Supplementary material for Brewer et al. (2000) Proc. Natl. Acad. Sci. USA, 10.1073/pnas.210392097
Appendix
The supplementary material here pertains to detailed numerical calculations for the analyses discussed in the article, the prevalence of prostitute women, prostitutes’ numbers of sexual partners, and the role of prostitution in accounting for the discrepancy in countries other than the United States.
Detailed Numerical Calculations.
This section provides the detailed numerical calculations for the sex discrepancy ratios (both unadjusted and adjusted for prostitution).General Social Survey (GSS), 1988–1991.
All calculations are based on weighting observations by the ADULTS (household size) variable. National population figures are from the U.S. Census.12-Month recall period, unadjusted discrepancy ratio:
Number of heterosexually active men in the U.S.
= 0.83931 (proportion from GSS) ´ 88,748,000 men in the U.S. ages 18 and older in 1990
= 74,487,000
Number of heterosexually active women in the U.S.
= 0.76370 (proportion from GSS) ´ 96,529,000 women in U.S. ages 18 and older in 1990
= 73,719,000
Number of partnerships reported by heterosexual men in the U.S.
= 2.0839 (mean number of partners reported by heterosexual men in GSS) ´ 74,487,000
= 155,220,000
Number of partnerships reported by heterosexual women in the U.S.
= 1.2136 (mean number of partners reported by heterosexual women in GSS) ´ 73,719,000
= 89,465,000
Unadjusted discrepancy ratio
= 155,220,000/89,465,000
= 1.7350
12-Month recall period, prostitution-adjusted discrepancy ratio:
Number of prostitutes expected in the GSS
= 0.001026 (proportion of women 18–44 in U.S. who were prostitutes in 1990 based on Census and Colorado Springs estimates) ´ 1,626 women in GSS ages 18–44 (who reported their numbers of partners in response to the PARTNERS or PARTOPEN questions)
= 1.6683
Expected number of prostitution-related partnerships in the GSS
= 1.6683 expected prostitutes ´ 694 expected partnerships per prostitute
= 1,157.8
Unrepresented number of prostitution-related partnerships in the GSS
= 1,157.8 expected prostitution-related partnerships – [60.5 partnerships reported by one observed prostitute ´ 0.508 (household size weight for this respondent)]
= 1,127.1
Prostitution-adjusted mean number of partnerships reported by heterosexual women
= {[1.2136 (mean number of partners reported by heterosexual women in the GSS) ´ 2,074 heterosexual women in GSS] + 1,127.1 unrepresented prostitution-related partnerships}/2,074 heterosexual women in GSS
= 1.7570
Prostitution-adjusted number of partnerships for women in the U.S.
= 73,719,000 heterosexual women in the U.S. ´ 1.7570 (prostitution-adjusted mean number of partners for heterosexual women)
= 129,520,000
Number of partnerships reported by heterosexual men in the U.S. after removing those commercial partnerships involving juvenile prostitutes
= 155,220,000 (number of partnerships reported by heterosexual men) – 740,000 (number of commercial partnerships in the U.S. attributable to juvenile prostitutes, based on Census and Colorado Springs estimates)
= 154,480,000
Prostitution-adjusted discrepancy ratio
= 154,480,000/129,520,000
= 1.1927
Proportion of prostitution-adjusted number of heterosexual women’s partnerships accounted for by prostitution
= (55,134 prostitutes in U.S. ´ 694 partners per prostitute)/129,520,000
= 0.2954
Proportion of expected number of prostitutes’ partnerships accounted for by clients’ total number of observed partnerships
= 169.43 (total number of partners reported by acknowledged clients of prostitutes in GSS)/1,157.8 (expected number of prostitution-related partnerships)
= 0.1463
5-Year recall period, unadjusted discrepancy ratio:
Number of heterosexually active men in the U.S.
= 0.8905 (proportion from GSS) ´ 88,748,000 (men in the U.S. ages 18 and older in 1990)
= 79,030,000
Number of heterosexually active women in the U.S.
= 0.8468 (proportion from GSS) ´ 96,529,000 (women in the U.S. ages 18 and older in 1990)
= 81,741,000
Number of partnerships reported by heterosexual men in the U.S.
= 4.5736 (mean number of partners reported by heterosexual men in GSS) ´ 79,030,094
= 361,450,000
Number of partnerships reported by heterosexual women in the U.S.
