: p-value in statistics
In: 10
A p-value is related to the null and alternative hypothesis. Those sound weird but they’re a little different than a hypothesis you might have seen in science class.
A null hypothesis says that there is nothing special about the relationship between two things. (Ex: Eating sugar does not increase obesity)
The alternative hypothesis says that there is something special about the relationship. (Ex: Eating sugar does increase obesity)
Now, we have to collect a lot of data from a bunch of people eating different amounts of sugar and seeing if they gain weight. But data is rarely clear. So they use what’s called a t-test to see if there is a legit difference between the group of people that ate sugar and the group of people that didn’t. The t-test spits out a number. This number, between 0 and 1, is compared to the p-value, a threshold. If the p-value is greater than 0.05, you reject the alternative hypothesis. So if we did this experiment and found that our p value was .06, we could say that random error or chance could cause variation of some people being obese and some not. Now we *could* set the p-value at 0.3 or 0.8, but it’s kind of defaulted at 0.5.
This doesn’t mean that the experiment failed or that the other side is “wrong”, it just means that in a world where sugar does not lead to obesity and the two sets of people were the same, here’s how likely you are to see the same results you got. So when that number is really small (p-value less than 0.5), it’s really not expected. (Note: this doesn’t mean impossible, just improbable)
Latest Answers