Would you please help? Assume you are to conduct a study in a population with different social statuses/ranks, where social hierarchy is the norm. The higher the rank, the more respect. Now that you are looking into sensitive health issues using PDAs (Personal Data Assistant – software for survey data collection, it look like a hand-sized computer) and you are worried about the biased responses from the high-ranking respondents. Will you give the PDAs to the latter to enter by themselves the sensitive data? Will you give PDAs to all respondents irrespective of the status to enter all their own sensitive data? Please advise.
This is, I think, more a methodological than an ethical question: the only ethical issue here, for me, relates to whether you will exclude certain ‘low-ranking’ and/or disabled groups through your chosen methodology of the PDA. In this case the ‘ethical’ response is to offer respondents a choice of method (e.g. paper quesitonnaires), even if this complicates your data analysis. In the UK, an ethics committee would typically make this sort of recommendation.
The question of bias is, as I said, one of method. The best way to address this is to pilot the study with a small representative sample and explore whether ‘bias’ exists in the ‘high-ranked’ individuals (or elsewhere). If your questionnaire had been validated with the general population as Gaussian (or normally distributed) but showed a skewed distribution in the pilot then this might be evidence of response bias (one example). Equally, the pilot might have high refusal rates in the ‘lower ranks’ which would also demonstrate a biased method.