Hi Maina,

Conducting research solo is a practice that may not be avoided. However solo researcher may be accused of introducing personal bias in their work. I think the first step to avoid bias in solo research is acknowledge the value of observing research ethics. This safeguards solo research from falling to the temptation of bias. When the research remembers that credibility is of essence in research every caution shall  be undertaken to ensure adequate checks are in place to safeguard against bias. Probably piloting the tools and discussing the pilot finding with other researchers shall help in highlighting possibilities of bias and dealing with them before the data is gathered. The issue of solo researchers is likely to generate interesting debate.

Thanks

Solomon

In any qualitative research, the researcher is the primary research instrument hence her views, ontological beliefs and epistemological underpinnings play a crucial role in the final data analysis and findings. The researcher 'constructs' meaning through her own experience thus the findings will heavily be influenced by such construction. As a result, while reporting such a study, a detailed description of these need to be added right at the beginning of the report so that the readers understand the philosophy behind the interpretations and claims.

Thank you for your feedback.

A limitation of conducting in-depth interviews are that the researcher’s presence may bias the participants’ responses, and that not all interviewees may be equally articulate and perceptive, which could impact on the responses given to questions asked. One means of addressing these limitations is to establish trustworthiness. One aspect of this relates to Confirmability. This ensures findings are a result of the enquiry not the biases of the researcher. Confirmability recognises that complete objectivity is not possible in social research but that social researchers need to demonstrate that they have acted in good faith, for example, not to have been influenced by personal values (see Bryman, 2012 social research methods). In naturalistic social research confimability should form part of the audit process (see Lincoln and Guba, 1985, Naturalistic Enquiry). See section on establishing trustworthiness in research.

See also Sparkes and Smith (2014) Qualitative research in sports exercise and health, for a more up to date perspective.

Bias is unavoidable, so you need to be transparent. Make explict the assumptions and theoretical framework guiding your analysis, and (more importantly) demonstrate what kind of conclusions you draw from your interviews by providing quotes and stating the conclusions you draw from them.

I agree with others above that in qualitative research the researcher bias is usually discussed overtly, and you state this at the beginning of your research. What is your position on the question you are asking? This transparency is often termed 'reflexivity' that is, always being aware of your position and questioning your own interpretation. Member checks with interviewees are sometimes (but not always) used. Other things you can do are: ask colleagues or mentors to independently check your analysis - this is what a supervisor would usually do; keep a reflexive journal in which you document your thinking (this is separate from any analysis that you do); and stay 'close' to your data - any findings that you have should be directly relatable (via your analytic framework) to the data - this is sometimes referred to as leaving an 'audit trail'. That is, a reader could easily track your analysis back from your findings to your data.

I think the question of how much influence the researcher has on the research process is different for data collection versus data analysis, especially for qualitative research. With regard to data collection, if just one person collects all the data, whether via surveys or qualitative interviews, that is certain to be some influence. The main thing to do to minimize this influence is to maintain a neutral approach to each interview.

When it comes to qualitative analysis, that is the main point where it is unavoidable that the researcher is the research instrument. Several other people have summarized techniques for ensuring confirmability and credibility in this regard. 

Dear Maina, 

Good topic. I did already several qualitative research projects and had similar dilemmas about bias. My answer would be you should take care of recurrence procedures. Get back into the field while you experience such collapse of not knowing if you are biased or this is the emic meaning and receive broader explanations directly from informants. This will influence credibility as well. Good luck ! 

Thank you all for your responses and the recommended sources. 

Hello Maina,

I agree that the ontological and epistemological underpinnings of the research should be made clear at the beginning of your report.  This will aid your reader in comprehending the logic of your data gathering and interpretation of meanings.  

To come up with a trustworthy and credible data and analysis, there is need for a good sampling method so that your research is inclusive of the multiple voices and experiences of the researched, Prior to the conduct of open-ended interview, you should have already established connection with the respondents to gain trust and rapport. From my experience, this is important in eliciting responses from respondents with different demographic characteristics.  If you get few responses, there is danger of getting data and analysis that is biased towards certain individuals/groups. 

In qualitative research, you bring bias in your assumptions, conceptual framework and research methods. It is necessary to make transparent in your report your subjective positioning that contributed to research findings.

In qualitative research the analysis would be generated through your personal interpretations as the researcher .. as you said it will be subjective and this is accepted in this kind of research. You, as the researcher will be the analysis tool here (as opposed to computerized programs in positivistic research) and your personal interpretations will shape the findings of the study. 

However there are ways that you can follow to ensure the credibility of the data captured, for e.g by member reflections (respondent feedback after analysis) and thick description (in-depth illustrations). 

