Academic Legal Writing - Eugene Volokh [188]
2. Extrapolating across places, times, or populations
* I derived these numbers from the General Social Survey datasets for 1991-2002. The GSS is generally seen as a well-conducted nationwide study, which isn't limited to one city and which involves a randomly selected sample (see Part XVII.G.2). I treated those respondents who reported having only same-sex partners in the last 5 years as homosexuals, and those having only opposite-sex partners in the last 5 years as heterosexuals. Few respondents had both same-sex and opposite-sex partners in the last 5 years; their median sexual partner count was 12 (again, since age 18). People who reported not having had any sexual partners in the past 5 years were not included in my analysis. (Of course, this analysis is skewed by the likelihood that some respondents weren't entirely candid, but that's a problem with all surveys.)
The study discussed in Edward O. Laumann, John H. Gagnon, Robert T. Michael & Stuart Michaels, The Social Organization of Sexuality: Sexual Practices in the United States (1994), also appears to be well-conducted and reliable, but unfortunately the Laumann book gives the average sexual partner counts rather than medians, and averages are less helpful than medians because they can be skewed by the behavior of a small fraction of the population. (Remember that the median is the number for which half the data is above it and half is below, while the average is the sum of the data divided by the number of data points—for instance, the average of 1, 1, 2, 2, 3, 4, 5, 22, and 50 is 10, while the median is 3.)
The Laumann book reports that the average number of sexual partners since age 18 for heterosexual men, defined as those who have had only female sexual partners in the last 5 years, is 17, while the average for homosexual and bisexual men is 27 (with a substantial margin of error). Other definitions of sexual orientation yield results of 16 vs. 43, 16 vs. 44, and 17 vs. 30. Id. at 315. So the raw partner counts aren't really comparable to the GSS results, since one can't directly compare averages and medians, but the ratios of homosexuals' partner counts to heterosexuals' partner counts in the two studies are quite consistent: they range from 1.6 to 2.75. And the ratios are not consistent with the claims that the median male homosexual has 250+ lifetime sexual partners, or, as the material quoted in the text a few paragraphs below suggests, 1000+.
3. Say how many cases the comparison is based on, and how small changes in selection may change the result
5. Beware of “10% of all Xs are responsible for 25% of all Ys” comparisons
*
20 For example, the Court relied upon Minersville School Dist. v. Gobitis, 310 U.S. 586, 595 (1940), which allowed the criminal prosecution of school children who refused to pledge allegiance to the flag. It failed to mention, however, that Gobitis was overturned three years after it was decided, by West Virginia State Bd. of Educ. v. Barnette, 319 U.S. 624 (1943). As one commentator noted, “[r]elying on Gobitis without mentioning Barnette is like relying on Plessy v. Ferguson without mentioning Brown v. Board of Education.”
21 The Court claimed that the seminal free exercise case of Wisconsin v. Yoder, 406 U.S. 205 (1972), which granted an exemption to Amish students from compulsory school laws, was decided not on free exercise grounds alone but in combination with the right of parents to direct their children's education. In the words of one who supported the outcome in Smith, the claim that Yoder “was decided on the basis of a ‘hybrid’ constitutional right ... is particularly illustrative of poetic license.” Marshall, supra note 15, at 309 n.3. See also McConnell, supra note 15, at 1121 (“[T]he opinion in Yoder expressly stated that parents do not have the right to violate the compulsory education laws for nonreligious