The study arrives at a moment of sustained demographic interest in why birth rates keep falling in high-income countries. After hovering around 2.0 children per woman through the 1990s and early 2000s, the U.S. total fertility rate peaked at 2.12 in 2007, then declined steadily in the wake of the Great Recession to reach 1.62 in 2023. Similar patterns of delayed and reduced childbearing have been documented across Europe and East Asia, making the United States part of a broad international trend rather than an outlier. Earlier work suggests that this decline cannot be explained simply by women wanting fewer children overall. Instead, researchers have pointed to postponement of childbearing and a shrinking likelihood of achieving the fertility goals people form early in life — a dynamic that matters biologically because the prospects for conception and healthy pregnancy decline with age. The new paper asks whether a third ingredient — uncertainty — helps complete the picture.

The authors draw a conceptual distinction between three dimensions of uncertainty in fertility goals. The first, "goal uncertainty," refers to difficulty forming a goal at all — the classic "don't know" response to questions about whether one intends to have children. The second, "realization uncertainty," concerns whether people who do report positive intentions believe those intentions will actually be achieved. The third, "intensity of goals," is not strictly uncertainty but a related attitude: how strongly a stated intention is held, measured here by whether a woman would be bothered if she ended up childless. The authors argue that these dimensions, though analytically separable, may travel together and jointly shape fertility behavior. A woman who doubts she can afford a child, for example, may simultaneously report an intention, express low confidence in achieving it, and care less about the outcome — a bundle of attitudes that standard projections, which typically count only the intention itself, cannot distinguish.

To measure these dimensions, the team used the National Survey of Family Growth (NSFG), a nationally representative repeated cross-sectional survey that has tracked fertility attitudes and behaviors in the United States since 1973 and is administered by the National Center for Health Statistics. The analysis pooled data releases from 2002 through 2018, yielding a weighted sample of 41,492 women aged 15 to 44, with representative estimates for 2002, 2008, 2012, 2014, 2016, and 2018. Because each survey round draws a fresh sample rather than following the same people, the design captures period trends across the population rather than individual change over time. The NSFG asks respondents stepwise questions: first whether they desire another child, then whether they intend one, then — for those who intend — how sure they are they will actually have a baby, with answers ranging from "very sure" to "not at all sure." Childless women were also asked whether it would bother them a great deal, some, a little, or not at all if they had no children. Because the survey's male questionnaire does not ask unpartnered men about intention certainty, the analysis was restricted to women.

The first headline result concerns goal uncertainty, which turned out to be rare. Only between roughly 1.5 and 2.1 percent of women answered "don't know" when asked whether they intended more children, and that proportion was stable across the study period. Notably, the NSFG does not offer "don't know" or softer categories like "probably yes" as pre-coded options, so respondents must volunteer hesitation on their own; this figure likely represents a lower bound on true goal uncertainty. The proportion of women who did not provide uncertainty answers varied little by age, parental status, partnership status, education, or income, even as the underlying proportions of "yes" and "no" intenders differed sharply across these groups. In other words, American women overwhelmingly manage to form an opinion about whether they want more children — the interesting variation lies in what happens after that opinion is formed.

Realization uncertainty, by contrast, was pervasive. Among women who intended at least one more child, up to 50 percent were not "very sure" they would follow through, including about 7 percent who said they were "not at all sure." The aggregate level was broadly stable over time, but subgroup patterns diverged in revealing ways. Childless women were more certain of realizing their intentions than mothers — about two-thirds versus one-half reporting they were very sure — but their certainty declined over the period, from 60 percent in 2002 to 52 percent in 2018, while mothers' certainty rose from 46 to 50 percent. The authors suggest this closing gap may reflect a growing sense among young, childless women that the conventional path to parenthood has become harder to complete. Younger women were more certain than women aged 30 to 44, of whom more than half expressed uncertainty and up to 17 percent reported being not at all sure in 2008. Partnered women were slightly more certain than unpartnered women, consistent with the practical advantages a stable partner provides for pursuing parenthood.

Socioeconomic status also mattered. Women with at least a bachelor's degree and those with household incomes at or above 200 percent of the poverty line reported higher certainty about realizing their intentions. Yet even this advantage appeared to be eroding: among college-educated women intending more children, the share who were "very sure" fell from 65 percent in 2014 to 54 percent in 2018, while the share "not at all sure" climbed back to 9 percent — matching the peak reached during the Great Recession in 2008. The authors suggest that generalized economic uncertainty may now be reaching even relatively advantaged groups. The post-2014 period encompasses not only continued labor-market precarity but also rising housing costs, student debt burdens, and the political turbulence of the late 2010s, any of which could shake confidence in long-term planning even among women whose incomes alone would seem to secure it.

Intensity told a more dynamic story. Among childless women who intended children, the share reporting weak intentions — saying they would be only a little bothered, or not bothered at all, if they never had a child — rose from 18 percent in 2002 to 26 percent in 2018, a 44 percent increase. The rise was driven almost entirely by young women aged 15 to 29, among whom weak intensity climbed from 17 to 25 percent. Low intensity was also more common among women with less than a bachelor's degree (28 percent in 2018, versus 20 percent for the college-educated) and among women below 200 percent of the poverty line (28 percent, versus 24 percent above it). The pattern echoes recent findings that childless American men are increasingly reporting disinterest in fatherhood, and it aligns with survey evidence that younger cohorts increasingly view parenthood as one acceptable life path among many rather than a default stage of adulthood.

The study then connected these dimensions to the specifics of fertility plans: how many additional children women intended, and when they intended to have the next one. Using regression models controlling for age, partnership status, race and ethnicity, education, employment, income, residence, religiosity, and survey year, the authors found that certainty and intensity were both associated with larger intended family sizes and tighter timelines. Childless women who were "very sure" of realizing their intentions reported intending 2.4 children on average, compared with 2.1 for the "somewhat sure" and 1.9 for the "not at all sure." Those who would be bothered by not having children intended 2.3 children, versus 2.1 for those who would not. The predicted probability of intending a next child within two years was 19 percent for the very sure, 11 percent for the somewhat sure, and 9 percent for the not at all sure, and 16 percent versus 13 percent for intense versus weak intentions. These are associations, not causal estimates, but they suggest that uncertainty and weak commitment are bundled with postponement and smaller intended families — and postponement, in demographic terms, frequently converts into forgone births.

The implications reach beyond measurement. Short-term intentions and intentions with concrete time frames are known to be more predictive of actual childbearing, and quantum intentions underpin U.S. fertility projections produced by statistical agencies and used for planning schools, housing, and social insurance. If high levels of realization uncertainty and low intensity persist even while headline intentions remain stable, aggregate fertility could continue to decline simply because intentions are never carried out. The findings also fit a broader "macro-micro" framework in which economic instability — from the Great Recession onward — makes it harder to plan for children, generating doubt about imagined futures that translates into postponed or abandoned childbearing. The authors note that other uncertainties, including worries about discrimination related to gender identity and sexual orientation, restricted access to reproductive health care, and political and civic disruption, may contribute to what they call "institutionalized uncertainty" surrounding childbearing decisions — a condition in which doubt about the future becomes a normal, expected feature of family formation rather than an exception.

The authors are candid about limitations. They rely on secondary survey data with fixed question wording, which shapes how uncertainty can be measured; the NSFG offers no direct "don't know" option for intentions, so goal uncertainty is likely underestimated. The cross-sectional design cannot track individuals over the life course or establish causal direction — it is possible that women planning children sooner simply feel more certain, rather than certainty producing sooner plans. The NSFG provides no measures of psychological traits or outlooks, no publicly available contextual economic indicators, and no information on male partners' goals, and state-level estimates are not possible. Future work, they argue, will need longitudinal data linking economic conditions, perceived uncertainty, and fertility goals to subsequent behavior.

