Tested the assumption that semantic negative and positive items measure the same construct in Likert-type scales on a sample of 501 8th and 9th graders in Jordan. Four item sets (i.e., "I feel happy," "I do not feel happy," "I feel sad," "I do not feel sad") were administered to 4 groups of Ss, who were stratified on the basis of gender and urbanization. The effects of the inclusion of negative and positive valence items on the homogeneity index of the scale and the validity of conclusions based on such tests were investigated. Correlational and variance analyses indicated that semantically negative and positive item contents did not measure essentially the same construct. Findings question the common practice of using both negative and positive valence items in a scale and making decisions on the basis of the students' total score on such items.