Regression model, Model Coefficients, and Statistical Significance
complete this exam with 4 questions
Time allowed: 3 hours Answer all questions. Marks for each indicated in square brackets. Question 1: [20 marks] Consider the following estimated regression results on a sample of 753 women: where the subscript i denotes the individual; hours is the annual hours of employed work; kidslt6 is the number of kids less than 6 years old; kidsge6 is the number of kids 6-18 years old; age is the woman’s age (in years); educ is the woman’s education (in years); and othfaminc is other family income (measured in thousands of dollars). The standard error for each estimated coefficient is listed in parentheses. (a) Interpret each of the estimated model’s coefficients and determine its statistical significance. (b) Are the signs (positive or negative) on each coefficient what you would expect to see? Explain. (c) Is the model statistically significant overall? Question 2: [5 marks] Consider the regression model: Explain in detail how you would test the following hypotheses: (a) (b) Question 3: [5 marks] Suppose that an individual’s wages depend positively on both their years of education and their ability. Years of education and ability are also positively correlated with each other. Ability, hoursi = 1346.75 (287.12) − 504.53 (63.96) ⋅ kidslt 6i − 82.49 (24.97) ⋅ kidsge6i −20.18 (4.54) ⋅ agei + 56.55 (13.94) ⋅ educi − 10.49 (2.71) ⋅ oth faminci R2 = 0.1180 F = 19.99 Prob > F = 0.0000 Yi = β0 + β1xi1 + β2xi2 + β3xi3 + ui H0 : β1 = β2 = 0 H0 : β1 = β2 however, is difficult to measure. The survey data you are using thus contain measures of only wages and years of education. Describe the consequences of running a regression model of wages (Y) on years of eduction (X) that omits ability from the model. Question 4: [10 marks] Suppose you have real estate sales data on price (in $), north (=1 if north location and =0 if south), and older (=1 if the house is 10 years of age or older; =0 if the house is less than 10 years old). (a) For the model , explain how you would interpret each coefficient. (b) Now modify the model in (a) to allow for the possibility that the effects of age differ by location. Explain how you would interpret each of this new model’s coefficients.
DECEMBER 20, 2020
Time allowed: 3 hours
Answer all questions. Marks for each indicated in square brackets.
Question 1: [20 marks]
Consider the following estimated regression results on a sample of 753 women:
hours i = 1346.75 − 504.53 ⋅ kidslt 6i − 82.49 ⋅ kidsge6i
(287.12)
(63.96)
(24.97)
−20.18 ⋅ agei + 56.55 ⋅ educ − 10.49 ⋅ othfaminc
(4.54)
(13.94)
(2.71)
R2 = 0.1180
F = 19.99 Prob> F =0.0000
where the subscript i denotes the individual; hours is the annual hours of employed work; kidslt6 is the number of kids less than 6 years old; kidsge6 is the number of kids 6-18 years old; age is the woman’s age (in years); educ is the woman’s education (in years); and othfaminc is other family income (measured in thousands of dollars). The standard error for each estimated coefficient is listed in parentheses.
* Interpret each of the estimated model’s coefficients and determine its statistical significance.
13465.75 is the y-intercept
The standard error (S.E.) measures average deviation of the prediction errors about the regression line
The t statistic, t= coefficient / standard error = 1346.75 / 287.12= 4.6905
The p-value is < 0.00001 and the result is significant at p < 0.05.
* The annual hours of employed work decreases by 504.53 when the number of kids less than 6 years old increases by 1
The t statistic, t= coefficient / standard error 504.53/ 63.96
t=7.89
Degrees of freedom (DF is (number of observations minus the estimated parameters) =753-6= 747
The p-value is < .00001 (One-tailed) and the result is significant at p < 0.05.
* The annual hours of employed work decreases by 82.49 when the number of kids 6-18 years old increases by 1
t=82.49/ 24.97= 3.3
The p-value is 0.000501 (One-tailed) The result is significant at p < 0.05.
* The annual hours of employed work decreases by 20.18 when woman’s age (in years) increases by 1
t=20.18/ 4.54 = 4.44
The p-value is < 0.00001(One-tailed)The result is significant at p < 0.05.
* The annual hours of employed work increases by 56.55 when the woman’s education (in years) increases by 1
t=56.55 /13.94 =4.06
The p-value is 0.000028 (One-tailed) The result is significant at p < 0.05.
* The annual hours of employed work decreases by 10.49 when the family income (measured in thousands of dollars increases by 1
t=10.49/ 2.71 =3.8708
The p-value is 0.000059(One-tailed)The result is significant at p < 0.05
* Are the signs (positive or negative) on each coefficient what you would expect to see? Explain.
The number of kids less than 6 years old is a negative coefficient as expected as mothers with young children are more likely to spend time caring for them, and this implies less time working
The number of kids 6-18 years old is also a negative coefficient as mothers take more time cari...
👀 Other Visitors are Viewing These APA Math Problem Samples:
- Economic Math Problem: Graded Homework Assignment4 pages/≈1100 words | No Sources | APA | Mathematics & Economics | Math Problem |
- Misleading Graph-Unemployment Rate under President Obama. Economics3 pages/≈825 words | APA | Mathematics & Economics | Math Problem |
- Problem set 10 Mathematics & Economics Math Problem3 pages/≈825 words | APA | Mathematics & Economics | Math Problem |
- Hypothesis Testing and Type Errors Module 5 Case10 pages/≈2750 words | No Sources | APA | Mathematics & Economics | Math Problem |
- Multiple Regressions Week 4 Assignment Chapter 12 Problem 641 page/≈275 words | 2 Sources | APA | Mathematics & Economics | Math Problem |
- Discussion: Break-Even Pricing Problem1 page/≈275 words | No Sources | APA | Mathematics & Economics | Math Problem |
- Individual assignment Mathematics & Economics Math Problem3 pages/≈825 words | 10 Sources | APA | Mathematics & Economics | Math Problem |