Methodological problems in follow‐up and replication studies and in using projective tests in research are discussed with particular reference to three fear‐of‐success studies: the original research by Horner in 1965, a replication in 1972, and a follow‐up of the original 1965 subjects carried out in 1974. The importance of each of the following is stressed: obtaining responses from the full original sample in follow‐up work; matching the settings in which projective tests are administered; establishing coder reliability between the data sets that are to be compared; and matching the stimulus value of the cue, which may not always mean using the