Basic Monthly Sample Notes
The CPS is a household-level survey; accordingly, most Basic Monthly CPS samples include information about everyone in surveyed households. Labor force items are asked only of those 14 and older from 1976 to 1988 and of those 15 and older from 1989 to the present. However, full demographic information (e.g., age, sex) is collected each month for all household members, regardless of age.
Some months, in addition to labor force and basic demographic data, include supplemental variables on specific topics such as food security, fertility, education, and many others.
1976-1981
CPS basic monthly data between 1976 and 1981 do not contain individuals under age 14. However, supplement data (delivered separately from the basic monthly data) contain records for individuals whose age was less than 14. IPUMS-CPS has included these child records in the months for which they are available. These months include October 1976, October 1977, September 1978, October 1978, November 1978, May 1979, June 1979, September 1979, October 1979, November 1979, December 1979, May 1980, June 1980, September 1980, October 1980, November 1980, May 1981, June 1981, September 1981, October 1981, December 1981.
Researchers who are linking individuals across months using CPSIDP should take note of the following. The child records in these months between 1976 and 1981 have a CPSIDP value of 0 and cannot be linked across time using this variable. Child records in these samples are not assigned a unique value of CPSIDP for two reasons:
- Limited linking possibility among child records
- Exacerbation of non-unique-within-household identifiers problem
Limited Linking
Many of the months in which child records are available are adjacent or are within linking distance of another month containing child records. While children could theoretically be linked across at least a couple of months, they are never able to be linked across eight months since child records are only in the months that have a supplement. For example, a child in month-in-sample 1 (see MISH) in May of 1979 could link to its month-in-sample 2 appearance in June of 1979 but not to its third or fourth months in the CPS (July or August of 1979) because these months do not contain child records. If we follow this example through all eight months of CPS, children who are first in the CPS in May of 1979 would only have observations in half of the months that his/her co-resident adults (14+) could have appeared
1979 | ||||||||
---|---|---|---|---|---|---|---|---|
... | May | Jun | Jul | Aug | Sep | Oct | ... | |
MIS1 | ... | Children | Children | Children | Children | ... | ||
MIS2 | ... | Children | Children | Children | Children | ... | ||
MIS3 | ... | Children | Children | Children | Children | ... | ||
MIS4 | ... | Children | Children | Children | Children | ... | ||
... | ... | Children | Children | Children | Children | ... |
1980 | ||||||||
---|---|---|---|---|---|---|---|---|
... | May | Jun | Jul | Aug | Sep | Oct | ... | |
... | ... | Children | Children | Children | Children | ... | ||
MIS5 | ... | Children | Children | Children | Children | ... | ||
MIS6 | ... | Children | Children | Children | Children | ... | ||
MIS7 | ... | Children | Children | Children | Children | ... | ||
MIS8 | ... | Children | Children | Children | Children | ... |
Non-unique Person Identifiers
In this period, there is no edited version of person line number (LINENO). This means that person line numbers are duplicated in some households and do not uniquely identify household members. In these instances, CPSIDP is assigned, but this value will not appear in any other months, so the records with duplicate person line numbers are not linkable. Adding child records exacerbated the duplicate person line number problem. We chose to not link children across months rather than to reduce the number of adults who can be linked across months.