Week 6 - Parametric statistics and test assumptions

Question 1

What are the three types of experimental design?

Answer 1

1) Independent groups
2) Matched pairs
3) Repeated measures

Question 2

What is independent groups design?

Answer 2

The sample is split into two groups. Each group does one of the experimental conditions. Also known as a between subject design

Question 3

What is matched pairs design?

Answer 3

Same as independent groups, however each participant is matched on important characteristics with someone in the other group

Question 4

What is repeated measures design?

Answer 4

The same group of participants takes part in both experimental conditions. Also known as a within-subject design

Question 5

What is the parametric test for independent groups?

Answer 5

Independent-samples t-test

Question 6

What is the non-parametric test for independent groups?

Answer 6

Mann-Whitney

Question 7

What is the parametric test for matched pairs?

Answer 7

Paired-samples t-test

Question 8

What is the non-parametric t-test for matched pairs?

Answer 8

Wilcoxon

Question 9

What is the parametric test for repeated measures?

Answer 9

Paired-samples t-test

Question 10

What is the non-parametric test for repeated measures?

Answer 10

Wilcoxon

Question 11

What is the difference between parametric and non parametric tests?

Answer 11

Parametric tests such as t-tests are calculated from the data using the mean and standard error

Non-parametric tests such as Man-Whitney or Wilcoxon are computed from ranked scores

Question 12

What is the assumption for parametric statistics?

Answer 12

Assume that the data you have collected come from a population that can be modelled on a normal distribution

Question 13

What is the assumption for non-parametric statistics?

Answer 13

Are sometimes referred to as distribution-free because they do not make that assumption

Question 14

What are the three assumptions that need to be met to use a parametric statistic?

Answer 14

1) The data needs to be at least interval level (continuous)

2) Assume that the data you have collected came from a population that can be modelled on a normal distribution

3) If you are comparing two groups it assumes that they have similar variance

Question 15

What is nominal data?

Answer 15

Information that is put into categories or just named (e.g. female/male, noise/no noise) (nominal)

Question 16

What is ordinal data?

Answer 16

Information or scores that are put in order or is ranked (e.g. level of education - secondary school to post-graduate) (ordinal)

Question 17

What is interval data?

Answer 17

There are equal intervals on a measurement scale (e.g. temperature - you can have minus numbers) (scale)

Question 18

What is ratio data?

Answer 18

Same as interval but there is a true zero point (0 = nothing) (e.g. height or memory) (scale)

Question 19

What does a normality test determine?

Answer 19

A normality test is used to determine whether sample data has been drawn from a normally distributed population

Question 20

What is a bell shape curve?

Answer 20

Symmetrical distribution around centre of all scores

Majority of scores rest around centre (one peak), with similar frequencies at the extremes

Question 21

What is a skew?

Answer 21

More developed on one side or in one direction than another

Question 22

What is kurtosis?

Answer 22

The sharpness of the peak of a frequency - Pointiness/heaviness

Question 23

What does a negative skew look like?

Answer 23

Pile of scores on the right so leans right

Question 24

What does a positive skew look like?

Answer 24

Pile up of scores on the left so leans left

Question 25

What does a negative kurtosis look like?

Answer 25

Flat

Question 26

What should a normal kurtosis value be around?

Answer 26

0

Question 27

What does a positive kurtosis look like?

Answer 27

Pointy

Question 28

What are the two broad methods of assessing normality?

Answer 28

1) Numerical methods such as skewness/kurtosis values and statistical tests

2) Graphical methods such as inspection of graphs

Question 29

What does the assumption of homogeneity of variance refer to?

Answer 29

Whether or not the two groups have a similar variance

Question 30

What is homogenous variance mean?

Answer 30

Both groups have similar variances

Question 31

hat does heterogenous variance mean?

Answer 31

The groups have different variances

Question 32

When should you use a non-parametric test?

Answer 32

When your sample size isn’t large enough to satisfy the requirements in the table above (more than 2 IV levels) and you’re not sure that your data follows the normal distribution

Question 33

What is an advantage of using a non-parametric test?

Answer 33

They can analyse ordinal and ranked data, and not be tripped up by outliers

Question 34

What is a negative about parametric tests?

Answer 34

They can only analyse continuous data and the findings can be affected by outliers