1.1. The Problem of Low Base Rates of Reproductive Suppression

Assessing the correlates of reproductive suppression is complicated by the problem of low base rates of actual reproductive suppression. According to the National Institute of Mental Health (2002), only about 0.5–3.7% of women experience anorexia nervosa (AN) in their lifetime. A partial solution to this low base rate problem is to identify variables with relatively high base rates of occurrence that are likely precursors of the low base rate variable of interest (e.g., Gottman & Levenson, 1992). In all likelihood, behaviors known to be predictive of excessive reduction in caloric intake (i.e., anorexia nervosa), and also are relatively high base rate behaviors, precede actual reproductive suppression as a result of reduced caloric intake, which is a low base rate occurrence. Using this approach, variables that predict increased disordered eating attitudes (i.e., dieting symptomatology) can be identified, assuming that women who display relatively higher levels of dieting symptomatology would be more likely to ultimately experience reproductive suppression than women who display low levels of dieting symptomatology. This is the approach we adopted in the present research.

1.2. Social Support and Dieting Symptomatology

Guided by the RSH, a number of cues to thin body ideals and disordered eating attitudes have been proposed, such as high levels of female–female social competition Mealy, 1999, Salmon et al., 2004, incongruity between sexual and psychological maturity due to early menarche (Surbey, 1987), value of female labor (Anderson, Crawford, Nadeau, & Lindberg, 1991), negative social consequences of pregnancy (Anderson et al., 1991), and stress, in particular from male sexual attention (Salmon et al., 2004). Stress-induced reproductive suppression has also been discussed by Dunbar (1985). All of these predictors of dieting symptomatology indicate a possible absence of support from others for the rearing of offspring. Primate females, in particular human females, are unique among mammals in requiring extensive social support for reproduction and childrearing Geary, 2000, Trevathan, 1987. When support is perceived to be minimal, women may be hesitant about raising offspring and subsequently engage in behaviors that could serve to temporarily suppress their reproductive capabilities (e.g., Lesk, 1996, Schmidt et al., 1997, Tiller et al., 1997, Wasser & Barash, 1983). The present research assumes that social support was imperative for successful reproduction in ancestral environments and that dieting behavior in the face of cues indicating low support would have been adaptive (e.g., Anderson & Crawford, 1992, Crawford & Salmon, 2002).

1.3. Present Research

The current research explores the association among dieting symptomatology, parental readiness, and perceived social support from partners, family, and friends. This research also assesses potential confounding variables, such as length of romantic relationship, felt stress, and Body Mass Index (BMI), and analyses were conducted to statistically control for scores on these variables. If perceiving low levels of support from partners, family, and friends has, over time, been linked with poor reproductive outcomes and dieting behaviors have ultimately allowed women to temporarily suppress ovulation, then perceptions of relatively low levels of social support should be associated with lower levels of perceived parental readiness (Hypothesis 1) and higher levels of dieting symptomatology (Hypothesis 2). While dieting could reflect an evolved reproductive suppression mechanism, parental readiness reflects individuals' emotional and cognitive awareness about being ready to reproduce at a given time. According to the RSH, dieting behavior is integrally related to a woman's self-evaluation as being unprepared for parenting. Therefore, parental readiness and dieting symptomatology should be negatively correlated, such that women who report not feeling ready to become a parent are expected to also report higher levels of dieting symptomatology (Hypothesis 3). The present research tests these hypotheses.

2.1. Participants

One hundred and two undergraduates at Simon Fraser University who were enrolled in Psychology and Archeology classes and were currently involved in a heterosexual romantic relationship participated in this study in exchange for course credit. Two participants were removed from the study because their age exceeded 35 years. The remaining 100 participants had a mean age of 20.6 years (S.D.=3.28). The average length of romantic relationships was 21.70 months (S.D.=26.08).

2.2. Procedure

Participants were tested in university classrooms, in groups of 10–25 individuals per session. They were given a set of questionnaires assessing demographic information, perceived support from partners, friends, and family, stress experienced in the last month, dieting symptomatology, and parental readiness.

2.3. Materials

Participants were asked to provide information regarding their age, weight, height, and duration and status of their romantic relationship.

