Summary.Few tests of spelling attainment have been validated against the criterion of spelling accuracy in everyday writing, which, it is contended, is the prime objective of teaching spelling. An investigation is described comparing the spelling performance of 506 children (median age 11: 6) in their written work and on dictated‐word (recall) and multiple‐choice (recognition) types of test. The dictated word test had a slightly higher concurrent validity, but other factors which should be taken into account in devising spelling tests are pointed