We used a regression tree method (RTM) to determine risks of depression in children/adolescents. The survey records of 4,143 children/adolescents in a study based in Mersin, Turkey served as data in this study, and multi-step, stratified, and cluster sampling were used. Effects of 24 variables (sex, smoking, parental problems, etc.) were evaluated on depression scores. The Child Beck Depression Inventory (CBDI) was used to determine the level of depression. Subjects were into 12 different groups based on magnitudes of mean depression scores. The interactions among 7 variables determined to be risk factors are shown on a schema. The STATISTICA (ver.6.0) package program was used for all computations. Although traditional statistical methods have often been used for analysis in this field, such approaches are associated with certain disadvantages such as missing values, ignorance of interaction effects, or restriction of the shape of the distribution. To avoid such disadvantages, we therefore suggest the use of the RTM in studies involving numerical-based outcome variables and for the investigation of a large number of variables and it may be more effective than traditional statistical methods in epidemiological studies which determine risk factors.