Spectacular new ways to slice and dice unemployment data since 2007
June 27, 2012 — When you’ve looked for monthly unemployment data, you have probably noticed that you can pretty readily find those data sliced by race or gender. Sometimes you can find the data arranged by age group or by educational attainment.
But each of those dimensions — race or ethnicity, gender, age, and educational attainment — operate at once to create demographic composites for which the unemployment rate looks very different. For example, Hispanic men of any age with a bachelor’s degree now have a 12-month moving unemployment rate of 5.8 percent. For Hispanic men who are 16 to 25 and who are not high school graduates, on the other hand, that rate is now 22.3 percent.
Remapping Debate has created a striking new set of interactive tools to allow you to explore how each of these factors work together.
We analyzed unemployment data from the Current Population Survey of the Census Bureau (the data that are used by the Bureau of Labor Statistics) from 2006 to the present. It is our intention to update these data each month.
In order to account for seasonal variation and to reduce the margin of error, the results reported for each month represent a moving average of that month and the preceding 11 months. Thus, for example, the first month represented in the visualization is Jan. 2007, comprising a moving average of the months from Feb. 2006 through and including Jan. 2007.
In the visualization below, you are able to create a demographic composite by making a selection for each of the dimensions discussed. If you wish to create another composite and compare results, we suggest that you open another window.
Note: We started with 300 demographic composites. When one of them had a standard error equal to 10 percent or more of the unemployment estimate for a 12-month moving average, that 12-month moving average does not appear in the visualization (you will see in a limited number of cases that the line graphing change in unemployment rate over time is not complete). In those cases where every month’s moving average was at or above our standard error threshold, you won’t be able to create the particular composite at all.
That said, there are more than 270 composites available. No 12-month moving average that is reported has a margin of error (+/-) of more than 4.5 percentage points (at a 90 percent confidence level). In most cases, that margin of error (+/-) is substantially less.
Along with the graph that tracks change over time, the visualization reports the overall average of unemployment for the entire period, the highest and lowest 12-month moving unemployment averages for the entire period, and the current 12-month moving average (May 2012 is the most recent reporting date.)
See a category in one dimension disappear when you make a selection in another dimension? That’s because you’ve come upon a composite with too high a standard error. Start selecting the “all” value in the filters until each of categories returns.
On the next page, the data are organized in table form and allow you to sort each of the available demographic composites by average, highest, lowest, and current 12-month moving averages of unemployment.
