Failure to Replicate Effects of Parent-Delivered Early Language Intervention: Evidence from a Randomized Controlled Trial with Implications for Universal Language Intervention

<b>Background </b>Parent-led, preschool language interventions may help to mitigate the risk of poor language and literacy outcomes associated with lower socioeconomic status. This study builds on two previous evaluations of a parent-led early language teaching programme, which demonstrated mixed findings. <br/><br/><b>Methods </b>A two-armed randomised controlled trial was conducted with 372 families and their children aged 3-4 years, recruited from 43 schools across the North-West of England. Families were randomly allocated to either the programme group (N=186) who delivered the 20-minute programme 5 days a week for 30 weeks, or the control group (N=186) who received storybooks at the end of nursery. Language and early literacy outcomes were assessed at baseline (t1), immediately after the 30-week programme (t2) and 10 months later (t3).<br/><br/><b>Results </b>Families delivered on average 65% (20/30 weeks) of the intended sessions. There were no significant group differences in language outcomes at immediate (Hedges’ g=-.05 to .11) or delayed post-test (g=-.13 to .10). Further, no significant group differences were found on measures of early literacy (g = -.11 to .12), the Home Learning Environment (g = -.0.07) and School Readiness (g = -.15). Effects of the programme on language skills were stronger at higher dosage (g=0.33 for those completing &gt;90% sessions) and for those with weaker language skills.<br/><br/><b>Conclusions </b>Findings showed no significant effects of intervention though indicated potential effects downstream, particularly for children with poorest language skills at outset and receiving more teaching. Reasons for failure to replicate previous findings of effectivenessnull findings are discussed with implications for universal interventions. Recommendations for future trials evaluating interventions of this nature, including targeting of families and children most in need of support, are discussed. <br/>