There are no guidelines available for diagnostic studies in patients with mental retardation (MR) established in an evidence-based manner. Here we report such study, based on information from original studies on the results with respect to detected significant anomalies (yield) of six major diagnostic investigations, and evaluate whether the yield differs depending on setting, MR severity, and gender. Results for cytogenetic studies showed the mean yield of chromosome aberrations in classical cytogenetics to be 9.5% (variation: 5.4% in school populations to 13.3% in institute populations; 4.1% in borderline-mild MR to 13.3% in moderate-profound MR; more frequent structural anomalies in females). The median yield of subtelomeric studies was 4.4% (also showing female predominance). For fragile X screening, yields were 5.4% (cytogenetic studies) and 2.0% (molecular studies) (higher yield in moderate-profound MR; checklist use useful). In metabolic investigations, the mean yield of all studies was 1.0% (results depending on neonatal screening programmes; in individual populations higher yield for specific metabolic disorders). Studies on neurological examination all showed a high yield (mean 42.9%; irrespective of setting, degree of MR, and gender). The yield of neuroimaging studies for abnormalities was 30.0% (higher yield if performed on an indicated basis) and the yield for finding a diagnosis based on neuroradiological studies only was 1.3% (no data available on value of negative findings). A very high yield was found for dysmorphologic examination (variation 39-81%). The data from this review allow conclusions for most types of diagnostic investigations in MR patients. Recommendations for further studies are provided.