= 1.8102 (mean number of partners reported by heterosexual women in GSS) ´ 81,741,000
= 147,970,000
Unadjusted discrepancy ratio
= 361,450,000/147,970,000
= 2.4427
5-Year recall period, prostitution-adjusted discrepancy ratio:
Number of prostitutes expected in GSS
= 0.001026 (proportion of women 18–44 in U.S. who were prostitutes in 1990 based on Census and Colorado Springs estimates) ´ 420 women in GSS ages 18–44 (who reported their numbers of partners in response to the PARTNRS5 or PARTOPN5 questions)
= 0.43092
Expected number of prostitution-related partnerships in the GSS
= 0.43092 expected prostitutes in GSS in 1 year ´ 694 (expected partnerships per prostitute) ´ 5 years
= 1,495.3
Unrepresented number of prostitution-related partnerships in the GSS
= 1,495.3 expected prostitution-related partnerships – 19.7 observed prostitution-related partnerships (total number of partnerships reported by heterosexual prostitutes in the GSS)
= 1,475.6
Prostitution-adjusted mean number of partnerships reported by heterosexual women
= {[1.8102 (mean number of partners reported by heterosexual women in the GSS) ´ 583 heterosexual women in the GSS] + 1,475.6 unrepresented prostitution-related partnerships}/583 heterosexual women in the GSS
= 4.3412
Prostitution-adjusted number of partnerships for heterosexual women in U.S.
= 81,741,000 heterosexual women in the U.S. ´ 4.3412 (prostitution-adjusted mean number of partners for heterosexual women)
= 354,850,000
Number of partnerships reported by heterosexual men in the U.S. after removing those commercial partnerships involving juvenile prostitutes
= 361,450,000 (number of partnerships reported by heterosexual men) – 1,478,600 (number of commercial partnerships in the U.S. in the last 5 years attributable to prostitutes who were juveniles throughout the last 5 years, based on Census and Colorado Springs estimates)
= 359,970,000
Prostitution-adjusted discrepancy ratio
= 359,970,000/354,850,000
= 1.0144
Proportion of prostitution-adjusted number of heterosexual women’s partnerships accounted for by prostitution
= (55,134 prostitutes in the U.S. ´ 694 partners per prostitute ´ 5 years)/354,850,000
= 0.5391
National Health and Social Life Survey (NHSLS), 1992.
All calculations are based on weighting observations by the HHSIZE (household size) variable. National population figures are from the U.S. Census.12-Month recall period, unadjusted discrepancy ratio:
Number of heterosexually active men in the U.S.
= 0.8748 (proportion from NHSLS) ´ 72,707,000 men in the U.S. ages 18-59 in 1992
= 63,604,000
Number of heterosexually active women in the U.S.
= 0.8648 (proportion from NHSLS) ´ 73,509,000 women in U.S. ages 18-59 in 1992
= 63,571,000
Number of partnerships reported by heterosexual men in U.S.
= 1.8899 (mean number of partners reported by heterosexual men in NHSLS) ´ 63,604,000
= 120,210,000
Number of partnerships reported by heterosexual women in U.S.
= 1.2906 (mean number of partners reported by heterosexual women) ´ 63,571,000
= 82,045,000
Unadjusted discrepancy ratio
= 120,210,000/82,045,000
= 1.4652
12-Month recall period, prostitution-adjusted discrepancy ratio:
Number of prostitutes expected in NHSLS
= 0.001043 (proportion of women ages 18–44 in U.S. who were prostitutes in 1992 based on
Census and Colorado Springs estimates) ´ 1,295 women in NHSLS ages 18–44 [who reported
their numbers of partners (PART12 variable)]
= 1.3507
Expected number of prostitution-related partnerships in NHSLS
= 1.3507 expected prostitutes ´ 694 expected partnerships per prostitute
= 937.39
Unrepresented number of prostitution-related partnerships in NHSLS
= 937.39 expected prostitution-related partnerships – 2.13 observed prostitution-related partnerships (total number of partners reported by heterosexual prostitutes in the NHSLS)
= 935.26
Prostitution-adjusted mean number of partnerships reported by heterosexual women
= {[1.2906 (mean number of partners reported by heterosexual women in NHSLS) ´ 1,385 heterosexual women in NHSLS] + 935.26 unrepresented prostitution-related partnerships}/1,385 heterosexual women in NHSLS
= 1.9659
Prostitution-adjusted number of partnerships for heterosexual women in U.S.
= 63,571,000 heterosexual women in the U.S. ´ 1.9659 (prostitution-adjusted mean number of partners for heterosexual women) ´ 0.972 (proportion of commercial partnerships involving men less than age 60, based on Colorado Springs estimates)
= 121,470,000
Number of partnerships reported by heterosexual men in the U.S. after removing those commercial partnerships involving juvenile prostitutes
= 120,210,000 (number of partnerships reported by heterosexual men) – 743,100 (number of commercial partnerships in the U.S. attributable to juvenile prostitutes, based on Census and Colorado Springs estimates)
= 119,470,000
Prostitution-adjusted discrepancy ratio
= 119,470,000/121,470,000
= 0.9835
Proportion of prostitution-adjusted number of heterosexual women’s partnerships accounted for by prostitution
= (56,355 prostitutes in U.S. ´ 694 partners per prostitute)/121,470,000
= 0.3220
Proportion of expected number of prostitutes’ partnerships accounted for by clients’ total number of observed partnerships
= 219.35 (total number of partners reported by acknowledged clients of prostitutes in NHSLS)/937.39 (expected number of prostitution-related partnerships)
= 0.2340
5-year recall period, unadjusted discrepancy ratio:
Number of heterosexually active men in the U.S.
