non-zero probability of selection. This prevents interviewers from unconsciously favoring certain individuals (e.g., the first person to open the door), thereby preserving the randomness of the sample at the household level.

How to Use This Tool

  1. Count the eligible members in the household and enter that number in the "eligible household members" field.
  2. Before interviewing, note the fixed order you will list members in (by sex, then age) in the note field, so the number you get maps to a real person.
  3. Select the Kish table printed on or assigned to this questionnaire, or use "assign a table by questionnaire proportion" to draw one at the standard shares.
  4. Read the selected respondent number from the large result and interview the member holding that position in your list.
  5. Copy the record line ("Table D, n=4 → respondent 3") into the questionnaire log.

FAQ

Q: Why use a Kish grid instead of just asking for "the adult who had the most recent birthday"?

A: A Kish grid assigns each questionnaire a fixed selection table and picks a member by position within the household, which is the random within-household selection step that reduces sampling bias — the tool reproduces that grid.

Q: What happens when a household has more than 6 eligible members?

A: The grid covers up to 6 members; the selection is read at min(n,6), so a household of 7 or more is indexed at the sixth column of the assigned table.

Q: How are the Kish tables A, B1, B2, C, D, E1, E2 and F shared out across questionnaires?

A: In the standard assignment they appear in the proportions 1/6, 1/12, 1/12, 1/6, 1/6, 1/12, 1/12 and 1/6 respectively, which is the distribution the "assign a table by questionnaire proportion" control draws from.

Q: Does it matter in what order I list the household members before applying the grid?

A: Yes — the output is a position, not a name. List members in a fixed order (by sex, then age) and the number returned points to the member in that position; a different order gives a different person.