Quantitative Surveys in India

Commission this when one of the questions below is holding up a decision. If your question is not on the list, say so in the brief — we would rather recommend a different method than sell you this one.

• What share of the market holds this behaviour or opinion, and within what margin?

• Is the difference between these two segments real, or inside the noise?

• How has the measure moved since the last wave?

• Which subgroups are large enough to build a plan around?

• What does the population look like once the sample is weighted properly?

Fieldwork in India fails quietly when nobody checks it. These controls are run on every quantitative surveys study, by a quality team that reports separately from the field team.

• GPS and duration stamps on every interview

  Each tablet record carries a location and a length. Interviews that are too short, or clustered at one convenient address, are pulled before they reach the dataset.

• Twenty per cent audio back-checks

  A fifth of all interviews are re-contacted or audio-reviewed by a quality team that does not report to the field supervisor who ran them.

• One supervisor to six interviewers

  Supervisors accompany each interviewer at least twice a week and sit through the full interview, not the doorstep introduction.

• Translation and independent back-translation

  Every instrument is translated into the field language, back-translated into English by a second linguist, then piloted before launch.

• Random starting points, not convenience routes

  Household selection follows a prescribed random walk from an assigned starting point, which is what keeps an urban sample from becoming a main-road sample.