Over the last decade the way in which children are taught to read in primary schools in England has fundamentally changed. The teaching of reading is now dominated by phonics, or more accurately “synthetic phonics”, where the focus is on teaching children the connection between phonemes (sounds) and letters. This teaching of the alphabetic code has always been part of learning to read, but our new research suggests it now dominates, to the exclusion of a more balanced approach involving reading whole texts and understanding the meaning of words. For example, children are given a statutory test when they are five or six called the “phonics screening check”, where they are given pseudo-words, or nonsense words, like “sut”, “blem” and “meck” to read. They do not need to know what these mean, but are judged to have “read” them correctly if they say the appropriate sounds. Yet this is decoding, not reading for understanding. Moreover, the inconsistencies in English spelling mean that children cannot only rely on phonics to read: common words like “the” and “she” have to be taught as exceptions. Our new research (written with Prof Dominic Wyse) found that the majority of teachers of children aged three to six were using a “phonics first and foremost” approach. In our survey of more than 2,000 teachers, teachers said that English as a lesson had been replaced by phonics, and that they were under pressure to “live and breathe phonics” to make sure their students did well in the phonics test. The problem is that the evidence does not back up this push towards phonics. Our review of systematic reviews (that combine evidence from multiple studies) and of robust longitudinal experimental trials, published in 55 research papers, concluded that contextualised teaching of reading is the most effective way for children to learn. This involves combining phonics with reading whole texts, and focusing on understanding words and sentences. There is no substantial research base that supports the dominance of synthetic phonics. Internationally, England is an outlier in using this policy approach to teaching reading, and data from international assessment tests suggest that teaching reading in England may have been less successful since the adoption of the synthetic phonics approach. So how did we get to a point where thousands of children are being taught to read in a way which has no basis in proper evidence? The “reading wars” have a long history, but the shifts in England’s policy in the 2010s marked a new era of phonics dominance in England. This grew out of the recommendation in 2006 by Jim Rose that phonics be taught “discretely” as a separate skill, and the inclusion of phonics within the primary national strategy. However, a huge boost was given to phonics by the introduction of the phonics screening check in 2012, as it brought the particular skill of decoding into the system of judging and comparing schools. Dedicated phonics lessons and a focus on nonsense words have been found in classrooms ever since. The government has recently ratcheted this up further by asking inspectors to check that teachers in reception classes “teach children to read systematically by using synthetic phonics and books that match the children’s phonic knowledge”. The Department for Education has revised the list of “approved” phonics schemes, so schools have had to buy new schemes and sets of “decodable” books. The debate about the teaching of reading relates to a wider problem of education policy that proclaims to be evidence-based but instead relies on limited evidence selected on the whims of ministers. In an open letter with over 250 signatories, we are calling on the government to act on the evidence, and take on a more balanced approach. Teaching phonics has a place – because who would disagree that children need to learn about letters and sounds – – but by focusing on it to the exclusion of other skills, we are failing our children. Alice Bradbury is a professor of sociology of education, University College London