Lecture 24: Chance ll

Question 1

What are p-values?

Answer 1

The probability of getting a study estimate (or a study estimate further from the null), when there is really no association, because of sampling error.

Question 2

Do we want to p-value to be high or low?

Answer 2

Ideally we want the p-value to be really low because then it means that it is unlikely that the estimate is due to sampling error (chance alone).

Question 3

The p-value tells us about

Answer 3

when there is no association in the population but in our sample, we find an association

Question 4

What is the null hypothesis?

Answer 4

When there is really no association in the population so the parameter equals the null value

Question 5

What is the alternative hypothesis?

Answer 5

when there is really an association in the population so the parameter does not equal the null value

Question 6

What is the threshold P-value?

Answer 6

0.05

Question 7

If the p-values is smaller than the threshold of 0.05, we reject the ______ hypothesis, we accept the ________ hypothesis. Is the association statistically significant?

Answer 7

null hypothesis
alternative hypothesis
yes, it is statistically significant

Question 8

If the p-values is greater than the threshold of 0.05, we ___ ___ reject the ______ hypothesis, we _____ the ________ hypothesis. Is the association statistically significant?

Answer 8

fail to 
null
reject
alternative
the association is not statistically significant

Question 9

What is a type 1 error?

Answer 9

when there is no association in the population but in our sample, we find an association

Question 10

What is a type2 error?

Answer 10

When when there is an association in the population but in our sample, we find no association

Question 11

In a type 2 error, is the p value bigger or smaller than it should be?

Answer 11

bigger (ie. greater than 0.05)

Question 12

Are you more likely to get a small p value with a bigger or smaller sample size?

Answer 12

bigger

Question 13

What is the relationship between the 95% confidence interval and the p value when the 95% CI includes the null value?

Answer 13

• the association is not statistically significant
• the p > 0.05
• we fail to reject the null hypothesis, reject the alternative hypothesis

Question 14

What is the relationship between the 95% confidence interval and the p value when the 95% CI does not include the null value?

Answer 14

• the association is statistically significant
• the p < 0.05
• we reject the null hypothesis, accept the alternative hypothesis

Question 15

What are three reasons that the p-value can be problematic?

Answer 15

• it is just an arbitrary threshold
• it is only about the null hypothesis
• is says nothing about importance

Question 16

What is meant by arbitrary threhold?

Answer 16

p = 0.05 is just a number chosen; is there really much difference between p = 0.04 and 0.06. This means that it is always useful to report p-values rather than just ‘statistically significant’ or ‘not statistically significant’

Question 17

What is meant by the p value being only about the null hypothesis?

Answer 17

It just gives evidence about the consistency with the null hypothesis and does not say anything about precision

Question 18

What is meant by the p-value being nothing about importance?

Answer 18

It looks at statistical significance not clinical significance which means it does not say anything about whether the results are valid, useful or correct