The backdrop to the research is one of the defining demographic shifts of our era: the systematic postponement of parenthood. Infertility is medically defined as the failure to conceive or maintain a pregnancy after twelve months or more of regular unprotected intercourse, and its global lifetime prevalence is estimated at 17.5 percent, according to a 2022 systematic review and meta-analysis. Because the biological capacity to reproduce declines with age—steeply for women and, to a lesser extent, for men—every year of delay raises the share of couples who eventually confront fertility problems. In Australia, where the study was set, an estimated 16 percent of heterosexual couples report difficulties conceiving. The timing of first births has transformed accordingly: in 1991, roughly a quarter of Australian women had their first child at age 30 or older; by 2020, more than half did. Meanwhile the two-child ideal has barely budged over two decades, even as completed family size has slid steadily downward—a widening gap between intended and achieved fertility that postponed and thwarted conception helps to explain.

Demographers have long known that fertility preferences are not fixed traits. Economic insecurity, career changes, new partnerships, separations and health shocks have all been shown to nudge childbearing desires and expectations up or down. Yet the uncertainty generated by infertility itself—a condition that is at once chronic, often persisting for months or years, and unplanned, arriving without warning—has remained largely uncharted territory. Earlier cross-sectional studies offered tantalising clues: among American women, those who identified as infertile reported lower fertility intentions but higher desires and a larger ideal family size, and infertile women attending gynaecology clinics in China expressed bigger ideal families alongside shakier confidence in achieving them. Because such single-moment snapshots cannot establish which came first, the altered preferences or the infertility, the causal story stayed murky. The new study is among the first to follow men as well as women through the transition into infertility and watch, statistically, what happens to their plans.

To do so, Lazzari turned to the Household, Income and Labour Dynamics in Australia (HILDA) Survey, a nationally representative panel launched in 2001 with more than 13,000 participants and wave-on-wave response rates above 90 percent. The analysis hinges on two waves separated by four years: 2015, designated Time 1, and 2019, designated Time 2. Respondents aged 18 to 45 for women and 18 to 50 for men answered two questions on 11-point scales: “How do you feel about having a child (more children) in the future?”, capturing desire, and “How likely are you to have a child (more children) in the future?”, capturing expectation, with the expectation scale running from very likely at zero to very unlikely at ten. Crucially, HILDA poses these questions to everyone of reproductive age, not merely to those who conceive easily—an inclusive design that sidesteps the routine exclusion of infertile respondents from fertility research, under the mistaken assumption that they lack fertility preferences. Infertility status was derived from a single question: “Based on medical advice, do you know of any physical or health reason that would make it difficult for you (and/or your partner) to have children?”

The methodological architecture is deliberately conservative. The cross-sectional sample numbered 6,520 respondents after the exclusion of cases with missing values, while the longitudinal core comprised 3,702 people who were fertile at baseline and traced to follow-up four years later; those already reporting infertility in 2015 were removed so that the analysis isolates the transition into the condition. Between the waves, 11.3 percent of women and 5.5 percent of men made that transition. Because infertility is not randomly assigned—people with strong fertility desires are more likely to forgo contraception and attempt conception, and therefore more likely to encounter infertility—Lazzari employed the regressor variable method, an ordinary least squares technique that predicts preferences at Time 2 while statistically holding constant each person’s baseline preference at Time 1. The models further adjusted for events unfolding between waves, including the birth of a child, changes in relationship status and shifts in financial satisfaction, together with age, education, parity, migration and Indigenous status, and area of residence. A supplementary analysis of 777 heterosexual co-resident couples incorporated partners’ own preferences, ages and education, since childbearing decisions are typically negotiated between two people rather than made by one.

