Edit and Allocation Procedures for the Minnesota Samples
(for all years prior to 1940 and for the Hispanic Oversample)

Go Back to Editing and Allocation Procedures Index

Introduction

This section describes the procedures for internal consistency checking, logical editing and missing data allocation for the Minnesota samples. Except where specifically noted, the same procedures were used for all years prior to 1940 as well as for the 1910 Hispanic Oversample.

To make this section easier to use, the information is organized by the IPUMS variable name with the variables presented in alphabetical order. The logic of data editing and allocation, however, required a different order for the actual procedures.

First the unedited data were run against a program that checked for consistency among the given responses within each person record and then within each household record. The checking was primarily to ensure consistency among the person-record variables RELATE (relationship to household head), MARST (marital status), AGE, SEX, EMPSTAT (labor force status) and the household-record variables GQ (group quarters) and GQTYPE (group quarters type). Responses that failed the internal consistency checks were corrected if sufficient information was available—either in the subject's own record or among those of other members of the subject's household—to make a logical edit. If not, the inconsistent entry was set aside for allocation later in the process. Specific internal consistency checks are described in the data quality procedures for the applicable variables.

Variables for which there was no response indicated—or for which the response was declared invalid in the consistency checking procedure—were assigned a value in a "hot-deck" allocation process. The allocation program searches the database for the nearest suitable "donor" record from which to draw a value for the missing variable. Donors must match subjects on key predictor variables such as SEX, AGE, RELATE and MARST. While donor records can contain entries that have been edited or hot-decked, the program checks to ensure that the donated value was not itself the product of a hot-deck procedure.

In the Minnesota samples, variables with missing values are allocated in the following order:

If a variable contains any responses which have been altered from the original enumeration entries, a related "flag" variable was created to inform users of the level of data quality available. All data quality variables begin with the letter "Q" (though please note that one non-flag variable, QTRUNEMP, also begins with a "Q.") Thus, the data quality flag for MARST is QMARST. In the Minnesota samples, a value of "3" on the flag (e.g., QMARST) indicates that the value given for the main variable (in this case, MARST) was the result of a logical edit as described above. A flag value of "4" indicates that the variable was "hot-decked" with a value donated from another record. A flag value of "0" indicates an original entry which was unaltered.

The "Data Quality Flag" section of IPUMS-2000, VOLUME 1, User's Guide, contains a list of all data quality flags in the IPUMS. This section also includes a table showing the percentage of values edited for each variable.

Variables

(P) indicates a person-record variable
(H) indicates a household-record variable

AGE:
Age (P)
Flag Name: QAGE

Internal Consistency Checks:
  1. If RELATE is head/wife, child-in-law, parent/parent-in-law then AGE must be 13 or more. If MARST is "not single", AGE must be 13 or more. If employed, AGE must be 13 or more. If AGE is inconsistent with only one of these variables, consider the other variable to be missing for subsequent allocation. If AGE is inconsistent with two or more of these variables, then consider AGE to be incorrect and declare it missing for subsequent allocation.
  2. Edit cases with RELATE code of parent who are too young to be parents of the head:
    1. If the AGE of the person listed as parent is less than 10 years older than the head but more than 10 years older than the wife, if the surname of the person listed as parent doesn't agree with that of the head, if the birthplaces of the wife's mother and father match those listed for persons identified as parents, AND if the MARST of the parent is married or ever married, then change RELATE from parent to parent-in-law of head.
    2. If the MARST of the parents is single (never married), or if the head and parent have different SURSIM values and there is also a BPL conflict, then RELATE is deemed incorrect and changed to missing for subsequent allocation.
    3. If there is not enough evidence to change RELATE under rules 2a or 2b, then at least one of the AGE classifications is incorrect (AGE of head or AGE of parent). Check first to see if head has any children present who are more than 50 years younger than head; if so, make head's AGE missing for subsequent allocation. If not, assume the parent's AGE is wrong and change it to missing.
  3. Count the number of AGE conflicts (gaps of less than 10 years) between parents and children in the household. If there is only one child conflict and only one parent conflict and if the person preceding the child with the AGE conflict is the head, spouse or another child, then declare AGE for the subject child with the conflict to be missing for subsequent allocation.
    Predictor Variables:

    • AGE
    • SEX
    • SEX of head
    • RELATE - Relationship to head
    • SURSIM - Surname similarity (same as previous person, not the same)
    • MARST - Marital status
    • EMPSTAT - Employment status (yes, in labor force; no, not in labor force)
    • OCC - Occupation
    • SCHOOL - School attendance
Universe of Donors:

All persons with an unallocated AGE entry.

Procedures:
  1. If the subject has a spouse and if the spouse's AGE is given, then assign the same value for AGE as the previous person with a spouse of the same AGE and SEX.
  2. If the subject is a household head and if the head has a child present in the household with a non-missing age, then assign the same value for AGE as the previous parent of the same SEX with a child the same AGE.
  3. If the subject is a child and if the previous person in the household is a child with a non-missing age, then assign the same value for AGE as the previous person who had a sibling of the same age listed before them and who also had the same value of SCHOOL, EMPSTAT, SURSIM and MARST.
  4. If the subject is a child and the previous person listed in the household is not a child, assign the same value for AGE as the previous child who had the same age difference from the head of household and who also matched on SEX of head, MARST, SCHOOL and EMPSTAT.
  5. If the subject is not the head and the head's AGE is known, assign the same value for AGE as the previous person who had the same AGE of head, RELATE, MARST, SEX, SCHOOL and EMPSTAT.
Minimum Donor Match, Steps 1-5:

/mnnsamp No variables are dropped in any of the above AGE allocation procedures.

  1. For any remaining subjects, assign the same value for AGE as the previous person with the same RELATE, MARST, SEX, SCHOOL and EMPSTAT.
  2. If no donor is found in step 6, drop MARST and look for a match.
Minimum Donor Match, Steps 6-7:

RELATE, SEX, SCHOOL and EMPSTAT.

Flag Value: 4

AGEMONTH:
Age in months (P)
Flag Name: QAGEMONT

Universe of Donors:
1850 - 1880:
Persons under 1 year of age with an unallocated AGEMONTH entry
1910:
Persons under 2 years of age with an unallocated AGEMONTH entry
1920:
Persons 5 or under with an unallocated AGEMONTH entry


Predictor Variables:
Procedures:
  1. Assign the same value of AGEMONTH as the previous person in the universe with the same BIRTHMO (1870 and 1880).
  2. If BIRTHMO is not available, assign the same value of AGEMONTH as the previous person in the universe.

Flag Value: 4

BIRTHMO:
Month of birth (P)
Flag Name: QBIRTHMO

Universe of Donors:
1870, 1880 & 1950:
Persons under age 1 with an unallocated BIRTHMO entry
1900:
All persons with an original (unallocated) BIRTHMO entry


Predictor Variables:
Procedures:
  1. Assign the same value of BIRTHMO as the previous person who matches the subject on AGEMONTH (1870 and 1880).
  2. If AGEMONTH is not available, assign the same value of BIRTHMO as the previous person in the universe.

Flag Value: 4

BIRTHQTR:
Quarter of birth (P)
Flag Name: QBIRTHMO

Internal Edit Rule
  1. Calculate from BIRTHMO.

Flag Value: 3 or 4, same value as QBIRTHMO.

BPL:
Birthplace (P)
Flag Name: QBPL

Internal Edit Rules:
  1. Assign the same value for BPL as the most proximate BPL (MBPL, FBPL) entry that has the same MTONGUE (MMTONGUE, FMTONGUE) as the subject. (Note that the donor is not necessarily the most proximate person. For example, if BPL is not given but MBPL and MMTONGUE are known and MMTONGUE is the same as MTONGUE, then the donor entry for BPL is the subject's own MBPL entry.)
  2. If the subject is a child and if the mother is residing in the household and was born locally (i.e., in the state of enumeration), assume that the child was also born locally and assign the child the same BPL as the mother.
  3. If the subject is a child with no mother residing in the household, and if the father is present and was born locally, then assume that the child was also born locally and assign the child the same BPL as the father.
  4. If the subject is a child and if the previous person in the household is an older child and if that child was born locally, then assume the subject child was also born locally and assign the same BPL as the previous child.
    Flag Value: 3
    Predictor Variables:

    • RACE-1 (white, black, Chinese, Japanese, Native American, other Asian or Pacific; other)
    • RACE (all categories)
    • AGE (0-4, 5-9...80+)
    • MOBILSELF - Mobility status of subject (internal migrant, immigrant, non-mover)
    • MOBILMOM - Mobility status of subject's mother (internal migrant, immigrant, non-mover)
    • MOBILDAD - Mobility status of subject's father (internal migrant, immigrant, non-mover)

Universe of Donors: All persons with an unallocated BPL entry.

Procedures:
  1. Assign the same value of MOBILSELF as the previous person with the same AGE, RACE-1, MOBILMOM and MOBILDAD. If the value of MOBILSELF is non-mover, assign the same value for BPL as STATEFIP. If the value of MOBILSELF is internal migrant or immigrant, assign the same value of BPL as that of the previous internal migrant or immigrant with the same value of RACE.
  2. If the subject has children present in the household and if MBPL or FBPL (depending on sex of the subject) is missing for those children, then make the children's missing MBPLor FBPL the same as subject's BPL.

Minimum Donor Match: AGE, RACE, and MOBILMOM/MOBILDAD.

Flag Value: 4

CHBORN:
Children ever born (P)
Flag Name: QCHBORN

NOTE: If CHBORN and CHSURV are both missing, allocate values from same donor woman.

Predictor Variables:
  • AGE (15-19, 20-24, 25-29...75+)
  • MARST - Marital status (married, separated or divorced, widowed)
  • RACE (white, black, other and Hispanic; Hispanicity overrides other race categories, e.g., if "black" and "Hispanic", race in this routine is "Hispanic")
  • NCHILD - Number of own children in the household

Universe of Donors: Ever-married women with an unallocated CHBORN entry.

Procedures:
  1. Assign the same value for CHBORN as the previous person with the same AGE, MARST, RACE and NCHILD.
  2. If no donor is found in step 1, drop MARST and look for a match.
  3. If no donor is found in step 2, drop MARST and RACE and look for a match.

Minimum Donor Match: NCHILD and AGE.

Flag Value: 4

CHSURV:
Children surviving (P)
Flag Name: QCHSURV

NOTE: If CHBORN and CHSURV are both missing, allocate values from same donor woman.

Predictor Variables:
  • AGE (15-19, 20-24, 25-29...75+)
  • MARST - Marital status (married, separated or divorced; widowed)
  • RACE (white, black, other and Hispanic; Hispanicity overrides other race categories, e.g., if "black" and "Hispanic", race in this routine is "Hispanic.")
  • CHBORN - Number of children ever born

Universe of Donors: Ever-married women with an unallocated CHSURV entry.

Procedures:
  1. Assign the same value for CHSURV as the previous person with the same AGE, MARST, RACE and CHBORN.
  2. If no donor is found in step 1, drop MARST and look for a match.
  3. If no donor is found in step 2, drop MARST and RACE and look for a match.

Minimum Donor Match: CHBORN and AGE.

Flag Value: 4

CITIZEN:
Citizenship status (P)
Flag Name: QCITIZEN

Predictor Variables:
  • ETHNIC - Ethnicity based on birthplace (US outlying or Canada; Latin America, including Mexico; Northern Europe, including England, Scotland and Wales; Ireland; Other Western Europe; Southern Europe; Germany; other Eastern and Central Europe; Asia; other. If Hispanic in the Hispanic Oversample, include with Latin America)
  • SEX (1920)
  • YRSUSA1 - Years in the US (categorized: 1, 2...9, 10-14, 15-19, 20-24, 25+; given for 1900; calculated from YRIMMIG for other census years)

Universe of Donors: All foreign-born persons with an unallocated CITIZEN entry in 1920; foreign-born males with an unallocated CITIZEN entry in prior years (before 1920, women were assigned the same citizenship status as their husbands).

Procedures:
  1. Assign the same value for CITIZEN as the previous person with the same ETHNIC, SEX(1920) and YRSUSA1.
  2. If no donor is found in step 1, drop ETHNIC and look for a match.
  3. If no donor is found in step 2 and subject is in 1920 sample, drop ETHNIC and YRSUSA1 and look for a match.

Minimum Donor Match: YRSUSA1 and SEX for 1920; YRSUSA1 for all other years.

Flag Value: 4

CITY:
City (H)
Flag Name: QCITY

NOTE: CITY, CITYPOP, SIZEPL and URBAN are allocated from the same donor.

Universe of Donors: All persons with an unallocated CITY code.

Procedures:
  1. Assign the same value for CITY as that of the previous person in the file.

Flag Value: 4

CITYPOP:
City Population (H)
Flag Name: QCITY

See QCITY: CITY, CITYPOP, SIZEPL and URBAN are allocated from the same donor.

Flag Value: 4

CLASSWKR:
Class of worker (P)
Flag Name: QCLASSWK

Internal Consistency Check:
  1. If CLASSWKR indicates subject is employed, AGE must be 13 or more to be consistent. If AGE is inconsistent with CLASSWKR but not inconsistent with RELATE and/or MARST, then CLASSWKR is changed to missing for subsequent allocation.
  2. If a legitimate OCC1950 value is listed and the value of CLASSWKR is "not applicable," make CLASSWKR "missing" for subsequent allocation.
  3. If subject case is in the 1910 or the Hispanic sample and is listed as employed in EMPSTAT and no CLASSWKR value is given, then make CLASSWKR "missing" for subsequent allocation.
Internal Edit Rules:
  1. In 1910, if CLASSWKR is missing and the EMPSTAT value is "not in labor force," make CLASSWKR "not applicable".
    Flag Value: 3
    Predictor Variables:

    • OCC1950 - Occupation in 1950 coding system
    • AGE (10-19, 20-29, 30-39, 40-49, 50-59, 60+)
    • RACE (white, black, other)
    • SEX

Universe of Donors: All gainfully employed persons with an unallocated CLASSWKR entry.

Procedures:
  1. Assign same value for EMPSTAT as the previous person with the same OCC1950, AGE, RACE and SEX.
  2. If no donor is found in step 1, drop AGE and look for a match.
  3. If no donor is found in step 2, drop AGE and RACE and look for a match.
  4. If no donor is found in step 3, drop AGE, RACE and SEX and look for a match.

Minimum Donor Match: OCC1950.

