Skip to main content
NCBI home page
Wiley Open Access Collection logoLink to Wiley Open Access Collection
. 2026 Jun 30;29(5):e70198. doi: 10.1111/desc.70198

The Kia Tīmata Pai Randomized Controlled Trial: ENRICH Early Childhood Teacher Training Improves Toddlers’ Oral Language and Self‐Regulation

Elaine Reese 1,✉, Tugce Bakir‐Demir 1, Louis J Moses 2,3, Sean Marshall 1, Elizabeth Schaughency 1, Karen Salmon 2, Mele Taumoepeau 2, Jimmy McLauchlan 4, Clair Edgeler 5, Hayley Guiney 6, Jesse Kokaua 7, Paul E Jose 2, Amanda Clifford 1, Natasha Maruariki 5, Yana Dorana 1, Yuxin Zhang 1, Scarlet Mollan 4, Stuart McNaughton 8
PMCID: PMC13318364  PMID: 42378677

ABSTRACT

Children's oral language and self‐regulation skills are critical for healthy development and life trajectories. ENRICH is a preventive intervention developed for early childhood teachers to enhance these skills in toddlers. Kia Tīmata Pai (Best Start study) is a longitudinal cluster randomized controlled trial in 136 New Zealand early childhood education centers testing the impact of ENRICH with 1875 teachers on 1481 toddlers’ (M = 20.6 months at recruitment; 72% European, 24% Māori, 23% Asian, 9% Pacific using a total response method; 7% low‐SES, 40% middle‐SES, 53% high‐SES) oral language and self‐regulation skills. After 18 months, ENRICH produced benefits for children's oral language and self‐regulation skills (parent and teacher ratings) and for early literacy and social skills (teacher ratings) compared to an active control condition. Oral language benefits (parent ratings) emerged earlier for children from low‐ and mid‐socioeconomic (SES) families and whose mothers had lower educational qualifications. Benefits for oral language and self‐regulation skills at age 3 held regardless of children's SES, ethnicity, bilingual status, initial language level, age, and maternal education. Benefits for oral language at age 3 were moderated by child sex and differed by informant, with teachers perceiving greater ENRICH benefits for boys, and parents for girls. The likely mechanism of ENRICH's effects is through sensitive serve‐and‐return interactions that support children's oral language and attention skills. ENRICH is a potentially powerful tool for supporting toddlers’ language, cognitive, and socioemotional development with possible downstream effects for academic achievement, health, and well‐being. A video abstract of this article can be viewed at https://youtu.be/iIV1l0‐cKIo.

Summary

  • ENRICH is a new oral language preventive intervention designed for early childhood education teachers of toddlers based on serve‐and‐return and rich conversations across the day.

  • A longitudinal cluster randomized controlled trial showed that ENRICH improved toddlers’ oral language, self‐regulation, emergent literacy, and social skills compared to an active control condition.

  • Oral language benefits emerged earlier for toddlers from low‐ and middle‐socioeconomic status (SES) families and whose mothers had lower educational qualifications.

  • Benefits for oral language and self‐regulation skills at age 3 held regardless of children's SES, maternal education, ethnicity, bilingual status, initial language level, and age.

Keywords: early childhood education, intervention, literacy, oral language, self‐regulation, toddlers

1. Introduction

Young children's oral language and self‐regulation skills are critical for healthy development. Oral language comprises vocabulary, grammar, pragmatic, and narrative skills. Such skills enable thinking, communication, and social relationships, as well as later reading, mathematics, and academic achievement (Lonigan and Shanahan 2010; Pace et al. 2019), and these in turn predict adult health and well‐being (Schoon et al. 2010). Self‐regulation, encompassing cognitive (attention, memory), behavioral (inhibition, task persistence), and emotion regulation skills, predicts later academic achievement, physical and mental health, and well‐being over the life course (Moffitt et al. 2011; Robson et al. 2020).

Both oral language and self‐regulation are developing rapidly—and thus are exquisitely sensitive to enrichment—in the toddler period from ages 1 to 3 (Bornstein et al. 2018; Cunningham et al. 2023). Moreover, these skills are significantly correlated across the toddler years (Masek et al. 2021; Thériault‐Couture et al. 2025). Although there are likely bidirectional influences between them over this period, our guiding theory is that early oral language potentiates self‐regulation (Reese et al. 2023a; Salmon et al. 2016). For example, oral language strengthens attention and memory skills, enables children to follow instructions, and to express negative emotions in words rather than actions (Salmon et al. 2016). Early oral language also predicts later self‐regulation to a greater degree than early self‐regulation predicts later oral language (Bruce and Bell 2022). The aim of the present research was to foster toddlers’ oral language through training of early childhood teachers, with expected benefits for both oral language and self‐regulation.

1.1. The Role of Adult‐Child Contingent Interactions in Language Development

The quantity and quality of adult language input shape toddlers’ oral language development (Gilkerson et al. 2018; Hart and Risley 1995; Masek et al. 2022; Anderson et al. 2021). Responsive back‐and‐forth interactions (“serve and return”) are considered especially critical for toddlers’ language and attention skills (Masek et al. 2021; Gilkerson et al. 2018; Hirsh‐Pasek et al. 2015). For instance, adult‐child exchanges that are contingent upon children's behaviors and responses are linked to multiple aspects of oral language growth, perhaps via increases in attention (Masek et al. 2021). The evidence for this hypothesis, however, is largely correlational and based on parental interactions (Anderson et al. 2021). One notable exception, at least with respect to the correlation issue, is an intervention with Australian parents of 1201 at‐risk toddlers called smalltalk focusing on linguistically rich and responsive contingent interactions. The smalltalk intervention produced short‐term benefits for parent‐child interactions and long‐term benefits for children's self‐regulation, assessed via parent‐ and teacher‐reported effortful control, at age 7.5 years (Bennett et al. 2025). Improvements in parents' skill at maintaining and extending children's focus in interactions 5 months after the intervention mediated these long‐term self‐regulation benefits. Little is known, however, about how such initiatives could benefit toddlers in early childhood education (ECE) settings because almost all existing large‐scale oral language interventions in those settings are with preschoolers aged 3–5 (Noble et al. 2019; but see Bleses et al. 2021).

1.2. Fostering Oral Language Development in Early Childhood Education

ECE attendance is increasing around the world from younger ages and for longer hours (OECD 2023). However, many ECE settings do not provide young children with a high‐quality language environment (Dickinson et al. 2014). It is especially imperative to improve teacher‐toddler interactions because increasingly high levels of screen time at home are linked to decreased quality of parent‐child interactions, language delays, and behavioral problems in young children (Barr et al. 2024; Brushe et al. 2024). Globally, only a few large‐scale randomized controlled trials (RCTs) in ECE settings to enhance toddlers’ oral language skills exist (Bleses et al. 2020, 2021; Landry et al. 2014). To our knowledge, only the 20‐week Danish trial, We Learn Together (N = 2327), showed immediate and long‐term benefits (1 year later) for toddlers’ oral language over and above control (curriculum‐as‐usual) (Bleses et al. 2020; 2024). We Learn Together also produced numeracy and socioemotional benefits but not self‐regulation benefits. The 36‐week US trial Responsive Early Childhood Curriculum (RECC, N = 542 low‐income toddlers), also found immediate socioemotional benefits but, unlike the Danish study, no oral language, literacy, or numeracy benefits over curriculum‐as‐usual (Landry et al. 2014). Self‐regulation was not assessed in the US study.

Both the Danish and US studies emphasized the use of responsive, contingent teacher‐child interactions. However, the two studies differed in key ways that could explain the different findings. First, the teacher:child ratio was much more favorable in Denmark (1:3 for toddlers) compared to the US study (1:12). Second, the teachers in We Learn Together held higher educational qualifications (60% with bachelor's degrees) than in the RECC study (8.4% with bachelor's degrees). Third, We Learn Together utilized a defined scope targeting oral language and math language, and a sequenced, yet unscripted approach of instruction. In contrast, the RECC study followed a sequence and scope only for socioemotional skills, not oral language (see Bleses et al. 2020).

1.3. The Current Study

The current study is a preregistered longitudinal cluster RCT in Aotearoa New Zealand (NZ) with 136 centers from a single national ECE provider (BestStart Educare). Centers were randomly assigned to either an oral language professional development condition (ENRICH) or an active control condition with 1481 toddlers over three assessment waves: baseline (at approximately 20 months) and 9‐monthly post‐tests (approximately 29 and 37 months). Our oral language intervention—ENRICH (Reese et al. 2023a) (Enhancing Rich Interactions)—is centered on teachers’ use of serve‐and‐return interactions (Fisher et al. 2016; Rowe and Snow 2020; Shonkoff and Bales 2011) with toddlers across the day's routines. Like the successful We Learn Together study, our approach followed a sequence and scope for oral language skills and was unscripted. We used a train‐the‐trainer model (Bleses et al. 2021) in which researchers conducted two workshops (at approximately 21 and 30 months, respectively) with six of the ECE provider's professional practice leaders (PPLs). PPLs were all experienced teachers who then delivered workshops to teachers in ENRICH centers. A non‐profit implementation service (Methodist Mission Southern) provided each ENRICH center with two new picture books every 4.5 months (total of four books in English and two translated into te reo Māori). The four English books contained conversation‐starters on each page, a feature that had boosted preschool children's oral language, literacy, and socioemotional skills in a prior parent intervention (Riordan et al. 2022; Reese et al. 2023; Schaughency et al. 2023; Clifford et al. 2024). ENRICH centers also received two sets of instructional cards with tips and examples for increasing serve‐and‐return interactions in five routines (book‐reading, diaper changes, mealtime, small‐group, play), along with video clips demonstrating techniques. Parents of children in ENRICH centers were invited to online workshops designed for home activities in reading, reminiscing, and sound play delivered by researchers. Teachers and parents in ENRICH centers were encouraged to conduct interactions in children's home languages when possible. Teachers and parents in Active Control centers received workshops on toddler nutrition delivered by PhD‐level experts.

