How the Analysis Was Done

The Washington Post asked how the region's economic expansion is benefiting people at different points on the income scale. One problem: Most government statistics that measure inequality in wages and incomes do not break down the data by cities.

Twice a year, though, the Bureau of Labor Statistics surveys 200,000 businesses and asks them how many employees they have in each of 801 occupations and how much they are paid.

The agency aims to give a snapshot of jobs and wages, not to produce data that can be compared over time. As a result, it changed the definitions of some occupations between May 2003 and May 2005, the span that The Post analyzed. To analyze how average wages changed, The Post had to ignore some job categories. Enough data remained to examine wage changes for 570 occupations in the Washington area and 693 nationally.

The Post ranked each job from lowest pay to highest; broke the occupations into five blocks, each of which accounts for about 20 percent of the region's jobs; then calculated how much, on average, those wages changed from 2003 to 2005. The data do not include the value of health or pension benefits, stock options, partnership income, or forms of pay other than cash wages and salary.

This method is not as detailed, thorough or statistically valid as data compiled every decade by the U.S. Census Bureau. It is based on a small subset of the businesses in the region and in the country, so there is sampling error. Nevertheless, the analysis offers a rough picture of how wages are changing in jobs at different levels of the income scale. The wage changes for specific jobs are less reliable.

-- Neil Irwin