Statistical Methods

by Kieran Healy on October 14, 2004

Maria’s “post about required statistics courses”:https://www.crookedtimber.org/archives/002676.html reminds me of a possibly apocryphal story. I _think_ it concerns one of the very early British social surveys of urban poverty by Charles Booth, or Mackintosh or one of those guys. The results were resisted by many for political reasons, and one strategy was to discredit the new-fangled methods they relied on. Thus, one critic in (I believe) the House of Commons asserted that he could not find the results credible because the report “only relied on a sample of the population — and a mere _random_ sample, at that.”

If anyone knows the source of this (doubtless mangled) story, let me know in the comments.

{ 6 comments }

1

Nathaniel 10.14.04 at 5:56 am

Kieran,

For a more recent example,:
http://www.census.gov/Press-Release/cb97-c25.html

– Of course I’m disgusted – haven’t you been paying attention?

Nathaniel

2

pj 10.14.04 at 2:00 pm

My favorite example in this vein is Supreme Court Justice Lewis Powell’s ridiculing any argument based on statistics as “numerology.” In this view, its far better to rely on a judge’s “gut” than data.

3

praktike 10.14.04 at 6:27 pm

George Bush’s “fuzzy math” allegation against Al Gore in the 2000 debates?

4

Ross 10.14.04 at 6:37 pm

Kieran,

I’ve not heard the quote, but from the context you provide, it sounds like it could have related to G. Udny Yule’s 1899 study of pauperism in England.

See here for more details on the study.

5

Donald A. Coffin 10.15.04 at 6:37 pm

This is not a response to Kieran’s specific question, but a reference to an extremely interesting and useful book for anyone interested in statistics. (I’d have posted this earlier, but have been working from home with a bad back and the book is in my office.) The Lady Tasting Tea: How Statistics Revolutionized Science in the Twentieth Century W. H. reeman, 2001), by David Salsburg, is about the uses (and some of the abuses) of statistics in helping us understand the world around us.

6

Don 10.16.04 at 3:16 pm

Nathaniel,

The debate about whether the Census should use statistical sampling was not merely about statistical ignorance, but rather about the very phrase in the Constitution which established the Census Bureau, which requires an ‘actual enumeration’. Not a statistically sampled ‘count’, but an actual physical count.

It may be gratifying to put the argument against sampling down to ignorance, but not entirely accurate I think….

Comments on this entry are closed.