The official statistics watchdog has said rising primary school test scores in England "substantially" overstated the rise in education standards.
Report says there is "no real alternative" to assess standards
The Statistics Commission said there was only "some" rise in standards between 1995 and 2000.
Official pronouncements should acknowledge other possible reasons for rising test scores, it said.
But the Department for Education said international evidence showed standards were "high and rising".
Standards over time
The commission, which operates independently of ministers and official statisticians, was contacted by Durham University professor Peter Tymms who operates different assessments for schools and education authorities.
His tests had not shown the rise in standards reported by the government.
Prof Tymms' concern was that the Key Stage 2 national curriculum test scores - which form the basis of the primary school "league tables" - were not suitable for monitoring trends in standards.
The proportion of children achieving Level 4
The commission has said it believes it has been established that the improvement in test scores between 1995 and 2000 "substantially overstates the improvement in standards in English primary schools over that period".
There was "some rise in standards".
"Ministers and others who may want to use the test scores in a policy context need to be made fully aware of any caveats about their interpretation," the commission said.
"There are a number of qualifications that need to be made. Yet government departments have usually failed to mention any caveats about other possible reasons for rising test scores in their public comments."
Official statements should acknowledge the effects of "teaching test technique and teaching to the test" on test scores following the introduction of a new test, as the Key Stage 2 tests were in 1995.
'No real alternative'
Initially, test results rose rapidly. The percentage of pupils getting what is known as Level 4 for English rose from 48% in 1995 to 75% in 2000, the period in question. The maths score went from 44% to 72%.
Since 2000 the upward trend has largely halted. Last year the results were 78% in English and 74% in maths.
The commission said the tests were not "a wholly inadequate" way of measuring standards over time - "not ideal", but there was "no real alternative at present".
The Qualifications and Curriculum Authority (QCA) said the Massey report which it had commissioned had reached broadly similar conclusions.
The phenomenon of a dramatic rise in test scores following the introduction of an important new public test was well known among measurement specialists.
But the implications of this had, to date, "not been well-presented to government, nor have issues of measurement error and the implications these have for performance tables".
Prof Tymms, director of the Curriculum, Evaluation and Management (CEM) Centre at Durham, told the BBC News website he was very pleased at the findings as he had been "banging on" about the issue for a long time.
But he said the important thing was to focus on "getting education right".
"I'm all for getting better and improving standards, I just want it done right."
A spokesperson for the department said England's performance in international benchmarks in maths and reading literacy gave it confidence that standards were "high and rising". And Ofsted had said that teaching had improved.
But the shadow education secretary, Tim Collins, said: "Parents and teachers will not be surprised that an independent report has shown ministers to have been selective and mendacious in their use of education statistics."
Statistics Commission, Report No 23: Measuring Standards in English Primary Schools.