Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2026 Feb 23.
Published in final edited form as: Dev Psychol. 2025 Jul 10;61(12):2396–2405. doi: 10.1037/dev0002023

Early and Later Susceptibility to Effects of Parenting: Exploring Frankenhuis and Panchanathan’s (2011) Evolutionary Commitment Hypothesis

Xiaoya Zhang 1, Laura Machlin 2, Katie McLaughlin 2,3, Jay Belsky 4
PMCID: PMC12925650  NIHMSID: NIHMS2138358  PMID: 40638297

Abstract

This study tests Frankenhuis and Panchanathan’s (2011) evolutionary commitment hypothesis by examining individual differences in susceptibility to the effects of parenting during early and late childhood. Drawing from the Future of Families and Child Well-Being Study (N = 4,898; 2,342 girls; White/Black/Latinx/other = 590/1,601/813/261) and the National Institute of Child Health and Human Development Study of Early Child Care and Youth Development (N = 1,364; 659 girls; White/non-Hispanic = 1,097, Black = 176, other = 91), we investigated whether the consistency of early-life conditions affects later susceptibility to parenting with respect to children’s cognitive functioning and problem behavior. We hypothesized that children with more consistent early-life experiences of threat will show decreased susceptibility to later parenting (H1), whereas those exposed to greater unpredictability of income will demonstrate increased susceptibility (H2). Susceptibility was measured using influence statistics. While empirical support was not forthcoming in the case of the National Institute of Child Health and Human Development Study, there was evidence in the Future of Families and Child Wellbeing Study consistent with H2 (but not H1). Despite the limited evidence for the evolutionary commitment hypothesis, more research is needed before strong conclusions can be drawn about the evolutionary commitment hypothesis.

Keywords: differential susceptibility, developmental plasticity, evolutionary commitment hypothesis, threat, unpredictability

Recent years have seen an outpouring of evidence documenting individual differences in susceptibility to both positive and negative environmental influences (i.e., for better and for worse). This is evident in work guided by Belsky’s differential susceptibility hypothesis (Belsky, 1997a, 1997b, in press; Belsky et al., 2007; Belsky & Pluess, 2009, 2013), Boyce and Ellis’s (2005) analysis of biological sensitivity to context (e.g., Ellis et al., 2011; Shakiba et al., 2020), and Aron and Aron’s (1997) theory of the highly sensitive person (e.g., Lionetti et al., 2018; Listou Grimen & Diseth, 2016). What remains much less addressed is the timing of differential susceptibility as a function of developmental experiences, especially with respect to whether or not early-life conditions prove more or less consistent over time.

There are clearly data showing that temperamentally difficult infants appear more susceptible to environmental influences, for better and for worse (e.g., Bradley & Corwyn, 2008; Mesman et al., 2009; Rioux et al., 2016; Zhang, Sayler, et al., 2022), as do more physiologically reactive children (e.g., Davies et al., 2011; Essex et al., 2011; Obradović et al., 2010; Skibo et al., 2020), and those carrying certain genetic variants (e.g., Bakermans-Kranenburg & van Ijzendoorn, 2011; Cicchetti et al., 2012; Kochanska et al., 2011; Sun et al., 2018; Windhorst et al., 2015). The most compelling evidence is to be found in experimental–intervention research in which environments are systematically manipulated to test hypotheses as to which children will benefit the most in the experimental group and which fare most poorly in the control group. For example, Beach et al.’s (2010) work with African American families showed that teens carrying the dopamine receptor D4 7-repeat allele exhibited greater and lesser substance use in control and treatment groups, respectively, than children carrying the 4-repeat allele. Similarly, Brett et al. (2015) found that children with a pair of short alleles of the serotonin-transporter-linked promoter region polymorphism had the lowest and highest externalizing behavior in, respectively, high-quality foster care and continued and neglectful institutional care. Consider also Shaw et al.’s (2019) longitudinal study finding that children with greater polygenetic susceptibility responded more favorably to the Family Check-Up intervention than those with lower genetic susceptibility; thus, the former were more likely to follow a persistently low pattern of conduct problems from early childhood to adolescence when enrolled in the intervention program but a persistently high pattern when randomized to the control group. More robust evidence supporting differential susceptibility from randomized controlled experiments can be found in several meta-analytic studies (e.g., Bakermans-Kranenburg & van IJzendoorn, 2015; Belsky & van Ijzendoorn, 2015; van Ijzendoorn & Bakermans-Kranenburg, 2015).

Despite the just cited differential-susceptibility-related evidence involving both young children and adolescents, the prevailing view regarding timing of developmental plasticity and thus susceptibility to environmental influences is that it is most pronounced in the opening years of life. The case has been made, however, that adolescence represents a second period of heightened developmental plasticity (e.g., Fuhrmann et al., 2015; Larsen & Luna, 2018; Steinberg, 2014). In light of such evidence, questions can be raised as to whether children might differ as to when they manifest the greatest susceptibility to environmental influences. Recent work based on a huge Danish population registry focused on susceptibility to effects of family dysfunction during early childhood and adolescence on problematic development in emerging adulthood found that even though most children proved consistently more or less susceptible across these developmental epochs, there remained a meaningful subpopulation which proved more susceptible during adolescence than in early childhood (Belsky & Andersen, 2022).

In light of such results, Frankenhuis and Panchanathan’s (2011) hypothesis that early life conditions influence the timing of susceptibility becomes of particular interest. These evolutionary-minded developmentalists posited that when early-life conditions are consistent and strong—be they reflective of developmental support or adversity—children essentially “commit” to a slow or fast life-history strategy. That is because these early-life conditions provide a kind of “weather forecast” as to what their future life is likely to be like, leading them to develop accordingly. In contrast, when early-life conditions are highly variable, making environmental signals unclear, ambiguous, and mixed, children defer committing to a particular developmental strategy.

Such theorizing would seem to lead to the prediction that consistent early-life conditions—whether reflecting high or low adversity—should attenuate susceptibility to later environmental conditions, as a developmental trajectory has been established, whereas inconsistent early-life conditions should leave the developmental program open to influence later in life. To our knowledge, only one inquiry has attempted to test this proposition empirically, the aforementioned Danish research (Belsky & Andersen, 2022), finding it wanting. Such null results could be due to limitations of population-register data on early-life conditions and reliance on a single developmental outcome. Here for the second time, then, we test the evolutionary–commitment hypothesis using two reasonably sized samples while relying on more refined environmental measurements and multiple developmental outcomes.

