Researchers have used both aggregate and individual-level approaches to study unrealised fertility, but in both cases, the concept is typically estimated using indirect techniques. Rather than simply asking people if they have fewer children than they would like, an approach advocated for by Yeatman and Smith-Greenaway (2025), researchers most often compare achieved fertility with fertility goals. Although aggregate approaches often yield estimates of unrealised fertility that are easier to understand and compare across contexts, these approaches fail to account for individual experiences and changes in fertility goals over the life course (Morgan and Rackin, 2010; Quesnel-Vallée and Morgan, 2003). To better understand the meaning and implications of unrealised fertility for individuals, greater conceptual clarity is needed about what it means to have a “fertility goal”, including how these goals are interpreted by survey respondents and articulated in demographic research. For comparison, I include one aggregate indicator of unrealised fertility (that uses fertility ideals) but focus the remainder of my analysis on measures of fertility goals at the individual level.

Due to wide-ranging approaches to the measurement of unrealised fertility, studies find a broad range of social factors associated with the phenomenon. Regional comparisons of indicators using ideal family size and desire for a(nother) child at the end of the reproductive career suggest that the prevalence is highest in sub-Saharan Africa and is lowest in South Asia (Casterline and Han, 2017). In Europe, the gap between intended and actual fertility varies across regions and educational levels, with the widest gap being observed among women who are highly educated or live in southern Europe or a German-speaking country and the smallest gap being found among women in central and eastern Europe (Beaujouan and Berghammer, 2019). Educational attainment and union status are also predictors of unrealised fertility: higher-educated women and those who delay marriage generally face higher risks of underachieving the fertility desires they reported at younger ages (Berrington and Pattaro, 2014; Nitsche and Hayford, 2020). In the U.S., rates differ across race and ethnicity; Hispanic women are less likely than non-Hispanic White women to achieve the fertility intentions they reported in early adulthood (Hartnett, 2014).

Beyond the social correlates of unrealised fertility, researchers have also considered the role of reproductive health factors, primarily parity and age at first birth. When fertility goals are compared at the end of the reproductive career, women who have more children overall and had their first birth at a younger age are less likely to have unrealised fertility (estimated by comparing ideal versus actual fertility and current desire for a(nother) child) (Casterline and Han, 2017). Recent studies have considered the role of infertility and other reproductive health conditions. By ages 42–50, men and women in Europe and Latin America who had ever experienced medical infertility had a higher likelihood of falling short of their family size ideals compared to individuals without a history of infertility (Lazzari and Beaujouan, 2026). In another study, researchers compared the gap between desires and intentions as a measure of unrealised fertility, finding that among women who desired a(nother) child, those with diagnosed reproductive health conditions (e.g. ovulation problems, endometriosis, uterine fibroids) were less likely to report an intention to have a child compared to women without a history of reproductive health issues (Guzzo and Broussard, 2026).

Reproductive health factors may directly contribute to unrealised fertility by limiting people’s capacities to attain their goals (e.g. due to infertility or pregnancy loss), or they may indirectly contribute to unrealised fertility by shaping people’s fertility goals. For instance, although women who have experienced infertility report a greater desire to have a baby and a higher ideal number of children, they tend to have lower fertility intentions (Shreffler et al., 2016). The contribution of such factors to unrealised fertility, however, remains underexplored.

Data for this analysis come from three cycles of the NSFG, a cross-sectional, nationally representative survey of men and women aged 15–49 in the U. S., which are publicly available from the National Center for Health Statistics. The analytical sample includes women from the 1995 cycle and from the 2015–2017 and 2017–2019 cycles. For all analyses, I use sample weights provided by the NSFG that account for survey design and nonresponse to calculate nationally representative estimates for 1995 and the midpoint of each two-year cycle (2016 and 2018, respectively). For the multivariable logistic regression models, observations with missing data are excluded (<1.3% of the sample), resulting in a final sample of 2715 respondents.

I investigate the relationship between unrealised fertility and a range of social and reproductive health characteristics. Weighted summary statistics for the variables included in the regression models are presented in Table 1.

I first compare the prevalence of unrealised fertility among women in the U.S. using four distinct approaches to measurement. Table 2 presents each measurement approach and the associated survey questions used to create each indicator.

