P value as cumulative probability

155 Views Asked by At

I apologize for having a very basic question that I've not been able to find the answer to. Using a cumulative probability table, I find that the cumulative probability of flipping 7 or fewer heads in 23 coin tosses ≈ 0.0465. However, I fail to reproduce this figure when I try to calculate the probability manually with a z table.

The standard deviation of a binomial distribution is sqrt(np(1-p)), in this case sqrt(23 x .5 x .5) ≈ 2.3979. The mean of the distribution is np = 11.5, so the z score of flipping 7 heads is (7 - 11.5) / 2.3979 ≈ -1.8766.

Looking up -1.8766 in a z table, I find that the p value of it ≈ 0.0307. My understanding is that the p value reflects the cumulative probability of obtaining a result less than or equal to the one that many (-1.8766) standard deviations away from the mean; thus, why is this p value so different from the 0.0465 that I found in the cumulative probability table?

I know that I'm just misunderstanding something very fundamental, but I've been reading all up on z tables and just can't put my finger on it. Thanks so much for any help!

1

There are 1 best solutions below

0
On BEST ANSWER

Perhaps the difference you're seeing is due to the continuity correction. Did you do the calculation with the continuity correction? In any case a p-value is a tail probability, usually one minus a cumulative for right tailed, but could also involve adding up both tails.