A semi-annual survey designed to produce employment and wage estimates for about 800 occupation classifications. Comparable statistics are available in every area, Metropolitan Statistical Areas and substate regions.
OES is a federal-state cooperative program between the U.S. Bureau of Labor Statistics and state agencies. Surveyed employers are asked about the number of wage and salary workers in detailed occupations and about the wage distribution for those workers. OES survey samples are drawn from the universe of non-farm employers covered by the Unemployment Insurance (UI) system. In Minnesota, about 6,000 employers participate in the survey each year.
Viewable tables, downloadable files
Occupations are classified using U.S. Department of Labor’s Standard Occupational Classification (SOC) typology.
The definition of wage for the OES program is straight-time gross pay, including base pay, incentive pay (commissions and production bonuses), cost-of-living allowances, guaranteed pay, hazardous-duty pay, on-call pay and tips. Excluded from the wage definition are overtime pay, shift differentials, non-production bonuses, holiday pay, meal and lodging payments, draw, severance pay, back pay, jury duty pay and tuition reimbursements.
For occupations where 40 hours per week and 52 weeks per year are not typical (i.e. teaching positions), only annual salary statistics are presented.
Program notes: There is a 10-month time lag between the start of the survey period and the public release of survey findings. The Minnesota Department of Employment and Economic Development adjusts wage estimates quarterly to account for wage inflation, using the Bureau of Labor Statistic’s Employment Cost Index (ECI), online at www.bls.gov/ncs/ect/.
These links lead to the website of the U.S. Bureau of Labor Statistics.
For over 30 years, the Occupational Employment Statistics (OES) survey has been the chief source of detailed occupational employment data and for the past 12 years, has also been the chief source of detailed occupational wage data in the United States. Wage and employment data are collected and published for approximately 800 occupations. Moreover, firms are classified into their appropriate industry using the North American Industry Classification System (NAICS), allowing for the publication of occupational employment and wage statistics by industry. The survey is conducted biannually in all 50 states and aggregated to the national level.
The following employees are included in the employment estimates: full- or part-time paid workers who are paid a wage or salary, workers on paid leave, workers temporarily assigned to other units, paid owners, officers, and staff of incorporated firms. Excluded are proprietors of unincorporated firms, other self-employed and contract workers, unpaid family workers, and workers on unpaid leave. If an employee's job requires multiple tasks which span occupations, he/she is recorded in the occupation requiring the highest level of skill. If there is no difference in the skill level of the occupations, she/he is recorded in the occupation in which he/she spent the most time.
Six wage estimate columns are also presented. The mean (or average) wage is calculated as the estimated total wages for an occupation divided by its estimated employment. The median wage, or 50th percentile wage, is the wage value at which 50 percent of workers in an occupation earn wages below, and 50 percent earn wages above the median wage. The 10th, 25th, 75th, and 90th percentiles are also shown.
The definition of a wage for the OES program is straight-time gross pay, exclusive of premium pay. Included are base pay, incentive pay (including commissions and production bonuses), cost-of-living allowances, guaranteed pay, hazardous-duty pay, on-call pay, and tips. Some exclusions from the wage are overtime pay, shift differentials, non-production bonuses, holiday premium pay, meal and lodging payments, draw, severance pay, back pay, jury duty pay, and tuition reimbursements. Hourly wage rates are given for almost all occupations, specifically for those where full-time is considered to be 40 hours per week and 52 weeks per year, or 2,080 hours per year. For occupations where 40 hours per week or 52 weeks per year are not typical, annual salary statistics are presented. These occupations include all Elementary, Secondary, and Post-Secondary Teacher occupations; Athletes, Coaches, Umpires, and Related Workers; Airline Pilots, Copilots, and Flight Engineers; and Commercial Pilots. Conversely, only hourly wages are produced for Actors; Dancers; Musicians and singers; and Entertainers and Performers, Sports and Related Workers, All Other due to the typically irregular employment and hours in these occupations.
The Bureau of Labor Statistics' OES program requires that occupational employment and wage estimates meet the following criteria before they can be published:
1. Occupational employment of at least 10 and more than .01 percent of total employment in the appropriate area/NAICS cell.
2. Relative standard error of less than 50 percent to publish employment or a relative standard error of less than 30 to publish wages.
3. At least three firms report data for an occupation, and no one firm reports more than 50 percent of the occupational employment, and no two firms report more than 75 percent of the occupational employment in the appropriate area/NAICS cell.
