Early operant learning is unaffected by socio-economic status and other demographic factors: A meta-analysis

https://doi.org/10.1016/j.infbeh.2012.02.005Get rights and content

Abstract

The relation between SES (socioeconomic status) and academic achievement in school-aged children is well established; children from low SES families have more difficulty in school. However, few studies have been able to establish a link between SES and learning in infancy, and thus the developmental onset of SES effects remains unknown. The limited studies that have been conducted to explore the link between SES and learning in infancy have generated mixed results; some demonstrate a link between SES and learning in infants as young as 6–9 months (Smith, Fagan, & Ulvund, 2002) while others do not. Further, studies examining the genetic as well as environmental contributors to learning in infancy and early childhood suggest that the effect of SES is likely cumulative and that as children develop, the effect of a low SES environment will become more pronounced (Tucker-Drob, Rhemtulla, Harden, Turkheimer, & Fask, 2011). Using aggregated data from 790 infants collected across 18 studies, we examined the contribution of SES and other demographic factors to learning of an operant kicking task in 2–4-month-old infants in a meta-analysis. Results indicated that, at least with respect to operant conditioning, an infant is an infant; that is SES did not affect learning rate or ability to learn in infants under 4-months of age. SES effects may therefore be better characterized as cumulative, with tangible effects emerging sometime later in life.

Highlights

► A meta-analysis of whether demographics are related to learning in young infants. ► Demographics include socioeconomic status, age and gender.► Regression models show no significant relation. ► This finding is discussed in terms of the onset of SES as a predictor of later cognitive factors.

Introduction

Socioeconomic status (SES) typically combines the educational achievement, economic success, and occupational prestige of an individual or family. A popular measure of SES was developed by Nakao and Treas (1992) and is frequently used today to classify participants in developmental studies. SES is related to scholastic preparation in toddlers and preschool children (Arnold & Doctoroff, 2003) and to academic performance in older children (Ginsburg and Bronstein, 1993, Pungello et al., 1996). SES differences in pre-school aged children have been observed in language development (Fish & Pinkerman, 2003), IQ, and other measures (see McLoyd, 1998 for review). These data regarding the link between SES and performance on multiple cognitive measures, as well as the demonstrated relationship between SES and many health-related issues (e.g., smoking incidence and weight issues in adolescence) were part of the motivation for federal funding agencies and journal editors to adopt a now long-standing policy requiring representational sampling and reporting of demographics (National Institute of Health Revitalization Act of 1993). In addition to sex and ethnicity, journals routinely ask that research reports investigating developmental questions include demographic data on the age and socioeconomic status (SES) of infants tested.

While the utility of data regarding socio-economic status overall is widely acknowledged, the evidence demonstrating an effect of SES in infancy is less clear than the evidence for its impact on cognition and educational outcomes in older children. For example, Tucker-Drob, Rhemtulla, Harden, Turkheimer, & Fask (2011) found that SES differences emerge around ten months, and these differences grow significantly across the first two years of life. Further, Jansen et al. (2009) reported that SES can predict activity level; specifically, 6-month-olds from low SES homes were more active than infants from higher SES homes. High activity as a trait has been associated with reactive temperament (Fagen et al., 1987, Rimm-Kaufman and Kagan, 2005, Rothbart and Ahadi, 1994). Conversely, McCall and Carriger (1993), in a meta-analysis of studies testing infants from newborn to twelve months of age, found that while a relationship between habituation, recognition memory and later IQ exists, infants show no predictive relation between SES and measures of habituation and recognition memory, a finding supported by Smith, Fagan, and Ulvund (2002). Rose, Feldman, Wallace, and McCarton (1989) reported a similar finding from tests of 7-month-olds.

Thus, overall, the presence of an SES effect on cognitive performance in infancy is debatable and would appear to be age-dependent. This difficulty in consistently demonstrating an effect of SES in the first year of life suggests that SES effects may be cumulative; that is, these effects may rely on a complex set of environmental factors dynamically affecting children over time (McLoyd, 1998). Persistently low SES children are exposed to more extreme environmental conditions and are more likely to be of poor health and live in homes with limited opportunities for cognitive stimulation. These conditions are chronic and may not have a measurable impact in infancy, but instead may impact development years later when the child is school-aged. The presence of SES effects in studies of older children and adults, as well as the inconsistency of findings on SES effects in infancy, make this investigation of the relation between SES and infant learning especially important. Chiefly, the question of when SES effects become impactful remains. In short, is it the case that SES is related to learning rate or learning success early in infancy, or does SES only affect learning following a significant stretch of developmental exposure?

