You report that failure to write an impact statement on a grant application does not affect the likelihood of obtaining funding, but your figures show the opposite ("Refusal to comply on impact failing to sink grant applications", 4 February). An application to the Engineering and Physical Sciences Research Council has a 1 in 11 chance of success with a statement, and only a 1 in 40 chance without. Moreover, a simple chi-squared test shows this is statistically significant at the 2.5 per cent level. On the other hand, an Arts and Humanities Research Council application has only a 1 in 18 chance of success with an impact statement and nearly 1 in 4 without. What does that tell us?
Robert Wilson, Queen Mary, University of London.