Author’s Note: Any characters (including myself) referred to in this article are fictional, and scenarios have been slightly modified in order to distance this piece from reality.
This article is borne out of a simple conundrum (mind the oxymoron) – can information collected for one purpose be used, at a later stage, for another? What happens to the context?
And if information is used, what usage restrictions should apply?
Moreover, will these restrictions mean that conclusions drawn are after-all limited or inconclusive?
Let me explain…
Over the past few months, I have been interviewing various participants for my research. The interviews, usually longer than an hour, have served the purpose I set out to achieve. However, they have also highlighted other relevant strands of information that open up new ideas for research. Some of these ideas can be researched within the given time frame, and some are now on the “Someday-Maybe” list – where time and interest permitting – these ideas will be explored once again, perhaps in greater detail.
Then there are some ideas that were not realized till yesterday. These ideas do not result from the interviews I conducted. Instead, these ideas came up in a brainstorming session for another set of papers that I and a co-author plan to write in future.
Interestingly, the ideas can be somewhat answered from the interviews I have conducted in the past. The interviews that were conducted 6 months ago could provide a foundation for one of the joint papers.
But is it fair?
Is it ethical?
Is it ok to use the data from interviews conducted 6 months ago to answer research questions that were raised yesterday?
Personally, I feel ill-at-ease to justifiably conclude a new research question based on past data.
Simply because the new research question brings up more interesting possibilities to research, and without fully exploring (within an interview) this research question, we cannot sufficiently conclude the research question.
In psychology, the term Hindsight Bias refers to the inclination to see events that have already occurred as being more predictable than they were before they took place. Could a similar effect be applied to data analysis from past research? Could I be more prone to fit the empty spaces in my research with out-of-context interviews from the past?
Many researchers perhaps assume that going back to old data to answer new questions is enough. However, this feels like forcing an answer when a thorough understanding of the research question might be necessary. For a conclusive and unquestionable result set, fresh set of interviews would be necessary, so the research question can be answered sufficiently without any doubt or hindsight bias.
But let’s say one decides to use old data anyway.
Forgoing the issue of its completeness and hence validity, what happens in case of joint articles?
Consider this scenario…
6 months ago, when I conducted the interview, I explicitly clarified (to the participant) that the audio and the transcription will be only made available to me and my supervisor(s). But now, if I decide to pursue a joint, collaborative qualitative research, should I share the resource with the co-author?
Of course, many will say a coherent and resounding NO – it is ethically wrong to do that. But I have often heard rumours of shared data outside of the channels initially agreed upon. Of course, these are simply rumours.
Well, let’s assume that within this qualitative study, I decline to share data. Ethically correct, right?
Am I jeopardizing the quality of research and hence the results?
Have I just hidden – perhaps the most important piece of information from my colleague? By denying her access to the audio and the transcription, I have just denied us possibilities of covering all aspects of this research in detail. Now, my colleague cannot offer suggestions based on the data which we are using for our research, which might affect our conclusions – conclusions that could make or break the research.
So while our research is now ethically justified, its quality has been compromised. Most qualitative assessments value independent analysis by multiple co-authors, but in our case, the results might be biased to opinions and objectives of one author – me.
To this – there can only be one conclusion (do share if you have more ideas)…
If one has a new research question, one should past data to understand and extrapolate it. However, if one wishes to study it in detail, one should collect and use new data, with clear ethical implications for the participants involved.