Relationship Satisfaction Can Help to Maintain the Positive Effect of Childbirth on Parental Self-Esteem

Abstract

The transition to parenthood is accompanied by declined self-esteem levels, which may be explained by parents’ relationship satisfaction. However, prior research examined self-esteem only shortly before and after childbirth and had no or only unmatched childless respondents as a control group, limiting the possibility to examine long-term adaptive processes and the causal interpretation of the associations. Thus, we used panel data (10 years, 4,075 individuals, and 16,122 observations) to compare self-esteem and relationship satisfaction trajectories of parents with matched childless respondents using propensity score matching. We found a quadratic trajectory for parents’ self-esteem, which declined and increased before birth and declined and returned to baseline levels after birth. In contrast, matched childless respondents’ self-esteem decreased linearly before childbirth and then recovered. The quadratic postpartum process in parents was significantly associated with reduced relationship satisfaction. Thus, a fulfilling relationship may help to maintain the positive effects of childbirth on self-esteem in parents.

Keywords

self-esteem relationship satisfaction parenthood propensity score matching childbirth

Self-esteem, the individuals’ subjective evaluation of their worth as a person (Donnellan et al., 2011), is an important marker of mental and physical health (Sowislo & Orth, 2013) and psychological variables such as relationship satisfaction (RS; Erol & Orth, 2014). Although self-esteem is relatively stable across time in terms of rank-order stability and can thus be viewed as a personality trait (Anusic & Schimmack, 2016; Donnellan et al., 2012), there are substantial mean-level changes across the life span that correspond to a very large cumulative effect size (Orth et al., 2018). The development often follows a curvilinear pattern, with self-esteem increasing between ages 4 and 11, being stable between ages 11 and 15, increasing strongly until age 30 and less strongly until the peak between ages 60 and 70, and declining afterward (Orth et al., 2018). In addition to the systematic development of self-esteem on an aggregated level, major life transitions play an important role on the individual level in shaping self-esteem’s development by impacting cognitions and behavior (Orth & Robins, 2014). However, research on self-esteem changes in response to major life transitions is relatively scarce, with research showing higher self-esteem levels due to transitions to the first romantic relationship between the ages of 23 and 25 (Wagner et al., 2015) or to college education (Chung et al., 2014). Importantly, research has demonstrated an increase in 2 years before and a decline in self-esteem 2 years after childbirth (Bleidorn et al., 2016). In the present research, we wanted to expand on this research by (i) extending the sample period to 5 years before and after childbirth to examine longer term effects, (ii) using a well-matched sample of individuals who did not become parents in the span of 10 years, and (iii) investigating whether RS can explain changes in self-esteem.

Changes in Self-Esteem During the Transition to Parenthood

Bleidorn et al. (2016) examined self-esteem changes over the course of the first 5 years of marriage by analyzing Dutch longitudinal data. Of 187 couples, 55 reported at the end of the study that they currently did not have children, which constituted the control group of nonparents. The results demonstrated that whereas nonparents showed stable levels of self-esteem over the 5 years, parents showed an increase in self-esteem before childbirth that they subsequently could not maintain, as evidenced by a decrease in self-esteem in the 2 years following childbirth. Similarly, research shows that although parenthood had a negative main effect on self-esteem, this effect was positively moderated by age, such that new parents showed more increases in self-esteem over time than individuals without any children (Chen et al., 2016). The authors speculated that the cumulative experience of raising a child might eventually lead to increased levels of self-esteem. Moreover, a study on 80,000 Norwegian mothers reported short-term increases in their self-esteem in the first 6 months after childbirth and declines in self-esteem 3 years after (van Scheppingen et al., 2018). However, self-esteem did not decline after the mother had a second child during the following 3 years, indicating that the declines might not continue after a second child. Consequently, the first goal of the present research was to examine the long-term effects of the transition to parenthood on self-esteem.

However, testing the causality of the reported associations between life transitions and self-esteem is difficult, given that experimental designs would be needed, which are not possible in the context of parenthood (Luhmann et al., 2014). Although the presented studies used longitudinal designs, which allow to establish predictive causality (Granger, 1969), it is still difficult to gain evidence about causality from these comparisons given that they used unmatched or no control samples. Consequently, any change in self-esteem could, in addition to transition, be caused by confounding variables that have not been controlled for (Luciano & Orth, 2017). Thus, the second goal of the present study was to control for a large number of sociodemographic variables by using propensity score matching (PSM) to better test the effect of parenthood on self-esteem.

Changes in RS During the Transition to Parenthood

So, what can explain the short-term decline in self-esteem during the transition to parenthood? Some theoretical accounts consider physiological and neurological changes during the transition to motherhood, which affect mother’s hormonal levels and body image (Tiggemann, 2004). However, although Bleidorn et al. (2016) only found a steep decline in self-esteem directly after childbirth in new mothers but not in new fathers, both new mothers and fathers experienced a slow decline in self-esteem during the following 2 years. Another explanation would be that the decline in self-esteem may be due to the stressful demands that newly parents are confronted with, which negates the self-esteem boost prior to childbirth. Prior research has shown that the first childbirth is among the strongest factors that impact the RS of couples, given that new parents must reorganize their family system (Belsky et al., 1983) and their work and family roles with regard to dividing labor and household tasks once the first child is born (Sanchez & Thomson, 1997). This might not only negatively affect postpartum RS when the new challenges are difficult to overcome but might also affect prebirth RS by increasing experiences of cohesion before the arrival of the firstborn (Doss et al., 2009).

