In the last decade, the wheat stem sawfly [Cephus cinctus Norton (Hymenoptera: Cephidae)] has resurged as an important pest of wheat in the Canadian prairies and continues to be a chronic pest in the Northern Great Plains of the USA. Ecological and management studies to determine egg and larval infestation, damage and parasitoid attack rates, require laborious dissections of stems collected at various spatial scales. We used a statistical simulation study to determine the minimum number of stems required to estimate these response variables at the level of a sub-sample (e.g., within a plot). The number of stems required to estimate sawfly cutting damage and parasitoid attack to larvae was strongly and negatively related with the response variable. At moderate to high levels of sawfly pressure where the stems cut by larvae exceeds 40%, it is possible to reduce stem counts to 50 stems; however, in the 10% cutting range, up to 200 stems are needed for accurate estimates. These values were similar for sample size required to estimate larval parasitism but egg infestation of stems, when levels surpass 70%, can be determined with as few as 30 stems.