Harnessing Bias and Ignorance: An Argument for a Diversity of Perspectives towards Better Collaboration
A Chinese philosopher once proposed that “the participant’s perspectives are clouded while the bystander’s views are clear.” (Unknown) This simple observation captures the core purpose of conducting interdisciplinary research; to leverage unique perspectives towards creating new knowledge at a necessary ever faster pace. However, perception can also be detrimental to research – our past experiences obscuring the implications of evidence perhaps clear to someone else (Chalmers, 1999, p. 7). By discussing the problems and benefits of perceptual bias, different types of expertise, and the notion of triangulation (INTEG220, 2011, p. Oct 27), I will show how an interdisciplinary approach to research can be used to overcome debilitating perceptual biases and how perceptual bias can be positively harnessed.
Priming is a phenomenon where previous knowledge skews our perception of reality. Bias is introduced in everything we observe because we are primed by previous experiences (Chalmers, 1999, p. 9), and often, the epistemological framework associated with a discipline. Nothing may be observed in perfect isolation and with complete objectivity. The philosophers Duhem and Quine contest it is impossible to test a single scientific hypothesis in isolation because any given hypothesis relies on background assumptions – other foundational hypothesis (INTEG220, 2011, p. Oct 13). Similarly, any observation also relies on background assumptions – other foundational observations which ultimately came to form an individual’s current state of knowledge. To consider or evaluate an observation in isolation without also considering the background assumptions that enabled or skewed the observation would unearth problems parallel to Popperian falsificationism of hypotheses. Popper’s falsificationism contests it is possible to refute a hypothesis in isolation provided the hypothesis is written in a testable manner, and has largely been superseded by theories that acknowledge the interconnected and multidimensional nature of scientific theories (INTEG220, 2011, p. Oct 13). It is this ‘skewing’ of an observation by background observations or knowledge that creates perceptual bias (Chalmers, 1999, p. 12); it is the sum of our perceptual biases which form our conceptual scheme – a subtly unique perspective formed by our environment, education and upbringing that permeates into our perception of the world (Brodie, 2011). Within the context of a discipline, the set of conceptual schemes carried by members of the community likely will form the discipline’s epistemological framework, or the methods by which the discipline acquires new knowledge.
A fundamental goal of science is to attain objectivity; to discover what actually is, what actually occurs. (INTEG220, 2011, p. Oct 27) If even observations – never mind analysis – are fundamentally prone to bias, how can we return to some semblance of objectivity? It seems considering an observation from multiple perspectives would be inherently valuable – perhaps to arrive at a mean in the set possibilities, or perhaps to simply have a path to a larger dataset on which to perform analysis. Someone who could temporarily suspend their own context and replace it with another would achieve some form of this – what we might refer to as ‘walking a mile in another’s shoes’ is a powerful tool. But to purposefully decide to switch contexts remains a cognitive decision, still limited by a fundamental bias-skewing observation. Will someone who is colour blind ever observe the world as someone who is not – or even as someone colour blind to a different degree or in a different way? Even our ability to consider alternate contexts is limited because of the very thing responsible for generating them: our past experiences and knowledge, which are not easily switched off. The first of several continuums I will introduce is concerned with the different degrees of interdisciplinary collaboration. Considering the difficulty in enabling, disabling or swapping our own contextual biases selectively and in an on-demand fashion, collaborating with others – fully acknowledging and welcoming biases based on their own unique set of experiences and knowledge – is perhaps the next best thing. Not all collaboration is created equal, however, and the degree of interdisciplinary collaboration will certainly define what will be possible to achieve. At one extreme is multidisciplinary – which perhaps should not even be considered a form of a collaboration. This is the case where many people from many backgrounds work on solving a single problem, but in complete insolation. By far the most common form bears the populist namesake interdisciplinary collaboration and is characterized by its ‘black-box nature’, where a project is subdivided into pieces and individuals take responsibility for the components – only concerned with the output required of themselves and the inputs required of other people. Transdisciplinary collaboration is the ideal form of collaborative teamwork. It transcends the core questions to involve contributors in defining the goals and outcomes of a project; specialists in other disciplines are not simply considered functional on-demand black-box resources to be engaged and disengaged as needed, but their expertise is incorporated as a primary concern at the earliest planning stages of research or a project. (INTEG220, 2011, p. Nov 3; Miller, 2008, p. Disciplinarities) It is within the space of transdisciplinary collaboration that the methods outlined in this paper will be most effective; it is within the context of transdiciplinary collaboration that perceptual bias may be best harnessed when conducting research in teams, especially teams involving people with limited domain expertise.
