A report for an independent inquiry into England's primary schools says standards of reading have risen little in 50 years.
What is the evidence for that claim?
The report's co-authors, Peter Tymms and Christine Merrell, are long-standing sceptics of the apparent rise in national curriculum test results that the government says shows standards have risen.
They are the director and project manager of the Curriculum, Evaluation and Management (CEM) Centre at the University of Durham.
The centre carries out research and provides monitoring systems for schools.
In their report for the Primary Review, they chart the national testing of reading that has been done since World War Two.
Between 1948 and 1979, there was a series of surveys of 11-year-olds' reading ability in England and Wales.
Watts-Vernon tests, as they were known, were used in 1948, 1952, 1956, 1961, 1964 and 1970-71, and National Survey 6 (NS6) tests in 1955, 1960, 1970-71, 1976 and 1979.
APU Language and Monitoring Project tests, using "authentic" tasks rather than multiple choice questions, were used in 1979, 1980, 1981, 1982, 1983 and 1988.
Scores gradually rose between 1948 and 1952 - thought to reflect the general post-war recovery - then remained constant.
The APU tests showed very little change between 1979 and 1988.
Surveys in Scotland between 1953 and 1963, then 1978 and 1995, showed slight changes "but nothing that could be said to be important educationally".
After 1990 there was a five-year hiatus following the introduction of the national curriculum and national testing in England, until the new system became established.
Small scale surveys during that time indicated a drop in standards - due possibly to changes in home and school contexts or the teachers' industrial dispute in the mid-1980s, or because the changes to the education system made heavy demands on teachers' time.
Since 1995 there has been controversy.
"The debate goes much deeper than an academic dispute as it links directly to public political claims," the report says.
The statutory tests (popularly known as Sats) were originally conceived to provide information to guide teachers' practice, as well as information about the levels of attainment.
"However, this rapidly shifted and tests became an accountability tool.
"The data enabled comparisons between children, schools and local authorities to be made and was in the public domain."
In other words, the "league tables" were born.
When Labour came to power in 1997, it stressed the basics of literacy and numeracy and introduced overarching strategies for improvement.
Scores from the tests at the end of Key Stage 2, taken in the final year of primary school by children aged 10 and 11, showed "remarkable" rises between 1995 and 2000.
The proportion of children who reached the standard of Level 4 - which is what the government expects of them at that age - rose from 48% to 75%.
But questions began to be asked. Prof Tymms challenged the apparent rise in standards in a report in 2004.
Having considered 11 different studies, most involving the individual testing of thousands of pupils, he concluded that there had been an improvement but the results had exaggerated it.
The then Statistics Commission - since abolished - confirmed that the official data had overstated the rise.
The then Department for Education and Skills - now the Department for Children, Schools and Families - disagreed.
The Primary Review report charts this controversy but concludes that standards of reading have remained more or less the same since the 1950s.
"There was a rise following the immediate post-war period and there was a slight drop followed by a recovery after the introduction of the National Curriculum, but in essence standards have remained constant."
One thing the government pointed to with pride was the major international study of reading known as the Progress in International Reading Literacy Study (Pirls) in 2001 which tested the reading of nine-year-olds.
England came third out of 35 countries, behind only Sweden and the Netherlands.
The then education secretary, Charles Clarke, said there was much to be proud of.
"It shows that the national literacy strategy we set up five years ago to raise standards in primary schools is working," he said at the time.
But later research challenged the Pirls methodology.
It questioned whether a single reading test could be translated into different languages and cultures.
The tests were said to be very like the English Sats. And - most damningly - England had excluded more children with special educational needs than other countries.
Nevertheless, the report for the Primary Review concludes that, as a general rule, children do learn to read in primary school.
"Most cannot read before they start school and the vast majority of children are able to do so by the time they move to secondary school.
"The general consensus is that this process is a successful one and it has been ongoing for generations."
The authors suggest it would be better to monitor standards over time using assessment that is not part of the accountability framework within which schools operate, and by testing samples of children rather than all children.
This echoes something that has been proposed by the head of the Qualifications and Curriculum Authority, Dr Ken Boston.
The government meanwhile is piloting a "testing when ready" system.
It is intended to allow schools and parents to monitor children's progress more regularly - rather than depending on the national tests at the end of key stages.