Multivariate + Regression

what is the meaning for each varible in Y=bX+a

Y= Predicted; b= slope; X= Predictor Variable; a=Y-intercept

What are the 3 Types of Regression?

Simple Regression, Multiple Regression, and Logistic Regression

What is the distance between the actual data and the regresiion line(predicted values) called?

residuals

If the correlation is stronger

than the better we can predict an accurate measure

If I am using class size to predict the end of term grades what type of regression am I using?

Simple Regression

If I am testing whether class size AND teacher’s year of experience predict end-of-term grades than what type of regression am I using?

Multiple Regression

what measures how far a set of data points spreads out from the sets mean

variance

If I’m measuring how much of the data sets results(variance) is caused by the predictor variable, then what would I use?

Coefficient of Determination(R²)

if I say that a model explains 25% of the variance than what is the value of R²?

R²= 0.25

If the data set creates a straight line pattern, what type of assumption is it?

Linearity

what is the difference what the model predicts and what actually happened

error

Which regression has one predictor variable and the outcome is measured by a continuous scale

Simple Regression

What Regression type has more than one predictor variable and the outcome is measured on a continous scale

Multiple Regression

What regression type has any number of predictor variables, and outcomes are categories rather than continuous numbers

Logistic Regreession

If the model is multiple regression, then what happens to the regression equation?

Y= (xb)^1 + (xb)^1 + a

How do you test Normality on SPSS

Analyze → Regression → Linear

Move variables to DV and IV Box

Click “Plots” Check: Normal Probability plot

Use histogram to see data set shape

How do you test Independence of errors on SPSS?

Analyze → Regression → Linear

Move variables to DV and IV Box

Click “Statistics"; Check the Durbin-Watson box

Is my Durbin-Watson value acceptable or not acceptable if it’s between 1.5 and 2.5?

it’s acceptable

If my Durbin-Watson value is close to 2 then what does that mean?

it means there is no serial correlation, or, the predictor variable and actual data set are indpedent of each other

How do I test No Multicollinearity on SPSS?

Analyze → Regression → Linear

Move variables to DV and IV

Click “Statistics”; Check: Collinearity Diagnostics

If the VIF value is greater than 5, what are my results for the No Multicollinearity test?

If the value is greater than 5 than it’s possible that on of your predictor variables( IV) is too highly correlated with another.

How do I test Linearity and Homoscedasticity on SPSS

Analyze→ Regression → Linear

Move variables to DV and IV

Click “Plots”; put standardized residuals in Y Box and standardized predicted. values in X box

What is the primary goal of multivariate statistics?

The goal is to understand the relationships and structres among multiple variables