Regression Match

Set 1: Mechanics of Regression & Correlation

Set 1 of 2
Data = Model + Error
Zero Correlation Consequence
Error (Residual)
Covariance
Correlation Significance
Intercept (a)
Pearson's r
Ordinary Least Squares (OLS)
Scatterplot Visualization
Slope (b)
The uncertainty remaining after fitting the model to the data
The value of the function when X is zero (crosses y-axis)
The fundamental form of most statistical models
Magnitude of change in f(x) resulting from a change in x
Should be judged by magnitude (effect size), not p-values
Optimization algorithm that minimizes squared errors to estimate parameters
Standardized measure of association bounded between -1 and 1
Necessary to detect non-linear relationships (e.g., U-shaped) missed by r
Unstandardized measure of shared variance limited by units
The regression model collapses to the Mean of Y (flat line)