Evidence for the notion that the reorganization presents parents with new stressors and conflicts that can negatively impact their RS comes from research showing that new mothers reported more conflicts with their partners over their involvement and how they divide labor and household tasks after childbirth, which strained their romantic relationship (Moller et al., 2008). Importantly, van Scheppingen et al. (2018) found that self-esteem and RS were positively related around and after childbirth, demonstrating that the impact of stress on self-esteem due to the transition to motherhood may be mediated by RS. Thus, the third goal was to test the notion that declines in self-esteem are associated with declines in RS by examining longer term effects before and after childbirth in parents in comparison to a matched childless control group. However, self-esteem and RS correlated differently depending on the time after childbirth (van Scheppingen et al., 2018), and the transition to parenthood affects parents’ life in a number of ways besides their romantic relationship, which takes an unknown time to adapt to. Therefore, we computed a moderated mediation model to examine the indirect effect of time after birth on self-esteem via postpartum RS.

Method

Sample

We used data from the German Family Panel (pairfam), release 11.0 (Brüderl et al., 2020) provided by the pairfam administration (Huinink et al., 2011). The structure of and the variables in the data set are described in the data (Brüderl et al., 2020) and scale manual (Thönnissen et al., 2020). Pairfam is an annually and nationally representative collected sample of participants from three different cohorts, born between 1971 and 1973, 1981 and 1983, or 1991 and 1993. The analyses of the present research are based on data from the anchor participants who were recruited through a randomly selected address pool and who provided permission to interview them and their family, that is, partner, parents, or children. Adherence to the study protocol declined from Wave 1 to 10, b = −737 individuals per wave, SE = 87.4, 95% confidence interval (CI) [−939, −536], which was countered by the inclusion of over 5,000 new respondents to all cohorts in Wave 11.

We excluded a number of respondents in the final sample based on the following restrictions (Figure 1). First, we grouped respondents into those who (a) were already parents in the first wave in pairfam, (b) became parents in any wave, and (c) did not become parents in any wave. We then excluded the parents at Wave 1 (Group a), given that we were interested in the transition to parenthood. In the following, the term parents and nonparents refer to respondents who became or did not become parents during the pairfam waves. Second, we discarded data from respondents who had missing covariate data used for PSM, including RS and sociodemographic characteristics. Finally, we restricted the observations to 5 years before and after childbirth in order to match the observations between parents and nonparents and performed PSM. In order to compare the period after birth between parents and nonparents, the “birth” was set to the meantime of childbirth in the parent sample for nonparents, which was in Wave 7.

Figure 1.
Participant flowchart explaining which exclusion criteria led to the final analysis sample.

To examine the changes in self-esteem during the transition to parenthood, an appropriate control group of nonparents is needed to disentangle the changes due to the transition from other factors such as age-related changes or selection effects (Luhmann et al., 2014). PSM estimates a propensity score for each respondent in the sample based on a chosen number of covariates, which indicates the likelihood of a respondent to transition to parenthood (Stuart, 2010). Then, it matches nonparents with parents with a propensity score as similar as possible to balance the parents and nonparents group.

We estimated the propensity scores in a logistic regression predicting the occurrence of transitioning to parenthood (1 = parent, 0 = nonparent) by eight categorical (cohort, gender, ethnicity, homosexuality, highest school degree attained at the time of interview, highest vocational degree attained at the time of interview, size of municipality, time in years before childbirth, and time in years after childbirth) and two continuous covariates (age, RS). We used the two-nearest-neighbor matching procedure with possible replacement, which matched one nonparent to one or more parents with the most similar propensity score (Austin, 2011). To match all available parents, we did not specify a caliper (i.e., a prespecified threshold of a propensity score deviation between the parent and nonparent that must be met).

The final sample included 4,075 respondents (53.8% females) with a mean age of M = 30.7 years (SD = 6.3) and 3.95 observations on average per individual. The matching procedure was successful, in that the pseudo explained variance was reduced from R ² = 26.1% in the unmatched sample to R ² = 1.1% in the matched sample, indicating there were no longer systematic differences in the distribution of covariates between both groups. The likelihood-ratio test also indicated a better match, with χ² = 7921.8, p < .001 in the unmatched and χ² = 239.1, p < .001 in the matched sample. Regarding the individual covariates, Table 1 shows that differences between parents and nonparents were greatly reduced, except for ethnicity (graphical evaluation can be found in Supplementary Figure S1). Given the satisfactory if not perfect covariate balance achieved by the PSM, it can be concluded that any differences in self-esteem in the following analyses are likely due to the transition to parenthood.

Table 1.
Differences Between Nonparents and Parents Before and After Propensity Score Matching.

Covariate Unmatched Sample Matched Sample

χ² df Cramer’s V χ² df Cramer’s V

Categorical variables

Cohort 3100 3 .33 5.58 2 .02

Gender 3.28 1 −.01 0.40 1 .00

Ethnicity 1.74 4 .01 40.8 4 .05

Homosexuality 111 1 −.06 13.6 1 −.03

Highest school degree 754 8 .16 74.2 8 .07

Highest vocational degree 2100 8 .27 82.6 8 .07

Size of municipality 50.8 6 .04 27.7 6 .04

Years before childbirth 481 4 .13 11.5 4 .03

Years after childbirth 964 5 .18 22.5 5 .04

Continuous Variables t df Cohen’s d t df Cohen’s d

Age −41.2 28529 −.54 2.29 16120 .04

Relationship satisfaction 1.53 28529 .02 −1.15 16120 −.02

Note. Estimates in bold are significant at p < .005.

Measures

Self-esteem

Self-esteem was annually evaluated from Wave 1 (2008–2009) to Wave 11 (2018–2019) with three to four items from the German version of the Self-Esteem Scale (Rosenberg, 1965), ranging on a scale from 1 (not at all) to 5 (absolutely). The original 10-item scale is a well-validated measure of self-esteem (Robins et al., 2001) and was reduced to three to four items in pairfam based on factor and reliability analyses. Given the multiple assessments per individual across the waves, we did not compute Cronbach’s α but report McDonald’s Ω (Geldhof et al., 2014; McNeish, 2018). Between-person reliability was very good (ω_between = .85), whereas within-person reliability was acceptable (ω_within = .62). With regard to sources of variances, self-esteem mostly varied between (ICC = 60.6%) and within respondents (ICC = 39.2%) but not much between cohorts (ICC = 0.02%).

