I’m working on the fourth edition of The Family, and since it’s an “accessible, data-driven introduction to contemporary sociological thinking on families,” I figured I’d share a few figures I worked on for the chapter on social class and poverty.
1. Parents’ education and educational attainment. The General Social Survey asks people what their parents’ education level was, so it’s great for measuring educational mobility. Well, sort of. It actually asks, “Were you living with both your own mother and father around the time you were 16?” and then only asks about the education level of whichever parent(s) they were living with (also, no word on how same-sex parents are treated — we’re working on it!). So, here are the results, using five education levels and two years of data:
2. Male occupational mobility. At the same time, the GSS asks people for their parents’ occupations. Using six occupational categories from the 2010 Census occupation codes, sort of ranked from higher to lower status, these are occupations of fathers and sons. The numbers shown are within-column ratios of observed percentage to total percentage. So, for example, 37.3% of sons with professional fathers are in professional occupations, compared with 22.5% of all sons, so the ratio is 1.7 (the highest value on the table). I put the Stata code for this one here: osf.io/mtwjd.
3. How rich is very rich? This one is not going in the book, but I made it while I was trying to decide how to describe the wealth of the wealthiest. What is a billion, what is a trillion? These are not easy to grasp. I thought one way to show it would be to compare wealth to overall consumption. The top 10 billionaires on Bloomberg’s list as of the other day — ten real, live individual Americans — together have just under a trillion dollars of wealth on paper. How much is that? It’s as much as all consumers spent on food eaten at home, and away from home, in the year 2020, according to the Consumer Expenditure Survey. Or, it’s as much as all consumers spent on rent, alcohol, gas and motor oil, tobacco products, and cereal and bread. Or every used car sold in a year, and all the NFL teams. I have no idea if this will be a useful device.
4. The US family income distribution, Lorenz curve. The Lorenz curve is a dry depiction of a dramatic reality. I like to use it because I want to introduce the math of inequality visually. (This is developed in 15-minutes in my blockbuster video, Measuring Inequality with the Gini Index.) That’s a little awkward because I don’t completely understand the math of the Lorenz curve. For example, one odd thing about this Lorenz curve is it looks symmetrical but it’s representing a very skewed income distribution. Apparently, it’s symmetrical when the distribution is log-normal, which is pretty common. I also shared the code for this, here, of course shortly before someone told me there is a Stata command that does this. Who knew!
Note on workflow. My work process has been pretty scattershot throughout the history of the book, with some analysis done by clicking on government websites, some done with SAS, some with Stata, some with Excel, some edited by adding text to previous versions, some by graduate assistants, and so on. There are different motivations for developing a replicable workflow for a project like this. One is to make it easier to write new editions of the book. Some figures I have repeated in each edition, as new data becomes available on a routine schedule. But it’s an intro textbook, not a research compendium. Sometimes the story actually changes. If all I did was update the figures, new editions of the book wouldn’t be as valuable; just adding new data points is not the point. Push-button revision is not the goal in many cases, even if the data structure allows it, and I never know if I’m making a figure for the last time. A second motivation is to be transparent and accountable, and to allow students and instructors to learn from the methods I use. Although not the core mission of the book, that’s also valuable, and I’m trying to improve this time around. And finally, a replicable workflow is a more reliable and robust workflow, less prone to errors and easier to correct. So it’s just good hygiene. But when I’m making 150 figures in 6 months, while doing my day job, adding an hour to make something replicable versus cobbling it together and moving on is a constant tradeoff debate.
Blog news. Actually, not blog news. Many of you follow these posts on Twitter, in addition to Facebook or by email subscription (free! sign up on this page). Without yet abandoning Twitter, I have recently started hanging around on Mastadon, experimenting with a new platform as the bird site swirls the drain. If you’re interested, you can find me there, here: https://mastodon.social/@philipncohen/.