SYNOPSIS
Objective.
To test three competing models of the nature and structure of maternal parenting practices with infants in U.S. national and multiple international samples. The three models were a one-factor dimensional model, a multi-factor style model, and a hybrid two-factor/six-domain model. Undertaking this evaluation of parenting with national and international samples permits a wide yet judicious analysis of culture-common versus culture-specific models of maternal parenting practices with young infants.
Method.
Basic caregiving practices of primiparous mothers with their 5-month-old infants during naturalistic interactions at home in nine different cultures were videorecorded, microcoded, and analyzed. Individual practices were organized into nurture, physical, social, didactic, material, and language domains.
Results.
In Study 1 using a U.S. national sample (N = 360), analyses of the structure of mothers’ parenting practices yielded a best-fitting two-factor/six-domain structure. In Study 2, using a 9-nation sample (N = 653), the two-factor/six-domain structure was largely replicated and partial metric invariance achieved.
Conclusions.
Mothers’ parenting in the middle of the first year of their infant’s life is commonly structured and adapted to the universal needs and developmental tasks of infants’ surviving and thriving.
Keywords: culture, parenting, mother-infant, structure of parenting, nature of parenting
INTRODUCTION
The Responsibilities of Parenting Human Infants at the Beginning of Life
UNICEF estimates that approximately 370,000 babies are born worldwide each day, leading hundreds of thousands of new mothers around the globe to experience the joys, reap the rewards, and assume the daunting responsibilities of becoming a new parent (Alhattab, 2021). New parents are fundamentally invested in infants -- their survival and subsistence, their socialization and education -- and from a relational developmental systems perspective parents and infants influence one another bidirectionally and transactionally (Lerner, 2018). However, a compelling tension pervades our understanding of this foundational human relationship. On the one hand, individual parent-infant dyads doubtlessly have idiosyncratic needs and goals, and parent and infant needs and goals are also shaped by the unique and specific cultures to which the two belong. On the other hand, human infants everywhere display many of the same biological needs and must succeed at many of the same early developmental tasks. Concomitantly, all parents share some similar responsibilities to guide their offspring to survive and thrive in their physical and social environments. Indeed, at the end of the day parents everywhere want physical health, mental achievement, social adjustment, and economic security for their children -- however those goals may be locally instantiated. At some level of analysis, then, parenting infants is likely to be similar regardless of the idiosyncratic dyad or the specific culture.1
Is parenting structured in any way to fulfill the responsibilities of infant childrearing and meet the developmental tasks of infants? Is any parenting structure common across cultures? The complementary pair of studies presented here was designed to address these two questions. We report on the videorecording, microcoding, analysis, and organization of basic maternal parenting practices as enacted during naturalistic interactions at home in 923 primiparous mothers with their 5-month-old infants in 9 different cultures. Undertaking an evaluation of parenting infants in the context of a broad international research design permits a comprehensive yet judicious analysis of maternal caregiving as well as a determination of the cultural commonality versus specificity of parenting practices with young infants. The particular cultural comparisons explored here set up direct contrasts among childrearing conditions by disentangling them (to the degree possible) from economic and educational, metropolitan-rural, modern-traditional, as well as ecological and climatic factors (van der Vijver & Leung, 1997). This study also contributes information about parenting infants in several still relatively underresearched populations and compares it to parenting infants in populations that have been more comprehensively studied.
Commonalities of Human Infancy and Parenting Infants
Infancy is an easily definable stage of life, based on biological status, mental data, as well as social convention. “A primary function of culture in shaping human experience is the division of the continuum of human development into meaningful segments, or ‘stages’. … [and] all cultures … recognize infancy as a stage of human development” (Harkness & Super, 1983, p. 223). Notably infants’ nature and actions influence parenting. For example, infant physiognomy attracts adults. The infant has a large head dominated by a disproportionately large forehead, widely spaced sizeable eyes, a small and snub nose, an exaggeratedly round face, and a small chin (Kringelbach et al., 2016). The ethologist Lorenz (1935/1970) famously contended that these facial features of “babyishness” universally provoke adults to reflexively express caregiving towards infants. In proposing this theory of Kindchenschema, Lorenz theorized innate releasing mechanisms for positive affect and solicitous adult parenting practices. Reciprocally, as an altricial species human infants not only are totally dependent on parents, but infants may be especially susceptible and responsive to external events and experiences. The dual phenomena of enhanced parental influence and prolonged susceptibility to experience in early childhood are thought by some to constitute prime evolutionary reasons for the extended neoteny of Homo sapiens (Bjorklund & Myers, 2019; Gould, 1977).
For their part, human beings appear to possess some intuitive knowledge about parenting, and some aspects of parenting infants may be wired into the common biological makeup of the human nervous system (Papoušek & Papoušek, 2002; Stark et al., 2019). For example, parents routinely speak to their infants even though they know that babies cannot understand language, and adults even speak to babies in a special speech register (“infant-directed speech”) that is thoroughgoing in its prosodic, simplicity, redundancy, lexical, and content distinctiveness from adult-directed speech (Golinkoff et al., 2015). Infant-directed speech is intuitive, nonconscious, and (essentially) universal (Soderström, 2007); even deaf mothers modify their sign language when interacting with their babies in much the same ways (Erting et al., 1994). Likewise, mothers in different cultures respond in similarly to infant cries, and the same regions in the brains of mothers of different cultures are excited by infant cries (Bornstein et al., 2017).
Most of young infants’ worldly experiences stem directly from interactions they have with their parents, and parents directly influence infant development by their cognitions and, perhaps more immediately salient in the phenomenology of the infant, by their practices (Bornstein, 2019). This study attempts to gain greater purchase on the nature and structure of mothers’ parenting practices and parent-provided experiences to infants by assessing both across diverse cultural settings. It does so by comparing three possible models of the structure of parenting infants.
Three Models of Maternal Parenting
Theoretically, parenting practices could be unstructured or structured in any number of ways. Three possible structures are tested here. One possible structure might conceptualize parenting as quantitatively unidimensional in character, that is as “good/bad,” “sensitive/insensitive,” or “engaged/unengaged” with infants. This orientation underpins familiar and prominent psychoanalytic, personality, ethological, and attachment perspectives that cast maternal behavior as, for example, “good,” “good enough,” “sensitive,” “warm,” “adequate,” and the like. An alternative possible structure might assert that parenting infants is constituted of multiple styles, that is qualitatively independent nominal classes of parenting. This orientation underpins equally familiar and prominent categorical approaches that commonly describe parenting as, for example, “authoritative,” “authoritarian,” “permissive,” or “neglectful.”
Likely, each of these two schools of thought about maternal parenting practices enjoys some kernel of truth. However, one-factor dimensional and multi-factor style theories of parenting are constrained for a variety of reasons. For example, these conceptualizations either do not admit to behavioral variability in parenting or they consider variability as noise around views of parents as unchanging (or at least consistent) across contexts, interactions, time, and (even) children. They also fail to acknowledge that parents are flexible, thinking beings who modify their childrearing based on dynamic considerations, ranging from transient emotional states to long-term caregiving goals, and, as a result, these conceptualizations tend to discourage nuanced and differentiated perspectives on parenting and developmental questions about parenting. Moreover, these conceptualizations overlook bidirectionality, the possibility that child characteristics (as individual differences or as temporally shifting) affect parenting. Yet, the “child effects” and relational developmental systems literatures (Bell, 1968; Lerner, 2018) both stress the nuanced adjustments in caregiving that parents normally make across contexts and interactions with respect to children’s age and gender, appearance and activity, temperament and cognition. Finally, students of family life recognize that parents themselves change in many ways in their parenting over time as they even do with the same child at two ages and between their first- and laterborn children (Holden & Miller, 1999).
A third hybrid possible structure that responds to these shortfalls conceptualizes parenting as organized into factors and domains. Everyday interactions between mothers and babies are characterized by a broad mélange of individual attitudes and actions concerned with infant surviving and thriving. By contrast with prevailing one-factor dimensional and multi-factor style approaches to parenting, a third approach proposes that parenting practices are structured in (1) a small number of higher-order factors that (2) overarch and are complemented by a larger number of separate and identifiable domains. Each level of this higher-order/multiple-domain structure would accommodate context, interaction, time, and child.
As to (1) higher-order factors, akin to the dimensional and style schools, parenting young children is often characterized by two conceptually independent, developmentally significant caregiving factors. They have been variously called social and didactic, affiliative and exploratory, animate and inanimate, affective and informational, primary and secondary intersubjective, and relationship-centered and environment-oriented (Ainsworth et al., 1978; Bornstein, 2019; Bowlby, 1969; Hastings & Grusec, 1998; Penman et al., 1983; Trevarthen & Aitken, 2001; Worthman et al., 2002). Generally speaking, one higher-order factor is dyadic and encompasses interactions that focus on the mother and infant and capitalize on visual, verbal, affective, and physical parenting practices that optimally engage infants in warm, nurturing, supportive interpersonal exchanges. The second higher-order factor is extradyadic and encompasses mother-infant verbal and physical exchanges which turn outward from the dyad to incorporate properties, objects, and events in the natural and designed environments and stimulate infants to engage and understand their surround.
As to (2) multiple domains of parenting practices, caregiving is asserted to consist of a number of separate and specific spheres of action. Out of the enormously varied repertoire of individual maternal parenting practices, a core census of six adaptive domains derives from the aim to meet fundamental developmental tasks of infancy. The six are rooted in theoretical, evolutionary, and developmental considerations and are based on extensive observations and analyses of mother-infant interactions in multiple cultures around the world (Bornstein, 2015, 2019, 2022; Bornstein et al., 2017). The six are nurturant, physical, social, didactic, material, and language (Appendix A, Column 1). Nurturant caregiving meets the biological, physical, and health requirements of infants through such practices as feeding, thermoregulation, grooming, and clothing. Physical caregiving promotes infants’ gross and fine psychomotor balance and movement. Social caregiving includes visual, verbal, affective, and physical activities that engage infants in interpersonal exchanges, express affection, and involve social play as well as regulate affect and emotions. Didactic caregiving consists of drawing the infant’s attention to properties, objects, and events in the environment as well as labeling and describing, demonstrating and stimulating. Material caregiving provisions and organizes infants’ home and local environments, including the number, variety, and composition of inanimate objects available to the infant, level of ambient stimulation, limits on physical freedom, and overall physical components of infants’ experiences. Language caregiving subserves verbal communication with infants and supports all aspects of parenting and infant development; maternal speech is a major channel through which mothers maintain contact, interpret infant cues, respond to infants, introduce experiences, and express affection.
In brief, three models of parenting infants have been proposed: a one-factor dimensional model, a multi-factor style model, and a hybrid two-factor/six-domain model. The dimensional model predicts that parenting practices aggregate into a single factor; the style model predicts that parenting practices disaggregate into multiple factors; the hybrid model predicts a two-level structure with two higher-order parenting factors overarching six separate domains. The present studies submit these three models of parenting infants to empirical test, first, with a large U.S. national sample of primiparous mothers of 5-month-old infants and, then, with a large international sample of primiparous mothers of 5-month-old infants from 9 countries. The overall strategy was initially to use U.S. parenting data (capitalizing on its large N to form calibration and validation samples) to determine the best-fitting model of the structure of mothers’ parenting toward infants and then to evaluate whether the best-fitting structure also fit the larger cross-cultural sample.
TWO STUDIES: GENERAL METHODS, PROCEDURES, AND ANALYTIC PLAN
To address two questions about the nature and structure of parenting infants -- Is parenting structured in any way to fulfill the responsibilities of infant childrearing and meet the developmental tasks of infants? Is any parenting structure common across cultures? -- three models of maternal parenting practices were tested. These tests were based on extensive and detailed standardized observations and systematic comparisons of naturalistic parenting practice repertoires of new mothers with their young infants.
Participants in the Two Studies
A total of 923 primiparous mothers and their healthy 5-month-old infants participated in the two studies. Study 1 recruited 360 mother-infant dyads from the United States; Study 2 recruited 139 mother-infant dyads from Argentina, 117 from Belgium, 40 from Brazil, 59 from France, 31 from Israel, 100 from Italy, 47 from Japan, 30 from Kenya, and a subsample of 90 from the original Study 1 sample of 360 mother-infant dyads from the United States. (Additional details about the samples appear in descriptions of Study 1 and Study 2 participants below, and sociodemographic characteristics of their nations in the Supplementary Information SI Table 1.)
Mothers were recruited from hospital or published birth notifications, patient lists of medical groups, newspaper advertisements, and targeted mailings. Mothers who expressed a willingness to participate in home-based naturalistic observations with their infants and who, with their infants, satisfied the following developmental and sociodemographic criteria were included in these studies on a first-come-first-recruited basis. Mothers were at least 16 years of age at their child’s birth, and the majority were living in intact families; infants were firstborn only children, born at term, more than 1500 g at birth, healthy, and 5 months of age on average at the time of the observation. Approximately equal numbers of girls and boys were enrolled into each country sample. In addition to the mother and infant dyads that met these inclusion criteria and participated, 9 dyads were excluded across the two studies: 2 dyads from France, 2 from Japan, and 1 from the United States were excluded because the durations of their digital observation records that could be coded totaled less than 42 min; 1 infant from Argentina and 3 from the United States were also excluded because they slept more than 5 min during the first 50 min of recording.
