Attachment Bonds and Baby Booms: A Cross‐Cultural Study of Reproductive Success

ABSTRACT

The goal of the current study was to explore attachment theory and reproductive success across the three cultural contexts of Japan, Canada and the USA. We investigate how adult attachment styles (secure, fearful, preoccupied and dismissive) are associated with reproductive success, as measured by the number of children. Using data collected from 15,120 participants via an online survey, we found fearful and preoccupied attachment styles were positively associated with reproductive success across all three samples, indicating that these insecure attachment styles may promote behaviours linked to larger family sizes. Conversely, secure attachment negatively predicted reproductive success in Canada and the USA but showed no significant relationship in Japan. Sex differences in attachment were observed, with women scoring higher in preoccupied attachment across cultures and men scoring higher in fearful attachment. We propose attachment styles are associated with reproductive strategies, with cultural norms moderating these associations. Collectivist values in Japan may attenuate individual attachment associations with family planning, while individualistic cultures such as Canada and the USA align secure attachment with deliberate, resource‐intensive reproductive strategies. This study advances the understanding of attachment theory by discussing its relevance to life history strategies and reproductive outcomes within diverse cultural settings.

Attachment theory, first articulated by Bowlby (1969), emphasises the formative influence of early caregiver interactions on the development of ‘working models’ of the self and others. They serve as cognitive and emotional frameworks that guide interpersonal behaviours and relational expectations throughout life. Building on Bowlby's work, Ainsworth et al. (1978) identified three primary attachment patterns in infants, which are secure, anxious‐resistant, and avoidant. These classifications have been integral to understanding how attachment processes evolve across the lifespan and influence outcomes across domains such as emotional regulation, relationships, and mental health (Cassidy and Shaver 2018).

Extending attachment to adult relationships, Hazan and Shaver (1987) demonstrated that patterns observed in childhood persist into adulthood, significantly influencing romantic partnerships. This continuity is explained through the operation of internal working models, which affect relational expectations, communication styles, and conflict resolution (Fraley and Shaver 2000). Bartholomew and Horowitz (1991) expanded the framework by distinguishing adult attachment styles along the two dimensions of positive or negative models of the self and others. Their classification includes secure, fearful, preoccupied and dismissing attachment styles. Research consistently links secure attachment to stable and satisfying relationships, while insecure attachment styles are associated with relational anxiety, avoidance or ambivalence (Mikulincer and Shaver 2007). Securely attached individuals excel at conflict resolution and tend to engage in adaptive relational behaviours, as their positive self‐ and other‐models enable both emotional intimacy and autonomy (Gillath et al. 2016). Secure attachment is regularly correlated with greater relationship satisfaction, stability and mutual support, which are foundational for successful co‐parenting and family formation (Kirkpatrick and Hazan 1994; Mikulincer and Shaver 2007). Individuals with secure attachment styles tend to have strong communication and relational harmony, creating environments conducive to planned family growth and resource investment in children (Mikulincer and Florian 1998).

Conversely, individuals with insecure attachment styles display varying patterns of relational dysfunction. Fearful attachment, combining negative views of self and others, is associated with difficulty forming and maintaining romantic partnerships due to distrust and rejection sensitivity, leading to relational instability (Simpson and Rholes 2017). Preoccupied attachment, characterised by an overreliance on others for validation, can result in conflict‐prone overdependent relationships that may harm relational satisfaction (Bartholomew and Horowitz 1991). Dismissing attachment, emphasising self‐sufficiency and emotional detachment, correlates with a reluctance to establish deep connections and avoidance of vulnerability, thereby undercutting the quality of romantic relationships (Bartholomew and Horowitz 1991; Ein‐Dor et al. 2011; Pietromonaco and Beck 2019). Moreover, these patterns extend beyond romantic relationships, influencing parenting behaviours and potentially shaping reproductive outcomes. For example, avoidant individuals often invest less in caregiving roles, while anxious individuals may display intrusive or inconsistent parenting (van IJzendoorn and Bakermans‐Kranenburg 2010).

Thus, romantic partnerships provide a context for expressing attachment styles, with effects that extend to broader life history strategies, with intention and family planning as important factors. Securely attached individuals, through their capacity for mutual support and collaboration, are more likely to establish planned, stable family structures that promote child development and well‐being (Belsky et al. 1991; Mikulincer and Shaver 2007). In contrast, the relational instability and conflict commonly associated with insecure attachment can disrupt pathways to planned reproduction (Schmitt et al. 2004). For instance, patterns linked with anxious or avoidant attachment may lead to shorter relationships, less stable family units, and potential challenges in co‐parenting (Mikulincer and Shaver 2007).

Belsky et al. (1991) argued that the attachment process is an evolved psychological mechanism that monitors local social conditions and directs individuals toward an effective reproductive strategy within those constraints. Therefore, insecure attachment styles may be a response to cues in the local social environment, suggesting that long‐term relationships are less advantageous than shorter, less committed ones. Thus, insecure attachment is the basis for the formation of one's fast life history strategy, signalled by early sexual maturation and experiences, the pursuit of short‐term mating, and low parental investment in subsequent children (see Dunkel et al. 2016). In contrast, those with more secure attachment during early childhood are more likely to adopt a slow life history strategy, characterised by later sexual maturation, stable pair‐bonding, and high parental investment (Dunkel et al. 2016).

These relational patterns may subsequently influence parenting behaviours. For instance, Del Giudice (2009) suggested that avoidant attachment styles align with reproductive strategies focused on short‐term relationships and minimal parental involvement, particularly among men. In contrast, women may adopt avoidant strategies in extremely challenging environments where paternal support is unlikely, but they may lean toward anxious attachment in moderately risky settings. Anxious attachment, marked by fears of abandonment and concerns about unreciprocated affection, may encourage greater commitment and support from partners or close relatives. Thus, according to Del Giudice (2009), men tend to display higher levels of avoidance and lower levels of anxiety compared to women, signalling distinct patterns in attachment and reproductive behaviours between the sexes.

1. Attachment Styles and Reproductive Outcomes

Individuals with secure attachment styles typically display relational stability, emotional regulation, and high levels of relationship satisfaction (Mikulincer and Shaver 2007). These traits often correlate with long‐term planning and a deliberate approach to family formation. As noted earlier, these relational patterns connect to broader life history strategies, with secure attachment aligning with slower strategies characterised by fewer offspring and greater per‐child investment, while insecure attachment styles are associated with faster strategies prioritizing reproductive output (Belsky et al. 1991; Chisholm 1996). Fearful attachment, characterised by a mix of anxiety and avoidance, may drive reproductive behaviours rooted in relational ambivalence, while preoccupied attachment, marked by increased dependency, could reinforce behaviours that promote larger family sizes.