Based on the evolutionary–developmental commitment hypothesis, we test the following predictions drawing on data from the Future of Families and Child Wellbeing Study (FFCWS) and the National Institute of Child Health and Human Development (NICHD) Study of Early Child Care and Youth Development (SECCYD). First, we hypothesize that children who experience high consistency of threat early in life—whether reflecting high or low levels—will prove less susceptible to the developmental effects of later-life conditions than other children. We focus on threat because it is an evolutionarily salient harsh life experience (Ellis et al., 2022; Sheridan & McLaughlin, 2014). Consistently, high or low experience of threat may indicate greater or less adversity in the developmental environment but, either way, provide a clear and strong signal of what future life may be like; this may lead to an earlier commitment to a developmental strategy and thus less susceptibility to later-life conditions. Second and in contrast, we predict that children whose early-life conditions are highly unpredictable reflected by income, another distinctive dimension of early-life adversity according to evolutionary developmental thinking (Ellis et al., 2022), will prove especially susceptible to later-life conditions. The first hypothesis is tested only using the FFCWS data because the NICHD SECCYD lacked sufficient information for measuring threat. When it comes to susceptibility, we focus on parenting during early and late childhood as predictors and three different developmental phenotypes: cognitive functioning and externalizing and internalizing problems.

Assessment of individual differences in susceptibility to parenting with respect to each developmental phenotype is estimated by means of standardized difference of the beta (DFBETAS), an influence statistic (Belsley et al., 1980; Cook & Weisberg, 1982) which has been used for this purpose in several recent investigations (Belsky et al., 2022; Sayler et al., 2022; Zhang et al., 2023; Zhang, Schlomer, et al., 2022). Individual differences in susceptibility to the effects of parenting on each of the three phenotypes are composited based on the well-established principle that composite measures are more reliable than individual ones, in this case reflecting susceptibility to environmental influence.

Method

Data

Data come from the FFCWS and the NICHD SECCYD. The FFCWS is a birth cohort study involving a stratified, multistage, probability sample of 4,898 children (boys = 2,556; White = 590, Black = 1,601, Latinx = 813, other = 261) born between 1998 and 2000 in 20 large U.S. cities, with an oversampling of births to unmarried parents, thereby insuring a diverse sample of lower income households and neighborhoods. (Details of the sampling methodology can be found at https://fragilefamilies.princeton.edu/documentation and in Reichman et al., 2001.) The present report involves data collected at birth and ages 1, 3, 5, 9, and 15 years.

The NICHD SECCYD recruited families through hospital visits to mothers shortly after the birth of a child in 1991 in 10 locations in the U.S. Following a conditional random sampling plan, a total of 1,364 families (boys = 705; White/non-Hispanic = 1,097, Black = 176, other = 91) with full-term, healthy newborns who completed a home interview when the infant was 1-month old became the study participants (NICHD Early Child Care Research Network, 2005). Upon enrollment, mothers had an average of 14.2 years of education, with 26% of the mothers having no more than a high school education; the average family income was 3.6 times the poverty threshold, with 21% of the household having incomes no greater than 200% of the poverty threshold. Data from the 1 month, 36 months, 54 months, 5th grade (~10-year-old), and 15 years of age assessments are used in the present report.

Measures

Measurements of both data sets are delineated. For replication purposes, we selected measures in the FFCWS and the NICHD SECCYD that were as closely aligned as possible. Available in the appendices are a list of the measurements (Appendix) and their details, including when they were measured, specific items used for predictor and outcome constructs, as well as indices of reliability (i.e., Cronbach α; see two columns in the Supplemental Material, one for each study, showing the items of each construct in each study).

Early-Life Conditions

Consistency of Threat.

Threat was measured using seven indicators in the FFCWS based on both primary caregiver-report and observations when children were 1 and 3 years of age (Miller et al., 2021; Sheridan & McLaughlin, 2014; see Appendix for the item list). These include the frequency of mother and father spanking their infant (birth to age 1), physical and emotional abuse (age 3; Straus, 1979), as well as exposure to domestic violence (ages 1 and 3) and community violence (age 3). These indicators were first standardized and then the standard deviations across these indicators were calculated for each child, reflecting individual differences in consistency of threat experiences. The resultant standard deviations were then reverse-coded so that higher standard deviations represented greater consistency in exposure to threat. The NICHD SECCYD data set lacked sufficient information to measure early-life experience of threat.

Unpredictability of Income.

Building on the work focusing on the effects of early-life harshness and unpredictability on later socioemotional and academic functioning by Li et al. (2018), we derived unpredictability scores based on repeatedly measured household income. For the FFCWS data, we relied on the mother- and father-reported income-to-poverty ratio, assessed at ages 1, 3, and 5 and averaged at each year (Mage1 = 2.06, SDage1 = 2.38; Mage3 = 2.21, SDage3 = 2.70; Mage5 = 2.22, SDage5 = 2.42). For the NICHD SECCYD, we based the income-to-needs ratio (calculated in the same way to the income-to-poverty ratio) on data collected when children were1 (M = 2.86, SD = 2.61), 6 (M = 3.66, SD = 3.10), 15 (M = 3.70, SD = 3.22), 24 (M = 3.72, SD = 3.04), 36 (M = 3.61, SD = 3.05), and 54 (M = 3.59, SD = 3.17) months of age. For both data sets, a ratio of 1.0 indicates that family income equals the federal poverty threshold adjusted for household size and higher scores indicate greater financial resources per person in the household.

Based on the repeatedly measured household income-to-needs ratio and following Li et al. (2018), unpredictability was operationalized as the degree of variation over time in the ratio after controlling for linear change. Toward this end, we first fitted growth curve models to estimate the linear change in the ratio over time. We then fitted a linear regression model for each child, with the repeated measures of income-to-needs ratio regressed on time. The generated residual standard deviation for each child was used as the index of early life experiences of unpredictability. Greater residual standard deviation reflects more variation around the predicted trend, indicating greater unpredictability over time (Hoffman, 2007).

Parenting Predictors

Early-Childhood Parenting.

Parenting was assessed using multiple items from the Home Observation for Measurement of the Environment (Bradley & Caldwell, 1984) in both studies. Early-childhood parenting was measured when children were age 3 (α = .76) and 5 (α = .78) for the FFCWS data set and at 36 (α = .67) and 54 (α = .66) months for the NICHD SECCYD data set. As the two studies used somewhat different versions of the Home Observation for Measurement of the Environment with different item sets, we sought to create a single-item set that was as similar across both studies as possible conceptually and proved internally consistent in both (see the Supplemental Material for a full list of items used in each data set). Each of these items is binary and scored as yes (1) or no (0). For each of the data sets, the scores of these items were averaged within and across age to create an early-childhood parenting composite score. Higher scores represent greater parental support.

Late-Childhood Parenting.

