Reported is the development and use of an instrument designed to measure intellectual and humanistic aspects of attitudes towards teaching science. The instrument was validated by a factor analysis technique. Test-retest reliability for the instrument was determined to be 0.87. Using this instrument, two methods of elementary science instruction were compared. The experimental treatment involved teaching methodology suggested by modern science curriculum projects. The control group was instructed in a traditional manner. The experimental group was instructed in a more humanistic manner. The author concluded that a more humanistic teaching strategy produced more positive attitudes towards teaching science. He recommends that university instructors utilize an approach more in concert with the views of the developers of modern science curriculum projects and that the development of affective behaviors toward science be a goal of elementary science instruction. (Author/EB)