Linear Regression

Question 1

Difference between Linear Regression & Correlation coefficient

Answer 1

LR has to

1. Specify DV & IV
2. Predict Y (DV) from X (IV)
3. Has additional parametric assumptions (residuals)

Question 2

Define Residuals

Answer 2

Difference between estimated and actual values of Y

Question 3

Example Null Hypothesis for Simple Linear Regression

Answer 3

The slope is zero, there is no linear relationship between running time and sex time

Question 4

Additional parametric assumptions for Simple Linear Regression? List former as well.

Answer 4

Check Scatterplot or Histograms of residuals:

1. No discernible pattern (appears scattered)
2. No outliers
3. Normally distributed

Question 5

Describe how Scatterplot for Residuals is plotted

Answer 5

1. Co-ordinates of each case is plotted
2. Regression line (line of best fit is added)
3. Line anchored at coordinate of two means (X & Y)
4. Angle of slope minimises sum of squared error

Question 6

What is the SLR equation? Describe each variable.

Answer 6

Y = a + bX
Y is DV (predicted value)
X is IV (specified value)
a is intercept/constant (value of Y when x = 0)
b is coefficient/slope of line associated with IV

Question 7

Caution when using Regression Equation

Answer 7

Dangerous to extrapolate outside range of data used to construct regression equation

Question 8

E2 means?

Answer 8

Move decimals two points to the right (1.3 to 130)

Question 9

Should we interpret R Square or Adjusted R Square?

Answer 9

ADJUST R SQUARE (more accurate)

Question 10

What is Standardized Coefficient (beta)?

Answer 10

Predicted effect on y if x increases by 1 SD