Recent research indicates that estimates of the effect of research and development (R&D) on total factor productivity growth are sensitive to different measures of total factor productivity. In this paper, we use establishment level data for the flat glass industry extracted from the Census Bureau's Longitudinal Research Database (LRD) to construct three competing measures of total factor productivity. We then use these measures to estimate the conventional R&D intensity model. Our empirical results support previous finding that the estimated coefficients of the model are sensitive to the measurement of total factor productivity. Also, when using microdata and more detailed modeling, R&D is found to be a significant factor influencing productivity growth. Finally, for the flat glass industry, a specific technical change index capturing the learning-by-doing process appears to be superior to the conventional time trend index.