Assessment of self-esteem in the pairfam data set switched from a computer-assisted personal interview in Wave 1 to computer-assisted self-administered interviews from Wave 2 onward. Possibly due to the switch, self-esteem scores in Wave 1 were significantly higher than in the other waves, b = 0.28, SE = 0.02, p < .001, 95% CI [0.23, 0.32]. Thus, we controlled for this method factor in all analyses by including a dummy variable (1 = Wave 1, 0 = all other waves).

RS

RS was annually assessed using computer-assisted self-administered interviews. All participants living in a partnership were interviewed about their RS, using one item (“All in all, how satisfied are you with your relationship”?) from the German version of the Relationship Assessment Scale (Sander & Böcker, 1993). The scale’s range was from 0 (very dissatisfied) to 10 (very satisfied). Variance in relationship was again mostly limited to variance between (ICC = 39.3%) and within respondents (ICC = 58.0%), with only very small variance between cohorts (ICC = 2.7%).

Analytic Approach

All analyses were performed in Stata Version 16 (Stata Corporation, College Station, TX), and the scripts for the sample compilation and analyses can be found at Open Science Framework (https://osf.io/bsmrn/). Given that the presented analyses are not preregistered and are thus affected by the high statistical flexibility such a rich data set provides, we used a conservative α = .005 for all statistical tests of significance to increase the robustness of our findings (Benjamin et al., 2018). Moreover, given the different response scales for self-esteem and RS, we used the percent of maximum possible (POMP) method (P. Cohen et al., 1999; Moeller, 2015) to facilitate the interpretation and comparison of the measures. For example, the self-esteem scale ranged from 1 to 5. To obtain the POMP score, 1 was subtracted from each self-esteem value, leading to a scale ranging from 0 to 4. Then, each score was divided by the maximum score (4) and multiplied by 100. All models were estimated with a random intercept for cohorts and respondents but no random slopes.

Changes in self-esteem before and after childbirth

Given the nested structure of the data, with observations i (Level 1) nested within respondents j (Level 2) nested within cohorts k (Level 3), we used multilevel modeling to model changes in self-esteem over time in relation to childbirth. To that end, we first predicted self-esteem by the two-way interaction between parents (0 = nonparents, 1 = parents) and the categorical variable time in years, ranging from 0 to 9 with childbirth occurring between Time Points 4 and 5 (Equation 1).

$\begin{array}{l} y_{i j k} = β_{000} + β_{100} {Parent}_{0 j k} + β_{200} {Time}_{i j k} + β_{300} Parent \times {Time}_{i j k} \\ + β_{400} Method {Correction}_{i j k} + μ_{00 j}^{Cohort} + μ_{1 j k}^{Respondents} + ε_{i j k} . \end{array}$
1

We then estimated and plotted the marginal means of self-esteem at each time point separately for parents and nonparents. To disentangle the change before and after childbirth, we replaced the time variable with two categorical piecewise variables, one concerning the change before and the other after childbirth (see Supplementary Table S1 for the coding process of the piecewise variables).

$\begin{array}{l} y_{i j k} = β_{000} + β_{100} {Parent}_{0 j k} + β_{200} {Prebirth}_{i j k} + β_{300} \\ {Postbirth}_{i j k} + β_{400} Parent \times {Prebirth}_{i j k} + β_{500} \\ Parent \times {Postbirth}_{i j k} + β_{600} Method {Correction}_{i j k} \\ + μ_{00 j}^{Cohort} + μ_{1 j k}^{Respondents} + ε_{i j k} . \end{array}$
2

We then estimated orthogonal polynomial contrasts (linear and quadratic) to model linear and nonlinear pre- and postbirth changes within parents and nonparents.

Changes in RS satisfaction before and after childbirth

The models for RS were the same as for self-esteem with the exception of the exclusion of the method correction dummy variable.

RS as a moderator of changes in self-esteem after childbirth

For the moderator analysis, we extended the model depicted by Equation 2 by including within-respondent-centered RS and the respective respondent mean (meanRS) to model three-way interactions in which each variable interacted with parents and time before or after birth (Equation 3):

$\begin{array}{l} y_{i j k} = β_{000} + β_{100} {Parent}_{0 j k} + β_{200} {Prebirth}_{i j k} + β_{300} \\ {Postbirth}_{i j k} + β_{400} Parent \times {Prebirth}_{i j k} + β_{500} \\ Parent \times {Postbirth}_{i j k} + β_{600} {centeredRS}_{i j k} + β_{700} \\ {meanRS}_{0 j k} + β_{800} centeredRS \times {Prebirth}_{i j k} + β_{900} \\ centeredRS \times {Postbirth}_{i j k} + β_{1000} meanRS \times {Prebirth}_{0 j k} \\ + β_{1100} meanRS \times {Postbirth}_{0 j k} + β_{1200} centeredRS \\ \times {Parent}_{i j k} + β_{1300} meanRS \times {Parent}_{0 j k} + β_{1400} centeredRS \\ \times Prebirth \times {Parent}_{i j k} + β_{1500} centeredRS \times Postbirth \\ \times {Parent}_{i j k} + β_{1600} meanRS \times Prebirth \times {Parent}_{i j k} \\ + β_{1700} meanRS \times Postbirth \times {Parent}_{i j k} \\ + β_{1800} Method {Correction}_{i j k} + μ_{00 j}^{Cohort} + μ_{1 j k}^{Respondents} + ε_{i j k} . \end{array}$
3

RS as a mediator of changes in self-esteem after childbirth

To test the indirect effect of time after childbirth on self-esteem via RS, we used a bootstrapping procedure (Preacher & Hayes, 2008) with 500 bootstrap samples, bias-correction, and acceleration of the CIs.

