Use and interpret income measures
Knowing when to use medians over averages for better analysis and mapping results.
Knowing when to use medians over averages for better analysis and mapping results.
Esri's Data Development team produces demographic data (known as Updated Demographics) for the United States using a variety of sources to update small areas, beginning with the latest U.S. Census base along with a mixture of other private sources to capture demographic change. Alongside Updated Demographics, Esri provides U.S. Census Bureau and American Community Survey (ACS) demographics as a point of reference for understanding growth in an area and to provide additional community details.
Data tutorials educate both the novice and the expert analyst to learn more about a topic to properly incorporate Esri Demographics that are accessible within various products. In this tutorial, you will learn about the following:
First, you'll learn why income data is used and how these measures are defined.
Household Income is one of the most important indicators of financial well-being routinely used by analysts across many sectors. Knowing how to use and interpret measures of a distribution like medians and averages, allows for improved analysis and more insightful decision making. Understanding summarized income data is critical for accurate analysis. Medians, averages, and the difference between the two measures can reveal a great deal about the distribution of wealth in an area.
Household Income of an area is usually summarized in one of two measures---Median or Average Household Income. Esri provides Household Income information for nine discrete income intervals upon which Median and Average Income are computed.
Median Household Income is one of the most widely used measures of income. Medians represent the midpoint value of a distribution. As depicted with the image shown right, values at the upper and lower ends of a distribution are not taken into account.
The strength of a median value is its immunity from the influence of extreme values in the tails of an income distribution. However, this can also be a weakness of the measure, as it does not capture the overall status of an area's income.
Average Household Income sums all of the Household Incomes in an area and divides by the number of Households with reported income in the area. As depicted with the image shown right, all values of a distribution are taken into account.
The primary strength of an average is its ability to capture the entire income distribution, providing more information than a median. However, averages can be heavily skewed by a top or bottom-heavy distribution.
All Households (depicted in green) are incorporated into an average.
Next, let's compare the median and average income measures to see how different the analytic outcomes can be. Shown are two map images of the U.S. by county using Esri Median and Average Household Income estimates. For demonstration purposes, the sample U.S. Median Household Income is $56,000, while the U.S. Average Household Income is $80,000. Toggle between the two slides to compare maps.
Notice how a majority of the counties are shaded in light green in this Median Household Income map. Areas shaded in the lightest green depict Median Income values less than $25,000. The darkest green areas represent values of $100,000 or more.
In comparison, the Average Household Income map reveals more counties that are shaded in darker green.
Areas shaded with the lightest green depict Average Income values less than $35,000. The darkest green areas represent Average Income values of $125,000 or more.
The take away from mapping these two income measures is that Average Household Income tends to fall in higher income categories (dark green shaded areas) more often than Median Household Income.
Another way to visualize differences between Median and Average Income is to rank the data. This chart shows which U.S. counties rank in the top 15 for Median and Average Household Income.
Notice how Nantucket County, Massachusetts ranks 14th in the country for Median Household Income, but ranks 50th for Average Household Income.
And Marin County, CA ranks 4th in the country for Average Household Income but 16th for Median Household Income.
Understanding the differences in the median and average can help you select the right statistic for your analysis. In the U.S., the differences in these types of measurements are largely driven by high income households that far exceed the uppermost reported income category. In other words, the upper tail of the income distribution influences the average, while the median, by definition, cannot capture this information. There are also areas in the U.S. that show the opposite pattern in which averages are driven lower than medians in areas dominated by lower income households.
In some analyses it is necessary to take advantage of both income measures.
Earlier we learned how Median and Average Household Income analysis can achieve varied results, when utilized as separate measures. But what happens when the measures are used together to analyze a trade area?
The New York-Newark-Jersey City, NY-NJ-PA Metropolitan Statistical area is home to five of the top fifteen U.S. counties as measured by Median Household Income, as shown here outlined in blue.
The Median Household Income for each of these counties exceeds $100,000. With the exception of Putnam County, average household incomes stand 30 to 40 percent higher, suggesting that top-heavy income distributions characterize these affluent counties.
Mapping Average Household Income reveals where these top-heavy income distributions reside.
