Attachment H - Source and Accuracy

ATTACHMENT 16
Source of the Data and Accuracy of the Estimates for the
August 2005 CPS Microdata File on Veterans

SOURCE OF DATA
The data in this microdata file are from the August 2005 Current Population Survey (CPS). The Census
Bureau conducts the CPS every month, although this file has only August data. The August survey uses
two sets of questions, the basic CPS and a set of supplemental questions. The CPS, sponsored jointly by
the Census Bureau and the U.S. Bureau of Labor Statistics, is the country’s primary source of labor force
statistics for the entire population. The Department of Veterans Affairs and the Department of Labor
jointly sponsor the supplemental questions for August.
Basic CPS. The monthly CPS collects primarily labor force data about the civilian noninstitutional
population living in the United States. The institutionalized population, which is excluded from the
population universe, is composed primarily of the population in correctional institutions and nursing
homes (91 percent of the 4.1 million institutionalized people in Census 2000). Interviewers ask questions
concerning labor force participation about each member 15 years old and over in sample households.
Typically, the week containing the 19th of the month is the interview week. The week containing the 12th
is the reference week (i.e., the week about which the labor force questions are asked).
The CPS uses a multistage probability sample based on the results of the decennial census, with coverage
in all 50 states and the District of Columbia. The sample is continually updated to account for new
residential construction. When files from the most recent decennial census become available, the Census
Bureau gradually introduces a new sample design for the CPS1.
In April 2004, the Census Bureau began phasing out the 1990 sample and replacing it with the 2000
sample, creating a mixed sampling frame. Two simultaneous changes occurred during this phase-in
period. First, primary sampling units (PSUs)2 selected for only the 2000 design gradually replaced those
selected for the 1990 design. This involved 10 percent of the sample. Second, within PSUs selected for
both the 1990 and 2000 designs, sample households from the 2000 design gradually replaced sample
households from the 1990 design. This involved about 90 percent of the sample. The new sample design
was completely implemented by July 2005.
In the first stage of the sampling process, PSUs are selected for sample. The United States is divided into
2,025 PSUs. The PSUs were redefined for this design to correspond to the Office of Management and
Budget definitions of Core-Based Statistical Area definitions and to improve efficiency in field
operations. These PSUs are grouped into 824 strata. Within each stratum, a single PSU is chosen for the
sample, with its probability of selection proportional to its population as of the most recent decennial
census. This PSU represents the entire stratum from which it was selected. In the case of strata
consisting of only one PSU, the PSU is chosen with certainty.
Approximately 71,000 housing units were selected for sample from the mixed sampling frame in August.
Based on eligibility criteria, 11 percent of these housing units were sent directly to Computer-Assisted
Telephone Interviewing (CATI). The remaining units were assigned to interviewers for Computer1
2

For detailed information on the 1990 sample redesign, please see reference [1].
The PSUs correspond to substate areas, counties or groups of counties that are geographically contiguous.
16-1

Assisted Personal Interviewing (CAPI).3 Of all housing units in sample, about 60,000 were determined to
be eligible for interview. Interviewers obtained interviews at about 55,000 of these units. Noninterviews
occur when the occupants are not found at home after repeated calls or are unavailable for some other
reason.
August 2005 Supplement. In August 2005, in addition to the basic CPS questions, interviewers asked
supplementary questions of veterans concerning year of discharge and disability.
Estimation Procedure. This survey’s estimation procedure adjusts weighted sample results to agree with
independently derived population estimates of the civilian noninstitutional population of the United States
and each state (including the District of Columbia). These population estimates, used as controls for the
CPS, are prepared monthly to agree with the most current set of population estimates that are released as
part of the Census Bureau’s population estimates and projections program.
The population controls for the nation are distributed by demographic characteristics in two ways:
•
•

Age, sex, and race (White alone, Black alone, and all other groups combined), and
Age, sex, and Hispanic origin.

The population controls for the states are distributed by race (Black alone and all other race groups
combined), age (0-15, 16-44, and 45 and over), and sex.
The independent estimates by age, sex, race, and Hispanic origin and for states by selected age groups and
broad race categories are developed using the basic demographic accounting formula whereby the
population from the latest decennial data is updated using data on the components of population change
(births, deaths, and net international migration) with net internal migration as an additional component in
the state population estimates.
The net international migration component in the population estimates includes a combination of:
•
•
•
•
•

Legal migration to the United States,
Emigration of foreign born and native people from the United States,
Net movement between the United States and Puerto Rico,
Estimates of temporary migration, and
Estimates of net residual foreign-born population, which include unauthorized migration.

