Lecture 3 Regression Flashcards
Errors are assumed to be…
Independent
Normally distributed with a mean of zero
Homoscedastic -
What is the variance explained?
How much can be predicted fro the straight line
R^2
Proportion of variance explained = 100%
R^2= 1.00
What does ANOVA tell us? F test
Tells us whether the variance explained is significantly different from zero
If it’s not significant, the regression is worthless. Predictor does not explain outcome variable at all.
What can b in regression tell us
It is the slope
It can tell us if there’s a relationship between X and y
-Ve if +Ve relationship
Estimate of selected change in y why X increases by 1
If standardised form (z scores) it will be the same as correlation coefficient
What is homoscedasticity and heteroscedasticity?
Homo - scatterplot varies equally around the line. Absense of any other pattern rand supports the line…
Hetero - scatterplot against predicted values.
How to write up regression results if
B =.67
Df =88
p
b=.67, t(88)=5.03, p
How to write up for ANOVA when
Df = 1 and 88
F = 25.3
p
R^2=0.223,F(1,88)=25.3, p
What is the regression model and what is it used for?
Y = a + bX + error
Outcome variable = intercept + slope X predictor variable
Want to find a and b which minimises error
Use to find line of best fit for a correlation scatterplot