To support good outcomes for mothers and babies, it is essential to understand the prenatal factors that place both at risk, such as birth trauma, the mother’s attachment history, and unplanned motherhood. However, for health professionals supporting mothers, interventions must target malleable intervening factors. One of these, we propose, is social identification as a mother. Among women who had become mothers in past 12 months (N = 317), we assessed prenatal risk factors, social identification as a mother, and three socio-emotional outcomes: mother’s depression, infant temperament, and mother-infant attachment. Consistent with hypotheses, all prenatal risk factors negatively predicted social identification as a mother, which was itself positively associated with socio-emotional functioning. Model fit was excellent. Reduced social identification as a mother may be a key mechanism through which prenatal risk factors produce poorer outcomes for mothers and babies. Interventions may be enhanced by supporting the development of strong and positive social identities.
Opmerkingen
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
The perinatal period and first 12 months of life is a critical window for determining long-term trajectories for both mother and child (Currie & Rossin-Slater, 2015; Saxbe et al., 2018). For mothers, the perinatal period is one of the highest risk periods for depression across the lifespan, and mental health disorders are the most common complication of pregnancy (Howard & Khalifeh, 2020; Munk-Olsen et al., 2006; Vesga-López et al., 2008). Suicide is the leading cause of death among perinatal women (Oates, 2003). However, even these statistics are dwarfed by the importance of this period in determining life trajectories for infants. The quality of parenting received in the first year of life has been linked to outcomes across the lifespan ranging from anxiety, to cognitive development, to antisocial behaviour (Davis et al., 2015; Kiernan & Huerta, 2008; Kochanska & Kim, 2012). For these reasons, it is crucial to understand what places mothers and babies at risk during this time, and what can be done to reduce these risks. The current study seeks to advance our understanding with a particular focus on social identification as a mother: the subjective, psychological sense of self-definition and affiliation that arises from motherhood.
Prenatal Risk Factors for Mother and Baby
In this project we focus on risk factors for poorer mother and child socio-emotional outcomes that occur prior to a child’s birth (i.e. prenatal risk factors). The first we will consider is traumatic experiences during labour and delivery, including both physical trauma as well as psychological trauma, the importance of which has received growing recognition especially in the last 10 years. For a significant minority of women, giving birth can be a highly traumatic experience where a woman fears for her life and/or that of her child (i.e. a Criterion A event; APA, 2013). Traumatic birth experiences are strongly linked to poor outcomes for the mother (especially depression; Ayers et al., 2016; Bay & Sayiner, 2021; Woolhouse et al., 2014). There is also both qualitative and quantitative evidence that birth trauma can undermine the quality of mother-infant attachment (Molloy et al., 2021; Smorti et al., 2020). In terms of infant outcomes, one meta-analysis found that birth trauma was associated with higher rates of child abuse and neglect, although causal links have not been convincingly established (Suomi et al., 2021). A systematic review also identified birth complications as a risk factor for difficult temperament (Takegata et al., 2021). Infant temperament is one of the most important early indicators of infants’ socio-emotional functioning. Temperament can then, in turn, become part of a complex and reciprocal feedback loop, affecting both parenting and maternal mental health (Belsky et al., 2009; Bridgett et al., 2009, 2011).
The second prenatal risk factor we will consider, and perhaps the most studied, is the mother’s attachment history. Attachment theory states that if a mother established a warm, loving, and secure attachment with her primary caregiver as an infant, this, in turn, is likely to be reflected in the quality of her bond with her own children (Della Vedova et al., 2011; Esposito et al., 2017; Fraley, 2002). That is, mothers who experienced consistent and positive emotional care in childhood are more likely to provide warm and sensitive parenting to their children (Belsky et al., 2009). It is through this mechanism that attachment styles (secure vs. insecure) are theorised to be transmitted from parent to child. A mother’s attachment insecurity is also a risk factor for other poor outcomes, such as poor wellbeing among infants, both in the short term and across the lifespan (Grant et al., 2012; Heider et al., 2006). There is also evidence that maternal attachment history may affect infant temperament (Priel & Besser, 2000). Attachment theorising proposes that the mediator of the intergenerational transmission of attachment style is caregiver sensitivity, such that caregivers with an insecure attachment history are less likely to provide timely, appropriate, and responsive care to their offspring. However, much has been written of the transmission gap, which is the relatively small amount of the variance between parental attachment and infant attachment that is accounted for by measures of caregiver sensitivity (Fonagy & Target, 2005; Verhage et al., 2016). This project focuses on providing new insights into the mechanisms underpinning the link between such prenatal risk factors and poor outcomes for mother and child.
