Survey Methodology

The sample for the National Compensation Survey (NCS) (Wages, Benefits, Compensation Cost Trends --Employment Cost Index (ECI) and Employment Costs for Employee Compensation (ECEC)) is selected using a three-stage design. The first stage involves the selection of areas. The NCS sample consists of 154 metropolitan and nonmetropolitan areas that represent the Nation's 326 metropolitan statistical areas and the remaining portions of the 50 States.

In the second stage, the sample of establishments is drawn by dividing the sample by industry and ownership. Each sample establishment is selected using a method of sampling called probability proportional to employment size. (See NCS: Occupational Wages in the United States, 2000 National Summary - Technical Note). Use of this technique means that the larger an establishment's employment, the greater its chance of selection.

The third stage of sampling is a probability sample of occupations within a sampled establishment. This step is performed by the field economist during an interview with the respondent using a method called Probability Selection of Occupations (PSO). During this process, the field economist obtains a complete list of employees with each selected employee representing a job within the establishment. As with establishment selection, the selection of a job is based on probability proportional to its size in the establishment. The greater the number of people working in a particular job, the greater the job's chance of selection. The field economist selects a certain number of sample occupations depending on the size of the establishment.

After job selection is complete, the field economist classifies each occupation under its corresponding major occupational group and occupational classification. Currently, NCS uses the 2000 Standard Occupational Classification (SOC) system.

Standard procedures are used to collect and tabulate all NCS data. The methodology and procedures used to estimate vary by product line. Wage series are computed by combining the wages for individual establishment/occupations. Before being combined, individual wage rates are weighted by: number of workers; the sample weight, adjusted for nonresponding establishments and other factors; and the occupation work schedule, which varies depending on whether hourly, weekly, or annual rates are being calculated.

Percentiles are calculated from average hourly wages for sampled establishment jobs with each occupation. The percentiles describe the distribution of an occupation's employment by the average wage rates for its jobs. For example, at the 10th percentile hourly wage for an occupation, one-tenth of the occupation's employment are found in sampled establishment jobs whose average wages are the same or less and nine-tenths are in jobs averaging the same or more. The calculations of the 25th, 50th, 75th, and 90th percentiles follow the same logic.

Benefit series have three weight-adjustment factors applied to the data. The first factor is introduced to account for establishment nonresponse. The second factor accounts for occupational nonresponse, and the third factor is introduced to adjust the estimated employment totals to actual counts of the employment by industry for the survey reference date. Appropriate employment or establishment totals are used to calculate the proportion, mean, or percentage that is desired.

To measure compensation costs free from the influence of employment shifts among occupations and industries, the ECI is calculated with fixed employment weights unlike the method with which wage series and benefit series are calculated. Since December 1994, 1990 employment counts from the Bureau's Occupational Employment Statistics (OES) program have been used. The ECI is a standard Laspeyres fixed-weighted index.

The ECEC estimates are based on data collected for the ECI. Unlike the ECI, ECEC estimates are weighted by the most recently available industry employment derived from data produced by the BLS Current Employment Statistics (CES) program.

Last Modified Date: November 16, 2006