Site hosted by Angelfire.com: Build your free website today!





A STATISTICAL SIDENOTE



Almost all forms of statistical analysis, including the simple presentation of percentages, are founded on the assumption that they are drawn from a random sample. A random sample is a selection in which every element of a set of data has an equal chance of being selected. To use Siena's work as an example, a random sample would mean that every venereal disease patient in a London workhouse wass equally likely to be labelled 'foul,' and thus to be visible for Siena's statement of percentages. It is very unlikely that this assumption of a random sample is true, as gender, age, and VD sufferers' attempts to disguise their condition likely had an impact on how they were labelled.

The sample is biased, not random, and furthermore we do not know how it is biased. For instance, we do not know whether older women were more or less likely to be classified as 'foul' than were younger ones. Therefore, the mathematical foundation on which a statement like "96 percent of all foul women were younger than 35" rests is deeply flawed (Siena, p. 164). It would be difficult to defend saying anything more concrete about the proportion of foul women under 35 than "probably most."

Although I am inclined to fault Siena for the use of rhetorical techniques and "the visual impact of graphs" (Hitchcock, p.236) that present his rough and tenuous estimates as nearly facts, it is possible, given the linguistic turn in history, to see this as only one of many devices that academics must use in order that their work may be accessible to readers. I estimate that at least 64% of my classmates have already been entirely turned off by my opaque and nearly-irrelevant esoteric discussion of this matter. There is little reason for Siena to lose 64% of his readership as well. In other projects intended for less academic and intimate audiences, I too have glossed over the esoterica of statistical method in the quest for accessibility. Being able to say "96%" rather than "probably most" makes your work both more convincing and more readable, and that is what writing has always been about.







relevance    reality     position     writing     statistics     questions     realityandrelevance