A third and final prenatal risk factor for the purposes of the current project is the degree to which pregnancy is planned or intentional, versus mistimed or unwanted. There is a high degree of stigma regarding unplanned parenthood, and highly variable access to contraception and abortion both across and within countries (Gipson et al., 2008). Even in areas where access to contraception and abortion is high, many parents do not find these acceptable or desirable alternatives to an unplanned pregnancy. While the stigma around unplanned parenthood creates barriers to rigorous study, research has suggested that unplanned pregnancies that are seen through to term are associated with several negative outcomes for mother and infant. In particular, children are at higher risk of poor developmental outcomes (Baydar, 1995), attachment is less likely to be secure (Goto et al., 2018), and mothers are at higher risk of depression (Cheng et al., 2009; Nakku et al., 2006; Qiu et al., 2020).
While prior research has persuasively demonstrated the importance of these risk factors, they are difficult for health professionals to target directly to improve outcomes for mothers and babies. They occur in the prenatal period, often prior to the involvement of maternal health services—even dating back to the mother’s own infancy in the case of mother’s attachment style. For unplanned parenthood, a further barrier to intervention is that its levers are primarily political rather than under the control of health service providers (Johnson-Mallard et al., 2017; Jozkowski & Crawford, 2016). While birth trauma is certainly something health professionals can and do seek to reduce (e.g. Taheri et al., 2018), by the time a new mother is presenting for support the experience of a traumatic event is part of a patient’s history. For this reason, it is crucial for research to investigate modifiable factors that intervene between risk factors and poor outcomes.
Social Identity as a Promising New Direction
Social identification as a mother is the psychological internalisation of motherhood; the degree to which a person subjectively feels like a mother. Social identity goes beyond merely an attitude (e.g. feeling positive about motherhood) or a relationship (e.g. friendship with other mothers). Instead, social identity is the degree to which self-concept is shaped by one’s membership of a particular group or social category, and so is best understood as both a social and cognitive construct. Because a social identity is part of our self-definition, it also shapes our thoughts, feelings, and actions. There are decades of evidence and thousands of publications to support the role of social identification in determining outcomes as diverse as protest, voting, discrimination, and employee retention (Cheng et al., 2016; Greene, 1999; van Zomeren et al., 2008). Crucially for the topic of this investigation though, there is also a large body of work that has found that social identification affects health and wellbeing. For example, a meta-analysis demonstrated a significant association between social identification and depression of approximately r = −0.15 across 76 studies (Postmes et al., 2019). There is also evidence that this link is causal. For example, experimental priming of social identities reduces depressive attributions following an experience of failure (Cruwys et al., 2015).
In the context of motherhood specifically, qualitative and quantitative research has documented the importance of social identity processes for maternal mental health. For example, Seppälä and colleagues (2022b) conducted three waves of semi-structured interviews with new mothers across a year. They found that a shared sense of identity with other mothers was key to combatting isolation and loneliness, and facilitated a sense of mastery. One correlational study with over 350 mothers of infants found a strong negative association between social identification as a mother and depressive symptoms (Seymour-Smith et al., 2017). Another study that sampled members of online parenting forums found that mothers who posted regularly in the postnatal depression forum had lower social identification as a mother and this declined over time, compared to mothers who posted in other forums on the same platform (Koschate et al., 2021). Finally, in a study with 150 mothers of children aged five years or less, Bentley and colleagues (Bentley et al., 2020) found that a mother’s total number of social identities (not necessarily related to motherhood) negatively predicted depression symptoms. The link between a mother’s social identification and her wellbeing is thus well-evidenced; however, there is not yet evidence for the role of mother’s social identification in supporting mother-infant attachment, or outcomes for infants (such as temperament). Nevertheless, we predicted that social identity would be protective for these outcomes given the central role of the primary caregiver – and the (social) relationship between mother and child – in determining trajectories of mother and infant socio-emotional functioning. There is also broader evidence that strong social identification supports one’s ability to enact the tasks associated with a particular identity (Cruwys et al., 2016), which in this case would include effective parenting. This study provided the first test of these pathways.
