Survey design

Filter results by

Search Help
Currently selected filters that can be removed

Keyword(s)

Type

1 facets displayed. 0 facets selected.
Sort Help
entries

Results

All (4)

All (4) ((4 results))

  • Articles and reports: 12-001-X198800114599
    Description:

    As part of the planning for the 1990 Decennial Census, the Census Bureau investigated the feasibility of adjusting the census for the estimated undercount. A test census was conducted in Central Los Angeles County, in a mostly Hispanic area, in order to test the timing and operational aspects of adjusting the Census using a post-enumeration survey (PES). This paper presents the methodology and the results in producing a census that is adjusted for the population missed by the enumeration. The methodology used to adjust the test census included the sample design, dual-system estimation and small area estimation. The sample design used a block sample with blocks stratified by race/ethnicity. Matching was done by the computer with clerical review and resolution. The dual-system estimator, also called the Petersen estimator or capture-recapture, was used to estimate the population. Because of the nature of the census enumeration, corrections were made to the census counts before using them in the dual-system estimator. Before adjusting the small areas, a regression model was fit to the adjustment factor (the dual-system estimate divided by the census count) to reduce the effects of sampling variability. A synthetic estimator was used to carry the adjustment down to the block level. The results of the dual-system estimates are presented for the test site by the three major race/ethnic groups (Hispanic, Asian, Other) by tenure, by age and by sex. Summaries of the small area adjustments of the census enumeration, by block, are presented and discussed.

    Release date: 1988-06-15

  • Articles and reports: 12-001-X198800114601
    Description:

    The Survey of Income and Program Participation (SIPP) is an ongoing nationally representative household survey program of the Bureau of the Census. The primary purpose of the SIPP is to improve the measurement of information related to the economic situation of households and persons in the United States. It accomplishes this goal through repeated interviews of sample individuals using a short reference period and a probing questionnaire. The multi-interview design of the SIPP raises methodological and statistical issues of concern to all panel surveys of families and persons. This paper reviews these issues as they relate to the SIPP. The topics reviewed are: 1) questionnaire design; 2) data collection, including respondent rules, data collection mode, length of reference period, and rules for following movers; 3) concepts, design, and estimation; and 4) response error.

    Release date: 1988-06-15

  • Articles and reports: 12-001-X198800114602
    Description:

    For a given level of precision, Hidiroglou (1986) provided an algorithm for dividing the population into a take-all stratum and a take-some stratum so as to minimize the overall sample size assuming simple random sampling without replacement in the take-some stratum. Sethi (1963) provided an algorithm for optimum stratification of the population into a number of take-some strata. For the stratification of a highly skewed population, this article presents an iterative algorithm which has as objective the determination of stratification boundaries which split the population into a take-all stratum and a number of take-some strata. These boundaries are computed so as to minimize the resulting sample size given a level of relative precision, simple random sampling without replacement from the take-some strata and use of a power allocation among the take-some strata. The resulting algorithm is a combination of the procedures of Hidiroglou (1986) and Sethi (1963).

    Release date: 1988-06-15

  • Articles and reports: 12-001-X198800114603
    Description:

    Most surveys have many purposes and a hierarchy of six levels is proposed here. Yet most theory and textbooks are based on unipurpose theory, in order to avoid the complexity and conflicts of multipurpose designs. Ten areas of conflict between purposes are shown, then problems and solutions are advanced for each. Compromises and joint solutions fortunately are feasible, because most optima are very flat; also because most “requirements” for precision are actually very flexible. To state and to face the many purposes are preferable to the common practice of hiding behind some artificially picked single purpose; and they have also become more feasible with modern computers.

    Release date: 1988-06-15
Data (0)

Data (0) (0 results)

No content available at this time.

Analysis (4)

Analysis (4) ((4 results))

  • Articles and reports: 12-001-X198800114599
    Description:

    As part of the planning for the 1990 Decennial Census, the Census Bureau investigated the feasibility of adjusting the census for the estimated undercount. A test census was conducted in Central Los Angeles County, in a mostly Hispanic area, in order to test the timing and operational aspects of adjusting the Census using a post-enumeration survey (PES). This paper presents the methodology and the results in producing a census that is adjusted for the population missed by the enumeration. The methodology used to adjust the test census included the sample design, dual-system estimation and small area estimation. The sample design used a block sample with blocks stratified by race/ethnicity. Matching was done by the computer with clerical review and resolution. The dual-system estimator, also called the Petersen estimator or capture-recapture, was used to estimate the population. Because of the nature of the census enumeration, corrections were made to the census counts before using them in the dual-system estimator. Before adjusting the small areas, a regression model was fit to the adjustment factor (the dual-system estimate divided by the census count) to reduce the effects of sampling variability. A synthetic estimator was used to carry the adjustment down to the block level. The results of the dual-system estimates are presented for the test site by the three major race/ethnic groups (Hispanic, Asian, Other) by tenure, by age and by sex. Summaries of the small area adjustments of the census enumeration, by block, are presented and discussed.

    Release date: 1988-06-15

  • Articles and reports: 12-001-X198800114601
    Description:

    The Survey of Income and Program Participation (SIPP) is an ongoing nationally representative household survey program of the Bureau of the Census. The primary purpose of the SIPP is to improve the measurement of information related to the economic situation of households and persons in the United States. It accomplishes this goal through repeated interviews of sample individuals using a short reference period and a probing questionnaire. The multi-interview design of the SIPP raises methodological and statistical issues of concern to all panel surveys of families and persons. This paper reviews these issues as they relate to the SIPP. The topics reviewed are: 1) questionnaire design; 2) data collection, including respondent rules, data collection mode, length of reference period, and rules for following movers; 3) concepts, design, and estimation; and 4) response error.

    Release date: 1988-06-15

  • Articles and reports: 12-001-X198800114602
    Description:

    For a given level of precision, Hidiroglou (1986) provided an algorithm for dividing the population into a take-all stratum and a take-some stratum so as to minimize the overall sample size assuming simple random sampling without replacement in the take-some stratum. Sethi (1963) provided an algorithm for optimum stratification of the population into a number of take-some strata. For the stratification of a highly skewed population, this article presents an iterative algorithm which has as objective the determination of stratification boundaries which split the population into a take-all stratum and a number of take-some strata. These boundaries are computed so as to minimize the resulting sample size given a level of relative precision, simple random sampling without replacement from the take-some strata and use of a power allocation among the take-some strata. The resulting algorithm is a combination of the procedures of Hidiroglou (1986) and Sethi (1963).

    Release date: 1988-06-15

  • Articles and reports: 12-001-X198800114603
    Description:

    Most surveys have many purposes and a hierarchy of six levels is proposed here. Yet most theory and textbooks are based on unipurpose theory, in order to avoid the complexity and conflicts of multipurpose designs. Ten areas of conflict between purposes are shown, then problems and solutions are advanced for each. Compromises and joint solutions fortunately are feasible, because most optima are very flat; also because most “requirements” for precision are actually very flexible. To state and to face the many purposes are preferable to the common practice of hiding behind some artificially picked single purpose; and they have also become more feasible with modern computers.

    Release date: 1988-06-15
Reference (0)

Reference (0) (0 results)

No content available at this time.

Date modified: