Do Local Sex Ratios Approximate Subjective Partner Markets? Evidence from the German Family Panel. Andreas Filser & Richard Preetz. Human Nature, Jun 19 2021. https://link.springer.com/article/10.1007/s12110-021-09397-6
Abstract: Sex ratios have widely been recognized as an important link between demographic contexts and behavior because changes in the ratio shift sex-specific bargaining power in the partner market. Implicitly, the literature considers individual partner market experiences to be a function of local sex ratios. However, empirical evidence on the correspondence between subjective partner availability and local sex ratios is lacking so far. In this paper, we analyzed how closely a set of different local sex ratio measures correlates with subjective partner market experiences. Linking a longitudinal German survey to population data for different entities (states, counties, municipalities), we used multilevel logistic regression models to explore associations between singles’ subjective partner market experiences and various operationalizations of local sex ratios. Results suggest that local sex ratios correlated only weakly with subjective partner market experiences. Adult sex ratios based on broad age brackets, including those for lower-level entities, did not significantly predict whether individuals predominantly met individuals of their own sex. More fine-grained, age-specific sex ratios prove to be better predictors of subjective partner market experiences, in particular when age hypergamy patterns were incorporated. Nevertheless, the respective associations were only significant for selected measures. In a complementary analysis, we illustrate the validity of the subjective indicator as a predictor of relationship formation. In sum, our results suggest that subjective partner availability is not adequately represented by the broad adult sex ratio measures that are frequently used in the literature. Future research should be careful not to equate local sex ratios and conscious partner market experiences.
Discussion
Imbalanced sex ratios have been linked to a wide range of social consequences, including family formation, economic decision-making, gender roles, partnership formation, fertility, personality, and sexuality (Bauer & Kneip, 2013; Feingold, 2011; Griskevicius et al., 2012; Harknett, 2008; Merli & Hertog, 2010; Pollet & Nettle, 2008; Schacht & Smith, 2017; Trent & South, 2012; Uggla & Mace, 2017). Sociodemographic and evolutionary mating market approaches have explained these findings by shifts in bargaining power based on differential mating opportunities for men and women in an imbalanced sex ratio environment (Filser & Schnettler, 2019; Guttentag & Secord, 1983; Kokko & Jennions, 2008; Pedersen, 1991; Schacht & Kramer, 2016). Individual partner market experiences might play a crucial role in these behavioral adaptations to local sex ratios given that subjective experiences provide critical guidelines for human behavior (Gilbert et al., 2016; Gintis, 2006; Kroneberg & Kalter, 2012). Yet, empirical evidence has been lacking on how closely individual experiences of partner market opportunities correspond to sex ratios of their local environment. To fill this gap, we analyzed associations between a variety of local sex ratio measures and subjective partner market experiences of female and male singles in a German panel survey.
In sum, the expected association between subjective partner market experiences and local sex ratios only held for selected, age-specific sex ratio measures. In particular, adult sex ratios based on broad age ranges as are commonly used in the literature did not prove to be significant predictors of subjective partner market experiences. This result was consistent across operationalizations of adult sex ratios as the proportion of men in the adult population (PMA) based on different age brackets at the level of states, counties, and municipalities. None of the adult sex ratio variants correlated with either men’s or women’s subjective experiences of surplus encounters with individuals of their own sex in a meaningful way. More granular, age-specific sex ratio measures (ASPM) that include only individuals of adjacent age cohorts were closer approximations of subjective partner market experiences. In particular, age-specific measures that also incorporated age shifts to reflect age hypergamy patterns proved to be better predictors of subjective partner market experiences. Nevertheless, only selected state-level, age-shifted sex ratios correlated with women’s surplus encounters with other women in a statistically significant way. The corresponding county-level age-shifted sex ratios yielded similar, yet smaller coefficients, which have to be interpreted with caution given that they did not reach statistical significance. For men, only county-level, age-shifted sex ratios significantly predicted associations with men’s subjective partner market experiences. Coefficients for state-level age-shifted sex ratios were similar in size but did not reach statistical significance. Overall, some reservations regarding the state-level findings seem warranted because the German states might be too large in geographic terms (with all but four being larger than 15,000 km2) to be considered a single partner market. Lengerer (2001:142) reports that 85% of future partners in Germany live within a 20 km radius of each other. Recent publications suggest that earlier recommendations to rely on smaller entities when operationalizing local partner markets continue to apply in the age of Internet dating (Bruch & Newman, 2019; Fossett & Kiecolt, 1991). Therefore, results for state-level sex ratios should be treated with caution.