Finally, another area that has not yet been explored in previous research is the determinants of social identification as a mother. There are theoretical reasons, though, to believe these may include prenatal risk factors such as a mother’s own attachment history. Whether a person incorporates a social group into their subjective psychological representation of self is influenced by both the fit with the social context (e.g. social contact with other mothers; Seppälä et al., 2022a), but also by what is called perceiver readiness (Turner et al., 1987). Perceiver readiness refers to the relative accessibility of a specific social category as a lens for making sense of the world. Experimental research has found that a person’s previous experiences related to an identity are a key component of perceiver readiness (Blanz, 1999; Smith et al., 1996). In this context, therefore, we might predict that negative experiences related to motherhood in the prenatal period could undermine a mother’s willingness to self-define on that basis. For example, birth trauma might lead one to find reminders of motherhood distressing, and lower social identity as a mother may thus be a form of psychological avoidance of such reminders. Similarly, insecure attachment to one’s own caregiver and unplanned motherhood may both make a woman less willing to or comfortable with self-defining as a mother. However, what makes social identity as a mother particularly interesting in the present context is that social identities are readily affected by social and contextual factors. Although only one experiment to our knowledge has manipulated social identity as a mother specifically (Study 4; Koschate et al., 2021), a large body of experimental research supports the dynamic nature of other kinds of social identities (e.g. Cruwys et al., 2022b; Greenaway et al., 2016; Kyprianides et al., 2019).
The Current Study
Drawing together these lines of evidence, we argue that prenatal risk factors reduce social identification as a mother, which, in turn, places maternal and infant socio-emotional functioning at risk. In other words, we conceptualise low social identification as a mother as a mechanism through which prenatal risk factors exert their negative consequences for mother and baby. This study provided the first investigation of the predictors of social identity as a mother, with a focus on birth trauma, insecurity of mother’s attachment history, and unplanned motherhood. In addition, we sought to assess the link between social identity as a mother and indicators of socio-emotional functioning for both mothers and children, including one outcome that has been previously investigated in this context (i.e. mother’s depression) and two outcomes that have yet to be linked to social identity but for which we had strong reason to expect a relationship (i.e. mother-infant attachment and infant’s temperament).
Our overarching prediction was that prenatal risk factors will work via reducing social identification as a mother to produce negative outcomes for mother and infant. More specifically, we hypothesised that birth trauma (H1a), insecurity of mother’s attachment to her primary caregiver (H1b), and the degree to which becoming a mother was unplanned (H1c) would each negatively predict social identification as a mother. Furthermore, we predicted that social identification as a mother would negatively predict mother’s depression (H2a), and positively predict infant’s temperament (H2b) and quality of mother-infant attachment (H2c). Finally, we predicted that the indirect effects of prenatal risk factors on mother and infant socio-emotional functioning via social identification as a mother would be significant (H3).
Method
Participants
Participants were eligible for the study if they identified as women, were English speaking, and had given birth, adopted, or become responsible for raising a child in the previous 12 months. Participants were from 18 countries, however, the majority were residing in the United Kingdom (49.7%, n = 149) or the United States (33.3%, n = 100), with much smaller numbers from other countries including Canada (2.3%, n = 7), Poland (2.0%, n = 6), and Portugal (1.3%, n = 4). Country of residence data were not available for 26 participants. The final sample included 317 women who were aged 18–42 (M = 30.31; SD = 4.76), whose youngest children were aged 0–11 months (M = 5.45; SD = 2.64). Sample demographics are summarised in Table 1.
Table 1
Demographics of the sample
Variable
Level
Percentage
n
Ethnicity
White
82.2%
249
Black
7.3%
22
Asian
5.6%
17
Hispanic/Latino/a
3.0%
9
Other
2.0%
6
Relationship Status
Married
73.0%
222
Defacto
17.4%
53
Single
7.6%
23
Divorced/Separated
2.0%
6
Education
High School or less
17.1%
52
Vocational Training
20.4%
62
Bachelor degree
42.1%
128
Graduate degree
20.4%
62
Annual Household Income
<£20,000
10.2%
31
£20,000-£30,000
17.5%
53
£30,000-£40,000
17.2%
52
£40,000-£50,000
12.2%
37
£50,000-£60,000
13.9%
42
£60,000-£70,000
11.2%
34
>£70,000
17.8%
54
How many children are you currently a mother to?