To measure unrealised fertility using survey-reported fertility ideals, I use a question that the NSFG last asked of respondents in 1995: “If you yourself could choose exactly the number of children to have in your whole life, how many would you choose?” In the absence of longitudinal data, I have constructed a synthetic cohort to compare the ideal fertility of women aged 20–24 in 1995 (n = 1518) to the completed fertility of women from the same birth cohort (n = 1920) in the pooled 2015–2017 and 2017–2019 cycles, when women in the first group would have reached ages 40–48 (Table 2, Indicator A). Synthetic cohorts have been used elsewhere to analyse the correspondence between fertility goals and outcomes, including with NSFG data (Guzzo and Hayford, 2023; Harknett and Hartnett, 2014). Although this represents the only aggregate approach employed in my comparative analysis, it offers insight into how measures using ideal fertility – particularly ideal fertility that is documented early in the reproductive life course – may fail to account for updated or revised preferences, resulting in higher estimates of unrealised fertility.

The remaining indicators of unrealised fertility consider fertility goals and attitudes at the individual level among women approaching the end of their reproductive careers (aged 40–49). Indicator B examines the proportion of women aged 40–49 who desire a(nother) child. Nonsterile women were asked: “Looking to the future, do you, yourself, want to have a(nother) baby at some time?” Meanwhile, sterile women were asked a more hypothetical question: “If it were possible, would you, yourself, want to have a(nother) baby at some time in the future?” (emphasis added). Indicator C investigates the proportion of women who intended to have a(nother) child by the end of their reproductive career. Finally, Indicator D offers a novel and more direct estimation of unrealised fertility by documenting the proportion of women aged 40–49 who anticipated having fertility regret, a term I use to refer to women’s reports that they would be bothered some or a great deal if they did not have any (additional) children. Women who reported that they would be bothered only a little or not at all are coded as not anticipating feeling fertility regret.

Dependent variables for the multivariable logistic regression models include fertility desires, intentions and anticipated fertility regret (Indicators B, C and D in Table 2). The proportion of the sample with unrealised fertility according to Indicator A are presented in Table 1. Because this indicator uses an aggregate approach, individual-level unrealised fertility cannot be identified. I instead present the proportion of the synthetic cohort who (a) reported an ideal family size of at least one child in early adulthood but ultimately had no children, and (b) reported an ideal family size of two or more children in early adulthood but ultimately had one child or no children. Similarly, because individual-level unrealised fertility cannot be calculated, Indicator A is not included in the regression models. For all models, respondents who said that they do not know their fertility desires or intentions or whether they anticipate having fertility regret are coded as zero, and respondents who refused to answer are excluded from analysis. This coding strategy ensures that the estimates of unrealised fertility are conservative.

I assess the relationship between each measure of unrealised fertility and two sets of independent variables: social characteristics and reproductive health characteristics. Social characteristics include race and ethnicity (non-Hispanic White, non-Hispanic Black, Hispanic or other race), educational attainment (less than high school, high school degree or some college), household income (coded as one if within 200% of the federal poverty level and as zero otherwise), relationship status (coded as one if married or cohabiting and as zero otherwise) and religiosity (coded as zero if the respondent reported no current religious affiliation, that religion is not important in her daily life or that she doesn’t know know/refuses; coded as one if the respondent indicated that religion is somewhat important in her daily life; and coded as two if the respondent reported that religion is very important in her daily life).

I consider how reproductive health characteristics may contribute to unrealised fertility by including several variables available in the NSFG: number of live births, previous diagnosis or treatment of a reproductive health condition (endometriosis, pelvic inflammatory disease, uterine fibroids or an ovulation or menstruation problem), sterility status and recent pregnancy loss (past five years). As both sterile and nonsterile women were asked about their fertility desires, sterility status (coded as zero if not sterile, as one if surgically sterile and as two if non-surgically sterile) is included in the models predicting the desire for a(nother) child. Women who were sterile were not asked about their fertility intentions and are thus coded as having no intentions to have a(nother) child. A sensitivity test that excluded sterile women from this analysis yielded similar results.