In sum, the results of this study suggest that previous findings regarding the social consequences of imbalanced sex ratios are unlikely to be mediated by conscious adaptations to partner scarcities or oversupplies. Adult sex ratios for fixed age brackets, such as the population aged 16–49 or 16–64, constitute the standard operationalization of local sex ratios in the literature (see Schacht et al., 2014; Pollet et al., 2017 for reviews). Our findings suggest that sex ratios for fixed adult age ranges are unlikely to correspond closely to subjective partner market experiences. Previous research has demonstrated that sex ratios correlate only moderately with each other when different age cutoffs are used (see Fossett & Kiecolt, 1991 for a discussion using US census data). Therefore, adult sex ratios are unlikely to be a well-suited summary measure of age-specific sex ratios. This is also supported by our dissimilar results for adult and age-specific sex ratios. In contrast to adult ratios, selected age-specific and age-shifted operationalizations significantly predicted subjective partner market experiences. In particular, the integration of age hypergamy into the sex ratio measures yielded significant results for predicting subjective partner market experiences. Future research should therefore consider focusing on age-specific, age-shifted sex ratio measures. Yet, although age-shifted sex ratios predicted men’s subjective partner market experiences, we only find weak evidence for a similar association for women. This difference between men and women might be due to a smaller sample size of women in our models. A further explanation could be related to sex differences in sexual strategies guiding partner market behavior. In particular, sexual strategies theory suggests that sexual selection favored antagonistic mating competition and preferences for multiple short-term mating in men (Buss, 1999; Schmitt, 2015; Trivers, 1972). This could also entail that men are more aware of marriage squeezes than women are.
A further finding of this paper is that subjective surpluses of same-sex encounters significantly predicted relationship formation. For both sexes, a subjective surplus of encounters with individuals of one’s own sex was significantly associated with a lower probability of entering a relationship. We are aware that survey questions on subjective partner market experiences may represent an excessive demand for respondents. However, the fact that the subjective indicator correlates with this specific partner market outcome supports the idea that the analyzed reports of surplus encounters with same-sex individuals constituted a valid approximation of individual partner market experiences. Concerning the local sex ratio measures, age-specific and age-shifted variants proved to be advantageous over adult sex ratios also when predicting relationship formation. None of the adult sex ratios significantly predicted relationship formation. Moreover, age-specific local sex ratios only yielded significant coefficients when incorporating age shifts. Specifically, relationship formation for women was significantly predicted by state- and county-level age-specific and age-shifted sex ratios. Yet, the probability of men entering a relationship was not predicted by local sex ratios, replicating similar asymmetric findings by Uggla and Mace (2017). With regard to the link to subjective partner market experiences, our findings suggest that subjective partner market experiences and local sex ratios should be considered distinct context variables rather than equivalent indicators. This is even true for detailed measures of local sex ratios. For instance, age-specific county-level sex ratios with a two-year age shift were a significant predictor of women’s relationship formation. Yet, we do not find conclusive evidence that these measures were correlated with women’s subjective partner market experiences. Consequently, these findings suggest that subjective and local sex ratios are not interchangeable operationalizations. Rather, they appear to be two separate dimensions of partner market circumstances. Researchers should be aware of this distinction when offering theoretical interpretations of results based on local sex ratios.