Mothers were chosen as the parent figure for this study because sociobiologists argue that almost all mammalian species are matrilocal (Wilson, 1975), and most animal and human families matriarchal. Theoreticians distinguish between bringing a new individual into the world (childbearing) on the one hand and caring for an existing individual (childcaring) on the other. Whereas species lower in the phylogenetic hierarchy are principally bearers of young, mammals are devoted child carers (Bjorklund & Myers, 2019), and early maternal care is more common than paternal care among land mammals (Geary, 2000), where males provide care in fewer than 5% of species (Moller, 2003). Cross-cultural surveys and meta-analyses alike attest to the central role that mothers play in infant development (Holden & Miller, 1999; Leiderman et al., 1977; Weisner & Gallimore, 1977). Furthermore, mothers and fathers do not share the same parenting investment strategies (Bjorklund & Myers, 2019; Parke & Cookston, 2019), and the maternal role is better articulated and defined than is the paternal role (Bögels & Perotti, 2011). Mothers participate in childrearing activities at significantly higher rates than do fathers (Pew Research Center, 2019) where, on average, mothers spend between 65% and 80% more time than do fathers in direct one-to-one interaction with their infants (Parke & Cookston, 2019). According to Konner (2010), there is not a single traditional society in which fathers devote more time to childcare than mothers. Given societal dictates that mothers are also ultimately responsible for their young children’s health and well-being, females have traditionally also become more expert caregivers than males (Negraia et al., 2018; Parke & Cookston, 2019).
General Procedures Applicable to the Two Studies
Procedures.
Approximately 1 h of naturalistic activities of mothers with their infants was videorecorded, microcoded, and analyzed for each dyad. Meta-analyses have indicated that maternal practices are most stable for observations lasting 30 min to 1 h compared to shorter or longer observations (Crockenberg & Litman, 1990; Holden & Miller, 1999; James et al., 2012); briefer observations can be unstable, and lengthier observations are likely to include samples of highly varied activities or contexts (Miller et al., 1998). In these studies, attempts were made to remain faithful to a principle of ecological validity by focusing on naturalistic interactions between mothers and infants in their home setting; that is, the aim was to observe spontaneous activities of the two under the most natural and unobtrusive conditions possible. Studying dyads at home presumably maximized their comfort and increased the validity of the observations. All observations were also conducted in a standardized way to render the data comparable across diverse samples. Briefly, mothers were asked to behave in their usual manner and to disregard the videographer’s presence insofar as possible; videographers were always young females native to the country; beside the videographer, only mother and infant were present; and observations took place at times of the day when infants were awake and alert. Only after a conventional period of acclimation to the presence of the videographer and the camera (as recommended in McCune-Nicolich & Fenson, 1984; Stevenson et al., 1986) did recording commence. At the conclusion of each visit, both mother and videographer independently evaluated multiple characteristics of the home visit (see the Supplementary Information); ratings of these characteristics were used descriptively and as covariates in analyses.
Maternal Parenting Practices.
Eighteen individual parenting practices and two context indicators constitute primary parenting tasks and performance competencies of mothers of young infants. Together, they aggregate into the six parenting practice domains listed in Appendix A Column 1. Each practice and context indicator is operationally defined in Appendix A Column 2. Development of this census of maternal parenting practices involved extensive observations and collaborative discussions. First, narrative observational accounts of maternal parenting practices (and infant behaviors) were made in the field. Field testing and refinement were then conducted. In this way, initial, unstructured descriptive data were shaped into more structured observations and, ultimately, quantitative data. Subsequently, formal operational definitions of maternal parenting practices were developed (see Hartmann et al., 2015) to facilitate coder accuracy and consistency. These definitions represent discrimination rules for coding the target practices, and they met three main criteria: (a) the definitions were objective and referred to directly observable target practices; (b) the definitions were clear, unambiguous, and easily understood so that trained coders could accurately use them; and (c) the definitions required little or no inference.
Coding.
The 18 parenting practices and two context indicators were coded from videorecords via computer entry (see Appendix A). Coding took place on individual passes through the videorecords and was continuous and comprehensive or time-sampled or consisted of counts and ratings. Continuous and comprehensive coding was implemented for 13 single parenting practices or mutually exclusive and exhaustive sets of conceptually related parenting practices. In mutually exclusive coding, only one practice can be coded for each time unit of observation; in exhaustive coding, one practice must be coded for each observation unit. During this kind of coding, only practice initiations are recorded because the onset (start) of a new practice automatically signals the offset (end) of the preceding practice (Sackett, 1978). A set of objective parameters was programmed such that the minimum duration of a parenting practice was set to .30 s, and an interruption of a parenting practice for less than 1 s did not constitute a new instance of the practice. Continuous and comprehensive coding is rigorous and powerful, yielding unbiased estimates of the frequency and duration of maternal parenting practices on the basis of their occurrence in the uninterrupted, natural time flow (Bakeman & Gottman, 1986). Mutually exclusive and exhaustive coding enjoys numerous conceptual and statistical advantages (see Bakeman & Gottman, 1986; Kauffeld et al., 2018; Lupia, 2018; Sackett, 1978). Five maternal parenting practices were coded using time sampling. In time sampling, whether or not a practice occurred during a fixed time interval (here 10 min) is recorded (Bornstein, 2002; Suen & Ary, 1989). Finally, two parenting practices used counts and ratings. With coded videorecords in hand, detailed analyses were undertaken.
Molecular observation categories code narrowly defined and specific activities, are generally easier for coders to learn and to apply, require less interpretation and inference, and can later be aggregated into molar categories for summary data analyses (as done here for domains). The parenting practices studied here are concrete and highly relevant to infant development, and they have been observed in each culture (albeit at varying levels; Bornstein et al., 2017). Molecular coding categorizes overt behaviors at high temporal resolution and captures detailed information from observational data. Microcoding and microanalysis are traditional in parenting studies (Feldman & Eidelman, 2007; Galligan et al., 2018; Koulomzin et al., 2002; Northrup & Iverson, 2020; Papaligoura & Trevarthen, 2001).
Coders.
A small cadre of trained coders was employed, and coders addressed issues of cultural sensitivity and coder bias. First, as data from different cultures were to be coded, the need for multiple trained reliable coders with multiple ethnic heritages and checks on measurement techniques arose as coders may have difficulty interpreting activities in persons from cultural backgrounds different from their own, or culture-based stereotypes might differ. To address these problems, coding focused on molecular-level counts of the frequency and duration of parenting practices rather than molar-level global ratings. Moreover, all coders were required to become reliable with a set of standardized reference codings. Shared meaning systems served to ensure consensus among coders, and judicious training helped to create and consolidate shared perceptions among coders.
Second, coders of activities may miss information—the human visual and auditory senses can be insensitive or unreliable in detecting certain activities -- and coders can also suffer from information overload. When a large number of target activities occur within a short period of time, a coder may have difficulty detecting or recording all of them. Sometimes coders harbor or develop (correct or incorrect) hypotheses about the nature and purpose of an investigation, how participants should behave, or even what constitute “appropriate” data. Thus, coders may make systematic errors in assessment and hold biases based on their information-processing limitations and expectations.
To address issues of coder bias and to maintain the accuracy of quantitative measures, coders adhered to standardization procedures (Hartmann et al., 2015): (1) Only trained and experienced but naïve coders were recruited, and stringent training criteria were employed. Only coders who possessed the ability to sustain attention, who had a propensity for detail and precision and a commitment to scientific detachment, and who were analytically minded were recruited and trained. (2) Prior to actual coding, coders were trained to criterion performance accuracy and consistency on a series of criterion videorecords which had varied and representative samples of the target practices. Coders learned to code accurately, and the pressure of time was eliminated. (3) Where possible, videorecords were randomly assigned to coders. (4) Coders were cautioned about the potential negative effects of bias, and they remained naïve to the specific scientific questions of these studies. (5) To the degree possible, precise low-inference operational definitions of maternal parenting practices were used. Coders learned the operational definitions and scoring procedures of the observation system as presented in a formal training manual. (6) Coding drift was corrected by means of regular reliability checks with experienced coders and standardized codings. Coders knew that their codings would be regularly checked for reliability but did not know which specific codings would be used. If “drift” away from coding accuracy occurred, re-training sessions were conducted (Dishion et al., 2017; Hallgren, 2012; Thornberry & Brestan-Knight, 2011).
Coding Reliability.
Different statistical metrics of coder reliability were employed. For all continuously coded practices, Kappa (κ; Cohen, 1960, 1968) was used. Kappa is observed agreement beyond that expected by chance as a proportion of the possible agreement beyond that expected by chance. Kappa was based on agreement in each 5-s interval for each parenting practice. Kappas were always evaluated relative to the prevalence index, the bias index, and the maximum attainable κ (Cicchetti et al., 2006; Sim & Wright, 2005). For the time-sampled parenting practices, the Intraclass Correlation (ICC; McGraw & Wong, 1996a, 1996b; Shrout & Fleiss, 1979) in two-way random effects models was used. After coders achieved initial reliability, at least every tenth videorecord that they coded was independently coded by second coders. Between 11% and 33% of each cultural sample (depending on country and domain) was coded independently by pairs of coders to monitor intercoder reliabilities. Supplementary Information SI Table 2 gives intercoder reliabilities of mothers’ parenting practices by country; averages for the 6 maternal parenting practice domains across the 9 countries were: Nurture κ = .89, Physical ICC = .72, Social κ = .67, Didactic κ = .73, Material ICC = .85, and Language κ = .70.
Data.
Coding generated multiple measures which are termed interim variables of frequency and duration, proportion, and variety, density, and consistency (Appendix A Column 3). Continuously coded practices generated both frequency and duration interim variables. Frequency is the number of discrete times a mother engaged in a practice or, more precisely, the number of times the parenting practice was initiated by a mother during the observation. Duration is the total time that a mother engaged in the practice. All practices in the nurture domain yielded only duration interim variables. Time-sampled practices yielded a single interim variable, the proportion of time units in which the practice was observed. Context indicators of quantity and quality of objects yielded counts and ratings of three interim variables: variety, density, and consistency (see Appendix A).
Appendix A Column 4 specifies how interim variables were aggregated into final indicator variables for each maternal parenting practice domain. Indicator variables for continuously coded domains are mean standard scores (z-scores where M = 0, SD = 1) of the frequencies and durations of the interim variables over the first 50 min of the observation. Indicator variables that were derived from time-sampled practices were also mean standard scores of interim variables. The two nurture indicator variables are sums of separate groups of interim practices (durations), and the single indicator variable of nurture is the duration of one interim variable. Indicator variables of mother physical derived from time-sampled practices are mean aggregated proportions of the time units in which an interim practice was observed.
Four maternal parenting practice domains (nurture, physical, social, material) were composed of two or more indicator variables. Except for mother nurture, domain scores were computed as the simple mean of the indicator variables for the domain, the indicator variables being equally scaled interim variables (mean standard scores, proportions). Because the indicator variables for mother nurture are duration scores of unequal means and standard deviations, the domain score was computed as the mean of the standard scores of the indicator variables. Appendix A Column 5 specifies how indicator variables were aggregated to form each maternal parenting practice domain score.
In summary, Appendix A Column 2 lists the maternal parenting practices that were coded and their operational definitions. Column 3 lists the interim variables derived from coding those practices. Column 4 lists the final indicator variables derived from the interim variables. Column 5 lists the mean standard score of final variables that constitute a domain and on which analyses were conducted. Column 1 lists the maternal parenting practice domain names. Because of the several advantages associated with microcoding, the videorecords were first coded at the molecular level of practice and context indicators (Bakeman & Gottman, 1986; Paul et al., 1986). To eliminate unnecessary complexity and enable clearer pictures of the nature and structure of maternal parenting practices, dependent variables were then aggregated at molar levels of domains, and analyses reported here are restricted to domains. Restricting analyses to domains takes full advantage of the multiple benefits of dimension reduction (Waggoner, 2021).
The concept of a practice domain used here is akin to an index and distinct from the concepts of a factor or latent variable. The latter refer to an unobserved construct that manifests itself in, and is inferred from, several theoretically and empirically related indicator variables. The parenting practice domains as used here consist of conceptually related practices that may, or may not, be empirically or statistically related (Bradley, 2004; Streiner, 2003). Justification for inclusion in a domain is conceptual coherence (qua an index), not necessarily empirical relatedness (qua a scale).
Preliminary Analyses and Analytic Plan.
Prior to all analyses, univariate distributions for all domain scores were examined for normality and outliers (Fox, 1997; Tabachnick & Fidell, 2019), and the distance of each case to the centroid was evaluated to screen for multivariate outliers (Bollen, 1987). To approximate normality and reduce the number and influence of outliers, variables were transformed as needed. Mother physical was re-expressed using a square-root transformation, and mother social and didactic were re-expressed using a log10 transformation. The same transformations were applied to calibration and validation U.S. samples and the cross-cultural data set. All analyses were conducted on transformed data.
The best-fitting factor structure of maternal parenting practices was evaluated in multiple steps using structural equation models (SEM). All SEMs were fit using maximum likelihood functions and followed the mathematical models of Bentler and Weeks (1980) as implemented in EQS (Bentler, 2008). Model fit was assessed using multiple, convergent indices, including the robust Satorra-Bentler (Satorra & Bentler, 1994, 2001) scaled χ2 statistic, robust comparative fit index (Robust CFI; Bentler, 1990), and the root mean square error of approximation (RMSEA; Browne & Cudeck, 1993) and its 90% confidence interval. To enhance the cross-validation adequacy of models, the Akaike Information Criterion (AIC; Akaike, 1987; Kaplan, 2000) was monitored for its decreasing value in all nested models. Cutoff values close to .95, .08, and .06 for CFI, SRMR, and RMSEA, respectively, are indicative of a relatively good fit between a hypothesized model and the observed data (Hu & Bentler, 1999). For models with relatively small sample sizes (Ns ≤ 250), the standardized root mean squared residual (SRMR; Browne & Cudeck, 1993) is reported.
Prior to fitting any SEM, bivariate plots were inspected to confirm that maternal parenting practice domains were linearly related and that no curvilinear effects existed between pairs of domains. In the course of fitting SEMs, Mardia (1970) coefficients of multivariate kurtosis and cases that contributed most to those estimates as well as the stability of parameter estimates and cases that contributed disproportionately to those estimates were evaluated. In the SEM, univariate measures of kurtosis and normalized estimates of Mardia’s multivariate coefficients indicated no significant problems of nonnormality.
