Differential effects of gender and ethnicity on children’s receptive language scores after a Canadian two-generation preschool program: follow-up to adolescence

ABSTRACT

Preschool programs for socially vulnerable children are believed to affect school readiness and language development but infrequently include follow-up to adolescence; observational measurement of receptive vocabulary is rare. The purpose of this longitudinal cohort study (55 children and 41 parents) was to report the trajectory of receptive vocabulary development for socially vulnerable children of diverse ethnicities (Indigenous, other Canadian-born, and Immigrant) who participated in a two-generation preschool program. English receptive vocabulary scores were measured at 5-time points: (a) program intake, (b) program exit, (c) age 7 years, (d) age 10 years, and (e) adolescence, using the Peabody Picture Vocabulary Test – Third Edition (PPVT-III). For all children, PPVT-III scores increased the most between program intake and exit and positive changes were sustained until adolescence. When the sample was examined by gender, PPVT-III scores were higher for boys than girls at all time points. When the sample was examined by ethnicity, other Canadian-born children scored higher than Indigenous and Immigrant children at all time points. Immigrant children scored lowest until age 7 years, and at age 10 years and adolescence scored higher than Indigenous children. Using mixed-effects modeling, PPVT-III scores increased when English was the primary language spoken at home at intake. PPVT-III scores increased slightly as caregiver age and caregiver Adverse Childhood Experiences score increased. PPVT-III scores were lower for Immigrant girls and Indigenous boys than other sub-groups of children. Differential effects of the two-generation preschool program by gender and ethnicity suggest tailoring programming to increase equitability of receptive vocabulary development for immigrant girls and Indigenous boys. Children with intergenerational adversity may require additional support.

KEYWORDS:

• Social vulnerability
• two-generation preschool program
• receptive vocabulary
• ethnicity
• gender

Introduction

Living with social vulnerability in early childhood (ages 3–5 years) is associated with poor outcomes in late adolescence (ages 15–19 years) (Huber et al., 2019; Johnson et al., 2022). Socially vulnerable families have minimal resources and barriers to effective family functioning and parenting (Perry et al., 2022). For this study, we defined social vulnerability as follows: (a) single-parent household; (b) mother with less than high-school education; and (c) low income (Statistics Canada, 2002). Children living with low income are at greater risk of poor outcomes (Cooper & Stewart, 2021; Kim et al., 2022). The strength of this association increases with the duration of low income (Hackman et al., 2010; McEwen & Stewart, 2014). The effects of low income on children’s school readiness persist over time and across cultures (Kim et al., 2022). School readiness is a transitional process including (a) children (physical well-being, motor development, social and emotional development, language, and cognition), (b) schools (bridging gaps between home and school), and (c) families (supportive and responsive parenting) (Britto, 2012).

Mechanisms underlying school readiness

Life course studies have postulated mechanisms through which adversity in early childhood exerts effects on school readiness (Britto, 2012), academic success (Kim et al., 2022; Pagani & Fitzpatrick, 2014; Romano et al., 2010), and employment (Parsons et al., 2011). Historically, there was a conceptual understanding of school readiness as an individual child characteristic (Janus & Gaskin, 2014). On most measures of school readiness (Britto, 2012), preschool girls score better than boys (Brandlistuen et al., 2021; Del Boca et al., 2019). There is increasing recognition of children’s environments (Britto et al., 2017) with responsive parenting, warmth and emotional availability, non-harsh discipline, and/or positive behavioral support contributing to school readiness (Kim et al., 2022; Madigan et al., 2019; Prime et al., 2021; Valcan et al., 2018). Responsive parenting encourages reciprocal engagement in parent–child interactions (Tamis LeMonda et al., 2014). Reciprocal engagement and positivity within interactions are key to supporting children’s learning (Carpendale & Lewis, 2004); children require contingent social interactions to learn vocabulary (Rowe & Weisleder, 2020). Socioeconomic status partially explains differences in vocabulary (Friesen et al., 2021). Thus, interventions to address health and social disparities in adulthood must be directed at improving environmental support for young children (Braveman & Barclay, 2009).