Results

Changes in Self-Esteem Before and After Childbirth

Figure 2 illustrates the changes in self-esteem from year to year for parents and nonparents. Piecewise multilevel analysis revealed significant differences in pre- and postbirth changes between parents and nonparents. For linear prebirth changes, we found a significant two-way interaction with parenthood, χ²(1) = 11.18, p < .001, such that only nonparents reported a linear decline in self-esteem in the 5 years before the matched transition point, b = −1.03, SE = 0.27, p < .001 99.5% CI [−1.80, −0.26], but not parents, b = 0.17, SE = 0.23, p = .468, 99.5% CI [−0.48, 0.82]. Parents’ trajectories could be described by a quadratic relationship, b = 0.56, SE = 0.22, p = .012, 99.5% CI [−0.06, 1.18], with a significant two-way interaction between the parent variable and prebirth change, χ²(1) = 9.34, p = .002. Thus, self-esteem first declined in parents and then recovered in the year before childbirth, which was further underlined by a significant jump from Time Points 3 to 4, the year before childbirth, b = 1.96, SE = 0.56, p = .001, 99.5% CI [0.37, 3.54].

Figure 2.
Estimated marginal means of self-esteem and their 99.5% confidence intervals as a function of time and parenthood. For the nonparents sample, the 10 assessments were centered around childbirth (i.e., between the two dotted vertical lines).

Regarding linear postbirth changes, the two-way interaction with parenthood was again significant, χ²(1) = 11.25, p < .001. Whereas nonparents’ self-esteem significantly returned to baseline levels after childbirth, b = 0.66, SE = 0.23, p = .005, 99.5% CI [<0.01, 1.31], parents’ self-esteem declined, b = −0.35, SE = 0.19, p = .063, 99.5% CI [−0.87, 0.18], but this decline was not significant. Instead, parents’ trajectories could again be best described via a quadratic relationship, such that self-esteem levels firstly declined but then slowly recovered in the last 4 years in the study period. Importantly, both the quadratic slope for parents, b = 0.50, SE = 0.18, p = .005, 99.5% CI [<0.01, 1.00], and also the two-way interaction between time after childbirth and parenthood, χ²(1) = 10.09, p = .002, were significant.

Changes in RS Before and After Childbirth

Figure 3 shows changes in RS from year to year for parents and nonparents. The trajectories were similar to the ones of self-esteem, with a significant difference in linear change between parents and nonparents, χ²(1) = 18.59, p < .001. For nonparents, there was a linear prebirth decline, b = −1.35, SE = 0.34, p < .001, 99.5% CI [−2.30, −0.40], but a small linear increase for parents, b = 0.60, SE = 0.30, p = .045, 99.5% CI [−0.24, 1.45], that also similarly fit a quadratic relationship, b = 0.70, SE = 0.29, p = .016, 99.5% CI [−0.12, 1.51]. Postbirth changes were best described as linear and significantly differed between parents and nonparents, χ²(1) = 78.15, p < .001. Whereas nonparents reported a small increase in RS after the matched transition point, b = 0.70, SE = 0.29, p = .026, 99.5% CI [−0.14, 1.49], which was not significant, parents reported a highly significant, steep decline of approximately 2.67% of the possible maximum score per year after childbirth, b = −2.67, SE = 0.24, p < .001, 99.5% CI [−3.34, −1.99]. The decline was strongest after childbirth at Time Point 5, b = −4.35, SE = 0.63, p < .001, 99.5% CI [−6.12, −2.58], and attenuated afterward.

Figure 3.
Estimated marginal means of relationship satisfaction and their 99.5% confidence intervals as a function of time and parenthood. For the nonparents sample, the 10 assessments were centered around childbirth (i.e., between the two dotted vertical lines).

Figure 3 demonstrates that RS levels were higher before childbirth and lower postpartum in parents compared to nonparents. To test these differences, we computed a multilevel model to predict RS by the two-way interaction between parenthood and time of childbirth (0 = before, i.e., Time Points 0–4, 1 = after, i.e., Time Points 5–9). The model yielded a significant interaction, b = −5.05, SE = 0.58, p < .001, 99.5% CI [−6.67, −3.42]. The analysis of the marginal means revealed that parents’ RS before childbirth (M = 84.0, SE = 0.85) was significantly higher than after childbirth (M = 78.6, SE = 0.81), b = 5.40, SE = 0.38, p < .001, 99.5% CI [6.15, 4.66], or than nonparents’ levels before (M = 79.2, SE = 0.83), b = 4.79, SE = 0.66, p < .001, 99.5% CI [3.49, 6.09], and after childbirth (M = 78.8, SE = 0.83), b = 5.15, SE = 0.66, p < .001, 99.5% CI [3.85, 6.44]. However, the postpartum differences between parents and nonparents were not significant, b = −0.25, SE = 0.61, p = .679, 99.5% CI [−1.46, 0.95].

RS as a Moderator of Changes in Self-Esteem After Childbirth

Next, we investigated whether RS can explain the increase in self-esteem before childbirth or can buffer the declining postbirth self-esteem levels. We found a positive two-way interaction between linear time after childbirth and respondent mean-centered RS, b = 0.04, SE = 0.01, p < .001, 99.5% CI [0.01, 0.07]. Although the three-way interaction with parenthood was not significant, χ²(1) = 0.70, p = .403, we found that the two-way interaction between linear time after childbirth and respondent-centered RS was only significant for parents, b = 0.05, SE = 0.01, p < .001, 95% CI [0.01, 0.08], but not for nonparents, b = 0.03, SE = 0.02, p = .120, 95% CI [−0.02, 0.08]. Thus, we will focus on the decline in self-esteem in parents after childbirth in the following analyses.

