An experiment was performed to evaluate the tenability of the assumption in the Elementary Secondary Education Act (ESEA) Title I proposed variance estimation procedures for criterion referenced tests. The assumption is that the ratio of the local to the national standard deviation for the national sample will be the same for the normed test as for the non-normed test. This assumption and the relevant procedure have been proposed only for the study of treatment effects where they depend on the use of aggregated data and not for evaluation at the local level. Two nationally normed reading achievement tests were used, and comparisons were made of actual standard deviations for norm groups with estimates of norm group standard deviations using the ESEA procedure separately for five different grade levels. All five estimates were above the actual values in error by eight percent to 57%. Until further study is made of the errors made in estimating population standard deviations, it appears that the use of non-normed tests in Title I evaluations is a questionable practice. (Author/CTM)