STUDY 1: MATERNAL PARENTING INFANTS IN THE UNITED STATES
The purposes of Study 1 were to develop and compare factor structure solutions in calibration and validation subsamples for mothers’ parenting of infants among three models – a one-factor dimensional model, a multi-factor style model, and a hybrid two-factor/six-domain model – and to test the robustness of the best fitting model to variations in infant gender as well as maternal age, education, and personality.
Study 1: U.S. Total Sample
The U.S. Total Sample consisted of 360 European American mothers and their 5-month-old infants, 162 mother-daughter and 198 mother-son dyads. Sociodemographic information for participating U.S. mothers and infants appears in Table 1. Mothers averaged 29.4 years of age (SD = 6.2, range = 16.3 to 43.1). Twenty-nine mothers had not completed high school, 43 had completed high school only, 76 had finished partial college, 104 had graduated from college, and 108 had enrolled in or completed university graduate programs. Families ranged from low to upper-middle socioeconomic status (SES; Hollingshead, 1975, 2011, Four Factor Index of Social Status, range = 14 to 66; see also Adams & Weakliem, 2011). Infants averaged 163.4 days of age (SD = 6.1, range = 141 to 195) when observed. An ethnically homogenous European American community sample was recruited, first, because a majority of the population of the United States identifies as European American (American Community Survey, 2018) and, second, as an initial step toward understanding the nature and structure of maternal parenting practices in advance of embarking on the more complex follow-up Study 2 and analysis with more culturally diverse samples (Bornstein et al., 2013; Jager et al., 2017).
TABLE 1.
Sociodemographic characteristics of the U.S. total and calibration and validation subsamples.
| Total Sample (N = 360) |
Calibration Subsample (n = 180) |
Validation Subsample (n = 180) |
||||
|---|---|---|---|---|---|---|
| M | SD | M | SD | M | SD | |
| Mother age (in years) | 29.4 | 6.2 | 29.4 | 5.8 | 29.4 | 6.5 |
| Mother educationa | 5.6 | 1.3 | 5.6 | 1.3 | 5.6 | 1.3 |
| Family socioeconomic statusb | 50.1 | 12.9 | 49.9 | 13.2 | 50.2 | 12.6 |
| Infant age (in days) | 163.4 | 6.1 | 162.9 | 5.7 | 163.8 | 6.5 |
| Infant sex (% girls) | 45 | -- | 50 | -- | 40 | -- |
| Infant birth weight (in grams) | 3493.3 | 508.1 | 3476.3 | 537.1 | 3510.3 | 478.4 |
Note.
Hollingshead (1975, 2011) 7-point scale.
Hollingshead (1975, 2011) Four Factor Index of Social Status.
Study 1: Procedures and Analytic Plan
Data were collected following the general procedures detailed above. Here, additional variables and main analytic considerations for Study 1 are described.
Additional Variables.
Infant gender and maternal age, education, and personality were included in the Study 1 analyses. Maternal personality was assessed using the Jackson Personality Inventory (JPI; Jackson, 1976). For purposes of this analysis, three principal components and two JPI subscales were used following Paunonen and Jackson’s (1996) model as measures of the “Big Five” personality factors. The Openness factor consisted of the JPI Breadth of Interest (M = 47.9, SD = 10.4) and Innovation scales (M = 50.5, SD = 9.7). The Neuroticism factor consisted of the Anxiety (M = 48.4, SD = 9.7), Interpersonal Affect (M = 48.2, SD = 8.6), and Conformity (M = 47.1, SD = 9.1) scales. The Extraversion factor consisted of the Self-esteem (M = 53.4, SD = 8.8) and Social Participation (M = 46.9, SD = 8.9) scales. The three factors were computed as the principal components of their constituent scales. The scale score Responsibility (M = 56.5, SD = 7.3) was taken as a single index of Trustworthiness, and the scale score Organization (M = 54.4, SD = 9.8) was taken as a single index of Conscientiousness.
Analytic Plan.
The factor structures of mothering in the United States for the three models were evaluated in five steps. In Step 1, the U.S. Total Sample was randomly split in half to form calibration and validation subsamples. Because the findings of a factor analysis can be sample-dependent, a best practice is to test and refine a model in a calibration sample and then cross-validate to test the replicability of the solution in an independent validation sample (Floyd & Widaman, 1995). Fit of the three competing a priori theoretical mothering models was tested in parallel searches for the best-fitting factor structure in the two subsamples (MacCallum et al., 1992). In Step 2, invariance across the calibration and validation subsamples was tested. In Step 3, the final model in the U.S. Total Sample was tested. In Step 4, the final model was tested for multiple-group invariance in mothers with infant daughters and mothers with infant sons. In Step 5, the final model was re-evaluated controlling for maternal age, education, and personality.
Study 1: Results
Goodness-of-fit statistics for all models appear in Table 2. Results of the five steps in the Analytic Plan are described.
TABLE 2 -.
United States model test goodness-of-fit statistics.
| Model | Name | χ2 | Value | p | CFI | AIC | SRMR | RMSEA | 90% CI | AASE |
|---|---|---|---|---|---|---|---|---|---|---|
| Separate validation and calibration samples | ||||||||||
| US1 | One-factor – calibration | S-B χ2(9) | 89.10 | < .001 | .58 | 71.10 | .13 | 22 | (.18, .27) | .08 |
| US2 | Multi-factor – calibration | S-B χ2(15) | 186.37 | < .001 | .00 | 156.37 | .24 | .25 | (.22, .28) | .18 |
| US3 | Hybrid – calibration | S-B χ2(9) | 46.24 | < .001 | .78 | 28.24 | .10 | .15 | (.11, .20) | .05 |
| US4 | Hybrid revised – calibration | S-B χ2(7) | 4.88 | .67 | 1.00 | −9.12 | .03 | .00 | (.00, .07) | .02 |
| US5 | Hybrid revised – validation | S-B χ2(7) | 4.58 | .71 | 1.00 | −9.42 | .03 | .00 | (.00, .07) | .01 |
| Invariance of hybrid revised across validation and calibration | ||||||||||
| US6 | Configural | χ2(14) | 8.75 | .85 | 1.00 | -- | -- | .00 | (.00, .04) | -- |
| US7 | Constrained | χ2(28) | 20.58 | .84 | 1.00 | -- | -- | .00 | (.00, .03) | -- |
| US7-US6 | Comparison | Δχ2(14) | 11.83 | .62 | -- | -- | -- | -- | -- | -- |
| Combined validation and calibration samples | ||||||||||
| US8 | Hybrid revised | S-B χ2(7) | 8.23 | .31 | 1.00 | -- | -- | .02 | (.00, .07) | .02 |
| Invariance of hybrid revised combined sample across boys and girls | ||||||||||
| US9 | Configural | χ2(14) | 13.61 | .48 | 1.00 | -- | -- | .00 | (.00, .07) | -- |
| US10 | Constrained | χ2(28) | 35.90 | .15 | .97 | -- | -- | .04 | (.00, .07) | -- |
| US10-US9 | Comparison | Δχ2(14) | 22.29 | .07 | -- | -- | -- | -- | -- | -- |
| Hybrid revised combined sample with covariates | ||||||||||
| US11 | Covariate | S-B χ2(38) | 45.58 | .19 | .98 | -- | .03 | (.00, .05) | .03 | |
Step 1: Testing the Three a priori Theoretical Models of Maternal Parenting in Calibration and Validation Subsamples.
(1) In a test of the one-factor dimensional model (Model US1), a model was fit to the calibration data in which all maternal domains load, without covarying errors, on a single mothering factor. The overall χ2 test, as well as the CFI, SRMR, and RMSEA, indicated that this model left considerable covariation unexplained, and the nurture domain had a nonsignificant negative loading on the single factor. The one-factor dimensional model of parenting was therefore rejected. (2) A formal test of fit of the multi-factor style model (Model US2) was conducted in which no maternal domains covary to the calibration data. Model US2 fit statistics indicated that this model left considerable covariation unexplained. The multi-factor style model of parenting was therefore rejected as well. (3) In the hybrid two-factor/six-domain model, the nurture, physical, social, didactic, and language domain scores were hypothesized to load positively on a factor representing a Dyadic focus in mother-infant interactions, the didactic and material domain scores were hypothesized to load positively on a factor representing an Extradyadic focus of mother-infant interactions, and the two factors were hypothesized not to covary. The didactic domain was hypothesized to cross-load on the Dyadic and Extradyadic factors because the didactic domain involves bringing the extradyadic world (environment) into dyadic interaction. In the didactic domain, mothers serve as mediators of the child’s interaction with their environment. Relative to the fit statistics for the one-factor model and those for the multi-factor model, the fit of the a priori two-factor model (Model US3) was improved; however, this model also failed to account for significant covariation. Examination of the Lagrange multiplier tests revealed substantial improvement in model fit would be gained by incorporating a single path specifying a negative loading of the domain score for nurture on Extradyadic: As mothers are engaged in feeding, grooming, and clothing their infants, they focus within the mother-infant dyad and draw focus of the dyad away from the extradyadic environment. Goodness of fit indexes for this re-specified model (Model US4) on the calibration data were excellent. All loadings of maternal practices on their respective first-order factors were significant at the .001 level or better. (4) Similar model misfit was found across the calibration and validation subsamples for the 3 a priori mothering models (data not shown). However, the final two-factor/six-domain model (Model US5) demonstrated an excellent fit to the validation data. All loadings of maternal parenting practices on their respective first-order factors were significant at the .05 level or better. Standardized factor loadings for the two subsamples appear in SI Table 3.
In brief, fit statistics indicated that the one-factor and multi-factor models each left considerable covariation unexplained. By contrast, fit statistics for a re-specified hybrid two-factor/six-domain model were excellent.
Step 2: Testing Invariance of the Hybrid Two-Factor/Six-Domain Parenting Model in the Calibration and Validation Subsamples.
For purposes of cross-validation, a series of nested multiple-group models was constructed across the calibration and validation subsamples that sequentially introduced constraints on factor loadings, factor variances, and unique variances (Bollen, 1989). Judgment of invariance was based on model fit indices of the constrained model, the deterioration of model fit when imposing constraints, and Lagrange multiplier tests of equality constraints. To avoid the problem of capitalization on chance due to the number of the equality constraints tested, a change in χ2 of 5 or more in the univariate Lagrange multiplier test was required to set the constraint free (Jöreskog & Sörbom, 1984; Scott-Lennox & Lennox, 1995).
A preliminary test in which no parameter estimates were constrained to be equal (Model US6) fit the data, suggesting that more restrictive models were appropriate. At each hierarchy of equality constraints, model fit indexes were evaluated and deemed appropriate; reported results are based on the most restrictive model. Results from multiple-group models in which factor loadings, factor variances, and unique variances were constrained to be equal (Model US7) yielded an excellent fit to the data. The difference in χ2 between this Model US7 and Model US6 which imposed no invariance constraints was not significant, suggesting that imposing total invariance across all estimated parameters had no deleterious effects on model fit. All equality constraints on factor loadings, factor variances, and unique variances were found to be tenable, probability values ranged from .15 to .89 for factor loadings, were .34 and .36 for factor variances, and ranged from .19 to .88 for unique variances.
In brief, the hybrid model was tested for its invariance across the calibration and validation subsamples. The calibration (Model US4) and validation (Model US5) subsamples had equivalent factor structures (Models US7-US6), and so they were combined.
Step 3: Fitting the Hybrid Two-Factor/Six-Domain Parenting Model in the U.S. Total Sample.
Figure 1 presents the standardized solution to the hybrid two-factor/six-domain model of parenting infants in the U.S. Total Sample (N = 360). All loadings of maternal parenting practices on their respective first-order factors were significant at the .001 level or better. (The correlation matrix, variances, and standardized residuals for the final model appear in SI Table 4.) To test replicability of the final two-factor/six-domain model, bootstrap procedures with 5000 full-size resamples were performed. The results confirmed that the model was robust and stable: The mean factor loadings ranged from .30 to .65, all ps < .001, with estimated standard errors ranging from .05 to .07.
Figure 1.
Standardized solution for the final model of the factorial structure of mothering based on the full U.S. sample (N = 360). (In this and/or subsequent figures, numbers associated with single-headed arrows are standardized path coefficients; numbers associated with double-headed arrows are standardized covariance estimates. Arrows associated with dependent variables are error or disturbance terms.)
In brief, examination of the factor loadings indicated that a Dyadic factor of maternal parenting practices was expressed in mothers’ nurturing infants, promoting infants’ physical development, social interaction with infants, didactic exchange with infants, and language to infants in that order; an Extradyadic factor was expressed in mothers’ didactic exchanges with infants and material provision of their infants’ environment while at the same time nurturing infants less. A significant amount of variance (38.7%) among the six parenting domain scores was explained by the two factors of Dyadic and Extradyadic parenting. That said, the error terms in Figure 1 for the six individual parenting domains indicate that a significant part of the variance in each domain was also unaccounted for by the parenting Dyadic and Extradyadic factors. Indeed, the range of unique variances across the six parenting domains is 31.4% to 86.5%, and an undefined part of each of the unique variances is potentially variance specific to the domain. The U.S. data are therefore consistent with a hybrid model of parenting in which two higher-order factors (Dyadic and Extradyadic) are instantiated in six separately distinguishable domains (nurture, physical, social, didactic, material, and language).
Step 4: Equality of Model Fit of the Two-Factor/Six-Domain Parenting Structure in Mothers with Daughters and Mothers with Sons.
