Models can always be improved by labelling cases that don’t fit as “outliers”, and removing them.
“Tropical storms” were excluded from the analysis, because they didn’t provide the desired result.
Simpler analyses were abandoned because they didn’t provide the desired result:
the authors experimented with using the hurricane name as a binary variable (0 for a male name, 1 for a female name) and also tried, as an explanatory variable, “years elapsed since the occurrence of hurricanes… However, this variable was dropped for the main analysis as its effect was nonsignificant in all models.”
