Definitions
adjusted R squared - a multiple correlation coefficient squared that has been adjusted for both the number of independent variables in the model and the number of observations.
Bayes theorem - a formula to combine conditional and prior probabilities to compute posterior probabilities.
cell mean - in a two factor design a cell mean is a mean of a dependent measure for a combination of levels of the two factors.
centroid - mean
coefficient of determination - the multiple correlation coefficient squared.
collinearity - the degree to which an independent variable can be predicted by other independent variables.
correlation matrix - a table of all possible correlation coefficients between a set of variables.
correlation matrix - a table of correlation coefficients.
cross-validation - a statistical procedure that checks the accuracy of a model by using the model created by a given sample to predict a different sample.
degrees of freedom - the number of values that are free to vary.
dependent variable - the variable to be predicted.
dependent variable - the variable that is predicted.
dichotomous group membership - the unit belongs to one of two groups.
dichotomous predictor variable - a categorical variable with two levels.
dichotomous variable - a variable with only two possible levels.
discriminability - ability to separate into groups.
discriminant function analysis - a procedure to predict group membership of preexisting groups based on linear combinations of interval variables.
dummy coding - recoding a categorical variable with more than two levels into a number of dichotomous variables so that they may be used in a linear model, such a linear regression
eigenvalues - combined variance.
full model - the regression model that contains all the independent variables.
general linear model - a general-purpose conceptual framework for many different statistical techniques utilizing a linear model, i.e. multiple regression, ANOVA, Discriminant function analysis, and canonical correlation
generating model - a theoretical probability model that is used to generate random data.
grand mean - in a two factor design the grand mean is the mean of all scores.
hierarchical regression - a sequential regression model procedure that enters the independent variables in a predetermined sequence.
hyperplane - a N-1 dimensional surface in an N dimensional space.
hyperspace - an N dimensional space.
independent variable - the variables used to predict the dependent variable.
independent variables - the set of variables used to predict the dependent variable.
interaction effect - a change in the simple main effect of one variable over levels of another variable or combination of variables.
interaction effect - is a change in the simple main effect of one variable over levels of the second.
main effect - changes in a dependent variable at different levels of another variable
main effects - differences in means over levels of one factor collapsed over levels of the other factor.
malingering - faking, usually pretending to be worse off than one really is.
marginal mean - in a two factor design a marginal mean is a mean of the dependent measure for a level of one of the two factors.
mean square - a measure of variability, calculated by dividing the sum of squares by the degrees of freedom.
mean squares within - the denominator of the F ratio. A measure of error within groups.
multicollinearity - the independent variables are highly correlated with one another, or a linear combination of a subset of independent variables correlates highly with another independent variable.
multiple correlation coefficient - the correlation coefficient between the observed and predicted dependent variables.
multiple R - the correlation coefficient between the observed and predicted Y values.
multiple regression - a statistical procedure used to predict a single dependent variable from one or more independent variables. The procedure uses a linear transformation of the independent variables to predict the dependent variable. The linear transformation is one that minimizes the sum of the squared differences between the observed and predicted values of the dependent variable.
multivariate outlier - a score that falls outside the standard range in a multivariate relationship.
orthogonal - independent, uncorrelated
outlier - a score that falls outside the range of the majority of scores.
outlier - a value that lies outside the range of the other values in the sample.
P(D/G) - the probability of the data given the group.
P(G/D) - the probability of the group given the data.
partial correlation coefficient - the correlation between that variable and the residual of the previous model.
plane - a two-dimension surface in a three-dimensional space.
posterior probabilities - the probability of belonging to a particular group given additional information about the unit.
predictor variable - see independent variable.
prior probabilities - the likelihood of belonging to a particular group given no other information is available. They are symbolized by P(G).
R square change - the difference in the unadjusted multiple correlation coefficient squared between a partial model and a full model.
R squared change - the difference in the multiple R squared between a full model and a partial model.
regression coefficients - the weights in a linear model that optimally predict a dependent variable.
residual - the difference between an observed and predicted value.
sequential regression models - a statistical procedure that uses multiple regression to examine how adding independent variables in stages affects the prediction equation.
shrinkage - the loss of predictive power in a model when a sample other than the sample used to create the model is used.
simple main effect - changes in a dependent variable at different levels of another variable, given the value of another variable remains constant
simple main effect - is a main effect of one factor at a given level of a second factor.
squared residual - the squared difference between an observed and predicted value.
standard error of estimate - a measure of error in prediction.
step-down regression - a multiple regression modeling procedure that sequentially subtracts variables from the regression equation based on how little additional predictive power the variable adds to the current prediction equation.
step-up regression - a multiple regression modeling procedure that sequentially adds variables to the regression equation based on how much additional predictive power the variable adds to the prediction equation.
suppressor variable - an independent variable that does not by itself correlate highly with the dependent variable, but when included in a set of independent variables causes the set as a whole to be more highly correlated with the dependent variable.
tolerance - the degree to which an independent variable cannot be predicted by other independent variables.
univariate analysis - analysis of each variable individually
unstandardized canonical discriminant function coefficients - weights in a linear model used to combine variables or scores to predict group membership. These weights are optimized to maximally discriminate among groups.
Wilks Lamba - a measure of relationship between group membership and interval variables.
