Sessile serrated adenoma (SSA) is the proposed precursor for microsatellite unstable colorectal carcinomas and some authorities recommend that SSAs should be managed similar to adenomas. The aim of our study was to determine whether serrated polyps can be classified with sufficient consistency to support current treatment recommendations. One hundred eighty-five serrated polyps were classified as hyperplastic polyp (HP), SSA, or traditional serrated adenoma (TSA) by 5 pathologists blinded to clinical data. The observers documented which histologic features they considered most helpful in reaching their diagnosis in each case. In a second round, the observers were provided with polyp site and size. After reaching a consensus on minimum criteria for SSA and TSA, the pathologists classified another set of 50 polyps. The interobserver concordance was calculated using kappa statistics. In the first round, the overall interobserver agreement was moderate (kappa=0.55). Concordance for HP and SSA was moderate whereas it was nearly perfect for TSA. In the second round, there was no improvement in the concordance. All observers relied more often on architectural features than on cytologic ones to distinguish SSA from HP and agreement was reached that architectural features should provide the basis for the diagnosis of SSA. Subsequently, interobserver concordance was slightly improved but remained moderate (kappa=0.58). Interobserver agreement for the diagnosis of serrated polyps is moderate. However, this level of variability is acceptable because the presence of SSA indicates increased risk of developing additional serrated polyps and carcinoma, and surveillance is appropriate.