Keywords

Family relationships are acknowledged as a primary context for the social and cognitive development of the child (Walsh 2012). Family is a construct that can encompass a wide variety of relational systems; for the purpose of this chapter, we focus on the mother-father-child triad as the first-level unit beyond the dyad that can be qualified as a family-level system. The studies dedicated to the relational dynamic within the triad have shown that cohesion in the coparental relationship (the part of their relationship concerning the child) and parental adjustment to the child’s characteristics (in particular, the child’s temperament ) are linked with positive outcomes in children, such as the ability to understand multiple perspectives and the development of positive cognitions about social relationships (Cummings and Davies 2010; Favez et al. 2012; Raikes and Thompson 2006). On the other hand, conflict and tensions between parents and the spillover of negative emotions from their relationship to parent-child relationships are predictive of various maladapted outcomes in children, such as anxiety or social withdrawal (Favez et al. 2006a; McHale and Lindahl 2011; Teubert and Pinquart 2010).

Numerous studies have aimed to identify the predictors of the quality of early family relationships, with a focus on the coparental relationship, theoretically considered as the core relational process for family functioning (Minuchin 1974). From the family life cycle standpoint, these predictors are seen as influencing the construction of family relations across time; that is, they have to be understood from a developmental perspective (McHale 2007). In this approach, longitudinal studies that start during pregnancy are of primary interest, in particular, because they allow one to disentangle the factors of influence pertaining to the parents and their relationship from the factors pertaining to the child, a much harder distinction to make after birth . Studies in this domain have highlighted different types of variables of influence at both the level of representations and the level of interactions. For example, prenatal interactions between parents that simulate a first encounter with their baby-to-be have been shown to be predictive of postnatal mother-father-infant interactions: In particular, coparental coordination and mutual support – or the absence thereof – can already be observed in the second trimester of pregnancy (e.g., Cairo et al. 2012; Carneiro et al. 2006; Favez et al. 2006b). On the other hand, recent studies have shown that parents’ representations of the quality of their couple relationship, their future family functioning, and coparenting collaboration are predictive of their effective coparental coordination after birth (Favez et al. 2013; Kuersten-Hogan 2017; McHale and Rotman 2007). These links are similar to those that have been observed between the representations that the parents have of their future baby during pregnancy and their parenting behavior after the baby is born (Stern 1991; Theran et al. 2005).

In this chapter, we specifically focus on a factor that has been rarely considered in studies on the transition to parenthood: parents’ gender-role orientation during pregnancy and its influence on the construction of coparental and family interactions. Gender roles are defined as “pre-determined schemas to which men and women were expected to adhere” (Donnelly and Twenge 2017, p. 556). It is well known that representations of gender roles are intimately linked to the way that each parent envisions the tasks that she or he is ready to assume and which tasks are considered to fall under the other parent’s responsibility and expertise (Goldscheider et al. 2015; Katz-Wise et al. 2010; Pape Cowan and Cowan 1992). Gender-role orientation should thus be one of the background variables that foreshadows the engagement of each parent and, in turn, coparental organization as early as during pregnancy. The gender role is not determined by the biological sex of the individual; both constructs have been diversely used in family theories according to the time period and schools of thought.

Biologically Determined Roles Versus Systemic Functions

The biological sex of the parent was traditionally considered as paramount in family functioning and in the development of the child by psychodynamic theories. For example, from a Freudian perspective (e.g., Freud 1924), children have to identify with the parent of the same sex and to differentiate themselves from the parent of the other sex in order to construct a gender identity that is congruent with their biological sex. Each parent was expected to play a predetermined and specialized role (not labeled this way in these theories) that was congruent with social expectations in two separated spheres: Women were considered to be naturally (i.e., biologically) drawn to taking care of children and providing affection, so that they had to assume family duties, whereas men were considered to be naturally determined to provide resources for the family and to ensure discipline in education, so that they had to be engaged in the workforce (Lamb and Lewis 2010; Perälä-Littunen 2007). These theories, while representing a window on family organization in the Western world at the beginning of the twentieth century, had the consequence of crystallizing the view of the family. Despite this configuration starting out as descriptive and representative of a time period, it became prescriptive, so that any deviation from this model was considered, in essence, dysfunctional. Anthropology has, however, long shown that this organization was neither universal nor essential (Harrell 1997).

