Empirical2
Theories should be consistent with facts, but the facts must be true. Researchers sometimes define variables poorly, measure them badly, collect data from biased samples, or make analytical or even typographical mistakes.
In 1870, Erich Wolf accidentally misplaced a decimal point when transcribing data from his notebook. As a result, spinach was reported to contain 35 rather than 3.5 milligrams of iron per serving, an error not corrected until 1937, by which time studio executives had chosen it as the source of Popeye’s strength (Arbesman 2012). Age, culture, personality, philosophy, politics, rank, religion, and sex of researchers and subjects are among variables that can bias results. Researchers may give undue weight to evidence supporting their own hopes or require a heavier burden of proof for propositions they dislike. An experiment on aggression using undergraduate psychology majors as subjects is likely to get very different results than the same experiment done with Navy SEALs as subjects. Thus, scientists replicate research under different circumstances.Replication validates and generalizes results but also protects against fraud. The problem is rare but significant, and exists in every field. Jayson Blair of The New York Times, Peter Arnett of CNN, and Michael Barnicle of the Boston Globe are among the high profile cases of fraudulent reporting that led to dismissal. Piltdown Man is the most infamous but not the only fraud in anthropological history. Not long ago, Manuel Elizalde costumed and passed off Philippine farmers as a tribe supposedly untouched by civilization.
Experts estimate that as much as 40% of the work of famous artists that appears on the auction block is forged or so heavily restored as to be worthless. Salvador Dali, in a stupor on his deathbed signed hundreds of sheets of blank paper. Giorgio de Chirico backdated pictures twenty years to when his work was popular, in effect forging himself.
Tom Keating specialized in forging Rembrandts, Eric Hebborn in Corots, and Hans van Meegeren in Vermeers. John Myatt did not specialize, crudely forging Braque, Chagal, Giacometti, Klee, Le Corbusier, Matisse and others. Yet he is the most successful art forger of all time because John Drewe convinced Myatt to focus on paintings that would not attract careful scrutiny. Then, posing as a scholar Drewe borrowed catalogs for exhibits at defunct art galleries, unstitched and then rebound them with photographs of Myatt forgeries, reset the title page and table of contents with computer matched print, and returned the forged, rebound catalogs to museum archives. Finally, Drewe created “Art Research Associates,” a company that authenticated Myatt’s forgeries for potential buyers by using the catalogs he had forged himself. (Mould 2010, Salisbury 2010).Three cases of fraud are directly relevant to the study of conflict. The worst case, because it led to the murder of thousands, involves the Protocols of Zion, repeatedly proven fraudulent, used in Russia to justify pogroms and violence against Jews and still cited occasionally as authentic by hate groups in the United States. Michael Bellesiles surrendered his prestigious 2001 Bancroft Prize in American history and resigned from Emory University when his study of gun ownership in the American colonies proved based on data he had fabricated. 3 Michael Moore received an Oscar for Bowling for Columbine and a Palme d’Or at the Cannes Film Festival for Fahrenheit 9/11. 4 Both are, to be polite, error-laden.
Social scientists, impressed by the success of the physical sciences since Galileo, tried to emulate their methods by emphasizing controlled experiments and mathematical analysis. It now is almost impossible to complete a degree in anthropology, communications, economics, political science, psychology, or sociology without at least one course in statistics. Cliometrics5 is a specialized subfield of history.
Social scientists have devised ways to measure almost every aspect of human behavior with varying reliability and validity. They have created elaborate but abstract models to guide their research. The result has been an emphasis on empiricism, and specifically on hypothesis testing, inferential statistics, and research design. John Gaddis, who wrote the leading textbook on the Cold War, had this to say about the effect on political science:Theorists believed the complexities of world affairs could be reduced to a few measurable variables that would allow them to explain the past and predict the future. Hans Morgenthau, a founding father, said it specifically: the study of international relations would ‘increase the reliability of prediction and thereby remove uncertainty from political action.’ That now looks a very bad prediction, for none of our major theories of world politics came close to anticipating the end of the Cold War or the manner of the breakup of the Soviet Union. It will not do to now claim that forecasting never was the intention because they repeatedly said it was. Nor can they argue that the Cold War was an inappropriate test: International Relations theory studied it more than anything else.
Wherever this sort of thing has occurred in the social sciences, it has tended toward oversimplification, in part because the theoreticians cannot reduce human behavior to the few variables that the statisticians can handle. In cases where they do make predictions, people learn what is expected and change strategy to gain an advantage. The fundamental flaw may be failing to integrate quantitative and qualitative facts. Quantitative models can be useful but require understanding of their limitations. Otherwise, conflict theory is likely to take its place alongside such dead ends as the ether, fluidium, and phlogiston.
This may be because the turn to mathematics was only part of the reason that the physical sciences succeeded so spectacularly. The choice of what to measure was even more important.
Galileo chose matter in motion, shifting to questions of how rather than why things happened. Not an obvious choice at the time, it led to rapid progress in science and technology. Social scientists have not yet been able to identify and settle on a few key variables with broad explanatory power, even within major disciplines. Some think human behavior or even any aspect of it such as conflict is too complex ever to identify a sufficiently small number of key variables to make the natural science model work. Some continue the search, but many seem oblivious to the problem and continue to measure everything in sight and subject the results to every conceivable statistical test in hope of finding something that will get them published and tenured.It is useful in studying conflict to think of empiricism as requiring objective determination and interpretation of facts. Leopold von Ranke (2011) established widely accepted principles for objective inquiry summarized as follows:
· Consider all evidence even if some contradicts your views
· Identify sources so others may judge your interpretation
· Limit reconstruction to what the evidence will bear
· Seek out all available evidence without alteration
· Test evidence for reliability and validity
David Brooks (2006) adds these useful criteria:
· Be dull rather than resorting to vitriol and ad hominems
· Process information honestly to identify patterns that may be there
· Remain modest and maintain self-control rather than emoting
· Seek truth rather than supporting friends or personal ideology
· Suspend judgment while collecting all relevant facts