Jump to the Navigation Bar

CPS Merged Outgoing Rotation Groups

• All the data you need to explore or explain wage determination, union wage effects, inter-industry wage differentials, wage inequality, and employment discrimination.
• Used in many famous papers.
• No copyright restrictions on extracts use.
• In Stata .dta format for PC, MAC or UNIX.
• If you don't have Stata, use STAT/Transfer to convert to other popular PC statistical package or database formats.
Updates and changes. Send questions to Jean Roth at jroth@nber.org


---

Frequently Asked Questions:

What is the CPS?
This is the monthly household survey conducted by the Bureau of Labor Statistics to measure labor force participation and employment. 50-60,000 households per month are queried.
Is this a panel survey?
Not really. Households are not followed if they move. But see the next question.
What are the Outgoing Rotation Groups?
Every household that enters the CPS is interviewed each month for 4 months, then ignored for 8 months, then interviewed again for 4 more months. Usual weekly hours/earning questions are asked only at households in their 4th and 8th interview. These outgoing interviews are the only ones included in the extracts. New households enter each month, so one fourth the households are in an outgoing rotation each month.
Is this the same as the March Annual Demographic Survey?
No, income and demographic questions asked in March are not available on the extracts. Weekly earnings are available, annual earnings are not. Basic demographic variables (age, sex, race, and marital status) are available, but not detailed information about children's ages (except 1984-1993) or annual income.
What do the weights sum to?
The total of weights for all the persons in the sample sums to the total population 16 years of age or greater. But only one fourth of the sample is in an outgoing rotation group. So 12 months of data sums to three times the total population.
Is software included on the extracts?
No.
Can I use a spreadsheet?
No, the files are much too large for that. You will need some experience with a statistical package or database language to make use of this data.
What format are the files?
We convert the original EBCDIC files to Stata binary save files for ease of use. The Stata files are compact and portable across operating systems and machine types. A year's data can be read off the disc in only a few seconds. If you prefer another statistical package, the program STAT/Transfer from Circle systems will convert Stata files into many popular PC formats.
Can I get just the most recent year or years?
No. When you order the most recent year or years, you get the previous years FREE.
What variables are included?
The following list is typical, there is some variation over the years:

Interview Month Month in sample State MSA/PMSA FIPS code PMSA ranking CMSA/MSA ranking MSA/CMSA size CMSA code Metropolitan status code Central city code Household ID Sex Veteran Highest grade attended Whether completed highest grade What was doing most of last week How many hours last all jobs Usually works >= 35 hrs at this job Why not at least 35 hours last week Class of worker Usual hours Paid by the hour Earnings per hour Usual earnings per week Union member Covered by a union contract Enrolled as a student full/part time relationship to reference person

Age Marital status Race How many hours last week? Reason less than or equal to 35 hours last w Why absent from work last week? 3-digit industry code 3-digit occupation code Class of worker Usual hours Paid by the hour Union member Ethnicity Labor force status recode Full-time or part-time status Detailed industry code Detailed occupation code Earnings eligibility flag Class of worker 2 Earnings per hour Earnings per week Final weight Earnings weight for all races Usual hours (I25a) allocation flag Paid by hour (I25b) allocation flag Earnings/hr (I25c) allocation flag Usl Earn/hr (I25d) allocation flag


Note: The variable "How many weeks a year do you get paid for?", is NOT on the extracts becuase it has considerably less utility than one might anticipate. The question is asked only of the 12% of earners who report their earnings as an annual amount. Almost 90% of these earners report working 52 weeks per year.

---