Statistics, damned statistics

Your latest student satisfaction survey contains glaring methodological shortcomings ("It's all about them", 14 January). Most striking is the minute sample size (in our university's case, 0.16 per cent - 41 students out of 24,000). The National Student Survey requires a minimum of 50 per cent respondents in each category before the data are taken into account.

Unlike the NSS, moreover, the THE table fails to take account of how different subject groups may have different expectations and responses to resources or staffing questions. To attempt to pass authoritative judgment on such flimsy statistics is intellectually dishonest.

Mohammad Dastbaz, Dean of the School of Computing, Information Technology and Engineering University of East London.