Like Father, Like Son: Empirical Insights into the Intergenerational Continuity of Masculinity Ideology

Scholarly work and public commentary point to the persistence of masculinity models characterized by a sense of entitlement, the exertion of dominance, and the justification of abusive behaviors. While there is abundant theoretical work on men and masculinities, fewer empirical studies have examined how young men develop their masculine ideals. In this study, we theorize the role of fathers’ adherence to masculinity ideology in influencing the development of young men’s masculine ideals. We then provide novel empirical evidence on intergenerational congruence between fathers’ and sons’ masculinities using unique data from an Australian national probability survey. Our results reveal moderate, positive associations between fathers’ and sons’ adherence to masculinity ideology. This pattern holds for an overall measure of masculinity, as well as for each of its subscales. Fathers’ religiosity amplified the magnitude of the intergenerational correlation. These findings suggest that interventions aimed at encouraging the development of healthy masculinities amongst young men should engage their paternal figures.

consequences-for example, through domestic violence, sexual misconduct, and aggression (Fleming et al., 2015;Fulu et al., 2013;Heilman et al., 2017;The Men's Project & Flood, 2018). Further, certain masculine norms are also associated with multiple markers of personal risk for the men who embody them-such as inhibited help-seeking behavior, addictions, suicidality, and poor mental health (Heilman et al., 2017;Milner et al., 2018;Seidler et al., 2016;The Men's Project & Flood, 2018). As such, recognizing the determinants that contribute to men's conformity to masculinity ideology is an important endeavor.
While there is abundant theoretical work about men and masculinity within sociology and the social sciences (see e.g., Connell & Messerschmidt, 2005;Gottzén et al. 2019;Schrock & Schwalbe, 2009), fewer empirical studies document how young men develop their masculine ideals (Nielson et al., 2022). Scholarship that uses social surveys and quantitative methods to identify the contributions of specific socializing agents is particularly scarce. Recent studies have considered factors such as individual experiences of rejection, social norms condoning male dominance, the quality of paternal and peer relations, or the relative roles of different socializing agents (Berke & Zeichner, 2016;DeKeseredy & In modern societies, men progressively express and enact their maleness in diverse ways, moving away from traditional models of masculinity (Anderson, 2009). Some emerging social ideals, for instance, encourage men to be caring and empathic, express their feelings and accept their emotional vulnerabilities, and celebrate men's involvement in conventionally female spheres-such as parenting and domestic work. However, scholarly work (Gottzén et al., 2019) and public commentary (Salam, 2019) point to the persistence of models of masculinity characterized by a sense of entitlement emerging from being a man, the exertion of dominance over women and other men, and the justification of abusive and aggressive behaviors. These enactments of masculinity can have profound interpersonal Schwartz, 2005;Heilman et al., 2017;Nielson et al., 2022). The present study contributes novel evidence on a likely central contributor that remains under-researched: the role of paternal adherence to masculinity ideology. In doing so, it adds to a growing body of sociological work documenting intergenerational correlations in cultural orientations and socio-political attitudes. Specifically, we theorize how fathers' adherence to masculinity ideology may affect the development of young men's masculine ideals. We then use unique data from an Australian national probability survey to provide novel empirical evidence on intergenerational correlations between fathers' and sons' masculinities.

Intergenerational Congruence in Socio-Economic Outcomes, Values, and Social Attitudes
Social scientists have long been interested in issues of intergenerational persistence, or the fact that children and their lives tend to resemble their parents and their parents' lives. Indeed, a long-standing body of sociological and economic work has documented intergenerational correlations in objective markers of socially valued outcomes, for example, educational attainment, occupational status, and earnings (Black et al., 2005;Ermisch et al., 2012;Torche, 2015). A parallel literature has addressed intergenerational correlations in subjective dispositions. For instance, studies from sociology, criminology, psychology and political science have identified parent-child associations in personality traits (Dohmen et al., 2012), externalizing behaviors and criminal activity (Besemer et al., 2017), religious practices (Bengtson et al., 2009), political and economic beliefs (Vollebergh et al., 2001), and attitudes to social issues-such as environmental problems (Grønhøj & Thøgersen, 2009) and gender inequality (Perales et al., 2021). The inheritance of social values and cultural orientations from one's parents is a plausible mechanism driving intergenerational correlations in objective behaviors and outcomes.
Social learning theory (Bandura & Walters, 1977) is routinely deployed to explain associations between parents' and children's social values and cultural orientations. This perspective proposes a vertical model of socialization that emphasizes the role of mothers and fathers as core socializing agents. Individuals learn how to interact with others and what constitutes acceptable and unacceptable social behavior by observing and communicating with their parents. This occurs both through role modelling-whereby children observe, internalize and replicate parental behaviors-and through direct teachings-whereby parents communicate to their children values and beliefs that they deem desirable (Trommsdorff, 2009). The bulk of this parental socialization process occurs during childhood and adolescence, when children spend the most time with their parents and are most receptive to social learning. For the formation of cultural values and social attitudes, more specifically, adolescence has been deemed a sensitive or critical period of the life course (Krosnick & Alwin, 1989;Vollebergh et al., 2001). While we acknowledge the plausible role of other socializing agents (e.g., peers or the media, see Nielson et al., 2022), in the current study, we draw upon this perspective to theorize intergenerational correlations in adherence to masculinity ideology between fathers and their sons.

