Four hundred nine tenth, eleventh, and twelfth grade high school students were surveyed to determine the percentage "at risk" for suicide and to compare the "at risk" group with those who were determined to be not at risk on the variables friendship, life stress, and self-esteem. The "at risk" group was determined by positive responses to questions about suicidal thoughts as well as by a high score (determined normatively) on the Hopelessness Scale (Beck, Weissman, Lester, & Trexler, 1974). The literature has shown that suicidal behavior can be represented by a continuum beginning with suicidal ideation and ending with the completed act so those who have thought about suicide are at greater risk than those who have not. Two primary analyses were conducted. In the first, two multivariate ANOVA's were carried out wherein high and low risk defined one independent variable and gender defined the other. Scores on self-esteem, life stress, and friendship served as the dependent variables. Those not classified as either high or low risk were excluded from this analysis. When compared to low risk individuals, high risk subjects were found to have significantly poorer quality friendships, lower self-esteem and had experienced more life stress in the previous year. When sex difference was the dependent variable, only life stress was significant with females reporting more stressful life events. For the secondary analysis, all subjects were included and multiple regression analysis was used to predict a) hopelessness, and b) suicidal ideation using self-esteem, life stress, and friendship as independent variables. The three variables in combination were better at predicting ideation (R² =.36) than hopelessness (R² =.24). Life stress was a significant predictor in both equations and was the best predictor for ideation. Self-esteem was the best predictor for hopelessness. Friendship entered into both equations, but only contributed significantly to the variance in hopelessness. / Ph. D.