The systemic approach considered a different view: The emphasis here was on equifinality, defined as the possibility of using different means to achieve a goal, and on functional roles, defined as a set of tasks necessary to the functioning of the whole system (Cox and Paley 1997). In the case of parenting, these tasks are to provide affection, protection, care, and education to the children; the sex of the parent is irrelevant, as long as the roles are fulfilled. Both parents – any adult in fact – can thus interchangeably take on parenting duties. Different types of families should thus be able to accomplish the functions of a family system and to provide a loving and nurturing context for children. This theoretical stance has been amply validated by empirical studies that have shown that in contemporaneous families, both parents – and any adult, irrespective of biological sex – are able to love, protect, and educate children and that parenting and coparenting are concepts that may be used for any adult team in charge of a child (Biblarz and Stacey 2010; Dufur et al. 2010; Hook and Chalasani 2008; McHale et al. 2002). However, the downside of this perspective was that it overlooked the fact that, even if mothers and fathers can both fulfill family roles, they may do so in different manners, even in the most egalitarian families. Studies have, for example, shown that parental interactive behaviors with an infant are notably different in women and in men, with a higher prevalence of vocal stimulations in mothers and of physical games in fathers (Nordahl et al. 2014; Power 1985). But how is it possible to be both similar and different? Gender-role theory is a perspective that takes into account both equifinality and specialization at the same time in order to explain these apparent contradictions.

The Gender Revolution and Gender-Role Theory

In the second half of the twentieth century, a gender revolution took place, and the idea of a possible interchangeability and sharing of tasks between parents came progressively to the fore in Western societies, even though the idea of a natural aptitude to accomplish certain roles still remains strong today. Women increasingly participated in the labor force and men increasingly participated in family work (Goldscheider et al. 2015). Studies in the 1960s and 1970s showed that mothers may be involved at work and fathers the primary caregivers without the child being harmed (Booth 1992; Favez et al. 2018). There was thus a shift from the conceptualization of task sharing based on biologically determined traits to the idea of tasks being determined according to socially determined gender roles. According to socio-constructivist theories, these roles are mainly determined by social norms and depend on a given era and cultural context . Such norms are thus likely to change and evolve through time. In Western societies, some qualities have historically been considered as masculine (such as instrumentality, autonomy, competitiveness, and productivity) and others as feminine (such as expressivity, orientation toward interpersonal relationships, dependency, and communication). Regarding parental roles, these constructs are congruent with the representations of fathers as being oriented toward pragmatism and discipline , and mothers as being oriented toward love and affection, the representation of feminine qualities being deeply rooted in the way laypeople assess the qualities needed for parenting (Hoffman and Moon 2000).

The gender schema theory (Bem 1974) is an emblematic example of a new way, in the 1970s, to conceptualize roles. According to this theory, any individual may have a masculine or a feminine gender-role orientation, or even one that is androgynous (both feminine and masculine, which are compatible with one another) or undifferentiated (having neither feminine nor masculine qualities). Masculinity and femininity are thus considered traits of personality. Flexibility is a central feature of this theory: being able to endorse both feminine and masculine roles without experiencing a lowering of self-esteem – an aptitude that is especially pronounced in androgynous personalities – allows for a richer personal and interpersonal experience and greater adaptation skills to life circumstances. In contrast, a strict masculine orientation in men or a strict feminine orientation in women restricts the behaviors that individuals may implement and experience (Bem 1981b; Bem et al. 1976). In this model, masculinity is defined as having dominant, assertive, and instrumental dispositions, while femininity is defined as having nurturing and expressive dispositions. Although this theory was proposed during the 1970s, studies have shown that identification with these gendered personality traits, as defined by Bem (1974), is still strong today; in a meta-analysis that included 34 studies completed between 1993 and 2012, the only difference found over time was that in recent years women tended to identify less with feminine traits (Donnelly and Twenge 2017).

Gender-Role Orientation and Parenting

A few studies have been specifically dedicated to examining whether orientations toward masculinity or femininity were linked with parenting. They have, for example, shown that in any given individual, an association of low femininity with high masculinity is less favorable to the implementation of parenting behaviors (Sanderson and Thompson 2002). In a study in Switzerland, our research group found that fathers who are involved in domestic and parenting tasks are less likely to have a strict masculine orientation than are fathers who are not (or who are less) involved in family life. Engaged fathers with a high masculine orientation also have a high feminine orientation; that is, they tend to match the androgynous profile described in the gender-schema theory. A high feminine orientation in fathers seems thus to be associated with a better aptitude for taking care of children (which is not incompatible with also having a high masculine orientation at the same time; Frascarolo et al. 1996).

Whereas gender-role orientation was not their main focus, studies on maternal “gatekeeping” have uncovered how social expectations influence parenting and coparenting. Indeed, it has been shown that a mother may inadvertently or purposefully engage in behaviors which restrict the father’s access to the child, or in any case lower the control the father can exert on child-related daily tasks. These behaviors have been tagged as “negative gatekeeping behaviors”; as a consequence of this form of maternal gatekeeping, the father may disengage himself from parenting or even from family life. This dynamic is thus unfavorable to coparenting cooperation. Gatekeeping behaviors are motivated, consciously or not, by maternal beliefs about mothers’ and fathers’ roles in the family; these beliefs are operating as early as pregnancy – and certainly even before (Allen and Hawkins 1999; Van Egeren 2003, 2004). According to traditional social representations, femininity is at the core of a mother’s role; thus, the higher the mother’s identification with feminine values, the more she might be likely to enact gatekeeping behaviors, at the expense of coparenting cooperation. However, to date, no study has been specifically dedicated to examine the links between gender-role orientation and coparenting.

