variables are measured without error(reliably).
the nature of education and social science research means that many variables the researchers are interested in are also difficult to measure,making measurement error a particular concern. In simple correlation and regression,unreliable measurement causes relationship to be under-estimated increasing the risk of Type 2 errors. In the case of multiple regression or partial correlation, effect sizes of other variables can be over-estimated if the covariate in not reliably measured, as the full effect of the covariate wounld not be removed.