304: Correlation and Regression

1. Floor Effects
1. This is where a test is so hard that everyone scores low and it does not differentiate between scores on the low end
1. Everyone would get an E despite some people putting in much less effort than others
2. Ceiling Effects
1. This is where a test is so easy that everyone scores high so it does not differentiate on the high end
1. So everyone would get an A even though some people may have worked much harder than others
2. Advil Experiment
1. Although the graph in this experiment goes down this is still a ceiling effect because the drug is measuring pain relief. 
3. Lets Get Luke Into Law School
1. What are the factors of getting Luke into Law School
1. Acceptance = (4.3)LSAT + (2.7)Letters of Recommendation + (2.2)GPA + (1.1)Experience + (3.1)Interview + (.7)Writing Sample - (5.6)Criminal Record - (.4)Demographic Information
1. This is a regression equation of all the factors for luke getting into law school.  The coefficients are to represent the different importance that the law school puts on different factors.
2. The question is...
1. When do you say I know enough and adding more items is just creating busy work?
4. Bivariate Regression
1. Regression refers specifically to prediction
2. Simple bivariate regression uses;
1. X - one predictor variable
2. Y - one dependent (criterion) variable
3. Equation: Y' = a + bX
5. Multiple Regression
1. Predicts one continuous dependent variable using a linear combination of two or more predictor variables
2. Equation: Y' = a +b1X1 + b2X2
6. Residuals in Prediction
1. A regression equation generates a predicted value of Y (Y') for each value of X
2. Residual
1. difference between the actual and the predicted value
7. Validity Shrinkage
1. When a regression equation is calculated using one group of subjects and used to predict performance in another group of subjects there is error in the second group