The last topic I want to look at is value freedom. There is a long tradition in the social sciences, as in science generally, of excluding value judgments from our investigations. In Miller's example, it would be impermissible to explain the fall of the Stuarts by appealing to Stuart injustice (though widespread belief that the Stuarts were unjust could figure in an explanation of the fall of the Stuarts). Moral features are not to be used in explanations. Miller identifies Max Weber as the originator of this view in the social sciences; this is surely oversimplified, as the view is surely implicit in earlier theorists than Weber (Adam Smith, anyone?) and adopted by people not much directly influenced by Weber, but Weber is of course pretty huge in the history of the social sciences, and did emphasize this point.
This is not to say that evaluations should play no role in investigations. Obviously, values determine what scientists will choose to investigate, and it would be absurd to try to do science without allowing values to play that role. Investigating the prevalence and spread of some terrible disease, or the responses to stress of some material commonly used in constructing human artifacts, not only does but should get vastly more attention than ever more precise measurements of exactly how many grains of sand there are in a particular cubic meter of a Saharan sand dune (to borrow an example from Ernest Sosa). What other than human values makes the former more worthy than the latter?
However, it is very important, according to the tradition, not to allow values to lead one to cook the books, as it were. If a scientist feels it would be good to get a particular result, for some reason, and the data doesn't seem to point that way, fudging is generally considered unacceptable. Richard Miller suggests that this is too closed-minded. His example involves anthropological investigations of non-white cultures around a century ago.
According to Miller, some anthropologists of the time were highly motivated by opposition to then nearly universal and unquestioned racism, and actively searched for evidence of cultural sophistication in non-white cultures in order to provide a counter to racist inclinations. It's important to be very careful about what this shows; if they set themselves the task of looking for sophisticated cultural patterns, that's perfectly compatible with Weberian standards. They only violated the rules if they did or would have fudged the results if their data had turned out to be negative. Did they do this, and should they have?
On the empirical question of whether they did, I have no clue. On whether they should have, I say absolutely not. It undermines their credibility. Creating an atmosphere in which fudging data is tolerated for any reason undermines science generally. Further, tolerance for any kind of cheating is going to benefit the entrenched interests more; there are more people willing to cheat for them. A zero-tolerance policy on cheating immunizes those opposed to the entrenched power structure from being criticized for cheating (sure, their critics will find other ways, or just lie, but while it's idealistic to suppose the truth always wins out, it is an advantage to have it on your side), and also avoids charges of hypocrisy when the radical critic attacks research favoring the status quo for failing to live up to standards of objectivity.
In Miller's own example, simply setting the question, simply deciding to investigate the cultural richness of non-white cultures, already helps the cause of radical politics. To bring this back to my positivist revival theme, the positivists favored meticulous value-freedom in the social sciences and very careful attention to logic and methodological issues at a time when the social sciences were dominated by Marxists. Those of Marxist inclination would, of course, be inclined to investigate how the existing capitalist ruling classes maintained power, what conditions were actually like for the poor, what obstacles existed to the poor improving their circumstances, and so forth. Objectively investigating questions like that would naturally produce results which would be of benefit to those engaged in radical politics. Again, injecting bias into the investigation beyond this choice of subjects would undermine the benefits of doing carefully chosen objective investigations. So it seems to me that the positivists were probably taking the optimal stance from the point of view of advancing their radical political agendas in emphasizing objective science and careful methodology.
Of course, just as encouraging particular lines of research can advance political agendas, so can discouraging lines of research. Furthermore, encouraging one line of research can ipso facto end up discouraging other lines, since the resources available for research are limited. So, for example, the radical critique Miller makes of contemporary economics, that its obsession with things like GDP growth is politically problematic, seems to me to be likely well founded. Injecting bias into the investigation of economic problems is not the solution, though; the solution is to find ways to encourage unbiased investigation of more relevant phenomena.