I would agree with what others are saying, that what is called 'bias' in quantitative research is recognised in qualitative research as inevitable differences in the perspective that different researchers bring to the data - there IS no 'God's eye' unbiased view on the data.  Instead it's useful to think about the validity and trustworthiness of interpretations - are they supported by the data & has this been made transparent?  Different researchers may make different interpretations that all have some validity & draw out different aspects of the data.  There's no way you can capture all these possibilities however many other researchers are involved, though obviously the fewer the researchers the more limited the range of perspectives.  However, I would think carefully about the extent to which member checks or respondent validation addresses the issue.  Participants may also bring vested interests to their interpretation of the data. If you are approaching your data with a hermeneutics of empathy and aiming to understand the participant's point of view as fully as possible (e.g. descriptive phenomenology) then respondent validation probably makes more sense.  However, if you are using a hermeneutics of suspicion & bringing a more critical perspective to the data, then your interpretations may differ from those of at least some of your participants.

Hello Maina,

I concur with the knowledge offered by m colleagues and suggest disclosing your biases within the study. There may be ethical bias, biases to the assumption, etc. Allowing the reader to know you may have biases that impede the study ensures trustworthiness as you are the research tool for interpreting the data.   

It's all about being transparent. Qualitative researchers often embrace their "bias" as sources of valuable expertise available to inform or direct the analysis. Some try to alleviate some bias by making the reader aware of their own sources of bias to allow the reader to be aware of the lens through which the data is analysed and presented. Scientific rigour through replicable methods and maximising attention to each unit of meaning in the interview transcripts are other ways some researchers achieve this. 

What you choose to do depends on the approach you take, and the theoretical and ontological background you are coming from so I would recommend reading up on this sort of thing and exploring your opinion on bias and which methodologies you are most aligned with.

Developing your research instrument, questionnaires for example, from multiple as well as contrasting theoretical positions helps to mitigate interpretative biases. 

Ideally, the data analysis should be conducted by another party who is not privy to the data collection.

I guess it depends on the type of research you are conducting. your philosophical positioning in the research will determine how you will deal with researcher's bias. Qualitative research is subjective in nature as human agency to get rich thick information from your data, however, to avoid this , one needs to continually re-evaluate the impressions and responses, and ensure that pre-existing assumptions are kept at bay. Keep the questions simple and be careful to avoid words that could introduce bias.Do not use leading questions that can prompt the participant to respond in favor of a particular assumption(Shah, 2019).

This is a really good question. It makes me wonder whether we should distinguish between a researcher's individual perspective and bias.

1) Applying a personal perspective in the collection and analysis of data is inevitable because in qualitative research, we reconstruct meaning. I don't see how this could be possible without applying a personal perspective. This personal perspective and its impact on interpretation should be made as explicit as possible. Three actions that support the explication of one's personal perspective to oneself (we are not necessarily aware of this perspective) and others are:

* Write down what you think aobut your object of study at the beginning of your investigation. Write down your assumptions about what is going on in the field and what you expect to find. Although it looks similar to formulating hypotheses, this is something else entirely. You just document your expectations in order to give yourself the opportunity to become aware of how these expectations affect your research.

* Communicate with other researchers about the data analysis. In many sociology research groups, "data sessions" have been established as a format in which a research can discuss data from an ongoing investigation and their interpretation. Discussing one's interpretations of data durring the research process is very important even though the 'discussants' are often outsiders, i.e. don't work in the research project.

* Explicate your perspective in your publications by providing interview quotes and your interpretations of them. As I wrote elsewhere on RG, interview quotes in publications of qualitative research are not just illustrations. They demonstrate how you interpreted your data, hence: your perspective.

2) Bias is trickier because it is about preconceptions rather than expectations, and preconceptions should not enter teh research process. They do nevertheless, but while our personal perspective is a research instrument, our preconceptions may lead to bias.

In principle, the steps described above should also help to discover bias. In addition, I suggest to (frequently) ask yourself about your normative position vis-a-vis the processes and people that constitute your research object. Are there people I (don't) like? Are there processes, rules, interactions I think are 'wrong' or 'bad'. These normative judgements are unavoidable too. The point, again, is to become aware of them and of their possible impact on your interpretations.

To avoid bias in survey research, a combination of data should be used, field data combined with interviews and other supporting data. Especially secondary data whose validity can be justified.

Working on a phone surveys for a research one institution we were allowed to end the questioning if the informant was rude, racist or Xenophobic. A woman begin describing to me the name of her grandparents favorite slaves and how "being in bondage was the best thing that happened to them."

My experience as a descendant of slaves who picked tobacco in Virginia and South Carolina as well as cutting sugar cane in Brasil and Louisiana was naturally triggering my ability to continue without being emotional.

This survey was marked incomplete. Data on the incomplete interviews was kept, but labelled with an explanation for the incompletion. This culling of the data allowed for the removal of surveys that were possible harbingers of research bias.