The subjective partner market indicator used in this study is not equivalent to the situational perception of the sex proportions in a group. Instead, it approximated the everyday interactions of individuals and therefore should not be interpreted as indicative of an inability to perceive sex ratios in set groups. Both Alt et al. (2017) and Neuhoff (2017) demonstrated that participants are able to give accurate sex ratio estimations based on short-term exposure to visual and auditory cues. Against the backdrop of these previous studies, one potential explanation for our findings could be that individual partner market experiences are not a direct representation of macro-structural conditions, i.e., local sex ratios (Blau, 1977; Rapp et al., 2015; Schwartz, 1990). Instead, individual partner markets may be structured in different “foci of activity,” such as workplaces, voluntary associations, or hangouts (Feld, 1981; Rapp et al., 2015). With this in mind, studying the consequences of sex ratios in interactive spheres such as workplaces (Åberg, 2009; Barclay, 2013; Svarer, 2007), industries (Uggla & Andersson, 2018), bars (Lycett & Dunbar, 2000), or colleges (Harknett & Cranney, 2017) would have the advantage of assuming that the individuals are actually interacting with one another. This is much more plausible than the same contention would be for local sex ratios. Consequently, individuals’ foci-specific sex ratios might give a more accurate impression of partner supply and demand within the respective foci rather than sex ratios of the local population, even for their specific age cohort.
This paper used a combination of administrative population information and survey data, which is crucial to this analysis. Studies relying on such data face a trade-off between the scope of the data and the ability to link survey data with survey-based partner market measures. The pairfam survey data constitute a unique combination of both ends of this spectrum. However, adult sex ratios in Germany may not have sufficient variation to allow for identifying a clear effect. This is particularly true for adult sex ratios at the state level, which only range between 96 and 108 men per 100 women (see Table 2). Consequently, nonsignificant findings for state-level sex ratios could also be due to the lack of variation at this level of aggregation. Internationally, local adult sex ratios may vary more substantially in selected regions, most notably in the male-skewed populations of China and India (Guilmoto, 2012). However, the county-level variation in adult sex ratios in the analyzed data was consistent with that of recent studies from other Western countries (e.g., Schacht & Kramer, 2016), and the ranges of age-specific sex ratios exceeded the ranges of adult sex ratios in our data.
A further limitation is that the findings are contingent on the validity of the subjective partner market indicator. While our complementary analysis demonstrated the predictive validity of the subjective indicator with respect to relationship formation, limitations persist. The directional verbalization of the indicator question introduced ambiguity, resulting in imprecise measurement of undecided and disagreeing answers. Specifically, respondents who met an equal number of men and women either might have reported disagreeing with the statement of predominantly meeting individuals of their own sex or might have given an undecided answer to express their experience of a balanced sex ratio. We explored this issue via fitting linear and multinomial models for different variants of the original indicator scale. These auxiliary results confirmed that the difference in probabilities for undecided and disagreeing answers was not significantly correlated to local sex ratios. However, agreement with the surplus same-sex contacts scale was related to selected local sex ratio measures (Fig. S6-S8, in the ESM). We therefore focused on the dichotomized indicator that summarized disagreeing and undecided responses. Nevertheless, our logistic regression results do not persist when taking a linear modeling approach, most likely because of measurement noise in disagreeing and undecided responses. Furthermore, the current analysis was limited to one global subjective indicator of opposite-sex encounters. A detailed survey of foci-specific sex ratios might give a closer approximation of subjective partner market experiences (cf. Rapp et al., 2015). This could reveal whether partner markets in specific foci actually correspond to local sex ratios, whereas partner markets in other foci do not. In particular, detailed information on job location could be of particular relevance, given that 60% of German employees cross municipality borders when commuting (Pütz, 2017). Consequently, adding sex ratios based on the place of work could yield a higher correspondence to subjective partner markets.
In conclusion, the sex ratio literature should be cautious regarding the assumption that individuals are consciously aware of local sex ratio skews. In particular, subjective and conscious partner market experiences do not appear to be a direct function of broad-range adult sex ratios but instead are correlated only with selected, age-specific measures. Researchers should consider this when interpreting findings based on local sex ratios. Although our findings shed some doubt on a direct link between conscious experiences and local sex ratios, this does not necessarily imply that local sex ratios do not capture partner markets. So far, very little is understood about how humans experience, remember, and process contextual sex ratios (Dillon et al., 2017). In particular, the relative importance of immediate interaction partners, local communities, and broader social contexts is yet to be explored (Maner & Ackerman, 2020).
This paper explored the relationship between a general indicator of subjective partner market experiences and local sex ratio measures. In sum, general sex ratio measures that are based on broad age ranges do not seem to capture conscious partner market experiences in a meaningful way. Future research will have to establish the role of unconscious factors, including endocrinal or network effects mediating contextual local sex ratios and adaptations in individual behavior.