What a P-value tells you in Statistics…

[stonenelson]

You made it clearer in a YouTube short than my professor has in a solid month. Bless you

[prod.kashkari]

One thing one of my professors did was tell as what the p value means visually. He showed the graph with the test statistic marked as a quantile, and the area shaded to the right, being the rejection region (we were doing a 1 sided t test). The rejection region represents values of your observed test statistic that are “extreme” ie are deep in the tails of the distribution and hence in the rejection region. Then, the p value is the probability to the right of the extreme test statistic.

If you have a p value very low, then this means that there was an extreme value in the tail of the distribution, and then hence the probability to the right of it is Tiny (since it’s in the tail and it’s extreme), but if it’s a large test statistic, it’s probability to the right is very large, so it’s probably less than the quantile of our test statistic, and hence not in the rejection region, so we fail to reject.

[dreaminderek]

Very noiceee tip

I usually prefer the computed statistic/critical value method but yes, P-values are important as well.

[chacmool]

More intuitively: the p-value is the probability that the Null Hypothesis is true given the observed values. So, knowing what the Null Hypo that is being tested says is of the essence. Sometimes you want a high p-value (e.g. Ljung-Box which tests the Null Hypo that the residuals are not correlated) and sometimes you don't (e.g. regression coefficients where the Null Hypo is that the coefficients are zero). 😊

[oyetundebello]

Correct me if im wrong but i think the p value is the probability of picking a sample the contradicts the alternative hypothesis, so if the probability of picking a sample that contradicts your own hypothesis is high then you have to go with the null and if it is vice versa you reject the null🤔

[tobysavory]

Small issue, I can’t figure out what the null hypothesis is meant to be, it’s the simple things that stump me

[ireneirnlg]

Thank youuu🥹 I have a Statistics2 exam tomorrow 🥲

[anubhavhazra]

Bro I have one
Please answer it.
Suppose x is a variable having values of heart rate (number of heart beats per minute), and the possible values are between 60 to 100 (integer values)
Is the variable "x" a discrete variable or a continuous variable?

[zachmorgan]

Yo man...idk what a p value is...but I love statistics

[someones.]

do you know any online course that you would recommend in statistics?