In line with our guiding theory (Salmon et al. 2016), the main expected outcomes of the ENRICH intervention were improvements in oral language and self‐regulation. ENRICH incorporates shared book‐reading and conversations about children's past, ongoing, and future experiences, including sound play and discussions linking book characters’ and children's feelings (Riordan et al. 2022; Reese et al. 2023; Schaughency et al. 2023; Clifford et al. 2024). Therefore, we also explored children's early literacy and socio‐emotional outcomes. Our preregistered hypotheses (timestamped prior to data analysis August 12, 2024) were as follows (https://osf.io/vkgxd/overview; see Table S32 for minor deviations from the preregistration):

  1. Children in ENRICH will show more advanced oral language at post‐tests (Waves 2 and 3) than children in Active Control.

  2. Children in ENRICH will show more advanced self‐regulation at post‐tests (Waves 2 and 3) than children in Active Control.

  3. Exploratory: Children in ENRICH will show reduced negative affect and increased positive affect/surgency (Wave 2) and higher early literacy and social skills (Wave 3) than children in Active Control.

  4. Exploratory: Benefits of ENRICH will be moderated by child and family characteristics: child age, sex, initial language level, bilingualism, ethnicity, maternal education, and family socioeconomic status (SES).

2. Method

Participants: We recruited 1496 families of toddlers and 1875 teachers from 136 early childcare centers in a single service provider across two cohorts to participate in the toddler phase of the study. Centers came from three regions (Northern = 39, Midland = 46, Southern = 51) across NZ. Children outside the desired age range of 13–30 months were excluded, yielding a total of 1481 children (721 girls and 760 boys) for analysis. Only 14 children/families (< 1%) formally withdrew from the study during the toddler phase; in addition, a total of 434 children left their study center so could no longer be assessed at some waves (see Figure 1). Cohort A children (n = 1126, M = 20.13, SD = 3.23) enrolled from May‐August 2021 and Cohort B children (n = 355, M = 22.08, SD = 3.44) from February to April 2022. The sample was roughly representative of the ethnic distribution of children aged 0–4 years in Aotearoa New Zealand at that time (72% European, 24% Māori, 23% Asian, 9% Pacific using a total response method) (Didham 2005), but with fewer low‐SES families due to government subsidies for childcare not beginning until age 3 (7% low‐SES, 40% middle‐SES, 53% high‐SES; Poulton et al. 2025). English was the most common home language for children (83%) with Hindi (6%), Mandarin (2%), and Cantonese (1%) the top home languages for bilingual children. At initial recruitment in 2021, teachers also represented a broad range of ethnicities (63% European, 24% Asian, 10% Māori, 4% Pacific) and educational backgrounds (62% with a university degree, with or without a postgraduate qualification; 75% with an early childhood qualification); 41% had fewer than 5 years of ECE teaching experience (Poulton et al. 2025).

FIGURE 1.

FIGURE 1

Open in a new tab

Participant recruitment flowchart. The flowchart outlines recruitment and attrition across three waves of data collection for Cohorts A and B, as well as teacher participants. Attrition is classified as either voluntarily withdrawn from study or left center due to a change in childcare or starting school.

Procedure: Assessment waves occurred at baseline (Wave 1; Mage = 20.60, SD = 3.38) and at approximately 9‐month intervals thereafter at Wave 2 (Mage = 29.91, SD = 3.25) and Wave 3 (Mage = 37.15, SD = 3.12). Parents and teachers completed questionnaires at each assessment wave via REDCap (Harris et al. 2009, 2019) or on paper forms that were then double‐entered into REDCap. Teachers received credit for professional development workshops and received release time to complete assessments at each wave. Participating parents who completed assessments were entered into raffles for 10 prizes of $250/each at each assessment wave. Parents also received a small gift of a book for those in ENRICH centers or a toy for those in Active Control centers. Centers that completed all child rating assessments were also entered into a raffle for 10 prizes of $250/each at each assessment wave.

Following baseline, centers were randomly assigned to one of four conditions (Reese et al. 2023a) (see Figure 1): ENRICH (oral language from ages 1 to 5 years); ENGAGE (self‐regulation from ages 3 to 5 years); ENRICH + ENGAGE (oral language from ages 1 to 5 years and self‐regulation from ages 3 to 5 years); and Active Control (curriculum‐as‐usual plus a total of five webinars on child development topics across the toddler and preschool waves). For the toddler phase from ages 1 to 3 years (the focus of this study), all centers were either in ENRICH (n = 67) or Active Control (n = 69).

ENRICH for toddlers is based on serve‐and‐return interactions across the day—during book time, diaper time, mealtime, group time, and play time—with a special focus on book‐reading, reminiscing, and sound play activities (Reese et al. 2023a). ENRICH was administered via a train‐the‐trainer procedure in which the research team trained a small group of professional practice leaders in the early childhood service to administer the 1.5‐h workshops to teachers online (Reese et al. 2023a). Parents were also offered similar ENRICH workshops soon after the teacher workshops were completed, but uptake was low (<10%). Two ENRICH workshops were delivered during the toddler phase: one after baseline (Wave 1) for younger toddlers and a second after Wave 2 for older toddlers. Teachers received ENRICH resources comprising six picture books (two in English, two bilingual in English and te reo Māori, and two in te reo Māori): two books after the first workshop and four books after the second. The English and bilingual books contained prompts on each page with conversation starters that were color‐coded as easy, medium, and challenging. Books were selected to become slightly more challenging with each phase of ENRICH. Teachers also received two sets of informational cards about children's language development at each phase with ideas for activities during routines across the day: book‐reading, diaper changes, mealtime, small‐group, and play. Finally, teachers were given links to videos demonstrating and explaining ENRICH techniques. Teachers and parents in the Active Control centers were offered two 1.5‐h webinars on childhood nutrition delivered by PhD‐level experts at the same time as the ENRICH workshops. Again, parent uptake was low (<10%).

Measures: Parents and teachers completed questionnaires at each assessment wave on their own demographics and educational qualifications, and on children's oral language, early literacy, positive and negative affect, self‐regulation, and social skills. Parents also answered questions about their child's demographics (Table S4). As Table 1 shows, across all three waves, the language, early literacy, self‐regulation, and social skills measures showed acceptable to excellent internal consistency (α = 0.70–0.99), indicating reliable assessment of children's development across time and informants (parents and teachers). The writing skills measure at Wave 3 showed lower internal consistency (α = 0.67) for parent and teacher reports. In addition, at each wave, and for each condition, parent and teacher reports of oral language and self‐regulation were significantly intercorrelated, as were parent and teacher reports of early literacy and social skills at Wave 3 (Tables S5–S10).

TABLE 1.

Summary of assessments (scale‐item internal consistency) across the three waves.

Language Self‐regulation Demographic information
Parent Teacher Parent Teacher Parent
Baseline (Wave 1) CDI

  • Gestures (α = 0.79)

  • 100‐word English vocabulary list (α = 0.98)

  • Syntax single‐item

 CDI

  • Gestures (α = 0.92)

  • 100‐word English vocabulary list (α = 0.98)

  • Syntax single‐item

 ECBQ

  • Effortful control (α = 0.73)

  • Surgency (α = 0.70)

  • Negative affect (α = 0.73)

 ECBQ

  • Effortful control (α = 0.84)

  • Surgency (α = 0.86)

  • Negative affect (α = 0.82)

 Demographic information

  • Child age

  • Child sex

  • Child ethnicity

  • Child language status

  • Family SES

  • Maternal education
Wave 2 CDI

  • 131‐word English vocabulary list (α = 0.99)

  • Syntax single‐item

 CDI

  • 131‐word English vocabulary list (α = 0.99)

  • Syntax single‐item

 ECBQ

  • Effortful control (α = 0.76)

  • Surgency (α = 0.74)

  • Negative Affect (α = 0.75)

 ECBQ

  • Effortful control (α = 0.85)

  • Surgency (α = 0.84)

  • Negative Affect (α = 0.84)
Wave 3 PROLL

  • Oral Language (α = 0.84)

  • Early Literacy (α = 0.75)

  • Writing (α = 0.67)

 TROLL

  • Oral Language (α = 0.91)

  • Early Literacy (α = 0.81)

  • Writing (α = 0.67)

 CBRS

  • Self‐Regulation (α = 0.88)

  • Social Skills (α = 0.72)

 CBRS

  • Self‐Regulation (α = 0.94)

  • Social Skills (α = 0.85)
Open in a new tab

Abbreviations: CBRS, Child Behavior Rating Scale; CDI, New Zealand Communicative Development Inventories; ECBQ, Early Childhood Behavior Questionnaire–very short form; PROLL, Parent Rating of Oral Language and Literacy; TROLL, Teacher Rating of Oral Language and Literacy; α, Cronbach's alpha.

2.1. Oral Language and Early Literacy Skills

At Wave 1 (baseline (Bakir‐Demir et al. 2026)) and Wave 2, children's language development was assessed with the NZ Communicative Development Inventories (CDI: a 100‐word English vocabulary list plus syntax question with both parent and teacher forms) (Peterson et al. 2017; Reese et al. 2018). The single question about the child's syntax asked whether children were combining words yet or not in any language. At Wave 1 only, a checklist of 12 gestures was also administered (Fenson et al. 1994). At Wave 2 only, an additional 31 mental state words (e.g., know, grumpy, love) were added to the English vocabulary list (Taumoepeau and Ruffman 2008). Composite language scores for teachers and parents were created by averaging the standardized scores for gestures, vocabulary (100 words), and syntax at Wave 1, and standardized scores for vocabulary (131 words) and syntax at Wave 2.

