Reliability of estimates

Measures of sampling error

Survey of Occupational Injuries and Illnesses (SOII)

Estimates from the Survey of Occupational Injuries and Illnesses (SOII) are based on a scientifically-selected probability sample, rather than a census of the entire population of establishments and cases. This sampling methodology makes it possible to collect data from a sample from which inferences can be made regarding the characteristics of the population from which the sample was selected. The sample used in any given year for the SOII was one of many possible samples, each of which could have produced different estimates. While these sample-based estimates may differ from the results obtained from a census of the population, on average, they will reflect the same inferences one would gain from a census. The variation in the sample estimates across all possible samples that could have been drawn is measured via the sampling error, which is presented for the SOII as the percent relative standard error (RSE).

Fatal Occupational Injury counts (CFOI)

Fatal Occupational Injury counts are based on a census and, as such, do not have any sampling error.

Fatal Occupational Injury rates (CFOI)

Fatal injury rates estimates rely on fatal injury counts from the CFOI and employment estimates from the Current Population Survey (CPS). CPS is a survey of households that is designed to represent the civilian noninstitutional population of the United States. Like the SOII (outlined above), sampling errors occur in the CPS because observations are made on a sample, not on the entire population. The impact of the CPS sampling error on the fatal injury rates can be measured by incorporating the RSE of the CPS estimates into an RSE computed for the fatal injury rates.

Confidence Intervals

The sampling error measured for either the SOII or the CFOI fatal injury rates can then be used to calculate a "confidence interval" around the sample-based estimate.

The 95-percent confidence interval is the interval centered on the sample-based estimate and includes all values within 1.96 times the estimate's standard error. If several different samples were selected to estimate the population value (such as an injury and illness incidence rate), the 95-percent confidence interval would mean that one would be 95-percent certain that the range of these sample-based estimates would include the true population value.

To calculate the 95-percent confidence interval given the percent RSE from a SOII estimate (or fatal injury rate):

  1. Divide the percent RSE by 100 and multiply the result by the SOII estimate (or fatal injury rate) to determine the standard error.
  2. Multiply the standard error by 1.96 to determine the margin of error (MOE).
  3. The SOII estimate (or fatal injury rate) plus or minus the MOE is the 95-percent confidence interval.

Examples:

The incidence rate of total recordable cases (TRC) of nonfatal injuries and illnesses of 6.9 cases per 100 full-time equivalent (FTE) workers for the nursing care facilities industry (NAICS 6231) in 2015 had an estimated relative standard error of 1.3 percent. Hence, we are 95-percent confident that the true TRC rate for the nursing care facilities industry lies in the interval between 6.7 and 7.1 (or 6.9 ± (1.96 x 6.9 x 0.013)).

If the fatal injury rate for an occupation is 15.6 per 100,000 FTE workers in 2015, and the margin of error reported is 0.4, we are 95-percent confident that the true fatal injury rate for this occupation lies in the interval between 15.2 and 16.0 (or 15.6 ± (0.4)).

Non-sampling error

Non-sampling errors will always occur when data are gathered. The inability to obtain information about all cases in the sample, mistakes in recording or coding the data, and definitional difficulties are general examples of potential causes of non-sampling errors in the survey. The Bureau has implemented quality assurance procedures to reduce non-sampling errors in the survey, including a rigorous training program for State coders, mechanical edits that identify questionable entries and coding inconsistencies, and a continuing effort to encourage survey participants to respond fully and accurately to all survey elements. See Data Quality Research for additional information regarding ongoing research to identify other potential sources of non-sampling errors affecting SOII/CFOI data.

For more information on SOII estimates please see the handbook of methods here: https://www.bls.gov/opub/hom/soii/calculation.htm#reliability-of-soii-estimates. For more information on the fatal injury rates, methodology, and caveats, please see the handbook of methods section here: https://www.bls.gov/opub/hom/cfoi/calculation.htm#fatal-injury-rate-methodology-and-calculation.

 

Last Modified Date: December 7, 2017