This was the official task:
This task is to estimate a Mincer earning function, including several
dummies such as for gender, region, colour, etc. You can use schooling, experience, experience squared, gender, race, union membership, wage in dollars per hour, age, occupation, sector and
marital status.
We can use the Statcrunch environment to do the calculations and to perform the statistical analysis. Steps that are typically in this process (not exhaustive, so find your own way).
1) First and above all: make sure that you all fully understand what this task entails.
What is exactly asked, what variables do we need, and what is the meaning of
these variables?
2) Go to the Statcrunch environment (www.statcrunch.com).
3) Within Statcrunch, find the CPS wage data from 1985. (Yes indeed, we are going to use the 1985 wage data based on US household data) and use the dataset from a user called sampleuser, May 25, 2007.
4) Open the dataset and find out what all columns mean. (see description in the overview of the datasets).
5) Find out what is needed for this task.
6) Create new variables (columns) that you need. What is the endogenous variable,
what are the exogenous variables (or what is the dependent variable (Y variable in Statcrunch), what are the independent (X) variables).
7) Please be careful on what dummy variables are and that you do not use categorical data as ordinary variables. Please think first on what this means.
See also a note on Canvas that is addressing this issue (Additional information on the use of dummies)
8) Estimate the most extensive equation (so including all independent variables).
9) Examine the result, leave out non-significant variables until you find something you are satisfied with. You can also make several variants and discuss them separately, as we often find in empirical papers.
10) As before, write a report on the methodology, data, steps taken and all results. Plus a conclusion on various aspects of the results, such as returns to education, returns to experience, wage difference due to gender, due to race etc. Outcomes of estimation results should be listed as we are used to in academic papers, so including significance level and/or standard error etc. plus a readable description of the dependent and independent variable(s). Please do not simply copy/paste the results from a software package into the document. Use a layout of tables as we see in the (research) papers that we address in this course.
The thing you need to do is to write an introduction of the report.
You must also tell some things about the graphs/box plots. You can find the graphs/box plots in the file i posted.
The regression results analysis will be done by someone else.
So you only need to write those parts, not the whole report!
Requirements:
Accurate grammar, spellings and punctuation; appropriate style and tone.
APA 7 style
Times New Roman 12
Include graphs/tables when necessary