Intergenerational Correlations in Gender-Related Issues
The present study is concerned with conceptualizing and estimating father-son intergenerational correlations in adherence to masculinity ideology. To our knowledge, only one previous study has examined intergenerational correlations in masculinity between fathers and sons. Using data on 113 fathers and their 12-year-old sons from Jyvaskyla (Finland), Huttunen (1992) reported bivariate correlations (r = .23) in fathers' and sons' responses to the Bem Sex Role Inventory (Bem, 1974). This study, however, relies on old and contextually specific data, a small and non-probability sample, and simple bivariate analyses. A cognate and more developed literature has considered father-son correlations in individual behaviors and attitudes that can be conceptualized as being part of masculinity ideology. This includes studies on trust and risk taking (Dohmen et al., 2012), externalizing behaviors (Besemer et al., 2017), sexism (Klann et al., 2018), and domestic violence (Li et al., 2020). These studies, however, are unable to compare intergenerational correlations across multiple dimensions of masculinity ideology, or to establish whether correlations emerge also for broader measures of overall adherence to masculinity ideology.
While research on intergenerational correlations in masculinity ideology is scarce, a cognate and relatively welldeveloped body of scholarly work on intergenerational correlations in gender ideology offers valuable insights. Here, gender ideology is defined as the degree to which individuals endorse traditional views about gender and gender divisions based on the notion of 'separate spheres', with men as 'breadwinners' and women as 'homemakers' (Davis & Greenstein, 2009). As explained below, endorsement of gender ideology overlaps with conceptualizations of masculinity ideology.
Studies conducted in the US (Davis & Wills, 2010;Moen et al., 1997), the UK (Burt & Scott, 2002;Platt & Polavieja, 2016) and a handful of other countries (Dhar et al., 2019;Kulik, 2002), have documented moderate-to-large correlations between parents' and children's gender ideologies. That is, parents who espouse traditional gender beliefs are more likely to have children who also espouse such beliefs. Parent-child correlations in gender attitudes are apparent during adolescence and adulthood, across different dimensions of gender ideology, and for male and female offspring (Burt & Scott, 2002;Dhar et al., 2019;Platt & Polavieja, 2016). While both maternal and paternal gender attitudes are associated with those among children, some studies have documented stronger correlations between same-gender parent-child dyads (Dhar et al., 2019;Kulik, 2002;Perales et al., 2021;Platt & Polavieja, 2016).
Only one previous study has examined intergenerational correlations in gender ideology in Australia, the country where the present study is situated. Using data from the Longitudinal Study of Australian Children, Perales and colleagues (2021) found substantial intergenerational correlations between fathers' and mothers' gender attitudes and those of their 14/15-year-old adolescents. Paternal and maternal attitudes exerted a similar degree of influence on children's attitudes, with some evidence of stronger correlations amongst mother-daughter dyads.
Based on this body of work, we expect the processes contributing to intergenerational associations in gender attitudes to overlap with those that may operate in the transmission of masculinity from fathers to sons. As such, the research findings reviewed in this section constitute plausible a priori expectations for the present study.