To test whether the hybrid model fit equally well for mothers of infant girls (n = 162) and mothers of infant boys (n = 198), a series of nested multiple-group models was constructed that sequentially introduced constraints on factor loadings, factor variances, and unique variances. A preliminary test in which no parameter estimates were constrained to be equal (Model US9) fit the data, suggesting that more restrictive models were appropriate. At each hierarchy of equality constraints, the model fit indexes were evaluated and deemed appropriate; reported results are based on the most restrictive model. Results from multiple-group models in which factor loadings, factor variances, and unique variances were constrained to be equal (Model US10) yielded a reasonable fit to the data. The difference in χ2 between this Model US10 and Model US9 which imposed no invariance constraints was not significant, suggesting that imposing total invariance across all estimated parameters had no deleterious effects on model fit. All equality constraints on factor loadings, factor variances, and unique variances were deemed tenable indicating that the samples of mothers with daughters and mothers with sons had equivalent factor structures.
In brief, the final hybrid two-factor/six domain model of parenting fit equally well for U.S. European American primiparous mothers with infant daughters and infant sons.
Step 5: The Two-Factor/Six-Domain Parenting Structure, Controlling for Maternal Age, Education, and Personality.
Last, whether the hybrid structure of maternal parenting practices would hold while controlling for maternal chronological age, educational achievement, and personality traits was tested. This model was based on 270 mothers who had complete sociodemographic and personality data. Maternal education was measured on the Hollingshead Index (1975, 2011) 7-point education scale, with values ranging from 1, less than 7th grade to 7, graduate/professional training.
Several covariates were related to one or more domain scores at the zero-order level. In the SEM, maternal age, education, Openness, Neuroticism, Trustworthiness, and Conscientiousness were therefore added as exogenous variables to the two-factor/six-domain structure and the factorial validity of this structure was re-evaluated in the context of these sociodemographic and personality factors. Trustworthiness was omitted from the final model because it did not relate to any of the maternal parenting practices when evaluated simultaneously with other variables in the model. Maternal age, r(268) = .12, p < .05, and Trustworthiness, r(268) = .13, p < .05, were related to nurture; maternal education, r(268) = .17, p < .01, and Openness, r(268) = .18, p < .01, were related to physical; maternal age, r(268) = .20, p ≤ .001, education, r(268) = .17, p < .01, and Neuroticism, r(268) = .15, p < .05, were related to social; Conscientiousness, r(268) = .14, p < .05, was related to material; and maternal age, r(268) = .41, p < .001, education, r(268) = .35, p < .001, and Openness, r(268) = .18, p < .01, were related to language.
Figure 2 shows that the final two-factor/six domain model (Model US11) for maternal parenting practices held when controlling for maternal age, education, and personality. All loadings of the maternal parenting practices on their respective first-order factors were significant at the .001 level or better. In addition, older mothers spoke more to their infants; mothers with more education focused more on dyadic interactions with their infants; mothers who rated themselves as more open in personality encouraged their infants’ physical development more; mothers who rated themselves as more neurotic engaged in more social exchanges with their infants; and mothers who rated themselves as more organized provided their infants with more objects and/or objects with more stimulating features for exploration.
Figure 2.
Standardized solution for the final model of the factorial structure of mothering based on the U.S. sample (n = 270), controlling for maternal age, education, and personality.
In brief, the final two-factor/six domain model for the structure of maternal parenting practices with infants held when controlling for maternal age, education, and personality.
Study 1: Discussion
European American first-time mothers’ parenting young infants has a hybrid bifactorial (Dyadic and Extradyadic) structure that is expressed in six parenting domains (nurture, physical, social, didactic, material, and language) that encompass 20 quotidian parenting practices. Figure 3A shows the structure of Dyadic parenting, its domains, and constituent practices, and Figure 3B shows the structure of Extradyadic parenting, its domains, and constituent practices. The overall structure of parenting infants appears to be robust and obtains for mothers with daughters and sons and over and above variation in maternal age, education, and personality. The General Discussion sets these results in a larger empirical and theoretical context.
Figure 3.
A. The structure of Dyadic parenting, its domains, and constituent practices. B. The structure of Extradyadic parenting, its domains, and constituent practices.
STUDY 2: MATERNAL PARENTING INFANTS ACROSS CULTURES
After more than a century of study, too little is still known about parenting as well as the life circumstances and experiences of parents and infants across a broad range of cultural settings. A survey of 10 international journals concerned with psychological aspects of infancy revealed that only 8% of articles published between 2002 and 2012 were based on populations not indigenous to Western Europe, North America, or Australasia (Tomlinson et al., 2014). With this circumstance in mind, and to test the wider generality of the hybrid two-factor/six domain model of maternal parenting practices, the nature and structure of mothers’ parenting of infants were studied in 9 contrasting cultural groups around the world. To the extent that research in developmental science is dominated by Western samples, it is challenging to distinguish universal mechanisms from Western-specific ones. Some universals in parenting likely exist, as parenting in different places likely draws on the same human neural, mental, and emotional machinery, just as infants likely elicit similar responses from parents that may be requisite for their wholesome development. However, human behaviors are known to differ (sometimes quite dramatically) across populations in different cultures, and human parenting is one prominently culturally variable behavior.
Cross-cultural research usually compares two (or just a few) cultures. To assume that one culture lies toward the opposite pole from another culture on some unitary dimension, and that the two are otherwise equivalent, is usually to assume in error; cultures are complex entities that differ from one another in manifold ways. However, the number of rival explanations of a common phenomenon can be reduced when the number of samples compared is increased (Campbell, 1986), so the larger the number of cultures studied the more compelling is the conclusion that any observed generic findings (e.g., about parenting) may be robust and culturally common. For Study 2, therefore, mothers and infants were recruited in South American, North American, European, African, Middle Eastern, and East Asian cultures that also varied on a variety of dimensions of possible cultural comparison (SI Table 1).
Study 2: Cross-Cultural Samples
Altogether 653 mothers and their healthy 5-month-old infants from Argentina, Belgium, Brazil, France, Israel, Italy, Japan, Kenya, and the United States participated. Because a test of invariance in factor loadings across 9 countries was planned in the analysis, it was desirable to include U.S. mothers in the cross-cultural comparison; to make the U.S. sample size comparable to sample sizes in the other countries, data were included from a subsample of the U.S. mothers (and infants) who participated in Study 1, and that subsample did not differ from the U.S. Total Study 1 Sample in the means and variances of three key sociodemographic characteristics (mothers’ age and education and family SES). As the same factor structure and comparable model fit indices were obtained with and without the U.S. mothers in the cross-cultural sample, reported results are based on the whole cross-cultural sample (including U.S. mothers).
Table 3 presents sociodemographic information for all participants in each country. Mothers averaged 27.7 years of age (SD = 4.6, range = 16.2 to 44.0). Because differences exist between countries in the duration, quality, and content of schooling, bicultural researchers adjusted mothers’ years of schooling in each culture so that all education scales were equivalent to the Hollingshead U.S. scale. Mothers’ average educational level as measured relative to the Hollingshead Index (1975, 2011) 7-point education scale was 4.6 (SD = 1.3, range = 1 to 7). Although the 9 samples were middle-class on average, they ranged from low to upper-middle SES (as measured by the Hollingshead: M = 43.2, SD = 12.6, range = 8 to 66). Infants averaged 161.1 days of age (SD = 8.0, range = 131 to 198).
TABLE 3.
Sociodemographic characteristics of the cross-cultural samples.
| Mother’s Age | Mother’s Educationa | Family SESb | Infant’s Agec | Infant’s Sex (% girls) |
Infant’s Birth Weightd | ||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| M | SD | M | SD | M | SD | M | SD | M | SD | ||
| Argentina (n = 139) | 25.3 | 4.9 | 3.9 | 1.4 | 34.5 | 14.6 | 165.1 | 8.3 | 46.8 | 3343.5 | 462.3 |
| Belgium (n = 117) | 29.3 | 3.6 | 5.2 | 1.1 | 47.6 | 11 | 157.1 | 9.7 | 47 | 3405.9 | 453.2 |
| Brazil (n = 40) | 25.8 | 5.9 | 3.9 | 1.9 | 36.4 | 13.9 | 156.7 | 5.6 | 52.5 | 3349.7 | 379.2 |
| France (n = 59) | 30.8 | 4.7 | 5.4 | 1.3 | 53.4 | 10.7 | 166 | 10.9 | 44.1 | 3252.5 | 366.5 |
| Israel (n = 31) | 28 | 3.5 | 5.4 | 0.7 | 50.9 | 6.7 | 166.3 | 4.7 | 51.6 | 3346.2 | 433.6 |
| Italy (n = 100) | 27.4 | 4.5 | 3.5 | 1.5 | 33.4 | 13.6 | 154.9 | 5.1 | 50 | 3247.5 | 388.4 |
| Japan (n = 47) | 29 | 2.9 | 5.6 | 0.9 | 52.8 | 11 | 162 | 9.1 | 51.1 | 3039.8 | 364.9 |
| Kenya (n = 30) | 21.7 | 3.4 | 2.5 | 1.7 | -- | -- | 159.1 | 9.1 | 50 | 2913.7 | 558.9 |
| United States (n = 90) | 29.6 | 6 | 5.6 | 1.3 | 50.3 | 12.9 | 164.1 | 6.3 | 43.3 | 3539.9 | 537.3 |
| Overall (n = 653) | 27.7 | 4.6 | 4.6 | 1.3 | 43.2 | 12.6 | 161.1 | 8.0 | 47.6 | 3317.7 | 446.1 |
Note.
Hollingshead (1975, 2011) 7-point scale.
Hollingshead (1975, 2011) Four Factor Index of Social Status.
days.
grams.
Study 2: Procedures and Analytic Plan
Data were collected following the general procedures detailed above. Here, additional analytic considerations for Study 2 are described. The hybrid two-factor/six domain structure of maternal parenting practices was evaluated in four steps for the cross-cultural sample. In Step 1, the hybrid factor structure based on the total cross-cultural sample is reported. In Step 2, the final model (in the total sample, collapsing across countries) was tested for multiple-group invariance in mothers with daughters and mothers with sons. In Step 3, the model was re-evaluated controlling for maternal age and education. In Step 4, invariance in factor loadings across the 9 cultures was tested.
Study 2: Results
Descriptive statistics of domain scores for mothers in the 9 cross-cultural samples appear in Table 4. The main analyses concern the factor structure of mothers’ parenting infants across cultures. Goodness-of-fit statistics for all models appear in Table 5.
TABLE 4.
Maternal domain scores for the cross-cultural samples.
| Maternal domains | Argentina (n = 139) |
Belgium (n = 117) |
Brazil (n = 40) |
France (n = 59) |
Israel (n = 31) |
Italy (n = 100) |
Japan (n = 47) |
Kenya (n = 30) |
United States (n = 90) |
|||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| M | SD | M | SD | M | SD | M | SD | M | SD | M | SD | M | SD | M | SD | M | SD | |
| Nurture | −.04 | .60 | −.11 | .63 | .10 | .47 | .15 | .69 | −.31 | .56 | .18 | .56 | −.23 | .54 | .41 | .51 | −.04 | .60 |
| Physical | .11 | .10 | .05 | .06 | .09 | .09 | .08 | .08 | .05 | .06 | .05 | .06 | .17 | .13 | .18 | .12 | .18 | .14 |
| Social | .36 | .97 | −.21 | .49 | .05 | .61 | .02 | .72 | −.18 | .56 | −.04 | .62 | −.02 | .61 | −.32 | .63 | −.10 | .50 |
| Didactic | .26 | .95 | −.39 | .59 | −.09 | .80 | .17 | .96 | −.05 | .78 | .06 | .94 | .20 | 1.17 | −.72 | .53 | .12 | 1.00 |
| Material | −.06 | .54 | .17 | .64 | −.36 | .45 | .26 | .59 | .63 | .76 | −.33 | .59 | −.25 | .52 | −.70 | .49 | .38 | .61 |
| Language | .04 | .95 | .12 | .75 | −.05 | .71 | .15 | .80 | −.12 | .68 | .12 | .86 | .04 | .83 | −1.35 | .61 | .05 | .89 |
Note. Physical domain score was mean proportion of intervals; all other domain scores were mean standard aggregate scores calculated across the total cross-cultural sample.
Table 5 –
Cross-cultural model test goodness-of-fit statistics.
| Model | Name | χ2 | Value | p | CFI | AIC | SRMR | RMSEA | 90% CI | AASE |
|---|---|---|---|---|---|---|---|---|---|---|
| CC1 | Hybrid | S-B χ2(7) | 67.53 | < .001 | .90 | -- | -- | .12 | (.09, .14) | -- |
| CC2 | Hybrid revised | S-B χ2(5) | 31.37 | < .001 | .96 | -- | -- | .09 | (.06, .12) | .02 |
| Invariance of hybrid revised across boys and girls | ||||||||||
| CC3 | Configural | χ2(10) | 36.02 | < .001 | .96 | -- | -- | .09 | (.06, .12) | -- |
| CC4 | Metric | χ2(26) | 53.01 | < .001 | .96 | -- | -- | .06 | (.03, .08) | -- |
| CC4-CC3 | Comparison | Δχ2(16) | 16.99 | .39 | ||||||
| Hybrid revised with covariates | ||||||||||
| CC5 | Covariate | S-B χ2(14) | 72.58 | < .001 | .94 | -- | -- | .08 | (.06, .10) | .03 |
| Invariance of hybrid revised across countries | ||||||||||
| CC6 | Configural | χ2(54) | 106.01 | < .001 | .98 | -- | -- | .12 | (.08, .15) | -- |
| CC7 | Metric | χ2 (101) | 177.59 | < .001 | .98 | -- | -- | .10 | (.08, .13) | -- |
| CC7-CC6 | Comparison | Δχ2(47) | 71.58 | < .05 | ||||||
| CC8 | Partial Metric | χ2(98) | 149.85 | < .001 | .98 | -- | -- | .086 | (.06, .11) | -- |
| CC8-CC6 | Comparison | Δχ2(44) | 43.84 | .48 | ||||||
Note. S-B χ2 = Satorra-Bentler (Satorra & Bentler, 1994, 2001) scaled χ2 statistic. RCFI = Robust Comparative Fit Index (Bentler, 1990). AIC = Akaike Information Criterion (AIC; Akaike, 1987; Kaplan, 2000). RMSEA = Root Mean Square Error of Approximation (Browne & Cudeck, 1993) and its 90% CI = confidence interval. AASE = Average Absolute Standardized Error.
