Problem Set 5
Due Friday, March 10, 2000
The following data is obtained from Statistics Canada's 1989 Survey of Consumer Finances. This Survey is attached to the Labour Force Survey, and uses the same basic sampling procedures as the monthly Labour Force Survey but is conducted only once each year. The set of numbers here is a small subset of the data in the Survey of Consumer Finances. The data given here represents 15 males, all living in Saskatchewan, all having occupation teaching, all having paid employment, all household heads between ages 30 and 60, and all having positive earnings. The particular set of 15 males shown here is a subset of the whole set of Saskatchewan males with the above characteristics. This subset was selected using the SAMPLE command in MINITAB. Age in years
is given in the first column and the earnings of each male in 1988 dollars are given in the second column. The earnings refer to earnings in 1988, with the Survey being conducted in April, 1989.
In terms of what is being examined in this problem, one ordinarily finds a typical profile of earnings by years of work experience. Human capital models provide some theoretical justification for this. For working class males in manual type jobs, the earnings profile
by years of work experience may be fairly flat, so that beginning wages are substantial, but may increase by relatively small amounts as the worker obtains more experience.
For males in white collar professional types of occupations, the earnings profile may be expected to increase more dramatically with experience. Since years of experience
is not directly measured in the Survey of Consumer Finances, here age is being used as a proxy for years of experience.
1. Compute the regression line relating earnings to age. Also compute R-squared, the standard error of estimate, the standard deviation of b, and the t-test. (In order to make the numbers a little more manageable, you could convert earnings into thousands of dollars and round to one decimal).
2. Draw the scatter diagram and draw in the regression line you compute.
3. Write a short note explaining the above results and your conclusions, based on these statistics and tests. Comment on any shortcomings you see in the data and methods,
possible violations of the assumptions (page 26 of Lewis-Beck), and any suggestions concerning how the sample or equation might be improved.