Quick Review Of Some Inference For Linear Regression Concepts

1. How do you know that you should attempt to use a linear regression method for predicting y values from x values?

When the moon is full.

When Mrs. Linner gives you the evil eye.

When the data look linear when graphed in a scatterplot.

When the relationship should (theoretically) be linear.

Either C or D will suffice.

2. When do you know that the model is a close fit to the data?

When I flip a coin and it lands heads up.

When I have run out of other models to try.

When the r^2 value is close to 1.

When the graph of the residuals(x, y - y-hat)forms a straight line.

When the residuals(x, y - y-hat)are small and scattered(no pattern).

3. When is the relationship between x and y too strong to assume that the slope of the best fit line (Beta) is possibly zero?

When the confidence interval for the slope does not contain the value 0.

When the p value of the linear regression t-test is very small.

When the residuals are small and scattered.

When the r^2 value is close to 1

Answers A or B will suffice.

4. Given b and SE of b you still need to know _______________ to calculate the t-statistic and the p-value.

R

N

Df

A

B or C

5. Given s, sx, and n you can calculate

Beta

Alpha

B

R

SE of b

6. How do you know that your linear model was a wise choice (the linear model is the correct model)?