Step 1: The Two-Factor/Six-Domain Parenting Structure in Cross-Cultural Perspective.
When the two-factor/six-domain model of parenting for the United States (see Figure 1) was fit to the cross-cultural data set, all maternal domains loaded on their respective first-order factors at the .05 level or better. However, the overall χ2 test and the RMSEA indicated that this model (Model CC1) left some covariation unexplained. Lagrange multiplier tests identified two correlated unique variances: one between mother material and language, χ2(1) = 16.76, p < .001, and one between mother physical and social, χ2(1) = 14.92, p < .001. That is, mothers who arranged richer environments for their infants also spoke to them more, and mothers who promoted their infants’ physical development also interacted socially with them more.
Figure 4 presents the standardized solution to the final, modified two-factor/six domain model (Model CC2), with the two correlated unique variances added. All loadings of the maternal domains on their respective first-order factors were significant at the .05 level or better. (The correlation matrix, variances, and standardized residuals for this final model appear in SI Table 5.) To test replicability of the two-factor/six-domain model of parenting, bootstrap procedures were performed with 5000 full-size resamples. The results confirmed that the model was robust and reasonably stable: The mean factor loadings ranged from .11 (p < .05) to .70 (p < .001), with estimated standard errors ranging from .04 to .06.
Figure 4.
Standardized solution for the final model of the factorial structure of mothering based on the full cross-cultural sample (N = 653).
Much of the variance (39.2%) among the six parenting domain scores was explained by the two factors of Dyadic and Extradyadic parenting. That said, the error terms in Figure 4 for each of the six parenting domains indicate that significant amounts of variance were again unaccounted for by the two parenting factors. The range of unique variances across the six parenting domains is 31.4% to 96.0%.
In brief, parenting young infants in 8 cultures around the world, as in the United States, is best characterized by a bifactorial (Dyadic and Extradyadic) structure that is expressed in six parenting domains (nurture, physical, social, didactic, material, and language).
Step 2: Equality of Model Fit of the Two-Factor/Six-Domain Cross-Cultural Parenting Structure in Mothers with Daughters and Mothers with Sons.
To test whether the final hybrid model fit equally well for mothers of infant girls (n = 311) and mothers of infant boys (n = 342), a series of nested multiple-group models was constructed that sequentially introduced constraints on factor loadings, factor variances, unique variances, and covariances. A preliminary test in which no parameter estimates were constrained to be equal fit the data (Model CC3), suggesting that more restrictive models were appropriate. At each hierarchy of equality constraints, the model fit indexes were evaluated and deemed appropriate; reported results are based on the most restrictive model. Results from multiple-group models in which factor loadings, factor variances, unique variances, and covariances were constrained to be equal yielded a reasonable fit to the data (Model CC4). All equality constraints on factor loadings, factor variances, unique variances, and covariances were deemed tenable indicating that the samples of mothers with daughters and mothers with sons had equivalent factor structures and that the two-factor/six-domain model of parenting fit equally well for both samples (Model CC4-CC3).
In brief, the two-factor/six-domain structure of mothers’ parenting infants is robust and obtains for mothers with infant daughters and for mothers with infant sons.
Step 3: The Two-Factor/Six-Domain Cross-Cultural Parenting Structure, Controlling for Maternal Age and Education.
In total, 640 mothers provided data on maternal age and education (personality was not collected in all countries). In the SEM, maternal age and education were added as exogenous variables and the factorial validity of the two-factor/six-domain structure of maternal parenting practices re-evaluated in the context of these two sociodemographic variables. Figure 5 shows that the final Model CC5 fit the data. The hybrid structure for mothering in the cross-cultural data set held controlling for maternal age and education. All loadings of maternal domains on their respective first-order factors were significant at the .05 level or better. Across cultures, older mothers spoke more to their infants, and mothers who attained more education exhibited a greater dyadic focus when interacting with their infants and provided their infants with more objects and/or objects with more stimulating features for exploration.
Figure 5.
Standardized solution for the final model of the factorial structure of mothering based on the cross-cultural sample (n = 640), controlling for maternal age and education.
In brief, the two-factor/six domain structure of maternal parenting practices with infants is robust and obtains for mothers over and above variation in maternal age and education.
Step 4: Testing the Two-Factor/Six-Domain Parenting Model for Invariance Across Cultures.
A multiple-group model in which the final model as depicted in Figure 4 was used as the baseline model for each of the countries, and no parameter estimates were constrained to be equal, yielded a reasonable fit to the data (Model CC6), with the exception that the RMSEA, which is the standardized difference between the observed covariance matrix and the covariance matrix under the model, indicated marginal fit. The RMSEA is an index of error in the approximation of the covariance matrix under the model for each degree of freedom in the model. Browne and Cudeck (1993, p. 144) stressed that their suggested conventions of a close fit (RMSEA ≤ .05) and a reasonable fit (RMSEA ≤ .08) were based on “practical experience” and “subjective judgment” and “cannot be regarded as infallible or correct.” A good SRMR but a larger RMSEA than desirable means that, although there is a relatively close fit in the observed and model-implied covariance matrices, there is enough instability in the estimates to warrant caution in interpreting the configural invariance model. To require full invariance of the model across 9 cultural samples may be unrealistic (Putnick & Bornstein, 2016). It was decided, therefore, to test a model of metric invariance; that is, to test that the Dyadic and Extradyadic factors were similar constructs with equal metrics in each country. As such, all factor loadings were constrained to be equal across cultures and were tested simultaneously (Model CC7). Compared to the model that imposed no invariance constraints, the difference in χ2 (Model CC7-CC6) was significant, indicating that the null hypothesis of full metric invariance should be rejected.
To test partial metric invariance (Byrne et al., 1989), the equality constraints that were significant were released sequentially, starting with the constraint that had the largest χ2 based on the Lagrange multiplier test statistics. Examination of the Lagrange multiplier tests revealed that the significant increase in χ2 was due to a lack of invariance of four factor loadings (out of the 48 factor loadings that were constrained to be equal). After all four constraints were released, goodness-of-fit indexes for the final partial metric invariance Model CC8 were adequate. The difference in χ2, when compared to the model that imposed no invariance constraints (Model CC8-CC6), was not significant. Given this nonsignificant difference, the goodness-of-fit indexes, and the absence of a change in χ2 of 5 or more in the univariate Lagrange multiplier tests in the final model with four equality constraints released, partial metric invariance was supported. (Standardized factor loadings estimated in this final partial metric invariance model appear in SI Table 6.) One constraint released regarded a factor loading that was not significant: physical on the Dyadic factor in Brazil. Three other constraints released concerned quantitative differences among countries; all three factor loadings (social on the Dyadic factor in Argentina and didactic on the Extradyadic factor in Brazil and Japan) were significant and were in the same direction as those in other cultural groups.
In brief, the two-factor/six-domain structure of maternal parenting practices with infants was partially invariant across 9 countries, with 44 of 48 (92%) loadings showing invariance.
Study 2: Discussion
Mothers’ parenting of young infants around the world has a higher-order bifactorial (Dyadic and Extradyadic) structure that is expressed in six parenting domains (nurture, physical, social, didactic, material, and language). This structure of parenting infants appears to be robust and obtains for mothers with infant daughters and sons and over and above variation in maternal age and education. The General Discussion sets these results as well in a larger empirical and theoretical context.
GENERAL DISCUSSION OF STUDY 1 AND STUDY 2
The central goal of the work reported here was to explore the nature and structure of mothers’ parenting of young infants within and across cultures. Three theoretical models were compared: a one-factor dimensional model, a multi-factor style model, and a hybrid two-factor/six-domain model. Analyses of maternal parenting practices with infants in both U.S. national and 9-country cross-cultural samples supported the two-factor/six-domain model. Here we discuss the two higher-order factors, the six parenting domains, strengths and limitations of the two studies, and conclusions and implications of this work.
Two Higher-Order Parenting Factors: A Culturally Common Structure of Parenting Infants
A higher-order two-factor structure of mothers’ parenting infants was identified, factors that can be thought of as conceptually separable and developmentally significant. The Dyadic factor in parenting infants encompasses interactions that focus on nurturing, physical, socioemotional, educational, and verbal exchanges between mother and baby. Achieving well-being – healthy physical development and wholesome emotional and social development by forging close relationships with other people (mainly parents) -- constitutes a central task of infancy, and parent-infant interaction is integral to success at that task (Ainsworth, 1977; Bowlby, 1982; Emde, 2013). The Extradyadic factor encompasses interactions that turn outward from the dyad to include foci on didactic exchanges and the material environment. Being introduced to, coming to understand, and beginning to negotiate the world outside the dyad are similarly central tasks for young infants. (It is likely that the Extradyadic factor does not include the language domain per se because coding didactics included mothers’ speech to the child; see Appendix A.) These two parenting factors therefore reflect independent, complementary, and elemental forces in caregiving human infants that support the infant’s development in two realms central to their early competencies, and the two factors characterized emphases mothers pursue in everyday rearing infants in different regions around the globe.
These twin complementary orientations, here documented empirically, accord with and confirm extant theorizing about parenting infants and young children. Many authorities have previously intuited, asserted, or derived one or both of these higher-order factors underlying parent-infant interaction qua social and didactic, animate and inanimate, affective and informational, primary or secondary intersubjective, and relationship-centered and environment-oriented (Ainsworth et al., 1978; Bornstein, 2019; Hastings & Grusec, 1998; Penman et al., 1983; Trevarthen & Aiken, 2001; Worthman et al., 2002). For example, a four-country study (Argentina, France, Japan, and the United States) analyzed mothers’ speech to infants at two ages in the first year of life and found that mothers use predominantly affect- and information-laden speech (Bornstein et al., 1992). Notably, the two-factor structure identified in the two studies presented here maps onto the twin behavior control systems (affiliative and exploratory) prominently associated with attachment (Bowlby, 1969) and primary and secondary intersubjectivity (Trevarthen & Aitken, 2001; Trevarthen & Hubley, 1978). This dual system also accords with evolutionary psychology which supposes that human adaptations display domain-specific modularity, as the Dyadic and Extradyadic functions do, and that each adaptation should be designed to accomplish a task that, given a natural developmental environment, will improve the individual’s survival (Tooby & Cosmides, 1992).
Two widely reported predictors of wholesome child development appear to be relationships with a caring prosocial caregiver and good intellectual functioning (Masten & Palmer, 2019). Competence ensues when fundamental systems foster child development or counteract threats to development. Parental warmth and emotional support (a dyadic focus) appear to be important for children’s future socioemotional competencies, and parental cognitive expectations and structures (an extradyadic focus) appear to be important for children’s future mental competencies. Enhancing socioemotional and mental competencies in children can have long-term cumulative protective effects. The dyadic-extradyadic balance in the infant-mother relationship therefore likely has vital and far-reaching consequences in the life of the child. To the extent that a mother effectively supports and promotes both affiliative and exploratory goals for her infant, her infant’s chances to develop both socioemotional and mental adaptive competencies are improved.
In brief, the dynamic of mother-infant dyadic-extradyadic interactions may have formative implications for understanding and predicting meaningful and enduring aspects of human development. The two systems are equally vital to wholesome growth and appear to work in tandem to foster children’s socioemotional relationships and mental capabilities. These proclivities are present from a surprisingly early point in the development of the mother-infant relationship and across an equally surprising diversity of cultural contexts. Together, the findings point to a basic reality of early caregiving.
Six Parenting Domains: The Modular and Specific Nature of Parenting
A two-factor/six-domain structure to parenting infants was identified. The Dyadic and Extradyadic factors accounted for significant amounts of variance in maternal caregiving. However, six domains of parenting -- nurturant, physical, social, didactic, material, and language -- each accounted for significant amounts of variance that were unaccounted for by the two higher-order factors. The residual variances associated with each of these domains mean that parents do not necessarily behave in uniform or consistent ways across these different domains of interaction. This differentiated and nuanced domain structure allows parents to emphasize one or another specific domian of parenting at one or another time with one or another child and in one or another context or culture, permitting cultures to prescribe and proscribe selected parenting practices.
The six-domain finding also accords with theories of modularity and specificity common to descriptions of cognitions and practices in evolutionary psychology and artificial intelligence (Bjorklund & Myers, 2019; Tooby & Cosmides, 1992). The principle of modular design states that any large computation can be split into a collection of smaller, nearly independent, specialized subprocesses or modules. Modularity is a fundamental property of living things at every level of organization, and proponents of modularity have offered several reasons for expecting biological, social, and mental processes to consist of multiple specialized systems. Domain specificity refers to the idea that a given system accepts or is specialized to operate on or process only specific classes of information. For example, a larger number of functionally specialized mechanisms are likely to perform more effectively and efficiently than is a smaller number of systems with more general functions. A virtue of such modular encapsulation is that it enables input systems to accomplish tasks quickly, and a system that needs to respond only to one specific class of stimulus can be tuned to respond with special efficiency.
The six-domain feature of the parenting structure also opens the possibility that individuals and cultures can vary in expressing their caregiving. Despite fundamentally similar childcare routines mothers around the world engage in with infants, and despite the fact that all mothers in this study engaged in all six parenting domains, clear between-cultural differences have been found across all domains (Bornstein et al., 2017). The cross-cultural invariant structure of parenting is independent of cross-cultural mean-level differences in parenting.