Family size serves as a proxy for reproductive success, measured by the number of children, which is an important indicator of evolutionary fitness. Međedović et al. (2022) report that among Serbians, an anxious attachment style is negatively associated with the number of children one has (as well as the motivation to have children). They posit that only secure attachment is adaptive, as it is positively related to reproductive outcomes (i.e., number of children, grandchildren) and other measures of fitness (e.g., self‐reported care for other genetically related family members). Avoidance was found to be linked to short‐term mating, but they argue that ‘this link cannot provide fitness benefits for avoidant individuals because short term mating is unrelated to the number of children and negatively related to parental investment and number of grandchildren’. These findings are consistent with other work reporting negative links between insecure attachment and fertility outcomes (Hadley et al. 2019; Međedović and Jovanov 2025).

We suggest that this conclusion warrants further scrutiny. As Ein‐Dor et al. (2010) propose, there may be advantages for attachment styles other than secure. The persistence of insecure attachment styles across human populations is consistent with several possible explanations, including functional adaptation to specific ecological or social contexts, as well as the persistence of neutral or mildly maladaptive traits through mutation‐selection balance (Zietsch 2024). We do not assume that prevalence alone implies adaptive purpose; rather, we treat the question of whether insecure attachment confers fitness benefits as an empirical one. By focusing predominantly on secure attachment as the optimal pattern, Međedović et al. risk overlooking how insecure attachment styles might be associated with fitness in varying environmental or cultural frameworks. Indeed, the theoretical framework proposed by Međedović et al. excludes the possibility that, as suggested by Ein‐Dor et al. (2010; but see also Frankenhuis 2010), insecure attachment traits may enhance survival and reproductive opportunities under specific conditions, such as heightened vigilance in preoccupied individuals or opportunistic behaviours in avoidant individuals.

Moreover, short‐term mating strategies, often associated with avoidant attachment traits, are dismissed as offering little to no evolutionary benefit within the framework presented by Međedović et al. However, short‐term mating behaviours are well‐documented across cultures (e.g., Schmitt and Shackelford 2008) and species (see for a review Schmitt 2015), indicating their potential evolutionary utility. Whether through genetic diversification, increased reproductive opportunities in fluctuating environments, or reduced pair‐bonding costs, these strategies may provide individuals with viable pathways to reproductive success that are not immediately evident in direct offspring counts. The absence of direct associations between avoidant attachment and reproductive outcomes does not rule out the possibility of adaptive benefits through indirect fitness pathways, which remain underexplored.

Cultural differences in attachment behaviours and reproductive strategies further indicate the need for a more comprehensive analysis. Secure attachment may align with long‐term family planning and regulated resource allocation, which may be advantageous in high‐income, industrialised contexts. However, in environments where unpredictability or resource scarcity prevails, alternative strategies may prove more successful. Additionally, cultural variance in individualism or collectivism, for example, may also influence behaviours that lead to reproductive success. Cultural variability in relational norms, reproductive priorities, and socioecological pressures necessitates a cross‐cultural approach to fully understand how attachment styles function as evolutionary strategies (e.g., Bakaraki et al. 2024; Keller 2013; Strand et al. 2019).

2. Sex Differences

Sex differences in attachment styles, while consistent across some cultures, also vary significantly depending on societal norms and values. Past research has indicated that women tend to score higher in anxious or preoccupied attachment styles, characterised by dependency and relational ambivalence, whereas men are more likely to exhibit avoidant or fearful styles, linked to emotional withdrawal and relational avoidance (Schmitt et al. 2003; Del Giudice 2011; Weber et al. 2022). Such patterns may be rooted in social roles, where women are expected to prioritise relationships and caregiving, while men are encouraged to emphasise independence and emotional restraint (see for review Brody and Hall 2010). These expectations often mirror broader cultural values. For example, individualistic societies, such as the USA and Canada, promote autonomy and self‐reliance, which may potentially amplify anxious attachment in women, who may internalise societal pressures to form and sustain close relationships (Schmitt et al. 2003). Men in these cultures may exhibit higher levels of fearful attachment, reflecting the tension between relational avoidance and societal expectations of independence (van IJzendoorn and Bakermans‐Kranenburg 2010). On the other hand, collectivist societies like Japan emphasise interdependence and communal harmony, which may reduce some of these sex differences while introducing unique patterns of attachment influenced by familial and societal interdependence (Li et al. 2009; Rothbaum et al. 2000; Keller 2013). For example, Japanese women may score higher on anxious and avoidant attachment dimensions due to cultural expectations of caregiving and family cohesion, while men might exhibit higher levels of fearful or dismissing attachment in response to pressures for emotional self‐sufficiency.

Existing research suggests that sociocultural environments interact with sex to influence attachment styles, but significant gaps remain in understanding how these patterns manifest across specific countries and cultural settings. Studies such as Schmitt et al. (2003) document the variability in attachment styles by sex globally, but these findings require greater contextualisation to account for specific cultural dynamics in countries like Canada, Japan and the USA. Investigating these patterns in cross‐cultural contexts provides an opportunity to disentangle the relative contributions of sex and culture to attachment behaviours, particularly in societies with contrasting values of autonomy and interdependence.

3. Cross‐Cultural Considerations

Cultural norms and societal structures may influence attachment orientations and reproductive behaviour. Japan, as a collectivist society, prioritises social harmony, interdependence, and conformity to group norms (Oyserman et al. 2002). Such values are reflected in demographic trends that include delayed parenthood, smaller family sizes, and lower fertility rates driven by significant economic pressures and shifts in roles. Women in Japan are often expected to balance traditional caregiving responsibilities with increasing participation in the workforce, which may discourage larger families (Retherford and Ogawa 2006). Contrariwise, individualistic cultures such as Canada and the USA emphasise autonomy, personal fulfilment, and self‐determination (Oyserman et al. 2002). These values align with secure attachment orientations, characterised by relational stability and deliberate approaches to family planning, as reviewed. The association of secure attachment with fewer children in these contexts could reflect a strategic prioritisation of resource allocation to ensure greater opportunities and wellbeing for offspring. For example, securely attached individuals in individualistic societies may favour smaller family sizes to maximise the quality of parental investment, mirroring slower life history strategies (Belsky et al. 1991).

4. Current Study

The present study examines the differential relationships between attachment styles and the number of children across the three cultural contexts of Japan, Canada, and the USA. It is hypothesised (H1) that secure attachment will be negatively associated with the number of children in Canada and the USA. This prediction follows directly from life history theory (Belsky et al. 1991), in that securely attached individuals in high‐resource, individualistic environments are expected to pursue slower reproductive strategies, prioritizing quality of parental investment over quantity of offspring. The negative association with raw child counts is therefore theoretically expected rather than paradoxical. However, we hypothesise that there will be no significant association in Japan, where documented demographic pressures (including delayed marriage, structural career demands, and declining fertility rates; Retherford and Ogawa 2006) may attenuate the expression of individual attachment dispositions on reproductive outcomes (Keller 2013; Rothbaum et al. 2000). Next, we hypothesise (H2) that fearful attachment will be positively associated with the number of children in Canada and the USA. We acknowledge that the prior literature, including studies linking anxious attachment and neuroticism to lower fertility (Penke and Jokela 2016), would lead to the opposite prediction. We nonetheless propose this hypothesis on the basis that faster life history is associated with fearful attachment (including relational instability and less deliberate family planning) which may produce higher actual child counts even in the absence of higher desired fertility. The same trend is expected in Japan, but the relationship is expected to be weaker than in Canada and the USA due to collectivist cultural norms that may buffer the effects of relational instability through external family and community support. We hypothesise (H3) that preoccupied attachment will be positively associated with the number of children in Canada and the USA. The dependency and relational ambivalence common among preoccupied individuals are expected to contribute to higher reproductive output, possibly to secure relational bonds or due to less deliberate family planning. The same pattern is predicted to occur in Japan, where collectivist values reinforce familial interdependence, amplifying the effects of preoccupied attachment on reproductive behaviour.