At Wave 3, because the CDI inventories were no longer age‐appropriate, the Parent Rating of Oral Language and Literacy (PROLL) (Bird et al. 2024) and the Teacher Rating of Oral Language and Literacy (TROLL) (Dickinson et al. 2003) were adapted for NZ ECE. Each scale consisted of an 8‐item Language Use subscale (henceforth Oral Language) that taps children's communication, vocabulary, sound awareness, and articulation skills in any of their languages; an 11‐item Books and Reading subscale (henceforth Early Literacy) that taps children's book behavior and early literacy skills; and a 6‐item Writing subscale of their early writing behavior. Because children were essentially at floor on the Writing subscale, we do not analyze it further. Items were rated on 4‐point scales describing different scenarios (corresponding to not yet, sometimes, often, very often) except for one item from the early literacy subscale (Does your child recognize their own first name in print?) that was rated as “not yet” or “yes.”

2.2. Self‐Regulation and Social Skills

At Waves 1 and 2, the parent and teacher versions of the Early Childhood Behavior Questionnaire: Very Short Form (Putnam et al. 2006) assessed children's temperament and self‐regulation (Bakir‐Demir et al. 2026). The questionnaire comprises 36 items measured on a 7‐point Likert scale (from never to always), yielding three scale scores: Effortful Control, Negative Affect, and Positive Affect/Surgency. Average scores were calculated for these subscales. Given our focus on self‐regulation, we were particularly interested in the Effortful Control subscale (12 items), which measures inhibitory control, attentional control, low‐intensity pleasure, and perceptual sensitivity.

At Wave 3, because the ECBQ was no longer age‐appropriate, we administered the Child Behavior Rating Scale (CBRS) (Bronson et al. 1990). The CBRS comprises 10 items tapping self‐regulation skills and seven items tapping social skills, all measured on a 5‐point Likert scale (never to always). The scale was created for teachers in a classroom setting; we slightly adapted the wording of three items for parents (e.g., we added “a parent” to the item “Does not fuss when he/she has to wait briefly to get attention from a parent or other adult; child may be asked once to wait by a parent or other adult”). Average scores were calculated for these subscales.

2.3. Demographic Information

2.3.1. Ethnicity

We used a total‐response approach to ethnicity (Didham 2005), allowing parents and teachers to nominate multiple ethnicities for themselves and their children: European, Māori, Asian, Pacific, MELAA (Middle Eastern, Latin American, or African), or Other (please specify).

2.3.2. Socioeconomic Status (SES)

Parent‐reported occupation was used to create a six‐level SES index based on the NZ Socio‐Economic Index (Boven et al. 2022). The initial six‐group classification was consolidated into three SES categories (low, middle, high) due to the small number of children in the lowest three categories.

2.3.3. Language Status

Children were classified as multilingual if their parents reported a language other than English as the primary language spoken at home. This is a conservative measure of multilingualism because some children whose primary language in the home is English may also hear other languages.

2.3.4. Sex, Age, and Parental Education

Parents provided information regarding the child's sex and age, and their own and the other parent's education levels. We used maternal education in analyses because it is a stronger predictor than paternal education of children's academic and developmental outcomes (Jeong et al. 2018).

2.3.5. Power Calculation

As noted in our protocol paper (Reese et al. 2023a), a power analysis of final‐phase primary outcomes at age 5 suggested a total sample size at baseline of approximately 1600 children (i.e., approximately 400 per condition, including control condition) to afford precision for estimates with alpha = 0.05 and power between 0.75 and 0.90. This sample size of 400 per condition allows for attrition of children over the early childhood phase of the study. We did not reach this goal of 1600 children recruited at baseline, yet for the purpose of the current toddler phase, we exceeded our sample size goal because at this phase there are only two conditions to compare (ENRICH = 751 children vs. Active Control = 730 children).

3. Results

3.1. Data Processing

Missing data increased across the waves of the study, largely as a function of children leaving centers (see Tables S1 and S2). Missingness was as follows for central oral language and self‐regulation outcomes. For parent‐reported oral language, percent missing in ENRICH was 15%, 34%, and 46% for Waves 1–3 respectively, and in Active Control 10%, 37%, and 45%; for teacher‐reported oral language, in ENRICH 11%, 18%, and 34%, and in Active Control 12%, 19%, and 31%. For parent‐reported self‐regulation, percent missing in ENRICH was 13%, 35%, and 46%, and in Active Control, 8%, 37%, and 46%; for teacher‐reported self‐regulation, percent missing in ENRICH was 15%, 19%, and 34%, and in Active Control, 17%, 20%, and 31%. Missingness was lower when calculated as percent of eligible children who remained in their original center (ranging from 7% to 24% for the above parent‐report variables across waves and 3%–17% for teacher‐report variables). Importantly, missingness did not differ as a function of condition nor was it systematically related to any covariates (e.g., ethnicity, SES, etc.). Hence, missing data were imputed (see https://osf.io/p7ex2/files/78s9m Tables S1–S3, S5–S10 for full imputation details as well as other preliminary data processing and analyses).

3.2. Analysis Plan

We conducted linear mixed‐effect models using SPSS v. 29 (IBM Corp 2022) and R v. 4.5.3 (R Core Team 2026) to account for the clustered structure of the data (i.e., children within centers). For each outcome, we ran separate models for parent and teacher reports at Wave 2 (controlling for the relevant Wave 1 outcome) and Wave 3 (controlling in separate analyses for the relevant Wave 1 and Wave 2 outcomes). Additionally, we controlled child age, sex, language status (monolingual vs. bi/multilingual), ethnicity, maternal education, family SES, and cohort in all analyses. All non‐binary covariates were grand‐mean centered; binary covariates were coded as –1/1. For ethnicity, multiple contrasts were possible (e.g., European vs. non‐European, Māori vs Non‐Māori, etc.). The pattern of results did not differ across these contrasts. We report results controlling European versus Non‐European ethnicity here; results controlling other contrasts are presented in the supplement (Tables S18–S20, S27–S29).

In addition to examining main effects of condition, we also explored potential moderator effects of child age, sex, initial language level, language status, ethnicity, maternal education, and family SES on ENRICH effects for primary outcomes of oral language and self‐regulation. When the difference between models with and without the moderator interaction term was not significant by a deviance test, we did not pursue the moderator effect further. When the interaction significantly improved the model, we added the random intercept and slope of the moderator to the model. When the difference between models with and without random slopes was not significant, the random slope was dropped. We did not conduct moderator analyses for exploratory outcomes such as early literacy and social skills.

For each set of analyses testing a priori predictions (e.g., predicted effects of ENRICH on language outcomes at Waves 2 and 3), we controlled the family‐wise error rate using Holm–Bonferroni corrections (pHB) (Giacalone et al. 2018) with αFW set to 0.05. Correspondingly, we adjusted Wald confidence intervals using these corrections (Ludbrook 2000). In contrast, for exploratory analyses (Hypotheses 3 and 4), we employed the conventional, unadjusted per‐comparison error rate of 0.05.

3.3. Overview

Figure 2 shows heat maps of condition effect sizes (Cohen's ds) for our focal outcomes of oral language and self‐regulation, separately by parent and teacher report, adjusted for covariates at each wave. In sum, no significant differences were present for these outcomes at baseline (Wave 1). At Wave 2, some intervention effects began to emerge. Intervention effects then became more widespread at Wave 3.

FIGURE 2.

FIGURE 2

Open in a new tab

Heat maps showing effect sizes (Cohen's d) by informant and assessment wave in oral language (Panel A) and self‐regulation (Panel B). Effect sizes were derived from covariate‐adjusted models comparing ENRICH and CONTROL conditions. Each cell represents the effect size for the given informant (Parent or Teacher) at each wave comparison: Wave 1, Wave 2 controlling Wave 1, Wave 3 controlling Wave 1, and Wave 3 controlling Wave 2. The black vertical line demarcates the baseline from intervention waves. Shades of red indicate the varying effects in favor of ENRICH; shades of blue indicate effects in favor of CONTROL (all non‐significant). Asterisks indicate statistically significant condition differences after Holm–Bonferroni correction (*pHB < 0.05; **pHB < 0.01; ***pHB < 0.001).

Hypothesis 1

Main effects of ENRICH on oral language.

Figure 3 illustrates ENRICH effects on oral language at Waves 2 and 3, controlling Waves 1 and 2, respectively (Table 2 for descriptives, S11 for full results).

FIGURE 3.

FIGURE 3

Open in a new tab

Boxplots showing Waves 2 and 3 language scores by condition (ENRICH vs. CONTROL), adjusted for all covariates. Horizontal bars within each box represent condition medians. Panels A and B display parent‐ and teacher‐reported Wave 2 scores controlling Wave 1 composite language, respectively. Panels C and D show Wave 3 scores controlling Wave 1 composite language. Panels E and F show the same outcomes controlling Wave 2 composite language. Asterisks indicate significant condition differences (*pHB  < 0.05; **pHB  < 0.01) after Holm–Bonferroni adjustment.

TABLE 2.

Descriptive statistics for child language outcomes across waves and informants by condition.