Because the latest available information on these components lags the survey date, it is necessary to make
short-term projections of these components to develop the estimate for the survey date.
ACCURACY OF THE ESTIMATES
A sample survey estimate has two types of error: sampling and nonsampling. The accuracy of an estimate
depends on both types of error. The nature of the sampling error is known given the survey design; the
full extent of the nonsampling error is unknown.
Sampling Error. Since the CPS estimates come from a sample, they may differ from figures from an
enumeration of the entire population using the same questionnaires, instructions, and enumerators. For a
3

For further information on CATI and CAPI and the eligibility criteria, please see reference [2].
16-2

given estimator, the difference between an estimate based on a sample and the estimate that would result
if the sample were to include the entire population is known as sampling error. Standard errors, as
calculated by methods described in “Standard Errors and Their Use,” are primarily measures of the
magnitude of sampling error. However, they may include some nonsampling error.
Nonsampling Error. For a given estimator, the difference between the estimate that would result if the
sample were to include the entire population and the true population value being estimated is known as
nonsampling error. Sources of nonsampling errors include the following:
•
•
•
•
•
•
•
•
•

Inability to get information about all sample cases (nonresponse).
Definitional difficulties.
Differences in the interpretation of questions.
Respondent inability or unwillingness to provide correct information.
Respondent inability to recall information.
Errors made in data collection, such as recording and coding data.
Errors made in processing the data.
Errors made in estimating values for missing data.
Failure to represent all units with the sample (undercoverage).

The Census Bureau employs quality control procedures throughout the production process including the
overall design of surveys, the wording of questions, the review of the work of interviewers and coders,
and the statistical review of reports to minimize these errors.
Two types of nonsampling error that can be examined to a limited extent are nonresponse and
undercoverage.
Nonresponse. The effect of nonresponse cannot be measured directly, but one indication of its potential
effect is the nonresponse rate. For the August 2005 basic CPS, the nonresponse rate was 8.0 percent. The
person-level nonresponse rate for the Veterans supplement was an additional 10.8 percent.
Since the basic CPS nonresponse rate is a household-level rate and the Veterans supplement nonresponse
rate is a person-level rate, we cannot combine these rates to derive an overall nonresponse rate. Since it is
unlikely the households missed by the basic CPS have the same number of persons as the households
successfully interviewed, combining these rates would result in an overestimate of the “true” person-level
overall nonresponse rate for the Veterans supplement.
Coverage. The concept of coverage in the survey sampling process is the extent to which the total
population that could be selected for sample “covers” the survey’s target population. Missed housing
units and missed people within sample households create undercoverage in the CPS. Overall CPS
undercoverage for August 2005 is estimated to be about 9 percent. CPS coverage varies with age, sex,
and race. Generally, coverage is larger for females than for males and larger for non-Blacks than for
Blacks. This differential coverage is a general problem for most household-based surveys.
The CPS weighting procedure partially corrects for bias from undercoverage, but biases may still be
present when people who are missed by the survey differ from those interviewed in ways other than age,
race, sex, Hispanic ancestry, and state of residence. How this weighting procedure affects other variables
in the survey is not precisely known. All of these considerations affect comparisons across different
surveys or data sources.
16-3

A common measure of survey coverage is the coverage ratio, calculated as the estimated population
before post-stratification divided by the independent population control. Table 1 shows August 2005 CPS
coverage ratios by age and sex for certain race and Hispanic groups. The CPS coverage ratios can exhibit
some variability from month to month.
Table 1. CPS Coverage Ratios {tc "CPS Coverage Ratios " \f D }: August 2005
Totals

