Economics homework help. Data Assignment #3 – Due MARCH 8 at 11:59 pm
For this assignment, you will pull data from the U.S. Census and/or the American Community Survey using the online data analysis tool created by IPUMS.org at the University of Minnesota. If you are curious, IPUMS stands for Integrated Public Use Microdata Sample, and if you weren’t curious, now you know anyhow.
For this assignment, you will choose an occupation, and examine the sex distribution of that occupation and gender earnings gap in that occupation for at least two separate years. Try to choose years that are at least 20 years apart. If you are enjoying this exercise, or feel slightly more ambitious, feel free to choose more years.
To make this easier, and to ensure that everyone is accessing the data and using the tool properly, I am providing these instructions. Note that you will need to create an account to access the FREE data. I am not providing those steps because I already have an account.
Go to https://www.ipums.org. The page should look something like this. If it does not, you are likely in the wrong place. Once you are on this page, click the “VISIT SITE” button in the “IPUMS USA” square on the top right
Click on the “Analyze Data Online” Button under “USE OUR ONLINE TOOL FOR ANALYSIS”
On the next screen, scroll all the way down until you see the title “Use data from multiple samples”. Click “United States 1850-2017”. The supercool and friendly folk at the Minnesota Population Center at the University of Minnesota have synthesized all of the data from the Decennial Census of the U.S. and the data from the American Community Survey for 1850-2017 for us (except of course for the year that some of the data was lost in a fire – anyone that tells me what Decennial Census year that was gets an extra point. They have also dealt with pesky variables that change over time as well so that the coding definitions are consistent over time.
STEP 4: Here is where they make you create a register (if you haven’t already), or log in. Sometimes they want to know what you are using the data for – “class assignment on occupational segregation over time” should suffice.
Once you have logged in, you should reach a page that looks like this. This page lets you create a table based on the underlying data. You will do this step at least twice (once for each separate year you choose).
Before you make your table, you will need to choose an occupation to analyze. To compare consistent occupations, you need to choose one of the following variables: OCC1950, OCC1990 or OCC2010. These variables are harmonized across years. Please refer to the following page in order to examine the IPUMS occupation codes for OCC2010:
On the left side, click on “Alphabetical Variable List” and scroll down to OCC2010. Click on “Variable Description” and then “Codes”. You can also do the same for whichever harmonized occupational variable you choose to analyze.
STEP 7: You will need to run this step twice – once for each year you plan to analyze
In the Row box, type “occ2010”
In the Column box, type “sex”
In the Selection Filter(s) box, type “year(year of your choosing)”, “occ2010(occupation code for chosen occupation)” – See the example below but obviously DON’T CHOOSE occ2010(1010)
Under Weight, perwt – Person weight should be selected
Under OUTPUT OPTIONS, Percentaging, make sure that Row is selected.
In the Title box, enter the occupation and the year you chose for this particular table.
STEP 8: Your output for each year should look like this. You are interested in the Results under “Frequency Distribution”. As you can see, in 1970 (I know that because even though I didn’t include it in the title, under Variables, it tells me that I applied a filter of year(1970)), there were 169,200 computer programmers in the US. Of these, 73.6 percent (124,500) were men and 26.4 percent (44,700) were women. I also reran the query for 2017. What do you think happened? Do you think the sex distribution of computer programmers changed over time?
Although I won’t show you the output for 2017, I bet you are curious so I will tell you. In 2017, there were 500,593 computer programmers in the U.S. – thus the number of computer programmers more than doubled in the past 47 years. But what about the sex distribution? Of these roughly half a million computer programmers, 77.5 percent were male (387,906) and 22.5 percent were female (112,687), so the sex distribution of computer programmers has not changed very much at all. Computer programming is still male dominated.
No, you aren’t done yet. I would like to see whether the gender wage gap in your chosen occupation has changed over time, and if so how. So you will need to do another query.
Fill out the Table Creator again.
For Row, type “incwage”
For Column, type “sex”
For Selection Filter(s), type “year(year of your choosing)”, “occ2010(occupation code for chosen occupation)” – See the example below but obviously DON’T CHOOSE occ2010(1010)
Under OUTPUT OPTIONS make sure that “Weighted” and “Summary Statistics” are selected
Once the table is generated, you will need to scroll down to the “Means”
As you can see, mean wages for male computer programmers were 9,430.24 compared to 6,389.60 for females. You should be able to calculate the gender wage gap based on these figures! You need to look at the same years you did before to determine whether and how the gender wage gap has changed in the occupation over time.
A neat table in Excel that includes the number and percent of women, men and total workers and gender wage gap in your chosen occupation in each year you examine. I would also like to see some sort of graph or chart that illustrates the data. If you have two years of data it would look something like this:
|Specify Year 1||Specify Year 2|
And of course I want you to write a paragraph or two discussing your results.