Perceptual bias is most effectively harnessed in a collaborative setting, with people from varying disciplines, and consequently depths of expertise in a particular matter. People at every depth of expertise can almost always contribute, though those contributing with little specialized knowledge in the field add value in significantly different ways than those with ten thousand hours of training. Specifically, those with contributory expertise add value stemming from their intimate knowledge of the domain at hand. This is different from how those with interactional expertise – knowing the language, jargon and terms – add value, which is likewise different from how those with primary source, popular understanding, or ‘beer-mat’ expertise may add value (Harry Collins, 2007, p. 14). Essentially a continuum – with contributory expertise mapping to our typical understanding of an expert, someone with doctoral credentials, to someone with ‘beer-mat’ expertise essentially having read a summary – each level of ‘expert’ has the potential to bring something valuable to an exchange, even if it is ignorance. Ignorance can be a powerful tool to force reanalysis of assumptions and justification of actions taken that otherwise would not likely have occurred. This reanalysis often results in alternative means to an end, and an awareness of alternatives is exceptionally important when making decisions – for the simple reason of otherwise being unable to choose the optimal path. An economics professor was once asked by a charitable foundation to help them decide, of all the possible requests for funding before them, which projects would best allocate their resources to benefit society. In touring the lab of a preeminent cancer researcher, it came upon the professor to ask if the researcher had considered any alternatives to his current approach, considering the significant amount of money requested. Economics being primarily concerned with the best possible allocation of resources to benefit society, it was natural for him to ask “Was there a better way to do this that would not cost as much?” The researcher responded that there was only one possible method. Again, the professor asked he had considered any alternatives – to which the researcher became quite agitated and the answer clear (Smith, 2011). Cancer research was a field the economics professor had a popular understanding of – knowledge from the media and consumer magazines, perhaps. Nonetheless, expertise from an entirely different domain enabled the economics professor to unearth a potentially fatal flaw in the cancer researcher’s approach to solving a very important problem – he had not considered alternative means to accomplishing his research objective. Alternatives posed by less than contributory experts are, in decreasing succession, less susceptible to thinking in ways confined by the epistemological framework – or conceptual scheme – attached to a discipline. Thus, the benefits of epistemological pluralism – where the epistemological approaches of disparate disciplines are recognized as being valuable – result from welcoming contribution from all disciplines and heights of expertise (INTEG220, 2011, p. Nov 3; Miller, 2008, p. Conclusion) staged within the context of transdisciplinary collaboration. While this alone will increase the effectiveness of research, we can also vary the design of the study itself to attain new perspectives.
Triangulation extends the concept of epistemological pluralism and builds on its benefits by intentionally varying the path to discoveries: the design of the study itself. Comparable to the divergent brainstorming process, triangulation acknowledges the existence of a continuum of study designs and advocates conducting multiple studies on a single research topic in anticipation of converging on the same conclusions, greatly strengthening the research position. This continuum contains manipulative studies at one extreme, through partial manipulation and natural experiments to observational surveys at the other end (Scheiner, 2004, p. 55). Manipulative studies allow the researcher control over all variables, granting the freedom to choose values to modify while maintaining controls. Partial manipulation and natural experiments simply acknowledge that some variables simply may not be contained and will likely vary naturally over the course of experimentation. While the defined goal of triangulation largely rests on corroboration of results, I contest there exists additional benefits to varying the study design in the path to arriving at conclusions. Just as the process of involving individuals from distinct backgrounds and significantly varying expertise results in novel approaches to problem solving, the subtle differences of study designs each will prompt different questions based on the nature of their processes. The activities involved with an observational study are much different from the activities required by manipulative or even natural designs, and it is these dissimilar activities which will prompt dissimilar questions about the same topic. A likeness to the benefits of epistemological pluralism becomes apparent, though arrived to by different means. If the method of thinking – or epistemological framework – normally affects which study design to employ, there could be significant value in decoupling the method of thinking from the study implementation, an almost certain inevitability when engaging triangulation. Conducting complimentary studies of multiple designs not only creates more rigorous conclusions, but encourages variance in the path to arriving at conclusions; it is this variance – the small details, the pieces that don’t quite mesh – which lead to larger questions and even more profound discoveries.
Perceptual bias and susceptibility to priming are unilaterally considered negative attributes within the objective world of science. I speculate that the notion of non-contributory individuals making meaningful contributions to a field is considered similarly absurd. However, ridding ourselves completely of perceptual bias is near, if not, impossible and expertise is a continuum that includes those without PhDs. I have argued that a biased perception can become an incredible asset if harnessed, and consequently demonstrated how even someone of limited knowledge can make valuable contributions to a specialist field. Further, by suggesting choice of study design itself lends to a particular perspective, I argued conducting multiple studies of varying design surrounding a single research topic delivers similar benefits. Granted, to implement a multi-study design approach to research – or even to intentionally involve non-contributory experts would imply significant change to the status quo. The financial burden of triangulation would be great; many would see it as simply replicating results and decry it as a waste of scarce resources. The time commitment required interacting and sharing with non-contributory experts is not insignificant, and requires noteworthy humility to acknowledge that someone of perhaps beer-mat knowledge could offer anything of value to highly specialized research. However, as the pace of research and innovation ever accelerates due to increasing accessibility to education and worldwide competitive pressures rise, we cannot simply maintain our current methods if we wish globally competitive. We must innovate how we come about innovations.
Much formatting was lost in translation. The full paper with references may be found here.