“Troubled by Heterogeneity? Opportunities for Fresh Views on Long-standing and Recent Issues in Biology and Biomedicine,” was a talk I gave on 13 Oct. ’10 (abstract). I sketched a number of cases to get the audience thinking about my underlying contention that research and application of knowledge resulting from research are untroubled by heterogeneity to the extent that populations are well controlled. Such control can only be established and maintained with considerable effort or social infrastructure, which invites attention to possibilities for participation instead of control of human subjects.
The pdf of the slides and the audio recording are downloadable. (By noticing when my voice rises in volume, which is when I approach the laptop, you can guess when I am clicking from one slide to the next.) Some of the sketches of cases have been addressed in previous posts (see links on the abstract). This blog post consists of some afterthoughts, including questions needing more thought, in response to discussion after the talk.
1. What am I saying researchers should do?
The contention underlying the talk (above) is at first descriptive. But it does assume that heterogeneity (of various types) is ubiquitous and is often not paid attention to. The title suggests that researchers could be troubled by heterogeneity. But what should they do? Given that my framework incorporates a social level of explanation, what should they change first—their thinking and methods, or the social situations that enable the knowledge they arrive at to become significant (including, to be implemented in policy and associated practice)? What would be their motivation for changing?—To get a better view of the world (one that applies in a wider range of circumstances)? Or because there is a cost in controlling populations (maintaining infrastructure, etc.)? Or because that control breaks down, especially in crises?
2. Varieties of heterogeneity
My exploration of heterogeneity in biology and biomedicine took off after I saw the overlooked significance of underlying heterogeneity in heritability studies. Most of the cases in the talk, however, revolved around a simpler form of heterogeneity, namely, variation around a mean. Should I—for expository and/or conceptual reasons—focus separately on the different kinds of heterogeneity. I have an evolving taxonomy of heterogeneities, http://sicw.wikispaces.umb.edu/heterogeneities.
3. Personalized medicine
The figure I used to discuss the issue of misclassification lacked a crucial element—a cut off point between OK and not OK medical outcomes.
|1 (not treated because not sick)||OK||OK|
|2 treated (with say drug X)||OK result||Not OK|
4. Isn’t simplification of complexity sometimes/often/always necessary for scientific progress? After researchers get a handle on the simplified situation, they add back variables that they had previous controlled.
Sometimes researchers add back variables; sometimes they continue to engineer the world so the control over those variables is maintained. They may come to see the world the same way as they control it and need ways to be reminded early and often of what has been left out. This is especially so regarding ecological complexity, where variables left out have dynamics of their own that interact with the variables in focus. Chapter 1B of my book, Unruly Complexity (U. Chicago Press, 2005), illustrates the problem of “apparent interactions” that arise. Indeed, I have come to see the “simplification is necessary for science” line as a way to define out of science many situations that deserve systematic study. What do philosophers and theoreticians think about getting to know situations that, from the start, are not amenable to control or are not the same thing if they are carved out from the whole?
Fresh perspectives can be brought to modern understandings of heredity and life-course development by examining the relationship between control and variation, particularity, or, more generally, heterogeneity. Broadly speaking, my contention is that research and application of resulting knowledge are untroubled by heterogeneity to the extent that populations are well controlled. Such control can only be established and maintained with considerable effort or social infrastructure, which invites attention to possibilities for participation instead of control of human subjects. Building on several recent publications of mine on heterogeneity and heritability, I explain why underlying heterogeneity warrants the attention of quantitative geneticists and critical commentators on nature-nurture debates (see post). I elaborate on my contention through brief sketches of cases from biomedicine, involving: genetic testing; gene-environment interaction; personalized medicine; IQ scores; racial-group membership; and life events and difficulties research. My goal is to stimulate wider exploration of heterogeneity and control in relation to biological and social theories and practice.