Sociocultural Correlates of Permissive Sexual Attitudes: A Test of Reiss's Hypotheses About Sweden and the United States - Statistical Data Included

Journal of Sex Research, Feb, 2000 by Martin S. Weinberg, Ilsa Lottes, Frances M. Shaver

The Swedish data were collected at a comparable university in Sweden (a public university that was large in size but not located in a metropolis nor considered to be socially or politically atypical). The questionnaire was initially translated and then checked and reworked. The distribution procedure was similar to the one in the U.S. except that the initial target population included students between the ages of 20 and 23. This difference in initial age strata was necessary because most 19 year old males in Sweden fulfill a year of compulsory military service. Thus, Swedish men are a year older than U.S. men when they begin their university education. The Swedish group included 736 students with a gender distribution of 65% women and 35% men--parallel to the group in the U.S. The distribution with respect to religious affiliation was 63% Church of Sweden (Evangelical-Lutheran), 6% Swedish Free Church, 1% Catholic, 4% other religion, and 26% no religion. The response rate was 52%. This is comparable to the response rate of 59% in the 1996 national interview and questionnaire study of Swedish adults described by Helmius (1998).

To eliminate a possible bias due to the age differences of students in the two countries, only students in the 20 to 22 year old age interval from both universities were included in the analyses for this paper. This included 570 Swedish students (359 women, 211 men) and 407 U.S. students (278 women, 129 men). Their mean ages were 21.0, 21.1, 21.0, and 21.1 respectively. For the U.S., 2% were married and 3% were in cohabiting relationships. For Sweden, no one was married and 9% were in cohabiting relationships.

Limitations

There were two major limitations in the data collection. One was that the study was restricted to students at a single university within each country. Second, there was the high rate of nonresponse (common in studies using mailed questionnaires). These limitations were considered in delineating the probability level required for statistical significance (cf. the analysis section). Questions regarding generalizability are addressed in the conclusion.

Measures

Because in some cases we were measuring new concepts defined by Reiss, we could not use established measures. Further, although several concepts could be considered multidimensional, we used only a few items when measuring the concepts due to constraints on the length of the questionnaire. As a consequence--since scales usually need more items to achieve a high alpha--some of our Cronbach alphas were in the low or middle range. In these cases, the scale items were used individually rather than in a composite form.(2)

Permissive sexual attitudes. To measure the dependent variable sexual permissiveness, we used six items. Respondents answered on a 5-point Likert continuum (fully approve, approve somewhat, tolerant toward, disapprove somewhat, fully disapprove). The items were: "15-year-old girls engaging in sexual intercourse with a steady boyfriend," "15-year-old boys engaging in sexual intercourse with a steady girlfriend," "15-year-old girls engaging in sexual intercourse with a person they're not significantly involved with," "15-year-old boys engaging in sexual intercourse with a person they're not significantly involved with," "18-year-old women having 10 or more sex partners in a year," "18- year-old men having 10 or more sex partners in a year." The items in a gendered pair were not placed in juxtaposition to one another. These topics covered the major subjects denoted by the traditional double standard: male privilege with regard to sex at a young age, sex outside a significant relationship, and number of partners. The age chosen for the items was based on the 15-year-old age of consent in Sweden and the result of pretests (testing 15- to 18-year-old age specifications) investigating what age produced the greatest variation in responses. Similar pretesting was also the basis for designing the specifics in the number of partners item. The average of these six items was used as a composite measure (Cronbach alpha = .91).