ENRICH (n = 751) Active Control (n = 730)
Language measures M SD Range M SD Range
Wave 1 Parent
Gestures 19.12 3.95 5–24 19.00 3.94 0–24
Vocabulary 34.04 24.32 0–100 32.12 23.46 0–100
Syntax 0.84 0.79 0–2 0.79 0.78 0–2
Composite Language 0.03 0.84 −1.95–1.85 −0.03 0.83 −2.27–1.85
Teacher
Gestures 16.82 5.48 0–24 16.28 5.97 0–24
Vocabulary 23.82 21.10 0–99 23.86 20.75 0–99
Syntax 0.69 0.75 0–2 0.73 0.74 0–2
Composite Language 0.01 0.85 −1.66–2.21 −0.01 0.85 −1.66–2.21
Wave 2 Parent
Vocabulary 73.52 21.55 3–100 72.54 22.06 2–100
Syntax 1.74 0.49 0–2 1.67 0.55 0–2
Composite Language 0.05 0.87 −3.09–1.05 −0.05 0.94 −3.1–1.05
Teacher
Vocabulary 54.22 26.85 0–100 55.49 27.05 0–100
Syntax 1.32 0.72 0–2 1.30 0.70 0–2
Composite Language 0.00 0.94 −1.88–1.49 0.00 0.92 −1.88–1.49
Wave 3 Parent
Oral language 17.06 3.90 0–24 16.40 4.03 0–24
Early literacy 14.34 3.70 0–27 14.34 3.59 1–28
Writing 2.17 1.74 0–14 2.00 1.39 0–9
Teacher
Oral language 14.05 4.96 0–24 13.16 5.23 0–24
Early literacy 11.70 3.68 2–28 10.70 3.95 0–28
Writing 1.69 1.31 0–5 1.62 1.29 0–5
Open in a new tab

Parent Reports : At Wave 2, by parent report, ENRICH children (M = –0.12)1 did not differ significantly in composite language from Active Control children, controlling Wave 1 composite language (M = –0.19), B = 0.06, 95% CI [–0.02, 0.14], pHB = 0.120, d = 0.10. In contrast, at Wave 3, ENRICH children (M = 16.65) had significantly stronger oral language than Active Control children (M = 16.16), B = 0.49, 95% CI [0.01, 0.98], pHB = 0.045, d = 0.17), controlling Wave 1 language. When controlling Wave 2, however, the condition difference was not significant, although in the predicted direction, B = 0.35, 95% CI [–0.05, 0.74], pHB = 0.096, d = 0.12).

Teacher Reports : At Wave 2, by teacher report, ENRICH children (M = –0.20) did not differ significantly in composite language from Active Control children, controlling Wave 1 composite language (M = –0.16), B = –0.04, 95% CI [–0.16, 0.08], pHB = 0.490, d = 0.06. Similarly, at Wave 3, ENRICH children (M = 13.33) did not differ significantly in oral language from Active Control children (M = 12.59) controlling Wave 1 language, although the effect was in the predicted direction, B = 0.74, 95% CI [–0.05, 1.53], pHB = 0.073, d = 0.18. Finally, when controlling Wave 2 language, ENRICH children (M = 13.89) had significantly stronger oral language than Active Control children (M = 12.99), B = 0.89, 95% CI [0.20, 1.58], pHB = 0.006, d = 0.23.

Hypothesis 2

Main effects of ENRICH on self‐regulation.

Figure 4 illustrates ENRICH effects on self‐regulation at Waves 2 and 3, controlling Waves 1 and 2, respectively (Table 3 for descriptives, S12 for full results).

FIGURE 4.

FIGURE 4

Open in a new tab

Boxplots showing Waves 2 and 3 self‐regulation scores by condition (ENRICH vs. CONTROL), adjusted for all covariates. Horizontal bars within each box represent condition medians. Panels A and B display parent‐ and teacher‐reported Wave 2 scores controlling Wave 1 self‐regulation, respectively. Panels C and D show Wave 3 scores controlling Wave 1 self‐regulation. Panels E and F show the same outcomes controlling Wave 2 self‐regulation. Asterisks indicate significant condition differences (**pHB  < 0.01; ***pHB  < 0.001) after Holm–Bonferroni adjustment.

TABLE 3.

Descriptive statistics for child self‐regulation outcomes across waves and informants by condition.

ENRICH (n = 751) Active Control (n = 730)
Self‐regulation Measures M SD Range M SD Range
W1 Parent
Effortful control 4.60 0.69 2.33–6.71 4.58 0.70 1.92–6.67
Negative affect 3.05 0.74 1.18–7 3.09 0.76 1.22–7
Surgency 5.22 0.68 2.64–7 5.21 0.72 2–7
Teacher  
Effortful control 4.69 0.83 1.78 – 6.75 4.69 0.86 1.17–7
Negative affect 2.75 0.83 1–6.75 2.60 0.82 1–5.91
Surgency 4.23 0.97 1.25 – 5.91 4.24 1.00 1.5–6.78
W2 Parent  
Effortful control 4.86 0.60 1.92–6.5 4.76 0.64 2.17–6.5
Negative affect 3.15 0.65 1.33–5.42 3.17 0.69 1.27–6.33
Surgency 5.40 0.58 3.33–7 5.40 0.61 2.67–6.92
Teacher  
Effortful control 4.75 0.86 1.86–7 4.80 0.84 1–6.83
Negative affect 2.61 0.74 1–6 2.65 0.78 1–6.22
Surgency 4.33 0.87 1.42–6.64 4.29 0.90 1.33–7
W3 Parent  
Self‐regulation 3.34 0.48 1.1–5 3.33 0.46 1.8–5
Social skills 3.46 0.37 1.43–4.71 3.45 0.38 1.86–4.86
Teacher  
Self‐regulation 3.29 0.63 1.2–5 3.12 0.62 1.1–5
Social skills 3.59 0.58 1.71–5 3.48 0.55 1.57–5
Open in a new tab

Parent Reports : At Wave 2, by parent report, ENRICH children (M = 4.83) had significantly better self‐regulation than Active Control children (M = 4.75), B = 0.08, 95% CI [0.02, 0.14], pHB = 0.010, d = 0.17, controlling Wave 1 self‐regulation. At Wave 3, however, ENRICH children (M = 3.3) did not differ significantly from Active Control children (M = 3.3), B = 0.00, 95% CI [–0.05, 0.04], pHB = 0.941, d = 0.00, controlling Wave 1 self‐regulation. The same was true controlling Wave 2 self‐regulation: ENRICH (M = 3.29) and Active Control (M = 3.32), B = –0.03, 95% CI [–0.08, 0.02], pHB = 0.387, d = 0.07.

Teacher Reports : At Wave 2, by teacher report, ENRICH children (M = 4.74) did not differ significantly in self‐regulation from Active Control children (M = 4.79), B = −0.05, 95% CI [−0.17, 0.06], pHB  = 0.356, d = 0.07, controlling Wave 1 self‐regulation. In contrast, at Wave 3, ENRICH children (M = 3.20) had significantly greater self‐regulation than Active Control children (M = 3.03), B = 0.16, 95% CI [0.07, 0.26], pHB  = 0.002, d = 0.33, controlling Wave 1 self‐regulation. The same was true controlling Wave 2 self‐regulation: ENRICH (M = 3.20) and Active Control (M = 3.02), B = 0.18, 95% CI [0.08, 0.27], pHB  = 0.001, d = 0.36.

Hypothesis 3

(exploratory). Main effects of ENRICH on children's emotion regulation, early literacy and social skills.

Full results for emotion regulation (negative affect and positive affect/surgency), and for early literacy and social skills are presented in Table S13.

Parent Reports : No significant condition differences by parent report were found for negative affect or positive affect/surgency at Wave 2 controlling relevant Wave 1 outcomes (these variables were not assessed at Wave 3).

Figure 5 illustrates ENRICH effects for early literacy at Wave 3, controlling Waves 1 and 2 composite language. At Wave 3, by parent report, ENRICH children (M = 14.12) did not differ significantly in early literacy from Active Control children (M = 14.29), controlling Wave 1 composite language, B = –0.18, 95% CI [–0.65, 0.30], p = 0.459, d = 0.05. The same was true controlling Wave 2 composite language: ENRICH (M = 14.25) and Active Control (M = 14.48), B = –0.24, 95% CI [–0.69, 0.22], p = 0.309, d = 0.07.

FIGURE 5.

FIGURE 5

Open in a new tab

Boxplots showing Wave 3 early literacy by condition (ENRICH vs. CONTROL), adjusted for all covariates. Horizontal bars within each box represent condition medians. Panels A and B display parent‐ and teacher‐reported scores controlling for Wave 1 composite language, respectively. Panels C and D show the same outcomes controlling for Wave 2 composite language. Asterisks indicate significant condition differences (**p < 0.01; ***p < 0.001).

Figure 6 illustrates the effects of ENRICH for social skills at Wave 3 (the only wave at which they were assessed). At Wave 3, by parent report, ENRICH children (M = 3.44) did not differ significantly in their social skills from Active Control children (M = 3.44), B = 0.00, 95% CI [−0.05, 0.05], p = 0.940, d = 0.01.

FIGURE 6.

FIGURE 6

Open in a new tab

Boxplots showing Wave 3 social skill scores by condition (ENRICH vs. CONTROL), adjusted for all covariates. Horizontal bars within each box represent condition medians. Panels A and B display parent‐ and teacher‐reported scores, respectively. Asterisks indicate significant condition differences (*p < 0.05).

Teacher Reports : No significant condition differences by teacher report were found for negative affect or surgency at Wave 2 controlling relevant Wave 1 outcomes.

At Wave 3, by teacher report, ENRICH children (M = 11.52) had significantly better early literacy than Active Control children (M = 10.65), B = 0.88, 95% CI [0.36, 1.40], p = 0.001, d = 0.27, controlling Wave 1 composite language. The same was true controlling Wave 2 composite language: ENRICH (M = 11.76) and Active Control (M = 10.82), B = 0.94, 95% CI [0.46, 1.43], p < 0.001, d = 0.30 (see Figure 5).

At Wave 3, by teacher report, ENRICH children (M = 3.55) had significantly better social skills than Active Control children (M = 3.45), B = 0.10, 95% CI [0.02, 0.19], p = 0.016, d = 0.21 (see Figure 6).

Hypothesis 4

(exploratory). Moderator effects of ENRICH on oral language and self‐regulation.

