Struggling with interpreting Output of a Quantile regression

[Guest]

Hello guys,

English is not my native language but I will do my very best to express myself as clear as possible.


I have to review the academic paper "Has the Quality of Accounting Education Declined?" by Paul E. Madsen. Until today I have attended an introduction intro econometric course and have little knowledge on STATA and its output. Unfortunately until today I have never heard about Quantile Regression and I am struggling to interpret the attached output. The regression used for this output is the following:

Inc%ile_it= α + β1 * Acct_it+ β2 * College4_it+ β3 * College5_it+ β4 * Hours41t48_it+ β5 * Hours49_it
+ β6 * Metro_it+ β7 * Age_it + β8 * Black_it+ β9 * Hispanic_it+ β10 * Asian_it + β11 * AmerIndianit + β12 * MiscRace_it+ β13 * Female_it+ β14 * Married_it+ β15 * WSD_it+ β16 * Child5_it+ β17 *ChildOlder_it+ β18 * Female_it* Married_it+ β19 * Female_it* WSD_it+ β20 * Female_it* Child5_it+ β21 * Female_it* ChildOlder_it+ ε

Note: Annual earnings are turned into percentiles because they are skewed to the right and are not comparable over time because in the price level.

where

Inc%ile_it: Percentile rank of the annual earnings of worker i in yeah t among the incomes of full-time workers in all occupations in a given year (higher incomes are near the 100th percentile and lower incomes are near the 0th percentile)
and Acct_it is a dummy variable weather the individual has an accounting education or not. So this regression tries to explain the effect of an accounting education on the annual earnings of a worker.
The other variables are control variables.

Now I am struggling to understand the output of the regression. For example: What does the 4 year college value of 0.017 and the 84% percentile tells me?
Here is a short Description of the Output:

Data summarized in this table come from the IPUMS-USA database. Columns labeled “Mean” and “Med” show mean and median values across all individual-level observations. Columns labeled “Accting” and “%ile” show the mean value for the “Accountants and Auditors” occupation and the percentile location of this value among values calculated similarly for all census occupations. Total earnings is the total annual earnings of individuals in thousands of nominal dollars. Accountant is a dummy variable equal to one for people in the “Accountants and Auditors” occupation and zero otherwise. 4 years college is a dummy variable equal to one for individuals that have attended four years of college and zero otherwise. 5+ years college is a dummy variable equal to one for individuals that have attended five or more years of college and zero otherwise. Working 41 to 48 hours is a dummy variable equal to one if the individual works, on average, between 41 and 48 hours a week and zero otherwise. Working 49+ hours is a dummy variable equal to one if the individual works, on average, 49 or more hours a week and zero otherwise. Metro location is a dummy variable equal to one for workers living in an area defined by the Census Bureau as a metropolitan area and zero otherwise. Age is individual’s ages. Black, Hispanic, Asian, Indian, and misc. race are dummy variables equal to one for individuals in each racial category and zero otherwise. Female is a dummy variable equal to one for females and zero otherwise. Married, and widowed, single, divorced are dummy variables equal to 1 for individuals with the applicable relationship status and zero otherwise. Number of children under 5 years old is a count of the number of each individual’s own children younger than five years old living with them. Number of children 5 or older is a count of the number of each individual’s own children five years old or older living with them.

If you are not able to help me in this specific case because I may left out some important informations, maybe you could give me some reading recommendations regarding this topic.

Thanks in advance
Guest

[Amin Sofla]

Note: This is not an output of a quantile regression. Rather, it is an output of descriptive statistics.

[Carlo Lazzaro]

Guest:
a reference in point is: https://www.cambridge.org/core/books...37390D44A328B1

[Guest]

Thank you so much for helping me out here.
@Amin Sofia: Well, that might be the reason why I have never seen such an output using STATA