Two-generation preschool programs and school readiness

The effects of many early intervention programs are greatest during participation and diminish or disappear over time (Bailey et al., 2020). Two-generation programs that include early intervention and support for parents may affect receptive vocabulary (RV) development Benzies et al., 2010, which is a proxy for school readiness (Ryan et al., 2013). Despite a two-generation preschool intervention, RV scores, as measured by direct observation of behaviours (Li et al., 2019), lagged in sub-groups of children. Immigrant girls started the program with the lowest RV scores and scored 10 points below the norm at age 10 years (Mughal et al., 2016) is unclear if these differences continued into adolescence. There is limited research about family factors, contributing to school readiness and sustained improvements in RV scores to late adolescence. The objective of this study was to explore trajectories of RV in sub-groups of socially vulnerable children between intake to a two-generation preschool program and adolescence. Our research questions were:

1. What are the trajectories of RV scores between program intake and adolescence for boys and girls? We hypothesized that RV trajectories would be steepest between intake and exit and flatten after age 7 years. Based on our previous research, we hypothesized the elevation of the trajectory of RV scores would be higher for boys than girls.

2. Do trajectories of RV scores differ by ethnicity? We hypothesized that immigrant children would have lower RV scores at intake than Indigenous and other Canadian-born children and their scores would continue to improve into adolescence.

3. What child and family characteristics are related to RV scores at each time point? We hypothesized that RV scores would be inter-correlated across time points and that social vulnerability would be correlated with lower RV scores.

4. What are the best predictors of RV scores at adolescence? We hypothesized that immigrant girls would have lower RV scores than Indigenous and other Canadian-born girls and all boys. We expected that living in a family with greater adversity would be associated with lower RV scores.

Theoretical framework

Vygotsky’s Zone of Proximal Development (Vygotsky et al., 1978), and Bronfenbrenner’s Bioecological Systems Theory (Bronfenbrenner, 2005), provided the theoretical framework. Vygotsky et al. (1978) postulated children’s development is best supported through social interactions with individuals who have more knowledge and expertise. Bronfenbrenner (2005) postulated a nested system of interactions within children’s environments with positive early childhood experiences in the family and two-generation preschool programs supporting development. Aligned with behavioral neuroscience approaches recommended by the Harvard Center on the Developing Child (Shonkoff & Fisher, 2013), high quality, two-generation preschool programs should mitigate the negative influence of low-income and a sub-optimal family environment and improve RV scores.

Materials and methods

Study design and setting

We conducted this longitudinal cohort study between January 2002 and December 2018 at CUPS (formerly Calgary Urban Project Society). CUPS is a not-for-profit community agency that provides integrated programs and services to build resiliency in more than 11,000 socially vulnerable clients each year https://www.cupscalgary.com/. CUPS is located in a large western Canadian city (population 1.3 million at end of study; City of Calgary, 2018). Clients are ethnically diverse, including Canadian-born, Indigenous, and immigrant. Indigenous peoples have suffered, historical trauma, including reserve systems, residential schools, economic destruction and loss of resources, suppression of cultural identity, and loss of traditional ways of knowing, language, and land, and has resulted in cross-generational transmission of poor mental and physical illness, suicide, substance use, family violence, sexual violence, incarceration, and child maltreatment (Nutton & Fast, 2015). At the time of this study, there were more than 7.5 million immigrants in Canada, representing over one-fifth of Canada’s total population (Statistics Canada, 2016). The lack of culturally meaningful programs and services and systemic barriers exacerbated by immigration status affect access to early childhood development programming (Brown et al., 2020).

The University of Calgary, Conjoint Health Research Ethics Board approved the study. We applied Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) to cohort studies (von Elm et al., 2007).

Participants

Families of preschool children were referred to CUPS by other community agencies and word of mouth. Children could enroll between age 2.5 and 5 years; each family could enroll more than one child. Children were eligible to enroll in the program if they lived in a family with low income, parental mental illness, addiction, or poor social support. Criteria to remain in the program were as follows: (a) child attended regularly, (b) caregiver attended a 6-week parent education course, and (c) caregiver volunteered on the school bus. Children were eligible to participate in the study if they attended the program for 3 months, which ensured a minimum exposure to the early intervention and caregivers completed the parenting education course. We examined reasons for attrition (Ginn et al., 2017) and used Facebook (Mychasiuk & Benzies, 2012) and birthday cards to sustain engagement.