1
54.6%
166
2
33.2%
101
3
7.2%
22
4
3.3%
10
5–7
1.3%
4
How did you become a mother?
My biological child (my egg)
95.7%
291
My partner’s biological child
3.0%
9
Adopted child
1.0%
3
Other
0.3%
1
Participants were all mothers to a child aged < 12 months
Percentages are based on the complete sample (valid %). Of 317 respondents, 13–14 responses were missing for each demographic variable
Procedure
The study was advertised via Prolific, an online recruitment platform based in the United Kingdom, to people who indicated in their Prolific pre-screening questionnaire that they identified as women and had become a parent in the past 12 months. Participants were remunerated £2.50 for their time. Prolific is a widely used online research recruitment platform that has been recommended over alternatives (including traditional university participant pools) due to its comprehensive screening process, high data quality, and low levels of dishonest and inattentive responding (Peer et al., 2017; Douglas et al., 2023). In accordance with best practice for online panel-based research (Abbey & Meloy, 2017), three attention checks were included throughout the questionnaire (described below).
The design was a cross-sectional survey that began with pre-screening questions: “Have you have become a mother to a child within the past 12 months?” and “How old is this child?” If participants answered ‘no’ and/or ‘over 12 months of age’ respectively, the study automatically terminated. Ethics approval was provided by the Australian National University’s Human Research Ethics Committee (Protocol 2019/887) and all participants provided informed consent.
Measures
Birth trauma
The 10-item Birth Experiences Questionnaire (BEQ, Saxbe et al., 2018) was included to measure birth trauma. This measure is designed to assess the degree to which mothers felt stress, fear, and partner support (reverse-scored) during birth (e.g. “Did you fear for your baby’s life?” from 1 = strongly disagree to 7 = strongly agree). It has demonstrated reliability and validity in community samples (Saxbe et al., 2018), which was confirmed in our sample, α = 0.83.
Insecurity of mother’s attachment history
To measure participants’ attachment history, we included the Parental Bonding Instrument (PBI; Parker et al., 1979; see also Ravitz et al., 2010). This 25-item, retrospective self-report questionnaire for adults is completed based on how participants remember their primary caregiver during their first 16 years. We focused on the 12-item ‘care’ subscale (e.g. “Did not seem to understand what I needed or wanted” from 0 = very unlike to 3 = very like; α = 0.95) because it assesses mothers’ perspectives of the type and degree of emotional care (e.g. sensitivity) that they received from their primary caregiver as a child, reflecting the overall quality of the attachment relationship (Van Bussel et al., 2010). The PBI has been validated for retrospective recall and has been found to be highly stable over time (Wilhelm et al., 2005). One validation study comparing the PBI to the Adult Attachment Interview (generally considered the criterion measure, but which requires in-person administration and so was not feasible here) found that the PBI performed well in distinguishing secure from insecure attachment (Manassis et al., 1999). The PBI has demonstrated strong psychometric properties in clinical and community samples (Safford et al., 2007; Tuladhar et al., 2020). We reversed the scale such that higher scores represented an insecure attachment history.
Unplanned motherhood
Participants were asked to indicate the degree to which taking on the role of mother to this child was planned on a five-point scale (1 = definitely yes to 5 = definitely not). Previous research that has investigated the degree to which motherhood was planned has often treated this as a binary variable (e.g. Turner-Zwinkels & Spini, 2020), however, we reasoned that the degree of planning and satisfaction with timing varies along a continuum (see also Santelli et al., 2003).
Social identification as a mother
The 14-item multicomponent group identification scale (Leach et al., 2008) was used to measure social identification as a mother (e.g. “Being a mother is an important part of how I see myself” from 1 = strongly disagree to 7 = strongly agree), α = 0.88. This measure has demonstrated strong psychometric properties in clinical and community samples (e.g, Bertschy et al., 2020; Sani et al., 2012).