Last, we hypothesise (H4) that sex differences in attachment styles will reflect cultural variations. In individualistic cultures like the USA and Canada, women are predicted to display higher preoccupied attachment, reflecting relational dependency, while men are likely to exhibit greater fearful attachment, aligning with ideals of independence and emotional detachment. However, in collectivist cultures such as Japan, women are predicted to score higher on both preoccupied and secure attachment styles, due to the cultural emphasis on caregiving and relational harmony. Men in Japan are predicted to show greater fearful or dismissing attachment, influenced by societal pressures for emotional restraint and self‐reliance.

We also explore, without a priori predictions, how attachment style is related to one's ideal number of children. It is possible that some participants desire more children than they currently have but are limited by economic and living situations, age at the time of the survey, medical concerns, or similar factors. Beyond logistical realities, the ideal number of children could also rely heavily on other ideal relationships, such as a more supportive family or a more invested partner. Whether the ideal number of children is based on current real‐life conditions or on other hypothetical factors is a research question. Such hypothetical matters do not lend themselves well to predictions, but an exploratory analysis can reveal insights that inform future, more firmly grounded studies.

5. Methods

An online survey was conducted to investigate the relationship between attachment styles and the actual and ideal number of children. We designed the experimental framework and questionnaires, and a well‐known survey company was hired to distribute the survey, perform attention checks, and collect responses. The respondents were registered users of the company living in Japan, Canada and the USA, and aged 20 years or older. Respondents provided consent and completed the survey anonymously.

The survey was conducted from February 10 to March 9, 2022, in all three countries. The demographic attributes of the respondents, such as age and sex, were based on the survey company's records. For the survey, we sampled responses so that the number of respondents was equal for each sex and each of five age bins: 20s, 30s, 40s, 50s and 60s and older, yielding 10 segments with the same number of responses. This equal stratification was adopted to ensure sufficient statistical power for age‐stratified analyses; sampling proportional to the population structure would have resulted in substantially smaller cell sizes in the youngest and oldest age groups, limiting the ability to make meaningful cross‐age comparisons. The questionnaires were prepared in Japanese (for those in Japan) and English (for those in Canada and the USA). Table 1 shows the total number of samples and the sex ratios for each country.

TABLE 1.

Basic attributes of the dataset.

Country		Japan	Canada	USA
Survey term		February 10–February 27, 2022	February 21–March 9, 2022	February 18–March 9, 2022
Number of responses	Men	2520	2520	2520
	Women	2520	2520	2520
	Total	5040	5040	3210

Adult attachment styles were measured using the four‐item Relationship Questionnaire (RQ) from Bartholomew and Horowitz (1991), with participants indicating how much each attachment style applies to them. Participants responded to a set of descriptive statements about their relational behaviours and attitudes, rating their agreement on a Likert‐type scale of 1 (disagree strongly) to 7 (agree strongly). Thus, each attachment style was rated as a continuous variable on a 7‐point Likert‐type scale, yielding secure, dismissing, preoccupied, and fearful styles based on scoring guidelines.

The four attachment descriptions from the RQ are as follows. The secure description was, ‘It is relatively easy for me to become emotionally close to others. I am comfortable depending on them and having others depend on me. I don't worry about being alone or having others not accept me’. The dismissing item was, ‘I am comfortable without close emotional relationships. It is very important to me to feel independent and self‐sufficient, and I prefer not to depend on others or have others depend on me’. The preoccupied item was, ‘I want to be completely emotionally intimate with others, but I often find that others are reluctant to get as close as I would like. I am uncomfortable being without close relationships, but I sometimes worry that others don't value me as much as I value them’. Finally, the fearful item was, ‘I am somewhat uncomfortable getting close to others. I want emotionally close relationships, but I find it difficult to trust others completely, or to depend on them. I sometimes worry that I will be hurt if I allow myself to become too close to others’.

Reproductive success was measured by asking, ‘How many children do you have?’ Respondents were asked to ‘please answer this question assuming your own children related to you by blood’. They were also asked, ‘What is the ideal number of children you would like to have?’

6. Results

Descriptive statistics for the current number of children and the ideal number of children were examined for each sample, see Tables 2, 3, 4. Attachment scores varied across countries, as shown in Table 5.

TABLE 2.

Japan sample: current number of children and ideal number of children desired (N = 5040).

Number of children	Currentchildren n (%)	Ideal number desired n (%)
0	2280 (45.2)	1359 (27.0)
1	821 (16.3)	376 (7.5)
2	1412 (28.0)	2163 (42.9)
3	466 (9.2)	1019 (20.2)
4	36 (0.7)	78 (1.5)
5	13 (0.3)	33 (0.7)
6	4 (0.1)	5 (0.1)
7 or more	8 (0.2)	7 (0.1)

Note: Percentages are based on the total sample (N = 5040). Categories with more than 6 children were combined into the ‘7 or more’ category to improve readability.

TABLE 3.

Canadian sample: current number of children and ideal number of children desired (N = 5040).

Number of children	Currentchildren n (%)	Ideal number desired n (%)
0	2333 (46.3)	1788 (35.5)
1	928 (18.4)	696 (13.8)
2	1101 (21.8)	1567 (31.1)
3	456 (9.0)	612 (12.1)
4	144 (2.9)	249 (4.9)
5	45 (0.9)	69 (1.4)
6	21 (0.4)	41 (0.8)
7 or more	12 (0.2)	18 (0.4)

Note: Percentages are based on the total sample (N = 5040). Categories greater than 6 children were combined into the ‘7 or more’ category to improve readability.

TABLE 4.

USA sample: current number of children and ideal number of children desired (N = 5040).

Number of children	Currentchildren n (%)	Ideal number desired n (%)
0	2038 (40.4)	1754 (34.8)
1	965 (19.1)	716 (14.2)
2	1074 (21.3)	1340 (26.6)
3	538 (10.7)	659 (13.1)
4	238 (4.7)	314 (6.2)
5	94 (1.9)	129 (2.6)
6	46 (0.9)	58 (1.2)
7 or more	47 (0.9)	70 (1.3)

Note: Percentages are based on the total sample (N = 5040). Categories with more than 6 children were combined into the ‘7 or more’ category to improve readability.

