Child-directed speech and early language development: A systematic review of reporting habits

Video walk through of poster presented at ICIS 2020: Child-directed speech and early language development: A systematic review of reporting biases. Styles & Yogarrajah (2020)

Abstract:
Since the well-known study by Hart and Risley in the 1990s, a substantial literature has arisen investigating the relationship between a child's language development and the speech they hear from their caregivers. In these studies, a recording of the child's auditory environment is analysed for the quantity and quality of parental input to the child (e.g., Quantity: number of word tokens, conversational turns; Quality: mean length of utterance, vocabulary diversity). The child's linguistic development is also assessed, using standardized measures, and correlations are used to establish whether higher quantity and quality of linguistic input in the early years is associated with higher language skills. When a positive relationship is observed, there is a tacit assumption that higher quantity or richer parental speech has been instrumental in the child's language development. However in correlational designs like this, other causal pathways are possible. We investigated the prevalence of different causal interpretations in the published literature, to find out whether the literature contains an interpretative bias.
We conducted a systematic review of peer reviewed journal articles from 1977 to 2018 that used a recording of parent/child interactions to establish input characteristics of the child's language environment, and independently assessed the language development of the child. A team of raters independently assessed and cross checked each peer reviewed article that met the inclusion criteria, looking for three causal pathways: Pathway A - the quantity/quality of parent talk influences child language outcomes (e.g., talking more helps a baby develop language skills); Pathway B - endogenous child language skills influence the quantity/quality of caregiver talk (e.g., a chatty child elicits more speech); Pathway C - external factors influence both parent talk and child language development independently (e.g., environmental factors such as poverty-induced stress). As in any correlation between two measured variables, all three pathways are possible, and it is not possible to establish which causal pathway (or which combination) is correct without direct intervention studies (Bishop 2014). The systematic review assessed which pathways were discussed as possible interpretations in each paper.
Following a preregistered PRISMA pathway, database searches returned 2,446 articles, of which 54 were eligible for inclusion (see Figure 1). Overall, 37 articles discussed just one possible causal pathway, 15 discussed two pathways, and only 2 discussed all three pathways. Figure 2 shows that Pathway A was the most commonly discussed interpretation, followed by Pathway B, then Pathway C. This pattern was observed most clearly from the 2000s onward, when publication rates on this topic began to increase. This finding demonstrates that reporting habits in this field are biased towards the interpretation that 'speaking more to a baby will enhance their language development'. This interpretive bias has potential implications for parents, educators and policy makers, as scarce resources may be invested in interventions targeting one causal pathway (e.g., talking more to babies), while others are overlooked (e.g., alleviating poverty in families most at risk). We argue that controlled studies of talk-based interventions will be critical for establishing the effectiveness of this causal pathway.