For both parent and teacher reports, no significant moderations of ENRICH effects were found on oral language at Waves 2 or 3 by children's age (Table S14), initial language level (Table S16), bilingualism (Table S17), or ethnicity (Tables S18–S20), or on self‐regulation by any moderator (Tables S23–S31). Moderator effects on oral language did emerge, however, for three other covariates (maternal education, family SES, and sex) at either Wave 2 or 3 as described below.

Maternal Education : At Wave 2 only, maternal education moderated parent‐reported ENRICH effects on oral language controlling Wave 1 language, B = –0.09, 95% CI [–0.13, –0.05], p < 0.001 (Figure 7 and Table S21). To explore the interaction, maternal education was categorized as low, middle, or high. At low levels of maternal education, children in ENRICH (M = –0.02) showed higher language scores than those in Active Control (M = –0.29), B = 0.27, 95% CI [0.11, 0.43], p = 0.001, d = 0.43. However, no significant condition differences were observed at middle levels: ENRICH (M = –0.12), Active Control (M = –0.24), B = 0.12, 95% CI [–0.01, 0.24], p = 0.060, d = 0.19; or high levels: ENRICH (M = –0.17), Active Control (M = –0.11), B = –0.05, 95% CI [–0.16, 0.05], p = 0.296, d = 0.09.

FIGURE 7.

FIGURE 7

Open in a new tab

Boxplots showing Wave 2 parent‐reported composite language scores by condition (ENRICH vs. CONTROL), adjusted for all covariates and Wave 1 composite language. Horizontal bars within each box represent condition medians. Panels A and B show moderation by maternal education and SES, respectively, at low, middle, and high levels. Asterisks indicate significant condition differences (*p < 0.05, **p < 0.01, ***p < 0.001).

Family SES : Similarly, at Wave 2 only, SES significantly moderated parent‐reported ENRICH benefits on oral language, B = –0.20, 95% CI [–0.31, –0.10], p < 0.001, controlling Wave 1 (Figure 7 and Table S22). To explore the interaction, SES was categorized as low, middle, or high (Boven et al. 2022). Parents reported that children in ENRICH had significantly higher scores than children in Active Control in both low‐SES (Ms = –0.04 vs. –0.36), B = 0.32, 95% CI [0.06, 0.58], p = 0.017, d = 0.51, and mid‐SES groups (Ms = –0.09 vs. –0.25), B = 0.16, 95% CI [0.06, 0.27], p = 0.003, d = 0.26, but not in high‐SES groups (Ms = –0.16 vs. –0.11), B = –0.05, 95% CI [–0.15, 0.05], p = 0.293, d = 0.08.

Sex : Finally, by parent report, sex significantly moderated ENRICH effects on oral language at Wave 3 controlling Wave 2 (but not Wave 1) composite language, B = –0.88, 95% CI [–1.46, –0.30], p = 0.003. Moreover, by teacher report, sex significantly moderated ENRICH effects on oral language at Wave 3 controlling both Wave 2 and Wave 1 composite language, B =  0.98, 95% CI (0.18, 1.78), p = 0.017; B = 1.12, 95% CI (0.25, 2.00), p = 0.012 (Figure 8 and Table S15). However, follow‐up analyses revealed rather different patterns of moderation by parent versus teacher report as detailed below.

FIGURE 8.

FIGURE 8

Open in a new tab

Boxplots showing Wave 3 oral language scores by condition (ENRICH vs. CONTROL) and sex, adjusted for all covariates. Horizontal bars within each box represent condition medians. Panels A and B show moderation by sex for parent‐reported oral language controlling Wave 2 composite language; Panels C and D show moderation by sex for teacher‐reported oral language controlling Wave 2 composite language; and Panels E and F show moderation by sex for teacher‐reported oral language controlling Wave 1 composite language. Asterisks indicate significant condition differences (**p < 0.01, ***p < 0.001).

By parent report, at Wave 3 controlling Wave 2, girls (M = 17.3) showed significantly higher oral language than boys (M = 16.6) in ENRICH, B = 0.61, 95% CI [0.20, 1.02], p = 0.004, d = 0.22, but girls’ oral language (M = 16.5) was not significantly different from that of boys (M = 16.7) in Active Control, B = –0.27, 95% CI [–0.69, 0.15], p = 0.209, d = 0.10. Relatedly, girls in ENRICH produced significantly better oral language than girls in Active Control, B = 0.79, 95% CI [0.35, 1.24], p < 0.001, d = 0.28, whereas boys’ oral language did not differ across conditions, B = –0.08, 95% CI [–0.52, 0.36], p = 0.711, d = 0.03.

In contrast, by teacher report, at Wave 3 controlling Wave 2, girls (M = 14.0) did not show significantly better oral language than boys (M = 13.8) in ENRICH, B = 0.17, 95% CI [–0.40, 0.74], p = 0.565, d = 0.04, whereas in Active Control girls (M = 13.6) did show significantly better oral language than boys (M = 12.4), B = 1.15, 95% CI [0.57, 1.72], p < 0.001, d = 0.30. Relatedly, boys in ENRICH showed significantly better oral language than boys in Active Control, B = 1.37, 95% CI [0.69, 2.05], p < 0.001, d = 0.36, whereas girls in ENRICH did not differ significantly from girls in Active Control, B = 0.39, 95% CI [−0.30, 1.08], p = 0.269, d = 0.10.

Finally, the pattern of results for teacher report at Wave 3 controlling Wave 1 was the same as that just described controlling Wave 2. Girls (M = 13.5) did not show significantly better oral language than boys (M = 13.1) in ENRICH, B = 0.37, 95% CI [–0.26, 0.99], p = 0.250, d = 0.09, whereas in Active Control girls (M = 13.3) did show significantly better oral language than boys (M = 11.9), B = 1.49, 95% CI [0.86, 2.12], p < 0.001, d = 0.36. Relatedly, boys in ENRICH showed significantly better oral language than boys in Active Control, B = 1.29, 95% CI [0.50, 2.09], p = 0.002, d = 0.31, whereas girls in ENRICH did not differ significantly from girls in Active Control, B = 0.16, 95% CI [−0.65, 0.98], p = 0.695, d = 0.04.

4. Discussion

This large‐scale pre‐registered oral language preventive intervention (ENRICH) with toddlers attending ECE is the first, to our knowledge, to show benefits simultaneously for oral language, self‐regulation, early literacy, and social skills at this young age. The benefits of ENRICH took time to develop; with a few exceptions, they were not apparent until approximately 1.5 years after baseline (roughly age 3), for both parents’ (oral language) and teachers’ reports (oral language, early literacy, self‐regulation, social skills). The benefits for self‐regulation were not moderated by any sociodemographic factors. The oral language benefits (parent‐report only) emerged earlier for children whose mothers possessed lower levels of educational qualifications and those from low‐ and middle‐SES families. By age 3, teachers reported greater ENRICH benefits for boys’ than girls’ language, whereas parents reported greater ENRICH benefits for girls’ than boys’ language. Overall, however, our findings for children's oral language were fairly consistent across parent and teacher reports (see Figure 2).

Our findings and effect sizes for oral language are also in line with those from another large‐scale ECE intervention with toddlers in Denmark (Bleses et al. 2020, 2021). We Learn Together is a shorter 20‐week intervention with toddlers of a similar age from a broad range of SES and language backgrounds showing long‐term maintenance of oral language, numeracy, and socio‐emotional benefits 1 year later at age 3 (Bleses et al. 2024). Denmark's ECE curriculum is similar to the holistic, play‐based emphasis of the NZ national ECE curriculum (Ministry of Education 2017). In contrast to our intervention, however, We Learn Together did not produce self‐regulation benefits. These successful language interventions are in contrast to the 36‐week RECC intervention in the United States (Landry et al. 2014), which resulted in socio‐emotional but not oral language or early literacy benefits for low‐income children.

We do not know for certain why the findings of our NZ study and the Denmark study differed from those of the US study, but the following factors may be crucial. First, the NZ and Denmark studies had substantially greater power than the US study (sample sizes 3–4 times as large). That said, there were not even suggestive benefits for children's oral language or literacy in the US study. Second, and notably, the teacher:child ratios in the US study were much less favorable (1:12) than those for toddlers in ECE centers in NZ (1:5) or Denmark (1:3). Third, our study included a similar percentage of teachers with a bachelor's degree or higher (62% at baseline) to the Denmark study (60%), both of which were much higher than in the US intervention (8.4%). Fourth, both ENRICH and We Learn Together trained all teachers at a center, whereas the RECC study focused on only one classroom per center. Training at the center level may foster collaboration and support of new practices across the entire team. Finally, two key differences to RECC are that, like We Learn Together (Bleses et al. 2020), ENRICH has a clear scope focused on oral language, and ENRICH is sequenced, with more challenging books and resources provided for teachers at the second phase as children age and their language skills increase.

In an observational sub‐study examining implementation of the new practices (n = 24 centers), we found increases in teachers’ language quality and children's language use during book‐reading and diaper‐change routines in 12 ENRICH centers relative to 12 Active Control centers 1 year after baseline (Video Project sub‐study; Swearingen et al. 2026; Zhang et al. 2026). Alongside the success of We Learn Together, these patterns suggest that improving children's oral language skills may require a sequenced program encouraging 1:1 or small‐group interactions delivered by teachers who possess stronger educational qualifications.

Our findings for oral language are also in line with studies documenting benefits of parents’ interactive shared book‐reading with toddlers (Dowdall et al. 2020) and of other contingent conversations, such as reminiscing about toddlers’ personal experiences (Reese and Newcombe 2007). At least one of these interaction‐focused parenting interventions, smalltalk in Australia, also found long‐term benefits for children's later self‐regulation in primary school (Bennett et al. 2025).

Our oral language findings held across children's ethnicity, initial language level, and bilingual status, yet were apparent earlier for children from lower‐ and middle‐SES families and with mothers of lower educational qualifications according to parent reports. These findings have practical, policy‐relevant implications for equity issues in children's language and literacy skills as a function of socioeconomic status (Phair 2021).