Similar to the early-childhood parenting measure, late-childhood parenting was measured using multiple items derived from several versions of the Home Observation for Measurement of the Environment instrument when children were 9 years old in the FFCWS (α = .64) and in Grade 5 (~age 10) in the NICHD SECCYD (α = .74; see the Supplemental Material for a full list of items used in each data set). Each of these items is binary and coded as yes (1) or no (0). Four items from the FFCWS data set (i.e., parent shouted at children, parent slapped or spanked child, parent scolded, derogated or criticized child, parent expressed overt annoyance with or hostility toward child) were reverse-coded (i.e., yes = 0 or no = 1). Again, item scores were averaged to create a late-childhood parenting composite index for each data set. Higher parenting scores reflect greater parental support during late-childhood.

Developmental Outcomes

Cognitive Functioning.

For both data sets, children’s cognitive functioning was measured using the Woodcock Johnson test of Applied Problems and Passage Comprehension subscales (Woodcock & Johnson, 1989). The test was administered at age 9 in the FFCWS (α = .77 and .75) and Grade 5 (~age 10) in the NICHD SECCYD (α = .81 and .87). The scores for the Applied Problems and the Passage Comprehension subscales were standardized and averaged to reflect children’s cognitive functioning in late-childhood. Higher scores reflect better cognitive functioning.

Externalizing Behavior.

Adolescents’ externalizing behavior was self-reported when the focal teens were age 15 in both studies—with somewhat different items (see the Supplemental Material for a full list of items used in each data set). For the FFCWS, externalizing behavior was measured by means of 13 items adopted from adolescents’ delinquent behavior measurement in the National Longitudinal Study of Adolescent to Adult Health (α = .74; Add Health; Harris et al., 2019); each item was rated using a 4-point Likert scale (1 = never;2 = 1 or 2 times; 3 = 3 or 4 times; 4 = 5 or more times). For the NICHD SECCYD, adolescents’ externalizing behavior was assessed by the Delinquent and Aggressive subscales of Youth Self Report (α = .80; Achenbach, 1991); each item was rated using a 3-point Likert scale (0 = not true,1 = somewhat or sometimes true,2 = very true or often true). For both studies, item scores were averaged to create a total score for externalizing behavior; higher scores reflected more externalizing problems.

Internalizing Behavior.

Adolescents’ anxiety and depression symptoms were self-reported when the focal children were age 15 in both studies (see the Supplemental Material for a full list of items used in each data set). For the FFCWS data set, adolescents’ internalizing behavior was assessed using the Brief Symptom Inventory (α = .76; Derogatis & Savitz, 2000) and the Center for Epidemiologic Studies Depression Scale (α = .80; Radloff, 1977). Both measures consist of items rated using a 4-point Likert scale ranging from 1 (strongly agree) to 4 (strongly disagree). All items except one (i.e., “I feel happy”) were reverse-coded so that higher averaged scores represent more internalizing behavior. For the NICHD SECCYD data set, adolescents’ internalizing behavior was measured using the Youth Self Report Anxious/Depressed subscale (Achenbach, 1991) and the short form of the Children’s Depression Inventory (Kovacs, 1992). As mentioned earlier, the Youth Self Report items were assessed using a 3-point Likert scale from 0 (not true) to 2 (very true or often true). The Children’s Depression Inventory was also based on a 3-point Likert scale in which 0 = normative behavior, 1 = a middle statement, and 2 = a depressive symptom after reflecting some of the items. The scores from different instruments were standardized and averaged to create an internalizing behavior composite score for each study. Higher scores reflect more internalizing problems.

Statistical Analysis

To increase the validity of the current work, the same analytic approach was applied to the FFCWS and the NICHD SECCYD data except when testing Hypothesis 1.

Preliminary Analyses

The preliminary analyses consisted of four steps. The first addressed missing data (the proportion of missingness ranges from 6.06% to 41.75%1 for the FFCWS and 0%–29.80% for the NICHD for the key variables). To allow missingness to be dependent on other variables in the data set given that the data may be missing at random, we conducted Multiple Imputation (Rubin, 1987) and created 30 imputed data sets, using multivariate imputation by chained equations (van Buuren & Groothuis-Oudshoorn, 2011). All reported results are thus based on the 30 imputed data sets for both data sets.

The second preliminary analytic step involved establishing bivariate associations between the two composite parenting predictors (i.e., early- and late-childhood parenting) and each of the three developmental outcomes (i.e., cognitive functioning, externalizing, and internalizing behavior). We purposefully chose not to implement demographic controls for two reasons. First, even a wide range of covariates does not guarantee the detection of causation in an observational study. Second, including covariates might lead to overcontrolling for meaningful variation in the parenting and early-life experience variables. Third, there are well-established cultural differences in parenting, which cannot be fully captured by demographics like race/ethnicity, so we followed guidance not to use these demographics to capture more complex social constructs that they reflect (Helms et al., 2005). Notably, while we write in terms of parenting “effects,” these are recognized as statistical associations only and cannot be presumed to be causal. Nevertheless, for purposes of exploring the issue of timing vis-à-vis susceptibility to potential parenting influences, we do employ language that, for heuristic purposes, highlights the potential influence of parenting early and later in development on child and adolescent functioning.

The third step evaluated the degree to which each child appeared susceptible to each of the six parenting effects (i.e., early- and late-childhood parenting each predicting each of the three outcomes) documented in the prior correlational analyses. Toward this end, we again relied on the DFBETAS (Belsley et al., 1980; e.g., Belsky et al., 2022; Zhang, Schlomer, et al., 2022) of correlation coefficients, a standardized influence statistic that relies on a leave-one-out procedure. This involves rerunning each of the bivariate predictor-outcome associations when excluding each case and repeating the process N times—so that the degree to which each child affects the full-sample association can be estimated. For example, a DFBETAS’ value of 0.1 (or −0.1) attributed to a specific child means that the corresponding correlation coefficient increased (or decreased) 0.1 standard error units when the child in question was included versus excluded from the sample and thus that the full-sample association was strengthened (or weakened) by the leave-one-out procedure.

To be more specific, whole-sample associations that are strengthened by including versus excluding a particular child suggest that the left-out child was more affected by the predictor than the sample as a whole, whereas associations that are reduced suggest the opposite that the child was less affected by the predictor. Thus, the more an association is strengthened (i.e., larger positive DFBETAS score) by including a particular child, the more susceptible the child appears to be to the effect under investigation, with the reverse being true the more it is weakened (i.e., more negative DFBETAS scores). Given the different implications of DFBETAS depending on the direction of an association, we reversed the sign of DFBETAS for theoretically anticipated and empirically documented negative associations (e.g., early-parenting support predicting adolescent externalizing) so that higher scores would reflect greater susceptibility.

