Karamanos, R., Flaten, D. N. and Stevenson, F. C. 2014. Real differences - Lessons from an agronomist's perspective. Can. J. Plant Sci. 94: 433-437. An experiment including a two penny treatment and an untreated check was established to show that implausible differences can be a statistical reality. An individual crop by location analysis showed that two pennies significantly (P<0.05) increased canola yield at one of the 19 locations, nearly (0.05<P<0.18) affected canola yield at one other location. A combined mixed model analysis showed that canola yield significantly increased by a small amount (0.1 t ha-1) with two pennies, whereas, a similar mixed model that accounted for residual variance heterogeneity showed that two pennies did not affect crop yield. Our results confirmed that the effect of a treatment not expected to cause a meaningful difference can be detected The results also highlight the importance of modeling all sources of variance, designing more efficient experiments, scrutinizing the size of treatment differences, and choosing an appropriate level of significance to ensure that only real differences are detected.