To interpret the significant two-way interaction in parents, we used the Johnson–Neyman technique (J–N), which is a generalization of the simple slope analysis (J. Cohen et al., 2003) and avoids picking usually arbitrary conditional values for probing the interaction. Instead, J–N explores the magnitude and significance of the relationship between the predictors and self-esteem at a wide range of conditional values of within-respondent-centered RS (Bauer & Curran, 2005). Figure 4 shows that the region of no significance, that is, the region where the upper and lower bounds of the simple slope intersect the x-axis, was between −19.1% and 15.9% of respondent-centered RS values. Thus, the association of self-esteem on postbirth change was negative and significant at values of RS smaller than −19.1%, not significantly different from zero at values of relation satisfaction between −19.1% and 15.9%, and positive and significant at values greater than 15.9%. Both the lower and upper region fell within the observed range of RS of −90.0% and 64.0%. Moreover, 330 of the 4,903 (6.7%) postbirth RS values in parents were greater than the upper bound of the region and 290 (5.9%) were smaller than the lower bound.

Figure 4.
J–N regions of significance (gray areas) and 99.5% confidence bands (gray lines) for the conditional relation between self-esteem and time after childbirth as a function of respondent-centered relationship satisfaction (black line).

RS as a Mediator of Changes in Self-Esteem After Childbirth

Given the significant two-way moderation between postbirth time and RS on self-esteem, we computed a moderated mediation model to examine whether RS can explain the quadratic changes in self-esteem in parents after childbirth. In our moderated mediation model, the effect of quadratic postbirth changes in self-esteem was mediated by RS and the path between RS and self-esteem was moderated by time after childbirth (Figure 5). As indicated in Table 2, the total effect of time on the quadratic postbirth changes in self-esteem was, as reported before, b = 0.020, and depended on the years after childbirth, with 40.5% of the total effect being mediated in the first year after childbirth to 87.1% in the fifth year.

Figure 5.
Moderated mediation model for the effect of postbirth time on self-esteem via relationship satisfaction, where the path between relationship satisfaction and self-esteem is moderated by postbirth time. Estimates in bold are significant at p < .005.

Table 2.
Indirect and Direct Effects of the Moderated Mediation Model With Bias-Corrected and Accelerated (bca) 99.5% Confidence Intervals.

Effect 1 Year After Childbirth 3 Years After Childbirth 5 Years After Childbirth

b (SE) 99.5% bca CI b (SE) 99.5% bca CI b (SE) 99.5% bca CI

Indirect effect .20 (.05) [.08, .36] .32 (.10) [.13, .60] .44 (.15) [.13, .89]

Direct effect .30 (.19) [−.19, .76] .18 (.20) [−.36, .63] .06 (.23) [−.53, .59]

Total effect .50 (.19) [−.03, .95] .50 (.19) [−.03, .95] .50 (.19) [−.03, .95]

Note. Estimates in bold are significant at p < .005.

Discussion

Prior research has found that the transition to parenthood is accompanied by a decline of self-esteem (Bleidorn et al., 2016), which could possibly be explained by the straining effects of parenthood on parents’ romantic relationship (van Scheppingen et al., 2018). However, the sample period in prior research was often relatively short, limiting the possibility to examine long-term processes. For example, the new challenges of parenthood could negatively impact parents’ self-esteem in the short-term but could also provide the potential for growth in the long-term when successfully and collectively adapting to the new challenges. Moreover, prior research had no or only unmatched childless respondents, which limits the causal interpretation of the association. By using data from the large 11-wave longitudinal family panel data set pairfam, we compared the trajectories of self-esteem and RS of parents with matched childless respondents using PSM. In the following, we summarize the key results of the present study and discuss how our research connects to research on self-esteem.

Our results showed that parents’ self-esteem levels jumped before childbirth and that postpartum levels followed a quadratic function, that is, parents’ self-esteem levels first declined before returning to baseline levels. Together with the relatively small changes, our results demonstrate that parents could not maintain the boost to their self-esteem before childbirth and that the declining levels of self-esteem during the 2 years after childbirth are best viewed as a return to their usual self-esteem levels. This pattern is in line with other research showing that life satisfaction peaked just after childbirth and then returned to baseline levels 2 years after (Dyrdal & Lucas, 2013). Given that prior research showed that self-esteem predicted life satisfaction 8 months later, but life satisfaction did not predict change in self-esteem (Ye et al., 2012), increases in self-esteem around childbirth could explain increases in well-being, too.

Interestingly, we found that nonparents’ self-esteem trajectory also followed a quadratic function: Whereas self-esteem levels were similar compared to matched parents at the beginning, self-esteem levels decreased in matched nonparents before the matched point of transition and then slowly recovered to baseline levels. Given that we matched parents to nonparents on the basis of a similar propensity to become parents, this could indicate that not becoming a parent in times when one’s peers transition to parenthood could potentially affect one’s self-esteem. This would be in line with research showing that parenthood is connected to the fertility of members of one’s social network (Bernardi & Klärner, 2014), such that the likelihood of transitioning to parenthood increases significantly after close others such as siblings give birth to a child (Lyngstad & Prskawetz, 2010). Thus, future research could pick up this potentially interesting pattern to the role of social networks in self-esteem development.

The trajectory of RS in parents was less positive than the trajectory of self-esteem: Parents reported higher levels of RS over the 5 years before childbirth compared to matched childless respondents. The differences of 5% of the possible maximum score in RS between parents and nonparents were highly significant and relatively pronounced. However, parents could not maintain these high levels after 5 years of childbirth with a decline of 5% of the possible maximum score on average and a very small and nonsignificant difference compared to nonparents.

We found that parents who could protect their romantic relationship against the straining challenges of early parenthood did not experience a postpartum decline in self-esteem and were, thus, able to maintain the boosting effect of anticipated parenthood prior to childbirth on self-esteem 5 years after childbirth. Furthermore, our results of the moderated mediation analysis showed that the indirect quadratic effect of time after childbirth on self-esteem via RS increased over time, demonstrating that the importance of RS for parents’ self-esteem did not decrease but increase with the time after childbirth. Thus, the increase in self-esteem in the 5th year after childbirth was very closely related to RS.