Somerset County is a prime example of a top-heavy income distribution. Average Household Income ($145,487) exceeds Median Household Income ($105,322) by over 38%.
More than 50% of Somerset County's households earn $125,000 or more.
Putnam County’s Average Household Income ($125,612) stands a notch below, only exceeding Median Household Income ($101,430) by 24 percent.
Identifying high income counties using only the Median Household Income will lead to affluent areas being missed in the analysis. New York County is a good example of this scenario.
As shown below, the Median Household Income for New York County is relatively lower at $77,442, while Average Income is 70 percent higher at $131,395.
Although Bergen County, New Jersey exhibits a large difference in Median and Average Household Income, ($87,806 vs. $125,284) it is not as significant as New York County. Why?
Bergen County isn't as influenced by households in the lowest income category as in New York County.
Typically, the more concentrated an income distribution is around the median, the closer the median and average values will be. Income values that stray from the median can pull the average upwards or downwards.
Medians and averages provide a snapshot of the underlying distribution and a means to easily compare two areas or even two characteristics. We've shown a few examples of how using two data measures can result in different analytical results. So which income measure is the right one to use? Here are three key points to consider when conducting your analysis:
Utilizing both medians and averages to characterize the underlying income distribution is a valuable and necessary process. Their strengths complement each other in areas with larger household bases and relatively normal distributions. In areas with small household bases use both measures to gain a complete insight into the area's income distribution.
The power behind using both the median and average measure for Household Income, is with comparative analysis. These summary measures are useful when comparing household income over time or when comparing two different geographic areas.
When analyzing demographic profiles in granular geographic areas such as census block groups, unexpected changes in medians can occur. While this is normal in areas that are undergoing rapid population change, this same outcome can be found in areas without much demographic churn. Averages are a more stable measure of change for small areas.
With any type of demographic analysis, many factors come into play. Gaining a better understanding of these income measures provides for a richer, more accurate analysis that will enhance your final results.
Summary measures of Household Income include medians and averages that are calculated from the distributions of income. In theory, all households (or points) would be in the distribution. Medians would be reported as the middle point when households are ranked in order of household income, while averages as aggregated income/by the count of households.
(Due to unavailable data, Esri applies methods to compute medians and averages of grouped data.)
A median represents the middle of the income distribution or the point that splits the distribution equally. A median is calculated from the income intervals of the distribution using Pareto interpolation, unless the median falls in the lowest (<$15,000) or highest (>$200,000) interval. For the lowest interval, linear interpolation is used.
(If a median falls in the upper interval, its reported as $200,001 due to the top-coding of households at $200,000.)
Averages are computed from estimates of aggregate income. Esri's process employs unique socio-demographic methods to model distributions and aggregates simultaneously. This top-down, bottom-up approach not only provides well-grounded small-area estimates but places emphasis on the relationship between medians and averages.
You can access Esri Demographics using Esri software and through apps like ArcGIS Business Analyst , ArcGIS for Excel , or ready to use maps from ArcGIS Living Atlas of the World . For use outside of the Esri platform data files are available in CSV, dBase, Excel, shapefile, or file geodatabase formats.
Contact an Esri data sales specialist with data questions at 800-447-9778 or send an email with your request to: datasales@esri.com.
In this tutorial, you learned about the basics of income measures, the strengths and weaknesses of each, how to compare, and select the right measures to use in your data analysis. Information covered on medians and averages in this data tutorial can be applied to any distribution. For example, Esri updates other distributions such as households by disposable income, households by net worth and by home value.
Additional data tutorials in two series are available. Click the links below for continued data exploration, learning, and ways to access the data.
Income estimates are developed using an extensive list of sources for household income trends that include both federal and proprietary sources as well as the review of national surveys including the ACS (both one-year and five-year estimates), the BEA's local personal income series, the CPS, and the BLS's Consumer Price Index. Represented as point-in-time estimates as of July 1, the data is available for Esri’s standard geographic areas and for any user-defined polygon such as a ring or drive time. Read the Esri Updated Demographics Methodology Statement for more information
Use our data reference page to help answer additional questions about Esri Demographics.
If you have a topic you would like covered in a data tutorial to help you better understand U.S. data, send us an e-mail with your topic idea.