Individuals may fall short of their childbearing goals for a range of reasons, but unrealised fertility due to infertility or miscarriage is qualitatively different from unrealised fertility that results from a voluntary revision of goals or the prioritisation of competing life ambitions, such as education or career. ¹ These divergent pathways to unrealised fertility likely have distinct emotional, social and policy implications, yet many existing studies treat them as analytically equivalent or do not attempt to distinguish between them.

To better reflect the complexity of trajectories that may result in unrealised fertility, I employ an analytical framework (Figure 1) that considers social, reproductive health and psychological determinants. I draw from established theories of fertility behaviour, including the Theory of Planned Behaviour (Ajzen & Klobas, 2013) and Miller’s Traits–Desires–Intentions–Behaviour framework (Miller, 2011). Both of these approaches consider how traits, social characteristics and life experiences influence fertility orientations, preferences and goals. Demographers have primarily focused on the social determinants of unrealised fertility (e.g. relationship status, educational attainment, financial well-being) (Beaujouan and Berghammer, 2019; Brauner-Otto and Geist, 2018; Guzzo, 2022; Morgan and Rackin, 2010; Nitsche and Hayford, 2020). Although a woman’s existing number of children is routinely included in analyses, research that considers how other reproductive health factors and experiences may contribute to unrealised fertility remains limited. Using a reproductive careers approach, which draws attention to how people “experience their reproductive lives as embodied and interconnected throughout the life course” (Johnson et al., 2023), I consider how both social and reproductive health characteristics and experiences may both shape individuals’ fertility goals and constrain their ability to achieve them.

In my analysis, I incorporate standard sociodemographic predictors and parity, but I also include markers of reproductive health that are often overlooked in demographic analyses, including recent miscarriage, sterility status and diagnoses of reproductive health conditions.

Fertility goals and orientations towards childbearing are key components of unrealised fertility, as without knowing what people want, we cannot observe whether they fall short of their goals. However, fertility goals are themselves informed by social and reproductive health factors. Figure 1 illustrates how each set of factors may directly contribute to unrealised fertility by constraining the ability to become a parent (see solid “direct effect” lines in Figure 1). For example, a person may desire three children but only have two due to social constraints, such as union dissolution, or reproductive health constraints, such as miscarriage or infertility. These factors may also indirectly contribute to unrealised fertility by reducing people’s desires and intentions to have children. This possibility is depicted by the dotted lines in Figure 1 that travel from the social and reproductive health factors through fertility goals. One example of this scenario is that of a woman who desires three children but who, in response to the physical and emotional effects of repeated miscarriages, adjusts her desired or intended fertility downwards to just two children. This experience may be described using the concept of relinquished intentions or desires (White and McQuillan, 2006).

Reproductive health factors may also be moderated by social factors. This process is illuminated by the concept of weathering, which describes the early deterioration of health experienced by structurally marginalised groups “as a consequence of the cumulative impact of repeated experience with social, economic, or political exclusion” (Geronimus, 2001). Meso- and macro-level social factors, such as structural racism and exposure to environmental stressors, may lead to increased rates of adverse reproductive outcomes among marginalised groups. For example, miscarriage rates are higher among women who are Black, work night shifts and are exposed to pollution and pesticides (Quenby et al., 2021). This pathway to unrealised fertility is represented in Figure 1 by the dotted line leading from social factors to unrealised fertility via reproductive health factors. Integrating these elements into the measurement of unrealised fertility allows for a more comprehensive and differentiated understanding of why individuals may not realise their fertility goals and how these experiences intersect with broader patterns of social stratification and inequality.

I first compare the four measurement indicators described in Table 2, showing how the estimated prevalence of unrealised fertility varies significantly across measures. Drawing from the analytical framework presented in Figure 1, I use multivariable logistic regression models to explore the social and reproductive health correlates of unrealised fertility across three separate measures among women aged 40–49: fertility desires, fertility intentions and fertility regret (Indicators B, C and D in Table 2).

