As a retired educator and?having?spent the great majority of my career in higher education I am always?curious?about how post-secondary education effects the lives of those who have?attained?a degree in higher education. A few days ago I learned about ?NSF SESTAT?2008 National Survey of Recent?College Graduates Public Use File?, a comprehensive and integrated?system of?information about the employment, educational and demographic?characteristics?of individuals with post-secondary degrees with a focus on scientists?and?engineers in the United States.
One?of the major?outcomes of attaining a postsecondary degree is its impact on
socio-economic status,?namely salary. With the SESTAT database in hand I examined salary?outcomes,?controlling for a number of covariates.
I. Variables
It?should be?noted that I restricted the analysis to exclude individuals who?responded that?they were not in the labor force.
The purpose of this study is to assess each of?the above covariate?s influence on the dependent variable, salary.
II. Model Specification and Findings
The following multiple regression equation was specified, producing the results described in this section:
Please?see?Appendix A for the multiple regression output.
The model explains 46?percent of?the variance in salary for degree holders (Prob > F =?0.0000) in the labor force. All of the covariates are statistically significant.
In the following section I provide graphs for each covariate effect on predicted salary,?controlling?for all the other covariates. In each case I plot the predictive margins?against age distribution on the horizontal axis.
1. Highest Attained Degree ? ?DGRDG?
Individuals who attain a doctorate or professional degree can expect a higher predicted salary, holding all other variables constant. ?Master degree, doctorate and professional degree holders are predicted to make $8,428, $16,144 and $19,328, respectively, more?than individuals with the baccalaureate degree as their highest degree.
The 95% confidence intervals are provided in the above table.
2. Job Satisfaction
Individuals who exhibit higher job satisfaction are more likely to enjoy a higher predicted salary. Individuals who are very satisfied with their job are predicted to make more than $6,700 per year than ?very dissatisfied? employees.?
?3. Attendance at a Community College
Individuals who at some point in their educational career attended a community college enjoy a $1,346 yearly advantage in salary. ?It?s been suggested that this outcome may be due ?to ?health field graduates? which includes many nursing students with bachelor?s degrees who subsequently get good paying positions in the health care industry.? This would be a good hypothesis to test but it appears the public release file that I am using doesn?t provide the level of detail to test the hypothesis.
4. Sex
Males are predicted, all other variables held constant, to make more than $4,500 per year throughout the age distribution.
5. Job Related to Degree
Individuals who are working in jobs closely related to their degree make $4,321?more than individuals who respond that their job is not related to their degree. Individuals with a job ?Somewhat related to their degree? experience only a $347 difference in yearly salary compared to individuals who are employed in jobs ?Closely related to their degree?.
6. Primary Work Activity
Individuals whose primary work activity is in computer applications enjoy higher predicted earnings than the other identified work domains, controllng for all other covariates. The teaching domain is associated with the lowest predicted salaries over the age distribution.
The regression analysis set the ?Research & Development? as the reference. Let?s see how the coefficients change when ?Teaching? is set as the reference group.
Holding all other covariates constant, on a yearly basis teachers make $5,371 less than ?Research & Development? personnel; $10,339 less than individuals working in the ?Management and Administration? domain; $13,454 less than individuals whose primary work activity is in computer applications; and $7,687 less than people working in domains classified as ?Other?.
7. Highest Degree Major
How does ?Highest Degree Major? influence predicted salaries? The multiple regression set ?Computer/Math? as the reference group. As the adjusted means suggest the ?Engineering? major enjoys the highest salaries, let?s set ?Engineering? as the reference group for ease of comparison. (In Stata version 12 this is done with the contrast command: contrast?rb5.MAJOR_HI_DEG, nowald effects.)
The closest yearly salaries to individuals with their highest degree in engineering are individuals with ? you guessed it ? ?S&E Related? (?science and engineering related?) degrees. ?Biology/Ag/Life Sciences? earn $17,134 less than individuals with their highest degree in engineering.
8. Full-Time Status
It should be no surprise that full-time working status predicts higher salaries over part-time working status. The gap is over $18,000 per year.
III. Conclusion
When I utilize a regression analysis to attempt to understand some phenomena I always attempt to remember George E. P. Box?s assertion that,??All models are wrong; the practical question is how wrong do they have to be to not be useful.??The specified model explains a considerable amount (46%) of the variance in salaries among post-secondary degree holders employed in the labor force. However, we know there are many other important variables that influence salary levels ? motivation, self-concept, locus of control, mother and father?s socio-economic background, to name just a few. Despite the limitations of this research I am hopeful it provides some utility for individuals interested in understanding the variance of post-secondary degree holder salaries in the workforce.
Appendix A
Outback Bowl Carly Rae Jepsen dallas cowboys Rose Bowl 2013 kim kardashian anderson cooper kim kardashian pregnant
No comments:
Post a Comment