= 0.8930 (proportion in the NHSLS) ´ 72,707,000 men in the U.S. ages 18–59 in 1992
= 64,927,000
Number of heterosexually active women in the U.S.
= 0.9004 (proportion in the NHSLS) ´ 73,509,000 women in the U.S. ages 18–59 in 1992
= 66,188,000
Number of partnerships reported by heterosexual men in the U.S.
= 4.5367 (mean number of partners reported by heterosexual men in the NHSLS) ´ 64,927,000
= 294,550,000
Number of partnerships reported by heterosexual women in the U.S.
= 2.1039 (mean number of partners reported by heterosexual women in the NHSLS) ´ 66,188,000
= 139,250,000
Unadjusted discrepancy ratio
= 294,550,000/139,250,000
= 2.1153
5-Year recall period, prostitution-adjusted discrepancy ratio:
Number of prostitutes expected in the NHSLS
= 0.001043 (proportion of women ages 18–44 in U.S. who were prostitutes in 1992 based on Census and Colorado Springs estimates) ´ 1,239 women ages 18–44 in the NHSLS (who reported their numbers of partners in response to the PARTNRS5 question)
= 1.2923
Expected number of prostitution-related partnerships in the NHSLS
= 1.2923 expected prostitutes ´ 694 expected partnerships per prostitute ´ 5 years
= 4,484.3
Unrepresented number of prostitution-related partnerships in NHSLS
= 4,484.3 expected prostitution-related partnerships – 120.35 observed prostitution-related partnerships (total number of partnerships reported by heterosexual prostitutes in the NHSLS)
= 4,364.0
Prostitution-adjusted mean number of partnerships reported by heterosexual women in NHSLS
= {[2.1039 (mean number of partners reported by heterosexual women in NHSLS) ´ 1,487 heterosexual women in NHSLS] + 4,364 unrepresented prostitution-related partnerships}/1,487 heterosexual women in NHSLS
= 5.0387
Prostitution-adjusted number of partnerships for heterosexual women in U.S.
= 66,188,000 heterosexual women in the U.S. ´ 5.0387 (prostitution-adjusted mean number of partners for heterosexual women) ´ 0.972 (proportion of clients of prostitutes under age 60 throughout last 5 years, based on Colorado Springs estimates) ´ 0.994 (reduction factor to account for clients ages 55–59 during the last 5 years but over age 59 at time of interview, based on Colorado Springs estimates)
= 322,220,000
Prostitution-adjusted number of partnerships reported by heterosexual men in U.S.
= 294,560,000 (number of partnerships reported by heterosexual men) – 1,486,200 (number of commercial partnerships in the U.S. in the last 5 years attributable to prostitutes who were juveniles throughout the last 5 years, based on Census and Colorado Springs estimates)
= 293,070,000
Prostitution-adjusted discrepancy ratio
= 293,070,000/322,220,000
= 0.9095
Proportion of prostitution-adjusted number of heterosexual women’s partnerships accounted for prostitution
= (56,355 prostitutes in the U.S. ´ 694 partners per prostitute ´ 5 years)/322,220,000
= 0.6069
Prevalence of Prostitute Women
To our knowledge, in the last 25 years only two other studies in North America have used systematic, empirical methods for estimating the prevalence of prostitute women. For each study, we made additional calculations based on the assumption that prostitute women in other North American communities were as mobile as those in Colorado Springs. Potterat et al.’s (1) estimate of the prevalence of full-time resident prostitutes in Colorado Springs takes into account prostitutes’ mobility. They assigned evanescent prostitutes (who worked intermittently in Colorado Springs for fewer than 40 days in a year) a weight of 0.1, short-term prostitutes (who worked in Colorado Springs no more than several months in a year) a weight of 0.5, and long-term prostitutes (who were present in Colorado Springs throughout a year, even if not working continuously) a weight of 1.0. In our additional calculations for the other studies, we multiplied the prevalence of prostitutes observed in a community by a certain proportion to obtain an estimate of the prevalence of full-time resident equivalent prostitutes. This proportion is 0.72 for a one-year period, which represents the mean ratio of the number of full-time equivalent prostitutes to the total number observed in a year in Colorado Springs between 1970 and 1988. For 2- and 3-year periods, the proportions are 0.64 and 0.59, which represent the ratios of the number of full-time equivalent prostitutes to the total number observed in Colorado Springs in 1987–1988 and 1986–1988, respectively.
