National Compensation Survey

National Compensation Survey: Occupational Earnings in the United States, 2006

U.S. Department of Labor
Elaine L. Chao, Secretary

U.S. Bureau of Labor Statistics
Philip L. Rones, Deputy Commissioner September 2007

Bulletin 2590



Occupational Earnings Tables: United States, December 2005 – January 2007 (average reference date June 2006)

Relative Standard Error (RSE) tables to accompany mean hourly, weekly, and annual earnings tables

Appendix A: Technical Note

Appendix B: Occupational Descriptions (PDF)

Appendix C: Survey Areas and Geographic Coverage (PDF)



The National Compensation Survey (NCS) provides comprehensive measures of occupational earnings, compensation cost trends, benefit incidence, and detailed benefit provisions. This bulletin presents estimates of occupational pay for the Nation. These national estimates originate from the NCS locality survey data and are weighted to represent the Nation as a whole. Data for more than one-half of the 152 individual NCS localities used for national estimates have been previously published. This includes pay estimates for workers in major sectors within the United States economy in 2006: Civilian, private industry, and State and local government; and by various occupational and establishment characteristics. The civilian economy, by NCS definition, excludes Federal Government, agricultural, and household workers.

Questions regarding these data and recent and historical NCS wage data can be addressed by calling the information line at (202) 691-6199 or by e-mail to This information is available to sensory-impaired individuals on request. Voice phone: (202) 691-5200; Federal Relay Service: 1 (800) 877-8339. Data requests also may be sent by mail to the U.S. Bureau of Labor Statistics, Division of Compensation Data Analysis and Planning, 2 Massachusetts Avenue, NE, Room 4175, Washington, DC 20212. Material in this publication is in the public domain and, with appropriate credit, may be reproduced without permission.

U. S. Bureau of Labor Statistics (BLS) field economists collected and reviewed the survey data. The Office of Compensation and Working Conditions, in cooperation with the Office of Field Operations and the Office of Technology and Survey Processing in the BLS National Office, designed the survey, processed the data, and prepared the survey for publication. The survey could not have been conducted without the cooperation of the many private businesses and government jurisdictions that provided pay data included in this bulletin. BLS thanks these respondents for their cooperation.

This edition of the National Compensation Survey (NCS) annual bulletin on occupational wages in the United States has undergone several major changes since its last publication. The most noticeable change is the bulletin's format. Formerly, the NCS annual wage bulletin was printed in soft-cover book format, with nearly 150 pages of data tables, as well as posted on the Bureau of Labor Statistics Web site in TXT and PDF formats. Starting with the 2006 national wage bulletin, the overview, explanatory text, data tables, and technical note are published exclusively in digital formats — on CD and on BLS Internet pages. The chartbook that holds the CD includes highlights of the data in the annual bulletin, depicted in a variety of charts and graphs.

This edition includes changes in data content, as well. Major changes are listed here and described below:

New occupational and industry classification of workers. Since its inception in 1996, NCS wage surveys have published data on occupations according to the 1990 Occupational Classification System (OCS). Beginning with locality bulletins published in September 2006, new NCS wage publications, including the 2006 national wage bulletin, classify occupations according to the 2000 Standard Occupational Classification (SOC) codes. The 2000 SOC system defines more than 800 detailed occupations and is designed to reflect the current occupational structure in the United States better than previous occupational systems. Detailed occupations are combined into broad occupations, broad occupations are combined into minor groups, and minor groups are combined into major groups. For details on the SOC occupational titles used for the NCS wage bulletins, see appendix B. The design of several levels of aggregation is intended to meet the widely varying needs of data users. For more information on the SOC classification system, see:

Also beginning with the locality bulletins published in September 2006, new NCS wage publications, including the 2006 national bulletin, classify the type of establishment in which employees work according to the 2002 North American Industry Classification System (NAICS). The NCS had formerly published its wage data according to the 1987 Standard Industrial Classification system (SIC) since the survey began in 1996. For more detail on the NAICS classification system, see: These occupational and industry classification changes are part of a major initiative to convert all Federal surveys to the NAICS and SOC classification systems. For more information on this transition, see

Imputation for temporary nonresponse of establishments. For the first time, the NCS wage program is imputing data for temporary nonresponse situations. The National Compensation Survey is voluntary, and a company official may refuse to participate in the initial survey or may be unwilling or unable to update previously collected data during a subsequent contact. For those situations where previous wage data cannot be updated, an estimate for the missing data is imputed using information obtained from similar establishments and occupations.