Flag Value: 4

COUNTY:
County (H)
Flag Name: QCOUNTY

NOTE: COUNTY, METRO, METAREA and SEA are allocated from the same donor.

Universe of Donors: All persons with an unallocated COUNTY code.

Procedures:
  1. Assign the same value for COUNTY as that of the previous person in the file.

Flag Value: 4

DURMARR:
Duration of marriage (P)
Flag Name: QDURMARR

Internal Edit Rule
  1. If subject's spouse has a valid DURMARR, assign this value to subject.

Flag Value: 3

Variables Used:
  • SEXMAR (male, no spouse; female, no spouse; married couple)
  • AGE (10-19, 20-29, 30-39, 40-49, 50-59, 60+)
  • RACE (white, black, other; Hispanic category added if part of Hispanic Sample or if Hispanic place of birth is indicated for subject or subject's parent/s)
  • SEX
  • DURMIN (minimum possible duration of marriage for this person: age of eldest biological child; otherwise, 0)
  • DURMAX (maximum possible duration of marriage for this person: age of youngest stepchild, otherwise 14)
  • RANGE (DURMAX minus DURMIN: 1,2,3...9, 10-11, 12-13...18-19, 20-24...40-44, 45+)

Universe of Donors: All currently-married persons with an unallocated DURMARR entry.

Procedures:
  1. Predict DIF (number of years difference between marriage and DURMIN)
    1. Assign same value for DIF as the previous person with the same SEXMAR, RANGE, AGE, RACE and SEX.
    2. If no donor is found in step 1a, drop RACE and look for a match.
    3. If no donor is found in step 1b, drop RACE and AGE and look for a match.
    4. If no donor is found in step 1c, drop RACE and AGE and SEXMAR and look for a match.
  2. Calculate DURMARR: Subtract DURMIN from DIF.
  3. If DURMARR > DURMAX, change DURMARR to DURMAX.
  4. If DURMARR < DURMIN, change DURMARR to DURMIN.

Minimum Donor Match: RANGE and SEX.

Flag Value: 4

EMPSTAT:
Employment status (P)
Flag Name: QEMPSTAT

Internal Consistency Check:
  1. If EMPSTAT indicates subject is employed, AGE must be 13 or more to be consistent. If AGE is inconsistent with EMPSTAT but not inconsistent with RELATE and/or MARST, then EMPSTAT is changed to missing for subsequent allocation.
  2. If subject case is in the 1910 or the Hispanic sample and the value of EMPSTAT is other than "not in the labor force", make EMPSTAT "missing" for subsequent allocation.
    Predictor Variables:

    • OCC1950 - Occupation in 1950 coding system
    • CLASSWKR - Class of worker (self-employed, wage and salary, missing, not applicable)
    • AGE (10-19, 20-29, 30-39, 40-49, 50-59, 60+)
    • RACE (white, black, other)
    • SEX

Universe of Donors: Employees with an unallocated EMPSTAT entry.

Procedures:
  1. Assign same value for EMPSTAT as the previous person with the same OCC1950, CLASSWKR, AGE, RACE and SEX.
  2. If no donor is found in step 1, drop AGE and look for a match.
  3. If no donor is found in step 2, drop AGE and RACE and look for a match.
  4. If no donor is found in step 3, drop AGE, RACE and SEX and look for a match.

Minimum Donor Match: OCC1950 and CLASSWKR

Flag Value: 4

FARM:
Farm (H)
Flag Name: QFARM

Predictor Variables:
  • AGE - Of head (19 or less; 20-29, 30-39, 40-49, 50-59, 60+)
  • RACE - Of head (white, black, other)
  • SEX - Of head
  • HOCC - Highest occupation in household [highest rank is farmer; if no farmers in the household, rank is determined by the first digit of the OCC1950 variable from "0" (professional/technical) to "9" (laborers); limited to people related to head]
  • SIZEPL - Size of place (unincorporated; <10,000; >10,000)

Universe of Donors: All households and group quarters with an unallocated FARM entry.

Procedures:
  1. Assign same value for FARM as the previous person with the same SIZEPL, HOCC, SEX, AGE and RACE.
  2. If no donor is found in step 1, drop SEX and look for a match.
  3. If no donor is found in step 2, drop SEX and AGE and look for a match.
  4. If no donor is found in step 3, drop SEX, AGE and RACE and look for a match.

Minimum Donor Match: SIZEPL and HOCC.

Flag Value: 4

FBPL:
Father's birthplace (P)
Flag Name: QFBPL

Logical Edit Rules:
  1. If FBPL is missing but the subject's father resides in the same household, assign the same value for FBPL that appears in the father's BPL.
  2. If the subject is a child and if the previous person in the household is also a child and if the previous child's record has a value for FBPL, make the subject's missing FBPL the same value as that of the previous child.
  3. If BPL is the same as MBPL and if FBPL is missing, make FBPL the same as BPL.
    Flag Value: 3
    Predictor Variables:

    • RACE-1 (white, black, Chinese, Japanese, Native American, other Asian or Pacific; other)
    • RACE (all categories)
    • AGE (0-4, 5-9...80+)
    • MOBILSELF - Mobility status of subject (internal migrant, immigrant, non-mover)
    • MOBILMOM - Mobility status of subject's mother (internal migrant, immigrant, non-mover)
    • MOBILDAD - Mobility status of subject's father (internal migrant, immigrant, non-mover)

Universe of Donors: All persons with an unallocated FBPL entry.

Procedures:
  1. Assign the same value of MOBILDAD as the previous person with the same AGE, RACE-1, MOBILSELF and MOBILMOM. If the value of MOBILDAD is non-mover, assign the same value for FBPL as STATEFIP. If the value of MOBILSELF is internal migrant or immigrant, assign the same value of BPL as the previous person with the same value of RACE.

Minimum Donor Match: AGE, RACE, and mobility status.

Flag Value: 4

FMTONGUE:
Father's mother tongue (P)
Flag Name: QFMTONG

Predictor Variables:
  • BPL - Birthplace
  • MBPL - Mother's birthplace
  • FBPL - Father's birthplace

Universe of Donors: All mother tongue/birthplace entries (MTONGUE/BPL; MMTONGUE/MBPL; FMTONGUE/FBPL) of foreign-born persons.

Procedures:
  1. Assign the same value for FMTONGUE as the most proximate FMTONGUE (MMTONGUE, MTONGUE) entry that has the same FBPL (MBPL, BPL) as the subject. (Note that the donor is not necessarily the most proximate person. For example, if FMTONGUE is not given but MTONGUE and BPL are known and FBPL is the same as BPL, then the donor entry for FMTONGUE is the subject's own MTONGUE entry.)

Minimum Donor Match: Entry for mother tongue/ birthplace combination (either own, mother's or father's)

Flag Value: 4

GQ:
Group quarters status (H)
Flag Name: QGQ

Internal Edit Rules:
  1. In 1880 and 1850, if the first person listed in a household is not the head but rather some relative of the head (other than spouse or parent), change GQ to fragment (code 6).
  2. In 1910, fix SAMPRULE and GQ:

    Determine the number of NONRELS in the household by subtracting (1) the number of family members as given in 1910 variable NFAMILY from (2) the number of persons in the household as counted on the schedule by the data entry operator.

    If the number of NONRELS is 21 or more but a family unit (a group of related individuals) can be identified within the household, change SAMPRULE to "related individuals within a large unit" (code 3) and change GQ to "other group quarters" (code 4).

