English Deutsch Français Italiano Español Português 繁體中文 Bahasa Indonesia Tiếng Việt ภาษาไทย
All categories

This is for a stats class. Im pretty sure its 0 < p < 1, but Im having doubts in the back of my mind Can someone verify this?

2006-11-22 10:38:20 · 5 answers · asked by Anonymous in Science & Mathematics Mathematics

5 answers

p is a probability. So you're almost right - the question is, can you have a p value of 0 or 1?
And the answer is yes. Although nearly every single statistical test you will use will never give 0 or 1, it is theoretically possible.

2006-11-22 10:41:52 · answer #1 · answered by stephen m 4 · 0 0

That's always the safe answer because any probability must be in that range. The p-value is the probability of getting a test statistic worse than what you've calculated. Note that worst depends on whether you're doing a one or two sided test. As an expample, lets say that you have a normal distribution and your test statistic in the mean. For a one sided test the p-value is 1/2 for a two sided test it's 1.

2006-11-22 10:44:16 · answer #2 · answered by modulo_function 7 · 0 0

The "p-value" of a test usually identifies the probability of the test results happening by luck or random chance. So the p-value is a probability, which can be zero or one, or any number in between.

2006-11-22 11:36:07 · answer #3 · answered by Polymath 5 · 0 0

We'd need to know your specific problem before we could say.

In general, the probability of any event is always somewhere between 0 to 1, inclusive; i.e., 0 ≤ P ≤ 1.

~ ♥ ~

2006-11-22 10:40:25 · answer #4 · answered by I ♥ AUG 6 · 0 0

the p-value is given the null is true, this is the probability that you will draw an extreme value from your sample. and yes for hypothesis testing the p-value must be greater than zero and less than 1

2006-11-22 12:29:38 · answer #5 · answered by jojothedogfacewonderbody 2 · 0 0

fedest.com, questions and answers