week 4 - one-way between subject ANOVA

Question 1

What is the purpose of correlation?

Answer 1

to find out if there is an association between two continuous variables.

Question 2

What is the purpose of simple regression?

Answer 2

to find out the extent of the linear relationship between the outcome variable and the predictor variable.

Question 3

Which test should you run for the following research question; are Gryffindor players better than Slytherin players?

Answer 3

between-subject t-test

Question 4

Which test should you run for the following research question; Is exam scores on Defence Against the Dark Arts associated with performance in Quidditch (game score)?

Answer 4

within-subject t-test

Question 5

Which test should you run for the following research question; Ron Weasley wanted to estimate and predict his Defence Against the Dark Arts exam score based on his Quidditch performance. What test should he run?

Answer 5

one way between-subject ANOVA

Question 6

Which test should you run for the following research question; Are Gryffindor players the best compared to the other 3 houses?

Answer 6

one way within-subject ANOVA

Question 7

What is ANOVA?

Answer 7

ANalysis Of VAriance. A statistical test for three or more groups or conditions (or a combination of groups and conditions).

Question 8

What are the two types of ANOVA?

Answer 8

Between-subject/Independent samples ANOVA and within-subject/related-samples ANOVA?

Question 9

When is between-subject/independent sample ANOVA used?

Answer 9

When there are three or more groups.

Question 10

When is within-subject/related-sampled ANOVA used?

Answer 10

When there are three or more conditions.

Question 11

ANOVA with both between-subject and within-subject factors is called…

Answer 11

A multi-factorial or mixed ANOVA.

Question 12

Explain groups and levels in; Are Gryffindor players the best compared to the other 3 houses?

Answer 12

groups: 4 levels: Gryffindor, Slytherin, Ravenclaw, Hufflepuff

Question 13

Explain conditions and levels in; Harry wanted to know if whoever uses Thunderbolt are better than when they use Firebolt or Nimbus 2000.

Answer 13

conditions: 3 levels: thunderbolt, firebolt, nimbus 2000

Question 14

What are F-tests?

Answer 14

Used by ANOVA to statistically test if the means are different across groups or conditions.

Question 15

What does F-test produce?

Answer 15

An F-ratio to inform us if the variability between group means is larger than the variability of the observations within the groups (i.e., error).

Question 16

ANOVA is an omnibus test, what does that mean?

Answer 16

it tests for an overall experimental effect. (i.e. we don’t know which groups or conditions differed.)

Question 17

Why not do multiple t-tests?

Answer 17

Increases chance of type 1 error.

Question 18

what is the family-wise error rate?

Answer 18

the greater the number of tests, the greater the chance of type 1 error.

Question 19

What does the F-ratio compare?

Answer 19

amount of systemic variance(variance due to what we know) to unsystematic variance(variance due to what we don’t know).

Question 20

What are the three types of variation (sum of squares)?

Answer 20

total sum of squares, model sum of squares, and residual sum of squares.

Question 21

What are post-hoc tests needed for?

Answer 21

with omnibus test we don’t know which groups are different specifically.

Question 22

What is the Tukey HSD test designed for?

Answer 22

So you can compare as many things as you want without inflating the family-wise error rate.

Question 23

How do you report Tukey HSD test results?

Answer 23

It produces no test statistic, only a p-value so we describe the finding in the text.

Question 24

What is cohen’s D for post-hoc testing?

Answer 24

difference between means divided by standard deviation.

Question 25

To calculate F in a one-way independent ANOVA, the variance between groups is divided by the variance within groups. To calculate the variance between groups you need to work out the residual sum of squares. To do this:

Answer 25

You take each individual score, subtract the mean for the group it belongs to, square the answer, and add all of these values together.