The number of students in the United States choosing science, technology, engineering or mathematics careers is declining at a time when demand for these occupations is rapidly increasing. Numerous efforts have been undertaken to reverse this trend, yet results are uncertain. One's attitude is key to many choices one makes, and this includes, for many, what career is pursued. Hence, teachers, informal science educators and researchers often wish to measure children's attitudes towards science using a pretest and a posttest to determine the effects of a curriculum, an activity or an intervention. However, measuring children's attitudes toward science has been problematic because of both the limited use of basic psychometrics in checking reliability and validity of instruments and the lack of a single construct of students' attitudes towards science being surveyed. This article reports the development and testing of an instrument for measuring students' science attitudes across several dimensions. Thirty-two scientists and teachers from the northeastern and south central United States participated in content validity trials. The instrument was field tested with 549 children (92 elementary-school students, 327 middle-school students and 130 high-school students) from 6 rural and suburban school systems located in the northeastern United States to determine inter-item reliability for each dimension. The resulting instrument, entitled My Attitudes Toward Science (MATS), has 40 items that measure four dimensions: (1) Attitude towards the subject of science; (2) Desire to become a scientist; (3) Value of science to society; and (4) Perception of scientists. The MATS, as a multidimensional instrument, can measure several facets of students' attitude toward science and is designed to be used across grades levels and to be scored easily.