  3. Calculate the number of unrelated persons in the household in 1900 from the original household structural variables (number of household members, presence of relatives of the head, ..... to head;) add together to get number of persons in the household;
  4. If a household has a GQ classification of fragment (GQ code 6) but also has a GQTYPE value of other non-institutional GQ (code 9), where possible, assign a GQ code based on the stated GQTYPE, i.e., let GQTYPE override GQ in the case of specific institutions.

    1. After all rules and allocation procedures have been applied, if subject case is in the 1880 through 1920 samples and:
    2. If the first person listed has a RELATE code of boarder, change the GQ code to other group quarters
    3. If the first person listed has a RELATE code of member of a religious order, change GQ to other group quarters.
    4. If the first person listed has a RELATE code of employee, change GQ to other group quarters.
    5. If the first person identified in 1a through 1c has relatives listed, make appropriate changes to their GQ codes.
  5. If the first person listed has a RELATE code of institutional inmate, assign everyone else in this household a GQ of institution.
  6. In 1910, if a household with a GQ code of fragment has 15 or more persons, change GQ to other group quarters (code 4).

Flag Value: 3

GQFUNDS:
Group quarters funding code (H)
Flag Name: QGQFUNDS

Predictor Variables:
  • GQTYPE

Universe of Donors: Group quarters units with an unallocated GQFUNDS entry.

Procedures:
  1. Assign the same value of GQFUNDS as the previous person with the same GQTYPE.

Minimum Donor Match: GQTYPE

Flag Value: 4

GQTYPE:
Group quarters type (H)
Flag Name: QGQTYPE

Internal Edit Rule:
  1. Identifying Missing Institution Codes. Count the number of (a) boarders and lodgers, (b) employees and (c) "other" (neither boarders, lodgers nor employees). If there are some boarders, no employees and no more than one "other" in the household, then assign a GQTYPE of boarding house. If there are no boarders and the rest are employees or no more than one "other" in the household, then assign a GQTYPE of "other worksites.
  2. When all rules and allocation procedures have been applied, if subject case is in 1880 through 1920:

    1. If the first person listed has a RELATE code of boarder, change GQTYPE to rooming house.
    2. If the first person listed has a RELATE code of member of a religious order, change GQTYPE to other religious (institution).
    3. If the first person listed has a RELATE code of employee, change GQTYPE to other worksites.
    4. If the first person identified in 2a through 2c has relatives listed, make appropriate changes to their GQTYPE codes.
Flag Value: 3
Predictor Variables:
  • AGE of head (under 16, under 65, 65 or over)
  • SEX of head
  • OCC of head (work camp, not employed, other)
  • HHSIZE - number of people in the household (number of people taken: under 60 in household; under 120; under 9000; 9000 or more)
  • RELTYPE - relationship to head (partner/roommate; boarder or other non-relative; inmate; employee)

Universe of Donors: Group quarters units with an unallocated GQTYPE entry.

Procedures:
  1. Assign the same value for GQTYPE as the previous person who has the same AGE, SEX, OCC, HHSIZW and RELTYPE.
  2. If no donor is found in step 1, drop SEX, AGE and OCC. Minimum Donor Match: HHSIZW and RELTYPE.
  3. If no donor is found in step 2, if the person is in an institution, then assign the value of 496 "institution combination or unknown" to GQTYPE. If subject is not in an institution, assign a value of 500 "non-institutional group quarters" to GQTYPE.

Flag Value: 4

IND:
Industry (P)
Flag Name: QIND

NOTE: Values for OCC1950, OCC, OCCSCORE, SEI, IND, and IND1950 are assigned together from the same donor. See OCC1950 for allocation procedure when OCC is missing.

Predictor Variables:
  • OCC1950 - Occupation in 1950 coding system
  • SEX

Universe of Donors: Persons of any age with an occupation recorded in OCC and an unallocated IND entry.
Procedure when OCC is Not Allocated:

  1. Assign same value for IND as the previous person with the same OCC1950 and SEX.
  2. If no donor is found in step 1, drop SEX and look for a match. Minimum Donor Match: OCC1950.

Flag Value: 4

IND1950:
Industry, 1950 Basis (P)
Flag Name: QIND

NOTE: Values for OCC1950, OCC, OCCSCORE, SEI, IND, and IND1950 are assigned together from the same donor. See OCC1950 for allocation procedure when OCC is missing.

LANGUAGE:
Language spoken (P)
Flag Name: QLANGUAG

Internal Edit Rule

If the subject is native born and both parents are native born, set LANGUAGE to English.
Flag Value: 3
Predictor Variables:

  • MMTONGUE - Mother's mother tongue
  • FMTONGUE - Father's mother tongue
  • BPL - Birthplace (native born, foreign born)
Procedures:
  1. Assign the same value for LANGUAGE as own MMTONGUE, if available. If the mother is native-born, assign the same value as FMTONGUE.

Minimum Predictors: MMTONGUE or FMTONGUE and BPL (foreign born status).

Flag Value: 3

LIT:
Literacy (P)
Flag Name: QLIT

Predictor Variables:
  • AGE (single years through age 30, then 5 year groups through age 79, 80+),
  • SEX
  • RACE (white, black, other),
  • OCC1950 - Occupation group (11 categories defined by the first digit of occ1950 with the 11th category being "not in labor force").
Universe of Donors:
1850 - 1860:
All persons age 20 and over with an unedited LIT entry
1870 - 1920:
All persons age 10 and over with an unedited LIT entry


Procedures:
  1. Assign same value for LIT as the previous person with the same AGE, SEX, RACE and OCC1950.
  2. If no donor is found in step 1, drop OCC1950 and look for a match.
  3. If no donor is found in step 2, drop OCC1950 and RACE and look for a match.

Minimum Donor Match: AGE and SEX.

Flag Value: 4

MARST:
Marital status (P)
Flag Name: QMARST

Internal Consistency Checks:
  1. If the subject has a RELATE value of child-in-law and MARST is never married, check to see if there is a wife or husband in the preceding record or a grandchild in the following record. If so, this is evidence that the child-in-law relationship is correct but MARST is inconsistent. Change MARST to missing for subsequent allocation.
  2. If the second person listed has a RELATE value of wife, the value for MARST must be married for both her and the head. If the values for MARST are different or missing, change them.

Logical Edit Rule: Persons under AGE 12 are given a MARST code of never married (code 6).

Flag Value: Universe check; record not flagged.

Predictor Variables:
  • AGE (single years from 0 to 79, 80+)
  • SEX
  • RELATE - Relationship to head
  • NCHILD - Presence of own children (0, 1+)
  • SURSIM - Surname similarity
  • RACE (white, black or other)

Universe of Donors: Persons aged 12 and over with an unallocated MARST entry.

Procedures:
  1. Assign the same value for MARST as the previous person who matches subject on AGE, SEX, RELATE, NCHILD, SURSIM, and RACE.
  2. If no donor is found in step 1, drop RELATE and look for a match.
  3. If no donor is found in step 2, drop RELATE and SURSIM and look for a match.
  4. If no donor is found in step 3, drop RELATE, SURSIM and NCHILD and look for a match.
  5. If no donor is found in step 4, drop RELATE, SURSIM, NCHILD and SEX and look for a match.

Minimum Donor Match: AGE and RACE.

