To design a self-administered screening questionnaire to inform men about their risk for testosterone deficiency. The screener was developed in two phases. First was a construction phase in which relevant risk factors and a scoring algorithm were defined from multiple logistic regression analyses of survey data. In the second phase, the screener's accuracy (based on sensitivity, specificity, and Receiver Operating Characteristic (ROC) curves) was tested using patients from a primary care clinic. All subjects provided blood samples for endocrine testing. Survey data from 1660 men aged 40-79 years participating in the Massachusetts Male Ageing Study (MMAS) were analysed in the first phase. The clinic sample consisted of 304 men aged 40-79 years presenting at a large Massachusetts primary health care clinic for routine check-ups or minor medical problems. The primary outcome was testosterone deficiency, defined as serum total testosterone below 12.1 nmol/l. Self-reported variables considered as potential risk factors included age, obesity, chronic diseases, health behaviour, the Jackson dominance scale, and symptoms of stress. The prevalence of testosterone deficiency was 20.4% in the MMAS and 42.1% in the clinic sample. An eight-item screener was developed based on age, body mass index, diabetes, asthma, headaches, sleep patterns, dominance preferences, and smoking status. The screener performed significantly better than chance in identifying men with low testosterone levels; the area under the ROC curve was 0.66 in the MMAS sample and 0.67 in the clinic sample. The self-scored screener developed in this study reliably detects men at risk of hypogonadism. The screener encourages at risk men to seek professional evaluation of their testosterone levels.