Limitations of The Two Studies
In terms of representativeness, these studies focused on maternal parenting practices in nonclinical community samples with typically developing firstborn infants of one specific age. Different patterns of results could emerge for mothers with multiple children, mothers with children of different ages or with special needs, single or divorced mothers, or mothers from clinical populations. The present studies focused on maternal parenting: The primacy of mothers in childrearing does not deny or minimize the considerable (and possibly increasing) contributions to infant care made by fathers and other caregivers both inside and outside of the family. Studies of fathers or other nonparental caregivers might yield similar or different parenting structures and so merit study. In these analyses maternal age, education, and SES were taken into account, but taken singly could moderate the structure of parenting infants. Samples from 9 different countries were recruited, including two in South America, one in North America, three in Europe, one in the Middle East, one in Africa, and one in East Asia. The two-factor/six-domain structure of mothering held in the United States as well as across 8 other cultures. Although these samples reflect a broad range of cultural styles, no pretentions are made that they are representative of either their nations or the world’s cultures. In addition, a principal component of the bioecological model of development (as formalized in Bronfenbrenner & Morris, 2006) includes time; the present investigation is cross-sectional, and the same factors might have similar or different implications depending on different temporal factors (e.g., age of child or historical period). All these considerations naturally constrain the generalizability of the parenting practices structure reported here.
Conclusions
These studies suggest that maternal parenting of infants is organized in a hybrid two-factor/six-domain structure, consistent with notions of cultural commonality as well as specificity and modularity in parenting (Bornstein, 2015, 2022). Parenting is bifactorial in the sense that parenting infants fits a factor structure characterized by overarching dyadic and extradyadic orientations, and parenting is domain specific in the sense that six domains instantiate the two parenting factors. Parenting is specific and modular in the senses that the two factors and six domains address separate parenting imperatives and infant needs, and factors and domains alike may manifest in idiosyncratic ways in individuals and cultures.
IMPLICATIONS FOR THEORY AND PRACTICE
Studies of parenting and mother-infant interactions have been undertaken at microanalytic and macroanalytic levels. Here, maternal parenting was assessed at a microanalytic level, focusing on operationally defined individual practices. However, dimensions of parenting can also be operationalized at a macroanalytic level, that is in terms of omnibus constructs like sensitivity. A structure organizing microanalytic maternal parenting practices emerged. By assessing these foundational in-the-moment building blocks of parenting, how they are structured, and how they manifest across cultures, these studies provide insights for future students of parenting theory and culture. This close analysis of everyday behaviors also has relevance for cognitive science, as such naturalistic tasks as basic caregiving require the coordination of multiple cognitive faculties (Newel, 1973). As Schultheis and Cooper (2022, p. 7) have observed, “everyday activities draw upon both precompiled or cached routines … and systems for general problem solving, including those involved in perception, motor and action planning, spatial reasoning, and the generation and maintenance of intentions.”
From a cross-cultural point of view, especially notable in these results is the emergence of a common two-factor/six-domain model of parenting infants. What might account for any widespread commonality in parenting? Actually, possible commonality is supported by arguments that derive from several quarters. Certain parenting practices could recur across even very different cultures on account of shared determinants endemic to evolution, to biology, to social history, or to children. First, some characteristics of parenting may be instinctual to a “parenting stage” in the human life cycle. On this argument, it is in the nature of being a human—as much as a parent—to optimize the success of one’s offspring and thereby to ensure the survival of one’s genes (Bjorklund, 2020; Dawkins, 2016). Maternal hormones and the maternal nervous system seem to have evolved to treat and respond to human infants in some uniform ways (Bornstein et al., 2017; Feldman, 2019; Stark et al., 2019). Basic physiology is mobilized to support parenting, and parenting might initially and naturally arise out of biological processes associated with pregnancy and parturition. “Intuitive parenting” involves inherent responses that are developmentally suited to the age and abilities of the child and have the goal of enhancing adaptation and optimizing development (Papoušek & Papoušek, 2002). Parents regularly enact such intuitive parenting programs in an unconscious fashion—that is, such programs do not require the time and effort typical of conscious decision making, and, being more rapid and efficient, they utilize less attentional and cognitive reserve. Second, shared environments may shape parents to think and act in similar ways (Harris, 1968). That is, certain economic or ecological factors are common even to different cultures on account of worldwide historically converging and homogenizing patterns of modernization, urbanization, Westernization, migration, or dissemination via media, and thereby shape a degree of uniformity in parenting infants. Third, by virtue of their helplessness or “babyish” characteristics, which are of course structurally universal, infants may trigger a parental care motivational system (Schaller, 2018) and so elicit a generic organization to their caregiving (Bell, 1968, 1979; Bell & Chapman, 1986; Kringelbach et al., 2016; Lorenz, 1935/1970). Separately or together, these several evolutionary, biological, historical, environmental, and interpersonal forces likely engender some similarities in parenting infants. It would be challenging, if not impossible, to determine the contributions of each. At the end of the day, different people (presumably) wish to socialize and educate their offspring, and they may do so in some manifestly similar ways.
The common two-factor/six-domain model of parenting infants unearthed in these studies speak to patterns of covariation that structure parenting practices. At the same time, these results do not address mean-level differences in parenting or culture-specific parenting that must be understood in the variances that are unique and unaccounted for by the two-factor/six-domain model. Cultural models of parenting include goals and ideals for child development, and the cognitions and practices that parents deploy to achieve those goals and ideals. Neither parenting nor infant development occurs in a vacuum. Both germinate and grow in the medium of culture. Presumably, specific patterns of childrearing -- within and between cultures -- are adapted to each specific culture’s settings and needs. Every culture promotes idiosyncratic ways of adapting to the exigencies and stringencies of its unique ecology, environment, and history and has developed parenting ethnotheories and mazeways to achieve its specific childrearing goals. A possible direction for future research would be to examine within- and between-culture mean-level variation in the parenting factors and domains identified here.
These studies aimed to learn more about human parenting during the significant period of the dyad’s initial mutual accommodation in the first half-year of the infant’s life by examining specific observable parenting practices of mothers and by recording, coding, analyzing, and comparing them in a large and diverse sample. From birth, human infants express and begin to read basic human emotions, develop individual personalities and social styles, form interpersonal bonds, and explore and attempt to understand objects, properties, and events in the world. Human parents guide their infants through all these dramatic firsts. Parenting an infant does not fix the trajectory or outcome of development, of course, but it makes sense that effects have causes and that the start exerts an impact on the end. Parenting is central to human development, to the continuity of culture, and to humanity’s long-term investment in the young. Two further practical implications of this work would be, first, to assess associations of these parenting factors and domains with infant and child development and, second, to examine possible variations in these parenting factors and domains in diverse at-risk populations of mothers.
Supplementary Material
Acknowledgements
We thank H. Azuma, S. Bali, L. Cote, A. De Houwer, C. Galperin, C.-S. Hahn, O. M. Haynes, C. Hendricks, M. Kabiru, M. Lim, S. Maital, M.-L. Moura de Seidl, M. Ogino, M.-G. Pêcheux, J.T.D. Suwalsky, P. Venuti, and A. Vyt. Our collaborators in each country are senior professionals and scholars who took active parts in the recruitment and conduct of this work. The ideas and opinions expressed herein are those of the authors alone, and endorsement by the authors’ Institutions is not intended and should not be inferred.
Funding
MHB, DLP, and the research were supported by the Intramural Research Program of the NIH/NICHD, USA; MHB by an International Research Fellowship at the Institute for Fiscal Studies, London, UK, funded by the European Research Council under the Horizon 2020 research and innovation programme (grant agreement No 695300-HKADeC-ERC-2015-AdG); and GE by Nanyang Technological University, NAP SUG 2015–2021 Grant.
Role of the Funders
None of the funders of this research had any role in the design and conduct of the study; collection, management, analysis, and interpretation of data; preparation, review, or approval of the manuscript; or decision to submit the manuscript for publication.
Appendix A. Mother parenting practice domains, definitions of constituent parenting practices, interim variables, final indicator variables, and mean standard score of final variables.
| Domain | Definition | Interim Variable(s) | Final Indicator Variable | Mean Standard Score of Final Variables |
|---|---|---|---|---|
| Nurture | Feed: Attempting to give the infant liquid or solid foods by cup, bottle, breast, or spoon | Duration of feed | Sum of durations of feed and burp/wipe | Mean of standard scores of duration of feed/burp/wipe, duration of bathe/diaper/dress, groom/other, and duration of hold |
| Burp/Wipe face or hands: Attempting to burp the infant in connection with a feeding or wiping the infant’s face, hands, or clothing at any time | Duration of burp/wipe | |||
| Bathe: Washing and drying the infant’s body and/or hair | Duration of bathe | Sum of durations of bathe, check/change diaper, dress, groom, and meet other health needs | ||
| Check/Change diaper: Checking if the infant needs a diaper change or changing the diaper | Duration of check/change diaper | |||
| Dress: Removing or putting an article of clothing on the infant | Duration of dress | |||
| Groom: Engaging in a practice designed to enhance the infant’s appearance (e.g., combs hair) | Duration of groom | |||
| Meet other health needs: Attending to other health needs of the infant (e.g., wipes or suctions the infant’s nose; gives medicine from a dropper or medicine spoon) | Duration of meet other health needs | |||
| Hold: Supporting some or all of the infant’s weight with the mother’s body | Duration of hold | Duration of hold | ||
| Physical | Physically encourage to sit: Placing the infant in a sitting position in which the infant’s back is not leaning against a firm surface | Proportion of consecutive 10-min time units in which physical encouragement to sit was observed | Mean proportion of consecutive 10-min time units in which physical encouragement to balance (sit or stand) was observed | Mean of mean proportions of encourage balance and encourage movement |
| Physically encourage to stand: Placing or holding the infant in a standing position so that there is some weight supported by the infant’s straightened legs | Proportion of consecutive 10-min time units in which physical encouragement to stand was observed | |||
| Physically encourage to roll: Physically assisting the infant to roll over | Proportion of consecutive 10-min time units in which physical encouragement to roll was observed | Mean proportion of consecutive 10-min time units in which physical encouragement to move (roll, crawl, or step) was observed | ||
| Physically encourage to crawl: Physically assisting the infant to move forward (on the belly or on hands and knees) by moving the infant’s arms and/or legs or by pushing rump or feet from behind | Proportion of consecutive 10-min time units in which physical encouragement to crawl was observed | |||
| Physically encourage to step: Holding the infant in a standing position and then moving the infant’s body to simulate stepping movements | Proportion of consecutive 10-min time units in which physical encouragement to step was observed | |||
| Social | Encourage attention to mother: Attempting to draw the infant into face-to-face social interaction with herself physically by intentionally moving her face toward the infant or moving the infant toward her face or verbally by making very specific comments about herself that are clearly designed to capture the infant’s interest (pauses of 2 ss or longer are coded as terminations of an ongoing practice) | Frequency of encouraging attention to mother Duration of encouraging attention to mother |
Mean standard score of frequency and duration of encourage attention to mother | Mean of standard scores of encourage attention to mother, social play, and express affection |
| Social play: Directing verbal or physical behaviors to the infant, the purpose of which appears to amuse the infant (i.e., to elicit smiles, positive vocalizations, laughter, or motoric excitement in the context of a primarily social dyadic interaction). Coding is discontinued when the mother has not interacted for 3 s and when she is no longer oriented to the infant and poised to continue the exchange; pauses of any duration when the mother clearly remains poised to continue are coded as part of the social play sequence. The types of exchanges coded as social play are: (a) physical contact with a fun-like quality (e.g., tickling); (b) introducing the element of surprise, suspense, or quick release of stimuli (e.g., peek-a-boo); (c) singing to the infant; and (d) playing a game that involves physical manipulation of the infant’s body (e.g., pattycake) | Frequency of social play Duration of social play |
Mean standard score of frequency and duration of social play | ||
| Express affection. Expressing affection or positive evaluation to the infant either physically (e.g., kissing, patting, stroking, or caressing) or verbally (using explicit phrases denoting praise or endearment) | Frequency of expressing affection Duration of expressing affection | Mean standard score of frequency and duration of expressing affection | ||
| Didactic | Encourage attention to object: Physically moving the infant or an object so that the infant can see or touch it or verbally referring to an object or an object-related event or activity that is no more than 12 feet from the infant (pauses of 2 s or longer are coded as terminations of an ongoing practice) | Frequency of encouraging attention to object Duration of encouraging attention to object |
Mean standard score of frequency and duration of encouraging attention to object | Mean standard score for encourage attention to object |
| Material | Quantity of objects provided infant: The number of toys, books, and household objects that are within the infant’s reach | Variety of objects provided: The number of different objects that are within infant reach during the total observation Density of objects provided: The mean number of objects within infant reach per consecutive 5-min time unit Consistency of objects provided: The number of consecutive 5-min time units in which any object was within infant reach |
Mean standard score of the quantity (variety, density, and consistency) of objects provided | Mean of standard score for quantity of objects |
| Quality (responsiveness) of objects provided infant: Ratings of all toys, books, and household objects within reach of infant on four dimensions: moving parts, change in shape or contour, noise production, and reflected image | Responsiveness of objects: Mean of sums of ratings for moving parts, change in shape or contour, noise production, and reflected image for each toy, book, and household object within infant reach Number of highly responsive objects: Number of toys, books, and household objects within infant reach that has a sum of ratings for moving parts, change in shape or contour, noise production, and reflected image >12 (on a scale of 4–16) Proportion of highly responsive objects: Proportion of toys, books, and household objects within infant reach that has a sum of ratings for moving parts, change in shape or contour, noise production, and reflected image >12 (on a scale of 4–16) |
Mean standard score of quality (responsiveness, number of highly responsive, and proportion of highly responsive) of objects provided | Mean of standard score for quality of objects | |
| Language | Speech to the infant: Mother’s speech steam directed to the infant in either adult- or infant-directed tones, including vocalizations of syllables, parts of words, words, and singing, was coded continuously. Changes and pauses lasting less than 1 s were not recorded, and vocalizations between pauses lasting more than 3 s were recorded as different | Frequency of vocalizations Duration of vocalizations |
Mean standard scores of frequency and duration of speech to the infant | Mean standard score for speech to the infant |
Footnotes
Conflict of Interest Disclosures
Each author signed a form for disclosure of potential conflicts of interest. No authors reported any financial or other conflicts of interest in relation to the work described.