Conceptualizing Masculinity Ideology
Gender norms are social norms that define the behaviors, roles, and attributes that are considered acceptable, expected, appropriate and valued for people of different genders within a particular society or cultural context (King et al., 2022), including normative understandings of 'masculinity' and 'femininity'. Simply put, masculinity refers to attitudes and values, behaviors and practices, and ways to approach interpersonal or social relations through which men express and enact their maleness.
A long tradition of sociological scholarship has underscored the existence of different models of masculinity, as well as the fluidity of masculinity at both the intrapersonal and societal levels. The concept of masculinity ideology encapsulates a dominant and idealized form of expressing maleness operating within a given society. Similar terms, such as hegemonic masculinity (Connell & Messerschmidt, 2005), traditional masculinity (The Men's Project & Flood, 2018), or dominant masculinity (Keddie et al., 2022) are also used in the literature, often interchangeably-for a discussion of definitions and approaches, see Smiler (2004) and Smiler and Epstein (2010). Generally speaking, these terms all describe a similar model of masculinity; one that encapsulates and values features such as stoicism and selfreliance, dominance over women and other men, risk-taking, heterosexuality, social status, and the endorsement of violence as a solution to problems (Mahalik et al., 2003).
Despite the increasing social acceptance of more positive and inclusive masculinities (Anderson & McCormack, 2016;Englar-Carlson & Kiselica, 2013) and the fact that some elements of traditional masculinity can be protective (Addis et al., 2016), many facets of masculinity ideology have been empirically connected to negative outcomes. This includes outcomes at the individual level (e.g., poor mental health and suicidality amongst men) and interpersonal level (e.g., domestic violence) (Fleming et al., 2015;Fulu et al., 2013;Heilman et al., 2017;The Men's Project & Flood, 2018).

Fathers, Sons, and the Transmission of Masculinity
Gender is a highly salient stratifying category and children learn about gender identity and gendered societal expectations from a very young age (Blakemore, 2003;Eagly, 2013). This means that, from the early years, boys are exposed to, and engage with, messaging about how to 'do gender' and enact masculinities that are perceived to be normative or privileged. Although this messaging comes from a range of social actors and institutions (Nielson et al., 2022), fathers occupy a prime position to-actively or passively-pass information and teachings about what constitutes appropriate masculine behaviors and attitudes onto their sons (see e.g., Kane, 2006). Given the salience of gender as a core aspect of identity, research has shown that fathers tend to identify more strongly with their sons compared to their daughters (Andreas et al., 2018;Cichy et al., 2007;Raley & Bianchi, 2006). As a result, fathers tend to make themselves more available to their sons and become more involved in their lives, creating opportunities to teach sons about 'desirable' gendered attitudes and behaviors (Perales et al., 2021). Sons, in turn, may look up to their fathers as aspirational male role models to emulate, or may feel pressure from their fathers-either real or imagined-to adopt particular gendered behaviors (Schroeder & Liben, 2021). Indeed, in comparison to mothers, research shows that fathers are less comfortable with their children not conforming to gender stereotypes (Our Watch, 2018), and make greater efforts to enforce gender stereotypes and boundaries with their sons (Kane, 2006).
Other factors may also influence the transferability of attitudes and values from parents to children. For example, research has documented how adolescents are more likely to accept parental socialization for values falling within the conventional domain (i.e., those governed by arbitrary social norms) or moral domain (i.e., those pertaining to others' rights or welfare) compared to the personal domain (i.e., those whose consequences affect only the actor) (Knafo & Schwartz, 2009). Likewise, intergenerational transmission is deemed easier if children perceive their parents as the best source of information on that domain and thus as 'legitimate transmitters' (Pinquart & Silbereisen, 2004). In our application, this suggests that conformity to masculine norms pertaining to issues where children don't see their parents as being the best source of information-for example, those related to sex and sexuality-may be subject to weaker transmission processes.
In addition, existing research on the transmission of social values suggests that the intergenerational transmission of masculinity may not occur uniformly across population groups. Rather, it may be facilitated or constrained by socio-demographic factors, including the characteristics of children, parents, and their social environment. As Schönpflug (2001, p. 174) puts it, "transmission may be enhanced by "transmission belts," that is, conditions favorable for transmission in a particular socioeconomic and cultural context, such as personal characteristics of the transmitter and the receiver (resources of education and age), and family interaction variables (parenting styles and parents' marital relationship)". In relation to masculinity, it is possible that fathers who are embedded in different social environments-for instance, due to their education levels or religious beliefs-place a premium on intergenerational continuity vs. discontinuity, and on the persistence of traditional values rather than their erasure. For example, given their commitment to a set of pre-determined moral principles and values, religious fathers may have a stronger predisposition towards intergenerational continuity and strive to transmit traditional models of masculinity to their children (Thornton et al., 1983). Other factors associated with less traditional worldviews, such as father's education, may exert the opposite effect.
Overall, the discussions in this section warn against making assumptions of uniformity in the intergenerational transmission of masculinity (see also Nielson et al., 2022). This underscores the need to examine potential heterogeneity in transmission processes across both dimensions of masculinity and socio-demographic groups. Based on this review of the literature, we expect to observe (i) significant intergenerational correlations in masculinity ideology, and (ii) some degree of heterogeneity in these correlations by masculinity domains and child and father characteristics. In the next Consistent with these notions, existing research demonstrates the impact of fathers' socializing practices on their children. For instance, evidence indicates that fathers who endorse different models of masculinity expose their children to these through their parenting practices, with fathers who adhere more closely to principles of masculinity ideology being more distant and punitive (Petts et al., 2018). Further, as previously noted, studies on the intergenerational correlation of attitudes and dispositions have documented strong correlations in how fathers and their children think and behave, including in issues pertaining to gender relations (Perales et al., 2021). The role of fathers as socializing agents to their sons has also been observed in qualitative studies concerned with religious values (Arweck & Penny, 2015), sexuality and sex education (Jerves et al., 2014;Thomas et al., 2022), and how fathers construct their own masculinity through their engagement with their children (Cooper, 2000;Della Puppa & Miele, 2015).
Altogether, based on these theoretical tenets and previous studies focusing on the transmission of gender attitudes, we expect to observe significant father-child correlations in conformity to masculine norms.