The two-generation program: CUPS

CUPS offers centre-based, early learning and care, parent education, and family support (Benzies et al., 2010). Children attended 4 days per week, 5 h per day. The curriculum built on children’s interests, with teacher/child ratios of 1:8. Children had access to on-site pediatrician and developmental specialists.

Measurement and procedures

See Table 1 for measures and data collection time points. PPVT-III scores were measured at 5-time points: (a) program intake, (b) program exit, (c) age 7 years, (d) age 10 years, and (e) adolescence. Data were collected at the preschool when children were enrolled in the program or during home visits for follow-ups. Caregivers completed questionnaires to provide demographic characteristics and adverse childhood experiences (ACE) scores. Caregiver consent was obtained at the program intake and affirmed at each time point; child assent was obtained at each time point after age 7 years. All participants accepted a $40 gift card to recognize their time.

Table 1. Description of measures.

Scale/Variable	Description
PPVT-III (Dunn & Dunn, 1997)	Observational screening assessment of English receptive vocabulary in individuals over the age of 2 years. Typically, 69 items depending on critical range. Standardized scoring with mean = 100 and standard deviation ±15; higher scores mean greater receptive vocabulary skill. Internal consistency reliability ranges from .61 to .88. PPVT-III scores are correlated with the McCarthy Scales of Children’s Abilities. Takes 10 to 15 minutes to complete. We used continuous scores measured at: (a) program intake, (b) program exit, (c) age 7 years, (d) age 10 years, and (e) late adolescence.
Time	Coded as 0 = program intake (reference category); 1 = program exit; 2 = age 7 years; 3 = age 10 years; 4 = late adolescence.
At Intake, caregiver-reported ethnicity	Coded as 1 = Indigenous; 2 = Canadian-born; 3 = Immigrant in last 10 years (reference category).
At Intake, caregiver age	Continuous variable.
At Intake, language spoken at home	Coded as 0 = not English; 1 = English.
At Intake, caregiver partnered	Coded as 0 = no; 1 = yes.
At Intake, child gender	Coded as 0 = boy; 1 = girl, as reported by caregiver.
At Intake, food insecurity	In the last 3 months, caregiver visited the food bank, ate a no cost meal, or received a food voucher coded as 0 = no; 1 = yes.
At Intake, caregiver involved with CWS for program child	Coded as 0 = no; 1 = yes.
At Time 4, caregiver Adverse Childhood Experiences (Felitti et al., 1998)	10-item self-report of exposures to abuse and neglect (psychological, physical, and sexual) and household dysfunction (substance abuse, mental health problems, violent treatment of mother, and criminal behaviour) before the respondent’s 18th birthday. The number of ‘yes’ responses are summed; higher scores indicate greater exposure to adversity in childhood. ACE scores ≥ 4 are associated with poor mental and physical health in adulthood (Anda et al., 2006; Felitti et al., 1998). Test-retest reliability kappa co-efficient ranged from .52 to .72 suggesting temporal stability of scores. The ACE Questionnaire takes 5 min to complete. Coded as 0, 1, 2, 3 and 4 + .
At late Adolescence, CWS file still open	Coded as 0 = no; 1 = yes.

PPVT-III = Peabody Picture Vocabulary Test – Third Edition. ACE = adverse childhood experiences. CWS = child welfare services.

Data analysis

To retain as many cases as possible, we used logical reasoning to replace missing values. For variables related to child welfare services (CWS), we replaced missing values for Indigenous and other Canadian-born groups with ‘1’ because all participants from those groups who responded were involved with CWS; we replaced missing values for Immigrants with 0 because no participants who responded were involved with CWS. Following similar reasoning, we replaced missing values in parental adverse ACE scores with the mean score for each group. To describe the characteristics of the sample and scale scores, we used descriptive statistics. To identify potential covariates for our model, we used Pearson’s correlations to estimate relationships between PPVT- III scores and child and family characteristics. To explore the association between PPVT- III scores and covariates, we used a mixed effect model. The dependent variable was mean PPVT-III scores at 5-time points, and the independent variables were parent age, partnered, English primary language at home, parent ACE score, parental involvement with CWS at intake, with ethnicity and child gender as an interaction term. We used StataIC (StataCorp 2015 Stata Statistical Software: Release 14. College Station, TX) with significance at p < .05.

