Teaching grammar and writing: A randomised controlled trial and implementation process evaluation of Englicious

70 schools randomised; test data from 1,246 pupils in 63 schools analysed

Year 2 pupils aged 6-7 in English primary schools - the youngest children ever studied with this design. The authors state that prior to this trial "no research worldwide had examined the impact of grammar teaching on the writing of six-year-old and seven-year-old pupils using this kind of robust design."

School-level randomised controlled trial with a qualitative implementation and process evaluation (teacher surveys plus visits to 12 case-study schools). The intervention was Englicious, a modern-linguistics-informed approach delivering England's national curriculum grammar requirements: 10 manualised lessons, each consisting of explicit teaching of grammatical terms followed by pupils applying that learning to a piece of their own writing. Primary outcome was a test of narrative writing; secondary outcome a sentence-generation test (generate a sentence from a two-word prompt), which is the more grammar-proximal measure. Both are researcher-designed. Reported in full as a Nuffield Foundation report ("Grammar and Writing in England's National Curriculum") and published in the Journal of Writing Research 18(1). Note the DOI is registered twice by the publisher (also as 10.17239/jowr-2026.17.04.04); the volume 18(1) registration is used here.

"There was effectively no impact on pupils' narrative writing (d=0.04) which was our primary outcome measure. This is consistent with previous studies in the field of grammar for writing which have not found an impact. The study finding does not offer support for grammar teaching to improve writing." The secondary, grammar-proximal outcome was larger but still null: sentence generation d=0.14, not statistically significant. The authors describe the sentence- generation result as "encouraging" relative to achieved power and hypothesise it reflects the manipulation of words, phrases and sentences in the intervention rather than the naming of grammatical terms - i.e. the sentence-construction component, not the grammar-labelling component. The report also records a developmental fact that bounds the whole question at this age: "The ability to combine sentences that use conventional grammar develops at around the age of seven for most pupils." Conclusion: the lack of robust evidence underpinning the grammar requirements in England's national curriculum is a concern and those requirements should be reviewed.

Low. School-level randomisation with an independent university evaluation team; treated and control pupils do not differ systematically in heritable ability.