His researchers' analysis, to be announced at teachers' union conferences next week, reveals that three out of four "high-performing" deprived schools also have special factors which their "bog standard" fellows cannot replicate.
Following a TES investigation last month showing that most "high-performing" schools in deprived areas either select some pupils or have extra funding, Mr Willis asked the Department for Education and Skills for its confidential free school meals tables.
These have been used by the Government to claim that pupils in schools with similar levels of deprivation perform very differently at GCSE. The DfES initially refused to give Mr Willis the figures.
The special factors affecting schools' performance include all-girl intakes, selection by banding or religious affiliation, specialist status bringing extra funding, and exceptionally low levels of pupils with special educational needs.
Once "special factor" schools are taken out of the equation, most schools with high levels of pupil deprivation - with more than 40 per cent receiving free school meals - perform at very similar levels at GCSE, with a third or fewer of their pupils gaining five grade A*-Cs. Of the 226 schools in the country with 40 per cent or more of their pupils on free school meals, 93 manage to get at least 30 per cent of children to that benchmark GCSE score. But of those 93 high-performing schools, 71 have a special factor affecting intake or funding. A mere 22 do not.
"These figures are a striking reminder that we have a very, very unequal school system," Mr Willis said. "The Government, while avowedly pursuing an inclusive agenda, has actually made things worse by targeted funding and the specialist schools programme."