Unlike prior research, which found that the transition to parenthood has a significant impact on especially mother’s self-esteem (Bleidorn et al., 2016), we did not find any gender differences in our examined associations (see Supplementary Table S2). One reason for the diverging results could be differences in parental leave schemes in the countries where the assessments took place. In Germany, where the pairfam data were collected, a new parental leave legislation was implemented in 2007, which aimed to nudge fathers’ participation in childcare (Reimer, 2019) and provided financial coverage when mothers or fathers care for their children in the first year after birth. However, Bleidorn et al. (2016) used data from Dutch individuals: The Dutch parental leave scheme was unpaid at the time of data collection and did not foresee a job guarantee upon return (Begall & Grunow, 2015). Consequently, it is likely that families in Germany more evenly divided their childcare and childrearing responsibilities, so that mothers and fathers were more equally affected by factors that impacted their self-esteem around childbirth. In fact, the proportion of fathers on paid parental leave in Germany has increased from 21% in 2008 to 36% in 2015 (Väterreport, 2018). Future research could pick up this line of research in order to test how different parental leave systems impact mothers’ and fathers’ self-esteem development in the transition to parenthood.

The present study has a number of advantages, such as a large and relatively diverse sample across age and gender, a relatively large sample period, and the use of a matched control group of childless respondents. Of course, such research is not without disadvantages, which should be acknowledged when evaluating the validity of the presented results. We focused on RS as mediator of the change in parents’ self-esteem, but parenthood does not only mean change in the parents’ relationship but gaining a new relationship to the newborn. How the parent–child relationship develops and is characterized might also play a role in how parents’ self-esteem changes. Thus, future research could include characteristics of this relationship and of the child, for example, the child’s temperament or parenting styles. Moreover, due to the restricted in panel assessments, some of the scales were shortened for usage in the pairfam project. For example, self-esteem was only assessed with three items of the original 10-item Rosenberg Self-esteem Scale (Rosenberg, 1965). These short scales limit the reliability of the scales, as evidenced by only acceptable internal consistencies. Finally, given the limitations of mediation analyses in general, we can only draw limited conclusions that RS is the mediator in the self-esteem trajectory during the transition to parenthood. Instead, we can be more certain that self-esteem and RS co-develop and influence each other over time. Thus, more research is needed that tests the directionality of this relationship and that examines third factors such as parental experiences, other environmental factors, or effects of genes, which is very difficult to do in observational research.

Taken together, our results demonstrate that couples who can maintain a healthy relationship during the rough early years of parenthood can sustain the positive effects of childbirth on their self-esteem and thus can use it as an additional source of self-esteem, which could receive increased attention in interventions to prevent disturbances as a result of reduced self-esteem in new parents.

Supplemental Material

Supplemental Material, Supplementary_Materials(SPP) - Relationship Satisfaction Can Help to Maintain the Positive Effect of Childbirth on Parental Self-Esteem

Supplemental Material, Supplementary_Materials(SPP) for Relationship Satisfaction Can Help to Maintain the Positive Effect of Childbirth on Parental Self-Esteem by Mario Wenzel, Doris Staab, Zarah Rowland and Manon A. van Scheppingen in Social Psychological and Personality Science

Covariate	Unmatched Sample	Matched Sample
Categorical variables
Cohort	3100	3	.33	5.58	2	.02
Gender	3.28	1	−.01	0.40	1	.00
Ethnicity	1.74	4	.01	40.8	4	.05
Homosexuality	111	1	−.06	13.6	1	−.03
Highest school degree	754	8	.16	74.2	8	.07
Highest vocational degree	2100	8	.27	82.6	8	.07
Size of municipality	50.8	6	.04	27.7	6	.04
Years before childbirth	481	4	.13	11.5	4	.03
Years after childbirth	964	5	.18	22.5	5	.04
Continuous Variables	t	df	Cohen’s d	t	df	Cohen’s d
Age	−41.2	28529	−.54	2.29	16120	.04
Relationship satisfaction	1.53	28529	.02	−1.15	16120	−.02

Effect	1 Year After Childbirth	3 Years After Childbirth	5 Years After Childbirth
Indirect effect	.20 (.05)	[.08, .36]	.32 (.10)	[.13, .60]	.44 (.15)	[.13, .89]
Direct effect	.30 (.19)	[−.19, .76]	.18 (.20)	[−.36, .63]	.06 (.23)	[−.53, .59]
Total effect	.50 (.19)	[−.03, .95]	.50 (.19)	[−.03, .95]	.50 (.19)	[−.03, .95]

Footnotes

Acknowledgments

This paper uses data from the German Family Panel pairfam, coordinated by Josef Brüderl, Sonja Drobnič, Karsten Hank, Franz Neyer, and Sabine Walper. pairfam is funded as long-term project by the German Research Foundation (DFG).

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The panel data received for this article was funded by the Deutsche Forschungsgemeinschaft (German Research Foundation).

ORCID iD

Mario Wenzel

Supplemental Material

The supplemental material is available in the online version of the article.

References

Anusic

Schimmack

(2016). Stability and change of personality traits, self-esteem, and well-being: Introducing the meta-analytic stability and change model of retest correlations. Journal of Personality and Social Psychology, 110(5), 766–781. https://doi.org/10.1037/pspp0000066

Austin

P. C.

(2011). An introduction to propensity score methods for reducing the effects of confounding in observational studies. Multivariate Behavioral Research, 46(3), 399–424. https://doi.org/10.1080/00273171.2011.568786

Bauer

D. J.

Curran

P. J.

(2005). Probing interactions in fixed and multilevel regression: Inferential and graphical techniques. Multivariate Behavioral Research, 40(3), 373–400. https://doi.org/10.1207/s15327906mbr4003_5

Begall

Grunow

(2015). Labour force transitions around first childbirth in the Netherlands. European Sociological Review, 31(6), 697–712.