White Only

Black Only

Residual Race

Hispanic

All
Age
Male Female Male Female Male Female Male Female Male Female
Group People
0-15
0.91
0.91
0.92
0.92
0.94
0.86
0.79
0.95
0.99
0.94
0.97
16-19 0.91
0.91
0.92
0.92
0.93
0.83
0.80
0.94
1.05
0.95
0.97
20-24 0.81
0.79
0.83
0.82
0.84
0.62
0.76
0.76
0.82
0.81
0.89
25-34 0.85
0.82
0.88
0.85
0.89
0.69
0.82
0.78
0.86
0.80
0.89
35-44 0.92
0.89
0.95
0.91
0.96
0.76
0.87
0.89
0.92
0.86
0.99
45-54 0.92
0.90
0.94
0.92
0.95
0.74
0.86
0.93
0.96
0.89
0.96
55-64 0.93
0.92
0.94
0.93
0.96
0.83
0.84
0.99
0.91
0.92
0.97
65+
0.95
0.96
0.95
0.96
0.95
0.98
0.92
0.84
0.89
0.86
0.92
15+
0.90
0.89
0.92
0.90
0.94
0.77
0.84
0.87
0.91
0.86
0.94
0+
0.91
0.89
0.92
0.91
0.93
0.79
0.83
0.89
0.93
0.88
0.95
Notes: (1)
(2)

The Residual Race group includes cases indicating a single race other than White or Black,
and cases indicating two or more races.
Hispanics may be of any race. For a more detailed discussion on the use of parameters for race and
ethnicity, please see the “Generalized Variance Parameters” section.

Comparability of Data. Data obtained from the CPS and other sources are not entirely comparable.
This results from differences in interviewer training and experience and in differing survey processes.
This is an example of nonsampling variability not reflected in the standard errors. Therefore, caution
should be used when comparing results from different sources.
Data users should be careful when comparing the data from this microdata file, which reflects Census
2000-based controls, with microdata files from March 1994 through December 2001, which reflect 1990
census-based controls. Ideally, the same controls should be used when comparing any estimates. In
reality, the use of same population controls is not practical when comparing trend data over a period of 10
to 20 years. Thus, when it is necessary to combine or compare data which are based on different controls
and/or different designs, users should be cognizant that changes in weighting controls and/or weighting
procedures can create small differences between estimates. See the discussion below for information on
comparing estimates derived from different sample designs.
Microdata files from previous years reflect the latest available census-based controls. Although this
change in population controls had relatively little impact on summary measures such as averages,
medians, and percentage distributions, it did have a significant impact on levels. For example, use of
Census 2000-based controls results in about a one percent increase from the 1990 census-based controls
in the civilian noninstitutional population and in the number of families and households. Thus, estimates
of levels for data collected in 2002 and later years will differ from those for earlier years by more than
what could be attributed to actual changes in the population. These differences could be
disproportionately greater for certain population subgroups than for the total population.

16-4

Note that certain microdata files from 2002, namely June, October, and November, contain both Census
2000-based estimates and 1990 census-based estimates and are subject to the comparability issues
discussed above. All other microdata files from 2002 reflect the 1990 census-based controls.
Users should also exercise caution because of changes caused by the phase-in of the Census 2000 files.
During this time period, CPS data are collected from sample designs based on different censuses. Three
features of the new CPS design have the potential of affecting published estimates: (1) the temporary
disruption of the rotation pattern from August 2004 through June 2005 for a comparatively small portion
of the sample, (2) the change in sample areas, and (3) the introduction of the new Core-Based Statistical
Areas (formerly called metropolitan areas). Most of the known effect on estimates during and after the
sample redesign will be the result of changing from 1990 to 2000 geographic definitions. Research has
shown that the national-level estimates of the metropolitan and nonmetropolitan populations should not
change appreciably because of the new sample design. However, users should still exercise caution when
comparing metropolitan and nonmetropolitan estimates across years with a design change, especially at
the state level.
Caution should also be used when comparing Hispanic estimates over time. No independent population
control totals for people of Hispanic ancestry were used before 1985.
A Nonsampling Error Warning. Since the full extent of the nonsampling error is unknown, one should
be particularly careful when interpreting results based on small differences between estimates. Even a
small amount of nonsampling error can cause a borderline difference to appear significant or not, thus
distorting a seemingly valid hypothesis test. Caution should also be used when interpreting results based
on a relatively small number of cases. Summary measures (such as medians and percentage distributions)
probably do not reveal useful information when computed on a subpopulation smaller than 75,000.
For additional information on nonsampling error including the possible impact on CPS
data when known, refer to references [2] and [3].
Standard Errors and Their Use. The sample estimate and its standard error enable one to construct a
confidence interval. A confidence interval is a range that would include the average result of all possible
samples with a known probability. For example, if all possible samples were surveyed under essentially
the same general conditions and using the same sample design, and if an estimate and its standard error
were calculated from each sample, then approximately 90 percent of the intervals from 1.645 standard
errors below the estimate to 1.645 standard errors above the estimate would include the average result of
all possible samples.
A particular confidence interval may or may not contain the average estimate derived from all possible
samples. However, one can say with specified confidence that the interval includes the average estimate
calculated from all possible samples.
Standard errors may also be used to perform hypothesis testing, a procedure for distinguishing between
population parameters using sample estimates. The most common type of hypothesis is that the
population parameters are different. An example of this would be comparing the percentage of men who
were part-time workers to the percentage of women who were part-time workers.
Tests may be performed at various levels of significance. A significance level is the probability of
concluding that the characteristics are different when, in fact, they are the same. For example, to
conclude that two characteristics are different at the 0.10 level of significance, the absolute value of the
16-5