Through snowball sampling, Decker (2) enumerated 39 active adult prostitutes in a small, seemingly typical Midwestern city of 28,000 residents. He noted that these women displayed relatively little mobility. On the assumption that all of these women were full-time residents of this city in the year prior to his study, the prevalence rate is 139 per 100,000. If the county population of 49,000 is taken instead as the denominator in the calculations, the prevalence is 80 per 100,000. If this county prevalence rate is adjusted for a level of prostitutes’ mobility the same as in Colorado Springs, the result is 57 per 100,000. These rates contrast with the 23 per 100,000 estimate of the annual prevalence in Colorado Springs.
Based on police arrest records from Vancouver, Canada, Rossmo and Routledge (3) estimated the prevalence of prostitute women who worked on the street during a 23-month period in 1986–1987. Eight hundred eighty-six different women were arrested for prostitution during this period. Rossmo and Routledge first assumed that the arrests of prostitute women resembled a capture–recapture process and that women who had been arrested once were less likely to be arrested again than women not already arrested as a result of deterrence (quitting prostitution or moving away). Next, they incorporated empirical information on the amount of time prostitutes spent working on the street and used an inverse Gaussian model to estimate the total number of prostitute women (arrested and unarrested) who worked in Vancouver during this period. Their analysis showed an estimated number of 1,610 prostitute women (95% confidence interval = 1,380–2,000). Based on population estimates for the Vancouver Census Metropolitan Area (4) and adjustments for mobility for a 2-year period, the prevalence rates are 41 per 100,000 and 74 per 100,000 (95% confidence interval = 63–92 per 100,000) for arrested and total estimated street prostitute women, respectively. In contrast, the 2-year prevalence in Colorado Springs in 1987–1988 is 30 per 100,000 (1).
In summary, other systematic, empirical estimates indicate prevalences substantially larger than that for Colorado Springs. We used the Colorado Springs estimate in our calculations because it is conservative and because Colorado Springs, of the three communities, most resembles the United States overall in demographic terms.
Prostitutes’ Number of Male Sexual Partners
Table 2 summarizes studies of prostitute women in the United States with respect to the number of reported male sexual partners for different recall periods. We excluded studies in which the number of partners was expressed in terms of number of partners per day/night without information about how frequently the sampled prostitutes worked.
The mean number of partners reported by prostitute women in the 1988–1991 Colorado Springs study (5) is conservative compared with the reports of prostitutes in other communities. For the 6-month and 5-year recall periods, the Colorado Springs mean is less than the mean for 10 of the 12 other communities. The two exceptions, Atlanta and Trenton, show means close to the Colorado Springs estimate. The higher estimates from the other communities seem to be due to less representative samples recruited primarily from streets and jails. Taken as a whole, the results from studies using other recall periods also suggest that the Colorado Springs estimate is reasonable, if not conservative. Although the maximum reported numbers for prostitute women in Los Angeles, Miami, and Newark seem exceptionally large, the median reported numbers for women in Los Angeles and Miami still exceed the mean from the 1988–1991 Colorado Springs study. In general, the distributions are fairly strongly skewed to the right, indicating that a small proportion of prostitute women account for a large fraction of commercial partnerships.
Short recall periods (such as a week or a month) are not suitable for estimating the numbers of partners prostitutes have in a year. Multiplying a mean for a short period by some factor to obtain an estimate for a longer period will lead to overestimation because many prostitutes do not work consistently throughout a year and repeat clients may be counted multiple times.
Table 2. Number of sexual partners reported by prostitute women in the United States for different recall periods
Location | Year | Sample sources | n | Mean | Median | IQR | Range |
Typical week |
|
|
|
|
|
|
|
Small Midwestern city | 1974 | Snowball sample (2) | 29 | 28* | — | — | — |
Seattle, WA | 1987–1988 | County jail (6), casual (nonrecurring) clients | 48 | 33 | — | — | 1–334 |
