The Subjectivity of Scientists and the Bayesian Approach

by Judith M. Tanur, S. James Press

Comparing and contrasting the reality of subjectivity in the work of history's great scientists and the modern Bayesian approach to statistical analysis
Scientists and researchers are taught to analyze their data from an objective point of view, allowing the data to speak for themselves rather than assigning them meaning based on expectations or opinions. But scientists have never behaved fully objectively. Throughout history, some of our greatest scientific minds have relied on intuition, hunches, and personal beliefs to make sense of empirical data-and these subjective influences have often aided in humanity's greatest scientific achievements. The authors argue that subjectivity has not only played a significant role in the advancement of science, but that science will advance more rapidly if the modern methods of Bayesian statistical analysis replace some of the classical twentieth-century methods that have traditionally been taught.
To accomplish this goal, the authors examine the lives and work of history's great scientists and show that even the most successful have sometimes misrepresented findings or been influenced by their own preconceived notions of religion, metaphysics, and the occult, or the personal beliefs of their mentors. Contrary to popular belief, our greatest scientific thinkers approached their data with a combination of subjectivity and empiricism, and thus informally achieved what is more formally accomplished by the modern Bayesian approach to data analysis.
Yet we are still taught that science is purely objective. This innovative book dispels that myth using historical accounts and biographical sketches of more than a dozen great scientists, including Aristotle, Galileo Galilei, Johannes Kepler, William Harvey, Sir Isaac Newton, Antoine Levoisier, Alexander von Humboldt, Michael Faraday, Charles Darwin, Louis Pasteur, Gregor Mendel, Sigmund Freud, Marie Curie, Robert Millikan, Albert Einstein, Sir Cyril Burt, and Margaret Mead. Also included is a detailed treatment of the modern Bayesian approach to data analysis. Up-to-date references to the Bayesian theoretical and applied literature, as well as reference lists of the primary sources of the principal works of all the scientists discussed, round out this comprehensive treatment of the subject.
Readers will benefit from this cogent and enlightening view of the history of subjectivity in science and the authors' alternative vision of how the Bayesian approach should be used to further the cause of science and learning well into the twenty-first century.

304 pages, Hardcover

First published April 16, 2001

Reviews

[Tom Schulte · Rating 3 out of 5]

In his book Clinical vs. Statistical Prediction: A Theoretical Analysis and a Review of the Evidence (University of Minnesota, 1954), psychoanalyst Paul Meehl gave evidence that statistical models almost always yield better predictions and diagnoses than trained professionals. The authors here examine the cases of selected pioneers in science and how bias-driven subjectivity played a significant role in directing their advancements. That role of subjectivity is the main thrust of this book. The formal application of Bayesian inference in deriving the posterior probability based on prior probability and a likelihood function derived from a statistical model for the observed data is not examined as deeply as the title may suggest. This is really a history of science via chapter-length biographies for the layman with a Bayesian introduction for the nonmathematician as almost an appendix.

...

[Look for my entire review at MAA Reviews]

[Ari · Rating 1 out of 5]

This might be the worst book I have ever read about the history of philosophy of science. The authors argue that science intrinsically relies on subjective judgments and that Bayesian reasoning lets us do this in a principled and rigorous way. They back this with biographical sketches of 12 great scientists (selected by a previous author) and 5 other more-controversial scientists.

The first problem with this book is that the authors have an idiosyncratic and unhelpful definition of "subjective"; they define it as "preexisting views or beliefs about entities that influence the gathering of data or its interpretation." That is, all theories used to design an experiment are subjective; the authors expressly treat all theoretical physics, even mathematical deduction, as "subjective." This is at best unusual and potentially wildly misleading.

The biographical sketches felt mostly redundant and in many places carelessly done. The authors refer to the Einstein letter about the atomic bomb, without mentioning that Einstein didn't write it. (Teller and Szilard did.) They refer to Pasteur injecting people with his rabies vaccine. (He didn't do the injections; he worked with a physician, who did the actual injection.)

The authors' references to past philosophers of science are painfully confused. They refer to "the Popperian approach (of falsification, or setting up a straw man hypothesis believed to be false and then showing it is indeed false.)" (p 214.). This is a wildly distorted portrayal of Popper. One wonders if they have in fact read him; if not, they have no business writing about the scientific method, since his is basically the consensus view of most working scientists.

I cannot understand what view the authors are rebutting. The straw-vulcan view, where a scientist should mechanically deduce a hypothesis from data, test it in a straightforward way, and repeat, is one that I have never seen anybody advocate for. Everybody has always asserted that science is a creative, imaginative field, with a lot of guessing and exploring of hunches. Objectivity matters because science aims to convince skeptical experts -- who might have different guesses. And those people have stories and guesses of their own and will only be swayed by math or empirical evidence. The point of objectivity is to convince skeptics, not to form hypotheses. To the extent that we exhort scientists to remain dispassionate and objective, it is precisely because everybody understands that the universal human tendency is otherwise.

But my biggest problem with the book was in the final chapter, on Bayesianism. The authors have a plodding description of high-school level bayesian inference (e.g., interpreting the posterior probability of a rare disease given a positive test result.) They give a few more examples along these lines, and then jump almost immediately to "and we could use this to weigh evidence for theories."

This is laughably shallow. The problems with this view have been known since the 19th century and are discussed in every serious book on the philosophy of science. To digress slightly -- (1) doing a Bayesian calculation requires computing the probability of an observation independent of the theory in question, or given that the theory is false. Nobody has any idea how to do this, or even what it would mean. We can't even imagine a practical process that enumerates over all theories. (2) Any real experiment has a vast number of auxiliary hypotheses, all of which are uncertain. So we need to have a whole bunch of priors and conditional probabilities, none of which can be measured or estimated directly. (3) All our best theories are literally false. Often we talk about models, not theories -- a model might be accurate or inaccurate, useful or useless, but it's never literally true. So a bayesian calculation of "how true is the theory given this evidence" is trivially useless, since there is likely to be a term pushing the probability of the theory being true all the way to zero.

All these points are old and familiar; I believe Poincaré and Duhem already made them in the 19th
century and most of 20th century philosophy of science tries to rebut or contain them. The authors here just ignore them.

I cannot imagine a reader for whom this book would be a good investment of time. The authors are fundamentally untrustworthy, and their view is not a new one -- it was abandoned and buried for good cause long ago and the authors give no hint of a reason to disinter it.