The snapshot results echoed earlier work with new precision. Among infertile men, 42 percent registered a desire score of zero and 45 percent an expectation score of zero, compared with 27 and 25 percent of their fertile counterparts; just under 40 percent of infertile women expressed the lowest possible expectation of having a child, against 23 percent of fertile women. Yet once age, parity, partnership status, education and finances were taken into account, a paradox surfaced: infertility was actually associated with higher fertility desires among women, and with lower expectations among both sexes. The expectation gap was no artefact of infertile respondents simply being older or more often parents—it persisted, statistically significant, after all observable differences were accounted for, in line with the study’s cross-sectional hypotheses.

The longitudinal analysis delivered the headline finding. Net of all controls, transitioning into infertility between 2015 and 2019 predicted a significant downward revision of fertility expectations for both men and women, confirming the hypothesis that expectations bend under the weight of the condition. Desires, by contrast, did not move significantly in either direction. Simple descriptive comparisons told the same story: by 2019, both groups had tempered their outlooks, but the downward adjustment in expectations was markedly larger among those who had experienced infertility, and for women the difference between infertile and never-infertile respondents was statistically significant in formal tests. A companion model restricted to co-resident couples, which additionally modelled each partner’s own wishes and relationship satisfaction, reproduced the pattern almost exactly: infertility reshaped expectations within couples, but left desires standing.

Two of the study’s predictions failed in illuminating ways. Lazzari had hypothesised that people of lower socioeconomic status, facing greater barriers to assisted reproductive technology, would revise their expectations downward more sharply, and that women, who the literature says bear the greater emotional toll, would show stronger effects than men. Neither held: education and financial satisfaction left the infertility–expectation link essentially unchanged, and interaction models revealed no gender-specific effects. The author points to Australia’s policy environment as a partial explanation: the national insurer Medicare subsidises assisted reproduction without restrictions on age, parity or the number of subsidised treatments, and general practitioners serve as gatekeepers to fertility clinics—circumstances that may blunt socioeconomic gradients, even though disparities in access persist and women remain about three times more likely than men to consult a doctor about fertility concerns. An alternative reading is that men who are aware of their infertility already constitute a self-selected, highly engaged group, compressing observable gender differences.

The dissociation between expectations and desires slots neatly into modern theories of fertility decision-making. In the traits–desires–intentions framework, desires represent an idealised vision of childbearing, relatively insulated from situational constraints, whereas intentions and expectations rationally incorporate perceived barriers and perceived behavioural control, as the Theory of Planned Behaviour also maintains. Cognitive models and the Theory of Conjunctural Action add that decisions are shaped as much by emotional associations and mental schemas as by explicit calculation. The Narrative Framework goes furthest: heightened uncertainty may darken expectations without extinguishing the imagined future self as a parent. Infertility is uniquely ambiguous on this score, since a medical diagnosis implies reduced odds of conception rather than impossibility—natural conceptions after unsuccessful assisted reproduction are far from rare—and because it destabilises identity, leaving people suspended between a parenthood they cannot assume and a childlessness they will not embrace, a state sociologists have described as a “spoiled identity.” Desires, in this reading, encode a deep psychological commitment to parenthood that survives even biological setbacks.

The implications cut two ways. For survey researchers the message is unambiguous: failing to measure infertility risks misreading the declining expectations of affected respondents as irrationality or unstable reproductive planning rather than a rational response to uncertainty. People may continue actively pursuing a child even as their stated expectations fall, deviating from what expectation-based forecasts alone would predict—an insight with consequences for projections of population fertility as parenthood is delayed further. The study also carries caveats: infertility was self-assessed and conditional on having received medical advice, likely undercounting disadvantaged and remote populations who consult doctors less often; two measurement points cannot capture self-perceptions that fluctuate over time; and HILDA records no information on assisted reproduction treatment, whose failures can darken expectations independently of the underlying condition. Still, the central lesson stands. Fertility expectations, the paper concludes, are more likely to be adjusted downward in the face of infertility, while desires tend to remain mostly unaffected—a quiet testament to how tenaciously the imagined future survives an uncertain present.