Heterogeneity in the Intergenerational Transmission of Masculinity
As explained before, masculinity is not a monolithic construct. Rather, it is best conceptualized as a set of values, behaviors, and worldviews. This poses the question of whether all of the dimensions that comprise the construct of masculinity ideology are intergenerationally transmitted from fathers to sons, and whether the transmission process is stronger for some masculinity domains than others.
Studies on the transmission of cultural values point to several factors pertaining to the content of values that can facilitate or inhibit intergenerational congruence. For example, values are less likely to be intergenerationally reproduced if parents are not motivated to socialize their children into those values or explicitly choose not to do so, or if their children actively resist the transmission process (Trommsdorff, 2009). These situations may occur disproportionately for parental values that have become dysfunctional in children's social environments, for instance, due to intergenerational discontinuity in social contexts or historical changes in social norms. Following this reasoning, it is possible that dimensions of masculinity that have become less socially acceptable over the past few decades may be less likely to be intergenerationally reproduced (e.g., the sanctioned use of violence, the expectation of heterosexuality, or status beliefs about male superiority).
young men who co-resided with their father or a father figure. The data do not allow identifying the kinship relations between the young men and their co-resident father figures. While in most cases these are father and biological son, in others they may be stepfather and non-biological son, brothers with an unusually large age difference, or grandfather and grandson. We argue that-irrespective of their specific kinship bonds-these substantially older co-resident males are likely to pass on their masculinity dispositions through direct interactions and indirect role modelling to the young men with whom they co-reside.
The following rules were used to identify the subsample of interest. First, we excluded cases in which respondents had missing data on their household identifier (n = 2), age (n = 325) or on a large portion of the masculinity questions (n = 262, details below)-as these are core, analytic variables. Second, we identified households in which there was at least one young man (age 15-20 years) and one father figure (age 30-55 years). These age ranges were imposed to constrain the young people to belong to the same survey cohort, and to exclude from the analyses mature-age males co-residing with older males (e.g., a 35-year-old man living with his father or father-in-law). A few cases in which the age difference between the identified young person and the in-scope father figure was smaller than 18 years were also excluded (n = 13). This decision was motivated by the fact that these are likely brother-brother pairings and it is questionable whether the processes of masculinity transmission described before operate similarly between siblings. Five in-scope households with young men contained two possible father figures (i.e., two males over 30 years of age). Upon closer inspection, these appeared to be multigenerational households with three generations of males. In these cases, the respondent that appeared to be the young man's father (rather than their grandfather) was used in the analyses as the father figure.
This sample selection approach yielded a final analytic sample of 839 young men and 743 father figures. Their respective age distributions are shown in Figure A1 in the Online Supplementary Materials. The data were subsequently re-arranged in a dyadic format for analysis, such that the responses from each young male and their father figure were contained within a single data row. Of the 743 in-scope households, 651 (or 87.6%) contained one young man, 88 (or 11.8%) contained two young men, and 4 (or 0.5%) contained three young men.

Measuring Masculinity: The Conformity to Masculine Norms Inventory
To measure masculinity ideology amongst both young men and father figures, we use the Conformity to Masculine section, we introduce the data and methods that we use to examine these propositions.

Method
This study was granted approval by the Human Ethics Committee at The University of Queensland.

