Week 5 (Hierarchical Multiple Regression)

Question 1

Why use hierarchical multiple regression

Answer 1

-Enables us to examine the contribution of additional values to the model when added in a separate step

Question 2

When to dummy code

Answer 2

Categorical or Nominal variable has 3+ categories

Question 3

Dummy code process

Answer 3

-Count number of groups you want and to recode and subtract 1
-Create as many new variables as that value
*Dummy variables
-Choose one group as baseline
*Usually control group
*All other compared against it
-Assign that group values of 0 for all dummy variables
-For dummy variable 1, assign the value 1 to first group, and all others 0
-For dummy variable 2, assign the value 1 to the second group, and all others 0
-Do this till you run out of dummy variables
-Include all dummy variables as predictors in regression analysis

Question 4

Difference in variables entered/ removed graph in hierarchical regression

Answer 4

You’ll see multiple rows, one for each step of the model.

Question 5

R squared change (Hierarchical)

Answer 5

-Section in the model summary SPSS output called “Change statistics”
-Tells us whether adding variable at step 2 improves the model, by increasing the proportion of variables accounted for.
-Change in r squared between 1 and 2, and whether it’s significant using F change + significance

Question 6

What would unstandardised beta of b = 3.93 mean

Answer 6

As hours in workshops increases by one unit (hour), stats exam mark increases 3.93 marks

Question 7

SPSS Excluded variables

Answer 7

Tells us what would have happened if we’d entered the variables at step 1 rather than step 2

Question 8

Casewise Diagnostic

Answer 8

-Tells us whether we have any extreme scores
-Looking for values greater than 2
-(Any residuals value greater than 2SD)