10. Exploring Assumptions

Question 1

steps to conducting statistical analyses

Answer 1

• explore your data
• check assumptions
• conduct statistical tests

Question 2

How do you explore your data?

Answer 2

• graphs

- run descriptive stats

Question 3

types of data

Answer 3

parametric

nonparametric

Question 4

assumptions about parametric data

Answer 4

• normally distributed
• homogeneity of variance
• at least interval data
• independence

Question 5

independent samples

Answer 5

data from different subjects are independent

Question 6

dependent samples

Answer 6

behavior of one subject doesn’t influence behavior of another

Question 7

What does it mean by normally distributed?

Answer 7

• sampling distribution and sample data are both normally distributed
• central limite theorem

Question 8

How to assess normality

Answer 8

• visually via graphs
• descriptive statistics
• comparison to normal distribution and assess for differences

Question 9

two graphs to use to assess normality

Answer 9

• histograms

- p-p plots

Question 10

How might histograms be useful for assessing normality?

Answer 10

• frequency distribution

- can add a normal distribution overlay

Question 11

What is a p-p plot?

Answer 11

probability-probability plot

Question 12

What does a p-p plot do?

Answer 12

• plots probability of a variable against the probability of a normal distribution

Question 13

What does a p-p plot convert scores to? Why?

Answer 13

z-scores

to compare against z-scores of normally distributed data

Question 14

data from the sample

Answer 14

actual/observed probability

Question 15

normally distributed data

Answer 15

expected probability

Question 16

What sort of descriptive stats would you use to assess normality?

Answer 16

• measures of central tendency
• measures of variability
• measures of shape

Question 17

What are the measures of shape used to assess normality?

Answer 17

skewness

kurtosis

Question 18

How do you compare data to a normal distribution?

Answer 18

2 tests can be used to determine

• Kilmogorov-Smirnov test
• Shapiro-Wilk test

Question 19

What is the benefit of the Shapiro-Wilk test?

Answer 19

more power to detect differences from normality

Question 20

With tests that compare to normal distribution (Kolmogorov-Smirnov and Shapiro-Wilk), what does P > 0.05 mean?

Answer 20

indicates that there’s no difference between the sample distribution and normal

Question 21

What are the limitations to tests that compare to normal distribution?

Answer 21

• not always accurate with large samples

- small changes can lead to significant test results

Question 22

What must you always do in addition to running tests?

Answer 22

graph the data

Question 23

What does homogeneity of variance mean?

Answer 23

the spread of scores around the mean should be similar in each group

Question 24

What type of design does homogeneity of variance apply to?

Answer 24

non-repeated measures designs

Question 25

How to test homogeneity of variance

Answer 25

• correlation

- comparison of means

Question 26

homogeneity of variance: correlation

Answer 26

uses graphs

Question 27

What test is used to test for homogeneity of variance?

Answer 27

Levene’s test

Question 28

What does Levene’s test assess?

Answer 28

assesses the null hypothesis that variances in different groups are equal

Question 29

For Levene’s test, what does P less than 0.05 mean?

Answer 29

• variances are different among groups

- assumptions have been violated

Question 30

What are limitations to assessing for homogeneity of variance using Levene’s test?

Answer 30

• subject to bias with large sample sizes

- small deviations produce significant Levene’s test with large samples

Question 31

Ways to deal with outliers

Answer 31

• remove the case
• transform the data
• change the score

Question 32

dealing with outliers: removing the case

Answer 32

• delete the data

- only done if there’s a good reason to believe it’s not from the population you intended to sample

Question 33

dealing with outliers: transforming the data

Answer 33

reduces skewness

Question 34

dealing with outliers: change the score

Answer 34

can be used if transforming data fails to normalize the distribution