Branches of 1.4 substitutions/site at depth 2 were underestimated by approximately 55%, compared to 30% underestimation for depth 1.
In a study of 147 fast food restaurant diners at food courts, Chandon and Wansink also found that people underestimated the calorie content of purchased meals, with larger underestimation for higher calorie meals and no association with recognition of nutritional information in the restaurants.
Consequently, branches of 1.4 substitutions/site were underestimated by approximately 35%, which is slightly greater than the underestimation for the same branch lengths when using the default prior (30%; Figure 2).
The expected rate of underestimation for depth 2 branches for these 4-taxon simulations was derived from underestimates for depth 2 branches for 8-taxon simulations.
Model performance was characterized by overestimation for ozone and underestimation for the other pollutants.
In the metric of transmission capacity, there still exists underestimation for the Matérn type II model.
Still, the relative deviation predicted an underestimation for the solar power generation of >5%.
Mean underestimation for EDV varied between26%and52%52% and for ESV between155% and 60%.
There was a slight tendency towards underestimation for low volumes and overestimation for high volumes.
In several cases, we found an underestimation for SAGE II derived values with respect to corresponding in situ values.
Applying Mika Mannermaa's ideas of making observations there could have been paradigm blindness, trend thinking and underestimation, for example.
