Because accountability may improve the comparability that is compromised by lenient grading, we compare exit exam outcomes in the same schools before and after a policy change that increased teacher accountability by anchoring grading scales. In particular, using a large administrative dataset of 364,445 exit exam outcomes for 72,889 students, we assess the effect of introducing centralized scoring standards into schools with higher and lower quality peer groups. We find that implementation of these standards increases scoring differences between the two school types by about 25 percent.