Null Hypothesis Testing (Foundational)

Question 1

What is null hypothesis testing used for?

Answer 1

A standard procedure for testing a claim (a hypothesis) about a population parameter or relationship

Question 2

Null Hypothesis

Answer 2

Statement of no difference, no relationship, no effect.

Question 3

Alternative Hypothesis

Answer 3

Typically associated with the claim.

Can be two-tailed: not equal, we’re not making a lesser greater claim. Simply not equal to.

left-tailed: less than

right-tailed: right tailed

Question 4

Are the alternative hypothesis arbitrarily chosen?

Answer 4

No! You can’t pick a test for no reason, there needs to be a specific reason for choosing a left or right tailed test.

Two-tailed is the default.

Question 5

How do we determine if the sample we choose is unusual?

Answer 5

We can create a binomial sampling distribution (dealing with a proportion), with a probability of success equal to our population probability of success, and n= sample size.

Then cumulative probability is assigned to our sample statistic (p-value), this probability tells us how far our sample is from the mean.

Question 6

What is the P-Value what does it tell us?

Answer 6

probability of observing sample results as extreme or more extreme (as defined by HA), if the model is correct.

Pvalue helps us determine if our sample statistic is unusual. (Values in the tails of the distribution will have a lower pvalue and be less probable)

Pvalue tells us if we should accept the null hypothesis

Question 7

The Rare Event Rule

Answer 7

If the probability—under a given assumption—of a particular observed result is small, then we have evidence that the assumption may not be correct.

Question 8

What should inform your decision of the ‘strength’ of

Answer 8

The representativeness of your sample data.
The validity of your data collection procedure (study design.

The possible implications of your conclusion on society and the environment.