The first official version of a household survey to determine employment status was the Monthly Report of Unemployment started by the Works Progress Administration in 1940. The Bureau of the Census took charge of this survey in 1942 and called it the Monthly Report on the Labor Force. In 1948, the name was changed to Current Population Survey. While the Bureau of the Census continues to conduct the survey each month, in 1959 the responsibility for calculating and reporting the labor force and unemployment data was taken over by the Bureau of Labor Statistics.

A Bit About the Households

Given that households generally have more than one resident, the 60,000 households included in the CPS actually generates information on almost 100,000 individuals 16 years of age or older each month. Individuals under 16 years of age are excluded from the survey due to child labor laws, mandatory school attendance, and social customs that frown on working children.

Once a household is selected for the survey, occupants are interviewed for 4 consecutive months, then not interviewed for 8 months, then interviewed again for 4 months before they are finally left alone to pursue a quiet life of spiritual and intellectual fulfillment free of Census Bureau inquiries. Using this procedure, one-fourth of the households included in the survey are replaced each month with a fresh group of faces. The Census Bureau implements the survey in this manner to: (a) maintain a degree of continuity from month to month, while at the same time, (b) not placing an excessive burden on a given household. In fact, only about 4 percent of the households included in this voluntary survey refuse to cooperate during any given month.

And A Bit About the Survey

This survey procedure is carefully devised to ensure reliability and to maintain consistency of the data collected.

• First, a week is used as the reference period (rather than a month or a day) because it is long enough to compensate for holidays and erratic, random, one-time aberrations, but short enough to give current employment status. It is also an easily defined time period.


• Second, the survey week immediately follows the reference week to ensure that respondents have an accurate recollection of their employment status. Should too much time pass between the survey week and the reference week, respondents might understandably forget what they did.


• Third, the reference week always includes the 12th day of the month to maintain consistency, to provide equal reporting periods, and to avoid monthly aberrations. Using different reference weeks each month--first week of one month, last week of the next month, then the second week of the third month--would, at the very least, provide an inconsistent reporting period. If, for example, should some employers make regular monthly employment decisions (that is, hiring new workers on the 15th of the month), then inconsistent reference weeks would measure these workers in some months but not in others.