Flag Value: 4

MBPL:
Mother's birthplace (P)
Flag Name: QMBPL

Internal Edit Rules:
  1. If MBPL is missing but the subject's mother resides in the same household, assign the same value for MBPL that appears in the mother's BPL.
  2. If the subject is a child and if the previous person in the household is also a child and if the previous child's record has a value for MBPL, make the subject's missing MBPL the same value as that of the previous child.
  3. If BPL is the same as FBPL and if MBPL is missing, make MBPL the same as BPL.

Flag Value: 3

Predictor Variables:
  • RACE-1 (white, black, Chinese, Japanese, Native American, other Asian or Pacific; other)
  • RACE (all categories)
  • AGE (0-4, 5-9...80+)
  • MOBILSELF - Mobility status of subject (internal migrant, immigrant, non-mover)
  • MOBILMOM - Mobility status of subject's mother (internal migrant, immigrant, non-mover)
  • MOBILDAD - Mobility status of subject's father (internal migrant, immigrant, non-mover)

Universe of Donors: All persons with an unallocated MBPL entry.

Procedures:
  1. Assign the same value of MOBILMOM as the previous person with the same AGE, RACE-1, MOBILSELF and MOBILDAD. If the value of MOBILMOM is non-mover, assign same value for MBPL as STATEFIP. If the value of MOBILMOM is internal migrant or immigrant, assign the same value of MBPL as that of the previous person with the same value of RACE.

Minimum Donor Match: AGE, RACE, and mobility status.

Flag Value: 4

METAREA:
Metropolitan AREA (H)
Flag Name: QCOUNTY

See COUNTY: COUNTY, METRO, METAREA and SEA are allocated from the same donor.
Flag Value: 4 ???

METRO:
Metropolitan Status (H)
Flag Name: QCOUNTY

See COUNTY: COUNTY, METRO, METAREA and SEA are allocated from the same donor.

Flag Value: 4 ???

MMTONGUE:
Mother's mother tongue (P)
Flag Name: QMMTONG

Predictor Variables:
  • BPL - Birthplace
  • MBPL - Mother's birthplace
  • FBPL - Father's birthplace

Universe of Donors: All mother tongue/birthplace entries (MTONGUE/BPL; MMTONGUE/MBPL; FMTONGUE/FBPL) of foreign-born persons.

Procedures:
  1. Assign the same value for MMTONGUE as the most proximate MMTONGUE (MTONGUE, FMTONGUE) entry that has the same MBPL (BPL, FBPL) as the subject. (Note that the donor is not necessarily the most proximate person. For example, if MMTONGUE is not given but MTONGUE and BPL are known and BPL is the same as MBPL, then the donor entry for MMTONGUE is the subject's own MTONGUE entry.)

Minimum Donor Match: Entry for mother tongue/birthplace combination (either own, mother's or father's).

Flag Value: 4

MORTGAGE:
Mortgage status (H)
Flag Name: QMORTGAG

Internal Edit Rule
  1. If OWNERSHP is "no," then MORTGAGE is "no." Flag Value: universe check; record not flagged.
    Predictor Variables:

    • HOCC - Highest occupation in household [highest rank is farmer; if no farmers in the household, rank is determined by the first digit of the OCC1950 variable from "0" (professional/technical) to "9" (laborers); limited to people related to head]
    • SEX - Of head
    • RACE - Of head (white, black, other)
    • FIRSTHH - First household in dwelling? (yes/no)

Universe of Donors: All owner-occupied housing units with an unallocated MORTGAGE entry.

Procedure for Remaining Cases (when OWNERSHP is other than "yes"):
  1. Assign same value for MORTGAGE as the previous person who matches subject on HOCC, FIRSTHH, SEX, AGE, SEX and RACE.
  2. If no donor is found in step 1, drop SEX and look for a match.
  3. If no donor is found in step 2, drop SEX and AGE and look for a match.
  4. If no donor is found in step 3, drop SEX, AGE and RACE and look for a match.
Procedure if both MORTGAGE and OWNERSHP are missing:
  1. Assign same value for OWNERSHP and MORTGAGE as the previous person with the same HOCC, AGE, SEX, RACE and FIRSTHH.
  2. If no donor is found in step 1, drop SEX and look for a match.
  3. If no donor is found in step 2, drop SEX and AGE and look for a match.
  4. If no donor is found in step 3, drop SEX, AGE and RACE and look for a match.

Minimum Donor Match: HOCC and FIRSTHH.

Flag Value: 4

MOUNEMP:
Months unemployed last year (P)
Flag Name: QQTRUNEM

NOTE: Only cases with illegible values are allocated; blanks are not allocated.

Predictor Variables:
  • OCC1950 - Occupation in 1950 coding system
  • RACE (white, black, other)
  • SEX

Universe of Donors: Persons who reported an occupation with an unallocated MOUNEMP entry.

Procedures:
  1. Assign same value for MOUNEMP as the previous person with the same OCC1950, RACE and SEX.
  2. If no donor is found in step 1, drop RACE and look for a match.
  3. If no donor is found in step 2, drop RACE and SEX and look for a match.

Minimum Donor Match: OCC1950

Flag Value: 4

MTONGUE:
Mother tongue (P)
Flag Name: QMTONGUE

Predictor Variables:
  • BPL - Birthplace
  • MBPL - Mother's birthplace
  • FBPL - Father's birthplace

Universe of Donors: All mother tongue/birthplace entries (MTONGUE/BPL; MMTONGUE/MBPL; FMTONGUE/FBPL) of foreign-born persons.

Procedures:
  1. Assign the same value for MTONGUE as the most proximate MTONGUE (MMTONGUE, FMTONGUE) entry that has the same BPL (MBPL, FBPL) as the subject. (Note that the donor is not necessarily the most proximate person. For example, if MTONGUE is not given but MMTONGUE and MBPL are known and MBPL is the same as BPL, then the donor entry for MTONGUE is the subject's own MMTONGUE entry.)

Minimum Donor Match: Entry for mother tongue/birthplace combination (own, mother's or father's)

Flag Value: 4

NUMPERHH:
Number of persons in the household (H)
Flag Name: QNUMPERH

Predictor Variables:
  • GQTYPE

Universe of Donors: All households and group quarters with an unallocated NUMPERHH entry.

Procedures:
  1. Assign the same value of NUMPERHH as the previous person with the same GQTYPE.

Minimum Donor Match: GQTYPE

Flag Value: 4

OCC:
Occupation (P)
Flag Name: QOCC

See OCC1950: Values for OCC1950, OCC, OCCSCORE, SEI, IND, and IND1950 are assigned together from the same donor. Blanks in the occupation variables indicate individuals who are not employed. Allocations for OCC and OCC1950 are made only for those entries that are illegible or otherwise unclassifiable.

OCC1950:
Occupation, 1950 basis (P)
Flag Name: QOCC

NOTE: Values for OCC1950, OCC, OCCSCORE, SEI, IND, and IND1950 are assigned together from the same donor. Blanks in the occupation variables indicate individuals who are not employed. Allocations for OCC and OCC1950 are made only for those entries that are illegible or otherwise unclassifiable.

Internal Edit Rule: 1. In 1920, if there is a valid CLASSWKR, but no OCC1950, make OCC1950 "illegible" for subsequent allocation.

