2.9. - Null hypothesis testing

Question 1

Who came up with the idea of p-values?

Answer 1

Fisher

Question 2

Who believed that there are 2 competing hypotheses?

Answer 2

Neyman and Pearson

Question 3

What are 2 competing hypotheses?

Answer 3

Alternative hypothesis

Null hypothesis

Question 4

What symbols represent the alternative hypothesis?

Answer 4

H1

Question 5

What symbols represent the null hypothesis?

Answer 5

Ho

Question 6

What does an alternative hypothesis state?

Answer 6

States that there will be a difference in outcome

Question 7

What does a null hypothesis state?

Answer 7

States that there won’t be a difference in outcome

Question 8

What are 2 types of alternative hypothesis?

Answer 8

Directional

Non-directional

Question 9

What is a directional hypothesis also called?

Answer 9

One-tailed hypothesis

Question 10

What is a non-directional hypothesis called?

Answer 10

Two-tailed hypothesis

Question 11

What does NHST stand for?

Answer 11

Null Hypothesis Significance Testing

Question 12

What is a test statistic?

Answer 12

A statistic for which we know how frequently different values occur

Question 13

How do we calculate test statistic?

Answer 13

effect / error
variance explained by model / variance not explained by model
signal / noise

Question 14

What is the cut-off point for confidence?

Answer 14

0.05, or 5%

Question 15

What does it mean if p is less than .05?

Answer 15

The test statistic is significant and we reject the null hypothesis

Question 16

Who identified type I and type II errors?

Answer 16

Neyman and Pearson

Question 17

What is another name for a type I error?

Answer 17

False posItIve

Question 18

When do type I errors occur?

Answer 18

When we think we found something, but in reality there was no effect. We wrongly reject the null hypothesis

Question 19

What is the value of alpha?

Answer 19

Usually 0.05

Question 20

What does the value of 0.05 probability mean?

Answer 20

If we replicated our data collection where we found no error 100 times, we could expect that 5/100 times we would obtain a significant test statistic. We only make a type I error 5% of the time.

Question 21

What is another name for a type II error?

Answer 21

False negative

Question 22

When do type II errors occur?

Answer 22

When we believe that there is no effect, when actually there is. We wrongly accept the null hypothesis

Question 23

What did Cohen suggest?

Answer 23

The maximum acceptable probability of a type II error is 0.2. This is known as beta.

Question 24

What does 0.2 probability mean when looking at type II errors?

Answer 24

20/100 times we would fail to detect effects in the population

Question 25

What factors affect the value of beta?

Answer 25

Strength of effect, type of design, sample size

Question 26

Who decides the value of alpha?

Answer 26

The researcher

Question 27

What does the term familywise error mean?

Answer 27

The increase in error rate across several statistical tests conducted on the same data.

Question 28

How do we calculate the familywise error?

Answer 28

1 - 0.95 to the power of the number of tests carried out on the data

Question 29

What is another name for familywise error?

Answer 29

Experimentwise error

Question 30

How do we prevent an increase in error rate between several statistical tests?

Answer 30

Adjust the level of significance by dividing the value of alpha by the number of comparisons

Question 31

Define statistical power

Answer 31

The ability of a test to find an effect

Question 32

How do we calculate statistical power?

Answer 32

1 - beta

Question 33

What is the general value for beta?

Answer 33

0.2

Question 34

What power do we aim to achieve?

Answer 34

0.8

Question 35

What does a power of 0.8 suggest?

Answer 35

0.8 is significant. Equates to an 80% chance of detecting an effect if one genuinely exists

Question 36

What factors affect the power of a statistical test?

Answer 36

How big the effect is How strict we are about deciding if an effect is significant Sample size

Question 37

What affect does the sample size have on the amount of sampling error?

Answer 37

The larger the sample, the less sampling error there will be