Research bleg from Daniel Klein: Seeking survey/dataset, with repeated question

Dear Readers,

Collaborators and I have administered a survey that posits a reform to immigration policy. One question asks:

  • “Implementing such a reform would, on the whole, yield positive net economic consequences for humankind (worldwide, present and future):”

[5 response options from “agree strongly” to “disagree strongly”]

Another question asks:

  • “The reform ought to be favored over the status-quo immigration policy:”

[again, 5 response options from “agree strongly” to “disagree strongly”]

The aim of our project is to explore the extent to which these two questions (and others in similar pairs) are practically the same. We address the matter by gauging the extent to which the individual person’s responses to the two questions are the same.

One way to assess the extent to which the two questions are practically the same is to see what happens in surveys where two questions are exactly the same.

If there are other surveys that asks the same question twice, ideally with other questions in between, then the intra-individual variability found in those surveys might provide a benchmark for assessing the intra-individual variability in our data for the two questions stated above.

So, we are seeking datasets for when the same question is put twice in a survey. The repeated question does not have to be about policy issues; it can be about personal well-being or personal opinion, for example.

We understand that marketing research sometimes administers a questionnaire that repeats a question, but with the aim of seeing whether the intervening questions move the respondent; in those questionnaires the intervening questions pertain to the question that is repeated. We, in contrast, do not want the intervening questions to bear on the repeated question.

We would be very grateful for leads on where to find such surveys – that is, a survey that repeats a question and corresponding datasets. Please email suggestions to

2 thoughts on “Research bleg from Daniel Klein: Seeking survey/dataset, with repeated question”

  1. The Health and Retirement Study survey administered by the University of Michigan Institute of Social Research had repeated questions–at least in earlier waves. I recall that there was a question that said something like “What is the chance that the stockmarket will be higher in one year?” with answers on a scale from 1-100. This question was generally administered to all takers. Then there were experimental modules at the end that only a subset of takers would get–a random subset (I think). In at least one of the experimental modules that question was repeated. IIRC the own-answer correlation between the two responses was something like 50% within a single survey wave, but you’d have to check me on that, since it was a long time ago. The data should be available publicly.

Comments are closed.