Dataset
This study uses data from Ten to Men: The Australian Longitudinal Study on Male Health (Currier et al., 2016;. These data are internationally unique in that they collect detailed information on the masculinity dispositions of young men and their co-resident father figures. Ten to Men is Australia's first national longitudinal study to focus exclusively on male health and wellbeing, following a cohort of ~ 16,000 males aged 10-55 years at baseline. The baseline survey took place in 2013/2014, with a second and third wave of data being collected in 2015/2016 and 2019/2020 respectively. The Ten to Men sample was identified using a stratified, multi-stage, cluster random sampling design and is largely representative of the Australian population of males in the target ages (Bandara et al., 2019). The sampling unit was the household, with all males within the target age range who satisfied the eligibility criteria (i.e., being Australian citizens/permanent residents and having sufficient Englishlanguage skills to complete the survey) being invited to participate in the study. The initial response rate was ~ 35%, and ~ 75% of participants were retained in Wave 2 of the study.
Ten to Men collects rich data on multiple aspects of male health and wellbeing, including physical and mental health, use of health services, health-related behaviors, health knowledge, and social determinants of health (Bandara et al., 2019). Given the large age range of the cohort, the study administered three different-although often overlappingsurvey instruments to boys (ages 10-14 years, n = 1,099), young men (ages 15-17 years, n = 1,026), and adults (18-55 years, n = 13,896). Data from boys were collected through a computer-assisted personal interview, whereas data from young men and adults were collected using a self-completed hard-copy questionnaire.

Sample Selection
We use data from Wave 1 of Ten to Men-when both young men and adults (but not boys) were asked questions about masculinity. Hence, the focus here is on males aged 15-55 observed in 2013/2014. Specifically, we are interested in the young men, the father figures, and their households; β 2 to β 4 are the respective vectors of model coefficients; and ε is the usual stochastic regression error term. The standard errors in all models were adjusted for the clustering of observations within households (Abadie et al., 2017). This course of action is necessary to ensure that the regression assumption of independence of observations is not violated by the fact that some households in the data contain two (n = 88) or three (n = 4) children.
The model controls include variables known or suspected to be correlated with-and to be causally prior to-the masculinity scores of young men and father figures. For the young men, the controls include age (in single years) and identifying as heterosexual (yes/no). For the father figures, they include age (in single years), a continuous-level measure of religiosity/spirituality (on a scale from 1 'Not important at all' to 5 'Extremely important'), partnership status (partnered/not partnered), being a father to more than one child (yes/no), and education level (university qualification/ professional diploma or certificate qualification/school Year 12/below school Year 12). For their common household, the model controls for income group (high: >= $150,000/ medium: $100,000-$149,999/low: < $100,000), residence in a major city (yes/no), and English being the main language spoken at home (yes/no).
There was a small amount of missing data on the control variables (n = 86, ~ 10.3%), particularly for the measures of household income (8%) and religiosity (1.8%). This was dealt with through multiple imputation by chained equations using Stata 17's mi package and 20 imputed datasets (Royston & White, 2011). Sample means and standard deviations for the core analytic variables are presented in Table 1. Those for specific CMNI-22 items and subscales are shown in Table S1 in the online supplement.

Intergenerational Correlations in Masculinity Ideology
The results of the main model specifications are shown in Table 2. The model coefficients give the estimated effect of a one-unit increase in the explanatory variables on the outcome variable (i.e., young men's CNMI scores, on a scale from 0 to 3). The key model coefficient is that on fathers' CNMI scores (range: 0.3), with positive coefficients denoting intergenerational continuity in masculinity ideology.
The estimates from the unadjusted model in Column 1 indicate that the masculinity scores of the father figures were positively and significantly correlated with the masculinity scores of the young men. A one-unit increase in Norms Inventory (Mahalik et al., 2003). This inventory has been shown to have desirable statistical properties, including strong internal consistency, good differential validity, good external validity, and high test-retest estimates across men of different ages (Kivisalu et al., 2015;Mahalik et al., 2003;Smiler, 2006). In this study, we use the CMNI-22 included in the Ten to Men dataset. This is a shortened version of the CMNI comprising 22 items grouped into 11 subscales capturing key dimensions of masculinity ideology: primacy of work, importance of being dominant, endorsement of risk-taking, salience of heterosexual presentation, salience of power over women, importance of emotional control, salience of playboy status, importance of social status, importance of winning, endorsement of violence as a solution to problems, and importance of self-reliance. The reliability of the CMNI-22 in the Ten to Men data was satisfactory (Cronbach's α = 0.65), with the measure having been used in recent studies (see e.g., Milner et al., 2018;Milner et al., 2019;. An overall measure of masculinity ideology for young men and father figures-the CMNI total score-was created by averaging their scores on the 22 items that comprise the CMNI-22, reverse coding items as appropriate so that higher scores always denote stronger adherence to masculinity ideology. The resulting measure ranges from 0 (lowest conformity to masculinity ideology) to 3 (highest conformity). In creating this composite variable, we only included respondents who provided valid answers to at least 15 of the 22 questions-or 99.86% of those who answered at least one question. Consistent with earlier research-see e.g., Wade's (2015) commentary-fathers and sons in our recent Australian sample both have CMNI total scores below (but close to) the scale's midpoint of 1.5: 1.32 units (SD = 0.26 units) for the young men and 1.32 units (SD = 0.26 units) for the father figures. The respective variable distributions are shown in Figure A2 in the Online Supplementary Materials (1.23 for fathers and 1.32 for sons).