estimated difference between characteristics must be greater than or equal to 1.645 times the standard
error of the difference.
The Census Bureau uses 90-percent confidence intervals and 0.10 levels of significance to determine
statistical validity. Consult standard statistical textbooks for alternative criteria.
Estimating Standard Errors. The Census Bureau uses replication methods to estimate the standard
errors of CPS estimates. These methods primarily measure the magnitude of sampling error. However,
they do measure some effects of nonsampling error as well. They do not measure systematic biases in the
data associated with nonsampling error. Bias is the average over all possible samples of the differences
between the sample estimates and the true value.
Generalized Variance Parameters. While it is possible to compute and present an estimate of the
standard error based on the survey data for each estimate in a report, there are a number of reasons why
this is not done. A presentation of the individual standard errors would be of limited use, since one could
not possibly predict all of the combinations of results that may be of interest to data users. Additionally,
variance estimates are based on sample data and have variances of their own. Therefore, some methods
of stabilizing these estimates of variance, for example, by generalizing or averaging over time, may be
used to improve their reliability.
Experience has shown that certain groups of estimates have similar relationships between their variances
and expected values. Modeling or generalizing may provide more stable variance estimates by taking
advantage of these similarities. The generalized variance function is a simple model that expresses the
variance as a function of the expected value of the survey estimate. The parameters of the generalized
variance function are estimated using direct replicate variances. These generalized variance parameters
provide a relatively easy method to obtain approximate standard errors for numerous characteristics. In
this source and accuracy statement, Table 3 provides the generalized variance parameters for labor force
estimates, Table 4 provides generalized variance parameters for characteristics from the August 2005
supplement.
The basic CPS questionnaire records the race and ethnicity of each respondent. With respect to race, a
respondent can be White, Black, Asian, American Indian or Alaskan Native (AIAN), Native Hawaiian
and Other Pacific Islander (NHOPI), or combinations of two or more of the preceding. A respondent’s
ethnicity can be Hispanic or non-Hispanic, regardless of race.
The generalized variance parameters to use in computing standard errors are dependent upon the
race/ethnicity group of interest. The following table summarizes the relationship between the
race/ethnicity group of interest and the generalized variance parameters to use in standard error
calculations

16-6

Table 2. Estimation Groups of Interest and Generalized Variance Parameters
Generalized Variance Parameters
To Use in
Standard Error Calculations

Race/Ethnicity
Group of Interest
Total Population

Total or White

Total White, White AOIC, or White Non-Hispanic Population

Total or White

Total Black, Black AOIC, or Black Non-Hispanic Population

Black

Total API, AIAN, NHOPI;
API, AIAN, NHOPI AOIC;
or API, AIAN, NHOPI Non-Hispanic Population

API, AIAN, NHOPI

Populations from Other Race Groups

API, AIAN, NHOPI

Hispanic Population

Hispanic

Notes: (1) API, AIAN, NHOPI are Asian and Pacific Islander, American Indian and
Alaska Native, Native Hawaiian and Other Pacific Islander, respectively.
(2) AOIC is an abbreviation for “Alone or In Combination.” The AOIC population for a race group of
interest includes people reporting only the race group of interest (“alone”) and people reporting
multiple race categories including the race group of interest (“in combination”).
(3) Hispanics may be of any race.