Predictor Variables:
  • EMPSTAT* - Employment status (employed, not employed, not in labor force, missing)
  • CLASSWKR* - Class of worker (self-employed, wage and salary, missing, not applicable)
  • AGE (10-19, 20-29, 30-39, 40-49, 50-59, 60+)
  • RACE (white, black, other)
  • SEX
  • WORKER - Is there evidence that the subject is a worker (yes/no); this variable distinguishes those for whom the occupation entry given is imprecise but clearly indicates gainful employment (e.g., "works") from those for whom the occupation entry is completely illegible.

* These variables are not available in all years

Universe of Donors:

Persons within age range with an unallocated OCC1950 entry:

1850:
Males age 15 and over
1860:
Persons age 15 and over
1870:
All persons
1880:
Persons age 10 and over and others with a regular occupation
1900:
Persons age 10 and over who worked or looked for work during previous 12 months, and non-working "capitalists"
1910 - 1920:
All persons


Procedures:
  1. Assign same value for OCC1950 as the previous person with the same EMPSTAT, CLASSWKR, AGE, RACE, SEX and WORKER.
  2. If no donor is found in step 1, drop AGE and look for a match.
  3. If no donor is found in step 2, drop AGE and RACE and look for a match.
  4. If no donor is found in step 3, drop AGE, RACE and SEX and look for a match.

Minimum Donor Match: EMPSTAT, CLASSWKR and WORKER in years when EMPSTAT and/or CLASSWKR are available; otherwise, WORKER.

Flag Value: 4

OCCSCORE:
Occupational income score (P)
Flag Name: QOCC

See OCC1950: Values for OCC1950, OCC, OCCSCORE, SEI, IND, and IND1950 are assigned together from the same donor. Blanks in the occupation variables indicate individuals who are not employed. Allocations for OCC and OCC1950 are made only for those entries that are illegible or otherwise unclassifiable.

OWNERSHP:
Ownership of dwelling (H)
Flag Name: QOWNERSH

Internal Edit Rule:
  1. If the subject's value for MORTGAGE is "no, own free and clear" or "yes--own and have a mortgage", then code OWNERSHP "yes." If the value for MORTGAGE is "not an owner", then OWNERSHP is coded "no."

Flag Value: 3

Predictor Variables:
  • HOCC - Highest occupation in household [highest rank is farmer; if no farmers in the household, rank is determined by the first digit of the OCC1950 variable from "0" (professional/technical) to "9" (laborers); limited to people related to head]
  • SEX - Of head
  • RACE - Of head (white, black, other)
  • FIRSTHH - First household in dwelling? (yes/no)
Universe of Donors:
1900:
Units with an unallocated OWNERSHP entry which are not boarders and lodgers
1910 - 1920:
Units with an unallocated OWNERSHP entry which were not sampled as a large unit.


Procedure for Remaining Cases (not covered by Rule when OWNERSHP is known):
  1. Assign same value for OWNERSHP or MORTGAGE as the previous person who matches subject on HOCC, FIRSTHH, SEX, AGE, SEX and RACE.
  2. If no donor is found in step 1, drop SEX and look for a donor.
  3. If no donor is found in step 2, drop SEX and AGE and look for a match.
  4. If no donor is found in step 3, drop SEX, AGE and RACE and look for a match.
Procedure if both MORTGAGE and OWNERSHP are missing:
  1. Assign same value for OWNERSHP and MORTGAGE as the previous person with the same HOCC, AGE, SEX, RACE and FIRSTHH.
  2. If no donor is found in step 1, drop SEX and look for a match.
  3. If no donor is found in step 2, drop SEX and AGE and look for a match.
  4. If no donor is found in step 3, drop SEX, AGE and RACE and look for a match.

Minimum Donor Match: HOCC and FIRSTHH.

Flag Value: 4

REGION:
Census Region and Division (H)
Flag Name: QSTATE

See STATEICP/STATEFIP: REGION and STATE are allocated from the same donor.

Flag Value: 4 ???

RELATE:
Relationship to household head - general code only (P)
Flag Name: QRELATE

Internal Consistency Checks:
  1. Misidentified Heads of Household: If the first person listed (normally the head) is identified as a wife, mother or father, change the relationship to head if it's a regular household (GQ code of 1 or 2). No change is made if the household is in Group Quarters since a household beginning with a relative of the head indicates a household fragment (GQ code 6).
  2. Non-relatives confused with Relatives: If the first person listed is a boarder or employee and not part of a regular household, see if there is anyone else with a relationship to head (e.g., wife, child, ). If so, change them to whatever the first person was (e.g., boarder).
  3. Never-married Child-in-Law: Check to see if there is a wife or husband in the preceding record or a grandchild in the following record. If there is no sign of a spouse the person might be married to, then change the relationship from child-in-law to adopted.
  4. Multiple Heads: If there is more than one head in a household, the second head and everyone subsequent is treated as missing on relationship and included in the allocation process.
  5. Inconsistent entries with MARST and SCHOOL: If the subject is in school and MARST is never married and RELATE is mother, father, or parent-in-law, then declare RELATE missing for subsequent allocation.
  6. RELATE inconsistent with SEX or MARST: If the wife is listed as male and the head is also male, change the wife's SEX code to female. If there's a wife, make her married and make her husband married, too.
  7. Wife confused with Mother: If the second person in the household is listed as a mother, if she's married, if her surname is the same as the head, if she is the opposite sex of the head, if the head is also married and if the age difference is less than 15 years, then she is really the head's wife not the head's mother and the RELATE code is changed to wife.
  8. AGE inconsistent with Other Variables: If RELATE is head/wife, child-in-law, or parent/parent-in-law, then AGE must be 13 or more. If MARST is other than single, AGE must be 13 or more. If employed, AGE must be 13 or more. If AGE is inconsistent with two or more of these variables, then consider AGE to be incorrect and declare it missing for subsequent allocation. If AGE is inconsistent with only one of these variables, consider the other variable to be inconsistent with age and declare it missing for subsequent allocation.
  9. Check and fix Wife not in second position: Look at the person immediately preceding the wife and if SURSIM, MARST, SEX and AGE difference are consistent with a wife, give the "wife" the RELATE code appropriate for the wife of the preceding person (e.g., daughter-in-law for son; sister-in-law for brother).
  10. Check and fix Child not preceded by another child or a head or a wife: If a subject has a RELATE code of child and is either male or single, if the previous person is not a child or a parent, if the subject's surname is different from the head OR if the age difference is implausible for a parent/child relationship (i.e., less than 15 years or greater than 69 years for father, less than 15 years or greater than 44 years for mother), then it is likely that the subject is not the child of the head. If the subject does have the same surname as the person right before them and the age difference is plausible for a parent/child relationship, then assign the RELATE code appropriate for child of the person immediately preceding the subject (i.e., grandchild for daughter; niece/nephew for brother).
  11. Eliminate non-polygamous wives: Women identified as polygamous (primarily in Utah) have a separate code for second wife. If there is more than one non-polygamous wife, the RELATE code for the second woman identified as wife and all relatives following the second wife, are declared missing for subsequent allocation.
  12. People listed as Parents who are too young to be parents of the head:

    1. If the AGE of the person listed as parent is less than 10 years older than the head but more than 10 years older than the wife, if the surname of the person listed as parent doesn't agree with that of the head, if the birthplaces of the wife's mother and father match those listed for persons identified as parents, AND if the MARST of the parent is married or ever married, then change RELATE from parent to parent-in-law of head.
    2. If the MARST of the parents is single (never married), or if there is a surname difference between head and parent and there is also a birthplace mismatch, then RELATE is deemed incorrect and changed to missing for subsequent allocation.
    3. If there is not enough evidence to change RELATE under rules 12a or 12b, then at least one of the AGE classifications is incorrect (age of head or age of parent). Check first to see if head has any children present who are more than 50 years younger than head; if so, make head's AGE missing for subsequent allocation. If not, assume the parent's AGE is wrong and change it to missing.
  13. Fix children too old relative to head: Look for AGE conflicts (gaps of less than 10 years) between parents and children.

