Labor Force Statistics
The SIPP universe is the noninstitutionalized resident population living in the United States. Field representatives interview eligible persons who are at least 15 years of age at the time of the interview. Not eligible to be in the survey are crew members of merchant vessels, Armed Forces personnel living in military barracks, institutionalized persons, such as correctional facility inmates and nursing home residents, and United States citizens residing abroad.
The SIPP sample for the 1991 panel is located in 230 Primary Sampling Units (PSUs) each consisting of a county or a group of contiguous counties.
For the 1991 panel, interviewing began in February, March, April, or May of 1991 for four random subsamples, respectively. For the remainder of the panel, interviews for each person occurred every 4 months for a total of 8 interviews. (One round of interviewing all 4 subsamples is called a wave.) At each inter- view, the reference period was the 4 months preceding the interview month.
Occupants of about 93 percent of all eligible living quarters participated in the first interview of the panel. For later interviews, field representatives interviewed only original sample persons and persons living with them. We followed respondents who moved during the panel. The Census Bureau automati- cally designated noninterviewed households at the first wave as noninterviews for subsequent waves.1
We classified a person as interviewed for the entire panel and both calendar years based on the following two definitions: 2
(1) Those for whom self, proxy, or imputed responses were obtained for each reference month of all eight interviews for the 1991 panel, and all three interviews for each calendar year; or
(2) Those for whom self or proxy responses were obtained for the first reference month of the interview period and responses exist for each subsequent month until they were known to have died or moved to an ineligible address (foreign living quarters, institutions, or military barracks).
Everyone else is considered noninterview.3
Some estimates are based on monthly averages from cross-sectional files. Nonresponse rates for the months on the file vary from 8 percent to 21 percent. Some respondents did not respond to some of the questions. Therefore, the overall nonresponse rate for some items, especially sensitive income and money related items, is higher than the person nonresponse rate.4Estimation
We used several stages of weight adjustments in the estimation procedure to derive the SIPP longitudinal person weights. We gave each person a base weight equal to the inverse of his/her probability of selection and applied adjustments to account for noninterviews.5
We performed an additional stage of adjustment to longitudinal person weights to reduce the mean square error of the survey estimates by age, sex, race, and ethnicity (Hispanic/non-Hispanic).1Accuracy of Estimates
We base SIPP estimates on a sample. The sample estimates may differ somewhat from the values obtained from administering a complete census using the same questionnaire, instructions, and enumerators. The difference occurs because a sample survey estimate is subject to two types of errors: nonsampling and sampling. We can provide estimates of the magnitude of the SIPP sampling error, but this is not true of nonsampling error. The next few sections describe SIPP nonsampling error sources, followed by a discussion of sampling error, its estimation, and its use in data analysis.
Nonsampling Variability. We attribute nonsampling errors to many sources; they include but are not limited to the following:
We used quality control and edit procedures to reduce errors made by respondents, coders, and interviewers.4
Undercoverage in SIPP resulted from missed living quarters and missed persons within sample households. It is known that undercoverage varies with age, race, and sex. Generally, undercoverage is larger for males than for females and larger for Blacks than for non-Blacks. Ratio estimation to independent age- race-sex population controls partially corrects for the bias resulting from survey undercoverage. However, biases exist in the estimates when persons in missed households or missed persons in interviewed households have character- istics different from those of interviewed persons in the same age-race-sex group. Further, we did not adjust the independent population controls for under- coverage in the census.6
Comparability with Other Estimates. Exercise caution when comparing data from this report with data from other SIPP publications or with data from other surveys. Comparability problems are from varying seasonal patterns for many characteristics, different nonsampling errors, and different concepts and procedures.4
Sampling Variability. Standard errors indicate the magnitude of the sampling error. They also partially measure the effect of some nonsampling errors in response and enumeration, but do not measure any systematic biases in the data. The standard errors mostly measure the variations that occurred by chance because we surveyed a sample rather than the entire population.Uses and Computation of Standard Errors
Confidence Intervals. The sample estimate and its standard error enable one to construct confidence intervals, ranges that would include the average result of alpossible samples with a known probability.
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.
The average estimate derived from all possible samples is or is not contained in any particular computed interval. However, for a particular sample, one can say with a specified confidence that the confidence interval includes the average estimate derived from all possible samples.