The current study addressed the relation between SES and learning in early infancy using a robust measure of associative learning (described next) and a large data set. Data collected from 2 to 4-month-old infants across multiple studies using a single operant learning procedure were used to investigate the extent of a relation between SES and several different measures of activity and learning at a point early in development. This approach enabled an investigation of whether SES effects are detectable at this early point in ontogeny, which in turn provided a data point asserting that, at this early point in development, any potential harm from a low SES environment is minimal at worst. Few studies have investigated this issue in young infants (see McCall & Carriger, 1993 for review) and the literature on SES and learning in infancy is limited by small samples and the use of a variety of different measures that make cross-study comparisons difficult. The current study differs in that performance was examined in an unusually large group of infants all tested using a single task. This large sample increases external validity when generalizing findings to a wider population of infants.

Data for this analysis were drawn from studies using the mobile conjugate reinforcement procedure, developed by Rovee-Collier (Rovee-Collier, 1996, Rovee and Rovee, 1969, Sullivan et al., 1979). This procedure involves training infants (typically across two daily sessions) to kick in the presence of a visual stimulus (an overhead crib mobile fixed with jingling bells), which moves and rings contingent upon the infant's kicks. Multiple studies have demonstrated (Adler et al., 1998, Guyla et al., 1999, Hill et al., 1988, Rovee-Collier, 1993, Rovee-Collier and Sullivan, 1980) that kick rates will increase across the two days of training as infants learn the contingency present in this procedure. Learning during the training sessions is defined by an increase in kicking of greater than 1.5 times the baseline-kicking rate (see Ohr, Fagen, Rovee-Collier, Hayne & Vander Linde, 1989 for a full description of this criterion). The learning rate is defined by the number of minutes of training needed in order to achieve this pre-defined increase.

This analysis, across 790 infants, offers a definitive description of the relation between SES and performance on an operant learning task early in life. The literature reviewed clearly shows that this is an issue of interest to the research community, but the dearth of data on early performance means that the starting point for the digression in performance between low-SES and higher-SES children is uncertain. Data on specific ages were included in the present analysis to permit examination of the possibility that even within the (limited) range tested, SES might begin to exert more of an influence at the higher end of this early point in development. The data presented below argue strongly that performance differences based on SES diverge at some point after the first few months of life, and therefore suggests that, as other work has shown, the lag seen in lower-SES children emerges only after some level of extended development within a lower-SES environment, and should be amenable to manipulations intended to ameliorate the performance decrements that have been reported later in life.

Section snippets

Participants

The present study included 1061 infants (560 males) from 2 to 4-month-old (age in days = 61–145, SD = 10.48). The parents of 814 participants reported SES information; the average SES score was 68.23 (SE = 0.51). The parents of 1024 infants reported ethnicity. The majority of the sample was Caucasian (N = 875). The remaining 14% of the sample was Hispanic, African American, Asian American and those who self-identified as “other” (N = 150). Parents of the remaining 36 infants did not report ethnicity.

Results

Two separate regression analyses were used to determine whether sex, age and SES (and interactions between these terms) predicted baseline rate of kicking as well as learning rate in the mobile conjugate reinforcement task. Stepwise regressions were used in order to evaluate the unique and additive contribution of SES and interactions between SES and other demographic variables. If SES could bolster the predictive power of the model after other predictive variables were entered (e.g. age and

Discussion

The outcome of these analyses showed that, for full-term infants, SES does not affect activity rate, learning rate, or probability of learning success. Indeed, given the poor fit of the model to the data, it is most reasonable to assert (because the sample was certainly large enough to detect even small effects) that there was no relation between any of the demographic variables (age in months, sex, or SES) and any of the measures of learning that were examined.

The results of this study are

References (31)

  • W.L. Hill et al.

    Continuities in infant memory

    Developmental Psychobiology

    (1988)
  • P.W. Jansen et al.

    Socioeconomic inequalities in infant temperament

    Social Psychiatry & Psychiatric Epidemiology

    (2009)
  • K.S. Kraebel et al.

    The influence of training views on infants’ long-term memory for simple 3D shapes

    Developmental Psychobiology

    (2007)
  • R.B. McCall et al.

    A meta-analysis of infant habituation and recognition memory performance as predictors of later IQ

    Child Development

    (1993)
  • V.C. McLoyd

    Socioeconomic disadvantage and child development

    American Psychologist

    (1998)
  • Cited by (13)

    • Learning abilities

      2020, Handbook of Clinical Neurology
      Citation Excerpt :

      Operant conditioning remains an important learning mechanism throughout infancy and the entire lifespan and some even argue that it is central to language acquisition, though this has been heavily debated (Sturdy and Nicoladis, 2017). In contrast to other, more complex forms of learning, operant conditioning in infants under 4 months of age has been found to be unaffected by socioeconomic status (Gerhardstein et al., 2012). The questions of how more complex associative and relational learning develops and when it can first be observed have been given much attention in infant research.

    • The Development of Infant Memory

      2022, The Development of Memory in Infancy and Childhood, Third Edition
    View all citing articles on Scopus

    The authors wish to express their appreciation to Carolyn Rovee-Collier, Rutgers University, for providing access to data for this research, and to the parents and infants for their participation. Stacie Miller is now at the Monell Chemical Senses Center, Philadelphia, Pennsylvania, United States.

    View full text