| 1987–1988 | County jail (6), regular clients | 48 | 14 | — | — | 0–75 |
|
|
|
|
|
|
|
|
Chicago, IL | 1990–1991 | Streets and jail (7) | 141 | 48 | — | — | 1–301 |
Rural Nevada | 1993 | Three legal brothels (8) | 41 | 111 | — | — | 40–336 |
Rural Nevada | 1995 | Two legal brothels (9) | 40 | 82 | 69 | — | 5–240 |
Five cities | 1992–1996 | Needle exchange programs in New York City, NY; Rochester, NY; Buffalo, NY; Los Angeles, CA; and Chicago, IL (10) | 235 | 20 | — | — | — |
Last 3 months |
|
|
|
|
|
|
|
Seattle, WA | 1991–1993 | Streets, shopping centers, motels, restaurants, bars, nightclubs, and escort service (locations selected based on extensive ethnographic research) (11)† | 57 | 59 | 20 | 7–55 | 2–434 |
Last 6 months |
|
|
|
|
|
|
|
Colorado Springs, CO | 1988–1991 | County STD clinic, HIV testing site, drug treatment program, outreach, and sexual, drug, and social contacts of other study participants (prostitutes and nonprostitutes) (5)† | 98 | 347 | 103 | 11–228 | 1–5,401 |
New York City, NY | 1991 | Dominican prostitutes from brothels, streets, restaurants, and nightclubs (12) | 77 | 1,208 | — | — | — |
Location | Year | Sample sources | n | Mean | Median | IQR | Range |
Last year |
|
|
|
|
|
|
|
New York City, NY | 1986–1987 | Respondent-referred prostitutes working in massage parlors or for escort agencies (13) | 78 | 256 | 200 | — | — |
Last 5 years |
|
|
|
|
|
|
|
Colorado Springs, CO | 1988–1991 | County STD clinic, HIV testing site, drug treatment program, outreach, and sexual, drug, and social contacts of other study participants (prostitutes and nonprostitutes) (5)† | 98‡ | 2,171 | 406 | 46–2,044 | 1–27,900 |
Atlanta, GA | 1986–1988 | Halfway houses, homeless shelters, correctional agencies, flyers, newspaper advertisements, streets, bars, clubs, houses of prostitution, and escort agencies (14)† | 123 | 1,567 | 300 | 33–1,050 | 2–23,042 |
Atlantic City, NJ | 1987 | Streets, STD clinic (14)† | 16 | 7,699 | 5,690 | 1,812–8,842 | 469–36,104 |
Colorado Springs, CO | 1986–1987 | STD, drug treatment, and HIV testing clinics (14)†§ | 98 | 3,743 | 1,380 | 128–5,097 | 2–31,642 |
Jersey City, NJ | 1986–1987 | Methadone maintenance clinic (14)† | 85 | 2,561 | 1,154 | 504–3,428 | 22–18,252 |
Los Angeles, CA | 1986–1987 | Jail (14)† | 308 | 9,909 | 7,201 | 3,600–13,377 | 21—90,010 |
Miami/Dade County, FL | 1986–1988 | Jail (14)† | 448 | 5,534 | 2,242 | 302–6,659 | 1–80,002 |
Rural southern Nevada | 1986–1987 | Brothels and applicants to become licensed brothel prostitutes (14)† | 38 | 3,833 | 1,167 | 285–5,265 | 30–32,869 |
Newark, NJ | 1986–1988 | Streets, STD clinic (14)† | 81 | 4,854 | 940 | 452–4,204 | 1–93,743 |
Paterson, NJ | 1987 | Streets, STD clinic (14)† | 2 | 2,433 | 2,433 | — | 2,011–2,855 |
San Francisco, CA | 1986–1988 | Streets (14)† | 212 | 3,760 | 1,184 | 205–6,062 | 1–27,250 |
Trenton, NJ | 1987 | Streets, STD clinic (14)† | 13 | 1,730 | 1,010 | 157–3,279 | 6–5,299 |
Location | Year | Sample sources | n | Mean | Median | IQR | Range |
Lifetime |
|
|
|
|
|
|
|
Atlanta, GA | 1986–1988 | Halfway houses, homeless shelters, correctional agencies, flyers, newspaper advertisements, streets, bars, clubs, houses of prostitution, and escort agencies (14†, 15) | 123 | 2,679 | 500 | 70–2,050 | 3–52,417 |
Atlantic City, NJ | 1987 | Streets, STD clinic (14)† | 16 | 12,081 | 10,450 | 5,006–18,409 | 469–36,104 |
Colorado Springs, CO | 1986–1987 | County STD, drug treatment, and HIV testing clinics (14)†§ | 98 | 6,458 | 3,243 | 294–7,198 | 5–84,005 |
Jersey City, NJ | 1986–1987 | Methadone maintenance clinic (14)† | 85 | 3,785 | 2,025 | 662–6,337 | 25–25,200 |
Los Angeles, CA | 1986–1987 | Jail (14)† | 305 | 13,488 | 7,210 | 2,885–18,002 | 1–99,999 |
Miami/Dade County, FL | 1986–1988 | Jail (14)† | 446 | 7,475 | 2,413 | 379–9,001 | 2–99,999 |
Rural southern Nevada | 1986–1987 | Brothels and applicants to become licensed brothel prostitutes (14)† | 38 | 5,709 | 2,712 | 505–7,899 | 31–32,869 |
Newark, NJ | 1986–1988 | Streets, STD clinic (14)† | 81 | 8,145 | 1,500 | 615–5,773 | 6–93,750 |
Paterson, NJ | 1987 | Streets, STD clinic (14)† | 2 | 2,435 | 2,435 | — | 2,010–2,860 |
San Francisco, CA | 1986–1988 | Streets (14)† | 213 | 6,126 | 2,250 | 268–9,999 | 8–42,590 |
Seattle, WA | 1991–1993 | Streets, shopping centers, motels, restaurants, bars, nightclubs, and escort service (locations selected on the basis of extensive ethnographic research) (11)† | 210 | 3,833 | 233 | 35–1,980 | 4–109,504 |
Trenton, NJ | 1987 | Streets, STD clinic (14)† | 13 | 4,961 | 3,844 | 157–6,711 | 20–19,434 |
STD, sexually transmitted diseases; IQR, interquartile range.
