PURPOSE: Drugs prescribed for the treatment of insomnia can be either benzodiazepine hypnotics or the newer z-hypnotics, zopiclone and zolpidem. This paper explores possible explanations for the choice made. METHODS: Data from the Norwegian Prescription Database covering the entire population was studied for incident users of hypnotics. Possible predictors were age, gender, previous psychotropic or analgesic drug use and prescriber speciality. RESULTS: Of the 73,163 incident users of hypnotics, 3876 were prescribed benzodiazepine hypnotics in 2006. The strongest predictors for being prescribed benzodiazepines were previous use of anxiolytics [odds ratio (OR) 1.8, 95% confidence interval (CI) 1.7-2.0] and male gender (OR 1.5, 95% CI 1.4-1.6). Other significant predictors were antipsychotic or opioid drug use and the prescriber being a psychiatrist. CONCLUSIONS: Z-hypnotics were commonly prescribed. Norwegian drug therapy recommendations also suggest a preference for z-hypnotics. The clear predominance of the shorter acting z-hypnotics may be due to the fact that only longer acting benzodiazepines are available in Norway. Reasons for prescribing benzodiazepines may be co-morbid psychiatric illness, such as anxiety, or a belief that benzodiazepine hypnotics are more effective than z-hypnotics.