Benchmarking of estimated employment. Post stratification, also known as benchmarking, has been introduced to adjust survey sample weights so that these weights reflect the current count of employment by industry. Initial weights are derived when the sample of establishments are selected, reflecting employment distribution by industry at that time. Those weights may be up to 7 years old for the oldest panel of five sample rotation panels at the time of publication. Benchmarking adjusts those weights to reflect the employment distribution by industry for the reference date of the data.

New tables and detail. The 2006 NCS national bulletin has added many new tables: wages for supervisory occupations, private industry sector, nonprofit establishments, and hospitals; wages by percentile; and other new details. For a full description of these changes, see "Change Comes to the National Compensation Survey Locality Wage Bulletins, Compensation and Working Conditions Online, available on the Internet, at: The tables are listed below.

(Back to top)

Occupational Earnings Tables:United States, December 2005 – January 2007 (average reference date June 2006)

The 2006 NCS national wage bulletin includes occupational earnings tables 1-41; relative standard errors of the estimates for tables 2-7, 12, 24-26, 30-34, and 40-41; and appendix tables 1 and 2. The relative standard errors tables are in the section below the occupational earnings tables; they are titled and numbered to correspond to their respective wage-estimates tables. Appendix tables 1 and 2 are part of appendix A.

Summary table. Table 1 presents an overview of all tables in this bulletin. Mean hourly earnings, weekly hours, and relative standard errors are given for Civilian, private industry, and State and local government workers by selected worker and establishment characteristics. Worker characteristics include high-level and intermediate occupational aggregation, full-time and part-time status, union and nonunion status, and time and incentive pay status. Establishment characteristics include goods producing, service providing, and size of establishment.

Full-time and part-time workers. Mean and median hourly, weekly, and annual earnings estimates are provided for full-time workers in each of the three major sectors, as well as mean weekly and annual hours worked. Further tables provide mean hourly earnings estimates for full-time and part-time workers by level within occupation.

Work levels. Levels are standardized measures of duties and responsibilities that apply to all occupations. The NCS National bulletin has previously published levels data, but this is the first time combined levels have been published for the Nation. Table 11 shows four bands I through IV, which combine levels 1 through 4; levels 5 through 8; levels 9 through 12; and levels 13 through 15; respectively. The publication of combined levels is intended to make the wage estimates more useful to compensation analysts.

Union and nonunion workers. Separate estimates on mean hourly occupational earnings are provided for union and nonunion workers for each of the three major sectors of the economy and for a detailed occupational groups as well as for detailed occupations.

Time and incentive workers. This table provides hourly earnings estimates for workers paid on a time or incentive basis for those who work in the Civilian and private industry sectors.

Percentiles. The 2006 National bulletin provides estimates on the mean hourly wage for the 10th percentile, the 25th percentile, 50th percentile (the median), the 75th percentile, and the 90th percentile of occupational wages separately for all Civilian, full-time Civilian, part-time Civilian, all private industry, full-time private industry, part-time private industry, all State and local government, full-time State and local government, and part-time State and local government workers.

Supervisory occupations. The set of tables on supervisory occupations are new to the national bulletin this year: Mean and median weekly and annual occupational earnings and mean weekly and annual hours for workers in management occupations by supervisory responsibility: team leaders, first-line, second-line, and third-line supervisory duties — provided separately for the Civilian, private industry, and State and local governments sectors.

Size of establishment. Estimates for mean hourly earnings for workers in major occupational groups by size of establishment — 1-49 workers, 50-99 workers, 100-499 workers, and 500 workers or more—are given separately for the Civilian, private industry, and State and local government sectors. There are also separate tables for detailed occupational estimates on mean and median hourly, weekly, and annual earnings and mean weekly and annual hours of full-time private industry workers in establishments with fewer than 100 workers and those in establishments with 100 workers or more.

Private industry sector. The 2006 national wage bulletin includes a new table on mean hourly earnings for workers by sector of private industry, for major occupational groups.

Nonprofit establishments. For the first time in the NCS national wage bulletin, mean and median hourly, weekly, and annual earnings and mean weekly and annual hours for private industry full-time workers who work in nonprofit establishments have been published by detailed occupation and level.

Hospitals. The bulletin provides a new table on mean and median hourly, weekly, and annual earnings and mean weekly and annual hours for full-time Civilian workers in hospitals, by detailed occupation and level. The bulletin also provides tables for private industry and State and local government sectors on mean hourly earnings estimates for all, full-time, and part-time workers in hospitals by detailed occupation.