Ethical Principles
The authors affirm having followed professional ethical guidelines in preparing this work. These guidelines include obtaining informed consent from human participants, maintaining ethical treatment and respect for the rights of human or animal participants, and ensuring the privacy of participants and their data, such as ensuring that individual participants cannot be identified in reported results or from publicly available original or archival data.
In this article, we refer to nations as the political entities where samples were recruited, but do not claim these samples are nationally representative. Every nation harbors different cultures. We refer to cultures as the entities in nations from which we recruited samples and to which our results are generalizable.
Contributor Information
Marc H. Bornstein, Eunice Kennedy Shriver National Institute of Child Health and Human Development, National Institutes of Health, 8404 Irvington Ave., Bethesda MD 20817, U.S.A.
Diane L. Putnick, Eunice Kennedy Shriver National Institute of Child Health and Human Development
Gianluca Esposito, University of Trento..
REFERENCES
- Adams J, & Weakliem DL (2011). August B. Hollingshead’s “Four Factor Index of Social Status”: From unpublished paper to citation classic. Yale Journal of Sociology, 8, 11–19. [Google Scholar]
- Ainsworth MDS (1977). Infant development and mother-infant interaction among Ganda and American families. In Leiderman PH, Tulkin SR, & Rosenfeld AH (Eds.), Culture and infancy (pp. 119–150). Academic Press. [Google Scholar]
- Ainsworth MDS, Blehar MC, Waters E, & Wall S (1978). Patterns of attachment: A psychological study of the Strange Situation. Erlbaum. [Google Scholar]
- Akaike H (1987). Factor analysis and AIC. Psychometrika, 52, 317–332. 10.1007/BF02294359 [DOI] [Google Scholar]
- Alhattab S (2021). New year’s babies: Over 370,000 children will be born worldwide on New Year’s Day – UNICEF. UNICEF. https://www.unicef.org/press-releases/new-years-babies-over-370000-children-will-be-born-worldwide-new-years-day-unicef [Google Scholar]
- American Community Survey. (2018). 2016 American Community Survey 1-year estimates. U.S. Census Fact Finder. [Google Scholar]
- Bakeman R, & Gottman JM (1986). Observing interaction: An introduction to sequential analysis. Cambridge University Press. 10.1017/CBO9780511527685 [DOI] [Google Scholar]
- Bell RQ (1968). A reinterpretation of the direction of effects in studies of socialization. Psychological Review, 75, 81–85. 10.1037/h0025583 [DOI] [PubMed] [Google Scholar]
- Bell RQ (1979). Parent, child, and reciprocal influences. American Psychologist, 34(10), 821–826. 10.1037/0003-066X.34.10.821 [DOI] [Google Scholar]
- Bell RQ, & Chapman M (1986). Child effects in studies using experimental or brief longitudinal approaches to socialization. Developmental Psychology, 22(5), 595–603. 10.1037/0012-1649.22.5.595 [DOI] [Google Scholar]
- Bentler PM (1990). Comparative fit indexes in structural models. Psychological Bulletin, 107(2), 238–246. 10.1037/0033-2909.107.2.238 [DOI] [PubMed] [Google Scholar]
- Bentler PM (2008). EQS 6 Structural Equations Program Manual. Multivariate Software (www.mvsoft.com). [Google Scholar]
- Bentler PM, & Weeks DG (1980). Linear structural equations with latent variables. Psychometrika, 45, 289–308. 10.1007/BF02293905 [DOI] [Google Scholar]
- Bjorklund DF (2020). Child development in evolutionary perspective. In Bornstein MH (Series Editor), Cambridge elements in child development. Cambridge University Press. doi: 10.1017/9781108866187 [DOI] [Google Scholar]
- Bjorklund DF, & Myers AJ (2019). The evolution of parenting and evolutionary approaches to childrearing. In Bornstein MH (Ed.), Handbook of parenting. Vol. 2. Biology and ecology of parenting (3rd ed., pp. 3–29). Routledge. 10.4324/9780429401459-1 [DOI] [Google Scholar]
- Bögels SM, & Perotti EC (2011). Does father know best? A formal model of the paternal influence on childhood social anxiety. Journal of Child and Family Studies, 20(2), 171–181. 10.1007/s10826-010-9441-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Bollen KA (1987). Outliers and improper solutions: A confirmatory factor analysis example. Sociological Methods and Research, 15, 375–384. 10.1177/0049124187015004002 [DOI] [Google Scholar]
- Bollen KA (1989). Structural equations with latent variables. John Wiley and Sons. 10.1002/9781118619179 [DOI] [Google Scholar]
- Bornstein MH (2002). Measurement variability in infant and maternal behavioral assessment. Infant Behavior and Development, 25, 413–432. 10.1016/s0163-6383(02)00143-1 [DOI] [Google Scholar]
- Bornstein MH (2015). Children’s parents. In Bornstein MH & Leventhal T (Eds.), Ecological settings and processes in developmental systems. Vol. 4. Handbook of child psychology and developmental science (pp. 1–78). Wiley. 10.1002/9781118963418.childpsy403 [DOI] [Google Scholar]
- Bornstein MH (2019). Parenting infants. In Bornstein MH (Ed.), Handbook of parenting. Vol. 1. Children and parenting (3rd ed., pp. 3–55). Routledge. 10.4324/9780429440847-1 [DOI] [Google Scholar]
- Bornstein MH (2022). Parenting, infancy, culture: Specificity and commonality in Argentina, Belgium, Israel, Italy, and the United States. Routledge. https://www.routledge.com/Parenting-Infancy-Culture-Specificity-in-Argentina-Belgium-Israel/Bornstein/p/book/9780367424671 [Google Scholar]
- Bornstein MH, Jager J, & Putnick DL (2013). Sampling in developmental science: Situations, shortcomings, solutions, and standards. Developmental Review, 33, 357–370. 10.1016/j.dr.2013.08.003 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Bornstein MH, Putnick DL, Park Y, Suwalsky JTD, & Haynes OM (2017). Human infancy and parenting in global perspective: Specificity. Proceedings of the Royal Society B: Biological Sciences, 284(1869), Article 20172168. 10.1098/rspb.2017.2168 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Bornstein MH, Putnick DL, Rigo P, Esposito G, Swain JE, Suwalsky JTD, Su X, Du X, Cote LR, De Pisapia N, & Venuti P (2017). The neurobiology of culturally common maternal responses to infant cry. Proceedings of the National Academy of Sciences, 114(45), E9465–E9473. Doi: 10.1073/pnas.1712022114 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Bornstein MH, Tamis-LeMonda CS, Tal J, Ludemann P, Toda S, Rahn C, Pécheux M-G, Azuma H, & Verdi D (1992). Maternal responsiveness to infants in three societies: The United States, France, and Japan. Child Development, 63, 808–821. 10.2307/1131235 [DOI] [PubMed] [Google Scholar]
- Bowlby J (1969). Attachment and loss: Vol. 1: Attachment Hogarth. [Google Scholar]
- Bowlby J (1982). Attachment and loss: retrospect and prospect. American Journal of Orthopsychiatry, 52(4), 664–678. 10.1111/j.1939-0025.1982.tb01456.x [DOI] [PubMed] [Google Scholar]
- Bradley RH (2004). Chaos, culture, and covariance structures: A dynamic systems view of children’s experiences at home. Parenting: Science and Practice, 4, 243–257. 10.1080/15295192.2004.9681272 [DOI] [Google Scholar]
- Bronfenbrenner U, & Morris P (2006). The bioecological model of human development. In Damon W & Lerner RM (Series Eds.) & Lerner RM (Vol. Ed.), Handbook of child psychology: Vol. 1. Theoretical models of human development (6th ed., pp.793–828). John Wiley. 10.1002/9780470147658.chpsy0114 [DOI] [Google Scholar]
- Browne MW, & Cudeck R (1993). Alternative ways of assessing model fit. In Bollen KA & Long JS (Eds.), Testing structural equation models (pp. 136–162). SAGE. [Google Scholar]
- Byrne BM, Shavelson RJ, & Muthén B (1989). Testing for the equivalence of factor covariance and mean structures: The issue of partial measurement invariance. Psychological Bulletin, 105, 456–466. 10.1037/0033-2909.105.3.456 [DOI] [Google Scholar]
- Campbell JA (1986). Scientific revolution and the grammar of culture: The case of Darwin’s origin. Quarterly Journal of Speech, 72(4), 351–376. 10.1080/00335638609383782 [DOI] [Google Scholar]
- Cicchetti D, Bronen R, Spencer S, Haut S, Berg A, Oliver P, & Tyrer P (2006). Rating scales, scales of measurement, issues of reliability: Resolving some critical issues for clinicians and researchers. Journal of Nervous and Mental Disease, 194, 557–564. 10.1097/01.nmd.0000230392.83607.c5 [DOI] [PubMed] [Google Scholar]
- Cohen J (1960). A coefficient of agreement for nominal scales. Education and Psychological Measurement, 20, 37–46. 10.1177/001316446002000104 [DOI] [Google Scholar]
- Cohen J (1968). Weighted kappa: Nominal scale agreement with provision for scaled disagreement or partial credit. Psychological Bulletin, 70, 213–219. 10.1037/h0026256 [DOI] [PubMed] [Google Scholar]
- Crockenberg S, & Litman C (1990). Autonomy as competence in 2-year-olds: Maternal correlates of child defiance, compliance, and self-assertion. Developmental Psychology, 26(6), 961–971. 10.1037/0012-1649.26.6.961 [DOI] [Google Scholar]
- Dawkins R (2016). The selfish gene. Oxford University Press. [Google Scholar]
- Dishion TJ, Mun CJ, Tein JY, Kim H, Shaw DS, Gardner F, & Peterson J (2017). The validation of macro and micro observations of parent–child dynamics using the relationship affect coding system in early childhood. Prevention Science, 18, 268–280. 10.1007/s11121-016-0697-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Emde R (Ed.). (2013). The development of attachment and affiliative systems. Springer Science & Business Media. [Google Scholar]
- Erting CJ, Prezioso C, and Hynes MO (1994). The interactional context of deaf mother—infant communication. In Volterra V & Erting CJ (Eds.), From gesture to language in hearing and deaf children (pp. 97–106). Gallaudet University Press. 10.1007/978-3-642-74859-2_9 [DOI] [Google Scholar]
- Feldman R (2019). The social neuroendocrinology of human parenting. In Bornstein MH (Ed.), Handbook of parenting: Biology and ecology of parenting (Vol. 2, 3rd ed., pp. 220–249). Routledge. www.taylorfrancis.com/books/e/9780429401459/chapters/10.4324/9780429401459-6 [Google Scholar]
- Feldman R, & Eidelman AI (2007). Maternal postpartum behavior and the emergence of infant–mother and infant–father synchrony in preterm and full-term infants: The role of neonatal vagal tone. Developmental Psychobiology, 49(3), 290–302. 10.1002/dev.20220 [DOI] [PubMed] [Google Scholar]
- Floyd FJ, & Widaman KF (1995). Factor analysis in the development and refinement of clinical assessment instruments. Psychological Assessment, 7(3), 286–299. 10.1037/1040-3590.7.3.286 [DOI] [Google Scholar]
- Fox J (1997). Applied regression analysis, linear models and related methods. Sage. [Google Scholar]
- Galligan RP, Beebe B, Milne D, Ewing J, Lee SH, & Buck KA (2018). Maternal into-the-face behavior, shared attention, and infant distress during face-to-face play at 12 months: Bi-directional contingencies. Infancy, 23, 538–557. 10.1111/infa.12234 [DOI] [Google Scholar]
- Geary DC (2000). Evolution and proximate expression of human paternal investment. Psychological Bulletin, 126(1), 55–77. 10.1037/0033-2909.126.1.55 [DOI] [PubMed] [Google Scholar]
- Golinkoff RM, Can DD, Soderström M, & Hirsh-Pasek K (2015). (Baby)talk to me: The social context of infant-directed speech and its effects on early language acquisition. Current Directions in Psychological Science, 24, 339–344. 10.1177/0963721415595345 [DOI] [Google Scholar]
- Gould SJ (1977). Ontogeny and phylogeny. Harvard University Press. [Google Scholar]
- Hallgren KA (2012). Computing inter-rater reliability for observational data: An overview and tutorial. Tutorials in Quantitative Methods for Psychology, 8, 23–34. 10.20982/tqmp.08.1p023 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Harkness S, & Super CM (1983). The cultural construction of child development: A framework for the socialization of affect. Ethos, 11, 221–231. 10.1525/eth.1983.11.4.02a00020 [DOI] [Google Scholar]
- Harris M (1968). The rise of anthropological theory: A history of theories of culture. Thomas Y. Crowell. [Google Scholar]
- Hartmann DP (1977). Considerations in the choice of interobserver reliability estimates. Journal of Applied Behavior Analysis, 10(1), 103–116. 10.1901/jaba.1977.10-103 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Hartmann DP (1982). Assessing the dependability of observational data. New Directions for Methodology of Social & Behavioral Science, 14, 51–65. [Google Scholar]
- Hartmann DP, Abbott CB, & Pelzel KE (2015). Design, measurement, and analysis in developmental research. In Bornstein MH & Lamb ME (Eds.), Developmental science: An advanced textbook (7th ed., pp. 113–214). Psychology Press. [Google Scholar]
- Hastings PD, & Grusec JE (1998). Parenting goals as organizers of responses to parent-child disagreement. Developmental Psychology, 34(3), 465–479. 10.1037/0012-1649.34.3.465 [DOI] [PubMed] [Google Scholar]
