This is a huge question, and my answer in a nutshell, is that you should not attempt to remain ‘unbiased’.
I’ll try to approach your question via the work of Alfred Schutz, a sociologist from the mid-20th century. He argued that, unlike the objects of study in the natural sciences, those studied in social research are active, sense-making human beings, who are engaged in interpreting and ascribing meaning to their world in interaction with each other. Yet this also applies to the social scientist herself, who is a further active interpreter of the same social world inhabited by those she would observe and understand.
A consequence for social science is that researchers need to acknowledge their own interpretative work as they analyse the social worlds they are researching, and recognise that in making sense of an actor’s sense-making, they themselves impose a second level of interpretation, that is itself subject to wht Weber called verstehen (understanding). This is a critical issue for social researchers, especially those using qualitative interpretivist approaches, as they must recognise that their human, rationalising, constructive activity is behind their analyses of actors’ life-worlds.
This analysis has a number of consequences. First, it acknowledges that people are engaged in an on-going project of producing the social world, and therefore that their sense-making must itself become part of the subject-matter of a social science, ruling out a simplistic limitation of study to ‘social facts’, and accepting the context-specificity of knowledge.
Second, it recognises that the tools of study in social science are human beings’ own capacities as interpreters of the world. As such, these instruments work by means of exactly the same processes of intersubjective meaning-attribution that the social scientist seeks to study. While there may be an aspiration to objectivity by the social scientist, this inheres only in her detachment from the practical commitments and interests of her subjects, not from some essential difference in her ability to interpret free from values, norms and so forth.
This leads to the third feature, the need for social scientists to be reflexive about their interpretative work, both to aspire to detachment but at the same time to accept its ultimate impossibility.
This ‘post-positivist’ analysis has led to two contrary perspectives on doing qualitative research, which can be broadly described as realist and constructivist. The former adheres to the notion that there is some objective reality to the social world, while acknowledging that the Schutzian analysis of social science as interpretative and therefore ultimately subjective sense-making, precludes the discovery of that reality once-and-for-all. All that can be achieved is the aspiration to knowledge through rigour, multiple data analysis and theory-building and testing. Constructivists, by contrast, consider not only that objective knowledge is impossible because of these problems of interpretation, but also that – given that the world is variously constructed by human beings with their context- and interest-specific views of the world anyway -- that reality is itself multiple, contingent and value-laden.
What follows is that, rather than trying to avoid bias, you need to acknowledge it, and if anything, use your reflexivity as a human being to add to the richness of the account you offer when you report the qualitative data you have collected and analysed.
Very good response Alexandru and Nick. You have both captured the essence that 'objectivity' in qualitative research is a 'double-edged sword'. It is indeed a complex situation - and logical positivists cannot often get their head around that notion. They often dismiss it as being subjectively bias as a whole.
On the one hand, the qualitative researcher wants to remain close and intimate to their collected narrative. On the other hand, hierarchies of evidence and trustworthiness criteria etc demand that the researcher remains distant and balanced on the collected data - especially with emic/etic approaches - such as grounded theory. Personally, I think that both count. To me, it's obvious if a qualitative study is 'made up' and subjectively bias - but implementing measures such as cross-checking, counter-checking, using qualitative software (where it is appropriate) etc - help to deter that illusion.
I've conducted phenomenological research previously where I have reported the 'phenomenological nod'. That is, the issue I am investigating is one that I have experienced previously myself. I distance myself from the collected narrative - but look at anything that emerges that resonates with my own experiences. It's not subjective - i am merely reporting when 'yes- this reflects m own personal experiences' - and 'no - this doesn't'.
I guess, at heart each researcher is essentially a writer trying to make sense of the world around him. Qualitative research allows for this opportunity in the true sense of the word. We are trying to give messages through this material that we have, at times yes we may be biased but by and large we mange to do some good.
A good response - but qualitative researchers should never be bias. There are so many methods and techniques to avoid and rule bias out. Yes - you can be 'attached' to your data and findings - but that is different to bias.
I guess "bias" deserves definition in an effort to ensure that we are on the same page, so to speak.
Sometimes "bias" is used to mean "a particular point of view." In this sense, bias can never be eliminated from research. I think this is where Qurratulann is coming from.
Alternately, bias can be seen as a predisposition or inclination that distorts processing or understanding. Of course, from this perspective, every scholar (I hope) would want to eliminate bias. Dean is using this definition, I believe.
I would add that transparency of method and interpretation is vitally important when reporting research. Readers should be able to evaluate the validity and reliability of your approach, and ultimately make a call about potential biases, no matter the source.
Correct Steven - I personally wouldn't recommend using the term bias to represent a 'point of view'. Probably, terms such as subjectivity, interpretivism, insider perspectives, interpretation etc are more acceptable. it might be semantics and they may amount to the same thing - but the term bias tends to conjure up the notion of 'purposeful manipulation' i.e. to manipulate the data to suit the researchers perspective and perhaps not what the data is really suggesting. In quantitative research, in particular, it amounts to 'error'.
Wow, I thought this would be a good question and now I see that many share the same thoughts. Everyone here provided great comments and I really appreciate the vast views. I realize that research must be transparent, valid, and replicable, which is why I wanted to get other thoughts on the depth of the issue. Knowing that there are ways to reduce bias in qual research, there still seems to be that small amount of interpretation that may get tainted based on one's personal experiences depending on the subject. To me; this is not simply a matter of bias, but one of ethics as well if the research ventures beyond that fine line in order to sway it towards the goals of the investigator. We must understand how to remove our own thoughts during investigation and reserve them for the very end. Only then shall we include the personal interpretations in the most objective way while also being ethically subjective on matters that do not invalidate the results.