Differences in Parenting Behavior According to the Sex of the Child

Finally, hints of the effects of gender might be found in studies that have shown that parenting practices vary according to the sex of the child. For instance, parents tend to educate their children according to gender stereotypes (Clearfield and Nelson 2006; Laflamme et al. 2002; Paquette et al. 2003). Coparental interactions also seem to be influenced by the expectations that parents have related to the sex of the child. In a study in Switzerland, we found that worsening of coparental interactions through the first year in primiparous families was observed only when the child was a boy (Favez et al. 2006b). In another US study, McHale et al. (2002) found that in families of boys, a tense relationship led to more conflictive coparenting, whereas in families of girls, it led to more skewed coparenting, with one parent, usually the father, withdrawing from family life. One hypothesis offered to explain this difference was that, because having a male child tends to be more socially valued, both parents stay engaged despite their mutual resentment; as a result, daily conflicts are more likely to happen. In all cases, expectations associated with the sex of the child and gendered representations explain the relational processes operating in these families, at least during the postpartum period. To date, no study has examined the role of the expected gender of the child-to-come with respect to prenatal coparental interactions or prenatal family dynamics.

What Should Be Concluded?

From a socio-constructivist perspective, gender revolution will be achieved following changes in individuals’ personality characteristics; that is, men preparing to take care of their baby should develop more feminine traits and women preparing to enter the workforce should develop more masculine traits (Eagly et al. 2000). At that moment, a shift will occur between the traditionally specialized roles, supposedly determined by biological sex, and contemporaneous roles that rely on equifinality: Both parents, or any adult, will be able to implement the feminine and masculine qualities necessary to raise and educate children.

For the time being, we are in an intermediate phase, with both traditional and contemporaneous contradictory forces exerting their influences on parents and families to varying degrees; as early as pregnancy, we expect parenting as well as coparenting to be influenced by the extent to which parents adhere to traditional versus more contemporary roles and by their gender-role orientation (Katz-Wise et al. 2010; Knudson-Martin 2012; Koivunen et al. 2009). There is evidence that couples’ more contemporaneous views during pregnancy often give way to more traditional views in the early postpartum period. Pape Cowan et al. (1985) discovered in their Becoming a Family Project that many men and women who held egalitarian views of parental roles during pregnancy adopted more gender-stereotypical attitudes and role divisions after the birth of their first child. These violations of couples’ prenatal egalitarian expectations were associated with postpartum decreases in couples’ marital satisfaction (Pape Cowan et al. 1985); on the other hand, couples who were able to share child care responsibilities more equally in the postpartum period experienced greater satisfaction with parental roles and couple relationship quality (Cowan and Cowan 1987).

Gender-Role Orientation and Family Interactions from Pregnancy to Toddlerhood: A Sample Study

We have conducted several longitudinal studies in which we focused on the development of family interactions through the transition to parenthood (see McHale et al. 2018, for a historical overview); one of them – the focus of this chapter – was specifically dedicated to identifying prenatal precursors of mother-father-baby interactions. We took several measures during pregnancy (at the fifth month), including gender-role orientation in both parents and prenatal interactions. Families in our study were expecting their first child, who was the target of the study, and were followed from pregnancy until the second half of their child’s first year of life (at 3, 9, and 18 months) in order to assess postnatal family interactions. Two later follow-up points, at 5 and 15 years, were also subsequently completed. In this chapter, we focus on the first 18 months of this longitudinal study. All of the families who took part in this study happened to have implemented traditional specialized roles (the mother as main caretaker), as is the case in most families in Switzerland. Even though the analysis of gender roles was not the main goal of the study, we were able to observe gendered effects. For example, in the first months after birth , fathers were more at ease playing with their babies when they had the opportunity to see mothers play first, whereas no such effects were observed for mothers (Frascarolo et al. 2003). This effect was especially pronounced in families with high coparental cohesion (Udry-Jørgensen et al. 2016). Our attention was therefore drawn to a possible interinfluence between parental roles, gender-role orientation, and the quality of family relationships.