Mother’s depression
To measure maternal depression, we included the seven-item depression subscale of the Depression, Anxiety and Stress Scales-21 (DASS-21; Lovibond & Lovibond, 1995), e.g. “I felt downhearted and blue”. Items refer to the extent that a respondent has experienced the symptom over the past week and is rated on a scale from 0 = never to 3 = almost always, α = 0.90. The DASS-21 has demonstrated strong reliability and validity in clinical samples (Page et al., 2007) as well as in mothers within 12 months of giving birth (Miller et al., 2006).
Infant temperament
As an indicator of infant temperament, we included the 37-item Infant Behavior Questionnaire-Revised (IBQ-R Very Short form; Putnam et al., 2014). We focused on orienting/regulatory capacity subscale (e.g. “When singing or talking to your baby, how often did s/he soothe immediately?”) for two reasons: first, because mother’s report of this was expected to be less contaminated by the mother’s own mental health and attachment history than was report of the infant’s affect, and second, because previous research has found particularly strong links between orienting/regulation and child development, including psychopathology (Bridgett et al., 2011; Kostyrka-Allchorne et al., 2020; Toffol et al., 2019). Mothers rated the frequency of specific behaviours observed over the past 1–2 weeks on a Likert scale (1 = never to 7 = always, or NA = does not apply), with higher scores indicating higher infant capacity for orientation and regulation, α = 0.78. It has demonstrated strong psychometric properties in both clinical and community populations (Peterson et al., 2017). Because the IBQ-R is only suitable for children between 3–12 months of age, parents of children less than 3 months old did not complete this scale.
Mother-infant attachment
To measure the emotional quality of the dyadic relationship between mother and infant (following Cooijmans et al., 2017), we used the 19-item Maternal Postnatal Attachment Scale (MPAS; Condon & Corkindale, 1998). This measure has three subscales: quality of attachment (e.g. “Over the last two weeks I would describe my feelings for the baby as:” with responses from 1 = dislike to 5 = intense affection), absence of hostility (e.g. “When I am caring for the baby, I get feelings of annoyance or irritation” with responses from 1 = very frequently to 5 = never) and pleasure in interaction (e.g. “I try to involve myself as much as I possibly can playing with the baby” with responses 1 = true or 2 = untrue). It is generally used as a unitary scale (e.g. Van Bussel et al., 2010) and the reliability of the total score in the current study supported this approach, α = 0.84. Following recommendations in the literature, items were recoded 1–5 so that all were weighted equally in the final scale.
Attention checks
Three attention checks were included, two of which asked participants to select a specific response from a Likert scale (e.g. “Please select the answer ‘applied some of the time’”), and one question asked, “how carefully are you reading these questions” with responses ranging from 0 = not at all to 5 = very carefully.
Demographics
Participants additionally provided their age, education, relationship status, ethnicity, household income, their youngest child’s date of birth, the number of other children that they had, and how they became a mother.
Analysis Plan
Hypotheses were evaluated using path analysis with manifest variables. This approach was considered most appropriate because it allowed us to simultaneously model multiple (mother and infant-related) outcome variables and provides an estimate of indirect effects (i.e. mediation). In addition to the hypothesised pathways via social identification as a mother, the first path analysis (Model 1) also included all direct pathways from prenatal risk factors to mother and infant socio-emotional functioning. Predictor variables were allowed to covary, as were the error terms for the outcome variables. Hypotheses were evaluated by considering the significance of the specific direct or indirect pathways in the final model. The model was refined and evaluated using (1) exclusion of non-significant pathways, and (2) six fit indices from across the three primary categories of fit; three absolute fit indices: χ2/df, Akaike’s Information Criterion (AIC), Standardised Root Mean Squared Residual (SRMR), two non-centrality-based indices: Root Mean Square Error of Approximation (RMSEA), comparative fit index (CFI), and a relative fit index: normed fit index (NFI); using cut-offs recommended by Hu & Bentler, 1999. A sensitivity analysis was also specified by adding six covariates which have been linked to the dependent variables in previous research: mother’s age, mother’s education, mother’s relationship status (dichotomously coded 0 = single/separated/divorced/widowed; 1 = married/defacto), child’s age, family income, and whether it was the mother’s first child.
Results
Exclusions
To ensure quality of the data, people who failed two or more of the attention checks were excluded (n = 4, leaving a sample of 317). The actual available N was smaller on some variables due to a small amount of missing data and the fact that some scales were only administered to a subset (i.e. the BEQ not administered to non-birth mothers, and IBQ was only suitable for infants aged 3–12 months). Full information maximum likelihood was used to manage missing data in the path analysis model, which is a sophisticated technique that has been recommended due to providing the least biased estimates of standard error (Larsen, 2011).