Belsky

Spanier

G. B.

Rovine

(1983). Stability and change in marriage across the transition to parenthood. Journal of Marriage and Family, 45(3), 567–577. https://doi.org/10.2307/351661

Benjamin

D. J.

Berger

J. O.

Johannesson

Nosek

B. A.

Wagenmakers

E.-J.

Berk

Bollen

K. A.

Brembs

Brown

Camerer

Cesarini

Chambers

C. D.

Clyde

Cook

T. D.

De Boeck

Dienes

Dreber

Easwaran

Efferson

… Johnson

V. E.

(2018). Redefine statistical significance. Nature Human Behaviour, 2(1), 6–10. https://doi.org/10.1038/s41562-017-0189-z

Bernardi

Klärner

(2014). Social networks and fertility. Demographic Research, 30(22), 641–670. https://doi.org/10.4054/DemRes.2014.30.22

Bleidorn

Buyukcan-Tetik

Schwaba

van Scheppingen

M. A.

Denissen

J. J. A.

Finkenauer

(2016). Stability and change in self-esteem during the transition to parenthood. Social Psychological and Personality Science, 7(6), 560–569. https://doi.org/10.1177/1948550616646428

Brüderl

Drobnič

Karsten

Neyer

F. J.

Walper

Alt

Bozoyan

Finn

Frister

Garrett

Gonzalez

Greischel

Gröpler

Hajek

Herzig

Huyer-May

Lenke

Minkus

Peter

… Wilhelm

(2020). The german family panel (pairfam). GESIS Data Archive, Cologne. ZA5678 Data file Version 11.0.0, https://doi.org/10.4232/pairfam.5678.11.0.0

10.

Chen

E. Y.-J.

Enright

R. D.

Tung

E. Y.-L.

(2016). The influence of family unions and parenthood transitions on self-development. Journal of Family Psychology: JFP: Journal of the Division of Family Psychology of the American Psychological Association (Division 43), 30(3), 341–352. https://doi.org/10.1037/fam0000154

11.

Chung

J. M.

Robins

R. W.

Trzesniewski

K. H.

Noftle

E. E.

Roberts

B. W.

Widaman

K. F.

(2014). Continuity and change in self-esteem during emerging adulthood. Journal of Personality and Social Psychology, 106(3), 469–483. https://doi.org/10.1037/a0035135

12.

Cohen

West

S. G.

Aiken

L. S.

(2003). Applied multiple regression/correlation analysis for the behavioral sciences (3rd ed.). Lawrence Erlbaum Associates.

13.

Cohen

Aiken

L. S.

West

S. G.

(1999). The problem of units and the circumstance for POMP. Multivariate Behavioral Research, 34(3), 315–346. https://doi.org/10.1207/S15327906MBR3403_2

14.

Donnellan

M. B.

Kenny

D. A.

Trzesniewski

K. H.

Lucas

R. E.

Conger

R. D.

(2012). Using trait-state models to evaluate the longitudinal consistency of global self-esteem from adolescence to adulthood. Journal of Research in Personality, 46, 634–645.

15.

Donnellan

M. B.

Trzesniewski

K. H.

Robins

R. W.

(2011). Self-esteem: Enduring issues and controversies. In Chamorro-Premuzic

von Stumm

Furnham

(Eds.), The Wiley-Blackwell handbook of individual differences (pp. 718–746). Wiley-Blackwell.

16.

Doss

B. D.

Rhoades

G. K.

Stanley

S. M.

Markman

H. J.

(2009). The effect of the transition to parenthood on relationship quality: An eight-year prospective study. Journal of Personality and Social Psychology, 96(3), 601–619. https://doi.org/10.1037/a0013969

17.

Dyrdal

G. M.

Lucas

R. E.

(2013). Reaction and adaptation to the birth of a child: A couple-level analysis. Developmental Psychology, 49(4), 749–761. https://doi.org/10.1037/a0028335

18.

Erol

R. Y.

Orth

(2014). Development of self-esteem and relationship satisfaction in couples: Two longitudinal studies. Developmental Psychology, 50(9), 2291–2303. https://doi.org/10.1037/a0037370

19.

Geldhof

G. J.

Preacher

K. J.

Zyphur

M. J.

(2014). Reliability estimation in a multilevel confirmatory factor analysis framework. Psychological Methods, 19(1), 72–91. https://doi.org/10.1037/a0032138

20.

Granger

C. W. J.

(1969). Investigating causal relations by econometric models and cross-spectral methods. Econometrica, 37(3), 424–438. https://doi.org/10.2307/1912791

21.

Huinink

Brüderl

Nauck

Walper

Castiglioni

Feldhaus

(2011). Panel analysis of intimate relationships and family dynamics (pairfam): Conceptual framework and design. Zeitschrift für Familienforschung, 23(1), 77–101.

22.

Luciano

E. C.

Orth

(2017). Transitions in romantic relationships and development of self-esteem. Journal of Personality and Social Psychology, 112, 307–328. http://dx.doiorg/10.1037/pspp0000109

23.

Luhmann

Orth

Specht

Kandler

Lucas

R. E.

(2014). Studying changes in life circumstances and personality: It’s about time. European Journal of Personality, 28(3), 256–266. https://doi.org/10.1002/per.1951

24.

Lyngstad

T. H.

Prskawetz

(2010). Do siblings’ fertility decisions influence each other? Demography, 47(4), 923–934. https://doi.org/10.1007/BF03213733

25.

McNeish

(2018). Thanks coefficient alpha, we’ll take it from here. Psychological Methods, 23(3), 412–433. https://doi.org/10.1037/met0000144

26.