The four approaches to measurement yield distinct rates of prevalence among women approaching the end of their reproductive years in the U.S. (Table 2). Indicator A reveals that while 97% of the sample of women in 1995 (aged 20–24) reported an ideal family size of at least one child, by 2015–2019 (when women in this cohort had reached ages 40–48), only 83.5% of the sample had at least one child. Thus, 13.5% of these women fell short of their ideal family size. Similarly, 89.2% of the sample reported an ideal family size of two or more children in early adulthood, but only 66.8% had two or more children by their forties, a difference of 22.4%. This is the highest overall estimate of unrealised fertility across the four indicators.

Indicator B, which categorises women aged 40–49 who desired a(nother) baby as having unrealised fertility, shows a prevalence rate of 16%. Because both sterile and nonsterile women were asked about their fertility desires, Table 2 presents the proportion for each group separately. Sterile women (85% of whom were surgically sterile) were more likely to report a desire for a(nother) baby (18.8% versus 13.4% of nonsterile women). As described elsewhere, this is likely due to selection: women who are sterile often already have children or may have discovered their sterility through prior attempts to achieve a desired pregnancy (Guzzo and Broussard, 2026; Shreffler et al., 2016). Alternatively, this may reflect differences in the question wording rather than differences in the true orientation towards future childbearing. While nonsterile women reported their actual desire for a(nother) child, sterile women reported a hypothetical desire: namely, whether they would want to have a(nother) baby if it were possible (emphasis added).

Indicator C examines the fertility intentions of women aged 40–49. This indicator yields the lowest estimated prevalence of unrealised fertility, with just 3.5% of women reporting that they intended to have a(nother) baby. Among those who reported a desire for a(nother) baby, less than a quarter (20.8%) said that they intended to have a(nother) child, providing evidence that desires and intentions, even at the end of the reproductive life course, are perceived as distinct concepts among survey respondents. This discrepancy suggests that fertility intentions are likely a poor indicator of lifetime unrealised fertility since many women who want or would want to have a(nother) baby may have already relinquished their intentions earlier in the life course.

Indicator D, a measure of anticipated fertility regret, asked how women how they would feel if they ultimately did not have any (additional) children. Although only 3.5% of women aged 40–49 said they intended to have a(nother) baby, more than twice as many (8.1%) said they would be bothered some or a great deal if they did not have any (more) children. This proportion is larger among women who reported a desire for a(nother) baby (30.5%) and among women who reported an intention to have a(nother) baby (42.2%). Surprisingly, however, the majority of women who reported having a positive desire or intention said they would be bothered only a little or not at all if they did not have a(nother) baby.

Table 3 presents the predicted odds of three indicators of unrealised fertility: fertility desires, fertility intentions and anticipated fertility regret at the end of the reproductive career. Models 1, 3 and 5 explore the association of each indicator with social characteristics. The second, nested models (Models 2, 4 and 6) explore the reproductive health correlates of each indicator of unrealised fertility.

The likelihood of desiring a(nother) baby among women aged 40–49 did not differ by social characteristics, except in the case of religiosity; respondents who deemed religion to be “very important” in their daily life had 52% higher odds of desiring a(nother) baby (Model 1). This relationship increased in magnitude after the addition of reproductive health factors in Model 2. Household income became significant in the full model: when controlling for reproductive health factors, women with a household income below the poverty level had 58% higher odds of desiring a(nother) baby near the end of their reproductive career. After adding reproductive health covariates independently (models not shown), the number of live births was found to be the primary driver of this shift in significance.

Three reproductive health characteristics were associated with fertility desires, including number of live births, recent experience of pregnancy loss and sterility status (Model 2). Respondents who had received diagnoses or treatment for reproductive health conditions had similar odds of desiring a(nother) baby compared to those with no conditions. For each additional live birth, a woman had 30% lower odds of desiring a(nother) baby. Those who had experienced pregnancy loss within the past five years had 3.2 times higher odds of desiring a(nother) baby. Compared to nonsterile women, surgically and non-surgically sterile women had significantly higher odds (74% and 58% higher, respectively) of reporting that they would want a(nother) baby if it were possible.

