Interocular Results

Fred Mosteller at Harvard University advocates a reality-oriented approach to statistics that is very applicable for the life sciences. He coined the term, ‘interocular differences’ in contrast to ‘statistically significant differences.’ His approach is for statisticians to go ahead and use statistically significant differences in research because they may find something worth further investigation. But don’t come to him until you’ve found interocular differences. With new innovative research we’re not trying to tease out small incremental benefits; we seek benefits that are obvious even without the use of statistics, and even in small patient populations. In other words, if the difference doesn’t hit me between the eyes, hence the term interocular, I don’t care about statistical significance.  Of course this approach does not apply for product safety concerns, which are only very rarely ‘interocular.’