Employment and wage estimates for occupations that do not meet the above publication criteria are not listed. For cases where the wages meet the above criteria but employment has a relative error of 50 percent or higher, the wages are listed but employment is suppressed. The reverse is true if the employment figures meet the criteria, but the wages do not. For those occupations for which a 40-hour-per-week/52-week-per-year schedule is not typical, the only wages available are annual wages. This includes all Elementary, Secondary, and Post-Secondary Teacher occupations; Music Directors, Singers, Composers, and Related Workers; Musicians, Instrumental; Producers, Directors, Actors, and Other Entertainers; and Athletes, Coaches, Umpires, and Related Workers; Airline Pilots, Copilots, and Flight Engineers; and Commercial Pilots.
The November 2008, May and November 2009, May and November 2010, and May 2011 OES samples were drawn from the universe of employers covered by the Minnesota Unemployment Insurance (UI) program during November 2007, May and November 2008, May and November 2009, and May 2010, supplemented with a list supplying establishment information on railroads. Self-employed and other workers not covered by unemployment insurance were not included in this survey. Industry coverage includes natural resources and mining; construction; manufacturing; trade (includes retail and wholesale trade), transportation, and utilities; information; financial activities; professional and business services; educational and health services; leisure and hospitality; and other services. The survey also covers government establishments.
The universe was stratified by MSA, 3-digit and 4-digit (in some cases) NAICS code, and employment size class. A probability sample was selected from each MSA/industry/size-class cell. UI units reporting 250 or more employees were sampled with certainty across the three-year cycle of the survey by including one-third of these firms in the sample each year. Firms in industries with less than four units in an MSA within every size class were also sampled with certainty across the three-year survey period. All state and federal government establishments were included in each year of the survey.
All establishments in the sample are initially sent an OES questionnaire. The OES questionnaires include a list of industry-specific occupational titles with corresponding columns for the employer to enter employment and wage data. For each occupation, employers were asked to indicate the total number of workers who were employed during the pay period that included May 12 for the May panel and November 12 for the November panel, and to indicate the number of workers in each of the following 11 possible wage ranges:
Under $9.25 Under $19,240
$9.25 – 11.49 $19,240 – 23,919
$11.50 – 14.49 $23,920 – 30,159
$14.50 – 18.24 $30,160 – 37,959
$18.25 – 22.74 $37,960 – 47,319
$22.75 – 28.74 $47,320 – 59,799
$28.75 – 35.99 $59,800 – 74,879
$36.00 – 45.24 $74,880 – 94,119
$45.25 – 56.99 $94,120 – 118,559
$57.00 – 71.49 $118,560 –148,719
$71.50 – 89.99 $148,720 -187,199
$90.00 and over $187,200 and over
For Minnesota's November 2008, all of 2009, all of 2010, and May 2011 survey rounds, 16,205 responses were received out of 20,696 establishments surveyed, yielding a response rate of 78.3 percent.
Using three years’ worth of data in the sample (November 2008, May and November 2009, May and November 2010, and May 2011) significantly reduces sampling errors and thus allows more reliable estimates at higher levels of disaggregation across geographic areas, industry types, and occupations. The wage estimates for each panel were escalated to the 4th quarter of 2011 via the appropriate ECI factors to provide a more current representation of the wage before being rolled together to produce estimates. This assumes that each occupation's wage changed by the same amount as the average movement of the wages in the appropriate occupational division (BLS produces ECI for ten broad occupational divisions). This "wage updating" procedure also assumes that there are no geographic differences in occupational wage growth, since the ECI factors are not currently produced at the state or MSA levels. In research conducted by BLS over the past several years, ECI wage updating has compared favorably with other modeling approaches. Current research results support the continued use of the ECI wage updating methodology.
A "nearest neighbor" or "hot deck" imputation procedure is used to impute for establishment non-response. In hot decking, units in the sample are stratified into year/state/5-digit NAICS cells. Even though the sample is drawn on a 3- and 4-digit NAICS level, each sampled firm is classified using a 6-digit NAICS code. This allows a stratification of sampled firms into less specific 5-digit NAICS groups. Within each cell, a responding unit ("donor") is selected to represent each non-respondent under the condition that a donor cannot be selected twice. The matching of donor to non-respondent is also subject to employment size comparison to ensure that an appropriate donor is chosen. Once a donor and non-respondent are matched, the proportional distribution of occupational employment from the donor is multiplied by the recipient's reference month employment (as reported to Unemployment Insurance). If a donor is not available in a specific year/state/5-digit NAICS cell, the procedure advances to successively higher level cells until a donor is found.
A variation of the above procedure is used to impute for the type of non-response that occurs when an establishment reports employment by occupation but omits wage information. In this case, units in the sample are stratified into year/MSA/5-digit NAICS cells. A wage-employment distribution is then calculated for each occupation with missing wage data based on the usable data in the year/MSA/5-digit NAICS cell. This wage-employment distribution is then used to estimate the missing wage-employment data.
