Measuring Race And Ethnicity Across The Decades: 1790-2010

Written By: Beverly M. Pratt, Lindsay Hixson and Nicholas A. Jones

Over the years, the U.S. Census Bureau has collected information on race and ethnicity. The census form has always reflected changes in society, and shifts have occurred in the way the Census Bureau classifies race and ethnicity. Historically, the changes have been influenced by social, political and economic factors including emancipation, immigration and civil rights. Today, the Census Bureau collects race and ethnic data according to U.S. Office of Management and Budget guidelines, and these data are based on self-identification.

A new interactive visualization released today shows how race and ethnicity categories have changed over time since the first census in 1790. This allows us to better understand the relationship between historical classifications and the present time. A static version of this same visualization was presented in April 2015 at the Population Association of America’s annual meeting.

We created this interactive timeline to establish a starting point for the public — including community stakeholders, academics and data users — to understand how race and ethnicity categories have changed over 220 years in the decennial census. This understanding is important as we interpret results from the 2010 Census Race and Hispanic Origin Alternative Questionnaire Experiment and the current middecade testing of race and ethnicity questions, including the 2015 National Content Test. The National Content Test will inform design changes for collecting data on race and ethnicity in the 2020 Census and other ongoing demographic and economic surveys conducted by the Census Bureau.



What can we learn from this visualization?

Some categories appeared, were removed from the census form and then reappeared throughout history. For example, the “other race” category existed in some form from 1790 through 1840, then disappeared between 1850 and 1900, and then reappeared in 1910. In addition, the visualization shows that detailed Asian groups first became part of the census form in 1860 and that more detailed Asian groups were added, such as “Hindu,” in the 1920, 1930 and 1940 censuses. We can also see when the Hispanic/Latino ethnicity category first appeared on the 1970 census sample form, while showing that in 1930, the category “Mexican” was used as a race.

Some race categories have changed while others have remained the same. The term “white” has been used since the very first census. In comparison, the descriptions of all other categories have changed and/or have been added over time.

Finally, our nation’s changing history is reflected in the categories. For example, after Alaska and Hawaii each received statehood in 1959, the 1960 Census saw the addition of the terms “Aleut,” “Eskimo” and “Hawaiian” for the first time.

How will the census form change in 2020 and beyond?

The Census Bureau is continually updating its approach to collecting, processing and categorizing all types of responses representing ethnic origins from nations around the world. The 2015 National Content Test, which is currently being conducted, will have dedicated areas so that people can report their specific nationality or ethnic origins in addition to providing responses to the standard OMB race and ethnic categories. This research and testing, along with continued input from our advisory committees and stakeholders, will ensure the 2020 Census form reflects how the U.S. population identify themselves.

Note: The terms displayed on the visualization have not been modified from their historical use.

Posted in Race | 6 Comments

Homegrowns and Rolling Stones

By Megan Benetsky, Charlynn Burd and Kin Koerber

Most people in the United States — about 59 percent — live in the state where they were born. Nevertheless, that leaves a large share of people who have moved from their home state to another part of country. This migration dynamic varies, as some states are more likely to have a mostly homegrown population while other states have large shares of people born elsewhere.

The Census Bureau released two tables last month that make it easy for researchers and local communities to explore the places where people were born compared to where they currently reside, using American Community Survey data. The first table crosses state of birth by state of current residence. The second table is a Historic Place of Birth Dashboard, an interactive table that shows the total population over time and the population born in a selected state. This is the first time these tables have been available from the American Community Survey.

As you can see below, states in the South and Midwest regions have among the largest shares of homegrown populations. States in the West are among the states with the smallest shares, or have the largest population of residents flowing in and out of the state.

Table 1. States among the largest and smallest proportions of current residents born in state
State Percent born in state of current residence
Louisiana 77.7%
Michigan 76.7%
Ohio 75.1%
Pennsylvania 73.4%
Mississippi 71.5%
Wyoming 41.0%
Arizona 39.0%
District of Columbia 36.2%
Florida 36.1%
Nevada 25.8%
Source: U.S. Census Bureau, 2014 American Community Survey


Table 1 presents data on people who currently live in the state where they were born, but there is another way to use this data. Instead of looking within a state to analyze who was born there and who was not, we can look at the distribution of all the states of current residence by the states where people were born (Figure 1). It is interesting to look at this table and ask ourselves: What varies among the set of states that makes for such different population distributions? It also begs the questions: Where do people born in these states move to and — particularly for the bottom five — where are they coming from?