TABLE 5.

Descriptive statistics for attachment styles across countries.

	Country
	Japan	Canada	USA
Attachment style
Secure	4.00 (1.41)	3.55 (1.74)	3.52 (1.91)
Fearful	3.89 (1.51)	3.90 (1.81)	3.81 (2.00)
Preoccupied	4.08 (1.40)	4.21 (1.70)	4.09 (1.87)
Dismissing	3.80 (1.37)	3.59 (1.66)	3.63 (1.80)

Note: A higher number indicates greater endorsement of the attachment style. Mean provided with standard deviations in parentheses.

A Poisson regression model was computed for the total sample of individuals to assess the association between attachment styles and the current, actual number of children. The model had current number of children as the dependent variable with the following independent variables: Country, relationship status, sex of participant, age, attachment styles, interaction of country and attachment styles, and interaction of sex and attachment styles. The regression model was a significant fit, X ²(15) = 4450.16, p < 0.001. As shown in Table 6, country, age, sex, fearful, preoccupied and dismissing attachment styles, the interaction of attachment style and the interaction of sex and attachment style were significant predictors of the number of current children.

TABLE 6.

Regression predicting number of actual children.

Predictor	B	SE	IRR	95% CI for IRR	p
Intercept	1.11	0.11	3.04	[2.46, 3.76]	< 0.001
Country (1 vs. 3)	−0.63	0.10	0.53	[0.43, 0.65]	< 0.001
Country (2 vs. 3)	−0.38	0.06	0.68	[0.61, 0.76]	< 0.001
Relationship status (1 vs. 6)	−0.58	0.08	0.56	[0.48, 0.66]	< 0.001
Relationship status (2 vs. 6)	0.62	0.08	1.86	[1.58, 2.19]	< 0.001
Relationship status (3 vs. 6)	0.38	0.10	1.46	[1.20, 1.77]	< 0.001
Relationship status (4 vs. 6)	0.58	0.10	1.78	[1.46, 2.16]	< 0.001
Relationship status (5 vs. 6)	0.42	0.09	1.52	[1.28, 1.80]	< 0.001
Sex (1 vs. 2)	−0.39	0.08	0.68	[0.58, 0.80]	< 0.001
Age	0.07	0.00	1.07	[1.06, 1.08]	< 0.001
Secure attachment	0.02	0.01	1.02	[0.99, 1.04]	< 0.166
Fearful attachment	0.04	0.01	1.04	[1.02, 1.07]	< 0.002
Preoccupied attachment	0.05	0.01	1.05	[1.02, 1.07]	< 0.001
Dismissing attachment	0.04	0.01	1.04	[1.02, 1.07]	< 0.001
Sex × Attachment style	−0.01	0.01	0.99	[0.98, 1.00]	< 0.019
Country × Attachment style	−0.01	0.00	0.99	[0.99, 1.00]	< 0.012

Note: IRR = incidence rate ratio. Reference categories were Country = 3, Relationship status = 6, and Sex = 2. Dependent variable: number of children.

To further examine the association between the number of current children and attachment styles, additional Poisson regression models were computed for each of the three Country samples. Note that the Country × Attachment interaction did not reach significance for the current number of children in the combined model (Table 6), so the following country‐specific models are presented as descriptive summaries to facilitate cross‐sample comparison rather than as formal interaction decompositions.

The regression model for the Japan sample was a significant fit, X ²(11) = 4045.13, p < 0.001, see Table 7. A 1‐unit change in relationship status is associated with an increase in the ideal number of children (IRRs range from 0.03 to 2.65, 95% CIs range from [0.01, 0.05 to 1.84, 3.82], p < 0.001). Married, widowed, and divorced individuals report having higher numbers of children. Also, older individuals report having more children than younger individuals (IRR = 1.07, 95% CI [1.06, 1.08], p < 0.001).

TABLE 7.

Japan sample: regression predicting number of children.

Predictor	B	SE	IRR	95% CI for IRR	p
Intercept	−1.14	0.20	0.32	[0.22, 0.47]	< 0.001
Relationship status (1 vs. 6)	−3.67	0.29	0.03	[0.01, 0.05]	< 0.001
Relationship status (2 vs. 6)	0.98	0.19	2.65	[1.84, 3.82]	< 0.001
Relationship status (3 vs. 6)	0.95	0.20	2.60	[1.76, 3.83]	< 0.001
Relationship status (4 vs. 6)	0.89	0.20	2.43	[1.62, 3.63]	< 0.001
Relationship status (5 vs. 6)	0.79	0.19	2.20	[1.51, 3.20]	< 0.001
Sex (1 vs. 2)	−0.02	0.03	0.98	[0.93, 1.03]	< 0.439
Age	0.06	0.01	1.07	[1.06, 1.08]	< 0.001
Secure attachment	0.01	0.01	1.01	[0.98, 1.03]	< 0.631
Fearful attachment	0.02	0.01	1.02	[1.00, 1.04]	< 0.121
Preoccupied attachment	−0.00	0.01	1.00	[0.98, 1.02]	< 0.861
Dismissing attachment	0.02	0.01	1.02	[1.00, 1.04]	< 0.077

Note: IRR = incidence rate ratio. Reference categories: Relationship status = 6 and Sex = 2. Dependent variable: number of children.

The regression model for the Canada sample was a significant fit, X ²(11) = 2083.28, p < 0.001, see Table 8. A 1‐unit change in relationship status is associated with an increase in the ideal number of children (IRRs range from 0.47 to 1.67, 95% CIs range from [0.36, 0.60 to 1.31, 2.12], p < 0.001). Married, widowed, and divorced individuals report having higher numbers of children. Also, men report having fewer the children, IRR = 0.84, 95% CI [0.80, 0.89], p < 0.001. Older individuals report having fewer children, IRR = 1.07, 95% CI [1.06, 1.08], p < 0.001. Lastly, securely attached individuals report having fewer children, IRR = 0.98, 95% CI [0.96, 1.00], p < 0.02.

TABLE 8.

Canada sample regression predicting number of children.

Predictor	B	SE	IRR	95% CI for IRR	p
Intercept	−0.455	0.137	0.63	[0.49, 0.83]	< 0.001
Relationship Status 1	−0.762	0.127	0.47	[0.36, 0.60]	< 0.001
Relationship Status 2	0.512	0.123	1.67	[1.31, 2.12]	< 0.001
Relationship Status 3	0.320	0.142	1.38	[1.04, 1.82]	< 0.024
Relationship Status 4	0.405	0.140	1.50	[1.14, 1.97]	< 0.004
Relationship Status 5	0.358	0.127	1.43	[1.11, 1.84]	< 0.005
Sex (1)	−0.175	0.027	0.84	[0.80, 0.89]	< 0.001
Age	0.067	0.005	1.07	[1.06, 1.08]	< 0.001
Secure Attachment	−0.020	0.008	0.98	[0.96, 1.00]	< 0.017
Fearful Attachment	−0.002	0.008	1.00	[0.98, 1.02]	< 0.858
Preoccupied Attachment	0.007	0.009	1.01	[0.99, 1.02]	< 0.439
Dismissing Attachment	0.008	0.008	1.01	[0.99, 1.02]	< 0.340

Note: IRR = incidence rate ratio. IRR values represent the multiplicative change in the expected count of children for a one‐unit increase in the predictor. Reference categories: Relationship Status = 6; Sex = 2.

