The purpose of present study was to develop an instrument to measure perceptions on good instruction for middle and high school students and to evaluate the scale¡¯s construct and criterion-related validity. The participants were 998 8th and 11th graders. I performed exploratory and confirmatory factor analysis for construct validity and analyzed and structure equation modeling among perspectives on good instruction, course satisfaction, and academic achievement for criterion reference validity. I found that the final version of PGI(perceptions good instruction scale) consisted of a five-factor solution (relationship, transmission, constructivist, product, and interest) and 30 items. Results of structure equation modeling revealed that perceptions on good instruction had direct and indirect effects on academic achievement mediated by course satisfaction.