To help answer these questions, we look at the two states at the extremes — Louisiana and Nevada. Of all people in the United States who were born in Louisiana, about 70 percent currently live there and about 30 percent are now scattered throughout the country. Most remain in the same general region. In comparison, only about one-quarter of people who were born in Nevada still live in Nevada. Just over 21 percent of Nevada’s current population was born outside the nation (i.e., in a U.S. island area or in a different country).



Using the second tool, we can look at the populations of Louisiana and Nevada going back to 1930. Below you can see a screen shot from the Historic Place of Birth Dashboard showing that from 1930 to 2010, Louisiana had a larger total population than Nevada, although Nevada’s population appears to be growing at a greater rate. When the ratio of the U.S. population born in a state to the state’s total population is greater than one, this indicates a net outmigration, or natives of the state have moved out at a higher rate than non-natives have moved into the state.

By using the data from Table 1 and the ratio of the total people born in a state to the state’s current population, we can make some assumptions about lifetime migration for Louisiana. About 77.7 percent of Louisiana’s population is made up of those who were born there, or “homegrowns,” (Table 1) meaning the remaining 22.3 percent are movers into Louisiana, or inmigrants. The ratio of 1.106 found on Figure 2 compares the population born in the state of Louisiana to Louisiana’s total population. This ratio means that homegrown Louisianans are about 10.6 percent larger a group than those currently residing in Louisiana. This indicates a net outmigration of homegrown Louisianans leaving compared to the number of inmigrants born elsewhere.

However, Nevada’s ratio of the population born in the state to Nevada’s current population is only 0.391. This implies that because Nevada’s current population is more than twice as large as the entire Nevada-born population, much of Nevada comprises people who moved to the state from somewhere else.

Using the 2014 state-to-state migration flows table, we know about 4.7 percent of Nevada’s population moved in from a different state in the last year, while the share of Louisiana’s inmigrants from other states and the national average is about 2.3 percent. In general, the migration rates for Nevada and Louisiana, including migration within the states, were 20.3 and 13.6 percent, respectively.

Together, these two products show how uniquely varied the populations are from state to state. Some states, like Louisiana, consist of people who are homegrown. Others, like Nevada, are made up of people from across the country.


You can find these data on place of birth and state-to-state migration flows, the Historic Place of Birth Dashboard, and more here.

Posted in Population | 1 Comment

Working in America: New Tables Detail Demographics of Work Experience

By: Braedyn Kromer and David Howard

More than seven in 10 people of traditional working age (16 to 64 years old) worked in 2014; for people 65 and over, at least one in five had worked in the past 12 months. In fact, 12.4 percent of people 70 and over continued to work.

Of these older workers, many worked full time, year round; nearly half of workers age 65 to 69 and nearly a third of workers 70 and over worked full time, year round. Among workers 65 and over, men worked longer hours each week than women; 57.6 percent of men compared with 45.5 percent of women 65 and over clocked in 35 or more hours per week.

For the first time, these findings are available from the Census Bureau’s American Community Survey through American FactFinder. A variety of interesting statistics across a number of geographies are available from two new tables. The first table (B23026) details the number of weeks worked per year and hours worked per week for men and women 65 and over (see figure 1).

Distribution of Hourse Worked by Sex for Workers 65 Years and Over, United States: 2014

The second table (B23027) details full-time, year-round work experience by age. Prior to the 2014 American Community Survey, data users could only see tabulated work experience information for the 16- to 64-year-old population as a whole. With this year’s data release, users are now able to obtain work experience data for detailed age groups, including the proportion of the population who worked in the past 12 months and the percentage of workers who worked full time, year round (see figures 2 and 3).

Proportion of the Population That Worked in the Past 12 Months (Workers) By Age, United States: 2014


Percent of Workers Who Worked full-time year round by Age, United States: 2014The statistics shown above are also available at a number of geographies. For example, the two maps below show full-time, year-round work status for workers age 16 to 24 (figure 4) and 65 and over (figure 5) by state. The proportion of workers 16 to 24 working full time, year round in each state ranged from 18.9 percent to 39.7 percent, with Rhode Island having one of the lowest and Alaska having one of the highest proportions. For workers 65 and over, the proportion of workers who worked full time, year round in each state ranged from 29.2 percent to 52.9 percent, with Maine having one of the lowest and Washington, D.C., having one of the highest proportions.