*Based on multiplying the mean number of clients per night by the mean number of nights/days worked per week.
†Based on unpublished data from this study.
‡Includes only those women who worked as prostitutes in the last 6 months.
§This sample includes 23 women who were prostitutes in the last 6 months in the 1988–1991 Colorado Springs study and 28 women who were prostitutes in the last 5 years in that study.
Role of Prostitution in Accounting for the Discrepancy in Other Countries
To our knowledge, Great Britain and the Ivory Coast are the only other countries for which national sex survey data and some empirical information related to both of the prostitution parameters are available. Because some of the empirical information on British prostitutes is limited, we cannot compute prostitution-adjusted discrepancy ratios directly. Therefore, by necessity, we approach the issue somewhat indirectly. The empirical information on Ivorian prostitutes is very limited. Consequently, our analysis of the Ivorian case is even more informal.
Great Britain
. Blakey and Frankland (16) reported that an HIV prevention outreach worker contacted 404 prostitute women working in Cardiff between 1990 and 1993. The outreach worker’s 3 years of intensive fieldwork focused on all sectors of prostitution. The authors noted that these women were mobile, with some working in Cardiff for brief periods before moving on to other areas (17). If the 404 women included all prostitute women in Cardiff and the population of South Glamorgan County is taken as the denominator, the 3-year prevalence of prostitute women is 98 per 100,000. If this prevalence is adjusted for a level of prostitutes’ mobility the same as in Colorado Springs, the resulting rate is 58 per 100,000.McKeganey et al. (18) conducted a 7-month capture–recapture study based on field intercepts with female street prostitutes in Glasgow in 1991. They estimated that 1,150 prostitute women worked on the streets in Glasgow during the 12 months of 1991. During this period, Glasgow prostitutes had limited mobility, as no more than 15% of sampled prostitutes in two Glasgow studies had worked as prostitutes outside of Glasgow in the last year (19, 20). If the population of the Strathclyde region (21) is taken as the denominator, the annual prevalence of prostitute women is 51 per 100,000. If this prevalence is adjusted for a level of prostitutes’ mobility the same as in Colorado Springs, the resulting rate is 37 per 100,000.
Sharpe (22) found 99 women in police records of a northern English city in 1993 who were known to be working or have worked recently as prostitutes. These women included virtually all known active prostitutes working in the county (K. Sharpe, personal communication). These women also were not very mobile, as only 7 (18%) of 40 prostitutes interviewed had ever worked as prostitutes in another city. With the county population as the denominator, the annual prevalence rate is 11 per 100,000 (23).
McCullagh, Syed, and Bellis (24) conducted a capture–recapture study of prostitute women in Liverpool between 1994 and 1995 based on records from outreach activities, drop-in clinics, and the police. They estimated that 557 prostitute women worked in Liverpool during the 2-year period. The authors did not report any information on prostitutes' mobility. If the Liverpool city population (25, 26) is taken as the denominator, the two-year prevalence is 118 per 100,000. If the Merseyside county population (25, 26) is taken as the denominator, the 2-year prevalence is 39 per 100,000. The latter rate accounts for mobility of Liverpool prostitutes within the county, although it does not account for prostitute women who worked in other Merseyside communities but not Liverpool. However, neither rate accounts for the fact that some Merseyside prostitutes may have worked much or most of the period outside of the county, which suggests that the latter rate may be more accurate.
Einon (27) reported point (1-day) estimates of the prevalence of call girls who advertise with cards in shop windows and telephone boxes in parts of London. However, these point prevalences cannot reasonably be converted to annual prevalences.
In Table 3, we summarize information on British prostitutes’ number of partners. The mean numbers of partners/clients for these studies parallel the mean numbers of partners reported by prostitutes in the United States for the same recall periods, as shown in Table 2. Although these results cannot reasonably be extrapolated to a 1-year period because of the studies’ short recall periods, they seem consistent with the number of partners necessary to eliminate the discrepancy. If the true annual prevalence of prostitutes in Great Britain were equal to the lowest empirical estimate [11 per 100,000 population, from Sharpe’s (22) northern English county], then prostitutes’ mean number of partners in a year would need to be between 524 and 655 to eliminate the discrepancy. This range is based on (1) the assumption that no prostitute women were included in the British survey (27, 28), (2) the unadjusted discrepancy ratios for 1-, 2-, and 5-year recall periods, (3) Wadsworth et al.’s (28) procedures for computing the discrepancy ratio, and (4) 1991 British census figures (29, 30). This range appears to be quite consistent with the numbers of partners shown in Table 3, even with conservative assumptions for extrapolating to a 1-year period. Thus, these results suggest that prostitution can account for the discrepancy in Great Britain.