    1. For households with more than one child conflict and more than one parent conflict: If there's a parent of the head present, assume the children belong to the parents of the head rather than to the head and change the RELATE codes to siblings (of head). If there are no parents of the head present, change the RELATE codes for all children with a gap under 16 years to missing for subsequent allocation.
    2. For households with more than one child conflict but only one parent conflict: If there's only 1 child with an age gap under 10 years, we assume the problem is not with the age of the child but either the relationship of child or the age of parent is wrong. If SURSIM is the same, then the problem is the parent's AGE and it is changed to missing. If SURSIM is different, the RELATE codes for all children with an AGE gap of less than 16 years is changed to missing.
    3. For households with only one child conflict but more than one parent conflict: If there are any children with a SURSIM code different from the head or from the wife, then change the RELATE codes of all children with an AGE gap of less than 16 years (from the AGE of the head) to missing.
    4. For households with only one child conflict and only one parental conflict: If the person preceding the child with an age gap under 10 years is not the head, spouse, or child then the problem is assumed to be the child's RELATE code; otherwise the problem is assumed to be the child's AGE. In either case, the value is changed to missing.
  14. After all other inconsistencies have been corrected, check 1880 cases to see if the first person listed has an AGE less than 12. If so, change GQ for everyone in these households to fragment (GQ code =6).
Internal Edit Rules:
  1. Where possible, infer RELATE from the detailed GQTYPE code (e.g., if GQTYPE is prison, assign inmate for RELATE).
  2. If the subject is the first person listed in the household and is not in group quarters, assign head for RELATE.
  3. If subject has no value for RELATE and the previous person is not related to the head, assign the same value for RELATE as the previous person.
  4. If the subject is not white and the head is white, then assign the same value of RELATE as the previous person with the same configuration of white head/nonwhite subject.
  5. If the subject is the second person listed in the household and the first person is the head, then if

    • the number of wives in the household is 0,
    • the SEX of the first person is not female and the SEX of the subject is not male,
    • MARST for both people is either married or missing,
    • SURSIM is the same for both,
    • AGE of the first person is over 16 and Age of the subject is over 15,
    • and if the first person is no more than 15 years younger nor 20 years older than the second person,

    then assign a RELATE of wife to the subject.

  6. If the subject and an adjacent person are of the opposite SEX, and if

    • both have a MARST of married,
    • both have the same SURSIM,
    • both have an AGE of 16 or more,
    • the AGE difference between them is within 15 years,

    then assume that the two are married and assign a RELATE appropriate relative to the adjacent spouse (e.g., if spouse is child, then subject is child-in-law; if spouse is a boarder, then the subject is also a boarder). Flag Value: 4 for Rule #4; 3 for all other rules.

Predictor Variables if Not in Group Quarters:
  • AGEDIF - between subject and head (-75, -74 ....+74, +75).
  • SEX
  • SEXOFHEAD
  • MARST
  • SURCODE - surname similar to head or any of head's relatives (yes/no)

Universe of Donors: All persons not in group quarters with an unallocated RELATE entry.

Procedure if Not in Group Quarters:
  1. Assign the same value of RELATE as the previous person with the same AGEDIF, SEX, SEXOFHEAD, MARST and SURCODE.
  2. If no donor is found in step 1, ignore MARST and look for a match on the remaining four.
  3. If no donor is found in step 2, replace MARST, ignore SEXOFHEAD and look for a match.
  4. If no donor is found in step 3, replace SEXOFHEAD, ignore SEX and look for a match.
  5. If no donor is found in step 4, replace SEX, ignore AGEDIF and look for a match.
  6. If no donor is found in step 5, drop AGEDIF, SEX, SEXOFHEAD and MARST and look for a match on SURCODE alone.

Minimum Donor Match: SURCODE.

Flag Value: 4

Predictor Variables if in Group Quarters:
  • GQTYPE
  • AGE
  • SEX

Universe of Donors: All persons in group quarters with an unallocated RELATE entry.

Procedure if in Group Quarters:
  1. Assign the same value of RELATE as the previous person with the same GQTYPE, AGE and SEX.
  2. If no donor is found in step 1, drop AGE and look for a match.
  3. If no donor is found in step 2, drop AGE and SEX and look for a match.
  4. When a donor is found for the first person missing RELATE in the household, then copy this value to all other cases in the block of five.

Minimum Donor Match: GQTYPE.

Flag Value: 4

Post-Allocation Consistency Checks:
  1. If the subject's SURSIM is different from the head and if the last person related to the head occurred prior to the subject, then change the subject's RELATE to non-relative.
  2. Check for inconsistencies in the AGE difference between the head and those with RELATE values of:

    1. Parent or child: children must be within 15-49 years younger than their mother and 15-59 years younger than their father.
    2. Brother or sister: the age of siblings must be within 30 years of the head.
    3. Grandchild: grandchildren must be at least 30 years younger than the head.
  3. Check non-relatives: if individuals have the same SURSIM and the same RACE as the head, then a RELATE value of non-relative is inconsistent.
  4. Correct inconsistencies identified in steps 2 and 3:

    1. If the subject is listed second in the household:

      1. If the head is male and the second person listed is female and if the AGE of the second person is greater than 16, change RELATE to wife.
      2. If not, change RELATE to child.
    2. If the subject is listed third or later in the household and a wife is present in the second position:

      1. If the age difference between the head and the subject is between 16 and 49 years, then change the subject's RELATE to child.
      2. If the age difference between the subject and the head is less than 16 years, change the subject's RELATE to sibling.
      3. If the subject is 16 to 49 years older than the head, change the subject's RELATE to parent.
      4. If the subject is more than 49 years younger than the head, change the subject's RELATE to grand-child.
      5. If the subject is more than 49 years older than the head, change the subject's RELATE to other relative.
    3. If there is no wife present in the second position

SCHOOL:
School attendance (P)
Flag Name: QSCHOOL

Predictor Variables:
  • AGE (single years through age 30, then 5 year groups through age 79, 80+)
  • SEX
  • RACE (white, black, other)
  • OCC1950 - Occupation group (11 categories defined by the first digit of OCC1950 with the 11th category being "not in labor force")

Universe of Donors: All persons with an unallocated SCHOOL entry.

Procedures:
  1. Assign the same value for SCHOOL as the previous person with the same AGE, SEX, RACE and OCC1950.
  2. If no donor is found in step 1, drop OCC1950 and look for a match.
  3. If no donor is found in step 2, drop OCC1950 and RACE and look for a match.

Minimum Donor Match: AGE and SEX.

Flag Value: 4

SEA:
State Economic Area (H)
Flag Name: QCOUNTY

See COUNTY: COUNTY, METRO, METAREA and SEA are allocated from the same donor.

