University of Florida Professor Misleads with Statistics
BY LEN CABRERA/MARCH 1, 2019
Vincent Adejumo, a lecturer in African-American Studies at the University of Florida, recently had a column in the Gainesville Sun titled “African-Americans still recovering from Great Recession”. The column is almost a parody of progressive victimology. This rebuttal is not to argue that blacks are not economically disadvantaged, but to point out flaws in the analysis that would not be acceptable in an introductory statistics course. We can’t talk about how to improve the situation until we understand it.
Adejumo starts with average income in 2007 ($55,265 for blacks, $86,732 for whites) with no source or context. The data come from Census Bureau Report P60-263 (Income and Poverty in the United States: 2017, Table A1: Households by Total Money Income, Race, and Hispanic Origin of Householder: 1967 to 2017). The table includes median income, which is a better measure than average income, given the skewed nature of incomes. (Large, outlier incomes pull the average up, so it doesn’t accurately represent the middle of the income distribution.) The dollar values are 2017 CPI-U-RS, so they are inflation-adjusted and can be compared year-to-year.
His conclusion that the average black family earned just 64 percent of a white, non-Hispanic family incorrectly implies that families earn less simply because they are black. That confuses correlation for causation and ignores many other factors that are related to household income.
From the same dataset (Table 1. Income and Earnings Summary Measures by Selected Characteristics: 2016 and 2017), the median household with a married couple in 2017 earned $90,386. The median earnings for a female head-of-household (no spouse) was $41,703, or 46% of the married household. (Table 1 only includes median income.) Because blacks are more likely to have single parent homes (CDC data show 70% black births to unwed mothers vs. 29% for whites), you cannot attribute the lower black earnings to skin color (or any single factor that correlates to income).
Adejumo almost gets to that near the end of the column when he cites other factors like “post-secondary education attainment, business and enterprise development, rates of divorce, single parent-headed households and health disparities,” but then attributes these differences to race alone. He should read my Nov 2018 column criticizing Alachua School Board’s push for equality:
“The emphasis on skin color at the county, state, and national levels is part of the problem. Rather than emphasizing other attributes that correlate to poor school performance − poverty, single family homes, truancy, drugs, etc. − they literally only look skin deep. What does it tell a student or parent when skin color is the main focus of inequality? They can’t control skin color, but they can control some of the actual factors that lead to poor performance.”
Adejumo is using data to paint a picture and push a narrative: the financial differences between blacks and whites are purely caused by skin color and racism, not by differences in number of household earners, parental involvement, parental education level, etc. There is no attempt to investigate any factor other than race.
For example, Adejumo laments that the proportion of black households under $15,000 is double that of whites (20.3% vs. 8.7%, using 2008 data). He does not mention the proportion of black households that make good incomes: 22.1% made over $75,000 in 2008. That number increased to 25.9% in 2017 (real dollars, so it’s not just a function of inflation). Rather than focus on white vs. black and risk an incorrect causal relationship with race, comparing characteristics of blacks with household incomes under $15,000 to those over $75,000 would provide better information on why some households earn higher incomes.
Adejumo goes on to compare net worth and home ownership rates and uses the difference to say blacks are more likely to incur credit card debt and are less able to meet financial obligations. He provides no data to substantiate that claim. He uses a 40-year projection to say black households will have zero net worth by 2053. This extrapolation would receive a failing grade in any statistics course and is as unreliable as any 10-year budget forecast coming out of D.C.
In fact, a 2017 study by CareerBuilder revealed that 78% of all Americans live paycheck to paycheck. This included people earning over $100K per year. Meeting financial obligations requires discipline and some knowledge of personal finance. I’ve talked to high school students who pay $4 ATM fees to cash out $20 pre-paid cards. That’s a poor life choice that will lead to poverty, and it has nothing to do with skin color.
According to a 2013 Brookings Institute study, the key to avoiding poverty is to finish high school, get a full-time job, and wait until age 21 to get married and have children. In their study, only 2 percent of people who followed those rules were in poverty, and nearly 75 percent earned over $55,000 (2013 dollars). A 2017 study by the American Enterprise Institute found a similar result.
Rather than investigate true sources of poverty or household earnings, Adejumo blames racism and legal barriers that “impede the upward mobility of African-Americans.” That ignores the millions of successful African-Americans and is a disservice to the students in his course. It is important to investigate disparities so we can determine whether there are policy solutions that will help, but the conclusion that the disparities are caused by race or racism is not supported by the data he references in the column.