Lecture 11

Question 1

Lecture 11:

What is Multiple Regression?

Answer 1

Study with multiple predictors (more than 1 Independent Variable) & a single dependent variable

Question 2

Lecture 11:

What is an example of Multiple Regression?

Answer 2

Predicting muscle fibre type from;
1.) rate of fatigue during 30sec cycle test (“wind gate test”)
2.) rate of force development maximal isometric contraction

Question 3

Lecture 11:

What are 3 reasons to choose a Multiple Regression?

Answer 3

1.) more than 1 (>1) independent variables will better predict the dependent variable
2.) total variance explained by >1 independent variables is greater than any single variable
3.) leads to a greater r^2 (explained variance)

Question 4

Lecture 11:

What is an example sentence where you would use multiple regression?

Answer 4

Using pre injury isokinetic strength (X1) & post surgical range of motion (X2) to predict time until running after ACL repair (Y)

Question 5

Lecture 11:

What is the letter sign for Multiple Correlation Coefficient & its range?

Answer 5

R = multiple correlation coefficient (not r for bivariate correlation)
- ranges from 0 to 1.0 *cannot be -‘ve

*0 = no relationship
*1.0 = perfect relationship

Question 6

Lecture 11:

What is Multiple Correlation Coefficient Squared & what does it explain?

Answer 6

R^2 & explains variance in the model
- eg; if R^2 = .72 than 72% of the variance in the DV (Y) accounted for by the IV (X1, X2,…)

Question 7

Lecture 11:

What is the optimal multiple regression?

Answer 7

The best model has the fewest independent variables (X1, X2,…) to predict the dependent variable
eg; 2 IV would be better predictors than 4 IV

Question 8

Lecture 11:

What are the 3 steps in the Forward Selection Method of Multiple Regression?

Answer 8

1.) start with a table with r between all X & Y variables (called correlation matrix) *wont have to perform one
2.) first X variable added to the model is the one with the highest correlation with the Y variable
3.) Further additions of X variables are added to the model in order of how much each variable can increase the R^2 value (maximize the explained variance)

Question 9

Lectrue 11:

What is an important thing to consider when choosing your variables for correlation matrix (multiple regression)?

Answer 9

Important to choose variables that are not highly correlated as it will not add enough information to the model & increases error if too correlated