The last preliminary analytic step examined whether children who appear more and less susceptible to early-childhood parenting, to varying degrees, also tend to be similarly susceptible to late-childhood parenting. This was accomplished in two major ways. The first involved separately averaging the three early-childhood and three late-childhood susceptibility scores (e.g., DFBETAS scores for early-parenting effects on cognition, externalizing, internalizing) to create two susceptibility composites (i.e., susceptibility to early- and to late-childhood parenting) and then correlating the two composites. To gain more insight into the association between susceptibility to early and later parenting, a second major way of evaluating consistency—or lack thereof—in susceptibility to parenting over time was implemented. It involved dividing the early and later susceptibility composite scores into terciles (i.e., low, moderate, high) and then cross-tabulating them (i.e., 3 × 3 cross-classification). Because this analysis was not independent of the prior one, no significance testing was conducted.

Primary Analyses

Primary analyses were conducted to test the two aforementioned hypotheses. The first addressed whether especially high consistency of early-life threat attenuates children’s susceptibility to later parenting (H1). Based on the evolutionary commitment hypothesis, we expected that such consistent signals would result in lower susceptibility to later parenting than would less consistent signals. Toward this end, we employed two approaches: first, we correlated the consistency-of-threat index with susceptibility to late-childhood parenting. Second, to gain further insight into the correlation findings, we categorized threat consistency into terciles and compared those in the top tercile of threat consistency with all others (i.e., low + moderate terciles) in terms of their susceptibility to later parenting using Welch’s t test, which is more robust for unequal sample sizes between groups than Student’s t test.

Relatedly, to evaluate whether unpredictability promotes children’s susceptibility to later parenting (H2), we adopted the same analytic strategy. Thus, after correlating unpredictability with susceptibility to later parenting, we binned unpredictability and compared those in the top quartile of unpredictability and others in terms of their susceptibility to later parenting using Welch’s t test.

Transparency and Openness

The study’s design and analysis were not preregistered. All analyses were conducted using R Version 4.2.1 (R Core Team, 2022). Supplemental Material includes the list of items used for each measure. The data and analytic syntax needed to replicate the analyses are available upon requests by emailing the corresponding author.

Results

After establishing bivariate associations between early- and late-childhood parenting with children’s cognitive functioning and externalizing and internalizing problems, the derived susceptibility scores based on the DFBETAS of the six sets of correlation coefficients were composited, as part of the preliminary analyses. These were then used to test, as the primary analysis, the two hypotheses guiding this report focusing on the timing of parenting effects: (a) consistent early-life experience of threat will decrease children’s susceptibility to later parenting, whereas (b) early-life experience of unpredictability will promote children’s susceptibility to later parenting.

Preliminary Analyses

Effects of Early- and Late-Childhood Parenting

In both data sets, standardized simple regression slope coefficients (or correlation coefficients) revealed, as expected, that greater parenting support during early- and late-childhood was associated with enhanced cognitive functioning; see Table 1 for associations for the FFCWS and Table 2 for the NICHD SECCYD. It was also the case, again as expected, that in both data sets greater early-parenting support predicted less externalizing behavior. Somewhat surprising, however, was that such anticipated negative associations did not emerge in the case of (a) late parenting predicting externalizing problems in the FFCWS, only in the NICHD SECCYD; and (b) that neither early- nor late-parenting predicted internalizing problems in the latter study, whereas both did in the FFCWS. Effects sizes of all detected significant effects were small to moderate in magnitude.

Table 1.

Associations of Early- and Late-Childhood Parenting and Developmental Outcomes for the Future of Families and Child Wellbeing Study

Predictor Cognitive functioning Externalizing behavior Internalizing problem
Early parenting .21*** −.07*** −.07***
Late parenting .15*** −.03 −.06**
Open in a new tab

Note. Early parenting means early-childhood parenting, whereas late parenting means late-childhood parenting.

**

p < .01.

***

p < .001.

Table 2.

Associations of Early- and Late-Childhood Parenting and Developmental Outcomes for the Study of Early Child Care and Youth Development

Predictor Cognitive functioning Externalizing behavior Internalizing problem
Early parenting .33*** −.06* .05
Late parenting .19*** −.07* −.04
Open in a new tab

Note. Early parenting means early-childhood parenting, whereas late parenting means late-childhood parenting.

*

p < .05.

***

p < .001.

Consistency of Susceptibility to Early- and Late-Childhood Parenting

We created two susceptibility composites for early- and late-childhood susceptibility to parenting, respectively, by averaging DFBETAS-based susceptibility scores with regard to children’s cognitive functioning and externalizing and internalizing problem behavior. For both studies, the intercorrelation of the two susceptibility-to-parenting composites yielded some evidence of cross-time consistency in susceptibility to youth outcomes: children more and less susceptible to early-childhood parenting proved (some-what) more and less susceptible to late-childhood parenting, respectively, as indicated by positive correlation coefficients with small-to-medium effect sizes, for FFCWS: r(4,896) = 0.18, p < .001; for NICHD SECCYD: r(1,362) = 0.12, p < .001.

To provide further descriptive insight into the association of susceptibility to parenting during early- and late-childhood, we cross-tabulated binned DFBETAS-based susceptibility scores, as shown in Tables 3 and 4. No significance testing was undertaken because of the nonindependence of the results of this cross-tabulation analysis and the prior correlational one.

Table 3.

Cross-Tabulation of Number of Children Classified as Low, Moderate, and High Susceptibility to Early- and Late-Childhood Parenting for the Future of Families and Child Wellbeing Study

Early susceptibility Later susceptibility
Low Moderate High
Low 669 515 449
Moderate 473 822 338
High 491 296 845
Open in a new tab
Table 4.

Cross-Tabulation of Number of Children Classified as Low, Moderate, and High Susceptibility to Early- and Late-Childhood Parenting for the National Institute of Child Health and Human Development Study of Early Child Care and Youth Development

Early susceptibility Later susceptibility
Low Moderate High
Low 206 141 108
Moderate 128 248 79
High 121 66 267
Open in a new tab

Inspection of tables indicates that even though about half of the participants (47.69% for the FFCWS; 52.86% for the NICHD SECCYD) were consistent in susceptibility over time, falling in the low, middle, or high tercile at both measurement occasions, almost one fifth of children (19.19%) in the FFCWS and one sixth (16.79%) in the NICHD SECCYD proved highly susceptible at one time and highly unsusceptible at another.

Primary Analyses

Hypothesis 1: Consistent Early-Life Experiences of Threat Will Attenuate Later Susceptibility

Hypothesis 1 was only tested using the FFCWS data set. One-tailed correlation analysis indicated that consistency of early experiences of threat was not associated with later susceptibility, r = −0.002, p = .438. In addition, using threat terciles, one-tailed Welch’s t test also showed that children with more consistent experience of threat (N = 1,633) did not have lower later susceptibility (M = −0.025, SD = 0.824) than others (N = 3,265, M = 0.012, SD = 1.08), t(4,110.7) = 1.339, p = .090. In sum, there is no evidence consistent with H1 based on the correlation analysis and t test.

