So if phonics doesn't work...what then?

With a quarter of ten and eleven year-olds in England unable to read to the expected standard (DfE, 2018), and with the majority of those children now having had phonics instruction for two years of their initial schooling perhaps Glazzard (2017) is correct when he asserts that if a child has received phonics instruction from aged five to seven and is still struggling to read then more phonics instruction is irrational; something different is needed not more of what has already failed.​ The Education Endowment Foundation (EEF) (educationendowmentfoundation.org.uk., 2018) stated in their online phonic toolkit that, ‘For older readers who are struggling to develop reading skills, phonics approaches may be less successful…children aged 10 or above who have not succeeded using phonics approaches previously require a different approach…’. They cited Ehri (2004), who suggested that the teaching of phonics to older pupils was complicated by their reliance on inefficient compensatory strategies that were difficult to reverse. If this were the case, then improvements in any code knowledge as a result of of specific teaching could be nullified by continued dependence on, and reversion to, the compensatory strategy. The EEF’s (2018) conclusion, however, did not account for Beck’s (1998) assertion that phoneme to grapheme mapping is the equivalent to reading as dribbling skills are to basketball: necessary but not sufficient to play the game. The implication from Beck (1998) was that without sufficient phonic knowledge, fluent reading may not be possible.

The fact that the DfE (2016) year 3 pilot screening of seven and eight-year-olds revealed that nearly 50% of those assessed had not reached the threshold for phonics mastery might suggest that Glazzard (2017) is indeed correct. Phonics just doesn't work for some children so something different is needed and the NAHT and other teaching unions were justified in crowing their success in having the year 3 phonics screening dropped (rather than hanging their heads in shame in achieving one of the darkest stains on national literacy assessment- exacerbated by the pilot's conclusions that it was poorer children who fared worse in the evaluation).

The impact of instruction in phonics on reading is significantly greater in the first two years of schooling (Ehri, 2004). However, effect sizes are considerably smaller when this instruction is introduced after the age of seven with fluency instruction and comprehension instruction producing far greater effect sizes (NIHCD, 2000). This led Ehri (2004) to suggest that beyond ages of seven, phonics instruction must be combined with other forms of reading instruction if maximum impact is to be attained. This is counter-intuitive to the Simple View of Reading (Gough and Tumner, 1986) which indicates that decoding mastery is directly correlated with reading comprehension (see here). Ehri (2004) makes this supposition directly from effect sizes and concedes to there being a paucity of research in this field. She surmises, that the diminishing effect sizes for phonics instruction in older pupils may be a result of the difficulties in altering students’ habits and compensatory strategies when attacking unknown words. This may require the suppression of guessing words based on initial phonemes and contextual cues and the closer examination of spellings when reading words (Ehri, 2004). For an older child who has failed to master phonics, the most efficacious strategy when attacking an unknown word is to wait until an adult decodes it for them (Connor, Morrison, Katch, 2004) - if you want to see this strategy in action, observe the majority of KS2 reading lessons.

Although specific research into the teaching of phonics beyond the age of seven is elusive, inferences can be drawn from wider studies. ‘The Reading First Program’, a reading intervention for struggling readers, established as part of the No Child Left Behind (2001) legislation in the United States, included substantial decoding instruction for older pupils (seven- and eight-year olds). Although the impact on reading fluency and comprehension was poor, the impact on decoding was significant (Kucan and Palincsar, 2011) indicating that phonics instruction for older children can be effective. The poor outcomes for comprehension have been blamed on a poorly conceived strategy for instruction with the approach lacking a systematic and consistent method (McKeown et al, 2009).

Further indication that phonics instruction in later years may be effective can be implied from Vaughn et al.’s research into a thirty-week intervention for forty-five eight-year-old children identified with reading problems. Although the intervention included both fluency and comprehension instruction, the early weeks were weighted heavily in favour of phonemic awareness and letter-sound relationships. Seventy-six per cent of the sample met the success criteria at the end of the intervention and further monitoring indicated that of those, seventy percent continued to be successful readers.

McCandliss et al.’s (2003) study of seven-year-olds whose word-attack strategy relied on initial consonant decoding also indicated that interventions that focused on phonemic manipulation by altering one letter at a time (a regular strategy employed in linguistic phonics programmes) resulted in participants significantly outperforming the control group in a decoding assessment. These results were supported by Harm et al.’s study into word building interventions that focused on participants writing letters rather than focusing on speech activities. The pairing of orthography and phonology was crucial, the study concluded, to enhance the knowledge of phonemic structure (Harm et al., 2003).A study of third grade children who were in the bottom 20% of readers (Blachman et al., 2004) found that after a year’s intervention they had made significant gains compared to the comparison group. The intervention included sisubstantial elements of sound to symbol correspondence instruction and the use of texts controlled for those learned correspondences. The positive effects were evident one year after the intervention and a follow up study ten years later also suggested moderate effect sizes and benefits (Blachman et al., 2014).