We note the somewhat mixed findings for oral language benefits of ENRICH as a function of child sex and informant (teachers vs. parents). According to teachers at Wave 3, boys benefitted relatively more from ENRICH than did girls. ENRICH narrowed the language gap between boys and girls that was present in the Active Control condition and in the complete sample at baseline (Bakir‐Demir et al. 2026). This moderation was present for teacher reports whether controlling children's oral language scores at Wave 2 or at Wave 1 (baseline). In contrast, at Wave 3 parents perceived greater benefits of ENRICH for girls than boys, although only when controlling language at Wave 2 and not at Wave 1. These differing sex‐moderation effects by informant are not yet understood and will need to be replicated.

Taken together, the benefits of ENRICH are consistent with our theory of self‐regulation skills being potentiated by earlier oral language skills (Reese et al. 2023; Salmon et al. 2016). Oral language skills enable toddlers to express their emotions, desires, and intentions through words rather than actions; oral language skills also sharpen cognitive skills critical to self‐regulation (Thériault‐Couture et al. 2025; Bruce and Bell 2022). Indeed, our strongest findings were for children's self‐regulation according to teacher reports at Wave 3. Thus, children's cognitive, behavioral, and emotional skills may improve through advances in their oral language and communication skills. However, we will need to conduct fine‐grained analyses with future waves of the study to ascertain the degree to which increases in self‐regulation are mediated by oral language. An alternative, complementary explanation is that our emphasis on teachers’ sensitive back‐and‐forth interactions may have had direct effects on children's attention span (Masek et al. 2021), which then supported their oral language and self‐regulation. The somewhat stronger effects we found for self‐regulation compared to oral language suggest that both mechanisms could be working simultaneously.

We propose that the primary benefits of ENRICH accrue via responsive turn‐taking, especially during interactive book‐reading and other small‐group and 1:1 conversations (Swearingen et al. 2026; Zhang et al. 2026). Similar to We Learn Together, ENRICH is an unscripted, flexible intervention in which teachers are encouraged to develop their own conversational routines and adapt them to children with different needs (Bleses et al. 2018, 2020). Even the conversation‐starter prompts in our ENRICH books are posed to teachers as examples that they are encouraged to adapt and extend for individual children. Using a model of adaptive, principled intervention design, our aim was to support teachers’ autonomy in their practice.

Unlike We Learn Together, however, we specifically coached teachers to engage in 1:1 conversations about children's personal experiences, using books as a springboard for reminiscing about experiences related to book themes (Riordan et al. 2022). In parent interventions with young children, coaching in elaborative reminiscing benefits children's oral language, autobiographical memory, self‐regulation, and socioemotional skills into adolescence and young adulthood (Salmon and Reese 2016).

Importantly, we achieved meaningful, broad benefits for toddlers given a relatively light‐touch intervention that was relatively low‐cost, easily delivered, and that teachers were able to implement despite ongoing disruptions related to the COVID‐19 pandemic across 2021–2023. Indeed, these disruptions could have contributed to the delayed impact of ENRICH, with few benefits apparent by Wave 2 and more extensive benefits by Wave 3. Although the effects took time to develop, they were medium to large for an educational intervention (Kraft 2020). These benefits may to lead to cascading differences over the preschool phase of the study, especially with the introduction of an advanced version of ENRICH and/or the self‐regulation intervention ENGAGE for children aged 3–5 years (Healey et al. 2022; Funder and Ozer 2019; Götz et al. 2021).

Limitations include a reliance on parent and teacher reports and not direct assessments of children's outcomes. However, past research shows that parent and teacher reports provide different perspectives and are moderately empirically convergent with observational and researcher‐administered measures (Fenson et al. 1994; Stucke and Doebel 2024). Consistent with our findings of stronger effects for teacher‐ than parent‐report, a new meta‐analysis shows that teacher reports of self‐regulation are stronger predictors of children's later behavior than parent reports (Stucke and Doebel 2024).

We also note that teachers in our study were not blind to their center's condition. In this open‐label trial, both teachers and parents were aware of the aims of the study, and teachers were aware of the ENRICH techniques. However, if teacher bias were driving these effects, we would have expected to see ENRICH benefits for teacher reports at Wave 2 but did not. Moreover, the pattern of findings for reports by teachers was similar to that of reports by parents, who were much less involved in the study. Finally, our informant reports are corroborated by filmed observations in the subset of 24 centers in the Video Project showing improved quality of teacher‐child interactions 1 year after baseline in ENRICH centers compared to Active Control centers (Swearingen et al. 2026; Zhang et al. 2026).

A further limitation is that, because the government childcare subsidy was unavailable for toddlers under age 3, all families had to pay for childcare. Hence, we were unable to recruit as many low‐income families as would be optimal. Nonetheless, the full range of SES was represented and, as noted earlier, we found stronger benefits for lower and middle SES families. In addition, in terms of generalizability, our ethnicity distribution was roughly representative of NZ toddlers nationally (Poulton et al. 2025; Stats NZ 2024).

Although our current focus was primarily on NZ English, ENRICH also encouraged teachers to use te reo Māori in line with the national languages of Aotearoa New Zealand and the national early childhood curriculum (Ministry of Education 2017; Mitchell et al. 2025). We will assess te reo Māori outcomes for the main trial in future analyses. Over the first year of the study, however, our Video Project sub‐study showed similar increases in teacher and child te reo Māori use for both ENRICH and Active Control centers (Mitchell et al. 2025).

In other future analyses with the current sample, we will assess implementation effects to determine minimum levels of ENRICH needed to achieve benefits. We will also conduct longitudinal mediation analyses of preschoolers’ self‐regulation via oral language skills. At ages 3–5, the study children are receiving the advanced ENRICH+ and/or ENGAGE or Active Control to test the separate and combined benefits of oral language and self‐regulation interventions for oral language, literacy, mathematics, self‐regulation, and social skills. In addition, with the subsample of 200+ children participating in the Brain and Behavior Development sub‐study, we will conduct longitudinal analyses of the impact of ENRICH and ENGAGE on brain development and researcher‐administered oral language, literacy, and self‐regulation tasks (Arnett et al. 2024).

Finally, all centers in our study were from a single provider of early childhood education with strong and stable leadership. In future research, we hope to replicate and extend these effects to a diverse range of early childhood education services. In an exciting development, the government of NZ recently announced in its 2025 budget that it planned to provide funding to scale up the ENRICH intervention to over 500 new early childhood centers over the next 4 years (New Zealand Treasury 2025).

The theoretical implications of this work are that supporting oral language at a young age can both generate notable language and literacy gains and spill over to enrich children's self‐regulation and socioemotional skills. The practical implications of intervening at the toddler period in ECEs are substantial for children's later reading, mathematics, academic achievement, and self‐regulation, alongside prevention of speech/language delays and behavioral difficulties. Ideally, these efforts will co‐occur with parent initiatives to boost shared book‐reading and back‐and‐forth conversations at home, alongside encouraging safe levels of high‐quality screen time (Barr et al. 2024; Brushe et al. 2024). Moreover, even small benefits for low‐income children's oral language and self‐regulation during early childhood could result in substantial societal and economic benefits across the life course (Phair 2021; Melhuish et al. 2015).

Conflicts of Interest

We have no conflicts of interest to disclose.

Supporting information

Supporting File 1: desc70198‐sup‐0001‐SuppMat.docx

DESC-29-e70198-s001.docx (1.1MB, docx)

Acknowledgments

This project was funded with a grant from the Wright Family Foundation to the Methodist Mission Southern. We thank staff and families at participating BestStart Educare centers, our Cultural Advisory Group, Julia Errington‐Scott, and all who helped with main trial data entry for Waves 1–3: Isabelle Swearingen, Yvonne Mitchell, Daniel Pushenko, Holly Stewart, and Eddy Grant. We are especially grateful to Professor Richie Poulton, who founded the study and who inspired us to continue the work after his death in September 2023.

Open Access publishing facilitated by University of Otago, as part of the Wiley ‐ University of Otago agreement via the Council of Australasian University Librarians

ENDNOTE

1

All means reported in the text are adjusted for covariates; in contrast, tabular means are raw means.

Data Availability Statement

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy and ethical restrictions. Full analytic syntax/code in SPSS/R is available: https://osf.io/p7ex2/files/rmvsh and https://osf.io/p7ex2/files/a9ukn.