Moeller

(2015). A word on standardization in longitudinal studies: Don’t. Frontiers in Psychology, 6. https://doi.org/10.3389/fpsyg.2015.01389

27.

Moller

Hwang

P. C.

Wickberg

(2008). Couple relationship and transition to parenthood: Does workload at home matter? Journal of Reproductive and Infant Psychology, 26(1), 57–68. https://doi.org/10.1080/02646830701355782

28.

Orth

Erol

R. Y.

Luciano

E. C.

(2018). Development of self-esteem from age 4 to 94 years: A meta-analysis of longitudinal studies. Psychological Bulletin, 144(10), 1045–1080. https://doi.org/10.1037/bul0000161

29.

Orth

Robins

R. W.

(2014). The development of self-esteem. Current Directions in Psychological Science, 23(5), 381–387. https://dx-doi-org.web.bisu.edu.cn/10.1177/0963721414547414

30.

Preacher

K. J.

Hayes

A. F.

(2008). Asymptotic and resampling strategies for assessing and comparing indirect effects in multiple mediator models. Behavior Research Methods, 40(3), 879–891. https://doi.org/10.3758/BRM.40.3.879

31.

Reimer

(2019). Why fathers don’t take more parental leave in Germany: Comparing mechanisms in different work organizations. Community, Work & Family, 1–20. https://doi.org/10.1080/13668803.2019.1608157

32.

Robins

R. W.

Hendin

H. M.

Trzesniewski

K. H.

(2001). Measuring global self-esteem: Construct validation of a single-item measure and the Rosenberg self-esteem scale. Personality and Social Psychology Bulletin, 27(2), 151–161.

33.

Rosenberg

(1965). Society and the adolescent self-image. Princeton University Press.

34.

Sanchez

Thomson

(1997). Becoming mothers and fathers: Parenthood, gender, and the division of labor. Gender & Society, 11(6), 747–772. https://doi.org/10.1177/089124397011006003

35.

Sander

Böcker

(1993). Die deutsche form der relationship assessment scale (RAS): Eine kurze skala zur messung der zufriedenheit in einer partnerschaft [The German version of the Relationship Assessment Scale (RAS): A short scale for measuring satisfaction in a dyadic relationship]. Diagnostica, 39(1), 55–62.

36.

Sowislo

J. F.

Orth

(2013). Does low self-esteem predict depression and anxiety? A meta-analysis of longitudinal studies. Psychological Bulletin, 139(1), 213–240. https://doi.org/10.1037/a0028931

37.

Stuart

E. A.

(2010). Matching methods for causal inference: A review and a look forward. Statistical Science : A Review Journal of the Institute of Mathematical Statistics, 25(1), 1–21. https://doi.org/10.1214/09-STS313

38.

Thönnissen

Wilhelm

Alt

Reim

Walper

(2020). Pairfam scales and instruments manual. Release 11.0. LMU Munich: Technical report. GESIS Data Archive, Cologne. ZA5678 Data File Version 11.0.0, https://doi.org/10.4232/pairfam.5678.11.0.0

39.

Tiggemann

(2004). Body image across the adult life span: Stability and change. Body Image, 1(1), 29–41. https://doi.org/10.1016/S1740-1445(03)00002-0

40.

van Scheppingen

M. A.

Denissen

J. J. A.

Chung

J. M.

Tambs

Bleidorn

(2018). Self-esteem and relationship satisfaction during the transition to motherhood. Journal of Personality and Social Psychology, 114(6), 973–991. https://doi.org/10.1037/pspp0000156

41.

Väterreport [Father report] . (2018). Bundesministerium für Familie, Senioren, Frauen und Jugend. https://www.bmfsfj.de/blob/127268/2098ed4343ad836b2f0534146ce59028/vaeterreport-2018-data.pdf

42.

Wagner

Becker

Lüdtke

Trautwein

(2015). The first partnership experience and personality development: A propensity score matching study in young adulthood. Social Psychological and Personality Science, 6(4), 455–463. https://doi.org/10.1177/1948550614566092

43.

K.-K.

(2012). A cross-lagged model of self-esteem and life satisfaction: Gender differences among Chinese university students. Personality and Individual Differences, 52(4), 546–551. https://doi.org/10.1016/j.paid.2011.11.018

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.44 MB

Covariate	Unmatched Sample			Matched Sample
Covariate	χ²	df	Cramer’s V	χ²	df	Cramer’s V
Categorical variables
Cohort	3100	3	.33	5.58	2	.02
Gender	3.28	1	−.01	0.40	1	.00
Ethnicity	1.74	4	.01	40.8	4	.05
Homosexuality	111	1	−.06	13.6	1	−.03
Highest school degree	754	8	.16	74.2	8	.07
Highest vocational degree	2100	8	.27	82.6	8	.07
Size of municipality	50.8	6	.04	27.7	6	.04
Years before childbirth	481	4	.13	11.5	4	.03
Years after childbirth	964	5	.18	22.5	5	.04
Continuous Variables	t	df	Cohen’s d	t	df	Cohen’s d
Age	−41.2	28529	−.54	2.29	16120	.04
Relationship satisfaction	1.53	28529	.02	−1.15	16120	−.02

Effect	1 Year After Childbirth		3 Years After Childbirth		5 Years After Childbirth
Effect	b (SE)	99.5% bca CI	b (SE)	99.5% bca CI	b (SE)	99.5% bca CI
Indirect effect	.20 (.05)	[.08, .36]	.32 (.10)	[.13, .60]	.44 (.15)	[.13, .89]
Direct effect	.30 (.19)	[−.19, .76]	.18 (.20)	[−.36, .63]	.06 (.23)	[−.53, .59]
Total effect	.50 (.19)	[−.03, .95]	.50 (.19)	[−.03, .95]	.50 (.19)	[−.03, .95]