The Bureau of Labor Statistics (BLS), using surveys from the Census Bureau as well as other sources, produces regular reports and news releases (like the monthly jobs report, officially known as the Employment Situation) that are a must for business and economics journalism. However, you have to be careful about the data definitions to keep your reporting accurate and clear.
Here are some of the tripping points I’ve seen in the media.
Median versus average
This is one of the most basic and important points about statistics you can face. Median is the middle value in a range. Average is the sum of the valued divided by the number of cases. In a so-called normal distribution — the classic bell curve — the median and average are the same.
Such is not the case in most economics, particularly given the levels of income inequality that exist in the U.S. Some numbers from the BLS are reported as averages while others are medians. The bureau reports both in some cases. The implications of the two can be significantly different.
Real versus nominal wages
Wages are always an important topic. They give insight into the state of most people, their ability to stimulate the economy, and a major cost pressure for business.
BLS, and economists in general, discuss wages in two ways: nominal and real. Nominal numbers are at face value. When, in August, the monthly jobs report noted average hourly earnings were up by 77 cents, or 2.9 percent, over the year, that was about nominal wages. The same is true for the average hourly gain of almost 4 percent between July and August.
Real wages are adjusted for inflation and the cost of living. Typically, the adjustment takes a significant bite out of any gains. BLS has a real earnings report that comes out about a week after the jobs report. The earnings increase from July to August was offset by a 0.2-percent increase in the Consumer Price Index. The real average hourly earnings gain was 1 percent instead of 4 percent.
When many outlets talked about the big boost in median earnings for August, I adjusted the numbers for inflation and showed that the yearly growth was more like 1 percent overall. Always check to see if figures are nominal or real. (A hint: the BLS notes when numbers are real. If a report doesn’t somewhere explicitly mention that wages are real, then they are nominal.)
Survey methods and definitions
Much BLS analysis is of data from extensive and regular Census surveys. Methodologies, assumptions, and definitions are critical to understand what the data says.
BLS has some extensive explanations on unemployment numbers that throw light on what the figures mean. For example, did you know that someone can work for no pay for a family business 15 or more hours a week and be counted as employed? Or that a high school student with an after-school job is placed in the category of a worker, with status as a student being secondary?
The more you understand what economic figures mean and how experts interpret them, the greater the context you can offer in your work.