Consequently, we specifically analyzed our data on gender-role orientation according to the two constructs of masculinity and femininity assessed during pregnancy. In accordance with the literature, we tested four hypotheses. First, we expected higher femininity in fathers during their partners’ pregnancy to be linked with better postnatal cooperation during family interactions, as more feminine fathers are theoretically more oriented toward parental tasks. Second, we expected higher femininity in pregnant mothers to be linked with lower postnatal cooperation, as more feminine mothers might “close the gate” to the father in order to be congruent with what they view as the traditional duties of mothers. Third, we expected the influence of gender-role orientation on interactions of families-to-be to be already observable during pregnancy. Fourth and finally, we expected the links between gender-role orientation during pregnancy and family cooperation in the postpartum to be different in families of boys versus families of girls. Family cooperation was assessed in terms of family alliance (FA), that is, the way the family is able to coordinate to achieve a task (Fivaz-Depeursinge and Corboz-Warnery 1999).

Method

Sample

The sample consisted of 50 families expecting their first child (27 boys, 23 girls), recruited through press announcements and in a maternity ward during pregnancy. Families were Swiss, French-speaking, clinically non-referred, and middle- to upper-middle class (Hollingshead Index of Social Position). The mean age was 30 years (SD = 3.2) for mothers and 32 years (SD = 5.4) for fathers.

Procedure

Families were studied in our laboratory in the University Hospital during the fifth month of pregnancy and then again when their child was 3, 9, and 18 months old. Family interactions were video recorded while families participated in the Lausanne Trilogue Play (LTP), a semi-standardized observation situation designed for the assessment of interactions during pregnancy (prenatal LTP) and after birth (standard LTP). A self-report questionnaire was also completed by the parents in the laboratory following the prenatal LTP to assess gender-role orientation.

The Prenatal LTP

In this situation, pregnant couples enacted an encounter with their child (Fivaz-Depeursinge et al. 2010), represented by a doll (with a baby’s body but an undefined face). The experimenter asked couples to imagine and enact the first time the three of them are alone together after birth . The procedure is described in detail in Chap. 3 of this book (Fivaz-Depeursinge et al.). The mean duration of the prenatal LTP was 5.10 min (SD = 1.46 min; range 2.15–9.25 min).

The Standard LTP

This play situation after birth involves the father, mother, and baby together and follows the same four-part scenario as the prenatal LTP (Corboz-Warnery et al. 1993). When the infant was 3 and 9 months old, the parents sat on chairs and the child in a baby chair, which could be oriented in three positions: toward one parent, toward the other, and between the two of them. When the child was 18 months old, the parents and child sat around a small round table, and various toys were at hand (wooden blocks, animals, a dinner set, a small hairbrush, a car). The instructions were as follows:

We ask you to play together as a family according to the following scenario in four parts. In the first part, one of you plays with the child, the other one being simply present. In the second part, roles are reversed. In the third part, you both play with the child together. In the last part, you will talk a while together; it will be the child’s turn to be simply present or playing on her own.

In 50% of cases, we asked the mothers to be the first to play, and in the other 50%, we asked the fathers to be first. On the other hand, we let the parents decide the duration of the play. The mean duration was 11.10 min at 3 months (SD = 2.58; range 4.57–17.58), 10.30 min at 9 months (SD = 2.69; range 5.92–17.00), and 13.48 min at 18 months (SD = 2.45; range 6.43–19.50). There were no significant differences in the durations of play interactions between the three observation times.

Assessment of Family Alliance (FA)

Prenatal FA

Prenatal FA was assessed by coding the prenatal LTP videos with five Likert scales, scores ranging from 0 (inappropriate) to 1 (partially appropriate) to 2 (appropriate; Carneiro et al. 2006). The videos were coded for the following: (a) coparental playfulness toward the task (the couple’s capacity to create a playful space and to co-construct the game; interrater reliability; Cohen’s kappa = 0.83), (b) structure of the play (the couple’s capacity to structure the play in four parts according to the instructions; kappa = 0.78), (c) intuitive parenting behaviors (use of intuitive parenting behaviors such as baby talk; kappa = 0.63), (d) couple’s cooperation (degree of active cooperation reached by the couple during the play; kappa = 0.64), and (e) family warmth (positive bond and mood between parents during play, including the infant and not at her expense; kappa = 0.64). The scores of the five scales (α = 0.79) were summed to obtain a global score between 0 and 10. The higher the score, the more functional the prenatal alliance.

Postnatal FA

The quality of the postnatal FA was assessed by coding the standard LTP videos with the Family Alliance Assessment Scales (Favez et al. 2011). This instrument consists of 11 scales that assess triadic interactive behaviors, scores ranging from 0 (inappropriate) to 1 (partially appropriate) to 2 (appropriate), for postures, gazes, role implication, task fulfillment, co-construction, parental scaffolding, family warmth, validation of the child’s emotional experience, authenticity of the expressed affects, communication mistakes during the game, and communication mistakes during transitions. A total score was then computed by adding the scores of these scales (0–22 points; α = 0.87). The higher the score, the higher the alliance.