Modelling Approach
We estimate correlations between young men's and father figures' masculinity ideology scores via the following ordinary least squares (OLS) regression model: where the y, f, and h subscripts stand for 'young man', 'father figure', and 'household', respectively; CMNI is a measure of overall conformity to masculine norms; α is the model's intercept; β 1 is the key parameter of interest capturing the intergenerational correlation in masculinity; the Xs are sets of control variables for observable characteristics of (β = 0.011, p < .10), and young men whose fathers had more than one child (β = 0.068, p < .10), and weaker amongst young men from English-speaking than non-Englishspeaking backgrounds (β= -0.055, p < .10), and young men with fathers who had professional (β = − 0.059, p < .05) or university (β = − 0.090, p < .01) qualifications, compared to less than school Year 12 education. Young men's age, fathers' age, household income, residence in a major city, and fathers' partnership status did not significantly predict young men's masculinity.

Subscale-Specific Correlations
To complement the results discussed in the previous section, we fitted regression models estimating intergenerational correlations for each of the 11 subscales that comprise the CMNI-22. These subscales are calculated by averaging two CMNI-22 items and therefore range from 0 to 3. These models are informative as to which domains of masculinity are more and less strongly correlated across generations and are summarized in Fig. 1 and Table S2 in the online supplement. In Fig. 1, the variables are ordered by the magnitude of the intergenerational correlation.
Significant adjusted associations between the masculinity scores of the father figures and the young men were identified for all 11 CMNI-22 subscales. However, the effect magnitude differed across subscales. It was largest for subscales capturing playboy status, heterosexual presentation, and endorsement of violence, and weakest for those capturing self-reliance, emotional control, social status, and power over women.
For readers interested in intergenerational associations across more nuanced attitudes and behaviors, similar analyses were performed on the individual CMNI items (n = 22, range: 0-3) and presented in Figure S3 and Table S2 in the online supplement. Significant adjusted associations between the masculinity scores of the father figures and the young men were identified for 20 of the 22 items that comprise the CMNI-22. The strongest associations were observed for statements concerning violent action being necessary, importance of heterosexual presentation, and desirability of having multiple sexual partners. On the other hand, the weakest correlations were found for statements capturing help-seeking behavior, justification of violence, respect for women, and primacy of social status.

Moderators of Father-Son Correlations in Adherence to Masculinity Ideology
In a final set of analyses, we explored whether the correlation between the masculinity ideology of the young men and their father figures was weaker or stronger at different fathers' masculinity ideology-on a scale from 0 to 3-was associated with a 0.23 increase in young men's conformity (β = 0.229, p < .01). The strength of the association is therefore moderate, with the Cohen's d associated with a onestandard-deviation increase in fathers' CMNI scores being 0.23.
The results from the adjusted model are presented in Column 2. The addition of the control variables to the model did not change the overall pattern of results: all else being equal, fathers' masculinity-ideology scores remained positively and significantly associated with young men's masculinity scores (β = 0.208, p < .01). This indicates that the raw associations observed in the unadjusted model are not the product of confounding by the model covariates. Altogether, these results are consistent with theoretical expectations suggesting that masculinity ideology is reproduced intergenerationally. Fathers who adhere to normative principles of traditional masculinity tend to have sons who also do so.
While not the focus of the present study, the coefficients on the model covariates are informative about other factors associated with young men's masculinity ideology. These were largely consistent with expectations. Conformity to masculine norms was stronger amongst heterosexual than non-heterosexual young men (β = 0.111, p < .001), young men whose father figures reported higher religiosity 17.0 1.7 Heterosexual, % (young man)) 86.9 Importance of religion/spirituality (father figure) 2

Discussion
Consistent with expectations, our analyses yielded significant associations between fathers' and sons' degree of adherence to masculinity ideology. Such associations were moderate in magnitude and persisted in the presence of an encompassing set of control variables capturing different parent, child, and family characteristics.