Due to the presumption that the absence of evidence is not evidence of absence, we conducted equivalence testing using the two one-sided tests procedure (Goertzen & Cribbie, 2010; Lakens et al., 2018; Rogers et al., 1993), with equivalence bounds set to the correlation the data had 80% power to detect. This enabled us to directly test for the absence of effects. With 4,989 participants, the data had 80% power to detect a correlation of .04. Equivalence testing results revealed a statistical equivalence, r(4,987) = −0.002, p = .004. The 90% confidence interval for the correlation ([−0.026, 0.022]) fell entirely within the equivalence bound (±0.04). This means we can reject the null hypothesis of the equivalence testing that the true correlation is large enough to lie outside these equivalence bounds. These results provide evidence that the correlation is small enough to be considered “equivalent to zero.”

Hypothesis 2: Early-Life Experiences of Unpredictability Will Increase Later Susceptibility

For both data sets, one-tailed correlation analysis and t test were used to evaluate whether early-life experiences of unpredictability of income increased later susceptibility. For the FFCWS, correlation analysis indicated that greater unpredictability significantly predicted greater later susceptibility, r = .032, p = .013. Relying on unpredictability terciles, single-tailed Welch’s t test showed that children who had greater experience of unpredictability (N = 1,632, M = 0.046, SD = 0.972) proved more susceptible to later parenting than others (N = 3,266, M = −0.023, SD = 1.010), t(3,382.1) = −2.318, p = .010. In the case of both analyses, the predictions, while statistically reliable, were small in magnitude.

Turning to the predicted associations of unpredictability of income with later susceptibility, associations proved insignificant for the NICHD SECCYD, r = 0.006, p = .416. Using the unpredictability terciles, one-tailed Welch’s t test indicated exactly the same (N = 454, M = 0.019, SD = 1.06 vs. N = 910, M = −0.009, SD = 0.968), t(834.26) = −0.478, p = .316. These null results again lead us to perform another equivalence testing to confirm the absence of evidence. With 1,364 participants, the data had 80% power to detect a correlation of .076. Therefore, the equivalence bound was set to ±0.076. Equivalence testing results revealed a statistical equivalence, r(1,362) = 0.006, p = .003. The 90% confidence interval for the correlation ([−0.047, 0.043]) fell entirely within the equivalence bound (within ±0.076), suggesting that the correlation is small enough to be considered “equivalent to zero” in this case.

Discussion

Recent advances in the evaluation of differential susceptibility to environmental influences have raised questions about whether it should be treated as a trait—with some children being more or less developmentally plastic than others across a variety of experiences and/or outcomes—or whether development is more complex than that (Belsky et al., 2022; Sayler et al., 2022; Zhang et al., 2023; Zhang, Schlomer, et al., 2022). Despite the widespread developmental view that children should be more susceptible to early than later developmental experiences, this questioning raises the possibility that some children may be more susceptible to later than early developmental experiences and exposures. Intriguingly, were this so, it could result from the thinking underlying what we have referred to as the “developmental commitment hypothesis.”

Frankenhuis and Panchanathan (2011) theorized that children adopt certain life-history strategies based on their early-life conditions, which function as a “weather forecast” to inform children about their future life conditions, thereby shaping their development trajectory. Thus, if early-life conditions are consistent and strong, be they enriched or adverse, children may “commit” to a particular developmental strategy earlier than others; in contrast, if early-life conditions are ambiguous, children may need prolonged time to discern the reliability of environmental cues before making such a commitment. Thus, drawing on data from two large studies, we tested the predictions that (a) early life experiences of consistency of threat will reduce children’s susceptibility to later parenting, whereas (b) early life experiences of unpredictability of family income to needs will promote children’s susceptibility to later parenting.

Only very limited support for these predictions emerged, in fact only two of six one-tailed tests—given the directional predictions—proved statistically reliable. And even then, the effect sizes were strikingly weak. In particular, we only found some support for the expected role of unpredictability of income—in seeming to promote greater susceptibility to later parenting—within the FFCWS. As this is only the second investigation to address empirically the “developmental commitment hypothesis” of Frankenhuis and Panchanathan (2011), it seems risky to embrace the null, despite the large sample sizes, the one-tailed statistical tests relied upon in this work, and follow-up testing confirming the null. Nevertheless, certainly more positive evidence is needed before the predictions of the developmental commitment hypothesis can be regarded as informative rather than just intriguing.

The null findings could arise for several reasons. First, all items except one (community violence) used to measure the consistency of threat focused on proximal processes within family at ages 1 and 3. Results might be different if we include threat experiences in other settings (e.g., childcare) and objective measures of macrolevel threat (e.g., census tract crime rates). Second, the domain of consistency in early-life conditions may matter in shaping how children calibrate their development. In line with Zhang et al.’s (2023) findings that susceptibility is domain-specific, at least for some individuals, it would seem worth investigating to what extent the developmental commitment hypothesis is itself domain-specific. If the domain in which children receive the signals that are used to calibrate their developmental trajectories (e.g., parenting warmth or harshness) is the one that remains consistent over time, might they “commit” to that domain’s conditions earlier than other domains? If that is the case, unpredictability in parenting might more specifically predict greater susceptibility to later parenting practices, whereas unpredictability in household income might be more predictive of susceptibility to future resource fluctuations.

It is also possible that environmental cues in one domain (e.g., unstable income) can spill over into other areas (e.g., parenting) due to stress contagion within families (Conger et al., 2010). Thus, income unpredictability might also influence how children respond to parenting, not only resource fluctuations. Integrating domain-specific measures of consistency/unpredictability in future research would provide clearer insight into the domain-specificity of the developmental commitment hypothesis and the timing of susceptibility.

It should be evident that there are many limitations to the work presented despite the obvious strengths, including the two large studies on which this report is based and the repeated observational measurement of parenting. Most notably, we only addressed the predictive power of parenting and many other exposures and experiences merit consideration given evidence that they can influence children’s development (e.g., domestic violence, neighborhood danger, marital conflict). The same, of course, applies to developmental phenotypes as well as early-life conditions that were not the focus of this inquiry (e.g., for phenotypes: executive function, prosocial behavior, academic achievement; for early-life conditions: deprivation, positive experiences). It is possible that results could be specific to the indices of threat and unpredictability considered herein. Future work could measure unpredictability using changes in family structure, residential moves, and parental employment changes, and/or measure threat using more macrolevel indicators (e.g., neighborhood crime rate).

In addition, we relied on DFBETAS to measure individual differences in developmental susceptibility, and we would be remiss not to make clear that like almost any other measurement, DFBETAS are subject to random errors. Also, we only assessed developmental susceptibility twice, in early and later childhood; having more measurement occasions to test the predictability of earlier and later experience might yield rather different results. For these reasons, it seems wisest to conclude that until further tests of the developmental–commitment hypothesis which yield more evidence consistent with it, it should be treated as just that, a theoretical prediction, not a documented fact.

