Chapter 7. Regression.
Chapter 7: Regression
Section 1: What Are Predictions?
Regression is a statistical technique used for predictions in various fields.
Common understanding of predictions refers to guesses about future events.
In statistics, predictions often refer to estimates about numbers, such as height, intelligence, or income.
An educated guess relates to uncertainties about future or existing situations.
Section 2: Predicting Individual Values Using Group Mean
Predictions involve estimating an individual's score using group averages (mean).
The variable being predicted is usually denoted as "Y" (the Y-variable or criterion).
The principle "Playing the Averages" suggests using the mean (πΜ ) of a group for the best estimate if no additional information is available.
Examples demonstrate calculations of predicted heights with mean based examples:
Example: Ben's height estimated at 62 inches (Y) if the average height is this value.
Section 3: Using Predictor Variables for Better Estimates
When another correlated variable (X) is known, it allows for improved estimates of Y.
Conditional means can provide more accurate predictions compared to group means.
Terminology: Y is the criterion, X is the predictor.
Example: Predicting weights of girls based on their grade using scatterplot data.
Section 4: The Regression Line
The regression line provides an efficient method to determine estimates and best fits conditional means.
Example calculations show using regression equations to predict percentages of overweight men from calorie intake with given data.
Section 5: Standardized Regression Equation
Standardized regression equations use z-scores for predictions and have properties, such as always having a slope equal to r (correlation).
Unstandardized regression equation includes slope (b) and intercept (a).
Section 6: Two Essential Facts About Standardized Regression Equation
Regression Fact 1: Standardized equation is always of the form π§Μπ = r * π§π.
Regression Fact 2: The line passes through the Point of Averages.
Section 7: Estimating Raw Scores and Percentiles
The standardized regression equation can also be used to estimate raw scores or percentiles from predictor variables.
Section 8: Finding the Unstandardized Regression Equation
The unstandardized regression equation consists of three steps: finding unstandardized slope, intercept, and plugging them into the general equation.
Section 9: R.M.S. Error of the Prediction
R.M.S. Error indicates how typically off predictions are from actual values, calculated through the sum of squared errors (SSE).
Section 10: Quick Formula for R.M.S Error
Quick Formula: R.M.S. Error = β(1 β rΒ²) * SDY.
Section 11: Sketching Approximate Regression Line
Steps for estimating regression line include dividing scatterplot into strips, sketching approximate conditional means, and finding coordinates for unstandardized slope and intercept.
Section 12: Data Set and Finding the Regression Equation
Exercises require practical application of calculating means, standard deviations, correlation, standardized and unstandardized regression equations, R.M.S. Errors, and plotting on scatterplots.