References

  1. Anderson, N. J. , Graham S. A., Prime H., Jenkins J. M., and Madigan S.. 2021. “Linking Quality and Quantity of Parental Linguistic Input to Child Language Skills: A Metaanalysis.” Child Development 92: 484–501. 10.1111/cdev.13508. [DOI] [PubMed] [Google Scholar]
  2. Arnett, A. B. , Guiney H., Bakir‐Demir T., et al. 2024. “Resting EEG Correlates of Neurodevelopment in a Socioeconomically and Linguistically Diverse Sample of Toddlers: Wave 1 of the Kia Tīmata Pai Best Start New Zealand Study.” Developmental Cognitive Neuroscience 65: 101336. 10.1016/j.dcn.2023.101336. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Bakir‐Demir, T. , Marshall S., Guiney H., et al. 2026. “Kia Tīmata Pai: Parent‐ and Teacher‐Reported Oral Language and Self‐Regulation in a Large Diverse Sample of New Zealand Toddlers.” Early Childhood Research Quarterly 76: 378–390. [Google Scholar]
  4. Barr, R. , Kirkorian H., Coyne S., and Radesky J.. 2024. Early Childhood and Digital Media. Cambridge University Press. 10.1017/9781108885751. [DOI] [Google Scholar]
  5. Bennett, C. , Westrupp E. M., Bennetts S. K., et al. 2025. “An Early Parenting Intervention Focused on Enriched Parent‐child Interactions Improves Effortful Control in the Early Years of School.” Child Development 96: 355–374. 10.1111/cdev.14166. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Bird, A. , E. Reese, E. Schaughency, et al. 2024. “Talking, Praising and Teaching: How Parent Interaction During a Learning Task Relates to Children's Early Learning.” Early Childhood Research Quarterly 66: 255–268. 10.1016/j.ecresq.2023.10.001. [DOI] [Google Scholar]
  7. Bleses, D. , A. Højen, L. M. Justice, et al. 2018. “The Effectiveness of a Large‐Scale Language and Preliteracy Intervention: The SPELL Randomized Controlled Trial in Denmark.” Child Development 89: e342–e363. 10.1111/cdev.12859. [DOI] [PubMed] [Google Scholar]
  8. Bleses, D. , Jensen P., Højen A., et al. 2024. “Examining the Effects of an Infant‐Toddler School Readiness Intervention in Center‐ and Family‐Based Programs: Are Results Generalizable?” Early Childhood Research Quarterly 67: 252–264. 10.1016/j.ecresq.2024.01.001. [DOI] [Google Scholar]
  9. Bleses, D. , Jensen P., Højen A., Slot P., and Justice L.. 2021. “Implementing Toddler Interventions at Scale: The Case of “We Learn Together”.” Early Childhood Research Quarterly 57: 12–26. 10.1016/j.ecresq.2021.04.008. [DOI] [Google Scholar]
  10. Bleses, D. , Jensen P., Slot P., and Justice L.. 2020. “Low‐Cost Teacher‐Implemented Intervention Improves Toddlers' Language and Math Skills.” Early Childhood Research Quarterly 53: 64–76. 10.1016/j.ecresq.2020.03.001. [DOI] [Google Scholar]
  11. Bornstein, M. H. , Hahn C. S., Putnick D. L., and Pearson R. M.. 2018. “Stability of Core Language Skill From Infancy to Adolescence in Typical and Atypical Development.” Science Advances 4: eaat7422. 10.1126/sciadv.aat7422. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Boven, N. , Shackleton N., Bolton L., Sporle A., and Milne B. J.. 2022. New Zealand Socio‐Economic Index 2018. COMPASS Research Center, University of Auckland. [Google Scholar]
  13. Bronson, M. B. , Goodson B. D., Layzer J. I., and Love J. M.. 1990. Child Behavior Rating Scale. ABT Associates. [Google Scholar]
  14. Bruce, M. , and Bell M. A.. 2022. “Vocabulary and Executive Functioning: A Scoping Review of the Unidirectional and Bidirectional Associations Across Early Childhood.” Human Development 66: 167–187. 10.1159/000524964. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Brushe, M. E. , Haag D. G., Melhuish E. C., Reilly S., and Gregory T.. 2024. “Screen Time and Parent–Child Talk When Children Are Aged 12 to 36 Months.” JAMA Pediatrics 178: 369–375. 10.1001/jamapediatrics.2023.6790. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Clifford, A. E. , Schaughency E., Das S., Riordan J., Carroll J. L., and Reese E.. 2024. “Tender Shoots: Effects of a Preschool Shared Reading and Reminiscing Initiative on Parent‐Child Interactions and for Socio‐Emotional and Self‐Regulation Outcomes After School Entry.” Learning and Individual Differences 112: 102443. 10.1016/j.lindif.2024.102443. [DOI] [Google Scholar]
  17. Cunningham, J. E. , Chow J. C., Meeker K. A., et al. 2023. “A Conceptual Model for a Blended Intervention Approach to Support Early Language and Socialemotional Development in Toddler Classrooms.” Infants Young Child 36: 53–73. 10.1097/IYC.0000000000000232. [DOI] [Google Scholar]
  18. Dickinson, D. K. , Hofer K. G., Barnes E. M., and Grifenhagen J. F.. 2014. “Examining Teachers' Language in Head Start Classrooms From a Systemic Linguistics Approach.” Early Childhood Research Quarterly 29: 231–244. 10.1016/j.ecresq.2014.02.006. [DOI] [Google Scholar]
  19. Dickinson, D. K. , McCabe A., and Sprague K.. 2003. “Teacher Rating of Oral Language and Literacy (TROLL): Individualizing Early Literacy Instruction With a Standards‐Based Rating Tool.” Reading Teacher 56: 554–564. http://www.jstor.org/stable/20205246. [Google Scholar]
  20. Didham, R. 2005. Understanding and Working With Ethnicity Data: A Technical Paper . Statistics New Zealand. 10.13140/RG.2.1.1347.0804. [DOI] [Google Scholar]
  21. Dowdall, N. , Melendez‐Torres G. J., Murray L., et al. 2020. “Shared Picture Book Reading Interventions for Child Language Development: A Systematic Review and Meta‐Analysis.” Child Development 91: e383–e399. 10.1111/cdev.13225. [DOI] [PubMed] [Google Scholar]
  22. Fenson, L. , Dale P. S., Reznick J. S., et al. 1994. “Variability in Early Communicative Development.” Monographs of the Society for Research in Child Development 59: 1–185. 10.2307/1166093. [DOI] [PubMed] [Google Scholar]
  23. Fisher, P. A. , Frenkel T. I., Noll L. K., Berry M., and Yockelson M.. 2016. “Promoting Healthy Child Development via a Two‐Generation Translational Neuroscience Framework: The Filming Interactions to Nurture Development Video Coaching Program.” Child Development Perspectives 10: 251–256. 10.1111/cdep.12195. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Funder, D. C. , and Ozer D. J.. 2019. “Evaluating Effect Size in Psychological Research: Sense and Nonsense.” Advances in Methods and Practices in Psychological Science 2: 156–168. 10.1177/2515245919847202. [DOI] [Google Scholar]
  25. Giacalone, M. , Agata Z., Cozzucoli P. C., and Alibrandi A.. 2018. “Bonferroni–Holm and Permutation Tests to Compare Health Data: Methodological and Applicative Issues.” BMC Medical Research Methodology 18: 81. 10.1186/s12874-018-0540-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Gilkerson, J. , Richards J. A., Warren S. F., et al. 2018. “Language Experience in the Second Year of Life and Language Outcomes in Late Childhood.” Pediatrics 142: e20174276. 10.1542/peds.2017-4276. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Götz, F. M. , Gosling S. D., and Rentfrow P. J.. 2021. “Small Effects: The Indispensable Foundation for a Cumulative Psychological Science.” Perspectives on Psychological Science 17: 205–215. 10.1177/1745691620984483. [DOI] [PubMed] [Google Scholar]
  28. Harris, P. A. , Taylor R., Thielke R., et al. 2009. “Research Electronic Data Capture (REDCap)—A Metadata‐Driven Methodology and Workflow Process for Providing Translational Research Informatics Support.” Journal of Biomedical Informatics 42: 377–381. 10.1016/j.jbi.2008.08.010. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Harris, P. A. , Taylor R., Minor B. L., et al. 2019. “The REDCap Consortium: Building an International Community of Software Platform Partners.” Journal of Biomedical Informatics 95: 103208. 10.1016/j.jbi.2019.103208. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Hart, B. , and Risley T. R.. 1995. Meaningful Differences in the Everyday Experience of Young American Children. Paul H. Brookes Publishing. [Google Scholar]
  31. Healey, D. , Milne B., and Healey M.. 2022. “Adaption and Implementation of the Engage Programme Within the Early Childhood Curriculum.” Scientific Reports 12, no. 1: 21580. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Hirsh‐Pasek, K. , Adamson L. B., Bakeman R., et al. 2015. “The Contribution of Early Communication Quality to Low‐Income Children's Language Success.” Psychological Science 26: 1071–1083. 10.1177/0956797615581493. [DOI] [PubMed] [Google Scholar]
  33. IBM Corp . 2022. IBM SPSS Statistics for Windows, Version 29.0. IBM Corp. [Google Scholar]
  34. Jeong, J. , Kim R., and Subramanian S. V.. 2018. “How Consistent Are Associations Between Maternal and Paternal Education and Child Growth and Development Outcomes Across 39 Low‐Income and Middle‐Income Countries?” Journal of Epidemiology and Community Health 72: 434–441. 10.1136/jech-2017-210102. [DOI] [PubMed] [Google Scholar]
  35. Kraft, M. A. 2020. “Interpreting Effect Sizes of Education Interventions.” Educational Researcher 49: 241–253. 10.3102/0013189x20912798. [DOI] [Google Scholar]
  36. Landry, S. H. , Zucker T. A., Taylor H. B., et al. 2014. “Enhancing Early Childcare Quality and Learning for Toddlers at Risk: The Responsive Early Childhood Program.” Developmental Psychology 50: 526–541. 10.1037/a0033494. [DOI] [PubMed] [Google Scholar]