Summary and Conclusion

We tested herein the evolutionary commitment hypothesis that more consistent early-life family experiences will attenuate susceptibility to later-life experiences, with the opposite being true of inconsistent early-life conditions, drawing on data from two different large-scale studies, the FFCWS and the NICHD Study. Even as some evidence emerged suggesting that children exposed to greater early-life unpredictability showed heightened susceptibility to parenting effects in late childhood, consistent with the evolutionary commitment hypothesis, the detected effect size was small— and lacked consistent support across data sets and analyses. Then, there was the fact that no statistically significant associations emerged between the consistency of threat experiences and susceptibility to effects of parenting. Given that this is only the second empirical test of the hypothesis under consideration, it is premature to dismiss it. Future research should incorporate more diverse environmental influences and developmental outcomes to assess susceptibility over multiple developmental stages in order that stronger conclusions regarding the hypothesis can be drawn.

Supplementary Material

supplement

Supplemental materials: https://doi.org/10.1037/dev0002023.supp

Public Significance Statement.

This study examines whether more and less consistency of exposure to threat and unpredictability in early life are associated with individual differences in developmental plasticity later in childhood. Such investigation will extend the understanding of when and how children remain more or less susceptible to environmental influences including interventions as they mature. Therefore, it holds the promise of guiding more targeted, developmentally sensitive strategies for promoting well-being among children with different early-life experiences.

Appendix. Measurement Selected for Each Data Set

Variable FFCWS SECCYD
Predictors
 Early-childhood parenting HOME at ages 3 and 5 (averaged) HOME at month 36 and 54 (averaged)
 Late-childhood parenting HOME at age 9 HOME at Grade 5
Outcomes
 Cognitive functioning WJ applied problems and passage comprehensive (averaged) at age 9 WJ applied problems and passage comprehensive (averaged) at Grade 5
 Externalizing behavior Teen self-reported delinquency at age 15 Youth Self Report (YSR) Delinquency and Aggression subscales at age 15
 Internalizing behavior Brief Symptom Inventory 18 and Center for Epidemiologic Studies Depression Scale at age 15 YSR Anxious/Depressed subscale and Children’s Depression Inventory at age 15
Correlates
 Threat Domestic violence exposure at ages 1 and 3
Mother spanking at age 1, Father spanking at age 1
Physical abuse at age 3, Verbal abuse at age 3
Community violence at age 3		 Not available
 Unpredictability Household income-to-poverty ratio at ages 1, 3, and 5 Household income-to-needs ratio at 1, 15, 24, 36, and 54 months
Open in a new tab

Note. FFCWS = Future of Families and Child Wellbeing Study; WJ = Woodcock-Johnson; SECCYD = Study of Early Child Care and Youth Development; HOME = Home Observation for Measurement of the Environment.

Footnotes

1

The percentage of missing data for the early-childhood parenting variable is 41.75%. Missingness was calculated at the variable level for all key constructs, including early- and late-childhood parenting, the three outcome variables, and the measures of early-life threat and income-based unpredictability.

