falsifying scientific data -- 3/30/22
Today's selection -- from The Knowledge Machine by Michael Strevens. Arthur Eddington of Cambridge was an early defender of Einstein’s new theory of relativity. He used photographs from a 1919 eclipse to examine a light shift which promised to provide evidence in support of that theory:
“How did Eddington and his collaborators reach the conclusion … that it was Einstein's predictions that came true?
"They had three sets of data [from photographs of a 1919 eclipse] at their fingertips. First, there were 2 photographs from Principe that dimly depicted stars through the clouds, and which according to some rather complex calculations by Eddington showed a shift of Einsteinian magnitude. Second, there were 7 photographs from the Brazilian 4-inch telescope that also showed an Einsteinian shift. Third, there were 18 photographs from the Brazilian astrographic telescope that showed the shift predicted by Newton's theory. Eddington's strategy was to argue that something had gone systematically wrong with this last set of photographs. They were, in fact, considerably blurrier than those produced by the 4-inch telescope, possibly (so he and his collaborators conjectured) because of distortions caused by the sun's uneven heating of the mirror that reflected the light from the eclipse into the telescope.
|One of Eddington's photographs of the total solar eclipse of 29 May 1919, presented in his 1920 paper announcing its success, confirming Einstein's theory that light "bends"|
"Certain of Eddington's contemporaries, however, found Eddington's argument to be rather fishy, as have many later historians of science. Eddington could explain the blurriness of the astrographic photos, but he gave no reason to think that they would systematically err so as to give Newtonian rather than Einsteinian values for light's degree of gravitational bending. Further, the crisp photos from the 4-inch telescope gave a value for gravitational bending that was considerably greater than that predicted by Einstein, to a degree that they could be considered to support Einstein's theory only if that telescope, too, was assumed to be systematically biased. Eddington appeared to be engaged in some rather special pleading, then: he assumed systematic errors in one direction for one of the Brazilian telescopes and in the other direction for the other telescope, so as to reach the conclusion that the results they delivered were quite consistent with Einstein’s theory of relativity. As. W.W. Campbell, an American astronomer and director of the Lick Observatory in San Jose, California, wrote about Eddington’s analysis in 1923: ‘the logic of the situation does not seem entirely clear.’
"If Eddington's reasoning was as murky as his Principe photographs, his aim was pellucid. He wanted very much for Einstein's theory to be true, both because of its profound mathematical beauty and because of his ardent internationalist desire to dissolve the rancor that had some Britons calling for a postwar boycott of German science. (Eddington, as a Quaker, was a committed pacifist; protesting against the proposed boycott, he wrote that 'the pursuit of truth ... is a bond transcending human differences.') These high-minded goals he pursued using the considerable political power at his disposal. He had recruited the Astronomer Royal, Sir Frank Dyson -- 'the most influential figure in British astronomy' -- to his cause early on; it was Dyson who, though he had no personal interest in relativity theory, proposed the eclipse expedition and then took the honorary position as principal author of the expedition's report, all at Eddington's behest.
"When the expedition presented its results, Eddington won an endorsement from the president of the Royal Society and qualified support from the president of the Royal Society and qualified support from the president of the Royal Astronomical Society. Other physicists were more dubious but also less influential and less institutionally powerful. Their reservations were written out of the story: in the wake of the eclipse, Eddington became the preeminent exponent of relativity theory in English, and his discussion of the eclipse experiments was regarded as the standard reference on the topic. While he details and celebrates the pro-Einsteinian measurements provided by the Brazilian 4-inch and Principe telescopes, the Brazilian astrographic results that favored Newton instead are perfunctorily dismissed. Those photographs were on the wrong side of history; consequently, they were entirely blotted out.
"I have begun this chapter with the story of Eddington and the eclipse in part because there is nothing remarkable: it is a rather typical (if unusually well documented) tale of complicated, confused, or ambiguous data, a certain selectivity in the interpretation or reporting of that data, and a concerted effort after the report is made to bend the course of consensus making in a direction favorable to the reporter’s intellectual, moral, or practical aspirations. This is the human mind operating according to its standard specifications, following a trajectory familiar to every student of history -- a pattern of partiality and politicking found in Thucydides’s description of the war between ancient Athens and Sparta, in Gibbon’s Decline and Fall of the Roman Empire in the intrigues of Renaissance Italy’s city-states, and in the backrooms and presidential palaces across the globe today.
"But scientific reasoning is supposed to be an antidote to these primeval inclinations -- and that is what is supposed to explain its extraordinary success. According to Karl Popper, the scientific knowledge machine is driven by an intense critical spirit and by the implacable principle of falsification. Neither is at all evident in Eddington's treatment of the eclipse. Eddington cosseted his own favored theory, shielding it from evidence that looked prima facie falsifying, while damning its rival using reasoning more redolent of the one-sided pleading of a criminal prosecutor than of the evenhanded and straightforward logic of falsification. According to Thomas Kuhn, what distinguishes scientific from ordinary inquiry is scientists' agreement to conduct their research in the framework of the prevailing paradigm, which both sets their goals and instructs them in the interpretation of the evidence. But there is little sign of such a rigid scaffolding in the case of the eclipse. Eddington used his scientific work to realize an aim that lay outside anything that might be dictated by a Kuhnian paradigm namely, a rapprochement between the British and the German scientific establishments. Further, he pursued this and his other aims by interpreting the data in a way that seems driven more by the desire to succeed than by some officially sanctioned, widely accepted procedure for bringing evidence to bear on theory, of the sort that a paradigm is supposed to prescribe. His subsequent political machinations and selective history writing equally seem more inspired by personal, albeit idealistic, ambition than by obeisance to a shared code of scientific conduct.
"Science is so exceptionally powerful, Kuhn argued, because the supremacy of the paradigm guarantees to scientists (so they believe) that their research has a certain fixed significance, underwritten by the goals, experimental methods, and rules for evaluating evidence that constitute the paradigm's core. Eddington's logical and political manipulations, however, disclose exactly the kind of flexibility of rule and pliancy of institutional framework that would set the significance of scientific results perpetually adrift. The Kuhnian paradigm is supposed to preclude such inconstancy. It did not.
"The 1919 eclipse is only a single example of the selective use of evidence. But the centuries since the Scientific Revolution are strewn with cases in which science's biggest names can be seen discarding or distorting difficult data so as to create the impression that experiment was in perfect harmony with their theoretical or other aims.
"Gregor Mendel, the founder of genetics, almost certainly massaged the statistics he presented in the 1860s in support of his thesis that genes lie at the root of biological inheritance. Ernst Haeckel embellished his careful drawings of animal embryos around the same time to support his thesis that ‘ontogeny recapitulates phylogeny’ -- that a human embryo, for example, pass through stages in which it takes on forms more or less identical to those of fish embryos, then amphibian embryos, then bird embryos. Robert Millikan in Robert Millikan, in pulling together the data from which he inferred the electric charge of a single electron -- work that earned him the 1923 Nobel Prize in Physics -- omitted many measurements that did not ‘look right,’ while claiming to have included everything. Even Isaac Newton manipulated certain empirical quantities to better fit his theories, tactics that in one case amounted, wrote his biographer Richard Westfall, to ‘nothing short of deliberate fraud.’"