Skip to interview prep
Apexed

53. Only respondents are included in a randomized training-program analysis, and response rates differ by group. What is the concern?

MediumStatistics & experiments

Technical
Show model answer

Random assignment does not guarantee a fair comparison after treatment-related dropout or nonresponse. Respondents in the two groups may represent different subsets, so their outcome difference can combine treatment effects with selection.

Compare response patterns, investigate reasons and seek permitted follow-up data. Use transparent sensitivity analyses or justified missing-data methods, and report the limitations. Do not claim randomization alone removes bias from analyzing only the observed respondents.

What interviewers listen for

  • Check post-assignment selection
  • Investigate response mechanisms
  • Report sensitivity to missing outcomes
Only respondents are included in a randomized training-program analysis, and response rates differ by group. What is the concern — Data & analytics interview qu