Lecture 7- Contrast Coding

Question 1

The f-statistic tests

Answer 1

The overall fit of the model

Question 2

Model parameters tell us about

Answer 2

Specific differences between means

Question 3

Dummy coding compares

Answer 3

Each category to the baseline

Question 4

What are the options for breaking down categorical predictors

Answer 4

• Orthogonal contrasts
• Post hoc tests
• Trend analysis

Question 5

Features of orthogonal contrasts

Answer 5

• Hypothesis driven
• Planned a priori
• Control type 1 error rate

Question 6

Features of post hoc tests

Answer 6

• Not planned (not hypothesis driven)
• Compare all pairs of means
• Multiple t-tests adjusted for the number of tests

Question 7

When is a Trend analysis useful

Answer 7

For ordered means

Question 8

With planned contrasts, the variability explained by the model (SSM) is due to

Answer 8

Participants being assigned to different groups

Question 9

With planned contrasts, the variability of the model (SSM) sometimes represents

Answer 9

An experimental manipulation

Question 10

With planned contrasts, the variability of the model (SSM) can be broken down further to test

Answer 10

Specific hypotheses about which groups might differ

Question 11

To control type 1 error rates contrasts must be

Answer 11

Independent

Question 12

With independent contrasts, if a group is singled out in a contrast then that group

Answer 12

Should not be used in any subsequent contrasts

Question 13

With k-1 contrasts you should always end up with

Answer 13

One less contrast than the number of groups

Question 14

With only 2 chunks each contrast should compare

Answer 14

Only 2 chunks of variation

Question 15

The first contrast will usually compare

Answer 15

Any control conditions (chunk 1) with any experimental ones (chunk 2)

Question 16

Groups coded with positive weights compared to groups coded with

Answer 16

Negative weights

Question 17

The sum of the weights for a comparison should be

Answer 17

Zero

Question 18

If a group is not involved in a comparison, assign it a weight of

Answer 18

Zero

Question 19

For a given contrast, the initial weight assigned to the groups in one chunk of variation should be

Answer 19

Equal to the number of groups in the opposite chunk of variation

Question 20

How to get the final weight

Answer 20

Divide the initial weights by the number of groups with non-zero weights

Question 21

In the absence of a specific hypotheses compare

Answer 21

All pairs of means to see where the specific differences lie

Question 22

How to avoid inflation in the type 1 error rate in post hoc tests

Answer 22

Adjust the alpha (or test statistic) to be more conservative

Question 23

What does trend analysis (polynomial contrasts) do

Answer 23

• Test for trends in the means

- Only makes sense for ordered groups