An 8-item loneliness scale was administered to a sample of high school students (N = 1,710) twice, about one year apart. The scale exhibited good internal consistency reliability (alphas of 0.78 and 0.79, respectively), was relatively independent of gender effects, and demonstrated evidence of construct validity. Factor analysis of the items yielded a two-factor solution which met methodological criteria, but over-all the data suggested a method artifact, e.g., positively and negatively worded items factored separately. Examination of the two factors indicated the subscale of four negative items (lack companionship, no longer close, left out, feel isolated) had an alpha reliability of .77, correlated .89 with the total scale score, and exhibited virtually an identical pattern of associations with a set of theoretically relevant variables as the full scale. This four-item scale may be useful in epidemiologic surveys in which over-all response burden is an issue.