Metropolitan and nonmetropolitan areas. The 2006 NCS national wage bulletin includes estimates of mean hourly earnings for Civilian, State and local government, and private industry workers by metropolitan, nonmetropolitan, and geographic areas; estimates of mean hourly earnings for Civilian workers in metropolitan and nonmetropolitan areas by worker and establishment characteristics; and estimates of mean hourly wages for Civilian workers in metropolitan and nonmetropolitan areas by geographic area. The geographic areas in these tables are Census divisions. See appendix C for more information. Detailed tables of occupational wages by Census region and division, which were published in previous NCS national bulletins, are not included here; however, they will be included in the 2006 NCS Census division wage bulletins, which are forthcoming.

Combined Wage Tables. Wages tables 1-15, 16-30, 31-41, and all wage tables are given here for ease of printing many tables at once.

(Back to top)

Relative Standard Error (RSE) tables to accompany mean hourly, weekly, and annual earnings tables:

Combined RSE Tables. All of the RSE tables are given here for ease of printing many tables at once.

Appendix A: Technical Note

  • Appendix table 1. Number of workers represented by the survey. (TXT) (PDF 10K)
  • Appendix table 2. Survey establishment response. (TXT) (PDF 10K)
  • This section provides basic information on the procedures and concepts. It is divided into three parts: Planning for the survey; data collection; and processing and analyzing the data. Although this section answers some questions commonly asked by data users, it is not a comprehensive description of all of the steps required to produce the data.

    Planning for the survey

    The overall design of the National Compensation Survey (NCS) includes questions of survey scope, sampling frame, and sample selection.

    Survey scope

    This survey covered establishments employing one worker or more in private goods-producing industries (mining, construction, and manufacturing); private service-providing industries (trade, transportation, and utilities, information, financial activities, professional and business services, education and health services, leisure and hospitality, and other services); State governments; and local governments employing 50 or more workers. Agriculture, forestry, fishing and hunting, private households, and the Federal Government were excluded from the scope of the survey. For purposes of this survey, an establishment is an economic unit that produces goods or services, a central administrative office, or an auxiliary unit providing support services to a company. For private industries in this survey, the establishment is usually at a single physical location. For State and local governments, an establishment is defined as all locations of a government agency within the sampled area.

    Sampling frame

    The list of establishments from which the survey sample was selected (sampling frame) was developed from State unemployment insurance reports. Due to the volatility of industries within the private sector, sampling frames were developed using the most recent month of reference available at the time the sample was selected. Approximately one-fifth of the sample is reselected each year.

    Sample design

    The sample for this survey area was selected using a two-stage stratified design with probability proportional to employment sampling at each stage. The first stage of sample selection was a probability sample of establishments. The sample of establishments was drawn by first stratifying the sampling frame by industry and ownership. The number of sample establishments allocated to each stratum is approximately proportional to the stratum employment. Each sampled establishment is selected within a stratum with a probability proportional to its employment. Use of this technique means that the larger an establishment’s employment, the greater its chance of selection. Weights were applied to each establishment when the data were tabulated so that it represents similar units (by industry and employment size) in the economy that were not selected for collection. The second stage of sample selection, detailed below, was a probability sample of occupations within a sampled establishment.

    Data collection

    The collection of data from survey respondents required detailed procedures. Field economists collected the data, working out of the U.S. Bureau of Labor Statistics (BLS) regional offices and visiting each establishment surveyed. Other contact methods, such as mail and telephone, were used to clarify and update data.

    Occupational selection and classification

    Identification of the occupations for which wage data were to be collected was a multi-step process:

  • Probability-proportional-to-size selection of establishment jobs
  • Classification of jobs into occupations based on the 2000 Standard Occupational Classification (SOC) system
  • Characterization of jobs as full-time or part-time, union or nonunion, and time or incentive
  • Determination of the level of work of each job
  • For each occupation, wage data were collected for those workers whose jobs could be characterized by the criteria identified in the last three steps. If a specific work level could not be determined, wages were still collected. In step one, the jobs to be sampled were selected at each establishment by the BLS field economist. A complete list of employees was used for sampling, with each selected worker representing a job within the establishment. As with the selection of establishments, the selection of a job was based on probability proportional to its size in the establishment. The greater the number of people working in a job in the establishment, the greater is its chance of selection. The number of jobs for which data were collected in each establishment was based on the establishment’s employment size. The number of jobs selected followed this schedule:


