Lecture 6 - Linear Regression Flashcards
Regression is a measure of what?
Relationship
Regression prediction is not what?
Causal
A regression line is best described as..?
A line of best fit for the data points.
The better the regressional model/regression line, the better the what?
The better the prediction we can make.
What is the full regression equation and what does each component refer to?
Y = a + bX + ɛ
a - intercept
b - slope
e - error
What are the two primary terms in the full regression equation?
a and b, intercept and slope.
What is the secondary term in the full regression equation?
e - error
How do you find the intercept of a regression line when it cannot be read from the plots?
The intercept (value of a) when x = 0: y = a + bx a = y - bx The intercept (a) would be the mean of y minus the mean of x times by b.
What are the three assumptions of linear regressions?
- The relationship is linear
- Y is normally distributed at all values of x.
- Y’s spread is the same at all values of x.
What is leverage?
Distance of a data point from the mean
What is the error?
The distance of a data point from the regression line
What is homoscedasticity?
The spread of y being the same for all values of x
An outlier with a large leverage and a large error will do what to the regression line?
Pull it towards the outlier so that it is in a skewed position, compared to where it would be without the outlier.
What is R square representative of?
The amount of variance explained by the model in the sample used in the study.
What is R square adjusted a representation of?
Estimate of the amount of variance explained in the population, within the bounds of the study’s sample.
