I'm not trying to "spin" your words, Michael. If I've misunderstood you, I apologize. I do think your initial post said something different than what you're saying now, though -- and I happen to disagree with what I understand you to be saying in both cases.
email@example.com (michael goldenberg) says: >Those of us who are skeptical of quantitative methods in the social >sciences are not skeptical of quantitative methods per se, a point I made >clearly in my post. In my case, I simply question the "objectivity" of >so-called hard data when it comes to most educational issues. Quantifying >height, weight, production yield, or SAT scores are equally easy, but not >equally meaningful.
Actually, your post I first responded to said educational data was not replicable, that each class was unique and thus no quantitative analysis could be made. *That's* hooey. Goodness, if it comes to that, each coin toss is unique. In some sense, statistical analysis is about filtering out those fluctuations to see ... what's left.
I do agree that one can dispute the meaning or usefulness of various measures. Sit down, fill-in-the-bubble tests may not be the right way to gather the useful information about student skill and understanding or the effectiveness of various teaching methods. That doesn't mean that there is no way to gather such information. And I still think one is on firmer ground disputing the meaning or relevance of a statistic than arguing about the right method with no statistics at all.
>For my money, even at its best "objective" science involves the >reading and unpacking of data in terms of theories. And since science >would appear to be socially constructed and contingent, as well as >constantly evolving, I question the notion that quantifying gives us >unquestionable truth.
Unquestionable truth, maybe not. Still, it is at least more concrete than one's own opinions, divorced from data, or supported only by small collections of non-quantified data. After all, one's own opinions are themselves "socially constructed, contingent, constantly evolving."
>I'm willing to look at data and consider its implications; I'm just not >fooled into thinking that I'm more of a "scientist" when I do so than I am >when I explore the implications of the ethnographies I collect.
I doubt I can convince you otherwise -- maybe this is *my* religious belief -- but, while I'm aware of the limitations of quantitative approaches, I think most of those limitations are shared by other methods.
You are no less a scientist when collecting ethnographies, but you are a lot more subjective than you might be when investigating quantitative data. Such conclusions are *more* susceptible to personal biases and *less* verifiable by other investigators. Subjectivity is not all-or-nothing---controlled observation and statistical analysis is simply a lot *less* subjective than free-form personal impressions.