This document contains an assignment with multiple questions related to econometrics and statistical analysis. It includes questions about interpreting regression results, testing hypotheses, and assessing significance of variables. The assignment involves analyzing data from different datasets and estimating several regression models.
This document contains an assignment with multiple questions related to econometrics and statistical analysis. It includes questions about interpreting regression results, testing hypotheses, and assessing significance of variables. The assignment involves analyzing data from different datasets and estimating several regression models.
Assignment #3: Due March 7, 2014 1. (4.1) Which of the following can cause the usual OLS t statistics to be invalid (that is, not to have t distributions under H 0 )? (a) Heteroskedasticity. (b) A sample correlation coecient of 0.95 between two independent variables that are in the model. (c) Omitting an important explanatory variable. 2. (4.3) The variable rdi:tc:: is expenditures on research and development (R&D) as a per- centage of sales. Sales are measured in millions of dollars. The variable jro):arq is prots as a percentage of sales. Using the data in rdchem.dta for 32 rms in the chemical industry, the following equation is estimated: \ rdi:c:: = 0.472 + 0.321 ln(:a|c:) + 0.050 jro):arq (1.369) (0.216) (0.046) : = 32, 1 2 = 0.099 (a) Interpret the coecient on ln(:a|c:). In particular, if :a|c: increases by 10%, what is the estimated percentage point change in rdi:tc::? Is this an economically large eect? (b) Test the hypothesis that R&D intensity does not change with sales against the alternative that it does change with sales. Do the test at the 5% and 10% levels. (c) Does jro):arq have a statistically signicant eect on rdi:tc::? 3. (4.5) Consider the following estimated regression, which can be used to study the eects of skipping class on college GPA. The variable co|G1 is GPA in college; /:G1 is high school GPA, CT is the score on the ACT test and :/ijjcd is the number of lectures missed per week. \ co|G1 = 1.39 + 0.412 /:G1 + 0.015 CT 0.083 :/ijjcd + n (0.33) (0.094) (0.011) (0.026) : = 141, 1 2 = 0.234 (a) Using the standard normal approximation, nd the 95% condence interval for ,
. (b) Can you reject the hypothesis H 0 : ,
= 0.4 against the two-sided alternative at
the 5% level? (c) Can you reject the hypothesis H 0 : ,
= 1 against the two-sided alternative at the
5% level? 1 4. (4.9) Consider the regression model :|ccj = , 0 + , 1 totnr/ + , 2 cdnc + , 3 aqc where :|ccj is a persons total amount of sleep (in minutes per week), totnr/ is his or her total amount of work (also in minutes per week), cdnc is years of education, and aqc is years of age. The estimated equation is [ :|ccj = 3638.25 0.148 totnr/ 11.13 cdnc + 2.20 aqc (112.28) (0.017) (5.88) (1.45) : = 706, 1 2 = 0.113 where the gures in parentheses are standard errors. (a) Is either cdnc or aqc individually signicant at the 5% level? Show your work. (b) Dropping cdnc and aqc from the equation gives [ :|ccj = 3586.98 0.151 totnr/ (38.91) (0.017) : = 706, 1 2 = 0.103 Are cdnc and aqc jointly signicant in the original equation at the 5% level? Justify your answer. (c) Does including cdnc and aqc in the model greatly aect the estimated tradeo between sleeping and working? (d) Suppose that the sleep equation contains heteroskedasticity. What does this mean about the tests computed in parts (a) and (b)? 5. (4.16; computer exercise) Use the data in mlb1.dta for this exercise. (a) Estimate the model ln(:a|arj) = , 0 +, 1 jcar: +, 2 qa:c:jr +, 3 /aq +, 4 /rn::jr +, 5 r/i:jr +n and then estimate it again, dropping the variable r/i:jr. What happens to the statis- tical signicance of /rn::jr? What about the size of the coecient on /rn::jr? (b) Add the variables rn::jr, )|djcrc, and :/a:c:jr to the model from part (a). Which of these factors are individually signicant? (c) In the model from part (b), test the joint signicance of /aq, )|djcrc, and :/a:c:jr. 2