    of employees


    of selected jobs


    Up to 4



    250 or more


    The second step of the process entailed classifying the selected jobs into occupations based on their duties. NCS uses the 2000 Standard Occupational Classification (SOC) system. A selected job may fall into any one of about 800 occupational classifications, from accountant to zoologist. When workers could be classified in more than one occupation, they were classified in the occupation that required the higher skill level. When there was no perceptible difference in skill level, the workers were classified in the occupation that described their primary activity. Each occupational classification is an element of a broader classification known as a major group. Occupations can fall into any of 22 major groups. Appendix B contains a complete list of all individual occupations, classified by the major group to which they belong. In step three, certain other job characteristics of the chosen worker were identified. First, the worker was identified as holding either a full-time or part-time job, based on the establishment’s definition of those terms. Then, the worker was classified as having a time versus incentive job, depending on whether any part of pay was directly based on the actual production of the worker, rather than solely on hours worked. Finally, the worker was identified as being in a union job or a nonunion job. For more detail, see the “Definition of terms” section, which follows.

    Occupational leveling

    In the last step before wage data were collected, the work level of each selected job was determined using a point factor leveling process. Point factor leveling matches certain aspects of a job to specific levels of work with assigned point values. Points for each factor are then totaled to determine the overall work level for the job. The NCS program is in the process of converting from a nine-factor to a four-factor occupational leveling system. The conversion is being phased in via annual NCS sample replenishment groups and will require several years for full implementation. The four occupational leveling factors are:

    Each factor consists of several levels, and each level has an associated description and assigned points. A knowledge guide for 24 families of closely related occupations contains short definitions of the point levels of knowledge expected for the occupations and presents relevant examples. The other three factors use identical descriptions for all occupational categories and contain a definition of each point level within each factor. The description within each factor best matching the job is chosen. The point levels within each factor are designed to describe the thresholds of distinct levels of work. When a job does not meet the full description of a point level, the next lowest point level is used. Points for the four factors are totaled to determine the overall work level. NCS publishes data for up to 15 work levels. Most supervisory occupations are evaluated based on their duties and responsibilities. A modified approach is used for professional and administrative supervisors when they direct professional work and are paid primarily to supervise. Such supervisory occupations are leveled based on the work level of the highest position reporting to them. For a complete description of point factor leveling, refer to the publication “National Compensation Survey: Guide for Evaluating Your Firm’s Jobs and Pay,” available on the Internet, at

    Combined work levels

    This bulletin includes a table which simplifies the presentation of work levels by combining them into four broad groups. The groups were determined by combinations of knowledge, job controls and complexity, contacts, physical environment, and supervisory duties, and are meant to be comparable across different occupations. The broad groups and the combined work levels are:

    Groupdesignation Levelscombined

    Group I

    Levels 1–4

    Group II

    Levels 5–8

    Group III

    Levels 9–12

    Group IV

    Levels 13–15

    Collection period

    Survey data were collected over a 13-month period for 60 metropolitan areas in the NCS program. For 20 small metropolitan areas and all 72 nonmetropolitan areas, data were collected over a 4-month period. For each establishment in the survey, the data reflect the establishment’s most recent information at the time of collection. The data for the National bulletin were compiled from locality data collected between December 2005 and January 2007. The average reference period was June 2006.


    Earnings were defined as regular payments from the employer to the employee as compensation for straight-time hourly work, or for any salaried work performed. The following components were included as part of earnings:

    The following forms of payments were not considered part of straight-time earnings:

    To calculate earnings for various periods (hourly, weekly, and annual), data on work schedules also were collected. For hourly workers, scheduled hours worked per day and per week, exclusive of overtime, were recorded. Annual weeks worked were determined. Because salaried workers who are exempt from overtime provisions often work beyond the assigned work schedule, their typical number of hours actually worked was collected.

    Definition of terms

    Full-time worker. Any employee whom the employer considers to be full time.

    Part-time worker. Any employee whom the employer considers to be part time.

    Time-based worker. Any employee whose earnings are solely tied to an hourly rate or salary.

    Incentive worker. Any employee whose earnings are tied, at least in part, to commissions, piece rates, production bonuses, or other incentives based on production or sales.

    Nonunion worker. An employee in an occupation not meeting the conditions for union coverage.

    Union worker. Any employee is in a union occupation when all of the following conditions are met:

    Level. A ranking within an occupation based on the requirements of the position.

    Processing and analyzing the data

    Data were processed and analyzed at the BLS national office following collection. This includes weighting, adjusting for nonresponse, imputation, benchmarking, and calculating wage estimates, variances, and other summary statistics.