Fertility intentions at the end of the reproductive life course varied across several social characteristics, including race, education and relationship status (Model 3), but not household poverty status or religiosity, as in Model 1. Black (versus White) women had more than twice the odds of intending to have a(nother) baby. Compared to those who had not completed high school, women who had completed high school or college had 3.5 to 4.5 times higher odds, respectively, of having positive fertility intentions (p < 0.05). Women who were married or cohabiting (versus neither) had approximately half the odds of intending to have a(nother) baby (OR = 0.53, p < 0.05). Unlike fertility desires, intentions to have more children did not vary according to household income.

The relationship between reproductive health characteristics and fertility intentions was similar to that of fertility desires: each additional birth was associated with 53% lower odds of intending to have a(nother) baby (p < 0.001), and women who experienced pregnancy loss in the past five years had 2.4 times higher odds of intending to have a(nother) baby (p < 0.01). Although reproductive health conditions were not associated with fertility desires in Model 2, respondents who had received a past diagnosis or treatment for a condition had 45% lower odds of intending to have a(nother) baby (p < 0.05) (Model 4). ² The relationship between two social characteristics (education and relationship status) and fertility intentions was no longer significant after the inclusion of reproductive health characteristics in the model. Again, I evaluated reproductive health covariates independently and found that the inclusion of the number of live births in the model was the cause for this shift in significance. Higher odds of intending to have a(nother) baby among Black (versus White women) persisted and increased in magnitude following the inclusion of reproductive health characteristics (Model 4).

In the final set of models, I explore the social and reproductive health factors associated with women’s anticipated fertility regret. Although Black and Hispanic women both reported higher odds of intending to have a(nother) baby (with only the odds ratio for Black women reaching statistical significance in Model 3), Black and Hispanic women were about half as likely as White women to say they would feel bothered if they did not have a(nother) baby. Respondents with a household income below the federal poverty limit had 64% higher odds of anticipating feeling fertility regret, while those who were married or cohabiting had 36% lower odds. Religiosity was also associated with higher odds of anticipating experiencing regret. Relative to those for whom religion was not important, respondents who deemed religion to be somewhat or very important were nearly twice as likely to anticipate feeling regret (somewhat important: OR = 1.62, p < 0.05; very important: OR = 2.00, p < 0.01).

Model 6 presents the odds of reporting fertility regret with the inclusion of social and reproductive health characteristics. For each additional live birth, respondents had 55% lower odds of anticipating feeling fertility regret. As with previous indicators (fertility desires and intentions), the inclusion of live births in the model was the primary driver of the loss in significance of social characteristics. In this case, race and relationship status were no longer significantly associated with fertility regret, but associations of fertility regret with household income and religiosity persisted and were magnified following the inclusion of reproductive health characteristics. As in previous models, a recent pregnancy loss was associated with higher odds of anticipating experiencing regret (OR = 2.53, p < 0.01). Although women who were sterile (both surgically and non-surgically) had higher odds of desiring a(nother) baby (Model 2), they were not more likely to say that they would be bothered if they did not have a(nother).

This study demonstrates that the prevalence and the correlates of unrealised fertility vary widely depending on how the concept is measured. Across four indicators – fertility ideals, desires, intentions and anticipated regret – estimates ranged from 3.5% to 22.4% among U.S. women aged 40–49. These discrepancies highlight the conceptual and methodological complexity of unrealised fertility and demonstrate the need for greater precision and standardisation in future survey research.

As expected, the aggregate-cohort indicator, which measured the correspondence of fertility ideals (reported earlier in the life course) to completed fertility, produced higher estimates of unrealised fertility (between 13.5% and 22.4%). Because this measure fails to account for potential revisions in individual fertility goals over the life course, a better estimate would compare completed fertility to the ideal family size reported at the end of the reproductive life course (cf. Assaf and Davis, 2021; Casterline and Han, 2017; Lazzari and Beaujouan, 2026; Yeatman and Smith-Greenaway, 2025). Unfortunately, the NSFG dropped this question after 1995. Nevertheless, this indirect estimate is informative in that it appears to better reflect an upper bound of unrealised fertility than other indicators.