Einon (27) argued that the prevalence of prostitute women in Great Britain required to eliminate the sex discrepancy in the reported number of lifetime sexual partners was improbably large. She based her conclusion on some of the annual and point prevalence estimates discussed earlier. Such estimates are inadequate for a lifetime recall period, as many different cohorts of prostitute women will have worked during that length of time. Indeed, in Colorado Springs, the cumulative total number of unduplicated full-time equivalent prostitute women between 1970 and 1988 is seven times greater than the number of full-time equivalent prostitute women in 1 year (1988) (1). We would expect similar or larger increases in prevalence from 1 year to longer time periods in Great Britain. We do not compute discrepancy ratios for the lifetime recall period because significant age differences between prostitutes and their clients (as noted in many studies cited here) and lack of sound empirical estimates of the prostitution parameters for previous decades make such calculations problematic.
Table 3. Number of sexual partners reported by prostitute women in Great Britain for different recall periods
Location | Year | Sample sources | n | Mean | Median | Range |
Last/typical week |
|
|
|
|
|
|
Birmingham | 1988 | Courts, outreach to prostitutes working in all major sectors of prostitution (31) | 258 | 22 | — | — |
Glasgow | 1991 | Streets (20) | 68 | 36* | — | — |
Glasgow | 1989–1990 | Drop-in center for street prostitutes (18) | 63 | 35* | — | — |
London | 1985 | Prostitutes attending an STD clinic, representing major sectors of prostitution (32) | 50 | 20 | — | 7–100 |
London | 1989–1991 | Prostitutes representing all major sectors of prostitution recruited from STD clinic, streets, courts, flats, agencies, saunas, telephone, and referrals from prostitutes (33) | 280 | — | 9 | 0–198 |
London | 1992–1993 | Heroin-addicted prostitutes recruited by interviewers who knew prostitutes or had access to them (34) | 51 | 7 | — | 0–30 |
London | 1994 | Prostitutes working out of apartments recruited from STD clinic (35) | 30 | 76 | — | — |
Middlesborough | 1997 | Drop-in center for prostitutes (36) | 33 | 25* | — | — |
Wolverhampton | 1993 | Streets, saunas† | 62 | 17 | 11 | 0–70 |
Typical month |
|
|
|
|
|
|
Glasgow | 1990–1991 | Drug-injecting street prostitutes (37) | 51 | 128 | — | — |
London | 1990–1991 | Drug-injecting prostitutes recruited from streets, pubs, cafes, private homes, and drug treatment clinics (38) | 34 | 56 | 47 | 1–300 |
*Based on multiplying the mean number of clients per night by the mean number of nights/days worked per week.
†Based on unpublished data (H. Kinnell).
Ivory Coast
. Gersovitz et al. (39) reported data indicating an unadjusted sex discrepancy ratio of 2.73 for a 1-year recall period on the assumption that all reported partners were of the opposite sex. They cited an unpublished manuscript by K. Kale for information on prostitute women in the Ivory Coast. In that manuscript, Kale noted that a 1991 census of 6 of the 10 districts of Abidjan (the capital and largest city) found 9,262 full-time professional prostitute women, in addition to an unknown number of part-time and occasional prostitutes. No information is given about how the census was conducted or the prostitutes identified. Kale also mentioned that prostitutes had an average of 12 clients per week and that two-thirds of prostitute women moved every 8 months, but did not describe how these estimates were derived. Without more detailed information about prostitutes, no calculations adjusting for prostitution can be done without making several arbitrary assumptions. We simply observe that if the 9,262 prostitutes enumerated in the 1991 partial Abidjan census represented all the prostitutes in the Ivory Coast, the prevalence rate would be 69 per 100,000 [based on 1993 population figures (see Encyclopedia Brittanica Online at http://www.eb.com:180/bol/topic?eu=125347&sctn=1)]. This prevalence suggests that prostitution quite likely can account for a substantial proportion or perhaps all of the sex discrepancy in reported number of sexual partners in the Ivory Coast as well.References
1. Potterat, J. J., Woodhouse, D. E., Muth, J. B. & Muth, S. Q. (1990) J. Sex Res. 27, 233–243.
2. Decker, J. F. (1979) Prostitution: Regulation and Control (Rothman, Littleton, CO).
3. Rossmo, D. K. & Routledge, R. (1990) J. Quant. Criminol. 6, 293–314.
4. Statistics Canada (1990) Postcensal Annual Estimates of Population for Census Divisions and Census Metropolitan Areas, June 1, 1990 (Component Method) (Catalogue number 91-212, Statistics Canada, Ottawa), Vol. 6.
5. Klovdahl, A. S., Potterat, J. J., Woodhouse, D. E., Muth, J. B., Muth, S. Q. & Darrow, W. W. (1994) Soc. Sci. Med. 38, 78–88.