Results

Participants

We recruited 79 caregivers of 134 preschool children; of these, we included 41 parents of 55 children at adolescence. Compared to those who dropped out, those who completed the adolescent follow-up were less likely to have visited the food bank within 3 months prior to intake or have a caregiver without a high-school diploma. See Table 2 for characteristics of children and caregivers at intake by ethnicity. At program intake, the average age of all parents was 30.54 (SD = 6.12) years. The average ages for Indigenous, other Canadian-born, and Immigrant caregivers were 26.33 (SD = 4.58), 31.29 (SD = 5.94), and 32.20 (SD = 6.28) years, respectively. At intake, the average age of all children was 45.15 (SD = 9.12) months. The average ages for Indigenous, other Canadian-born, and Immigrant children were 44.08 (SD = 8.90), 46.67 (SD = 9.49), and 44.27 (SD = 9.09) months, respectively.

Table 2. Characteristics of children and caregivers at program intake for full sample and by ethnicity

	Full Sample		Indigenous		Canadian		Immigrant
	N	Frequency (%)	n	Frequency (%)	n	Frequency (%)	n	Frequency (%)
Child
Gender (% boys)	55	26 (47.3)	12	6 (50.0)	21	9 (42.9)	22	11 (50.0)
Caregiver
Relationship status (% partnered)	40	22 (55.0)	8	3 (37.5)	17	7 (41.2)	15	12 (80.0)
Education (% with high school)	41	27 (65.9)	9	4 (44.4)	17	12 (70.6)	15	11 (73.3)
Language spoken at home (% English)	42	25 (59.5)	10	10 (100.0)	17	12 (70.6)	15	3 (20.0)
Food insecurity (% yes)	40	19 (47.5)	9	6 (66.7)	16	7 (43.8)	15	6 (40.0)
As child, involved with CWS (% yes)	40	11 (27.5)	9	7 (77.8)	17	4 (23.5)	14	0 (0.0)
As caregiver, involved with CWS (% yes)	40	13 (32.5)	9	5 (55.6)	17	8 (47.1)	14	0 (0.0)
If yes, file still open (% yes)	13	5 (38.5)	5	3 (60.0)	8	2 (25.0)	-	-

CWS = Child Welfare Services.

PPVT-III scores by time and gender

Children scored at or below the standardized norm (100 points) on the PPVT-III at all time points. See Figure 1 and supplementary table. For the full sample, PPVT-III scores showed the steepest increase between intake and exit, flattened after exit, and were relatively stable in adolescence, which supported our first hypothesis. PPVT-III scores differed by child gender, with boys scoring higher than girls at all time points. After exit, scores for boys decreased to age 10 years and increased slightly to adolescence. Similarly, scores for girls decreased slightly to age 7 years, but continued to increase slightly to adolescence. There were group differences in average PPVT-III scores by gender at all time points (p < .05), except at age 10 years (p > .05).

Figure 1. Average PPVT-III scores between intake and adolescence for all children and by child gender.

PPVT-III = Peabody Picture Vocabulary Test – Third Edition.

PPVT-III scores by time and ethnicity

There were group differences in children’s PPVT-III scores by ethnicity at intake, F(2,48) = 7.30, p =.002, exit, F(2,50) = 3.95, p =.026 and age 7 years F(2,41) = 4.26, p =.021, but not at age 10 years, F(2,38) = 2.86, p =.070, or adolescence, F(2,48) = 0.38, p =.685. See Figure 2 and supplementary table. These results partially supported our second hypothesis. In post hoc analyses, at intake both Indigenous and Immigrant children scored significantly lower than other Canadian-born children. At the exit, only Immigrant children scored lower than Indigenous and other Canadian-born children, and this pattern held at ages 7 and 10 years. At adolescence, there were no longer any group differences in PPVT-III scores by ethnicity.

