Module 5

Question 1

What is internal validity?

Answer 1

Findings are free of chance, bias and confounding. IS there something making the study findings wrong?/Are there other explanations for the results?

Question 2

What is a parameter?

Answer 2

True values of the measure in the population

Question 3

What is an estimate?

Answer 3

The value found in the study sample

Question 4

What happens if you repeatedly sample randomly from the same population?

Answer 4

Most of the time you would get a sample with a similar composition to the population you sampled from but some of the samples would be quite different just by chance. This is sailing error which is a form of random error and is commonly just called chance

Question 5

What can reduce sampling error?

Answer 5

Increasing sample size

Question 6

What is external validity?

Answer 6

• whether findings can be applied to broader populations
• generalisability
• judgement call about what has being studied applies to

Question 7

How is a confidence interval interpreted?

Answer 7

We are (confidence level), that the parameter lies between (lower bound) and (upper bound)

Question 8

What is interpreted first?

Answer 8

Measure of association and then the confidence interval

Question 9

What happens if the whole CI is above the clinically important value?

Answer 9

Then the study findings are clinically importance

Question 10

What happens if part of the CI is above the clinically important value?

Answer 10

The study findings may have clinical importance

Question 11

What happens if the CI excludes the clinically important value?

Answer 11

The study findings have no clinical importance

Question 12

What are p values?

Answer 12

The probability of getting a study estimate (or one further from the null), when there really is no association due to sampling error. If p is really low then it is unlikely the estimate is due to sampling error

Question 13

What is the null hypothesis?

Answer 13

There is no association and the parameter = null value

Question 14

What is the alternative hypothesis?

Answer 14

there is an association and the parameter does not equal the null value

Question 15

How is a p value interpreted?

Answer 15

The probability fo finding a (sample statistic) of (value) or further from the null when the null hypothesis is true is (p-value)

Question 16

What is set for a type 1 error?

Answer 16

A threshold which is usually 0.05

Question 17

What happens if the p value is less than 0.05?

Answer 17

Since the p-value is less than 0.05 the association is statistically significant. We reject the null hypothesis and accept the alternative hypothesis. Chance is an unlikely explanation of the study finding

Question 18

What happens if the p value is more than 0.05?

Answer 18

Since the p-value is more than 0.05 the association is not statistically significant. We fail to reject the null hypothesis and reject the alternative hypothesis. The study finding is consistent with chance as an explanation

Question 19

what is a type 2 error?

Answer 19

When we incorrectly fail to reject the null hypothesis when should have (p should have been less than 0.05 but it was more)

Question 20

What is a type 2 error typically due to?

Answer 20

Having too few people in the study

Question 21

What can statisticians do?

Answer 21

Calculate power to find out how many participants are needed to minimise chance of a type 2 error

Question 22

Why are p values problematic?

Answer 22

Arbitrary threshold, only about the null hypothesis and nothing about importance

Question 23

What is meant by and arbitrary threshold?

Answer 23

Sometimes it isn’t too far from 0.05 so the p-value should be reported as well as saying it is statistically significant or not statistically significant

Question 24

What is meany by it is only about the null hypothesis?

Answer 24

It just gives evidence about consistency with the null hypothesis nothing about precision which is why they are best presented with confidence intervals

Question 25

What is meant by there is nothing about importance?

Answer 25

Statistical significance is not the same as clinical significance, doesn’t say anything about whether results are valid, useful or correct in terms of confounding and bias. Absence of a statistically significant association does not mean absence of a real association