Interpartner concordance of self-reported sexual behavior among college dating couples

Journal of Sex Research, Wntr, 1997 by David Wyatt Seal

Unlike many scientifically studied phenomena, most human sexual behavior is neither ethically nor pragmatically amenable to direct observation, given the highly private nature of this behavior. Thus, our understanding of human sexuality is derived primarily from self-reported sexual behavior. This reliance on self-reported behavior has led many critics to suggest that the scientific study of sexuality lacks both reliability and validity. Not surprisingly, then, much research has been focused on factors increasing the reliable assessment of sexual behavior (see Blumstein et al., 1990; Catania, Gibson, Chitwood, & Coates, 1990). Spurred in part by the continuing AIDS epidemic, researchers have made significant progress toward developing rigorous methodology for collecting reliable sexual behavior data (Blumstein et al., 1990; Catania et al., 1990). However, determining whether these data are valid has been much more elusive. To date, few objective markers of validity exist within sex research. Although some researchers suggested that measures of the interpartner concordance of self-reported sexual behavior may serve as an indirect measure of data validity (Clark & Wallin, 1964; Coates et al., 1988; Kinsey, Pomeroy, & Martin, 1948; Kinsey, Pomeroy, Martin, & Gebhard, 1953; Padian, Aral, Vranizan, & Bolan, 1995; Seage, Mayer, Horsburgh, Cai, & Lamb, 1992), other authors cautioned against the interpretation of interpartner concordance as a validity coefficient of self-reported sexual behavior without validation against some type of objective index (e.g., biological markers; Blumstein et al., 1990; Catania et al., 1990). Unfortunately, even "objective" indices such as biological markers are "imprecise, cumbersome, or expensive to apply in large population-based studies" (p. 341; Catania et al., 1990). For example, linking biological markers to non-vaginal intercourse sexual behavior may be difficult, if not impossible. Other biological markers such as population prevalence of HIV, sexually transmitted diseases (STDs), or pregnancy do not necessarily have direct one-on-one correspondence with individually reported sexual behavior.

Although interpartner concordance may be an inadequate marker of validity, Jacobson and Moore (1981) pointed out that if partner "reports of the events in their relationship are considered important in their own right, independent of their accuracy, then validity remains an issue unencumbered by standard reliability considerations" (p. 275). These authors suggested that the independent observations of partners constitute a viable area of study in their own right, as partners "are not passive and objective observers" of behavior, but rather "they are participants with biases, and these biases need to be understood" (p. 276; Jacobson & Moore, 1981; see also Catania et al., 1990; Julien, Bouchard, Gagnon, & Pomerleau, 1992). An understanding of the biases in partner reports of their own sexual behavior has utility for both clinicians and researchers. Methodologically, most dyadic sexual behavior assessment is not directly derived by data collection from both members of dyads. Rather, researchers and clinicians often ask participants to provide proxy information about their partner's sexual behavior and his or her perceptions of the relationship, or they use a single partner's self-reported behavior to represent the relationship. The extent to which these commonly used assessment methodologies are reliable may depend in part on the degree to which a single person's behavioral reports are concordant with his or her partner's perceptions. Within heterosexual couples, society's scripts and expectations for normative heterosexual behavior may differentially bias mens and womens self-reports. Other factors such as partners' sexual satisfaction (Clark & Wallin, 1964; Levinger, 1966), relational distress (Jacobson & Moore, 1981), relational status (married versus unmarried; Bromberg, Maher, Wang, Grijalva, & Cordell, 1993), and awareness of a partner's HIV status (Seage et al., 1992) may also influence interpartner concordance. In sum, the study of interpartner concordance has both empirical and clinical relevance.

Unfortunately, researchers' assessment of the concordance of partner reports of sexual behavior has been limited in scope. Researchers have historically focused on the interpartner concordance of the reported occurrence and/or frequency of sexual intercourse (Blumstein & Schwartz, 1983; Booth & Welch, 1978; Clark & Wallin, 1964; Jacobson & Moore, 1981; Levinger, 1966). More recently, behavioral AIDS researchers have been interested in the interpartner concordance for the use or nonuse of condoms (Bromberg et al., 1993; Miller, Bartholow, Schoenbaum, & Webber, 1993; Padian, 1990; Padian et al., 1995; Upchurch et al., 1991). In general, researchers have found moderate to good interpartner agreement for both dichotomous (e.g., never/always) and continuous (e.g., frequency or percentage of occasions) reports of condom use and penetrative intercourse (i.e., oral, vaginal, and anal) across gay (Blumstein & Schwartz, 1983; Coates et al., 1988; Seage et al., 1992), married and unmarried heterosexual (Blumstein & Schwartz, 1983; Booth & Welch, 1978; Clark & Wallin, 1964; Guimaraes et al., 1993; Jacobson & Moore, 1981; Kinsey et al., 1948, 1953; Levinger, 1966; Miller et al., 1993; Schopper, Doussantousse, & Orav, 1993; Seage et al., 1992; Upchurch et al., 1991), and HIV-affected samples (Bromberg et al., 1993; Padian, 1990; Padian et al., 1995). With the notable exceptions of Kinsey's landmark research (Kinsey et al., 1948, 1953) and Coates et al.'s (1988) project with gay men, few researchers have assessed the interpartner concordance of reported nonpenetrative sexual behaviors. Researchers have also failed to consider interpartner agreement about intercourse-related attitudes and behaviors (e.g., alcohol use, safer sex talk). Finally, past researchers have provided little insight into factors influencing increased versus decreased interpartner concordance for reported sexual behavior.