Figure 2. Average PPVT-III scores between intake and adolescence for all children and by ethnicity.

PPVT-III = Peabody Picture Vocabulary Test – Third Edition.

Bivariate correlations

As expected, children’s PPVT-III scores were strongly, positively inter-correlated across all 5-time points. See Table 3. The strongest correlation was between intake and exit, r = 0.771, and the weakest between intake and adolescence, r = 0.349. Correlations were stronger between proximal time points. Except at adolescence, PPVT-III scores had moderate-to-large correlations with caregiver ACE scores.

Table 3. Bivariate correlations between average PPVT-III scores between intake and late adolescence and variables considered for inclusion in the final model

	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16
1. PPVT-III Intake	.771**	.687**	.583**	.349*	.389**	.299*	−.395**	.198	−.180	.434**	.244	−.023	.328*	.129	−.069
2. PPVT-III Exit	-	.753**	.741**	.608**	.502**	.246	−.452**	.212	−.143	.437**	.443**	−.036	.340*	−.149	−.238
3. PPVT-III Age 7		-	.725**	.567**	.686**	.181	−.525**	.229	−.198	.351*	.214	.080	.498**	−.091	−.208
4. PPVT-III Age 10			-	.714**	.497**	.460**	−.283	.336*	−.058	.180	.280	−.030	.298	.137	−.041
5. PPVT-III Adolescence				-	.281	.327*	−.357*	.102	.024	.252	.134	−.108	.205	−.041	−.130
6. Caregiver ACE					-	−.026	−.652**	.109	−.486**	−.505**	.390*	.531**	.488**	.059	−.142
7. Caregiver Age						-	−.068	.403**	.339*	−.384**	−.199	−.350*	−.117	.196	−.155
8. Caregiver Partnered							-	.023	.334*	−.443**	−.286*	−.242	−.666**	.021	.161
9. Highschool								-	.077	−.045	.158	.024	−.039	.249	−.015
10. Ethnicity									-	.654**	−.299*	−.664**	−.428**	−.017	−.013
11. English at home	.									-	.407**	.520**	−.445**	−.104	−.025
12. Food insecure											-	.320*	.274	−.050	−.216
13. CWS for Caregiver												-	.150	.010	.111
14. CWS for Child													-	.108	−.197
15. Child Age														-	−.029
16. Child Gender															-

PPVT-III = Peabody Picture Vocabulary Test – Third Edition. ACE = Adverse Childhood Experience score; CWS = Child Welfare Services. N = 55 children; N = 41 caregivers. *p < .05. **p < .01.

Except at exit and age 7 years, PPVT-III scores were moderately, positively correlated with caregiver age. Except at age 10 years, PPVT-III scores were moderately to highly and negatively correlated with their caregiver being partnered. Only at age 10 years, PPVT-III scores were moderately and positively correlated with caregiver completion of high school. There were no significant correlations between PPVT-III scores at any time point and ethnicity. PPVT-III scores were moderately, positively correlated with English the primary language spoken at home at intake, exit and age 7 years, but not at age 10 years or adolescence. Only at exit were PPVT-III scores moderately, positively correlated with food insecurity. Thus, PPVT- III scores were correlated with adversity, which partially supported our third hypothesis.

Association between PPVT-III scores and covariates

The likelihood ratio test was significant, χ² = 17.80, p < .001, indicating that significant correlations exist within children’s PPVT-III scores and a mixed effect model was preferred over a linear regression model. The intraclass correlation coefficient was 0.25. That is, 25% of the total variance in PPVT- III scores was accounted for by the clustering of repeated observations among each child. Table 4 shows that PPVT-III scores increased slightly as caregiver age and ACE scores increased. PPVT- III scores increased when English was the primary language spoken at home at intake. Finally, being an Immigrant girl or Indigenous boy was associated with decreased PPVT-III scores at adolescence, which partially supports our fourth hypothesis.