    Weighting and nonresponse

    Sample weights were calculated for each establishment and occupation in the survey. These weights reflected the relative size of the occupation within the establishment and of the establishment within the sample universe. Weights were used to aggregate data for the individual establishments or occupations into the various data series. Some of the establishments surveyed could not supply or refused to supply information. If data were not provided by a sample member during the initial interview, the weights of responding sample members in the same or similar cells were adjusted to account for the missing data. This technique assumes that the mean value of data for the nonrespondents equals the mean value of data for the respondents at some detailed cell level. Responding and nonresponding establishments were classified into these cells according to industry and employment size. Responding and nonresponding occupations within responding establishments were classified into cells that were additionally defined by major occupation group. If average hourly earnings data were not provided by a sample member during the update interview, then missing average hourly earnings were imputed by multiplying prior average hourly earnings by the rate of change in the average hourly earnings of respondents. The regression model that takes into account available establishment characteristics is used to derive the rate of change in the average hourly earnings. Establishments that were determined to be out of business or outside the scope of the survey had their weights changed to zero.


    The wage series in the tables are computed by combining the wages for each sampled occupation. Before being combined, individual wage rates are weighted by the number of workers; the sample weight, adjusted for nonresponding establishments and other factors; and the occupation’s scheduled hours of work. The sample weight reflects the inverse of each unit’s probability of selection at each sample selection stage and four weight adjustment factors. The first factor adjusts for establishment nonresponse and the second factor adjusts for occupational nonresponse. The third factor adjusts for any special situations that may have occurred during data collection. The fourth factor, post-stratification, also called benchmarking, is introduced to adjust estimated employment totals to the current counts of employment by industry. The latest available employment counts were used to derive average hourly earnings in this publication. Not all calculated series met the criteria for publication. Before any series was published, it was reviewed to make sure that the number of observations underlying it was sufficient. This review prevented the publication of a series that could have revealed information about a specific establishment. Estimates of the number of workers represent the total in all establishments within the scope of the study, and not the number actually surveyed. Because occupational structures among establishments differ, estimates of the number of workers obtained from the sample of establishments serve to indicate only the relative importance of the occupational groups studied.


    The percentiles presented in tables 15 through 23 are computed using earnings reported for individual workers in sampled establishment jobs and their scheduled hours of work. Establishments in the survey may report only individual-worker earnings for each sampled job. For the calculation of percentile estimates, the individual-worker hourly earnings are appropriately weighted and then arrayed from lowest to highest. The published 10th, 25th, 50th, 75th, and 90th percentiles designate position in the earnings distribution within each published occupation. At the 50th percentile, the median, half of the hours are paid the same as or more than the rate shown, and half are paid the same as or less than the rate shown. At the 25th percentile, one-fourth of the hours are paid the same as or less than the rate shown, while at the 75th percentile, one-fourth of the hours are paid the same as or more than the rate shown. At the 10th percentile, one-tenth of the hours are paid the same as or less than the rate shown, while at the 90th percentile, one-tenth are paid the same as or more than the rate shown.

    Data reliability

    The data in this bulletin are estimates from a scientifically selected probability sample. There are two types of errors possible in an estimate based on a sample survey, sampling and nonsampling.

    Sampling errors occur because observations come only from a sample and not from an entire population. The sample used for this survey is one of a number of possible samples of the same size that could have been selected using the sample design. Estimates derived from the different samples would differ from one another. A measure of the variation among these differing estimates is called the standard error or sampling error. It indicates the precision with which an estimate from a particular sample approximates the average result of all possible samples. The relative standard error (RSE) is the standard error divided by the estimate. RSE data are provided alongside the earnings data in many of the bulletin tables. The standard error can be used to calculate a confidence interval around a sample estimate. As an example, table 1 shows mean hourly earnings for all workers of $19.29 per hour and a relative standard error of 1.2 percent for this estimate. At the 90-percent level, the confidence interval for this estimate is from $18.91 to $19.67 ($19.29 x 1.645 x 0.012 = $0.3808, rounded to $0.38); ($19.29 - 0.38 = $18.91; $19.29 + 0.38 = $19.67). If all possible samples were selected to estimate the population value, the interval from each sample would include the true population value approximately 90 percent of the time.

    Nonsampling errors also affect survey results. They can stem from many sources, such as inability to obtain information for some establishments, difficulties with survey definitions, inability of the respondents to provide correct information, or mistakes in recording or coding the data obtained. Although they were not specifically measured, the nonsampling errors were expected to be minimal due to the extensive training of the field economists who gathered the survey data, computer edits of the data, and detailed data review.

    (Back to top)

    Appendix B: Occupational Descriptions (PDF)

    Appendix C: Survey Areas and Geographic Coverage (PDF)