The regression model for the USA sample was a significant fit, X ²(11) = 1963.52, p < 0.001; see Table 9. A 1‐unit change in relationship status is associated with an increase in the ideal number of children (IRRs range from 0.41 to 1.31, 95% CIs range from [0.01, 0.05 to 1.84, 3.82], p < 0.001). Married, widowed, and divorced individuals report having higher numbers of children. Also, older individuals report having more children than younger individuals, IRR = 1.07, 95% CI [1.06, 1.08], p < 0.001. Also, men report having fewer children, IRR = 0.76, 95% CI [0.72 0.79], p < 0.001. Older individuals report having fewer children, IRR = 1.04, 95% CI [1.03, 1.04], p < 0.001. Lastly, securely attached individuals report having fewer children, IRR = 0.98, 95% CI [0.97, 0.99], p < 0.004.

TABLE 9.

USA sample: Poisson regression predicting number of children.

Predictor	B	SE	IRR	95% CI for IRR	p
Intercept	0.332	0.102	1.39	[1.14, 1.70]	0.001
Relationship Status 1	−0.881	0.095	0.41	[0.34, 0.50]	< 0.001
Relationship Status 2	0.267	0.092	1.31	[1.09, 1.56]	< 0.004
Relationship Status 3	0.021	0.116	1.02	[0.81, 1.28]	< 0.855
Relationship Status 4	0.209	0.101	1.23	[1.01, 1.50]	< 0.039
Relationship Status 5	0.171	0.094	1.19	[0.99, 1.43]	< 0.070
Sex (1)	−0.281	0.025	0.76	[0.72, 0.79]	< 0.001
Age	0.034	0.004	1.04	[1.03, 1.04]	< 0.001
Secure Attachment	−0.020	0.007	0.98	[0.97, 0.99]	< 0.004
Fearful Attachment	−0.005	0.007	1.00	[0.98, 1.01]	< 0.435
Preoccupied Attachment	0.010	0.007	1.01	[1.00, 1.02]	< 0.172
Dismissing Attachment	−0.010	0.007	0.99	[0.98, 1.00]	< 0.150

Note: IRR = incidence rate ratio. IRR represents the multiplicative change in the expected count of children for a one‐unit increase in the predictor. Reference categories: Relationship Status = 6; Sex = 2.

A logistic regression was conducted to examine whether attachment style ratings differed systematically by sex across the total sample. Attachment styles and their interaction with country were entered as predictors; sex of participant (1 = men, 2 = women) served as the outcome variable. This framing is intended to characterise sex differences in attachment, not to imply that attachment determines biological sex. The model was significant X ²(5) = 20766.87, p < 0.02, R ² = 0.01, Nagelkerke R ² = 0.02. Significant predictors were secure attachment (b = 0.05, p < 0.001, Odds ratio = 1.05, Wald = 26.41), fearful attachment (b = −0.06, p < 0.001, Odds ratio = 0.95, Wald = 28.59), preoccupied attachment (b = 0.10, p < 0.001, Odds ratio = 1.10, Wald = 76.47), and dismissing attachment (b = 0.08, p < 0.001, Odds ratio = 0.1.08, Wald = 53.82). Men were higher in fearful attachment, while women were higher in the other attachment styles.

The number of current children and ideal number of children were correlated, Spearman r = 0.65, p < 0.001, see Table 10 for each country. Note that the ideal number of children is confounded by existing parity. In all three samples, correlations were positive and statistically significant, indicating that individuals who currently have more children also tend to report higher ideal family sizes, although the relationship was moderate, suggesting these constructs are related but not redundant.

TABLE 10.

Spearman correlations between the number of actual child(ren) and ideal number of children.

Country
Japan	Canada	USA
0.65***	0.67***	0.65***

^*** p < 0.001.

A Poisson regression model was computed for the total sample of individuals in the dataset who reported not having children to assess the association between attachment styles and the ideal number of children. The Ns for each country were as follows: Japan = 2280, Canada = 2333, and USA = 2038, for 6651 in total. The model had the ideal number of children as the dependent variable with the following independent variables: Country, relationship status, sex of participant, age, attachment styles, interaction of country and attachment styles, and interaction of sex and attachment styles. The regression model was a significant fit, X ²(15) = 263.77, p < 0.001. As shown in Table 11, country, age, fearful attachment style, and the interaction of attachment style were significant predictors of the ideal number of children.

TABLE 11.

Total sample: regression predicting ideal number of children.

Predictor	B	SE	IRR	95% CI for IRR	p
Intercept	0.77	0.15	2.17	[1.61, 2.92]	< 0.001
Relationship Status (1)	0.18	0.11	1.20	[0.96, 1.50]	< 0.104
Relationship Status (2)	0.40	0.12	1.49	[1.18, 1.87]	< 0.001
Relationship Status (3)	0.03	0.17	1.03	[0.74, 1.43]	< 0.876
Relationship Status (4)	0.53	0.19	1.70	[1.18, 2.47]	< 0.005
Relationship Status (5)	0.08	0.14	1.08	[0.83, 1.41]	< 0.549
Sex of Participant	0.01	0.12	1.01	[0.80, 1.27]	< 0.934
Country (1)	−0.59	0.14	0.55	[0.42, 0.73]	< 0.001
Country (2)	−0.49	0.07	0.61	[0.53, 0.71]	< 0.001
Sex Age	−0.17	0.01	0.85	[0.84, 0.86]	< 0.001
Secure Attachment	0.02	0.02	1.02	[0.99, 1.06]	< 0.194
Fearful Attachment	0.09	0.02	1.09	[1.05, 1.12]	< 0.001
Preoccupied Attachment	0.04	0.02	1.04	[1.01, 1.08]	< 0.013
Dismissing Attachment	0.05	0.02	1.05	[1.02, 1.09]	< 0.003
Sex × Attachment Style	−0.01	0.01	0.99	[0.98, 1.01]	< 0.383
Country × Attachment Style	−0.02	0.00	0.98	[0.97, 0.99]	< 0.001

Note: B = unstandardised coefficient; CI = confidence interval; IRR = incidence rate ratio (Exp(B)); SE = standard error. Reference categories were relationship status (6), sex of participant (2), and country (3).