Table 4. Mixed-effects model showing association between the PPVT-III scores at adolescence and other variables

Variable	Coefficient	SE	p value	95% CI
Caregiver Age at Intake	0.99	0.23	<.001	0.54, 1.43
Caregiver Partnered at Intake
No	Reference
Yes	−1.83	3.62	.61	−8.93, 5.27
English primary language
No	Reference
Yes	9.40	3.26	.004	3.01, 15.79
Caregiver ACE score	0.97	0.48	.04	0.04, 1.90
CWS as Caregiver at Intake
No	Reference
Yes	6.09	3.80	.11	−1.36, 13.53
Culture x Child Gender
Immigrant x boy	Reference
Immigrant x girl	−7.15	3.51	.04	−14.02, −0.28
Indigenous x boy	−11.94	4.88	.01	−21.51, −2.38
Indigenous x girl	−8.15	4.98	.10	−17.91, 1.61
Other-Canadian born x boy	−6.56	4.02	.10	−14.44, 1.32
Other-Canadian born x girl	−4.59	3.64	.21	−11.72, 2.54

PPVT-III = Peabody Picture Vocabulary Test – Third Edition; ACE = Adverse Childhood Experience; CWS = Child Welfare Services. SE = Standard Error; CI = Confidence Interval.

Discussion

In this longitudinal cohort study of children who participated in a two-generation preschool program, we found that the combination of ethnicity and gender influenced English RV scores between intake and adolescence. Contrary to other studies that girls outperform boys with language development (Brandlistuen et al., 2021; Chilosi et al., 2023; Del Boca et al., 2019), in our study we found that boys consistently outperformed girls. By examining the interaction effect of ethnicity by gender over time, we were able to explain our novel result. Compared to other Canadian-born boys and girls, Indigenous girls, and Immigrant boys, RV scores decreased for Immigrant girls and Indigenous boys between intake and adolescence. In a systematic review of barriers to formal education for girls in India, Thamminaina et al. (2020) identified prioritization of household work over education, and perception of limited opportunities after school, but concluded the evidence is weak. It remains unclear if these barriers will impede the success of girls from other countries in preschool education programs after immigration to Canada. Our results about gender differences are consistent with a study of Immigrant children that reported being a girl from Hong Kong or Philippines and living a single parent family were significant predictors of lower academic performance (Oxman-Martinez et al., 2012). Not all Immigrant girls in our study were from South Asia. It may be that Immigrant girls require special programming about the equitability of gender roles to improve academic success in Canada (Meherali et al., 2021).

As hypothesized and consistent with the literature, overall improvements in RV scores were greatest between intake and exit (Schweinhart et al., 1993). Except for other Canadian-born children whose RV scores reverted to intake levels, gains achieved during the program were sustained through to adolescence. For Immigrant children, RV scores continued to increase after program exit to adolescence. This result is consistent with Beiser et al. (2002) who reported Immigrant children struggle in school for about 10 years and then do better than the average Canadian-born child.

Our results suggest that early family environment is influential in children’s RV development. We found that (a) English as the primary language spoken at home, (b) caregiver adversity in their own childhood, and (c) older caregiver age contributed to better RV at adolescence. These results are consistent with studies that suggest a shift away from traditional conceptualizations of school readiness as a within-child phenomenon (Christensen et al., 2022; Janus & Gaskin, 2014). Our results align with the concept of cumulative (dis)advantage, where dis-favorable and favorable societal positioning (such as ethnicity, gender, and income levels) accumulate over time, affecting educational trajectories across generations (Cohen et al., 2022). An unexpected result of our study was that higher childhood adversity, as measured by caregiver ACE scores, was associated with higher RV scores at adolescence. An age 15 follow-up of the Fragile Families Study identified a three-dimensional integrative framework of adversity: (a) deprivation linked to health problems and cognitive ability, (b) threat linked to aggression, and (c) unpredictability linked to substance use and sexual risk-taking (Usacheva et al., 2022). The preceding constructs of adversity were related, and the authors suggest environmental influences as influential across the dimensions.

