Yessenia Case

My friend Yessenia has been considering taking a year off her job to further her education and then returning to her current employer. Yessenia is aware of my proficiency is econometrics and asked me to consult her on her decision. In order to consult her decision numerically, I decided to run a regression where I considered the variables I considered most relevant to her situation. In order to do this I used data gathered by National Longitudinal Survey of Youth in 2000 and a statistical package named EViews to run the regression and interpret the data.
Before running a regression I had to restrict the data to only female and west coast samples to make the regression analysis most relatable to Yessenia. Then I had to consider the variables most relevant to Yessenia, I decided that years of schooling, years of experience, years working for current employer (tenure), and age were the
independent variables that most effected and related most to the regression ran for Yessenia’s situation. The goal of the regression would be to see the effect of these independent variables on earnings, our dependent variable and the central idea behind Yessenia’s decision to choose between school and her job. In equation form the regression I ran looks like:
Earnings=βschooling*(years of Schooling) + βexperience*(years of Experience) + βtenure*(years of Tenure) + βage *(age)
*where β represent the numerical effect of that variable on earnings
After running the regression I concluded that an additional year of schooling marginally increases earnings by $1.13/hour while an additional year at her job increases her earnings by $0.55/hour (individually: experience ($0.39/hour) and tenure ($0.16/hour)); creating a difference in earnings of $0.58/hour favoring a year of schooling. Before testing the hypothesis that βschooling >βexperience + βtenure I first tested the null hypothesizes βschooling =0, βexperience =0, βtenure =0 and after calculating T-Statistic values of 17.99, 7.03, 3.25 respectively I was able, at a 95% significance level (1.96 critical value), reject the null hypothesis. It is important to note that I left out βage from my hypothesis test because its p-value of .1328 led me to conclude it was statistically insignificant when considering its effect on earnings. Then I tested the hypothesis that βschooling >βexperience + βtenure (or βschooling – (βexperience + βtenure) =0 in EViews) to conclude whether or not the year of school adds more value to earnings than experience and tenure does. My F-tested concluded that you would reject the hypothesis that βschooling – (βexperience + βtenure) =0 favoring a year of schooling over experience and tenure. Given these results I recommend to Yessenia that she take a year off work to gain the additional year of schooling. Now Yessenia asked me to consider to the same scenario; however, now she will not be returning to the same company, therefore losing her tenure.
She asked me to consider if a year of school still outweighs experience and tenure considering the presence of ongoing tenure. In order to execute this I considered a similar equation as before but adding in existing tenure (creating a multiplier effect with βtenure) βschooling >βexperience + tenure*βtenure. After evaluating the EViews of hypothesis test of βschooling =βexperience + 3.87*βtenure (using 3.87 because the mean tenure from our sample is 2.87 and you must consider the additional year being forgone), the results conclude that you would fail to reject the null hypothesis, indicating that a year of schooling does not outweigh a year of education and tenure, while considering existing tenure as a multiplier to the coefficient βtenure. Considering my regression and test results I would recommend Yessenia not leave her existing job for a year of schooling, and instead stay at her job where she can accumulate experience and
tenure.