BLS Handbook of Methods

Chapter 9.
Occupational Safety and Health Statistics

Estimation Procedures

Weighting
Sample units. By means of a weighting procedure, sample units are constructed to represent all units in their size class for a particular industry. The weight is determined by the inverse of the sampling ratio for the industry/employment-size class from which the unit was selected. Because a small proportion of survey forms are not returned, weights of responding employers in a sampling cell are adjusted to account for nonrespondents. The respondents are then shifted into the estimating cell determined by the reported employment. Data for each unit are multiplied by the appropriate weight and nonresponse adjustment factor. The products are then aggregated to obtain a total for the estimating cell.

Lost worktime cases. Each case involving days away from work is weighted by the sample unit weight with which it is associated and the industry benchmark (see below) in which the associated sample unit resides. In addition, each case is weighted to adjust for case subsampling and case nonresponse for those establishments which did not provide information on all cases with days away from work which occurred in their establishment in the survey year.

Benchmarking
Because the universe file which provides the sample frame is not current to the reference year of the survey, it is necessary to adjust the data before publication to reflect current employment levels. This procedure is known as benchmarking. In the annual survey, all estimates of totals are adjusted by the benchmark factor at the estimating cell level. The benchmarking procedure requires a source of accurate employment data which can be converted into annual average employment figures for the cell level in which separate estimates are desired. Because industry/employment-size data are required for national estimates, benchmark factors are applied to the size class "blow up" estimates.

Incidence rate calculation
Incidence rates are calculated using the total obtained through the weighting and benchmarking procedures. The adjusted estimates for a particular characteristic, for example injury and illness cases involving days away from work, are aggregated to the appropriate level of industry detail. The total is multiplied by 200,000 (the base of hours worked by 100 full-time employees for 1 year). The product is then divided by the weighted and benchmarked estimate of hours worked as reported in the survey for the industry segment.

The formula for calculating the incidence rate at the lowest level of industry detail is:

Incidence rates for higher levels of industry detail are produced using aggregated weighted and benchmarked totals. Rates may be computed by industry, employment size, geographic area, extent or outcome of case, and case characteristic category. Rates for illnesses and rates for case characteristic categories are published per 10,000 full-time employees, using 20,000,000 hours instead of 200,000 hours in the formula shown above. Rates per 10,000 workers can be converted to rates per 100 workers by moving the decimal point left two places and rounding the resulting rate to the nearest tenth.

Next: Reliability of Estimates