The Average Is Nobody's Result

A 2013 mammography study found that computer-aided detection (CAD) had a mixed effect on radiologists' performance. The average result was a null, but further analysis showed that CAD helped less skilled radiologists and hindered more skilled ones. This highlights the dangers of relying on averages, which can mask underlying effects. As a result, IT teams should be cautious when interpreting metrics and consider the individual effects on different team members. This can help avoid misinterpreting the impact of new tools and processes.