Standard Errors of Estimated Numbers. The approximate standard error, sx, of an estimated number
from this microdata file can be obtained by using the formula:
s x = ax 2 + bx

(1)

Here x is the size of the estimate and a and b are the parameters in Table 3 or 4 associated with the
particular type of characteristic. When calculating standard errors from cross-tabulations involving
different characteristics, use the set of parameters for the characteristic that will give the largest standard
error.
Illustration 1
Suppose there were 1,597,000 unemployed non-veterans, aged 40 – 54, in the civilian labor force. Use
the appropriate parameters from Table 3 and Formula (1) to get
Illustration 1
Number of unemployed nonveterans in the
civilian labor force (x)
a parameter (a)
b parameter (b)
Standard error
90% confidence interval

16-7

1,597,000
-0.000016
3,096
70,000
1,482,000 to 1,712,000

The standard error is calculated as

s x = − 0.000016 ×1,597,000 2 + 3,096 ×1,597,000 = 70,000
The 90-percent confidence interval is calculated as 1,597,000 ± 1.645 × 70,000.
A conclusion that the average estimate derived from all possible samples lies within a range computed in
this way would be correct for roughly 90 percent of all possible samples.
Standard Errors of Estimated Percentages. The reliability of an estimated percentage, computed using
sample data for both numerator and denominator, depends on both the size of the percentage and its base.
Estimated percentages are relatively more reliable than the corresponding estimates of the numerators of
the percentages, particularly if the percentages are 50 percent or more. When the numerator and
denominator of the percentage are in different categories, use the parameter from Table 3 or 4 as indicated
by the numerator.

The approximate standard error, sx,p, of an estimated percentage can be obtained by using the formula:
s x ,p =

b
p(100 − p )
x

(2)

Here x is the total number of people, families, households, or unrelated individuals in the base of the
percentage, p is the percentage (0 # p #100), and b is the parameter in Table 3 or 4 associated with the
characteristic in the numerator of the percentage.
Illustration 2
Suppose there were 1,476,000 Vietnam-era war veterans aged 40 to 54 in the civilian labor force, and 2.8
percent were unemployed. Use the appropriate parameter from Table 4 and Formula (2) to get
Illustration 2
Percentage of Vietnam-ear war veterans, aged
40-54, who are unemployed (p)
Base (x)
b parameter (b)
Standard error
90% confidence interval

2.8
1,476,000
3,096
0.76
1.55 to 4.05

The standard error is calculated as
s x ,p =

3,096
× 2.8 × (100.0 − 2.8) = 0.76
1,476,000

The 90 percent confidence interval for the estimated percentage of unemployed Vietnam-era veterans in
the civilian labor force aged 40 to 54 is from 1.55 to 4.05 percent (i.e., 2.8 ± 1.645 × 0.76).

16-8

Standard Errors of Estimated Differences. The standard error of the difference between two sample
estimates is approximately equal to
sx−y = sx + sy
2

2

(3)

where sx and sy are the standard errors of the estimates, x and y. The estimates can be numbers,
percentages, ratios, etc. This will result in accurate estimates of the standard error of the same
characteristic in two different areas, or for the difference between separate and uncorrelated
characteristics in the same area. However, if there is a high positive (negative) correlation between the
two characteristics, the formula will overestimate (underestimate) the true standard error.
Illustration 3
Suppose that of the 1,476,000 Vietnam-era war veterans in the civilian labor force between 40 - 54 years
of age, 2.8 percent were unemployed, and of the 47,021,000 nonveterans in the civilian labor force
between 40-54 years of age, 3.4 percent were unemployed. Use the appropriate parameters from Table 4
and Formulas (2) and (3) to get
Illustration 3
Vietnam-era
Nonveterans (y)
Veterans (x)

Percentage unemployed,
aged 40 - 54 (p)
Base (x)
b parameter (b)
Standard error
90% confidence interval

Difference

2.8

3.4

0.6

1,476,000
3,096
0.76
1.55 to 4.05

47,021,000
3,096
0.15
3.15 to 3.65

0.77
-0.67 to 1.87

The standard error of the difference is calculated as
s x − y = 0.76 2 + 0.15 2 = 0.77

The 90-percent confidence interval around the difference is calculated as 0.6 ± 1.645 × 0.77. Since this
interval does include zero, we cannot conclude with 90 percent confidence that the percentage of
unemployed Vietnam-era veterans in the civilian labor force between 40 - 54 years of age is less than the
percentage of unemployed nonveterans in the civilian labor force between 40 - 54 years of age.
Standard Errors of Quarterly or Yearly Averages. For information on calculating standard errors for
labor force data from the CPS which involve quarterly or yearly averages, please see the “Explanatory
Notes and Estimates of Error: Household Data” section in Employment and Earnings, a monthly report
published by the U.S. Bureau of Labor Statistics.
Technical Assistance. If you require assistance or additional information, please contact the
Demographic Statistical Methods Division via e-mail at [email protected].