Flag Value: 4 ????

SEI:
Duncan Socio-Economic Index (P)
Flag Name: QOCC

See OCC1950: Values for OCC1950, OCC, OCCSCORE, SEI, IND, and IND1950 are assigned together from the same donor. Blanks in the occupation variables indicate individuals who are not employed. Allocations for OCC and OCC1950 are made only for those entries that are illegible or otherwise unclassifiable.

SEX:
Sex (P)
Flag Name: QSEX

Internal Consistency Check:
  1. If the wife is listed as male and the head is also male, change the wife's SEX code to female.
Internal Edit Rules:
  1. Check to see if the NAME given is on the categorized list of names by SEX. If so, assign the value of SEX most commonly associated with that name.
  2. If subject's value for RELATE is wife, make SEX female and make the head of this household male if missing.
  3. If subject's value for MARST is married and if the SEX of the partner is given, make the subject the opposite sex.

Flag Value: 3

Predictor Variable:
  • OCC

Universe of Donors: All persons with an unallocated SEX entry.

Procedures:
  1. Assign the same value of SEX as that of the previous person with the same OCC.

Minimum Donor Match: OCC.

Flag Value: 4

SIZEPL:
Size of Place (H)
Flag Name: QCITY

See CITY: CITY, CITYPOP, SIZEPL and URBAN are allocated from the same donor.

Flag Value: 4 ????

STATEICP / STATEFIP
State (ICPSR code) (H)
State (FIPS code) (H)
Flag Name: QSTATE

Universe of Donors: All persons with an unallocated STATEICP/STATEFIP code.

Procedures:
  1. Assign the same value for STATEICP/STATEFIPas that of the previous person in the file.

Flag Value: 4

SURSIM:
Surname similarity (P)
Flag Name: QSURSIM

Predictor Variables: None.

Universe of Donors: All persons in households or related groups with an unallocated SURSIM entry.

Procedures:
  1. If RELATE is missing, and AGE is under 15 or SEX is female, assign the same value of SURSIM as the previous person in the household. Everyone else (e.g., adult men) gets new SURSIM code.

Flag Value: 3

URBAN:
Urban/rural status (H)
Flag Name: QCITY

See QCITY: CITY, CITYPOP, SIZEPL and URBAN are allocated from the same donor.

Flag Value: 4 ????

WKSUNEMP:
Weeks of unemployment last year (P)
Flag Name: QQTRUNEM

NOTE: Only cases with illegible values are allocated; blanks are not allocated.

Predictor Variables:
  • OCC1950 - Occupation in 1950 coding system
  • RACE(white, black, other)
  • SEX

Universe of Donors: Wage workers (and others who responded to the weeks unemployed question) with an unallocated WKSUNEMP entry.

Procedures:
  1. Assign same value for WKSUNEMP as the previous person with the same OCC1950, RACE and SEX.
  2. If no donor is found in step 1, drop RACEand look for a match.
  3. If no donor is found in step 2, drop RACEand SEX and look for a match.

Minimum Donor Match: OCC1950

Flag Value: 4

YRIMMIG:
Year of immigration (P)
Flag Name: QYRIMM

Logical Rule:
  1. Calculate YRIMMIG by subtracting the census year from YRSUSA1.

Flag Value: 3

YRNATUR:
Year naturalized (P)
Flag Name: QYRNATUR

Predictor Variables:
  • YRSUSA1 - Years in the US (given for 1900; calculated from YRIMMIG for other census years)
  • ETHNIC - Ethnicity based on birthplace (US outlying or Canada; Latin America, including Mexico; Northern Europe, including England, Scotland and Wales; Ireland; Other Western Europe; Southern Europe; Germany; other Eastern and Central Europe; Asia; other. If Hispanic in the Hispanic Oversample, include with Latin America)

Universe of Donors: Foreign-born citizens with an unallocated YRNATUR entry

Procedures:
  1. Assign the same value for YRNATUR as the previous person with the same YRSUSA1 and ETHNIC.
  2. If no donor is found in step 1, drop ETHNICand look for a match.

Minimum Donor Match: YRSUSA1.

Flag Value: 4

YRSUSA1:
Years in the United States (P)
Flag Name: QYRIMM

Internal Edit Rules
  1. If YRSUSA1 (1910) is missing but there is a value for YRIMMIG, calculate YRSUSA1 from YRIMMIG.
  2. If a person was married before arrival (i.e., if DURMARR < YRSUSA1) and the spouse's record has no value for YRSUSA1, assume that the spouse arrived at the same time and fill in the missing YRSUSA1.
  3. If a woman immigrated and any of her children still living with her have missing values for YRSUSA1, assume they arrived with the mother and fill in the children's missing YRSUSA1.
  4. If there is no mother in the household but there is a man who immigrated and any of his children are still living with him have missing values for YRSUSA1, assume they arrived with the father and fill in the children's missing YRSUSA1.
  5. If YRSUSA1 is more than one year older than AGE, change YRSUSA1 to "missing" for subsequent allocation.

Flag Value: 3 unless marked for allocation.

Variables Used:
  • YRSUSA1 - Years in the US (given for 1900; calculated from YRIMMIG for other census years)
  • YRSUSMIN- Minimum possible years in USA (age of oldest native-born biological child born in the USA; if no native-born children, age of youngest foreign-born child; if no children, 0; if subject has native-born spouse and DURMARR < YRSUSA1, YRSUSMIN equals DURMARR; if citizen, set YRSUSMINequals to 6)
  • YRSUSMAX - Maximum possible years in USA (age of youngest foreign-born biological child; if YRSUSMAX< YRSUSMIN, make YRSUSMAX the same as YRSUSMIN)
  • RANGE - YRSUSMAX minus YRSUSMIN(1,2,3...9, 10-11, 12-13...18-19, 20-24...40-44, 45+)
  • SEX
  • AGE (10-19, 20-29, 30-39, 40-49, 50-59, 60+)
  • ETHNIC- Ethnicity based on birthplace (US outlying or Canada; Latin America, including Mexico; Northern Europe, including England, Scotland and Wales; Ireland; Other Western Europe; Southern Europe; Germany; other Eastern and Central Europe; Asia; other. If Hispanic in the Hispanic Oversample, include with Latin America)

Universe of Donors: Foreign-born persons with an unallocated YRSUSA1 entry.

Procedures:
  1. Predict DIF (number of years difference between immigration and YRSUSMIN):
  2. Assign same value for DIFas previous person with the same SEX, RANGE, AGE and ETHNIC.

    1. If no donor is found in step 1a, drop ETHNICand look for a match.
    2. If no donor is found in step 1b, drop ETHNICand AGE and look for a match.
    3. If no donor is found in step 1c, drop ETHNIC, AGE and SEX and look for a match.
  3. Calculate YRSUSA1 by adding DIFto YRSUSMIN.
  4. If calculated YRSUSA1 > YRSUSMAXK, change YRSUSA1 to YRSUSMAX.
  5. If calculated YRSUSA1 < YRSUSMIN, change YRSUSA1 to YRSUSMIN.
  6. If calculated YRSUSA1 is one year older than AGE, change YRSUSA1 to AGE.

Minimum Donor Match: RANGE.

Flag Value: 4

Go Back to Editing and Allocation Procedures Index