Hypothesis Testing
Bivariate correlations are presented in Table 2. Interestingly, the three predictor variables were not significantly associated with one another, while the three outcome variables were significantly correlated (0.16 < |r| < 0.39). In the initial path analysis model (Model 1), all indirect pathways were significant, but many of the direct pathways were not—speaking to the importance of social identification as a mediator. This model was fully specified, and so model fit parameters were not available. Therefore, we removed the six non-significant direct pathways to yield Model 2 (see Fig. 1). Model fit indices suggested excellent fit (see Table 3).
Table 2
Bivariate correlations (Pearson’s r)
Variable
1
2
3
4
5
6
7
8
9
10
11
1. Mother’s age
-
-
-
-
-
-
-
-
-
-
2. Education
0.26*
-
-
-
-
-
-
-
-
-
3. Income
0.19*
0.37*
-
-
-
-
-
-
-
-
4. Child’s age
0.02
−0.01
−0.07
-
-
-
-
-
-
-
5. First child
−0.20*
0.19*
0.10+
−0.06
6. Birth trauma
−0.09
0.08
−0.02
0.01
0.25*
-
-
-
-
-
7. Insecurity of mother’s attachment to her primary caregiver
0.09
−0.08
−0.06
0.03
−0.21*
0.03
-
-
-
-
8. Unplanned motherhood
−0.26*
−0.16*
−0.16*
0.07
0.03
0.10+
0.05
-
-
-
9. Social identification
0.00
−0.09
−0.02
0.00
−0.07
−0.16*
−0.30*
−0.14*
-
-
10. Mother’s depression symptoms
−0.01
−0.05
−0.08
0.11+
−0.15*
0.19*
0.37*
0.12*
−0.25*
-
11. Infant temperament
0.05
−0.05
−0.03
−0.05
−0.01
−0.12+
−0.16*
−0.02
0.25*
−0.24*
12. Mother-infant attachment
−0.11+
−0.20*
−0.18*
0.01
0.09
−0.10+
−0.26*
−0.08
0.50*
−0.47*
0.37*
+p < 0.10; *p < 0.05
Table 3
Model fit indices
χ2/df
RMSEA
CFI
NFI
AIC
Model 1: All direct effects included
[other fit indices could not be calculated due to model saturation]
70.00
Model 2: Main model (six non-significant direct paths deleted).
6.35/6 = 1.08, p = 0.386
0.014
>0.99
0.98
64.35
Model 3: Model 2 plus six covariates.
19.78/12 = 1.65, p = 0.071
0.045
0.99
0.97
203.78
SRMR standardised root mean squared residual, RMSEA root mean square error of approximation, CFI comparative fit index, NFI normed fit index, AIC Akaike’s information criterion
×
Consistent with H1, birth trauma (β = −0.14, p = 0.010), insecurity of mother’s attachment to her primary caregiver (β = −0.29, p < 0.001) and the degree to which becoming a mother was unplanned (β = −0.11, p = 0.047) were all negatively associated with social identification as a mother. Consistent with H2, social identification as a mother negatively predicted mother’s depression (β = −0.13, p = 0.013) and positively predicted infant temperament (β = 0.24, p < 0.001) and quality of mother-infant attachment (β = 0.47, p < 0.001). Finally, all of the indirect effects of perinatal risk factors on mother and infant outcomes via social identification as a mother were significant for birth trauma and insecurity of mother’s attachment to her primary caregiver (see Table 4). However, the three indirect effects for unplanned motherhood were all non-significant. Therefore, there was partial support for H3. Overall, Model 2 explained 12% of the variance in social identification as a mother, 6% of the variance in infant temperament (orienting/regulation), 17% of the variance in mother’s depression, and 26% of the variance in mother-infant attachment.
