(20) 5. A study was done to see how some of the characteristics of a worker's...
(20) 5. A study was done to see how some of the characteristics of a worker's employer (like the size of the employer, or how long it has been in business) affect the wages of that worker. The data for the study came from a survey of workers that asked about the worker's personal characteristics, but also got the name and address of the worker's employer so that employer characteristics could be measured as well. The sample size was 1067. Some of the employer characteristics included in the sample were SITESIZE=number of employees working at the location where the person worked (Mean=58.5, standard deviation=8.8) EMPSIZE=total employer size, meaning the number of people working at all locations for the company where the person worked (Mean=290.0, standard deviation=22.2) EMPAGE=number of years the person's employer had been in business (Mean=40.1, standard deviation=32.1) In order to look at the effect of these employer characteristics on the hourly wages of workers, the following regression was estimated (standard errors are in parentheses under the estimates): Ln (Hourly Wage) = constant + .049*Ln(SITESIZE) (.0010) 1.011) + .121*Ln (EMPSIZE) +.0022*EMPAGE 4 (.008) (.0010) a. Test, with a=.10, the hypothesis that other things equal, companies with more total employees tend to pay higher wages. (Specify the null and alternative hypothesis, and explain the basis upon which you made your decision.) b. Interpret the coefficient on Ln(SITESIZE). + c. Does EMPAGE have a statistically significant effect on wages? If so, is this effect large enough to be practically (or economically) important? Explain. - d. As mentioned above, the researchers also had information on employee characteristics, including measures of employee quality like education, tenure, and experience. What they really wanted to know was whether two workers of identical quality would receive different wages if they worked for employers with different characteristics. When they added measures of employee quality to the regression, the coefficient on EMPAGE became -.0001, with a standard error of .0002. True, False, Explain (an answer with no explanation will receive no points): This change in the coefficient of EMPAGE is what you would expect if older firms tended to hire better quality workers. e (5) 6. True, False, Explain. Adding a variables to a regression that are highly correlated with the independent variables already included but not with the dependent variable will increase your chance of committing type II errors when conducting tests of statistical significance on the estimated coefficients.