In the UK, we have a game called ‘bowls’ in which the aim is to roll a black ball about the size of grapefruit over grass towards a small white ball called a jack: the person who is the closest wins. The problem is that the black balls have a weight in them that makes them roll in a curve. The good thing is that this bias is consistent, so it can be compensated for by an experienced player.
In qualitative research, a consistent bias like that can also be accounted for, and in the various answers to this question it’s been suggested that we recognise and account for our subjective biases, and even value them as adding something to the findings (given that a researcher may be part of the culture they are investigating).
There is a risk however, I feel, in treating this process as an entirely rational one, as if human researchers are able to step back from their own beliefs, values and emotions. First, we are not always aware of our biases (for instance, a deep-seated political or religious value that is so rooted in us that it makes us what we are). Second, we may be aware of them, but not be willing to admit to them because to do so would be emotionally painful (for instance, an aversion to people who are sexist or racist, because of something that happened in our earlier life that we prefer not to think about). These ‘non-rational’ responses are going to be hard to entirely eradicate, though part of learning the business of social research is becoming more reflective and insightful about one’s own motivations and desires.
As researchers, we have to accept that if we use human beings as measuring instruments (which is what we do when we gather, analyse and interpret qualitative data), then we introduce a load of both rational and affective baggage into the research process. Social research is messy and is never going to be as rational and controlled as natural science experiments. But let’s celebrate that, rather than trying to turn social inquiry into a sterile activity conducted by automata!
Rick, if you follow Steven Venette's comments with regards "transparency of method", and being "able to evaluate the validity and reliability of your approach" in non-postivistic terms, there is no issue. However your comment that you "realize that research must be transparent, valid, and replicable" - especially with the addition of "replicable" - makes use of a positivistic language and thus frames things differently to what I believe Steven was pointing out.
In my area of management research, specifically human resource management, it would be hard pressed to argue that a study be "replicable" in the same sense of natural sciences. People change, circumstances change, contexts differ.
Instead of a qualitative study being "replicable", I find the notion of "recoverability" in such research methodology as "Soft Systems Methodology" a much better notion to look into. Recoverability is similar to an "audit trail" where the reader of your research is able to retrace the steps and actions you took in your research to gain, and thus, "recover" the experience that unfolded throughout your study. Peter Checkland, the founder of Soft Systems Methodology, states this much better than I can and provides a strong philosophical base.
I personally use constructive grounded theory. To add to the discussion on "bias", an issue that I found in doing field research is that as I was so passionate about the topic (diversity and inclusion in companies in Japan) that I was conducting interviews with people in companies, I noticed that my analysis of the interview was too sympathetic with the interviewee's plight. To overcome this I created what I have tentatively termed "analytic distance", where I followed the rationale of the what the interviewee was saying and started to notice contradictions and potential issues. Maybe I don't need to use my made-up term and "reflexivity" is the proper term. But what I want to say is that in follow-up interviews, I used the findings of this analysis as a basis for some of my questions with the interviewee . As a result, the interviewee opened up and frankly started so started to reflected on what they were saying during the actual interview! However, this is still a work in progress.
I would be interested in what others in this thread would like to point out to me- especially my last comment. Everyone in this thread offers such wonderful insight! :) And Rick, I hope you could understand what I was trying to point out to you. :)
Darren, thank you for that great rendition of the variations of "replicable" and "recoverability". I actually really like that description. Replicable research to me does seem like it would lean more towards quantitative methods than qualitative. I understand exactly why you described it in the way that you did and I will definitely keep that thought with me from now on. In the back of my mind I had felt that there was something off-putting about the word "replicable" in qual research (at least in most respects).
Nick - you must be a SCIENTIST in the real social sense. Everything you said is totally understood and I appreciate your strong rational. Fantastic sir!!!!
Thank you all for the great feedback. This has been a wonderful discussion. I have been away for a while and want to thank all of you for the responses. Everyone has fantastic rationale and opinions that will make my research more clear when including my interpretation.
The Functionalist paradigm holds an objective view of reality, concerned with explaining how organizations and society maintain order, Burrell and Morgan (1979).
Positivism, which is one view as to how knowledge is created and which is highly favoured by researchers, theorizes that behaviour can be predicted and that cause-and-effect relationships are clear and pervasive Burrell and Morgan (1979). Functionalism strongly emphasises the pre-eminence of the social whole over the individual parts (human actors).
This contrasts sharply with the Interpretivist paradigm which holds a subjective view of reality and its anti-positivist epistemology which is concerned with explaining individuals’ perception of organizations and society and which seeks to understand the subjective individual and organizational processes that shape and control behaviour. In the interpretivist paradigm, explanations of human behaviour and all aspects of organizational life, are socially constructed, also known as the perspective of constructivism.
Consequently, to understand the processes that constructed an individual’s behaviour, researchers must gather data that reveal an individual’s subjective experiences, the interpretive lenses that give meaning to that experience, as well as the organizational factors and contexts that create these expressions.
Knowledge therefore flows from the objectivist or subjectivist metal frames of individuals. Individuals construct "small-scale models", in their minds, of the entities and relationships which they perceive to form the external reality in order to anticipate and explain events, Craik, (1943). As illustrated paper attached, functionalists take an objective perspective and tend to have a realist ontology, a positivist epistemology, a deterministic vew of individuals, and a nomothetic methodology. Interpretivists on the other hand, take a subjective perspective and tend to have a nominalist ontology, an anti-positivist epistemology, a voluntarist view of individuals, and a ideograpic methodology, (Grant and Perrin, 2002).