Table 4
Standardised indirect effects (and 95% confidence intervals) via social identification as a mother (Model 2)
Dependent variable
Independent variable
Mother’s depression
Infant’s temperament
Mother-infant attachment
Mother’s birth trauma
0.02 (CI: 0.001, 0.05)
−0.04 (CI: −0.07, −0.01)
−0.07 (CI: −0.12, −0.02)
Insecurity of mother’s attachment history
0.04 (CI: 0.004, 0.08)
−0.07 (CI: −12, −0.03)
−0.13 (CI: −0.20, −0.08)
Unplanned motherhood
0.02 (CI: −0.001, 0.05)
−0.03 (−0.07, 0.008)
−0.06 (CI: −0.13, 0.005)
Finally, Model 3 added covariates, which were allowed to covary with one another and with the prenatal risk factors. Model fit was still excellent, and all hypothesised pathways were still significant. There were four significant covariate effects: income negatively predicted mother-infant attachment (β = −0.14, p = 0.008), education negatively predicted mother-infant attachment (β = −0.14, p = 0.010), having only one child positively predicted mother-infant attachment (β = 0.14, p = 0.005), having only one child negatively predicted mother’s depression (β = −0.14, p = 0.010).
Discussion
This study sought to explore whether prenatal risk factors (specifically, birth trauma, insecurity of mother’s attachment to her primary caregiver, and the degree to which becoming a mother was unplanned) exert their negative effects by reducing social identification as a mother. Consistent with hypotheses, the model had excellent fit and all the prenatal risk factors were negatively associated with social identification as a mother, which in turn significantly predicted maternal depression, mother-infant attachment, and infant temperament. Not only were all of these hypothesised pathways significant, but additionally social identification fully accounted for these associations in many cases—for example, while birth trauma was negatively correlated with infant temperament, this relationship was entirely accounted for by the fact that women who had experienced greater birth trauma felt less socially identified as a mother. Indeed, for infant temperament, there were no direct effects of the prenatal risk factors that were not mediated via their mother’s social identification.
The indirect effects of unplanned motherhood on our dependent variables were all nonsignificant. However, this should be interpreted in the context of the larger standard error for pathways involving this predictor. Furthermore, unplanned motherhood significantly predicted reduced social identification as a mother, which in turn significantly predicted each dependent variable. Putting these two things together, it is likely that our measure of unplanned motherhood—which was the only measure we used that was not previously validated in the literature—was less reliable than the other measures and thus caution is warranted in interpreting findings relating to this predictor.
Finally, we considered six potentially confounding variables: mother’s age, education, household income, relationship status, child’s age, and whether it was the mother’s first child. Several of these significantly predicted the dependent variables in the sensitivity analysis. Specifically, mother’s depression tended to be lower if she had only one child, while mother-infant attachment was lower among women with multiple children, higher education, and higher income. Some of these associations are not consistent with the typical pattern found in the literature, where demographic indicators of disadvantage tend to be associated with poorer mother-infant attachment (e.g. Gedaly & Leerkes, 2016). However, it is important to note that this study included a community sample rather than an at-risk sample (i.e. there were overall low levels of demographic risk factors in the present sample). Controlling for these demographic characteristics did not appreciably change the hypothesised results, building confidence in the robustness of the findings.
Implications
The central role of social identification as an intervening factor between prenatal risk and mother and child socio-emotional functioning is not merely an interesting fact in the abstract. It holds substantial potential for intervention efforts to support mothers during this period of elevated risk to their own and their children’s wellbeing. Prenatal risk factors are challenging (in some cases, impossible) to modify for the healthcare professionals who are supporting a mother. Social identification, on the other hand, is malleable and a promising target for intervention. A recent meta-analysis reviewed 26 interventions to build social identity, finding that these were effective at improving both mental and physical health (Steffens et al., 2021). These social identity building interventions are quite diverse in nature, ranging from design groups in a residential care home (Haslam et al., 2014) to a 10-day boat journey for disadvantaged youth (Scarf et al., 2016). One social identity intervention has been manualised, Groups 4 Health, and the benefits of this psychotherapeutic group programme have been found to persist for up to a year (Cruwys et al., 2022a), and to be robust to future unexpected threats to social connectedness (Cruwys et al., 2021). While no social identity interventions tailored to new mothers specifically have yet been evaluated, emerging quantitative evidence suggests that those mothers who are able to build shared social identities with other mothers have better wellbeing (Seppälä et al., 2022a). This research speaks to the wide-reaching potential that social identity interventions may have in this context and the need for future investment to develop and trial such approaches.
