12. Analysis of associations Part II: Linear Regression

Question 1

When do we use regression?

Answer 1

to predict values of one variable from another variable

we manipulate the explanatory variable and see how it effects the response

Question 2

what questions do we ask about regression models?

Answer 2

intercept significantly different to 0

slope of lien is Signiant different to 0
(expect a flat line)

Question 3

what is the approach of a linear regression?

Answer 3

estimates m and c
estimates SE

uses SE to determine if all terms are necessary

determines variation explained by model

Question 4

What are residuals?

Answer 4

the difference between an observation and its predicted variables

y - (y-arrowhat)

Question 5

What does the variance of residuals describe?

Answer 5

describes the spread or scatter of points
around the regression line

we use it to estimate the standard error of the slope

Question 6

What d.f. do we use when we do a t-test when testing the hypothesis of slopes and intercepts.

Answer 6

2

Question 7

how do we work out the variation explained by a linear regression model?

Answer 7

ANOVA