A reading improvement scheme allocated almost 拢100 million in government funding may have helped participating schools weather the impact of Covid-19 disruption, Department for Education research suggests. However, found no firm 鈥渃ausal鈥� link between participating in the English hubs programme and improved outcomes in the phonics screening check. English hubs were launched in 2018 with 拢26 million in set-up funding. Each of 34 鈥渉ubs鈥� is a school selected for their 鈥渆xpertise鈥� in teaching early reading. They work with selected schools in their area to improve teaching. A further 拢48 million has been spent since then, and 拢19.8 million has been allocated for next year. Government researchers analysed results from the phonics screening check, which is conducted in year 1, for schools that participated in the English hubs scheme and those that did not. They found that between 2018-19, the last year the phonics check was conducted before pandemic disruption, and 2021-22, the proportion of pupils in partner schools meeting the 鈥渆xpected standard鈥� rose by around 1 percentage point. In non-partner schools, the proportion meeting the benchmark fell by around 7 percentage points. Supported schools ‘appear to have fared better’ The DfE said this meant schools supported by English hubs 鈥渁ppear to have therefore fared better than other schools, on average, in dealing with the negative impacts of the pandemic on PSC attainment鈥�. The research said this was 鈥減ossibly鈥� due to the English hubs programme. However, there are 鈥渓imitations鈥� to the findings. To interpret the results as 鈥済enuine programme impacts鈥� relies on 鈥渟trong assumptions about what would have happened to partner schools鈥� attainment in 2021-22 if they had not been enrolled into the programme鈥�. A causal interpretation 鈥渁lso requires the assumption that the results are not being biased by variables we do not or cannot control for鈥�, such as schools implementing interventions that were not at the behest of the English hubs scheme. 鈥淯ltimately, we cannot definitively rule out that omitted variable bias is driving part or all of the estimates and therefore none of these results can be interpreted as causal. 鈥淗owever, given the size of the coefficients across the various methodologies and sensitivity analysis carried out, it seems more likely than not that the [programme] is having an impact on phonics results.鈥