To further examine the association between the ideal number of children desired and attachment styles Poisson regression models were then computed across the three countries. For the Japanese sample, the regression model was significant, X ²(11) = 142.82, p < 0.001; see Table 12. A 1‐unit change in relationship status is associated with an increase in the number of ideal number of children (IRRs range from 2.38 to 6.94, 95% CIs range from [0.321, 1.41 to 1.25, 2.62], p < 0.002 to 0.001). Additionally, men desired a larger number of children than women (IRR = 1.10, 95% CI [0.006, 0.19], p < 0.04). Older individuals also desire fewer children (IRR = 0.93, 95% CI [−0.09, −0.05], p < 0.001). Lastly, fearfully attached individuals (IRR = 1.05, 95% CI [0.014, 0.08], p < 0.004) and individuals with a dismissing attachment style desire a higher number of children (IRR = 1.05, 95% CI [0.011, 0.08], p < 0.01).

TABLE 12.

Japan sample: predicting ideal number of children desired.

Predictor	B	SE	IRR	95% CI for IRR	p
Intercept	−1.08	0.30	0.34	[0.19, 0.61]	< 0.001
Relationship Status (1)	0.87	0.28	2.38	[1.38, 4.11]	< 0.002
Relationship Status (2)	1.25	0.28	3.49	[2.01, 6.05]	< 0.001
Relationship Status (3)	1.32	0.34	3.75	[1.91, 7.37]	< 0.001
Relationship Status (4)	1.94	0.35	6.94	[3.51, 13.74]	< 0.001
Relationship Status (5)	1.24	0.30	3.44	[1.92, 6.15]	< 0.001
Sex of Participant	0.10	0.05	1.10	[1.01, 1.21]	< 0.036
Sex Age	−0.07	0.01	0.93	[0.92, 0.95]	< 0.001
Secure Attachment	−0.03	0.02	0.97	[0.94, 1.00]	< 0.083
Fearful Attachment	0.05	0.02	1.05	[1.02, 1.08]	< 0.004
Preoccupied Attachment	0.01	0.02	1.01	[0.98, 1.04]	< 0.589
Dismissing Attachment	0.05	0.02	1.05	[1.01, 1.08]	< 0.010

Note: B = unstandardised coefficient; CI = confidence interval; IRR = incidence rate ratio (Exp(B)); SE = standard error. Reference categories were relationship status (6) and sex of participant (2).

For the Canadian sample, the regression model was also significant, X ²(11) = 641.93, p < 0.001, see Table 13. Divorced individuals and widowed individuals desire fewer children (IRRs range from 0.57 to 1.91, 95% CIs range from [0.52, 1.24, to −1.08, −0.05], p < 0.03). Additionally, men desired a larger number of children than women (IRR = 1.11, 95% CI [0.007, 0.21], p < 0.04). Older individuals also desire fewer children (IRR = 0.78, 95% CI [−0.28, −0.23], p < 0.001). Last, securely attached individuals (IRR = 0.96, 95% CI [−0.07, −0.007], p < 0.02) desire a smaller ideal number of children.

TABLE 13.

Canada sample: regression predicting ideal number of children desired.

Predictor	B	SE	IRR	95% CI for IRR	p
Intercept	0.94	0.22	2.57	[1.68, 3.91]	< 0.001
Relationship Status (1)	−0.10	0.19	0.90	[0.62, 1.32]	< 0.599
Relationship Status (2)	−0.01	0.20	0.99	[0.67, 1.46]	< 0.943
Relationship Status (3)	−0.38	0.28	0.69	[0.40, 1.19]	< 0.178
Relationship Status (4)	0.65	0.30	1.91	[1.05, 3.47]	< 0.033
Relationship Status (5)	−0.57	0.26	0.57	[0.34, 0.95]	< 0.031
Sex of Participant	0.11	0.05	1.12	[1.01, 1.24]	< 0.036
Sex Age	−0.25	0.01	0.78	[0.76, 0.80]	< 0.001
Secure Attachment	−0.04	0.02	0.96	[0.94, 0.99]	< 0.016
Fearful Attachment	−0.02	0.02	0.98	[0.95, 1.01]	< 0.243
Preoccupied Attachment	−0.01	0.02	0.99	[0.96, 1.02]	< 0.615
Dismissing Attachment	0.02	0.02	1.02	[0.99, 1.05]	< 0.327

Note: B = unstandardised coefficient; CI = confidence interval; IRR = incidence rate ratio (Exp(B)); SE = standard error. Reference categories were relationship status (6) and sex of participant (2).

Finally, for the USA sample, the regression model was significant, X ²(11) = 671.11, p < 0.001, see Table 14. Widowed individuals desire a smaller number of children (IRR = 0.14, 95% CIs [−0.76, 0.08], p < 0.008). Additionally, men desire a larger number of children than women (IRR = 1.17, 95% CI [0.05, 0.26], p < 0.004). Older individuals also desire fewer children (IRR = 0.80, 95% CI [−0.25, −0.21], p < 0.001). Lastly, securely attached individuals (IRR = 0.97, 95% CI [−0.06, −0.005], p < 0.02), individuals with: a dismissing attachment style (IRR = 0.94, 95% CI [−0.09, −0.04], p < 0.001), and a preoccupied attachment style (IRR = 0.97, 95% CI [0.07, −0.007], p < 0.02), desire a smaller number of children (IRR = 1,05, 95% CI [0.011, 0.08], p < 0.01) while individuals with a fearful attachment style desire (IRR = 1.07, 95% CI [0.04, −0.09], p < 0.001) a higher number of children.

TABLE 14.

USA sample: regression predicting ideal number of children desired.

Predictor	B	SE	IRR	95% CI for IRR	p
Intercept	1.00	0.18	2.71	[1.89, 3.88]	< 0.001
Relationship Status (1)	0.00	0.16	1.00	[0.73, 1.37]	< 0.997
Relationship Status (2)	−0.02	0.18	0.98	[0.70, 1.38]	< 0.911
Relationship Status (3)	−0.55	0.30	0.58	[0.32, 1.05]	< 0.072
Relationship Status (4)	−1.94	0.73	0.14	[0.04, 0.60]	< 0.008
Relationship Status (5)	−0.34	0.21	0.71	[0.47, 1.08]	< 0.110
Sex of Participant	0.15	0.05	1.17	[1.05, 1.29]	< 0.004
Sex Age	−0.23	0.01	0.80	[0.78, 0.82]	< 0.001
Secure Attachment	−0.03	0.01	0.97	[0.94, 1.00]	< 0.022
Fearful Attachment	0.06	0.01	1.07	[1.04, 1.10]	< 0.001
Preoccupied Attachment	−0.04	0.01	0.97	[0.94, 0.99]	< 0.016
Dismissing Attachment	−0.07	0.01	0.94	[0.91, 0.97]	< 0.001

Note: B = unstandardised coefficient; CI = confidence interval; IRR = incidence rate ratio (Exp(B)); SE = standard error. Reference categories were relationship status (6) and sex of participant (2).

7. Discussion

The present study examines how adult attachment styles are associated with reproductive success across various cultural contexts and between the sexes. By analysing samples from Japan, Canada and the USA, this research explores the associations between attachment patterns, cultural norms and reproductive strategies (as measured by reproductive success or actual number of children and ideal number of children).