  37. Landry, S. H. , Smith K. E., Swank P. R., and Guttentag C.. 2008. “A Responsive Parenting Intervention: The Optimal Timing Across Early Childhood for Impacting Maternal Behaviors and Child Outcomes.” Developmental Psychology 44: 1335–1353. 10.1037/a0013030. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Lonigan, C. J. , and Shanahan T.. 2010. “Developing Early Literacy Skills: Things We Know We Know and Things We Know We Don't Know.” Educational Researcher 39: 340–346. 10.3102/0013189x10369832. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Ludbrook, J. 2000. “Multiple Inferences Using Confidence Intervals.” Clinical and Experimental Pharmacology & Physiology 27: 212–215. 10.1046/j.1440-1681.2000.03223.x. [DOI] [PubMed] [Google Scholar]
  40. Masek, L. R. , McMillan B. T. M., Paterson S. J., et al. 2021. “Where Language Meets Attention: How Contingent Interactions Promote Learning.” Developmental Review 60: 100961. 10.1016/j.dr.2021.100961. [DOI] [Google Scholar]
  41. Masek, L. R. , Weiss S. M., McMillan B. T. M., et al. 2022. “Contingent Conversations Build More Than Language: How Communicative Interactions in Toddlerhood Relate to Preschool Executive Function Skills.” Developmental Science 26: e13338. 10.1111/desc.13338. [DOI] [PubMed] [Google Scholar]
  42. Melhuish, E. , Ereky‐Stevens K., Petrogiannis K., et al. 2015. A Review of Research on the Effects of Early Childhood Education and Care (ECEC) on Child Development . CARE—European Early Childhood Education and Care. [Google Scholar]
  43. Ministry of Education . 2017. Te Whāriki: He Whāriki Mātauranga mō ngā Mokopuna o Aotearoa: Early Childhood Curriculum. Ministry of Education. [Google Scholar]
  44. Mitchell, Y. A. , Thomas B.‐A., Clifford A., et al. 2025. Kia Tīmata Pai Video Project: Observed Te Reo Māori Usage Within Early Childhood Education in Aotearoa New Zealand Following Professional Development . Manuscript under review.
  45. Moffitt, T. E. , Arseneault L., Belsky D., et al. 2011. “A Gradient of Childhood Self‐Control Predicts Health, Wealth, and Public Safety.” Proceedings of the National Academy of Sciences of the United States of America 108: 2693–2698. 10.1073/pnas.1010076108. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. New Zealand Treasury . 2025. Budget 2025: Summary of Initiatives. New Zealand Treasury. https://budget.govt.nz/budget/pdfs/summary‐initiatives/b25‐sum‐initiatives.pdf. [Google Scholar]
  47. Noble, C. , Sala G., Peter M., et al. 2019. “The Impact of Shared Book Reading on Children's Language Skills: A Meta‐Analysis.” Educational Research Review 28: 100290. 10.1016/j.edurev.2019.100290. [DOI] [Google Scholar]
  48. OECD . 2023. Education at a Glance 2023. OECD Publishing. [Google Scholar]
  49. Pace, A. , Alper R., Burchinal M. R., Golinkoff R. M., and Hirsh‐Pasek K.. 2019. “Measuring Success: Within‐ and Cross‐Domain Predictors of Academic and Social Trajectories in Elementary School.” Early Childhood Research Quarterly 46: 112–125. 10.1016/j.ecresq.2018.04.001. [DOI] [Google Scholar]
  50. Peterson, E. R. , Waldie K. E., Mohal J., et al. 2017. “Infant Behavior Questionnaire–Revised Very Short Form: A New Factor Structure's Associations With Parenting Perceptions and Child Language Outcomes.” Journal of Personality Assessment 99: 561–573. 10.1080/00223891.2017.1287709. [DOI] [PubMed] [Google Scholar]
  51. Phair, R. 2021. International Early Learning and Child Well‐Being Study Assessment Framework. No. 246. OECD Publishing. 10.1787/af403e1e-en. [DOI] [Google Scholar]
  52. Poulton, R. , Bakir‐Demir T., Guiney H., et al. 2025. “Cohort Profile: The New Zealand Best Start Study (Kia Tīmata Pai).” International Journal of Epidemiology 54, no. 4: dyaf103. 10.1093/ije/dyaf103. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Putnam, S. P. , Gartstein M. A., and Rothbart M. K.. 2006. “Measurement of Finegrained Aspects of Toddler Temperament: The Early Childhood Behavior Questionnaire.” Infant Behavior and Development 29: 386–401. 10.1016/j.infbeh.2006.01.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. R Core Team . 2026. R: A Language and Environment for Statistical Computing . R Foundation for Statistical Computing. https://www.R‐project.org/. [Google Scholar]
  55. Reese, E. , Barrett‐Young A., Gilkison L., et al. 2023. “Tender Shoots: A Parent Book‐Reading and Reminiscing Program to Enhance Children's Oral Narrative Skills.” Reading and Writing 36, no. 3: 541–564. 10.1007/s11145-022-10282-6. [DOI] [Google Scholar]
  56. Reese, E. , Keegan P., McNaughton S., et al. 2018. “Te Reo Māori: Indigenous Language Acquisition in the Context of New Zealand English.” Journal of Child Language 45: 340–367. 10.1017/S0305000917000241. [DOI] [PubMed] [Google Scholar]
  57. Reese, E. , Kokaua J., Guiney H., et al. 2023a. “Kia Tīmata Pai (Best Start): A Study Protocol for a Cluster Randomised Trial With Early Childhood Teachers to Support Children's Oral Language and Self‐Regulation Development.” BMJ Open 13, no. 9: e073361. 10.1136/bmjopen-2023-073361. [DOI] [PMC free article] [PubMed] [Google Scholar]
  58. Reese, E. , and Newcombe R.. 2007. “Training Mothers in Elaborative Reminiscing Enhances Children's Autobiographical Memory and Narrative.” Child Development 78: 1153–1170. 10.1111/j.1467-8624.2007.01058.x. [DOI] [PubMed] [Google Scholar]
  59. Riordan, J. , Reese E., Das S., Carroll J., and Schaughency E.. 2022. “Tender Shoots: A Randomized Controlled Trial of Two Shared‐Reading Approaches for Enhancing Parent‐Child Interactions and Children's Oral Language and Literacy Skills.” Scientific Studies of Reading 26, no. 3: 183–203. 10.1080/10888438.2021.1926464. [DOI] [Google Scholar]
  60. Robson, D. A. , Allen M. S., and Howard S. J.. 2020. “Self‐Regulation in Childhood as a Predictor of Future Outcomes: A Meta‐Analytic Review.” Psychological Bulletin 146: 324–354. 10.1037/bul0000227. [DOI] [PubMed] [Google Scholar]
  61. Rowe, M. L. , and Snow C. E.. 2020. “Analyzing Input Quality Along Three Dimensions: Interactive, Linguistic, and Conceptual.” Journal of Child Language 47: 5–21. 10.1017/S0305000919000655. [DOI] [PubMed] [Google Scholar]
  62. Salmon, K. , O'Kearney R., Reese E., and Fortune C. A.. 2016. “The Role of Language Skill in Child Psychopathology: Implications for Intervention in the Early Years.” Clinical Child and Family Psychology Review 19: 352–367. 10.1007/s10567-016-0214-1. [DOI] [PubMed] [Google Scholar]
  63. Salmon, K. , and Reese E.. 2016. “The Benefits of Reminiscing With Young Children.” Current Directions in Psychological Science 25: 233–238. 10.1177/0963721416655100. [DOI] [Google Scholar]
  64. Schaughency, E. , Linney K., Carroll J., Das S., Riordan J., and Reese E.. 2023. “Tender Shoots: A Parent‐Mediated Randomized Controlled Trial With Preschool Children Benefits Beginning Reading 1 Year Later.” Reading Research Quarterly 58, no. 3: 450–470. 10.1002/rrq.500. [DOI] [Google Scholar]
  65. Schoon, I. , Parsons S., Rush R., and Law J.. 2010. “Children's Language Ability and Psychosocial Development: A 29 Year Follow‐Up Study.” Pediatrics 126: e73–e80. 10.1542/peds.2009-3282. [DOI] [PubMed] [Google Scholar]
  66. Shonkoff, J. P. , and Bales S. N.. 2011. “Science Does Not Speak for Itself: Translating Child Development Research for the Public and Its Policymakers.” Child Development 82: 17–32. 10.1111/j.1467-8624.2010.01538.x. [DOI] [PubMed] [Google Scholar]
  67. Stats NZ . 2024. 2023 Census Population Counts by Ethnic Group, Age, and Māori Descent, and Dwelling Counts . Stats NZ. [Google Scholar]
  68. Stucke, N. J. , and Doebel S.. 2024. “Early Childhood Executive Function Predicts Concurrent and Later Social and Behavioral Outcomes: A Review and Meta‐Analysis.” Psychological Bulletin 150: 1178–1206. 10.1037/bul0000445. [DOI] [PubMed] [Google Scholar]
  69. Swearingen, I. , Van Bergen P., Schaughency E., Moros S., and Reese E.. 2026. “Kia Tīmata Pai (Best Start) Video Project: Effects of an Oral Language Professional Development Program (ENRICH) on Educator‐Toddler Conversations.” Early Childhood Research Quarterly 76: 261–271. 10.1016/j.ecresq.2026.02.003. [DOI] [Google Scholar]
  70. Taumoepeau, M. , and Ruffman T.. 2008. “Stepping Stones to Others' Minds: Maternal Talk Relates to Child Mental State Language and Emotion Understanding at 15, 24, and 33 Months.” Child Development 79: 284–302. 10.1111/j.1467-8624.2007.01126.x. [DOI] [PubMed] [Google Scholar]
  71. Thériault‐Couture, F. , MatteGagné C., and Bernier A.. 2025. “Child Vocabulary and Developmental Growth in Executive Functions During Toddlerhood.” Developmental Science 28: e70010. 10.1111/desc.70010. [DOI] [PMC free article] [PubMed] [Google Scholar]
  72. Zhang, Y. , Bakir‐Demir T., Taumoepeau M., Salmon K., and Reese E.. 2026. “Kia Tīmata Pai Video Project: Impact of an Oral Language Intervention (ENRICH) on Linguistic Aspects of Educator‐Toddler Interactions.” Teaching and Teacher Education 179: 105596. 10.1016/j.tate.2026.105596. [DOI] [Google Scholar]

Associated Data

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

Supplementary Materials

Supporting File 1: desc70198‐sup‐0001‐SuppMat.docx

DESC-29-e70198-s001.docx (1.1MB, docx)

Data Availability Statement

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy and ethical restrictions. Full analytic syntax/code in SPSS/R is available: https://osf.io/p7ex2/files/rmvsh and https://osf.io/p7ex2/files/a9ukn.

Articles from Developmental Science are provided here courtesy of Wiley

RESOURCES