Yesterday, Stephen Pincus of Yale, History, gave a talk to the Law/History seminar on the English Revolution of 1688. His point was that it was not a revolution of modernism vs. an entrenched past, but rather of two possible paths to modernism (land vs. human capital). I asked him how everyone else had gotten it wrong. First, he pointed out that early on they did not get it wrong at all--it was violent, and the conflicts were well understood. But somewhere in the eighteenth century, it was taken as the more conventional conflict, and Macaulay's History entrenched that national myth (although M's research notes are much richer and clearly M understood the whole story). What has made it possible to tell the new story is the availability of archives, archives that were once privately held by families but then were given to the state by families to deal with taxes, AND that Pincus being American has no need to maintain the conventional civics history.

He pointed out as well, that the conflicts in the American Revolution were quite alive in England, and so the Amer Revolution is not us vs. them, but in fact from the point of view of world history were going on in India, the West Indies, and in England, and they were us vs. us.

The lesson here is salutary: If you have limited access to data and information, and you always will be in that position!, you may well miss something BIG. If you have an interest in a particular story (that is what national histories always are), you may be perpetrating myth rather than a more complex story. And if you don't broaden your point of view, you may succeed but in the end your work is much less useful or true. In particular, if you do quantitative analysis, it does help to talk to people. You can't write about elderly people's housing preferences by thinking about what YOU might like when you are older (and you are 35!), nor just talk to your relatives. You have to do some actual fieldwork--or at least find out what others who have done such fieldwork have discovered--even if you have wonderful quantitative data sets and terrific theory. You just might be wrong. Of course, the smell test always applies: If the results seem to be too good to be true, be sure they aren't. You can't measure things to 1 in a 1000 in social science research, reliably that is!

The task, always, is to find out about the world in a reliable way. It is only secondary that it be published, in the best journals. People can go very wrong while doing methodologically impeccable work.