Studies using tests scores as the dependent variable often report point estimates in student standard deviation units. We note that a standard deviation is not a standard unit of measurement since the distribution of test scores can vary across contexts. As such, researchers should be cautious when interpreting differences in the numerical size of point estimates when comparing across contexts. We empirically assess the importance of this issue both within the United States and internationally.
Available at: http://works.bepress.com/jeffrey-schiman/8/