The relationship between secure attachment and reproductive success varied across cultural contexts, supporting the first hypothesis (H1). Consistent with predictions, secure attachment was negatively associated with the number of children in Canada and the USA, suggesting that securely attached individuals in these countries employ reproductive strategies characterised by deliberate long‐term family planning and a focus on resource‐intensive parenting. These findings align with life history theory, which posits that secure attachment is associated with slower life history strategies, prioritizing offspring quality over quantity (Belsky et al. 1991). However, in Japan, secure attachment showed no significant relationship with reproductive success, as predicted. We propose that this result reflects Japan's collectivist cultural context, where external socio‐economic and cultural factors overshadow individual factors such as attachment orientation. The cultural emphasis on professional and economic stability, combined with de‐emphasis on family planning, an ageing population and declining fertility rates, may diminish the relevance of individual‐level attachment traits in predicting reproductive behaviour (Retherford and Ogawa 2006).

The second hypothesis (H2) predicted that fearful attachment would positively correlate with reproductive success across all three countries, but only to a weak extent in Japan due to the collectivist sociocultural environment. The results partially supported this hypothesis. Fearful attachment was positively associated with reproductive success across all three nations, affirming its link to fast life‐history strategies and relational instability. Contrary to predictions, however, the association was stronger in Japan than Canada and the USA. This unexpected finding may be attributable to the collectivist cultural emphasis on communal support and familial interdependence, which could mitigate the negative effects of relational instability commonly associated with fearful attachment. Thus, individuals with fearful attachment may be better able to overcome some relational barriers and achieve greater reproductive success in a collectivistic culture than in individualistic cultures. This finding shows the significance of cultural context in exploring the association between attachment styles and reproductive strategies. The interaction between insecure attachment and reproductive behaviour in collectivist settings may involve external support structures, such as extended family networks, that reduce the burden of relational instability on family planning (Rothbaum et al. 2000).

Hypothesis three (H3) predicted a positive association between preoccupied attachment and reproductive success in all three cultural contexts. This hypothesis was confirmed. Across Japan, Canada, and the USA, preoccupied attachment consistently predicted higher numbers of children. This result points to the potential role of preoccupied attachment in driving relationship dependency and behaviours oriented toward increasing family size as a means of relational security. This finding is consistent with prior research identifying preoccupied attachment as a mechanism for reinforcing relational bonds and securing parental investment, particularly in contexts of relational ambivalence or insecurity (Bartholomew and Horowitz 1991). However, the effect was weaker in Japan compared to the other two countries. This pattern could again be explained by Japan's collectivist values, which may reduce reliance on individual‐level relational behaviours in favour of broader familial or societal structures that shape reproductive decisions.

The fourth hypothesis (H4) proposed that sex differences in attachment styles would vary across cultural contexts, reflecting the relationship between societal norms and cultural values. The findings supported this hypothesis, as women in Canada and the USA scored higher in preoccupied attachment, while men scored higher in fearful attachment. These patterns align with norms in individualistic societies, where women are often socialised to place importance upon relational bonds and engage in caregiving, while men are encouraged to be independent (Brody and Hall 2010). Women scoring higher in preoccupied attachment may align with tendencies to prioritise relational bonds and dependency, sometimes at the expense of emotional independence (Simpson and Rholes 2017). Fearful attachment in men in individualistic societies might correspond to patterns of relational instability and avoidance, reflecting social norms that discourage open emotional dependency (Cassidy and Shaver 2018). The lack of significant sex differences in secure attachment supports the possibility that secure attachment is a universal adaptive orientation possessed by the majority of individuals (Scharfe 2016).

The results for Japan revealed a more complex pattern, as women scored higher in preoccupied and secure attachment, consistent with the cultural emphasis on caregiving and relational harmony. They also scored higher in dismissing attachment, a result not anticipated in the hypothesis. This finding likely reflects the tensions Japanese women face in balancing traditional caregiving expectations with increasing societal pressures to follow professional and economic pursuits (Sugimoto 2011). Japanese men scored higher in fearful attachment, as predicted, emphasising the cultural expectations of emotional restraint and relational caution (Brody and Hall 2010). These findings suggest that collectivist values play a role in shaping attachment, where interdependence and communal harmony may enhance certain attachment dimensions in women while promoting emotional self‐sufficiency in men (Rothbaum et al. 2000).

Taken together, the findings clearly show interactions between culture and sex in shaping adult attachment styles. Japan's collectivism may increase sex differences in attachment by emphasising relational interdependence in women, while men may internalise norms of emotional independence and fear of rejection, reflected in higher levels of fearful attachment. In Canada and the USA, individualism could dilute such distinctions but still reflectsex differences, with women demonstrating relational ambivalence (preoccupied attachment) and men expressing relational unease (fearful attachment) (Keller 2013). These patterns indicate the importance of contextualising attachment theory within sociocultural frameworks to better understand how sex and culture influence emotional and relational patterns.

While we had no a priori predictions about how attachment styles might influence actual versus desired number of children, we did find similar patterns. The positive correlation obtained suggests that as the ideal number of children increases, so does the actual number of children. This may indicate that a person's ideal number of children is based more on their own real‐life experiences than hypothetical scenarios. Also, in general, more secure attachment scores were associated with a lower actual number of children, as well as ideal numbers. Fearful and preoccupied scores were associated with higher actual and ideal number of children, while dismissive attachment was also related to higher ideal numbers in Japan and Canada (though not the actual number of children). This pattern indicates that while similar factors are likely to influence the desire for more children as well as the realisation of that desire across cultures, the constraints on actual behaviour and offspring production (e.g., income, culture, social support, work‐life trade‐offs) can result in less children produced than the number desired (and vice‐versa, as some women may have more children than they desire due to partner desires, cultural pressures, and poor access to birth control). Adaptive mechanisms do not always maximise reproductive success across all environments. It is also notable that dismissive attachment was the only attachment dimension that showed a significant association with the ideal number of children, but not with the actual number of children. This discrepancy may reflect the possibility that reproductive desires are suppressed by social constraints or implicit gender role expectations in more collectivistic or conformity‐oriented societies such as Japan and Canada. In contrast, individuals in the more individualistic cultural context of the United States may find it easier to translate ideal reproductive goals into actual behaviour, being relatively less constrained by external social norms.

The results of this study converge with and extend previous findings in several ways. The observed sex differences in attachment styles are consistent with global patterns identified by Schmitt et al. (2003), who documented that women tend to score higher in anxious or preoccupied attachment, while men more often exhibit avoidant or fearful orientations. Additionally, the association between secure attachment and lower reproductive success in individualistic cultures corroborates life history perspectives emphasising slower strategies among securely attached individuals (Belsky et al. 1991; see also Dunkel et al. 2016).

