Objective: Experience sampling (ES) involves participants rating ambulant phenomena at unpredictable intervals. Despite its perceived benefits, there is little published research evaluating the limitations of this method. Method: Predictors of compliance were investigated across three studies using ES in psychotic populations. Results: Regression analyses indicated that none of the demographic or clinical variables significantly (p.05) predicted the number of valid reports or compliance as defined by traditionally used cut-off points (33%). Conclusions: We conclude that it is difficult to predict compliance in ES research, which is likely to be an accessible methodology, even for people experiencing distressing symptoms.