258 • Big Data Analytics
13%. In other words, the probability of having at least one Type I error
tops 87%. In fact, we probably have more than one such error. By the time
we have 100 independent hypothesis tests, then the odds of having at least
one Type I error exceed 99%. In fact, we can expect somewhere around
ve (100 trials times 5% of a Type I error).
e outcome of our statistical test can be conceptually similar to
the odds of winning the lottery. e odds of winning the Powerball (a
US-based lottery) grand prize are 1 in 175,223,510.
15
e odds of being
struck by lightning, in comparison, are 1 in 280,000.
16
Most people do
not know anyone who has been struck by lightning, yet the odds of being
struck by lightning are approximately 6 ...