If you do calculations by hand, keep at least 4 significant figures. That way these applets will be able to tell whether you are right or not.

1. To see if you did your logarithms correctly, compare your spreadsheet
with the picture on page 3 of the Assignment 5 pdf.

2. You had to convert into logarithms the S.C. values of the independent variables.
To check this, what is the logarithm
of the anticipated S.C. UE_RATE?

If that's OK, probably all your logarithms are OK.

3. When you run the regression in LS you will get a table of estimated coefficients.
What is your estimated LNUE_RATE coefficient?

4. You are asked to test hypotheses on two
coefficients, to see if their true values might be 1.
Let's see how you did with the second of those, the hypothesis
that the true coefficient of LNUE_RATE is 1.
For this, you do not use the P-Value in the regression results. That
P-Value is for testing the hypothesis that the coefficient is 0.
Instead, you must calculate the t-statistic using this formula: [beta-hat - 1]/[standard error of beta-hat]. (The 1 is there because that's your hypothesized value for beta.)
What is your calculated t-statistic?

5. Let's move to predicting the AFDC-UP caseload.
The first step in this to get LS to calculate a
predicted value for LNAFDCUP%. What did you get?

6. If you got that right, let's see if you successfully translated that
prediction for LNAFDCUP% into a number of families. You must covert that
number into its antilog. This gives you an estimate for AFDCUP%.
You divide that by 100 (because it's a percentage) and multiply by the
size of the SC labor force, 1,700,000.
What is your prediction for the number of AFDC-UP families?

The views and opinions expressed in this page are strictly those of
the page author. The contents of this page have not been reviewed or approved
by the University of South Carolina.