Coding Strategy

The coding was done by four pairs of independent coders, one pair at each time point, so that eight coders were involved. At each time point, one of the coders coded all of the LTPs and the other double coded 30% of the LTPs. Intraclass coefficients for the prenatal LTP scores ranged between 0.68 and 0.94, for an average of 0.82; at 3, 9, and 18 months, coefficients ranged between 0.83 and 0.96, for an average of 0.88. Coders at a given time were blind to the coding of the other times. All coders were trained by senior coders of our team.

Assessment of Gender-Role Orientation

Each partner’s gender-role orientation was assessed during pregnancy by using the Bem Sex Roles Inventory (BSRI ; Bem 1974, 1981a). This questionnaire assesses sexual identity, defined as “the representation people have of themselves regarding their sexual roles.” Two dimensions are assessed: masculinity and femininity. The standard version of the questionnaire includes 60 items: 20 are considered feminine (e.g., affectionate, compassionate, gentle), 20 are considered masculine (e.g., act as leader, aggressive, competitive), and 20 are neutral, designed to assess social desirability. Each item is assessed on a 7-point scale ranging from 1 (never) to 7 (always). In the present study, we used the 38-item version of the inventory, with 19 masculine items and 19 feminine items. From the standard version, the two items “I am masculine” and “I am feminine” were removed, as they were considered to be too related to biological sex (Bem 1979; Frascarolo et al. 1996). The 20 neutral items were also not retained (Holt and Ellis 1998). A mean of the 38 items was then computed to obtain a total score of masculinity and femininity for each respondent (α = 0.85 and 0.82, respectively, for mothers; α = 0.88 and 0.82, respectively, for fathers). The higher the score, the higher the person ranks on the considered dimension. In the original model, respondents are categorized in one of four categories of gender-role orientation: androgynous, masculine, feminine, or undifferentiated. In this study, we used the scores of the two dimensions of femininity and masculinity as continuous variables, rather than using the categories, and we used the cross-product of the two dimensions as a measure of androgyny.

Statistical Analyses

As a preliminary step, we used structural equation modeling to investigate the evolution of postnatal FA by specifying a growth curve model with two factors, an intercept and a slope factor, which represent families’ baseline scores and change in FA scores between 3 and 18 months (Model 1), respectively. To investigate whether the evolution of FA was similar or different according to the gender of the child, we then used multigroup analyses and tested the same model separately in families of boys and girls. In a first model (Model 1a), the mean and variance of the intercept and slope factors were freely estimated in each group (assuming differences between families of boys and girls), whereas in a second model (Model 1b), these parameters were set to be equal in families of boys and girls (assuming equivalence between families of boys and girls). The adjustment of these two nested models was compared by using a likelihood ratio test (LRT).

Then, to test for the influence of prenatal variables on postnatal FA, we specified a model (Model 2) in which the FA intercept and slope factors were regressed on gender-role orientation variables, namely, masculinity, femininity, and androgyny in mothers and in fathers. Due to the difference in the nature of the measures between the two coding instruments, we included prenatal FA as a covariate in the model and not as a variable of the same level as the other time points. We centered masculinity and femininity variables and created the androgyny variable as a cross-product of these centered variables.

Again, to test for potential differences in families of boys versus families of girls regarding the influence of parental gender-role orientation on the evolution of FA, we used multigroup analyses and tested the same model separately in families of boys and girls. In a first model, the regression weights that linked gender-role orientation variables to the FA intercept and slope factors were freely estimated in each group (assuming differences between families of boys and girls), whereas in the second model, these parameters were set to be equal in both groups (assuming equivalence between families of boys and girls). These two nested models were compared by using an LRT.

All statistical analyses were performed with IBM SPSS 24 software and Mplus version 7. We used chi square, comparative fit index (CFI), root mean square error of approximation (RMSEA), and standardized root mean square residual (SRMR) to assess the fit of the models. Models were estimated by using a maximum likelihood estimator with robust standard errors and a mean- and variance-adjusted test statistic (MLMV estimator), which is robust to non-normality (Maydeu-Olivares 2017).

Results

Descriptive Statistics

Descriptive statistics (see Table 9.1) showed that there were values close to both ends of the continuum in prenatal FA and in postnatal FA at each time point. Mean scores tended to be in the partially appropriate – appropriate range (a score of 5 for prenatal FA and of 11 for postnatal FA could reflect an average of partially appropriate scores, that is, a score of 1 on each dimension of both coding systems). There was thus a general trend toward a functional alliance in our sample.

Regarding the BSRI , the means were 5.10 for femininity and 4.45 for masculinity in mothers and 5.07 for femininity and 4.87 for masculinity in fathers. Comparison with a reference sample of 447 couples from a previous study in Switzerland (Frascarolo et al. 1996) showed no significant difference between this sample and the reference sample.

