Frequently Asked Questions

---

How can a family have no income in the census?

The no income family population is small - 32,937 out of 3,662,100 families in the 5% Public Use sample for 2000 [unweighted counts]. The universe is defined as hht >=1 and hht <=3.

The finc=0 population can be divided into 3 types.

Type 1. The family has no income, but the household income is positive because a non-relative living in the household has income. This accounts for 4,224 of the cases.

Type 2. The second type, which is very small, N=97, is where the family members have a mix of positive and negative incomes, which sum to 0.

Type 3. The largest group, N=28,619, is defined as hinc=0 and finc=0. All sources of income for persons >14 have values of '0.'

These people could be living on income 'in-kind' or money from other sources of non-regular or lump sum payments.

Receipts from the following sources are not included as income: capital gains, money received from the sale of property (unless the recipient was engaged in the business of selling such property); the value of income ''in kind'' from food stamps, public housing subsidies, medical care, employer contributions for individuals, etc.; withdrawal of bank deposits; money borrowed; tax refunds; exchange of money between relatives living in the same household; and gifts and lump-sum inheritances, insurance payments, and other types of lump-sum receipts.

What is the difference between OCCSOC and OCCCEN?

For the first time, the Census Bureau is releasing two occupation variables with the census microdata. The items are:

   OCCCEN  Occupation (census code)
   OCCSOC  Occupation (Standard Occupational Classification code)
 

The two codes correspond perfectly with each other. Every code in OCCCEN is represented by one and only one code in OCCSOC. And, every code in OCCSOC is represented by one and only one code in OCCCEN.

However, the perfect correspondence is due to some collapsing of OCCSOC codes on the part of the Census Bureau. For instance, the OCCCEN code for computer scientists and system analysts is 100. This maps to the following SOC codes:

 15-1011    Computer and information scientists, research
 15-1051    Computer system analysts
 15-1099    Computer specialists, all other

For the time being, the Census Bureau has collapsed the above three OCCSOC codes to 15-10XX.

Any time there is an XX in OCCSOC, this means the Census Bureau collapsed some of the detail in OCCSOC to maintain correspondence with OCCCEN.

Users need to become familiar with OCCSOC. Future products will be released with this item. The following is the implementation schedule for using OCCSOC.

Details on the OCCSOC Coding Scheme

The standard occupation classification (SOC) was updated in 2000. The last revision was in 1980.

There are two differences between the 1980 and 2000 standard. On a technical note, there is more detail in 2000. The 2000 code is a six-digit code, whereas in 1980 it was a four-digit code. The dash between the third and fourth digits is for presentation purposes only. The coding scheme is as follows:

Description     Location        Number of categories
Major group     First two digits           23
Minor group     Third digit            96
Broad occupation    Fourth and fifth digits       449
Detailed occupation Sixth digit               821

For example, "Life, Physical and Social Science Occupations" (19-0000) is divided into four minor groups, "Life Scientists" (19-1000), "Physical Scientists" (19-2000), "Social Scientists and Related Workers" (19-3000), and "Life, Physical and Social Science Technicians" (19-4000). Life Scientists contains broad occupations such as "Agriculture and Food Scientists" (19-1010), and "Biological Scientists" (19-1020). The broad occupation Biological Scientists includes detailed occupations such as "Biochemists and Biophysicists" (19-1021), and "Microbiologists" (19-1022).

The second change is the coding scheme is substantive. In 2000, people who work together in the same job family are grouped together in their major occupational group. There are 23 major groups. For instance, the major group for doctors, nurses, technicians, etc. is:

29-1000 Health Diagnosing and Treating Practitioners.

An exception to the rule about grouping workers in a job family together is supervisors. First-line managers and supervisors of production, service, and sales workers who spend more than 80 percent of their time performing supervisory activities are classified separately in the appropriate supervisor category, since their work activities are distinct from those of the workers they supervise.

Thus, all farm workers are not in the same major group. The farmer or rancher is coded as 11-9012, where 11-0000 is [Management occupations], while farm workers are in the following major group: 45-0000 [Farming, Fishing, and Forestry occupations].

The following are some additional Standard Occupational Classification resources. They include crosswalks between the census and SOC codes, downloadable versions of the codes, coding instructions for surveys, frequently asked questions, etc.

Census Bureau site

Bureau of Labor Statistics site

---

Differences in item detail across the National (1%) and State-level (5%) Files

Due to concerns about confidentiality, the Census Bureau had to make a trade-off between level of detail in items and level of geographic detail. See link for the operational plans for the 2000 microdata release.

The national file (1%) has more item detail on variables while the state-level file (5%) has better geographic detail. This note discusses the difference in item detail.

For most variables there is no difference across the two files. However, the following variables have more detail for the 1% file: language, occupation, first ancestry, second ancestry, and detailed race recode.

For each variable, there is a spreadsheet that shows the code, description, and case counts for the 1% and 5% files. If the descriptions differ across the two files, the row for that code is highlighted in turquoise. Sometimes, these differences are very minor language differences. Other times, several codes are collapsed into one for the 5% file. It is up to the user to determine whether these differences are important.

If an item is not available for one of the files, the cell is highlighted in gray.

These are some examples of coding differences across the 1% and 5% language codes.

