Author
Summary
When outcome distributions are unimodal - symmetric, what is Expected is what is Most Likely, but when they are not it is not. Indeed, in non-unimodal symmetric situations Expected and Most Likely outcomes can be very different with expected outcomes being the less likely prospects, which begs the question: why not use Most Likely or Modal values in uncertain situations? The issue is generic and pertinent in many diverse contexts. Here, to emphasize that diversity, it is explored in the context of four quite disparate examples: Canadian Life Expectancy, UK Health Outcomes, Spanish Parental Circumstances and Portfolio Choice in US Stock Markets. The broad conclusion common to each of these non-symmetric unimodal situations is that employing Most Likely values makes a substantial difference to the analysis which should give investigators pause for thought.