FA Through the Transition to Parenthood

The results for the growth curve model of the evolution of FA along the first 18 months showed that Model 1 had a good fit, χ2 = 0.200, df = 1, p = 0.655, CFI = 1.000, SRMR = 0.011, RMSEA = 0.000, 90% confidence interval (CI) [0.000, 0.334]. The estimated means of the intercept and slope factors were significant (M = 12.671, p < 0.001, and M = 2.308, p < 0.001, respectively). These results suggested that the average FA score was different from 0 at baseline and tended to increase linearly from 3 to 18 months. The variance of the intercept factor was significant (σ2 = 32.575, p < 0.001), suggesting that there was a significant heterogeneity in FA scores at baseline. In contrast, the variance of the slope factor was not significant, indicating homogeneity in the way that FA scores increase from 3 to 18 months.

Concerning potential differences in the postnatal evolution of FA in families of boys or girls, the results of the estimation of Model 1a and Model 1b, assuming differences or equivalence, respectively, in FA evolution in families of boys or girls, showed that both models had a good fit, χ2 = 1.582, df = 2, p = 0.454, CFI = 1.000, SRMR = 0.028, RMSEA = 0.000, 90% CI [0.000, 0.430] and χ2 = 5.348, df = 6, p = 0.500, CFI = 1.000, SRMR = 0.186, RMSEA = 0.000, 90% CI [0.000, 0.283], respectively. All fit indices showed a good fit for both models, except for the SRMR of Model 1b, which was higher than 0.05, a finding not uncommon in small samples. The results of the LRT were nonsignificant (χ2 = 3.786, df = 4, p = 0.436), which suggested that Model 1b should be preferred, as it was more parsimonious, but not statistically different from Model 1a.

Prenatal Gender-Role Orientation as Predictor of FA

The results for the test of Model 2 (see Fig. 9.1) showed that this model had a good fit, χ2 = 6.926, df = 8, p = 0.545, CFI = 1.000, SRMR = 0.016, RMSEA = 0.000, 90% CI [0.000, 0.178], with a nonsignificant chi square, a CFI above 0.95, and an SRMR and RMSEA below 0.05.

Fig. 9.1
figure 1

Results of the parameter estimation for Model 2. Model fit: χ2 = 6.926, df = 8, p = 0.545, comparative fit index = 1.000, standardized root mean square residual = 0.016, root mean square error of approximation = 0.000, 90% confidence interval [0.000, 0.178]. All estimates are non-standardized. FA Family alliance

Full size image

Parameter estimates showed that the FA intercept factor, that is, the baseline level of FA in the postnatal period, was predicted only by lower maternal androgyny. On the other hand, the FA slope factor, that is, the evolution of the alliance throughout the postpartum period, was negatively associated with paternal masculinity and positively associated with maternal femininity and androgyny.

Concerning potential differences according to child gender, the results of the estimation of Model 2a and Model 2b, assuming differences or equivalence, respectively, in the influence of parental gender-role orientation on the evolution of FA in families of boys and girls, showed contrasting results, particularly in the adjustment of the models. Model 2a could be considered to have a moderate to good fit according to the different fit indices, χ2 = 23.394, df = 16, p = 0.104, CFI = 0.930, SRMR = 0.054, RMSEA = 0.160, 90% CI [0.000, 0.291]. The nonsignificant chi square indicated good fit, and the CFI value between 0.90 and 0.95 indicated an acceptable rather than an excellent fit. The SRMR, at just above 0.05, indicated a correct fit, whereas the RMSEA indicated a poor fit. Model 2b could also be considered to have a moderate to good fit according to the different fit indices, but globally showed a slightly lower adjustment, χ2 = 40.264, df = 28, p = 0.063, CFI = 0.88, SRMR = 0.070, RMSEA = 0.156, 90% CI [0.000, 0.257]. The nonsignificant chi square indicated good fit and the SRMR, at below 0.08, was acceptable. In contrast, the CFI and RMSEA values indicated a poor fit. The result of the LRT comparing these two models was not significant (χ2 = 14.721, df = 12, p = 0.257) and indicated that Model 2b should be preferred, assuming equivalence between families of boys and girls in the influence of parental gender orientation on postnatal FA evolution.

Prenatal FA, as a covariate, was specifically positively linked with femininity in fathers and negatively with androgyny in mothers; it was also strongly positively linked with the FA intercept factor, suggesting a continuity in the quality of alliances from the prenatal to the postnatal period.