Perceived support from partners was measured with the Partner-Specific Investment Inventory (Ellis, 1998). This questionnaire consists of 52 items divided into 10 subscales (e.g., Expressive/Nurturing, Giving of Time, Socially Attentive). The inventory consists of two parts: the first part involves questions regarding the partner's behavior in the last 6 months on a discrete frequency scale from 0=never to 4=very often. The optional response N/A (not applicable) was removed for this study. The second part consists of descriptive statements regarding the partner's behavior in the last 6 months on a five-point Likert scale ranging from 1=strongly disagree to 5=strongly agree. A total score was computed by aggregating z scores for the subscales. The internal consistency for this scale was very high (Cronbach's α=.85).

Support from family and friends was assessed with the Perceived Social Support Scale (Procidano & Heller, 1983), which measures the extent to which respondents perceive that their needs for support, information, and feedback are fulfilled by friends and family. Responses are restricted to yes (1), no (0), and don't know (0) choices. The internal consistency was high for the scale assessing support from family (Cronbach's α=.93) and also for the scale assessing support from friends (Cronbach's α=.86).

Felt stress was measured with the 14-item Global Measure of Perceived Stress (Cohen, Kamarack, & Mermelstein, 1983). This scale assesses the level of life stress, daily hassles, and availability of coping resources experienced in the last month. Responses are measured on a frequency scale ranging from 0=never to 4=very often. The internal consistency for this scale was very high (Cronbach's α=.89).

Dieting symptomatology was measured with three subscales from the Eating Disorders Inventory (EDI; Garner, 1990), including nine items from the Body Dissatisfaction subscale, seven items from the Drive for Thinness subscale, and eight items from the Maturity Fears subscale. Scores were averaged across the three subscales of the EDI to create an index of dieting symptomatology. The EDI is a widely used self-report measure of symptoms commonly associated with AN and bulimia nervosa, with higher scores reflecting the increased presence of disordered eating attitudes as well as eating pathology. Items are on a discrete frequency scale ranging from 0=never to 6=always. The internal consistency for this scale was very high (Cronbach's α=.83).

The Parental Readiness Questionnaire is an unpublished questionnaire designed by Ward (2000) and consists of an eight-item Guttman scale that measures a woman's current physical, emotional, and instrumental readiness for pregnancy and childrearing (see Appendix A for the items that comprise this scale). The internal consistency for this scale was moderate (Cronbach's α=.65).

4.1. Alternative Explanations and Caveats

It could be argued that women in this study, who perceived less available support and more stress in their lives, were anxious about not being in a situation that stereotypically precedes parenthood and were concerned about their future and their likelihood of finding a good partner. The results may therefore reflect a general anxiety about not having a committed partner and good family relations, with increased dieting symptomatology possibly reflecting a bargaining tactic to elicit greater support from family and friends (e.g., Hagen, 2002). It is possible though that such a general anxiety may represent a proximate expression of the reproductive suppression mechanism, but the current research was not designed to test this possibility.

This research examined dieting symptomatology or a set of attitudes and behaviors with a relatively high base rate as a proxy for reproductive suppression, an occurrence with a relatively low base rate. Future research needs to assess the links between perceptions of support, dieting symptomatology, and actual reproductive suppression. Furthermore, reproductive suppression through means other than reducing caloric intake should be assessed in women that perceive less available support (e.g., use of birth control, sexual activity, putting children up for adoption, excessive exercise).

This research also utilized a correlational design and therefore directional relationships between the variables cannot be inferred from the results. Longitudinal research that tracks women over time could provide more definitive evidence regarding the direction of these relationships. Additionally, all measures reflect self-reports and the results may thus reflect a suite of integrated perceptions that are not necessarily accurate reflections of reality.

Whereas the present research focused on factors that may trigger dieting symptomatology in women, recent theoretical work has suggested that the distinctive symptoms of AN, such as restricting food and hyperactivity, are evolved psychological adaptations that facilitated the movement away from resource poor environments (Guisinger, 2003). This model, however, does not specifically address factors related to the onset of AN. Future research that focuses on factors that precede the onset of AN as well as factors that emerge after the onset of these symptoms would certainly help forge a greater understanding of the evolution of disordered eating attitudes in women.