- Holden GW, & Miller PC (1999). Enduring and different: A meta-analysis of the similarity in parents’ childrearing. Psychological Bulletin, 125, 223–254. 10.1037/0033-2909.125.2.223 [DOI] [PubMed] [Google Scholar]
- Hollenbeck AR (1978). Problems of reliability in observational research. In Sackett GP (Ed.), Observing behavior. Vol. 2. Data collection and analysis methods (pp. 79–98). University Park Press. [Google Scholar]
- Hollingshead AB (1975). The four-factor index of social status [Unpublished manuscript]. Yale University. [Google Scholar]
- Hollingshead AB (2011). Four factor index of social status. Yale Journal of Sociology, 8, 21–51. [Google Scholar]
- Hu L-T, & Bentler PM (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling, 6, 1–55. 10.1080/10705519909540118 [DOI] [Google Scholar]
- Jackson DN (1976). Jackson personality inventory manual. Research Psychologists Press. [Google Scholar]
- Jager J, Putnick DL, & Bornstein MH (2017). More than just convenient: The scientific merits of homogeneous convenience samples. Monographs of the Society for Research in Child Development, 82, 13–30. 10.1111/mono.12296 [DOI] [PMC free article] [PubMed] [Google Scholar]
- James DM, Wadnerkar MB, Lam-Cassettari C, Kang S, & Telling AL (2012). Thin slice sampling of video footage for mother/child interaction: Application to single cases. Journal of Psychopathology and Behavioral Assessment, 34(3), 351–360. 10.1007/s10862-012-9282-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Jöreskog KG, & Sörbom D (1984). LISREL VI: Analysis of linear structural relationships by the method of maximum likelihood. Scientific Software, Inc. [Google Scholar]
- Kaplan D (2000). Structural equation modeling: Foundations and extension. SAGE. [Google Scholar]
- Kauffeld S, Lehmann-Willenbrock N, & Meinecke AL (2018). The advanced interaction analysis for teams (act4teams) coding scheme. In Brauner M, Boos M, & Kolbe M (Eds.), The Cambridge handbook of group interaction analysis (pp. 422–431). Cambridge University Press. 10.1017/9781316286302.022 [DOI] [Google Scholar]
- Konner M (2010). The evolution of childhood: Relationships, emotions, mind. Belknap Press. [Google Scholar]
- Koulomzin M, Beebe B, Anderson S, Jaffe J, Feldstein S, & Crown C (2002). Infant gaze, head, face and self-touch at 4 months differentiate secure vs. avoidant attachment at 1 year: A microanalytic approach. Attachment & Human Development, 4(1), 3–24. 10.1080/14616730210123120 [DOI] [PubMed] [Google Scholar]
- Kringelbach ML, Stark EA, Alexander C, Bornstein MH, & Stein A (2016). On cuteness: Unlocking the parental brain and beyond. Trends in Cognitive Sciences, 20(7), 545–558. 10.1016/j.tics.2016.05.003 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Leiderman PH, Tulkin SR, & Rosenfeld A (Eds.). (1977). Culture and infancy: Variations in the human experience. Academic Press. [Google Scholar]
- Lerner RM (2018). Concepts and theories of human development (4th ed.). Routledge. 10.4324/9780203581629 [DOI] [Google Scholar]
- Lorenz K (1935/1970). Studies in animal and human behavior. Methuen. 10.4159/harvard.9780674430389 [DOI] [Google Scholar]
- Lupia A (2018). How to improve coding for open-ended survey data: Lessons from the ANES. In Vannette D & Krosnick J (Eds.), The Palgrave handbook of survey research. Palgrave Macmillan. 10.1007/978-3-319-54395-6_16 [DOI] [Google Scholar]
- MacCallum RC, Roznowski M, & Necowitz LB (1992). Model modifications in covariance structure analysis: The problem of capitalization on chance. Psychological Bulletin, 111, 490–504. 10.1037/0033-2909.111.3.490 [DOI] [PubMed] [Google Scholar]
- Mardia KV (1970). Measures of multivariate skewness and kurtosis with applications. Biometrika, 57, 519–530. 10.1093/biomet/57.3.519 [DOI] [Google Scholar]
- Masten AS, & Palmer AR (2019). Parenting to promote resilience in children. In Bornstein MH (Ed.), Handbook of parenting. Vol. 5. The practice of parenting (3rd ed., pp. 156–188). Routledge. 10.4324/9780429401695-6 [DOI] [Google Scholar]
- McCune-Nicolich L, & Fenson L (1984). Methodological issues in studying early pretend play. In Yawkey TD & Pelligrini AD (Eds.), Child’s play: Developmental and applied (pp. 81–124). Erlbaum. [Google Scholar]
- McGraw KO, & Wong SP (1996a). Forming inferences about some intraclass correlation coefficients. Psychological Methods, 1, 30–46. 10.1037/1082-989X.1.1.30 [DOI] [Google Scholar]
- McGraw KO, & Wong SP (1996b). “Forming inferences about some intraclass correlations coefficients”: Correction. Psychological Methods, 1(4), 390. 10.1037/1082-989X.1.4.390 [DOI] [Google Scholar]
- Miller PC, Shim JE, & Holden GW (1998). Immediate contextual influences on maternal behavior: Environmental affordances and demands. Journal of Environmental Psychology, 18(4), 387–398. 10.1006/jevp.1998.0111 [DOI] [Google Scholar]
- Moller AP (2003). The evolution of monogamy: Mating relationships, parental care, and sexual selection. In Boesch C (Ed.), Monogamy: Mating strategies and partnerships in birds, humans, and other mammals (pp. 29–41). Cambridge University Press. [Google Scholar]
- Negraia DV, Augustine JM, & Prickett KC (2018). Gender disparities in parenting time across activities, child ages, and educational groups. Journal of Family Issues, 39(11), 3006–3028. [DOI] [Google Scholar]
- Newell A (1973). You can’t play 20 questions with nature and win: Projective comments on the papers of this \symposium. In Chase WG (Ed.), Visual information processing. Academic Press. 10.1016/B978-0-12-170150-5.50012-3 [DOI] [Google Scholar]
- Northrup JB, & Iverson JM (2020). The development of mother–infant coordination across the first year of life. Developmental Psychology, 56(2), 221–236. 10.1037/dev0000867 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Papaligoura Z, & Trevarthen C (2001). Mother–infant communication can be enhanced after conception by in-vitro fertilization. Infant Mental Health Journal, 22, 591–610. 10.1002/imhj.1021 [DOI] [Google Scholar]
- Papoušek H, & Papoušek M (2002). Intuitive parenting. In Bornstein MH (Ed.), Handbook of parenting Vol. 2 Biology and ecology of parenting (2nd ed., pp. 183–203). Lawrence Erlbaum Associates. [Google Scholar]
- Parke RD, & Cookston JT (2019). Fathers and families. In Bornstein MH (Ed.), Handbook of parenting Vol. 3: Being and becoming a parent (3rd ed., pp. 64–136). Routledge. 10.4324/9780429433214-3 [DOI] [Google Scholar]
- Paul GL, Mariotto MJ, & Redfield JP (1986). Sources and methods for gathering information in formal assessment. In Paul GL (Ed.), Assessment in residential treatment settings: Principles and methods to support cost-effective quality operations (pp. 127–162). Research Press. [Google Scholar]
- Paunonen S, & Jackson DN (1996). The Jackson Personality Inventory and the five-factor model of personality. Journal of Research in Personality, 30, 42–59. 10.1006/jrpe.1996.0003 [DOI] [Google Scholar]
- Penman R, Cross T, Milgram-Friedman J, & Meares R (1983). Mother’s speech to prelingual infants: A pragmatic analysis. Journal of Child Language, 10, 17–34. 10.1017/S0305000900005109 [DOI] [PubMed] [Google Scholar]
- Pesch MH, & Lumeng JC (2017). Methodological considerations for observational coding of eating and feeding behaviors in children and their families. International Journal of Behavioral Nutrition and Physical Activity, 14(1), 1–14. 10.1186/s12966-017-0619-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Livingston G, & Parker K (2019). 8 facts about American dads. Pew Research Center. https://www.pewresearch.org/fact-tank/2019/06/12/fathers-day-facts/ [Google Scholar]
- Putnick DL, & Bornstein MH (2016). Measurement invariance conventions and reporting: The state of the art and future directions for psychological research. Developmental Review, 41, 71–90. 10.1016/j.dr.2016.06.004 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Sackett GP (1978). Measurement in observational research. In Sackett GP (Ed.), Observing behavior: Vol. 2. Data collection and analysis methods (pp. 25–44). University Park Press. [Google Scholar]
- Satorra A, & Bentler PM (1994). Corrections to test statistics and standard errors in covariance structure analysis. In Von Eye A& Clogg CC (Eds.), Latent variable analysis: Applications for developmental research (pp. 399–419). Sage Publications. [Google Scholar]
- Satorra A, & Bentler PM (2001). A scaled difference chi-square test statistic for moment structure analysis. Psychometrika, 66(4), 507–514. 10.1007/BF02296192 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Schaller M (2018). The parental care motivational system and why it matters (for everyone). Current Directions in Psychological Science, 27, 295–301. doi: 10.1177/0963721418767873 [DOI] [Google Scholar]
- Schultheis H, & Cooper RP (2022). Everyday Activities. Topics in Cognitive Science. doi: 10.1111/tops.12603 [DOI] [PubMed] [Google Scholar]
- Scott-Lennox JA, & Lennox RD (1995) Sex-race differences in social support and depression in older low-income adults. In Hoyle RH (Ed.), Structural equation modeling: Concepts, issues, and applications. Sage Publications. [Google Scholar]
- Shrout PE, & Fleiss JL (1979). Intraclass correlations: Uses in assessing rater reliability. Psychological Bulletin, 86(2), 420–428. 10.1037/0033-2909.86.2.420 [DOI] [PubMed] [Google Scholar]
- Sim J, & Wright CC (2005). The Kappa statistic in reliability studies: Use, interpretation, and sample size requirements. Physical Therapy, 85(1), 257–268. 10.1093/ptj/85.3.257 [DOI] [PubMed] [Google Scholar]
- Soderström M (2007). Beyond babytalk: Re-evaluating the nature and content of speech input to preverbal infants. Developmental Review, 27(4), 501–532. 10.1016/j.dr.2007.06.002 [DOI] [Google Scholar]
- Stark EA, Stein A, Young KS, Parsons CE, & Kringelbach ML (2019). Neurobiology of human parenting. In Bornstein MH (Ed.), Handbook of parenting. Vol. 2. Biology and ecology of parenting (3rd ed., pp. 250–284). Routledge. 10.4324/9780429401459-7 [DOI] [Google Scholar]
- Stevenson MB, Leavitt LA, Roach MA, Chapman RS, & Miller JF (1986). Mothers’ speech to their 1-year-old infants in home and laboratory settings. Journal of Psycholinguistic Research, 15, 451–461. 10.1007/BF01067725 [DOI] [PubMed] [Google Scholar]
- Streiner DL (2003). Being inconsistent about consistency: When coefficient alpha does and doesn’t matter. Journal of Personality Assessment, 80, 217–222. 10.1207/S15327752JPA8003_01 [DOI] [PubMed] [Google Scholar]
- Suen HK, & Ary D (1989). Analyzing quantitative behavioral observational data. Erlbaum. [Google Scholar]
- Tabachnick BG, & Fidell LS (2019). Using multivariate statistics (7th ed.). Pearson. [Google Scholar]
- Thornberry T, & Brestan-Knight E (2011). Analyzing the utility of the dyadic parent-child interaction coding system (DPICS) warm up segments. Journal of Psychopathology and Behavioral Assessment, 33, 187–195. 10.1007/s10862-011-9229-6 [DOI] [Google Scholar]
- Tomlinson M, Bornstein MH, Marlow M, & Swartz L (2014). Imbalances in the knowledge about infant mental health in rich and poor countries: Too little progress in bridging the gap. Infant Mental Health Journal, 35, 624–629. 10.1002/imhj.21462 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Tooby J, & Cosmides L (1992). The psychological foundations of culture. In Barkow JH, Cosmides L, & Tooby J (Eds.), The adapted mind: Evolutionary psychology and the generation of culture (pp. 19–136). Oxford University Press. [Google Scholar]
- Trevarthen C, & Aitken KJ (2001), Infant intersubjectivity: Research, theory, and clinical applications. Journal of Child Psychology and Psychiatry, 42, 3–48. 10.1111/1469-7610.00701 [DOI] [PubMed] [Google Scholar]
- Trevarthen C, & Hubley P (1978). Secondary intersubjectivity: Confidence, confiding and acts of meaning in the first year. In Lock A (Ed.), Action gesture and symbol (pp. 183–229). Academic Press. [Google Scholar]
- van de Vijver FJR, & Leung K (1997). Cross-cultural psychology series, Vol. 1. Methods and data analysis for cross-cultural research Sage Publications, Inc. [Google Scholar]
- Waggoner P (2021). Modern dimension reduction (Elements in quantitative and computational methods for the social sciences). Cambridge University Press. 10.1017/9781108981767 [DOI] [Google Scholar]
- Weisner TS, & Gallimore R (1977). My brother’s keeper: Child and sibling caretaking. Current Anthropology, 18, 169–190. 10.1086/201883 [DOI] [Google Scholar]
- Wilson EO (1975). Sociobiology. Harvard University Press. [Google Scholar]
- Worthman CM, DeCaro J, & Brown R (2002). Cultural consensus approaches to the study of American family life (MARIAL Working Paper 13). CCM and Working Families. http://anthropology.emory.edu/home/documents/worthman-lab/Cultural%20consensus.pdf [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.