Indicators B and C (desires and intentions to have a(nother) child, respectively) examined women’s fertility goals at the end of their reproductive careers. In terms of estimating unrealised fertility, fertility desires were closer to fertility ideals, while intentions to have a(nother) child were significantly lower, thus providing a lower bound estimate of unrealised fertility. This low estimate likely obscured the relinquishment of previously held fertility goals. Specifically, women who were sterile were not asked this question, and were subsequently recoded in the NSFG as having no intentions to have a(nother) baby. Sterilisation remains a leading form of contraception in the U.S., but nearly a quarter of women who are sterilised express a desire for reversal, and sterilisation regret is particularly high among women who undergo the procedure at age 30 or younger (Eeckhaut and Sweeney, 2018). Thus, this is a potentially important oversight. Nevertheless, Indicator C may provide a useful benchmark for comparing changes in the proportion of women striving to become pregnant at older ages, particularly as women continue to delay reproduction until later in the life course.

Indicator D represents the anticipated emotional consequences of unrealised fertility and thus reflects the most person-centred indicator in the study. As a measure of unrealised fertility, it differs from other indicators by directly identifying the proportion of the population for whom not having a(nother) child is perceived as consequential. As with other fertility goals, this measure at the end of the reproductive life course may not fully capture fertility regret that occurred earlier in life, but since the estimated prevalence of this indicator (8%) falls between that of the intentions (3.5%) and desires (16%) indicators, it is likely capturing a wider range of the population with unrealised fertility for whom it will have prolonged emotional impacts.

Consistent with prior demographic research, I found that unrealised fertility in the U.S. is socially patterned, but the significance and the direction of the association differed across indicators. The analytical framework presented here describes how social and reproductive health factors may independently or jointly influence whether fertility goals are met, revised or relinquished. At the end of the reproductive life course, lower-income and religious women were more likely to desire a(nother) baby and anticipate feeling fertility regret, even after controlling for reproductive health factors, which illustrates that these indicators are embedded within moral and socioeconomic contexts. Compared to White women, Black women were more likely to intend to have a(nother) baby at the end of their reproductive career even after accounting for reproductive health, which suggests that unmeasured social factors, such as cultural norms or structural conditions, may sustain higher fertility aspirations during a period in the life course when pregnancy is less attainable.

Although researchers have extensively studied the role of education in shaping fertility goals and the achievement of desired fertility (Barber, 2001; Beaujouan and Berghammer, 2019; Berrington and Pattaro, 2014; Guzzo and Hayford, 2024), the women in this sample had similar odds of reporting a desire for a(nother) child and anticipating feeling fertility regret across educational levels. Higher-educated women had greater odds of intending to have a(nother) child at the end of their reproductive career, but these effects disappeared with the inclusion of achieved parity, which suggests that delayed childbearing, rather than unmeasured sociocultural characteristics, was the primary driver of unrealised fertility among this group.

Two reproductive health factors proved to be strong and consistent correlates of unrealised fertility. Each additional birth reduced the odds of experiencing unrealised fertility across all measures, and recent pregnancy loss increased those odds two- to threefold. Reproductive health diagnoses were not associated with fertility desires but were negatively associated with intentions at the end of the reproductive career. This distinction supports recent arguments that desires and intentions are shaped by different sets of constraints (Badolato, 2025), and that people with health challenges may still want to have children but feel unable to act on that desire due to biological constraints (Badolato, 2025; Guzzo and Broussard, 2026; Lazzari, 2025; Lazzari and Beaujouan, 2025). The findings presented in this study suggest that fertility desires may better reflect the unrealised fertility of people with reproductive health conditions since they may no longer hold positive fertility intentions at the end of their reproductive career. However, a recent study showed that people who experience infertility may also adjust their ideal family size downwards over time (Lazzari and Beaujouan, 2026), highlighting the limitations of any measure of unrealised fertility that operationalises survey-reported goals. The analytical framework used in this study can motivate future research investigating how a broader range of reproductive health factors – which may themselves be shaped by social and structural conditions – influence fertility goals as well as achieved fertility.

