# Simple Regression

?
• Created by: Chloe
• Created on: 30-04-15 14:20
• Simple Regression
• Making predictions of a criterion variable (Y) from a predictor variable (X)
• The best prediction is given by the average of all the Y values for any given value of X - the higher the R value the better the prediction will be because the point cloud will be much narrower
• Using the linear regression equation to make predictions
• Two separate regression lines: (1) The regression of Y on X (predicting Y from X values) (2) The regression of X on Y (predicting X from Y)
• For any correlation other than r=1 or r=-1, low & high scores on X are associated with Y values that are closer to the mean
• Extreme scores tend to become more average on the second testing - a problem when subjects are selected on extreme scores
• High scorers more likely to include overestimated scores & low scorers underestimated
• If r=1, all points lie on the regression line (Y has no error)
• The linear regression equation: is used to predict the value of Y from the value of X
• Y = a + bx
• Y = the predicted value of Y, a = the intercept (the value of y when x =0), b = the shape of the line (the amount by which y increases for every one unit change in x)
• Prediction & Casuality
• Causality: viewing violent TV leads to increased aggression in play
• Reverse causality: children who get involved in aggressive play at school have a kind of residual excitement that leads to them watching violent TV
• Extraneous variables: effect the outcome but not the predictor
• Mediation or intervening variables: intervene between predictor & outcome variables in casual linkage