Despite concerted efforts to retain participants (Ginn et al., 2017; Mychasiuk & Benzies, 2012), our study is limited by attrition and small sample. Testing children in English provides a partial understanding of children’s RV in other languages and may underestimate their ability (Umbel et al., 1992). Future studies should consider using tests that have been translated to the language understood by the child. Except for a few studies (Kim et al., 2022), limited research exists surrounding the complex relationships between gender and ethnic variation in neurobiological development (Chilosi et al., 2023; Qu et al., 2021). Future research using a transdisciplinary approach to examine the intersection of RV development in different ethnicities may explain similarities and differences in development across the lifespan (Causadias et al., 2023). Our study was conducted through a single, two-generation preschool program in Canada. Results may not be generalizable to other jurisdictions. Finally, this study used a longitudinal cohort design. Future studies should employ more rigorous designs with a control or comparison group.

Implications for early childhood education policy and practice

Our results have implications for policymakers and practitioners who wish to improve outcomes for children who attend two-generation preschool programs. Although universal programs have the advantage of reducing stigma for socially vulnerable populations, disadvantages include large implementation costs, without expected outcomes (McLuckie et al., 2019). Greater precision in identifying target populations and delivery of tailored early interventions may improve desired outcomes for specific subgroups of children. The shift towards the use of behavioural neuroscience frameworks is consistent with the increased understanding of the interdependencies of child characteristics and socioeconomic status on school readiness (Micalizzi et al., 2019). In two-generation preschool programs, it may be useful for practitioners to apply behavioural neuroscience frameworks (Center on the Developing Child, 2018) that associate reciprocal parent–child interactions as supporting children’s learning. It is likely that responsive parent–child interactions (rather than positive affective qualities) are most important for child language development (Madigan et al., 2019).

Highlights

• Follow-up studies to adolescence after two-generation preschool intervention programs are limited.

• Longitudinal, observations of receptive vocabulary development are limited.

• Exploration of differences in receptive vocabulary by gender and ethnicity is rare.

• Focused preschool intervention is recommended for Immigrant girls and Indigenous boys.

• Children with intergenerational adversity may require additional supports.

Acknowledgment

We are grateful to the parents and young people who participated.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

Data are available from the first author upon reasonable request from qualified researchers.

This study was performed in accordance with the ethical standards as laid down in the 1964 Declaration of Helsinki and its later amendments. The Conjoint Health Research Ethics Board at the University of Calgary (REB13–0750) approved the study. Written, informed consent was obtained from parents or legal guardians of children, which included consent to publish aggregated data.

Additional information

Funding

This work was supported by the Max Bell Foundation under Grant 15-B-30.

Notes on contributors

Karen M. Benzies

Karen M. Benzies, PhD, RN is a Professor with the Faculty of Nursing, and Adjunct Research Professor in the Departments of Pediatrics and Community Health Sciences in the Cumming School of Medicine. She is also the Director of the Social Innovation Initiative, Vice-President (Research) Office, University of Calgary and member of the Alberta Children’s Hospital Research Institute. Benzies has been conducting research with CUPS (formerly Calgary Urban Project Society) since 2001. She conceptualized the study, supervised longitudinal data collection and analyses, and wrote the first draft of the manuscript.

Arfan R. Afzal

Arfan R. Afzal, PhD, was a biostatistician with the Faculty of Nursing at the time of data analysis. He conceptualized the approach to data analysis and analyzed the data. He wrote the sections about data analysis and assisted with interpretation of results.

Carla Ginn

Carla Ginn, PhD, RN is an Associate Professor with the Faculty of Nursing. She assisted with retention of participants, data cleaning, and interpretation of the results.

Robert Perry

Robert Perry, MBA, was the Director, Strategic Partnerships, Research, and Technologies with CUPS throughout longitudinal data collection. He assisted with design of the two-generation preschool program, recruitment, data collection, and interpretation of results.

Carlene Donnelly

Carlene Donnelly, MBA, is the Executive Director with CUPS and supported research to ensure evidence-informed programs and services for vulnerable children and families. She was instrumental in ensuring that behavioral neuroscience theory permeated research and practice at CUPS. All authors read and approved the final version of the manuscript.