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

References

  1. Achenbach TM (1991). Integrative guide for the 1991 CBCL/4–18, YSR, and TRF profiles. Department of Psychiatry, University of Vermont. https://www.scienceopen.com/document?vid=03b6041c-03bb-4e15-acf4-23723344d502 [Google Scholar]
  2. Aron EN, & Aron A (1997). Sensory-processing sensitivity and its relation to introversion and emotionality. Journal of Personality and Social Psychology, 73(2), 345–368. 10.1037/0022-3514.73.2.345 [DOI] [PubMed] [Google Scholar]
  3. Bakermans-Kranenburg MJ, & van Ijzendoorn MH (2011). Differential susceptibility to rearing environment depending on dopamine-related genes: New evidence and a meta-analysis. Development and Psychopathology, 23(1), 39–52. 10.1017/S0954579410000635 [DOI] [PubMed] [Google Scholar]
  4. Bakermans-Kranenburg MJ, & van IJzendoorn MH (2015). The hidden efficacy of interventions: Gene×environment experiments from a differential susceptibility perspective. Annual Review of Psychology, 66(1), 381–409. 10.1146/annurev-psych-010814-015407 [DOI] [PubMed] [Google Scholar]
  5. Beach SRH, Brody GH, Lei M-K, & Philibert RA (2010). Differential susceptibility to parenting among African American youths: Testing the DRD4 hypothesis. Journal of Family Psychology, 24(5), 513–521. 10.1037/a0020835 [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Belsky J (in press). The nature of nurture. Harvard University Press. [Google Scholar]
  7. Belsky J (1997a). Theory testing, effect-size evaluation, and differential susceptibility to rearing influence: The case of mothering and attachment. Child Development, 68(4), 598–600. 10.2307/1132110 [DOI] [PubMed] [Google Scholar]
  8. Belsky J (1997b). Variation in susceptibility to rearing influences: An evolutionary argument. Psychological Inquiry, 8(3), 182–186. 10.1207/s15327965pli0803_3 [DOI] [Google Scholar]
  9. Belsky J, & Andersen SH (2022). Individual differences in timing of susceptibility to adverse effects of family dysfunction. Journal of Family Psychology, 36(7), 1229–1239. 10.1037/fam0000970 [DOI] [PubMed] [Google Scholar]
  10. Belsky J, Bakermans-Kranenburg MJ, & van IJzendoorn MH (2007). For better and for worse: Differential susceptibility to environmental influences. Current Directions in Psychological Science, 16(6), 300–304. 10.1111/j.1467-8721.2007.00525.x [DOI] [Google Scholar]
  11. Belsky J, & Pluess M (2009). Beyond diathesis stress: Differential susceptibility to environmental influences. Psychological Bulletin, 135(6), 885–908. 10.1037/a0017376 [DOI] [PubMed] [Google Scholar]
  12. Belsky J, & Pluess M (2013). Beyond risk, resilience, and dysregulation: Phenotypic plasticity and human development. Development and Psychopathology, 25(4), 1243–1261. 10.1017/S095457941300059X [DOI] [PubMed] [Google Scholar]
  13. Belsky J, & van Ijzendoorn MH (2015). What works for whom? Genetic moderation of intervention efficacy. Development and Psychopathology, 27(1), 1–6. 10.1017/S0954579414001254 [DOI] [PubMed] [Google Scholar]
  14. Belsky J, Zhang X, & Sayler K (2022). Differential susceptibility 2.0: Are the same children affected by different experiences and exposures? Development and Psychopathology, 34(3), 1025–1033. 10.1017/S0954579420002205 [DOI] [PubMed] [Google Scholar]
  15. Belsley D, Kuh E, & Welsch R (1980). Regression diagnostics: Identifying influential data and sources of collinearity. Wiley. 10.1002/0471725153 [DOI] [Google Scholar]
  16. Boyce WT, & Ellis BJ (2005). Biological sensitivity to context: I. An evolutionary-developmental theory of the origins and functions of stress reactivity. Development and Psychopathology, 17(2), 271–301. 10.1017/S0954579405050145 [DOI] [PubMed] [Google Scholar]
  17. Bradley RH, & Caldwell BM (1984). The HOME Inventory and family demographics. Developmental Psychology, 20(2), 315–320. 10.1037/0012-1649.20.2.315 [DOI] [Google Scholar]
  18. Bradley RH, & Corwyn RF (2008). Infant temperament, parenting, and externalizing behavior in first grade: A test of the differential susceptibility hypothesis. Journal of Child Psychology and Psychiatry, 49(2), 124–131. 10.1111/j.1469-7610.2007.01829.x [DOI] [PubMed] [Google Scholar]
  19. Brett ZH, Humphreys KL, Smyke AT, Gleason MM, Nelson CA, Zeanah CH, Fox NA, & Drury SS (2015). Serotonin transporter linked polymorphic region (5-HTTLPR) genotype moderates the longitudinal impact of early caregiving on externalizing behavior. Development and Psychopathology, 27(1), 7–18. 10.1017/S0954579414001266 [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Cicchetti D, Rogosch FA, & Thibodeau EL (2012). The effects of child maltreatment on early signs of antisocial behavior: Genetic moderation by tryptophan hydroxylase, serotonin transporter, and monoamine oxidase A genes. Development and Psychopathology, 24(3), 907–928. 10.1017/S0954579412000442 [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Conger RD, Conger KJ, & Martin MJ (2010). Socioeconomic status, family processes, and individual development. Journal of Marriage and Family, 72(3), 685–704. 10.1111/j.1741-3737.2010.00725.x [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Cook RD, & Weisberg S (1982). Residuals and influence in regression. Chapman and Hall. [Google Scholar]
  23. Davies PT, Sturge-Apple ML, & Cicchetti D (2011). Interparental aggression and children’s adrenocortical reactivity: Testing an evolutionary model of allostatic load. Development and Psychopathology, 23(3), 801–814. 10.1017/S0954579411000319 [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Derogatis LR & Savitz KL (2000). The SCL–90–R and Brief Symptom Inventory (BSI) in primary care. In M. Maruish E (Ed.), Handbook of psychological assessment in primary care settings (pp. 297–334). Lawrence Erlbaum Associates Publishers. [Google Scholar]
  25. Ellis BJ, Sheridan MA, Belsky J, & McLaughlin KA (2022). Why and how does early adversity influence development? Toward an integrated model of dimensions of environmental experience. Development and Psychopathology, 34(2), 447–471. 10.1017/S0954579421001838 [DOI] [PubMed] [Google Scholar]
  26. Ellis BJ, Shirtcliff EA, Boyce WT, Deardorff J, & Essex MJ (2011). Quality of early family relationships and the timing and tempo of puberty: Effects depend on biological sensitivity to context. Development and Psychopathology, 23(1), 85–99. 10.1017/S0954579410000660 [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Essex MJ, Armstrong JM, Burk LR, Goldsmith HH, & Boyce WT (2011). Biological sensitivity to context moderates the effects of the early teacher-child relationship on the development of mental health by adolescence. Development and Psychopathology, 23(1), 149–161. 10.1017/S0954579410000702 [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Frankenhuis WE, & Panchanathan K (2011). Individual differences in developmental plasticity may result from stochastic sampling. Perspectives on Psychological Science, 6(4), 336–347. 10.1177/1745691611412602 [DOI] [PubMed] [Google Scholar]
  29. Fuhrmann D, Knoll LJ, & Blakemore SJ (2015). Adolescence as a sensitive period of brain development. Trends in Cognitive Sciences, 19(10), 558–566. 10.1016/j.tics.2015.07.008 [DOI] [PubMed] [Google Scholar]
  30. Goertzen JR, & Cribbie RA (2010). Detecting a lack of association: An equivalence testing approach. British Journal of Mathematical and Statistical Psychology, 63(3), 527–537. 10.1348/000711009X475853 [DOI] [PubMed] [Google Scholar]
  31. Harris KM, Halpern CT, Whitsel EA, Hussey JM, Killeya-Jones LA, Tabor J, & Dean SC (2019). Cohort profile: The national longitudinal study of adolescent to adult health (add health). International Journal of Epidemiology, 48(5), 1415–1415k. 10.1093/ije/dyz115 [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Helms JE, Jernigan M, & Mascher J (2005). The meaning of race in psychology and how to change it: A methodological perspective. American Psychologist, 60(1), 27–36. 10.1037/0003-066X.60.1.27 [DOI] [PubMed] [Google Scholar]
  33. Hoffman L (2007). Multilevel models for examining individual differences in within-person variation and covariation over time. Multivariate Behavioral Research, 42(4), 609–629. 10.1080/00273170701710072 [DOI] [Google Scholar]
