Question 1 Which coding scheme will typically account for the largest percentage of variance in a multiple regression analysis?

a.Dummy coding

b.Effect coding

c.They will all account

for the same percentage of variance

d.It depends on the

specific data in the analysis

Question

2 Your colleague is performing a multiple regression and has coded

ethnicity into a series of variables. However, she has not told you the coding

method she used. You look at the variables and all of the have values are

either 0 or 1. Which coding scheme did she use?

a.Dummy coding

b.Effect coding

c.It could be either

approach

d.If you are performing

a multiple regression, you shouldn't code ethnicity using any of those

approaches -- instead, it should be treated as a continuous variable.

Question

3 You are going to code information regarding the state in which a

student lives as part of a multiple regression analysis. You wish to compare

students in New Hampshire, Vermont, and Massachusetts to students in Maine.

Which of the following coding schemes would be the most logical choice?

a.Dummy coding

b.Effect coding

Question

4 You wish to examine the effect of different textbooks on student

learning of mathematics. You are going to code information regarding the book a

student uses, but don't have a referent group or any specific comparisons you

wish to make. Instead, you want to examine how a results for a book compare to

the overall or mean-of-the-mean scores. Which of the following coding schemes

would be the most logical choice?

a.Dummy coding

b.Effect coding

Question

5 Participants in your study belong to one of four ethnic groups.

If you wished to dummy code this information, how many dummy variables would

you need?

a.1

b.2

c.3

d.4

Question

6a This problem includes four sub-questions based on the following

scenario: You conducted an analysis predicting mathematics scores based on

student's state of residence (Maine, New Hampshire, Vermont, and

Massachusetts). State of residence was effect coded with

the variable names based on the corresponding state's 2-character postal code

abbreviation (e.g., Maine = ME). The resulting multiple regression equation was

(note that " * " is used for multiplication instead of

"x"):

Math = 34.2 + 2.2*ME - 1.1*NH

+ 1.8*VT

Based on this equation,

please calculate the predicted score students in each of the four states.

The predicted score for Maine

is:

Question

6b Continuing with the previous question and the formula...

Math = 34.2 + 2.2*ME - 1.1*NH

+ 1.8*VT

The predicted score for New

Hampshire is:

Question

6c Continuing with the previous question and the formula...

Math = 34.2 + 2.2*ME - 1.1*NH

+ 1.8*VT

The predicted score for

Vermont is:

Question

6d. Continuing with the previous question and the formula...

Math = 34.2 + 2.2*ME - 1.1*NH

+ 1.8*VT

The predicted score for

Massachusetts is: