This paper outlines the development of a quantitative instrument designed to measure students' multiple achievement goals and key aspects of their cognitive engagement in Australian educational settings. The paper demonstrates the use of congeneric measurement models in assessing and improving the validity of the scales comprising the Goal Orientation and Learning Strategies Survey (GOALS-S), and it also illustrates the use of cross-validation as a way to assess the stability of the GOALS-S scales. Three academic goals, four social goals, three cognitive strategies, and three metacognitive strategies were defined and operationalized in the developed GOALS-S instrument. Preliminary factor analyses were used to refine the measure. Its 13 scales were tested with 380 Australian urban students. Once collected, data from a randomly chosen subsample (half the original sample) were subjected to confirmatory factor analysis using 13 separate one-factor congeneric models. Modified models that showed sufficient fit in the first half were tested with the second half of the sample. The initial modeling process supported the construct validity of only four of the original scales, but relatively minor modifications resulted in considerably better fit for the other nine scales. Results also demonstrate support for the reliability of at least eight of the GOALS-S scales. (Contains 6 tables and 88 references.) (SLD)