The growing recognition that many reading difficulties are revealed beyond the early years of schooling (Chall, 1967) has not corresponded with research studies in this area. However, Leach et al. (2003) studied older pupils with late-emerging reading difficulties (eight and nine-year olds) and concluded that word recognition, decoding and spelling are significant impediments to progress in reading achievement beyond early school grades. They suggested that late emerging reading difficulties are being overlooked by educators and that a more forensic assessment protocols are required by schools.


In England, Edwards’ (2006) small scale study of sixteen 14 to 16-year-olds found significant improvements in word reading following a systematic phonics intervention. These results were supported by Jeffes’ (2016) phonics-based intervention for 30 secondary pupils which showed significant improvements in decoding and word recognition for participants. It also had a qualitative element which indicated generally positive views of the intervention from instructors which, with Jeffe’s (2016) study taking place after the introduction of the PSC this may indicate the importance of KS2 training in phonics.


Jennings (2008) noted that many pupils in England entered KS2 with ‘much of the phonics input from KS1 still unlearned’ (2008:32). Jennings (2008) also found that polysyllabic decoding was often the most in deficit. Her study provided phonics intervention for 16 Year 5 children with the most pronounced shortfall finding that they made four times the progress of the average child. However, Jennings (2008) notes that the intervention was not a pure phonics intervention. It is also noteworthy that the pupils had been taught early reading under the National Literacy Strategy (DfEE, 1998) and thus had not been exposed to SSP, as is the fact that Jennings’ (2008) study focuses on readers with chronic reading issues and had a very small sample size.


Results from a meta-analysis that included adolescents with reading difficulties, concluded that phonics training for poor readers was effective (McArthur et al.,2012). These outcomes were supported by a further meta-analysis of randomised controlled studies that included adolescents and found that only phonics instruction produced significant positive effect sizes (Galuschka et al., 2014). The study concluded that systematic instruction in letter-sound correspondences and decoding strategies were the most effective for improving literacy skills for children and adolescents with reading difficulties.


Gorard et al.’s (2015) research studied 433 pupils who had recently entered the English secondary school system (aged 11) but had achieved below the expected English standard. Two hundred and twelve pupils were removed from their English lessons for three hours every week for 22 weeks, and in small groups received phonics and word recognition practice. The study found that the intervention group made the equivalent of three months additional progress in standardised comprehension scores compared to the comparison group and concluded that there was considerable promise from using phonics as an intervention for these older pupils.


Summarising research into struggling readers, particularly in later years, Kucan and Palincsar (2011,) conclude that, ‘We need to focus our efforts on minimizing the bottle-neck effects of the decoding problems experienced by some struggling readers…’ (Kucan and Palincsar, p. 354).

With the poverty of research into the value of phonics instruction in later years it is apposite to review studies into adult literacy improvement for any clues as to the efficacy of an approach which seeks to advance decoding strategies where some non-phonic compensatory schemas may be established. Despite there being little academic investigation in this area of reading prior to the 1970s (Brooks, 2011), Kruidenier’s (2002) analysis of random controlled trials into adult illiteracy indicated that phonemic awareness and word analysis instruction led to an increase in achievement for poor adult readers. Burton’s (2007) small-scale study of adult illiterates supported this with its conclusion that phonics instruction enhanced students’ progress. A follow-up study by Burton (2008) found that there was a positive correlation between students’ progress and the amount of phonics training that their teachers had received. Brooks notes the lack of research into the efficacy of systematic synthetic phonics instruction for older learners despite the positive indications for very young readers and suggests that this approach ‘awaits convincing demonstration…’ (2011, p.192).

That the efficacy of phonics instruction for older learners has been questioned posits a crucial line of inquiry. There is little explicit research into this area of pedagogy, but what implicit research exists suggests that it is possible and effective. If phonics mastery is a threshold that must be traversed for reading fluency to be possible then the unpicking of compensatory strategies will be essential along with the establishing of an effective instructional programme in phonics for older learners.


And finally, the EEF have removed their statement from their website. Nice to know that they catch up eventually. However, the Australian equivalent still retains the verbatim advice...


What all this means for schools:

  • Assess the full code for all pupils until mastery - and don't rely on the PSC (it does not assess the full code).

  • Keep teaching the code until mastery is achieved - that is likely to be into KS2.

  • Ensure your phonics programme has a substantial amount of polysyllabic level decoding.

  • Spelling has a greater effect on reading than reading has on spelling and spelling is linked to the code (more here)

  • Post COVID 19 many secondary schools are receiving pupils with a chronic code deficit - you'll need a phonics programme and interventions.

Follow the reading ape on twitter - @thereadingape