  34. Kochanska G, Kim S, Barry RA, & Philibert RA (2011). Children’s genotypes interact with maternal responsive care in predicting children’s competence: Diathesis-stress or differential susceptibility? Development and Psychopathology, 23(2), 605–616. 10.1017/S0954579411000071 [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Kovacs M (1992). Children’s Depression Inventory. Multi-Health Systems. [Google Scholar]
  36. Lakens D, Scheel AM, & Isager PM. (2018). Equivalence testing for psychological research: A tutorial. Advances in Methods and Practices in Psychological Science, 1(2), 259–269. 10.1177/2515245918770963 [DOI] [Google Scholar]
  37. Larsen B, & Luna B (2018). Adolescence as a neurobiological critical period for the development of higher-order cognition. Neuroscience and Biobehavioral Reviews, 94, 179–195. 10.1016/j.neubiorev.2018.09.005 [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Li Z, Liu S, Hartman S, & Belsky J (2018). Interactive effects of early-life income harshness and unpredictability on children’s socioemotional and academic functioning in kindergarten and adolescence. Developmental Psychology, 54(11), 2101–2112. 10.1037/dev0000601 [DOI] [PubMed] [Google Scholar]
  39. Lionetti F, Aron A, Aron EN, Burns GL, Jagiellowicz J, & Pluess M (2018). Dandelions, tulips and orchids: Evidence for the existence of low-sensitive, medium-sensitive and high-sensitive individuals. Translational Psychiatry, 8(1), Article 24. 10.1038/s41398-017-0090-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Listou Grimen H, & Diseth Å (2016). Sensory processing sensitivity: Factors of the highly sensitive person scale and their relationships to personality and subjective health complaints. Perceptual and Motor Skills, 123(3), 637–653. 10.1177/0031512516666114 [DOI] [PubMed] [Google Scholar]
  41. Mesman J, Stoel R, Bakermans-Kranenburg MJ, van IJzendoorn MH, Juffer F, Koot HM, & Alink LR (2009). Predicting growth curves of early childhood externalizing problems: Differential susceptibility of children with difficult temperament. Journal of Abnormal Child Psychology, 37(5), 625–636. 10.1007/s10802-009-9298-0 [DOI] [PubMed] [Google Scholar]
  42. Miller AB, Machlin L, McLaughlin KA, & Sheridan MA (2021). Deprivation and psychopathology in the Fragile Families Study: A 15-year longitudinal investigation. Journal of Child Psychology and Psychiatry, 62(4), 382–391. 10.1111/jcpp.13260 [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. NICHD Early Child Care Research Network. (Ed.). (2005). Child care and child development: Results from the NICHD study of early child care and youth development. Guilford Press. [Google Scholar]
  44. Obradović J, Bush NR, Stamperdahl J, Adler NE, & Boyce WT (2010). Biological sensitivity to context: The interactive effects of stress reactivity and family adversity on socioemotional behavior and school readiness. Child Development, 81(1), 270–289. 10.1111/j.1467-8624.2009.01394.x [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. R Core Team. (2022). R: A language and environment for statisticalcomputing. R Foundation for Statistical Computing. https://www.R-project.org/ [Google Scholar]
  46. Radloff LS (1977). The CES-D scale: A self-report depression scale for research in the general population. Applied Psychological Measurement, 1(3), 385–401. 10.1177/014662167700100306 [DOI] [Google Scholar]
  47. Reichman NE, Teitler JO, Garfinkel I, & McLanahan SS (2001). Fragile families: Sample and design. Children and Youth Services Review, 23(4–5), 303–326. 10.1016/S0190-7409(01)00141-4 [DOI] [Google Scholar]
  48. Rioux C, Castellanos-Ryan N, Parent S, Vitaro F, Tremblay RE, & Séguin JR (2016). Differential susceptibility to environmental influences: Interactions between child temperament and parenting in adolescent alcohol use. Development and Psychopathology, 28(1), 265–275. 10.1017/S0954579415000437 [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Rogers JL, Howard KI, & Vessey JT (1993). Using significance tests to evaluate equivalence between two experimental groups. Psychological Bulletin, 113(3), 553–565. 10.1037/0033-2909.113.3.553 [DOI] [PubMed] [Google Scholar]
  50. Rubin DB. (1987). Multiple imputation for survey nonresponse. Wiley. 10.1002/9780470316696 [DOI] [Google Scholar]
  51. Sayler K, Zhang X, Steinberg L, & Belsky J (2022). Parenting, peers and psychosocial adjustment: Are the same—Or different—Children affected by each? Journal of Youth and Adolescence, 51(3), 443–457. 10.1007/s10964-022-01574-9 [DOI] [PubMed] [Google Scholar]
  52. Shakiba N, Ellis BJ, Bush NR, & Boyce WT (2020). Biological sensitivity to context: A test of the hypothesized U-shaped relation between early adversity and stress responsivity. Development and Psychopathology, 32(2), 641–660. 10.1017/S0954579419000518 [DOI] [PubMed] [Google Scholar]
  53. Shaw DS, Galán CA, Lemery-Chalfant K, Dishion TJ, Elam KK, Wilson MN, & Gardner F (2019). Trajectories and predictors of children’s early-starting conduct problems: Child, family, genetic, and intervention effects. Development and Psychopathology, 31(5), 1911–1921. 10.1017/S0954579419000828 [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Sheridan MA, & McLaughlin KA (2014). Dimensions of early experience and neural development: Deprivation and threat. Trends in Cognitive Sciences, 18(11), 580–585. 10.1016/j.tics.2014.09.001 [DOI] [PMC free article] [PubMed] [Google Scholar]
  55. Skibo MA, Sturge-Apple ML, & Suor JH (2020). Early experiences of insensitive caregiving and children’s self-regulation: Vagal tone as a differential susceptibility factor. Development and Psychopathology, 32(4), 1460–1472. 10.1017/S0954579419001408 [DOI] [PubMed] [Google Scholar]
  56. Steinberg LD (2014). Age of opportunity: Lessons from the new science of adolescence. Houghton Mifflin Harcourt. [Google Scholar]
  57. Straus MA (1979). Measuring intrafamily conflict and violence: The conflict tactics (CT) scales. Journal of Marriage and Family, 41(1), 75–88. 10.2307/351733 [DOI] [Google Scholar]
  58. Sun Y, Fang J, Wan Y, Hu J, Xu Y, & Tao F (2018). Polygenic differential susceptibility to cumulative stress exposure and childhood obesity. International Journal of Obesity, 42(6), 1177–1184. 10.1038/s41366-018-0116-z [DOI] [PubMed] [Google Scholar]
  59. van Buuren S, & Groothuis-Oudshoorn K (2011). mice: Multivariate imputation by chained equations in R. Journal of Statistical Software, 45(3), 1–67. 10.18637/jss.v045.i03 [DOI] [Google Scholar]
  60. van Ijzendoorn MH, & Bakermans-Kranenburg MJ (2015). Genetic differential susceptibility on trial: Meta-analytic support from randomized controlled experiments. Development and Psychopathology, 27(1), 151–162. 10.1017/S0954579414001369 [DOI] [PubMed] [Google Scholar]
  61. Windhorst DA, Mileva-Seitz VR, Linting M, Hofman A, Jaddoe VW, Verhulst FC, Tiemeier H, van IJzendoorn MH, & Bakermans-Kranenburg MJ (2015). Differential susceptibility in a developmental perspective: DRD4 and maternal sensitivity predicting externalizing behavior. Developmental Psychobiology, 57(1), 35–49. 10.1002/dev.21257 [DOI] [PubMed] [Google Scholar]
  62. Woodcock RW, & Johnson MB (1989). Woodcock-Johnson Psychoeducational Battery-Revised. Riverside. [Google Scholar]
  63. Zhang X, Sayler K, Hartman S, & Belsky J (2022). Infant temperament, early-childhood parenting, and early-adolescent development: Testing alternative models of Parenting × Temperament interaction. Development and Psychopathology, 34(3), 784–795. 10.1017/S0954579420002096 [DOI] [PubMed] [Google Scholar]
  64. Zhang X, Schlomer GL, Ellis BJ, & Belsky J (2022). Environmental harshness and unpredictability: Do they affect the same parents and children? Development and Psychopathology, 34(2), 667–673. 10.1017/S095457942100095X [DOI] [PubMed] [Google Scholar]
  65. Zhang X, Widaman K, & Belsky J (2023). Beyond orchids and dandelions: Susceptibility to environmental influences is not bimodal. Development and Psychopathology, 35(1), 191–203. 10.1017/S0954579421000821 [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

supplement
NIHMS2138358-supplement-supplement.docx (26.6KB, docx)

RESOURCES