The effect on achievement when the prescribed maximum time limits of a standardized achievement test are exceeded was studied using students nested in block-by-treatment combinations. The analysis was a hierarchical (nested) randomized blocks analysis of covariance involving 408 fifth graders taking the Stanford Achievement Test in Lansing, Michigan. Results indicated that excess testing time had a significant positive linear effect on reading comprehension and total reading scores and a significant nonlinear effect on the same outcomes. The benefit of excess time was most pronounced between maximum time and 5 minutes more and between 10 and 15 minutes excess time. No significant interactions were found between excess time and selected demographic variables. It was clear that decisions made at the classroom level were substantially more sensitive to the effects of excess time than were student level decisions. Results highlight the importance of excess time at various levels of aggregation. (Contains 9 tables and 24 references.) (SLD)