Table 9.1 Descriptive data of study variables (N = 50)
Full size table

Discussion

Following studies on gender-role orientation and parenting, we expected gender-role orientation as assessed during pregnancy in both parents to be linked with the development of FA throughout the first 18 postpartum months. Firstly, we expected higher femininity in expectant fathers to be predictive of better FA, as femininity in fathers has been shown to be favorable to paternal engagement; secondly, we expected higher femininity in expectant mothers to be predictive of lower FA, as maternal femininity may be one of the factors motivating gatekeeping behaviors. Whereas our results showed that gender-role orientation is generally related to the development of FA, the specific links between these variables differed somewhat from our expectations.

Our results show that androgyny in mothers is linked with an overall lower postnatal alliance; even though we expected femininity and not androgyny in mothers to be unfavorable to FA, this finding nevertheless makes theoretical sense. It may be the combination of femininity and masculinity in more androgynous mothers that leads them to engage in gatekeeping behaviors, with the consequence of lowering cooperation between them and their partners. Indeed, the feminine characteristics of androgynous mothers may lead them to stick to traditional roles, whereas their masculine characteristics may provide them with the necessary assertiveness to affirm their will. On the other hand, androgyny and femininity in mothers are paradoxically both predictive of an increase of alliance across the first 18 months, a picture that is congruent with a traditional family organization that relies on the centrality of maternal engagement. With respect to fathers’ gender-role orientation, although their femininity is not linked with postnatal family alliance, their masculinity is predictive of a lowering of the alliance over time. This latter finding is congruent with a contemporaneous organization of the family that relies on more egalitarian and less dominant paternal behaviors.

In sum, the evolution of FA is predicted by a blend of traditional and contemporaneous tendencies in parental gender-role attitudes. Notably, prenatal FA, used as a control variable in our study described in this chapter, is strongly linked to the level of postnatal FA, showing that similar patterns of interactions are already operating before the birth of the couples’ first child – the organization of the actual triad is foreshadowed by the imaginary triad, a result we have noted previously (Favez et al. 2013). Moreover, in accordance with our third hypothesis, we found that prenatal FA is linked to two variables related to gender-role orientation. Androgyny in expectant mothers is associated with lower prenatal alliance, as is the case with postnatal alliance. Maternal gatekeeping behavior, influenced at least in part by expectant mothers’ gender-role attitudes regarding motherhood, may thus already be operating during pregnancy. On the other hand, we also found that femininity in fathers is associated with better prenatal family alliance , whereas this is not the case with postnatal alliance. In fathers with more feminine gender-role orientations, this may reflect a process of their investment in a role traditionally assigned to mothers; their readiness to collaborate with mothers and willingness to endorse more feminine-gendered traits may be part of their effort to build a parental identity prior to their children’s births (Eggebeen and Knoester 2001; Favez and Frascarolo 2019). However, once the baby is born, this effect seems to fade, perhaps due to the joint influence of maternal gatekeeping and the social context in which the young family is developing or due to changing gender-role attitudes after the transition to parenthood (Pape Cowan et al. 1985).

How can these mixed results and the coexistence of multiple influences on the development of postnatal FA be explained? We suggest that they hint at parallel historical and contemporaneous influences that place today’s parents in Western cultures such as Switzerland within a period of transition; traditional parental role divisions are still influential while more egalitarian parental roles are also operating to shape couples’ ideas of their families envisioned during pregnancy as well as their developing families during the early postpartum months. Although the gender revolution has brought a change toward more egalitarian parenting practices between mothers and fathers (Cabrera et al. 2000; Trifan et al. 2014), the process is still in a transitional phase (Goldscheider et al. 2015). This intermediate era in which traditional views of parenting have not been quite abandoned and new views fully embraced by all families may explain interindividual variations between families: Some families may continue to endorse traditional roles, while others may have switched to the new way of endorsing masculinity and femininity, and still others switch between these different gender-role orientations. This may be why, as a group, the families in our study showed these multiple influences. Moreover, contradictory influences may also coexist within the same family at an intraindividual and dyadic level; parents may be torn between old and new roles. There might indeed be specific processes operating during the transition to parenthood that lead to the adoption of more traditional roles after the baby is born – which would explain the differences in the influence of gender-role orientation on FA we observed between the prenatal and the postnatal periods, especially in fathers.

