At 06:51 PM 8/28/97 -0700, Tom Walker wrote, inter alia: >>Why? >> . . . we can measure the >>"ideological temperature" of the environment (specific communities) to know >>what individual members on average think. > >I'm not sure I understand what you're getting at here. But "the ideological >temperature of individuals on average" doesn't sound like information I'd be >particularly interested in. Sorry for the too cryptic of a metaphor. What I intended to say was that researcher/interviewer is often and correctly (cf. Foucault) perceived as a some sort of an authority figure, and the Rs tend to to throw back at him/her what they hear from the offcial sources: the government, the media, the pulpit etc. Many so-called "public policy" issue as formulated in such a way that have little direct relationship to people's everyday's lives -- consequently people have no personal stakes in this or that responses. In that context, giving the "politically correct" answers may be dictated by really trivial considerations, such as the desire to please the researcher by giving him/her the answers the Rs think they want to hear, the desire not to draw suspicion by giving "politically incorrerct" answers, or simply by a desire to get the researcher off one's back by giving him/her the "standard" answers. However, situation changes quite dramatically when people have personal stakes in an issue. Take the example of abortion. If formulated in abstraction, people may associated it with the way it is often portrayed from the pulpit or by government officials: as abstract "liberated" women having it for the heck of it, or to destroy 'family values' and 'our way of life' etc. And it is quite obvious what kind of answer are they likely to give to the question "do you support abortion on demand?" The situation changes quite dramatically when the Rs herself, or his/her daughter or relative become pregnant -- in such a case the options (abortion vs continuation of preganancy) are formulated in a way that reflects that person's personal stakes. So the person may say "no" to the survey question "do you support abortion on demand," yet seek abortion if she or her daughter becomes pregnant -- and what even more important do not see any contradiction between these two position. She may pe4rceive herself as different from 'those women" she heard from her preacher. This is reflected by the following anecdote from the Russian Revolution. A peasant wanted to join the CP. To test his sincerity, the commissar asked him "Would you give your cow to the Party, if the Party asked you?" "Sure" answered the peasant. But that did not convinced the suspicious commissar so he continued "Would you also give your horse to the Party?" "Of course!" "And how about your pig, would you give it too?" "In an instant!" "And your sheep?" "My sheep!!?? -- No way, sir!" Nonplussed, the commissar said "You just told me that you would give your horse, your cow and your pig, yet you do not want give your sheep that is of much smaller value than the other animals. I do not understand that. Why???" "Because I have a sheep" --answered the peasant. The funniest part of this anecdote, though, is not the thinly veiled opportunism of the peasant, but the contradictions in the commissar's behaviour: on the one hand, he expected a "true consciousness" =lack of attachment to private property on the part of the peasant, yet despite his professed vanguard position, he displayed quite a bit of "false consciousness" =attachement to bourgeois methods of inquiry. A true revolutionary would have known the peasant's unwillingness to give up his meager private property, beacuse the possession of such property was the condition of his survival under the capitalist mode of production. In short, it does not matter what the Rs think at the moment, because what they think reflects their living conditions. Change their living conditions, and the content of their thinking will change. The bourgeois social scientists, OTOH, want to do it the other way around -- "testing" the content of people's thinking to find the (supposedly) best living conditions. >I can only reiterate that Q methodology isn't in the slightest interested in >what the "majority" of people support or even in what some given proportion >of people think. Q methodology seeks to interpret coherent varieties of >opinion on an issue. Well, from your description I gather that q uses factor analysis to find commonalities among different indicators. Factor invariably uses linear regression, and the core of regression is the prediction of the mean value of the dependent variable. The only difference is that unlike in ordinary regression, the values of "dependent" variable or presumed "common factor" are not known in factor, instead -- they are predicted from the values of indicator variables and "loadings" (that are simply standardized regression coeffcients of several regression equations in which the values of indicator variables serve as x- values or 'idependent' variables ). Therefore, the "commonality" discovered that way is nothing more but the predicted mean value of the 'dependent' variable, just like your y^ in ordinary regression. Now, since the "mean predicted value" implies, in my mind, a number of observations representing different people (we are simply predicting their average value on the dependent variable knowing how they scored on the indpependent variables, or indicator items) -- I somehow cannot see how q is not interested in "majority" opinions except in the most technical sense, where "majority" refers to count-based measures (such as percentages) rather than continuous ones. Again, from your description I gather that the strength of q lies in the completeness of the list of indicators used to find common factors -- a nontrivial thing for regression-based methods that guarantees that an important explanatory variable is not missing, that is 'hidden' among the unexaplained variance (or error) which is supposed to be uncorrelated to the dependent variable (not including an important predictor obviously violates that assumption). But the completeness of the list of indicators (or independent variables) does not guarantee that any given item is interpreted in the same way by all the Rs. Using my previous example, a R who has not come into a personal contact with the abortion issue may interpret the whole battery of abortion-related questions in a much different way from a person who personally encountered that problem in this or another way. Perhaps there is something I missed, but I simply cannot see how a method using standardized cues (ie. the same cues given to a number of Rs - and every survey by definition must rely on that, because it gurantees its "generalizability") can avoid giving anything but standardized answers that ignore individual idiosyncrasies. regards, wojtek sokolowski institute for policy studies johns hopkins university baltimore, md 21218 [EMAIL PROTECTED] voice: (410) 516-4056 fax: (410) 516-8233 POLITICS IS THE SHADOW CAST ON SOCIETY BY BIG BUSINESS. AND AS LONG AS THIS IS SO, THE ATTENUATI0N OF THE SHADOW WILL NOT CHANGE THE SUBSTANCE. - John Dewey