Code		Description (1% file)	Population Count (1%)	Description (5% file)	Population Count (5%)
	610	Flemish	144,821	Dutch	149,354
	607	Swiss	1,366,470	German, includes Luxembougian	1,386,128
	624	Cajun	17,805	Cajun	15,966
	714	Taiwanese	80,729	Formosan	85,277
	734	Balinese	568
	988			Other Pacific	23,454

In all cases, when an item is not available on the 5% file, the weighted population total on the 1% file is less than 10,000 persons nationally.

The only items not available on the 1% file are the combination items such as "Other Indo-European" or "Other Pacific." These can be produced from the 1% data by collapsing several codes into a new item.

The following sections detail the differences across the 1% and 5% files:

Language

---

What is Allocation?

The allocation variables indicate whether or not the respondent provided an answer to a question or not (such as age, sex, earnings, etc.). If a respondent left an item blank, the Census Bureau imputes a response. There are three different imputation methods: (a) based on other household characteristics [consistency allocation], (b) based on another person or housing unit [hot deck allocation], or (c) based on a pre-determined distribution [cold deck allocation]. Respondents who refuse to answer questions are not random and the questions that respondents fail to answer are not evenly distributed. For instance, in the 2000 national file (1%), 17 percent of the answers to wage and salary income are allocated, while only 2 percent of the responses to sex are allocated.

How can an intercensal estimate change?

I have an estimate from a P-25 report for July 1, 1977 and it does not agree with what is on the Census Bureau estimation web site.

The Census Bureau issues a set of "intercensal" estimates after census that are consistent with the most recent census. The intercensal estimates are produced using a mathematical formula to interpolate between two census dates.

The July 1, 1977 estimate on the Census Bureau web site is an intercensal estimate that was produced to be consistent with the 1970 and 1980 census results. The July 1, 1977 estimate in the P-25 report, no. 708 was produced before the 1980 census.

For further details on the formula used by the Census Bureau, see this draft written by Sam Davis of the Estimation Branch at the Census Bureau. Computation Technique for Intercensal Estimates

What does data defined person mean? It is an item in the microdata for 2000.

A person is considered data defined if at least two of the following items have legal entries: Name, Relationship, Sex, Age or Data of Birth, Hispanic Origin, Race.

What is augmentation?

There are two augmentation variables in the 2000 census sample. The augmentation flag for the person record is PAUG, and HAUG for the household record. Augmentation of the Census 2000 sample occurred in a relatively small number of weighting areas where the realized sample size was determined to be inadequate. A systematic sample of person and housing unit records were selected and sample data was imputed for these records. (p. 91 of the Technical Documentation of the Public Use Sample, 2000).

Another way in which corrections were made during the computer editing process was substitution. Substitution assigned a full set of characteristics for people in a household. If there was an indication that a household was occupied by a specified number of people but the questionnaire contained no information for people within the household, or if the occupants were not listed on the questionnaire, the Census Bureau selected a previously accepted household of the same size with the same demographic characteristics and substituted its full set of characteristics for this household. (p. 33 of the Technical Documentation, Public Use Sample, 2000)

Further details from the IPUMS staff describing augmentation in the 1990 census:

Augmented Person
In rare instances during the implementation of the sample weighting process, the sample size was considered inadequate to make estimates of sample data. In collection block groups with a designated 1-in-2 sampling rate, augmentation was employed if the 100% housing unit count was at least 6 and the observed sampling rate was less than 1-in-12. In collection block groups with a designated 1-in-6 or 1-in-8 sampling rate, augmentation was employed if the 100% housing unit was at least 12 and the observed sampling rate was less than 1-in-30. Augmentation was performed separately for group quarters persons using the same criteria as for the 1-in-6 or 1-in-8 designated sampling rates. Augmentation was achieved by selecting a sample of housing units (or group quarters persons) to increase the observed sampling rates to at least 1-in-12 or 1-in-30. Using the 100% characteristics, the selected households (or group quarters persons) were matched by a hot deck procedure to similar housing units (or group quarters persons) with sample data. The sample data were then copied to the augmented housing units (or group quarters persons). The augmentation rate was very small. Most augmentation occurred for group quarters persons.

• Research
  • Overview
  • Research Areas
  • Academic Disciplines
  • Affiliated Programs
• Publications
  • Overview
  • Recent PSC Publications
  • Search
  • Browse
  • Series Descriptions
  • Featured Books
  • Order
• People
  • U-M Researchers
  • Off-Campus Researchers
  • Fellows
  • Trainees
  • Staff
  • Honors
  • In the News
• Training
  • Overview
  • Predoctoral
  • Postdoctoral
  • International
    • Fogarty
  • Courses
  • Faculty
  • Jobs
  • Placements
  • Funds
    • Fogarty
    • Freedman
    • Hermalin
    • Mueller
    • Weinberg
• Data & Information Services
  • Overview
  • Data Services
  • Information Services
  • Census Support
• Events & News
  • Overview
  • Brownbag Schedule
  • Event Archive
• About
  • Center Overview
  • UM Ties
  • Employment
  • Directions
• Intranet

Contact . Census Support . Intranet . Population Studies Center . U of M . © 2008
xhtml . css