STAT 440 - Forecasting
Lecture Review #4 - Regression Error

Review questions:

1) How are the slope and intercept of a regression model computed? What do these numbers mean?

2) How do we compute the error variance (s_e²) for a regression model? What does the number mean?

3) What is meant by partitioning the sum of squares in regression?

4) What is the coefficient of determination? What does the number mean?

Computational exercises:

Consider the following data set:

1) Find the correlation for these data.

2) Find the slope and intercept for these data.

3) Find the error variance for these data.

4) Find the total sum of squares (SST), the sum of squares for regression (SSR) and the error sum of squares (SSE) for these data.

5) What is the coefficient of determination? Verify that it can be obtained from the sums of squares.

FIRST, answer the above questions by working through the problem "by hand." THEN use Microsoft Excel to obtain the same answers. Submit both your written work and the Excel printout.

SOLUTIONS:
1) r = .9759
2) slope = 10, intercept = 20
3) s_e2 = 200
4) SST = 8400. SSR = 8000. SSE = 400.
5) r²=(.9759)²=.9524, which is also 8000/8400