These findings notably diverge from Međedović et al. (2022), who suggested that only secure attachment is adaptive for reproductive success based on a Serbian sample. We propose that the positive associations between fearful and preoccupied attachment styles and higher numbers of children in the current study suggest that insecure attachment styles may confer important fitness advantages in specific contexts (Ein‐Dor et al. 2010; but see Frankenhuis 2010), and that the findings need to be replicated in more cultures.

There remain several issues to address with future work. For example, Schmitt (2005) examined the relationship between attachment styles, personality traits, and short‐term mating across various cultures. His findings revealed a moderate association between insecure attachment and short‐term mating, as well as a strong link between insecure attachment and maladaptive personality traits. However, short‐term mating appeared largely independent of maladaptive personality. Notably, in certain cases, particularly among young men, short‐term mating was linked to adaptive personality traits, such as higher self‐esteem. Schmitt argued that while insecure romantic attachment may play a role in short‐term mating behaviours, the causal connection between early parent–child attachment and short‐term mating was unclear. More recently, Uhlich et al. (2022) examined the effects of attachment primes on mating preferences, finding that exposure to secure cues increased preferences for long‐term mating strategies while reducing preferences for short‐term strategies. This supports life history theory, which suggests that secure attachment is associated with stable, long‐term relational goals. Additionally, insecure attachment styles remain strongly associated with short‐term mating strategies. Avoidant attachment, in particular, is linked to a preference for casual sexual relationships and lower emotional intimacy (Schmitt 2005; Uhlich et al. 2022). Sex differences have also been observed, with men displaying avoidant attachment often favouring short‐term mating strategies, while women with anxious attachment exhibit ambivalence toward casual relationships.

The positive relationship between fearful or preoccupied attachment styles and larger family sizes is consistent across the three nations examined. This universality may stem from basic psychological mechanisms associated with insecure attachment. For instance, preoccupied and fearful orientations often correlate with heightened relationship dependency or ambivalence, potentially leading to relational and sexual behaviours that prioritise immediate attachment needs over long‐term reproductive planning (Bartholomew and Horowitz 1991; Chisholm 1996). However, the effect of these behaviours on reproduction may be modulated by cultural values. For example, greater emphasis on familial interdependence in collectivist cultures like Japan may strengthen the association between insecure attachment styles and family expansion, while in individualistic cultures, it could manifest differently based on personal and relational dynamics. Further research is required.

One final area of future inquiry relates to sex differences in the association between life history and attachment. Dunkel et al. (2016) reported sex differences with women exhibiting higher levels of preoccupied attachment, and with slow life history strategies positively correlating with this attachment style. For men, slow life history strategies were associated with secure attachment, while negatively correlating with both fearful and dismissing attachment styles. Del Giudice (2009) further proposed sex‐specific predictions, suggesting that slow life history strategies correlate with secure attachment for both sexes, whereas fast life history strategies align with dismissing attachment in men and preoccupied attachment in women. While Dunkel et al. (2016) observed a positive link between slow life history strategies and secure attachment, this relationship reached statistical significance only in men. Importantly, hierarchical regression analyses indicated no significant sex differences in how life history strategies predict secure attachment. These findings present gaps and opportunities for further research. Future studies could explore why the sex‐specific predictions of Del Giudice (2009) are only partially supported, particularly regarding the weaker associations in women. Additionally, examining cultural and developmental factors that may moderate how life history strategies influence attachment styles could be helpful, particularly using data from diverse populations or longitudinal designs that cross‐sectional studies have not captured.

A further limitation concerns the use of life history theory (LHT) as an interpretive framework. Although LHT informs our discussion of the associations between attachment styles and reproductive outcomes, the present study did not directly measure any life history variables, such as environmental harshness, pubertal timing, somatic investment, or mating effort. Our use of cultural context (individualism–collectivism indices; national fertility statistics) as a distal proxy for ecological conditions is theoretically motivated but is not equivalent to direct LHT measurement. Accordingly, our invocation of LHT should be understood as a theoretical interpretive lens rather than a tested mechanistic explanation. Future research should incorporate direct life history measures alongside attachment assessments to provide a more rigorous test of LHT‐informed predictions about attachment and reproductive outcomes.

A second limitation concerns the measurement of attachment. The four‐item Relationship Questionnaire (Bartholomew and Horowitz 1991) has several known constraints. First, because each attachment style is represented by a single prototype item, internal consistency reliability cannot be evaluated and measurement precision may be lower than that of multi‐item scales. Second, previous research has noted potential overlap among the four attachment prototypes, which may limit discriminant validity and make it difficult to cleanly distinguish between styles. Third, while the original RQ used categorical classification, the current study employed 7‐point Likert‐type ratings to treat attachment as continuous, which partially addresses sensitivity concerns but does not resolve the single‐item limitation. Future research should employ validated multi‐item attachment scales to replicate and extend these findings.

A third limitation concerns the ideal number of children measure. This variable is confounded by respondents' existing parity, as individuals who already have two children are unlikely to report an ideal of fewer than two, meaning the measure partly reflects post hoc rationalisation of actual fertility rather than an independent reproductive goal. Although we include this variable as part of an exploratory analysis and acknowledge its value for understanding the gap between intentions and outcomes, the results should be interpreted with caution. Future studies should control for parity when modelling ideal fertility, or use longitudinal designs that assess ideal family size prior to reproductive outcomes.

In the current study, we found attachment styles, culture, and sex converge in their associations with reproductive success. Secure attachment, while adaptive for resource‐intensive parenting, was associated with lower reproductive success in individualistic cultures. By contrast, insecure attachment styles, particularly fearful and preoccupied orientations, consistently were associated with higher reproductive success across all three nations, pointing to their potential role in faster life history strategies. These patterns varied by culture, with collectivist values in Japan associated with differences in attachment, number of children, and sex differences. Overall, the current research demonstrates the need to contextualise evolution‐informed research on attachment theory within broader cultural frameworks to fully understand its implications for human behaviour and reproductive outcomes.

Author Contributions

Maryanne L. Fisher: conceptualization, investigation, writing – original draft, writing – review and editing. T. Joel Wade: conceptualization, investigation, writing – original draft, writing – review and editing, formal analysis. Hidenori Komatsu: supervision, formal analysis, methodology, writing – original draft, writing – review and editing, funding acquisition, investigation, conceptualization. Nobuyuki Tanaka: supervision, investigation. Hiromi Kubota: supervision, investigation. Rebecca Burch: writing – original draft, writing – review and editing, conceptualization, investigation. Catherine Salmon: conceptualization, investigation, writing – original draft, writing – review and editing. David Widman: investigation, writing – original draft, writing – review and editing, conceptualization.

Ethics Statement

The study was approved by the Ethics Committee of the Strategy and Planning Division within the Grid Innovation Research Laboratory at the Central Research Institute of Electric Power Industry (Approval Number G2022004). This study was conducted in accordance with the Declaration of Helsinki and its later amendments.

Consent

The survey company obtained informed consent from all adult participants in the study.

Conflicts of Interest

The authors declare no conflicts of interest.

Data Availability Statement

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.