Others have described the transition to parenthood as a period of changes in gender-role attitudes toward more specialized roles (Katz-Wise et al. 2010; Pape Cowan et al. 1985; Pape Cowan and Cowan 1992); in a way, congruence with traditional roles allows parents to follow a familiar pathway with well-established landmarks so that uncertainties associated with parenthood may be reduced. Pape Cowan et al. (1985) suggest that in the face of the multitude of changes impinging on the couple during the transition to parenthood, men and women may fall back on familiar role models they encountered in their families of origin which is why their roles may become more gender stereotypical and increasingly different after birth of their first child (Pape Cowan et al. 1985). Moreover, parenthood is a domain in which pressure to conform to gender stereotypes is particularly high; this pressure may not be explicit, but rather a consequence of the socialization of fathers and mothers. The traditional model is thus so internalized that it seems natural to follow it; as a consequence, families may feel the need to conform to social representations of what a family should be, at least to a certain extent. Finally, the demands of the social environment in most present-day European and other cultures are contradictory in themselves: Both parents are supposed to be at the same time efficient professionals and nurturing caregivers; that is, they are expected to follow both contemporaneous and traditional gender-role standards. Combining active involvement with children with their role as financial provider may be difficult for fathers, especially as the employment world is not yet ready to offer men the conditions that allow them to manage both their jobs and their family lives (McGill 2014). This is especially the case in Switzerland where paternity leave is almost non-existing. Similarly, mothers have to combine engagement in the workforce with child care and family responsibilities, a double agenda that may induce feelings of guilt and over-burden for mothers (Borelli et al. 2017; Craig 2006; Donnelly et al. 2016; Milkie et al. 2002).

Does this mean that after the gender revolution has been completed and this transitional phase has passed, family systems will be in a “perfect state of equifinality” (each parent fulfilling parental tasks necessary to family functioning) and that the masculine and feminine dimensions in parenting will have no importance? While no one can predict the answer to this question with certainty, we believe that whichever terms are used to designate them, the qualities typically understood as feminine or masculine are still important for adaptive family functioning, regardless of how these qualities are divided among the caregivers of children. Based on our findings with families led by two, heterosexual parents, it seems best if both parents are able to incorporate these qualities equally, but this does not imply that fathers will behave exactly like mothers and that mothers will behave exactly like fathers (Hook and Chalasani 2008). Theories that highlight the importance of the qualities traditionally attributed to fathers and to mothers for the development of the child thus have to make a shift in order to associate these qualities not with the biological sex – as most classic theories have done – but with “postgender” roles.

Finally, we also expected the links between gender roles and FA to differ according to the sex of the child. This fourth hypothesis was not confirmed, as our model in which equality between families of boys and girls is constrained allows a better explanation of the data than our model in which the influence of the sex of the child is freely estimated. Even though differences in the type of non-cohesive behaviors have been observed between families of boys and girls (McHale et al. 2002), the effect of gender-role orientation in parents is not primarily determined by the sex of the child and seems to be instead linked to features of the parents themselves – features which are already influential during pregnancy – and to their engagement in a collaborative coparental relationship.

In conclusion, although we did not find a clear pattern that links gender roles and FA (or perhaps precisely because we did not find this pattern), this study shows that representations of gender roles need to be considered in the development of early family relationships, as early as during pregnancy. Parental representations of the baby, of the future family, and of the self as a parent have all been shown to be linked to the way each parent behaves during family interactions. Ambivalence regarding roles and contradictory influences such as those highlighted in this study have been frequently described in surveys of the transition to parenthood. Studies have shown that parents are willing to be egalitarian but continue to organize themselves, unconsciously, according to traditional roles. Postgender attitudes are thus not yet completely achieved (Knudson-Martin and Mahoney 2005), a previously reported finding that our results support as well. Cultural expectations, cultural norms, social norms, individual expectations, and behaviors may change at different paces, creating discrepancies between and within families (Fox et al. 2000; Milkie et al. 2002). Taking these variables pertaining to the larger social environment into account would allow us to better understand this tension parents have to face during the transition to parenthood. For example, a welfare state regime greatly influences the transition in gender norms, as shown in Europe with the differences observed in the balance of family/work time in fathers and mothers according to the national policies of different countries (Neilson and Stanfors 2014).

Several limitations of this study have to be mentioned. First, this is a secondary study embedded in a larger study on the transition to parenthood. Since the influence of gender role was not the primary variable of interest in our study, we lack complementary measures to strengthen our results. Secondly, we used the BSRI questionnaire, whose construct validity has been debated over the years. Other instruments (see Moradi and Parent 2013) could be used to confirm and further test the validity of the psychological qualities that we identified as feminine and masculine (Lips 2017). Moreover, we stuck with the tradition that considers gender role as being close to a trait. However, gender also has state-like qualities; that is, the expression of gender is affected by the specifics of the context for a given individual (Keener and Strough 2017), so that multiple contexts should be taken into account for a comprehensive assessment of gender-role orientation in any individual. Finally, alternative methodologies and additional data would be needed to understand the exact process by which representations of gender roles may affect the relationship between parents and their mutual interactive behaviors during pregnancy and through the first years of life of the child. For example, it would be interesting to take into account each parent’s perception of the other parent’s gender-role orientation and the expectations associated with these representations.

Even though our study has raised more questions than it has answered, it shows the importance of taking gender-role orientations of parents-to-be as well as new parents into account in explorations of coparenting interactions. Understanding the role of gender-role orientation allows for a more comprehensive depiction of the contextual development of family relationships across the transition to parenthood.