16-9

Table 3. Parameters for Computation of Standard Errors for Labor Force Characteristics:
August 2005
Characteristic

a

b

Civilian Labor Force, Employed
Not in Labor Force
Unemployed

-0.000016
-0.000009
-0.000016

3,068
1,833
3,096

Civilian Labor Force, Employed, Not in Labor Force, and Unemployed
Men
Women
Both sexes, 16 to 19 years

-0.000032
-0.000031
-0.000022

2,971
2,782
3,096

-0.000151
-0.000311
-0.000252
-0.001632

3,455
3,357
3,062
3,455

-0.000141
-0.000253
-0.000266
-0.001528

3,455
3,357
3,062
3,455

-0.000346
-0.000729
-0.000659
-0.004146

3,198
3,198
3,198
3,198

Total or White

Black
Civilian Labor Force, Employed, Not in Labor Force, and Unemployed
Total
Men
Women
Both sexes, 16 to 19 years
Hispanic
Civilian Labor Force, Employed, Not in Labor Force, and Unemployed
Total
Men
Women
Both sexes, 16 to 19 years
API, AIAN, NHOPI
Civilian Labor Force, Employed, Not in Labor Force, and Unemployed
Total
Men
Women
Both sexes, 16 to 19 years

Notes: (1) These parameters are to be applied to basic CPS monthly labor force estimates.
(2) API, AIAN, NHOPI are Asian and Pacific Islander, American Indian and Alaska Native, and Native
Hawaiian and Other Pacific Islander, respectively.
(3) For foreign-born and noncitizen characteristics for Total and White, the a and b parameters should be
multiplied by 1.3. No adjustment is necessary for foreign-born and noncitizen characteristics for
Blacks, Hispanics, and Asians.
(4) Hispanics may be of any race. For a more detailed discussion on the use of parameters for race and
ethnicity, please see the “Generalized Variance Parameters” section.
(5) For nonmetropolitan characteristics, multiply the a and b parameters by 1.5. If the characteristic of
interest is total state population, not subtotaled by race or ancestry, the a and b parameters are zero.

16-10

Table 4. Standard Error Parameters for Estimated Numbers and Percentages
for the CPS August 2005 Veterans Supplement

Characteristic

a

b

Total Employed and Nonagriculture Employed in Labor Force, Occupations, and Disability
Status of Employed

Total or Men
All Veterans
War Veterans
Other Service Veterans
Vietnam Era Veterans

-0.000131
-0.000175
-0.000520
-0.000394

3,068
3,068
3,068
3,068

Women
All Veterans
War Veterans
Other Service Veterans
Vietnam Era Veterans

-0.001863
-0.002772
-0.005687
-0.011988

3,068
3,068
3,068
3,068

Total or Men
All Veterans
War Veterans
Other Service Veterans
Vietnam Era Veterans

-0.000132
-0.000177
-0.000525
-0.000398

3,096
3,096
3,096
3,096

Women
All Veterans
War Veterans
Other Service Veterans
Vietnam Era Veterans

-0.001880
-0.002797
-0.005738
-0.012097

3,096
3,096
3,096
3,096

Unemployed, Duration of Unemployment

Notes: (1) These parameters are to be applied to the August 2005 Veterans Supplement data.
(2) For foreign-born and noncitizen characteristics, the a and b parameters should be multiplied by 1.3.
(3) For nonmetropolitan characteristics, multiply the a and b parameters by 1.5. If the characteristic of
interest is total state population, not subtotaled by race or ancestry, the a and b parameters are zero.

16-11

References

[1]

Bureau of Labor Statistics. 1994. Employment and Earnings. Volume 41 Number 5, May 1994.
Washington, DC: Government Printing Office.

[2]

U.S. Census Bureau. 2002. Current Population Survey: Design and Methodology. Technical
Paper 63RV. Washington, DC: Government Printing Office.
(http://www.census.gov/prod/2002pubs/tp63rv.pdf)

[3]

Brooks, C.A. and Bailar, B.A. 1978. Statistical Policy Working Paper 3 - An Error
Profile: Employment as Measured by the Current Population Survey. Subcommittee on
Nonsampling Errors, Federal Committee on Statistical Methodology, U.S. Department of
Commerce, Washington, DC. (http://www.fcsm.gov/working-papers/spp.html)

16-12