Discussion of Key Findings
Our findings are consistent with social learning theory and other perspectives that emphasize the role of fathers as socializing agents to boys and young men (Perales et al., values of the control variables. To do so, we estimated an OLS model of young men's total CMNI scores in which each of the control variables was interacted with the focal explanatory variable capturing fathers' total CMNI scores. Results from these models indicated that only one such interaction effect was (marginally) statistically significant, namely that for the variable capturing the importance of religion to the father (see Table S3 in the online supplement). The model estimates for this interaction are visualized in Fig. 2, which shows a stronger intergenerational correlation in masculinity ideology when fathers are more (compared to less) religious.

Model
(1) (2) CMNI total score (father figure) 0   continuity across population groups defined by their sociodemographic characteristics. For example, the degree to which fathers and sons had similar masculinity ideology was not affected by whether fathers had higher or lower education levels, their sons were heterosexual or not, or they both lived in urban or rural areas. Parental religiosity was however an exception, being the only variable that significantly-albeit only marginally-moderated the intergenerational correlations. Specifically, fathers for whom religion was important to their lives were more likely to pass on their masculinity onto their sons. Given that these fathers report stronger conformity to masculine norms than non-religious fathers, this means that their children are also more likely to adhere to such norms. This finding is consistent with those of earlier studies focusing on other types of dispositions, such as gender attitudes (Thornton et al., 1983), and may reflect a greater motivation on the part of religious fathers to pass on their worldviews onto their children. The intricate connections between religiosity, tradition, and parenting may explain these patterns (Arweck & Penny, 2015). More broadly, this finding adds to a body of work on 'complex religion' pointing to the potential for religion to not only to exert direct effects on individuals' socio-political attitudes, but also to modify the effects of other socio-demographic factors on such attitudes (Perales et al., 2019).

Limitations and Future Research Directions
Some analytic and scope limitations must be borne in mind when interpreting our findings. First, there were issues stemming from the properties of the Ten to Men data. These included the absence of direct information on kinship relationships between sons and father figures, and a relatively modest reliability score on the items comprising the CMNI total score. We also note that the CMNI is one of many ways of operationalizing masculinity. Additionally, the data only allowed us to consider intergenerational correlations in masculinity for young men who co-resided with their fathers. Research shows that a father's residence (or lack thereof) in the family home has implications for sons' gender-role development (see e.g., Mandara et al., 2005), suggesting that the father-to-son transmission of masculinity may operate differently in non-co-residential families (see Marcell et al., 2011). Further, the data only capture information on masculine norms at a single time point, precluding longitudinal analyses. Repeated measurement of masculinity ideology would help ascertain whether sons' attitudes continue to resemble their fathers' as they grow older and enter new life-course stages. Future studies should replicate our findings using datasets that are unaffected by these data-driven limitations.
2021 ;Schroeder & Liben, 2021). They also align with findings from an earlier Finnish study by Huttunen (1992) and with previous studies documenting father-son intergenerational correlations in moral values and socio-political attitudes (see e.g., Vollebergh et al., 2001), gender ideology (see e.g., Davis & Wills, 2010), sexism (Klann et al., 2018), and individual personality traits associated with masculinity-such as risk-taking or use of violence (see e.g., Dohmen et al., 2012). In fact, the unadjusted association between fathers' and sons' conformity to masculine norms observed here (β = 0.23) is comparable in magnitude to the association between fathers' and sons' gender-related attitudes reported by a recent Australian study (β = 0.18; Perales et al., 2021). Notwithstanding, while important, these correlations are not exceedingly large (Cohen's d = 0.23). This finding is consistent with the long-standing observation that "intergenerational discontinuities are considerably stronger than intergenerational continuities" (Rutter, 1998(Rutter, , p. 1270, illustrating the potential for disruption in familial cycles that foster psychosocial risk. Inspired by previous studies focusing on social values, our analyses moved beyond the estimation of overall patterns and probed into potential heterogeneity in intergenerational correlations in masculinity ideology. We considered potential disparities across both the specific dimensions being transmitted and the characteristics of the actors implicated in the transmission process. Concerning dimension-based heterogeneity, significant intergenerational associations were observed for all 11 subscales of the CMNI-22. These results demonstrate that the intergenerational correlations observed for the CMNI total score are not driven by a handful of indicators but are rather indicative of broader transmission processes that applies to multiple behaviors and attitudes. There were nevertheless some disparities in the strength of the father-son associations across subscales. For example, associations appeared stronger for dimensions of masculinity regarding playboy status, heterosexual presentation, and endorsement of violence, and weaker for dimensions regarding self-reliance, emotional control, and social status. These findings appear inconsistent with the claim that masculine norms that are becoming less tolerated, or that may be more 'taboo' and hence feature less in father-son exchanges, are less likely to be transmitted. One possible explanation is that certain domains (e.g., being perceived as heterosexual) may be seen as less relevant or less important for the younger cohort or are more open to influence from peers or the media than fathers (Nielson et al., 2022). Alternative ways of conceptualizing dimension-heterogeneity in the transmission of masculinity ideology are thus needed.
Concerning group-based heterogeneity, our results painted a picture of uniformity in the degree of intergenerational

Conclusion
Using Australian national survey data, this study has generated first-time evidence on the intergenerational continuity of masculinity ideology. The results underscore the relevance of paternal conformity to masculine norms as a factor contributing to the persistence of masculinity ideology amongst young men. These findings point to useful avenues to promote the development of alternative masculinity models that reduce the burden of unhealthy masculinities within contemporary societies.

Acknowledgements
The authors would like to thank Ms Heidi Hoffmann for excellent research assistance.
Author Contribution Francisco Perales led the study conception and design, with input from Michael Flood, Tania King, and Ella Kuskoff. Material preparation, data collection and analysis were performed by Francisco Perales. The first draft of the manuscript was led by Francisco Perales with contributions by all authors, and all authors commented on and edited previous versions of the manuscript. All authors read and approved the final manuscript.
Funding Open Access funding enabled and organized by CAUL and its Member Institutions. The research was supported by the Australian Research Council Centre of Excellence for Children and Families over the Life Course (project number CE200100025).
Data Availability This paper uses data from Ten to Men: The Australian Longitudinal Study on Male Health. The Ten to Men project is funded by the Australian Government Department of Health and Ageing (DHA) and is managed by the Australian Institute of Family Studies (AIFS). The findings and views reported in this paper, however, are those of the author and should not be attributed to either DHA or AIFS. Syntax code to replicate the analysis can be made available by the authors upon request.

Declarations
Ethical Approval and Consent to participate Informed consent from participants was obtained from AIFS upon collection of the Ten to Men data.

Human and Animal Ethics
The study was granted approval by the Human Ethics Committee at The University of Queensland (2020000178).

Consent for Publication
The publication of this manuscript has been approved by all co-authors.

Competing Interests
The authors have no conflicts of interest, competing interests, and/or financial or non-financial interests to declare.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the Second, we are unable to identify the specific mechanisms that may underpin the observed father-son associations in masculinity ideology. Here, we have adopted a similar framework as earlier studies and theorized parental transmission (e.g., through role modelling or direct teachings) as a likely pathway. However, we acknowledge that other factors that remain unaccounted-or only partially accounted for-in our analyses are also likely to play a role. These may include male peers and media channels (Nielson et al., 2022), shared social environments subjecting fathers and sons to similar external socializing influences (Trommsdorff, 2009), and bidirectional processes whereby young men also shift their fathers' expressions of masculinity (Della Puppa & Miele, 2015). Further research unpacking the relative contributions of these other processes is warranted.

Practice Implications
Despite these limitations, the findings of this study bear implications for theory, policy and practice. At a basic level, the results confirm a theoretical premise that has long been assumed-often uncritically-in social-science research, namely that young men's enactments of particular forms of masculinity resemble those of their fathers. More speculatively, our findings suggest that fathers' own masculinity ideology is a likely contributor to their sons' expressions of masculinity. This simple yet important observation has implications for preventive or remedial programs aimed at fostering healthy masculinities amongst young men. Chiefly, it suggests that any such interventions may be more likely to be successful if they involve, or also target, fathers.
This suggestion is consistent with a growing body of programs focused on engaging fathers in positive parenting and father-child relationship interventions (Henry et al., 2020;King et al., 2019). The modest degree of domain or group heterogeneity in our results further suggests that such programs may work for individual masculinity traits or specific population groups. At the same time, the intergenerational associations in father-child masculinities that we observe underscore the potential long-term effects of successful intervention. If a program manages to help young people develop positive masculinities, it is likely that-as they themselves become fathers-their own children's masculinities are also positively affected. Further research identifying practical ways to foster healthier masculinities is warranted. source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons. org/licenses/by/4.0/.