In industrialized populations, the average woman ovulates roughly 300 times in her lifetime. In traditional natural fertility populations, women may ovulate fewer than 100 times in a lifetime on average; they are pregnant or breastfeeding during much of their reproductive years (Strassmann, 1999). Only during a few days leading to ovulation can copulation result in conception and, therefore, pregnancy (e.g., Dunson et al., 2001). In total, these days amount to a small proportion of the average lifespan of a woman in a natural fertility population. Every human ever born was conceived during these very brief windows, successfully implanted in the uterine endometrium, and survived gestation.

Not surprisingly, then, natural selection has shaped adaptations that foster successful ovulation, conception, implantation, and gestation. The reproductive cycle entails substantial physiological changes, the timing of which are orchestrated by ovarian hormones, estrogens and progesterone. Psychological and behavioral adjustments likely accompany these physiological changes, further enhancing the prospects of successful reproduction. During the few days when copulation could result in conception, women's sexual interests often heighten (e.g., Hill, 1988; Regan, 1996; Roney and Simmons, 2013). Other competing motivations, such as appetite, diminish (e.g., Roney and Simmons, 2017; Gorczyca et al., 2016). Unlike most mammalian females, however, women have sex even during non-conceptive phases, which likely serves functions other than conception, such as strengthening the pair-bond and enhancing the likelihood a partner will invest in offspring (e.g., Strassmann, 1981; Righetti et al., 2020; Gangestad et al., 2021). These non-conceptive periods are much more prevalent than the narrow conception window. It is therefore plausible that women have psychological adaptations that have specific functions during non-conceptive periods. These non-conceptive phases are not simply a “switching off” of the psychological responses to conceptive status. Behavior is a critical component of reproduction, so estrogens and progesterone are involved not only in physiological changes, but also psychological adjustments that support goals relevant to both the conceptive and non-conceptive phases of the menstrual cycle.

Along these lines, Reynolds et al. (2018) investigated associations between women's progesterone levels across the cycle and their concerns about their levels of social support. Ancestrally, women's ability to successfully raise a child would have been enhanced by support from other people – their primary partners but also, perhaps, members of broader social networks of potential alloparents. Reynolds et al. (2018) proposed that, post-ovulation, and extending into early pregnancy, women should become attuned to and worry about the levels of social support they receive. Progesterone, which increases around the time of implantation and early pregnancy, may produce these effects. Two studies (N = 100 and 61) examining associations of progesterone levels with measures of anxiety in naturally ovulating women yielded qualified support for this proposal. In both studies, women's average progesterone across sessions positively covaried with their anxiety levels. In Study 2, within-woman variations in progesterone levels also predicted scores on a measure specifically tapping attachment anxiety (e.g., “I am afraid to lose my partner's love”). In Study 1, however, within-woman associations of progesterone levels covaried weakly and non-significantly with a single-item rating of anxiety level (“I feel anxious today”). In sum, these studies offered consistent evidence for an association between women's average progesterone levels and their anxiety (i.e., between-subjects effect). Evidence pertinent to whether women's level of anxiety tracks their progesterone levels, as they change across the cycle (i.e., within-subjects effect), was inconsistent. However, the measure of anxiety in the study was limited to a single item, and one not specifically targeted at women's worries about interpersonal support. This is insufficient to reject the hypothesis that progesterone drives within-cycle changes in anxiety about potential interpersonal support.

Reynolds et al. (2018) were careful to note limitations of their studies and findings. As they discussed, results could not speak to the direction of causality. Although progesterone may promote vigilance to levels of interpersonal support, it is also possible that stress-inducing experiences, perhaps especially ones that are interpersonal in nature, may lead to increased levels of progesterone (see Dinh et al., 2021). The latter possibility is perhaps most consistent with weak or variable associations of anxiety with within-woman progesterone levels. Progesterone secreted during stressful experiences is produced in the adrenal glands, where it is a biosynthetic precursor to cortisol (e.g., Herrera et al., 2016). Yet, the vast majority of within-woman variation is almost certainly of ovarian origin—progesterone levels massively increase post-ovulation.1 Within-woman variations in progesterone produced during stressful experiences may be difficult to detect against this background of massive cyclic variation. Still, additional empirical work is needed before firm conclusions about true empirical patterns or their likely causal explanation can be drawn.

Hahn et al. (2020) performed a partial conceptual replication of Reynolds et al.'s work. They specifically examined hormonal associations across the cycles of naturally cycling women with anxious jealousy, a construct likely related to anxious attachment (sample item, “I worry that my partner might leave me for someone else”). They found no evidence for a within-woman association between progesterone levels and anxious jealousy (see also Stern et al., 2023, who found no within-woman association between progesterone levels and a general measure of jealousy; neither study reported associations of these outcome measures with women's average levels of progesterone). Though jealousy likely relates to concerns about partner involvement, social support, and more general anxiety, it pertains to a particular component of it.

Shifts in mood across the menstrual cycle have been of long-standing interest to researchers, especially in the context of premenstrual symptoms. The premenstrual phase (the last several days of the cycle) is not associated with the highest progesterone levels; indeed, during this phase, progesterone levels diminish. In any event, reviews of the literature offer no clear conclusion about how negative or positive mood shifts across phases (e.g., Romans et al., 2012; Stern et al., 2025). One recent study found a descriptively (and non-significant) negative within-woman association between progesterone and negative affect (Stern et al., 2025, Study 2). The study did not report associations with women's average progesterone levels.

In the current study, we further investigated associations between progesterone and women's anxieties about social support and their causes. In light of Reynolds et al.'s findings, we were interested in examining the robustness of both between-women (cycle average) and within-women associations of progesterone with worries about social support and negative emotionality. Though these associations are correlational in nature, their patterns can speak to underlying causal scenarios, to the extent that different causal scenarios expect different patterns of association (e.g., Grosz et al., 2020). Once again, if progesterone facilitates an adjustment in women's attunement to their social support, both between-women and within-woman associations are to be expected. If it exists, a robust association of average progesterone levels and worries about partner involvement or negative emotionality, even in the absence of within-woman associations, is an empirical observation that demands explanation. With the current sample, we were able to perform additional unique tests to probe the causal scenario at play. Associations were tested in a sample of 213 naturally ovulating women in heterosexual romantic relationships, of whom 180 participated in multiple lab sessions conducted across a month. Of the 180 women attending multiple lab sessions, 135 attended four sessions, 15 attended 3 sessions, and 30 attended 2 sessions, resulting in 645 total observations. During each session, women completed a questionnaire that included (a) a 9-item assessment of women's positive and negative emotional states over the past two days (e.g., “distressed,” “nervous,” “happy,” “sad”), and (b) a 3-item measure specifically targeting women's worries about their partners' levels of investment and involvement (e.g., “I worried about my partner's level of involvement in our relationship”). We assayed estradiol, progesterone, cortisol, and testosterone (or their metabolites) in urine collected the day of each session. We hence were able to assess women's average and within-cycle associations between hormone levels and psychological states, including both general negative affect and worries about partner investment specifically. In subsequent analyses, we examined whether effects held after accounting for individual differences, including neuroticism and perceived family support, as these individual differences could potentially drive associations between women's average progesterone levels and outcome variables. In addition, based on the results of initial analyses, we separately examined associations with mean progesterone levels during the follicular and luteal phases, as the source of the former is largely adrenal whereas the source of the latter is largely ovarian (e.g., Herrera et al., 2016). Especially as adrenal progesterone may be secreted in response to stress (Herrera et al., 2016), differential associations could help illuminate potential causal scenarios.