Hypothesis Testing. One may also use standard errors for hypothesis testing. Hypothesis testing is a procedure for distinguishing between population char- acteristics using sample estimates. The most common type of hypothesis tested is (1) the population characteristics are identical versus (2) they are different. One can perform tests at various levels of significance, where a level of sig- nificance is the probability of concluding that the characteristics are different when, in fact, they are identical.
Unless noted otherwise, all statements of comparison in the report passed a hypothesis test at the 0.10 level of significance or better. This means that, for differences cited in the report, the estimated absolute difference between parameters is greater than 1.645 times the standard error of the difference.
Note that as we perform more tests, more erroneous significant differences will occur. For example, at the 10-percent significance level, if we perform 100 independent hypothesis tests in which there are no real differences, it is likely that about 10 erroneous differences will occur. Therefore, interpret the significance of any single test cautiously.
Standard Error Parameters and Tables and Their Use.
Most SIPP estimates have greater standard errors than those obtained through a simple random sample because we sampled clusters of living quarters for the SIPP. To derive standard errors at a moderate cost and applicable to a wide variety of estimates, we made a number of approximations. We grouped estimates with similar standard error behavior and developed two parameters (denoted "a" and "b") to approximate the standard error behavior of each group of estimates. The standard errors we computed from these parameters provide an indication of the order of magnitude of the standard error for any specific estimate.
Methods for using these parameters and tables for computation of standard errors are given in the following sections. To calculate standard errors for estimates of persons ever participating or persons participating all of two years, use a = -0.0000483 and b = 8,912. The bases for percentages are found in appropriate text tables.
Standard Errors of Estimated Numbers. Approximate s using the formula,
s = SQRT(ax2 + bx).
Here x is the size of the estimate.
Illustration. As shown in text table E, the 1991 SIPP estimates approximately 1.7 million labor turnover actions occurred in the retail trade industry in an average month during 1991. The appropriate "a" and "b" parameters are a = -0.0000483 and b = 8,912.
Using the above formula, the approximate standard error is
s = SQRT((-0.0000483)(1,737,0000)2 + (8,912)(1,737,000)) = 123,832
The 90-percent confidence interval is from 1,533,296 to 1,940,704. Therefore, 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 samples.
Standard Errors of Estimated Percentages. The reliability of an estimated percentage, computed using sample data for both numerator and denominator, depends on the size of the percentage and its base. Approximate the standard error by the formula:
S = SQRT(b/x (p)(100-p)
Here x is the total number of persons in the base of the percentage and p is the percentage (0 <= p >= 100). Illustration. As shown in text table F, the 1991 SIPP estimates that the average monthly labor turnover rate for men age 25 to 54, was 4.9% in 1991. To find the base for the percentage, use text table F. In this example, the base is 39,892,000. The appropriate "b" parameter is b = 8,912.
Using the above formula, the approximate standard error is
S = SQRT((8,912/39,892,000)(4.9)(100-4.9)) = 0.32 percent
The 90-percent confidence interval is from 4.4 to 5.4 percent. Therefore, a conclusion that the average percentage derived from all possible samples lies within a range computed in this way would be correct for roughly 90 percent of all samples.
1 Details on nonresponse and Hispanic controls are in "SIPP 91: Source and Accuracy Statement for the Longitudinal Panel File REVISION," dated October 19, 1994
2 Details on interview-status classification are in "Weighting of Persons for SIPP Longitudinal Tabulations," paper by Judkins, Hubble, Dorsch, McMillen and Ernst in the 1984 Proceedings of the Survey Research Methods Section, American Statistical Association.
3 Details on patterns of nonresponse are in "Weighting Adjustment for Partial Nonresponse in the 1984 SIPP Panel.." paper by Lepkowski, Kalton, and Kasprzyk in the 1989 Proceeding of the Survey Research Methods Section, American Statistical Association.
4 For more discussion on nonresponse and the existence and control of nonsampling errors in the SIPP, see the Quality Profile for the Survey of Income and Program Participation, May 1990, by T. Jabine, K. King and R. Pertoni. Available from Customer Services, Data User Services Division (301-763-1139)
5 For more details on noninterview adjustment for longitudinal estimates, see Nonresponse Adjustment Methods for Demographic Surveys at the U.S. Bureau of the Census, November 1988, Working Paper 8823, by R. Singh and R. Petroni.
6 More detailed discussions of the population controls are in the SIPP Dynamics of Economic Well-Being: Labor Force and Income, 1990 to 1992, Report P70-40, by Wilfred Masumura and Paul Ryscavage.