Sterility status provides an instructive example of how survey design may affect estimates. The alternative question wording (“if it were possible…”) may have elicited responses closer to the women’s fertility ideals than their prospective desires, which demonstrates the need to harmonise question phrasing and clarify the concept of unrealised fertility (Table 2). Women who were sterile had higher odds of desiring a(nother) baby but were automatically recoded by the NSFG as having no fertility intentions. In the U.S., nearly a quarter of surgically sterile women express a desire to have their procedures reversed (Eeckhaut and Sweeney, 2018), which suggests that some women in this group may still have positive childbearing intentions, and thus may not be accounted for in estimates of unrealised fertility that rely on intentions-based measures. Because socioeconomically disadvantaged women and women of colour are more likely to both undergo sterilisation and desire a sterilisation reversal (Johnsen and Sweeney, 2022), the inclusion of sterile individuals in measures of unrealised fertility is critical for providing a clearer picture of who fails to achieve their fertility goals and why.

The wide-ranging estimates of unrealised fertility observed in this study have implications for future demographic measurement and policy. As debates over fertility decline intensify, statistics about the “fertility gap” may be invoked to justify pronatalist interventions, including those that may limit access to contraception and abortion (Forati and Bartz, 2024; Gietel-Basten et al., 2022). Because the present study reveals strong variation in the prevalence and distribution of unrealised fertility across measures, public claims about fertility shortfalls should be subject to greater scrutiny. Indicators based on early-life ideals may overestimate unrealised fertility, while those grounded in current desires, intentions and anticipated regret may better reflect the lived realities and consequences of unrealised fertility but fail to capture the experiences of individuals who may have relinquished their fertility goals earlier in life. Refining these indicators should involve efforts to document people’s reproductive histories, current preferences and the broad range of factors, both social and reproductive, that inhibit people from achieving the families they desire.

Direct, retrospective measures of fertility regret are especially promising indicators of unrealised fertility, yet such questions are rarely asked in population-based surveys (for notable exceptions, see Greil et al., 2022 and Yeatman and Anderson, 2025). The question in the NSFG is unique, as it prospectively asks respondents how they would feel if they did not have any (additional) children. Results from the full model are largely consistent with patterns found in the model predicting fertility desires, which suggests that these indicators may be measuring similar constructs and orientations towards parenthood. Future studies should investigate not only retrospective accounts of fertility regret, but also the consequences, both material and emotional, of unrealised fertility. For example, how does unrealised fertility affect women’s relationships, their mental health and the care and support they receive at older ages? Women who experience reproductive health constraints, such as infertility and pregnancy loss, report significant levels of psychosocial distress (McQuillan et al., 2003; Quenby et al., 2021). In contrast, little is known about the consequences of unrealised fertility, particularly when it is induced by social constraints.

In addition to the reproductive characteristics analysed in this study, future research should explore how other reproductive experiences contribute to unrealised fertility, including the physical, financial and emotional challenges associated with infertility, IVF, pregnancy, childbirth and the postpartum period. These studies could investigate the relationship between specific events or conditions (e.g. birth trauma, medical interventions like Caesarean section, hyperemesis gravidarum, postpartum depression) and fertility goals, fertility outcomes and unrealised fertility.

This study is subject to several limitations. First, the four indicators developed for comparative analysis were limited to the specific variables available within the NSFG. Consequently, these measures do not reflect the full range of indicators that have been used to estimate unrealised fertility. Second, the results reflect the fertility goals and outcomes of a specific cohort of U.S. women and may not be generalisable to men, gender-diverse individuals or populations in different cohorts or social contexts. Nevertheless, by explicitly comparing four indicators within a single nationally representative dataset, the study results demonstrate significant heterogeneity in estimates of unrealised fertility and provide a replicable framework for comparative studies in other contexts.

Second, the reproductive health determinants of unrealised fertility should be interpreted with caution, as infertility and pregnancy loss are often underreported in surveys (Lindberg and Scott, 2018). Similarly, as many reproductive health conditions go unrecognised or undiagnosed, knowledge of the impact of reproductive health factors on desired and achieved fertility is limited. The reproductive health characteristics analysed in the present study reflect those of women aged 40–49 and may not have been present earlier in the life course when most women initiate childbearing. Future surveys and novel data collection efforts (Nobles et al., 2026) that track fertility goals prospectively and document a broad range of social and reproductive health experiences would allow for a more accurate assessment of how various factors influence the achievement and revision of fertility goals.