Workers 16 to 24 years old who worked full time, year round in the US: 2014

Workers 65 years and over who worked full time year round in the US in 2014

Posted in Economy, Uncategorized | Tagged , | Leave a comment

Statisticians: A Goodness of Fit Test

By One or More Census Bureau Statisticians

In honor of World Statistics Day, let’s talk about statisticians. While lots of people use statistics in their jobs every day (such as sports writers, stock analysts, weather reporters, biologists, engineers, economists, sociologists, epidemiologists, practicing physicians, nuclear physicists and lots more), only 45,145 or so (standard deviation 2,150) describe themselves as “statisticians” on the American Community Survey.

Who are these number crunchers? It seems like we should look at some numbers.

Statisticians are equally likely to be male or female. Their mean age is statistically indistinguishable from the labor force (40.8 years and a sizeable margin of error, compared with 41.4 years). Yet that mean is misleading, as they so often can be. Referring to the attached histogram, the reader will discover statisticians are concentrated far more heavily in their late 20s and early 30s than the general labor force population (Figure 1).


As part of the science, technology, engineering and math (STEM) workforce, statisticians have high educational attainment — 93.0 percent have a bachelor’s degree or more (59.5 percent have an advanced degree) among 25- to 64-year-olds compared with 35.1 percent of the comparable total labor force. They have correspondingly high earnings (median earnings of full-time, year-round statisticians of $83,461 compared with $43,545 for the comparable total working population). Statisticians also have a high probability of employment, with an employment-to-population ratio of 92.5 in 2014, compared with 68.0 for the working-age population.

And what use do they find for their statistics? Well, about one quarter of them work in government (and not all at the Census Bureau!), plus all sorts of industries employ them. All told, about nine in 10 statisticians work in 10 major industries (Figure 2).WSDRS2

Perhaps most interesting to some readers, only 54.5 percent of these often young, well-employed, highly educated men and women are presently married — barely more than half. About one third have never been married at all.

Enjoy World Statistics Day.

Posted in About the Census Bureau | 2 Comments

Women Now at the Head of the Class, Lead Men in College Attainment

Written By: Kurt Bauman and Camille Ryan

In 1940, under 5 percent of the U.S. population held a bachelor’s degree. Men, at 5.5 percent, were more likely than women at 3.8 percent, to have a college education. Although the 1.7 percentage point gap may appear small, it was big relative to the portion of women with bachelor’s degrees (it would have taken a 45 percent increase among women for them to match men).

Now, nearly 75 years after the Census Bureau began collecting these statistics, the educational attainment of our population has increased to 30 percent -and the gender balance has shifted. For the first time since measurement began in 1940, women were more likely than men to have a bachelor’s degree.

Figure 1 shows how things have changed in the last decade (from 2005 to 2014). Data from the American Community Survey show that in 2005, 28.5 percent of men had a bachelor’s degree or higher, compared with 26.0 percent of women. In 2014, the percentage for men was 29.9, while that for women was 30.2, marking the first year that women’s college attainment was statistically higher than men’s college attainment.


The trend toward higher education started with younger women. A 2004 publication noted that since 1996, young women age 25 to 29 have had higher college attainment rates than young men. However, the pattern for older men and women is different. Figure 2 shows that among women aged 25 to 34, 37.5 percent have completed a bachelor’s or higher degree, compared with 29.5 percent of men. Among women 65 and older, 20.3 percent had a bachelor’s degree, compared with 30.6 percent of men.


Figure 3 shows how women’s attainment stands relative to men across the 50 states. In some states, such as Alaska, North Dakota and Mississippi, a greater portion of women than men have completed a bachelor’s degree or higher. In other states, like Washington, Idaho and Utah, men are more likely to have a college degree.


What will the future bring? In the near future, women’s college attainment will continue to grow relative to men’s attainment, simply as a result of younger people replacing older generations. On the other hand, there is some evidence that young men may have started to keep up with young women. In the past five years, college attainment of people aged 25 to 34 has increased about as much for both women and men (2.4 percentage points for men, 2.6 percentage points for women – these estimates are not statistically different from one another).

For more information on education products, including data on school enrollment, educational attainment and field of bachelor’s degree, visit the Census Bureau’s education home page.

Posted in Education | 1 Comment