I’ve been feeling kind of betrayed lately. And, to be frank, I blame the groundhog.
You can’t just tell people that there are only going to be six more weeks of winter. And then pummel them with 20+ inches of snow. That’s right. Over. Twenty. Inches. In Central Park, the snow amounted to more than a third of my height.
Now that. Is just unethical.
So then the real question becomes. Where’s an IRB when you need one? And has anyone ever calculated the statistical significance of this Groundhog Day phenomenon? Because, not to burst anyone’s bubble, but it really seems to me like the p-value for this whole operation is definitely less than 0.05. Meaning, obviously (at least to those of us who were just forced to take a course in epidemiology and biostatistics), that the association between winter’s length and whether or not Punxsutawney sees his shadow. Is entirely up to chance. No correlation. No positive predictive value (or negative predictive value). Pure serendipity.
Although we’d have to calculate the beta error to be sure. Odds are. We just don’t have enough statistical power.