Concurrent with the drawing of the sample, is the assigning of a weight for each establishment calculated as the inverse of the probability of sample selection. For example, if four units out of 12 in an MSA/3-digit or 4-digit NAICS/size-class cell were sampled, each of the sampled units was assigned a weight of three. Then benchmark factors are calculated that ensure that estimates of employment approximately match the population employment at each level of aggregation. The source of the population employment data is the refined Minnesota Quarterly Census of Employment and Wages (QCEW) file.
Four different benchmark factors are used. One benchmark factor is used to match the OES-estimated employment within an MSA/3-digit or 4-digit NAICS/size-class cell with the QUI employment for the same cell. The second benchmark factor is calculated so as to match OES-estimated employment to the QUI at the statewide 3-digit or 4-digit NAICS level. Two other factors are calculated similarly, one at the statewide 3-digit industry level and one at the statewide 2-digit (or sector) industry level. Within each MSA, the estimated employment for an occupation at the 3-digit or 4-digit NAICS/wage interval level is calculated by multiplying the weighted employment by its four benchmark factors.
Thus, within each MSA, the estimated employment for an occupation at the 3-digit or 4-digit NAICS/wage interval level was given by summing the individual unit's weighted employment, which is its reported employment times its sample weight times its four benchmark factors. The estimated employment for an occupation at the all-industry level was obtained by summing the occupational interval employment estimate across all industries within an MSA reporting that occupation. The employment and wage data for federal government workers in each occupation were calculated separately and added to the survey-derived data.
The mean wage is the estimated total wages for an occupation divided by its weighted survey employment. For the upper open-ended wage interval of greater than $90.00 an hour (or greater than $187,200 annually), the mean hourly wage value was calculated from the BLS National Compensation Survey. For the other (closed) intervals, it is assumed that the mean hourly wage of all workers in that interval is the interval's midpoint. This method is supported by research conducted by the Office of Compensation and Working Conditions. For each occupation, total weighted wages in each interval (i.e., the interval mean wage times weighted employment) were summed across all intervals and divided by the occupation's weighted employment to obtain a mean wage.
The median wage is the estimated 50th percentile of the distribution of wages; 50 percent of workers in an occupation earn wages below, and 50 percent earn wages above the median wage. Similarly, 25 percent of workers in an occupation earn wages below, and 75 percent earn wages above the 25th percentile. The wage interval containing the median wage was located using a cumulative frequency count of employment across wage intervals. After the targeted wage interval was identified, the median wage rate was then estimated by a linear interpolation procedure. The 10th, 25th, 75th, and 90th percentiles were also determined by first finding the target wage interval, and then using linear interpolation.
The occupational employment and wage rates in this report are estimates derived from a sample survey. Two types of error are possible in an estimate based on a sample survey; sampling error and non-sampling error.
Sampling error occurs because a sample of observations may not perfectly reflect the population from which they are drawn. The particular sample used in this survey is one of a large number of possible samples of the same size that could have been selected using the same sample design and population file. Occupational employment and wage rate estimates derived from different samples will differ from one another. Sampling error is the difference between an estimate obtained from one sample and the true value that would be obtained if the entire universe had been surveyed.
Sampling error can be measured. The standard deviation or standard error of an estimate is one such measure of the precision with which an estimate from a particular sample approximates the true value of that population characteristic. Another measure, the relative standard error of the mean, is the standard deviation divided by the mean. Using statistical theory, the probability is 68.26 percent that the interval from the mean plus or minus one standard deviation from the mean includes the true population mean. Likewise, the probability is 95.5 percent that the wage is within two standard deviations from the actual population value. For example, consider an occupation with a mean wage of $10.00 and a standard deviation of $1.00 (thus a relative error of 10 percent). Then the probability that the interval from $9.00/hour to $11.00/hour includes the true mean wage is 68.28 percent, while the probability that the interval $8.00/hour to $12.00/hour includes the true mean wage is 95.5 percent. Instead, if the standard deviation is $3.00, meaning the relative error is 30 percent, then the probability is 68 percent that the actual mean wage is between $7.00/hour and $13.00/hour. As this example demonstrates, estimates with lower relative errors are more precise than those with higher relative errors.
The standard errors of the estimates of employment are also calculated, with the same interpretation applying to the calculation of these intervals, and with the same implication that higher confidence can be associated with lower relative standard errors. Employment and wage estimates must have relative standard errors of less than 50 percent in order to be published.
Non-sampling error cannot be explicitly measured. These errors may be due to non-response, faulty questionnaire design, occupational coding errors, errors in respondents' employment and wage data, transcription and data editing errors, non-response adjustment errors, and estimation errors. These errors would also occur if a complete census were conducted under the same conditions as the sample survey. To eliminate as many of these types of errors as possible, the survey responses are checked and edited, and non-respondents are contacted via follow-up mailings and phone calls to try to solicit their responses.