This study also has theoretical implications for our fundamental understanding of social relationships. While social identity is a group process and attachment an interpersonal one, they clearly have overlap as constructs which capture our deepest forms of commitment to and enmeshment with others. It is interesting, then, that this is one of the first studies to simultaneously examine both social identity and attachment processes. Our findings suggest that these processes are not at all independent, but instead can mutually reinforce or undermine one another, both within an individual and from mother to child. Indeed, the fact that social identification as a mother accounted for a substantial proportion of the variance in the effect of mother’s attachment history on mother-infant attachment suggests that this is a promising avenue for research seeking to explain the mechanisms of the intergenerational transmission of attachment style (the so-called “transmission gap”; Verhage et al., 2016). Our findings are also consistent with other recent work suggesting that social identity processes are relevant for the success of dyadic relationships (Cruwys et al., 2022b) and that social identities can function as a form of “secure base” (Klein & Bastian, 2022). Our findings add a voice to these calls for renewed attention on how we conceptualise social relationships and how these different levels of analysis interact with one another.
Strengths and Limitations
Strengths of the current study included the recruitment methodology and sophisticated management of missing data, which allowed us to access a diverse sample of mothers and provided adequate power to test our hypotheses. Many studies of new mothers rely on referrals from hospitals, community health services, or mother’s groups. These methods may exclude the least connected of mothers (Stuart et al., 2022), or alternatively, bias recruitment towards those with more complex health issues. Online research platforms, by contrast, show a wide distribution of incomes and backgrounds and the quality of data often compares favourably to recruitment using more traditional methodologies (Eyal et al., 2021). Another strength was the use of reliable and validated measures for the focal constructs. We also considered outcomes related to both mother and infant socio-emotional functioning, rather than only one or the other, with this being the first social identity-oriented study to look at infant-related outcomes.
A limitation of this study was its correlational design. We were conscious of this in choosing our predictor variables, which intentionally focused on risk factors prior to or during the infant’s birth so that we could be more confident of the direction of these effects. Nevertheless, these findings are best interpreted in the context of a broader literature that includes experimental and clinical trials evaluating the causal effects of social identity on health and wellbeing (e.g. Cruwys et al., 2022a). Another limitation was the reliance on mother’s self-report. While this is the most valid way to assess many of our constructs (e.g. depression; social identification), it does mean that common method variance may have been responsible for some proportion of the relationship among our constructs of interest. It also precluded the use of the most widely recommended instruments in some cases (e.g. for measuring attachment).
Conclusions
The first 12 months of life is a critical period in terms of trajectories for both mother and child. While a variety of risk factors for poor outcomes during this period have received attention in previous research, several problems remain, including (1) a “transmission gap” in understanding how prenatal risk factors come to influence infant functioning, and (2) the prenatal nature of such risk factors means that they are largely beyond the influence of maternal health professionals. The current study sought to address this with a focus on a candidate mediator—social identification—which is known to be highly responsive to intervention and has already shown links to key maternal indicators of socio-emotional functioning. Our findings suggest that social identity as a mother is a key mechanism through which risk factors affect post-natal trajectories for mothers and infants. Indeed, social identification as a mother fully accounted for the effect of risk factors on infant temperament and for the effect of birth trauma on mother-infant attachment. This speaks to the need to support women to perceive their identity as a mother in ways that are positive, efficacious, and self-defining. While this study focused on social identity as a mother, given the broader evidence for benefits from other social identities, alternative group memberships may also have a role to play, especially for women who feel ambivalent about their identity as mothers. Future research can capitalise on the malleability of social identity to develop tailored interventions that improve outcomes for mothers and babies.
Compliance with Ethical Standards
Conflict of interest
The authors declare no competing interests.
Ethics approval and consent
The Australian National University’s Human Research Ethics Committee approved the study (Protocol 2019/887) and all participants provided informed consent.
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 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/.
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Met BSL Psychologie Totaal blijf je als professional steeds op de hoogte van de nieuwste ontwikkelingen binnen jouw vak. Met het online abonnement heb je toegang tot een groot aantal boeken, protocollen, vaktijdschriften en e-learnings op het gebied van psychologie en psychiatrie. Zo kun je op je gemak en wanneer het jou het beste uitkomt verdiepen in jouw vakgebied.