Statistics for Zombies

I have a confession to make. Lately I’ve been avoiding writing about Statistics (PSYCH 105), though it is one of the courses in which I’m enrolled. It seemed a difficult subject about which to write in any sort of, shall we say, non-boring fashion. The thing is, though, statistics is a massively important part of psychology. Without proper, standardized, evaluative measures any sort of “experiment” could only be written up in the form of anecdote, and any shocking advances in the subject would needlessly be relegated to the realm of urban legend, because how would we know if the experimenter weren’t engaging in some serious confirmation bias? For the following example, I will attempt to veer from the beaten path for the sake of education.

 

Consider a sample of n = 36 zombie hunters who’ve taken a zombie survival class for one year prior to the zombocalypse. They average a kill-count of M = 5 zombies a day, with the added benefit of not getting eaten. The entire rest of the population from which this sample is taken may or may not have taken such a class and has a mean kill-count of μ = 3 per day, and probably get eaten. Assume a standard deviation of σ = 1 and standard level of significance of α = 0.05 with a two-tailed test. Finding the z-score (or, here, zombie score) is now a simple matter of formula.

z = (M - μ) / (σ / √ n)

(5 - 3) / √36 = 0.333 = z

 

Such a low z-score cannot possibly recommend the course as being statistically more significant. The null hypothesis cannot be rejected, so one must assume that those who took the course are equally as likely to be eaten as those who didn’t take the course.

 

Thank goodness I took statistics before the world blew up.