6. Wollaston, H. A. (1988) M.P.H. thesis (Univ. of Washington, Seattle).
7. Fritz, R. B. (1998) J. Appl. Soc. Psychol. 28, 888–911.
8. Albert, A. E., Warner, D. L., Hatcher, R. A., Trussell, J. & Bennett, C. (1995) Am. J. Public Health 85, 1514–1520.
9. Albert, A. E., Warner, D. L. & Hatcher, R. A. (1998) Am. J. Public Health 88, 643–646.
10. Paone, D., Cooper, H., Alperen, J., Shi, Q. & Des Jarlais, D. C. (1999) AIDS Care 11, 269–280.
11. Kasprzyk, D., Montano, D. E. & Fishbein, M. (1998) J. Appl. Soc. Psychol. 28, 1557–1583.
12. Deren, S., Sanchez, J., Shedlin, M., Davis, W. R., Beardsley, M., Des Jarlais, D. & Miller, K. (1996) AIDS Educ. Prev. 8, 444–456.
13. Seidlin, M., Krasinski, K., Bebenroth, D., Itri, V., Paolino, A. M. & Valentine, F. (1988) J. Acquired Immune Defic. Syndr. 1, 150–154.
14. Darrow, W. W. & the Centers for Disease Control Collaborative Group for the Study of HIV-1 in Selected Women (1990) in AIDS, Drugs, and Prostitution, ed. Plant, M. (Tavistock/Routledge, London), pp. 18–40.
15. Leonard, T., Sacks, J. J., Franks, A. L. & Sikes, R. K. (1988) J. Med. Assoc. Ga. 77, 162–164, 167.
16. Blakey, V. & Frankland, J. (1995) Health Educ. J. 54, 131–142.
17. Blakey, V. & Frankland, J. (1994) HIV Prevention for Women Prostitutes and Their Clients: A Preliminary Report on a Pilot Project in South Glamorgan, Briefing Report No. 5 (Health Promotion Wales, Cardiff, U.K.).
18. McKeganey, N., Barnard, M., Leyland, A., Coote, I. & Follet, E. (1992) Brit. Med. J. 305, 801–804.
19. Green, S. T., Goldberg, D. J., Christie, P. R., Frischer, M., Thompson, A., Carr, S. V. & Taylor, A. (1993) AIDS Care 5, 321–335.
20. McKeganey, N. & Barnard, M. (1992) AIDS Care 4, 395–407.
21. General Register Office for Scotland (1993) 1991 Census Report for Scotland (Her Majesty’s Stationery Office, Edinburgh).
22. Sharpe, K. (1998) Red Light, Blue Light: Prostitutes, Punters, and the Police (Ashgate, Aldershot, U.K.).
23. Central Statistical Office (1995) Regional Trends 30, 198.
24. McCullagh, J., Syed, Q. & Bellis, M. (2000) in Proceedings of the Conference on Changing Perspectives on Female Prostitution, February 1998 (Univ. of Liverpool, Liverpool, U.K.), in press.
25. Office for National Statistics (1996) Regional Trends 31, 211.
26. Office for National Statistics (1997) Regional Trends 32, 166–167.
27. Einon, D. (1994) Ethol. Sociobiol. 15, 131–143.
28. Wadsworth, J., Johnson, A. M., Wellings, K. & Field, J. (1996) J. R. Stat. Soc. A 159, 111–123.
29. Office of Population Censuses and Surveys (1993) 1991 Census: Age, Sex, and Marital Status (Her Majesty’s Stationery Office, London).
30. Office for National Statistics (1999) Population Trends 96, 63.
31. Kinnell, H. (1989) Prostitutes, Their Clients, and Risks of HIV Transmission in Birmingham (Central Birmingham Health Authority, Birmingham, U.K.).
32. Barton, S. E., Underhill, G. S., Gilchrist, C., Jeffries, D. J. & Harris, J. R. (1985) Lancet 8469/70, 1424.
33. Ward, H., Day, S., Mezzone, J., Dunlop, L., Donegan, C., Farrar, S., Whitaker, L., Harris, J. R. & Miller, D. L. (1993) Brit. Med. J. 307, 356–358.
34. Gossop, M., Powis, B., Griffiths, P. & Strang, J. (1994) Addiction 89, 961–970.
35. Whittaker, D. & Hart, G. (1994) Sociol. Health Ill. 18, 399–414.
36. Opaneye, A. A. & Surtees, M. (1998) Int. J. STD/AIDS 9, 245–246.
37. Taylor, A., Frischer, A., McKeganey, N., Goldberg, D., Green, S. & Platt, S. (1993) Addiction 88, 1561–1564.
38. Rhodes, T., Donoghoe, M., Hunter, G. & Stimson, G. V. (1994) AIDS Care 6, 269–276.
39. Gersovits, M., Jacoby, H. G., Dedy, F. S. & Tape, A. (1998) J. Am. Stat. Assoc. 93, 875–863.