Assessment of science process skills in the elementary school is essential to a hands-on science program. Concern arises with the realization that there are few performance assessment instruments of science process skills, and most of the available instruments do not have documented reliability and validity standards. This study examines the reliability and validity of a set of science performance assessment instruments through the application of currently available statistical methods. The Performance Assessment of Science Skills (PASS) tests, which were formulated by S. M. Johnson as an alternative assessment of science process skills in the elementary school grades, are the subject of investigation.Third-grade students completed PASS instruments which measured the process skills of observation, binary classification, and measurement, and fifth-grade students completed PASS instruments which measured the science process skills of prediction, designing a controlled experiment, and hierarchicalclassification.Reliability and validity for the PASS instruments focused on alternate form reliability, internal reliability, inter-rater reliability and construct validity. This study revealed that statistical reliability and validity can be documented for some alternative assessments of science process skills. / Department of Biology