Please enable JavaScript.
Coggle requires JavaScript to display documents.
simple linear regression - Coggle Diagram
simple linear regression
individual differences
height
weight
personality
intelligence
autistic spectrum
sometimes related
taller people tend to be heavier
helps makes predictions
visualising relationships
best visualised on a scatter plot
allows us to predict a score on variable Y (DV) from a known score on variable X (predictor)
method of least squares
can't just draw any line through the data
need line of best fit
least squares method aims to minimise difference between the line and observed data points
The maths behind regression
we can predict Y from X using the equation for a straight line
Y' = a + bX
Y' = predicted height
a = value of the intercept
b = the slope/gradient of the line
x = the value we know
how good are our predictions?
just because we have a line of best fit does not mean its a good fit
the better our line fits the data the more accurate the predictions
we can measure degree of fit using Rsquared
Rsquared of 0 = terrible fit
Rsquared of 1 = perfect fit
limitation of simple regression
can only use 1 variable to predict outcome
think of most useful predictor