Burning Glass Technologies (BGT): The team used Burning Glass Technologies data from 2010 to 2019 to understand the level of education sought by employers filling open positions. BGT collects online job advertisements and populates a database that can be used for labor market research. We analyzed the occupations associated with the job ads, the metro areas and states of the employers, and the minimum level of education listed in the ads to calculate the share of jobs accessible to sub-baccalaureate workers in a given metro area for a given occupation. It is not known whether the lowest level of education mentioned by the employers is required or simply preferred.
To ensure sufficient observations to calculate the sub-baccalaureate share, we combined three years of BGT data to calculate shares where we found at least 100 job ads per occupation and area combination. For instance, 2012 estimates are based on job ads between 2010 and 2012, 2013 estimates are based on job ads between 2011 and 2013, and so forth. Where this approach yielded insufficient observations, we calculated the average sub-baccalaureate share for each occupation-area combination across all years (2010–17), and we also calculated this share for each occupation across the entire United States for these years. We then divided the states or metro areas into three groups based on the ratio of that area-occupation sub-baccalaureate share across all years with the national share across all years. An area where the ratio is below 0.95 is considered "least accessible to sub-baccalaureate workers." An area where the ratio is between 0.95 and 1.05 is considered "average accessibility to sub-baccalaureate workers." And an area where the ratio is greater than 1.05 is considered "most accessible to sub-baccalaureate workers." Finally, we use the average sub-baccalaureate share across an area's group in cases where we had fewer than 100 observations for a certain occupation.
U.S. Bureau of Labor Statistics (BLS) Occupational Employment Statistics (OES): The team accessed information on total employment and 10th, 25th, 50th, 75th, and 90th percentile wages by occupation for both states and metro areas between 2012 and 2019. OES data are updated annually. Note that information on some occupations in specific geographic areas is missing, so that total employment and median wage numbers do not reflect 100 percent of occupation data in a given year.
The metro area names and map displayed in the tool reflect the Office of Management and Budget’s (OMB) 2013 metro area definitions. Between 2012 and 2014, however, the composition and names of certain metro areas differed. While the tool displays the newer metro area’s name and geographic outline, the data displayed for those years represent the original metro area composition. This affects 64 metro areas in the tool: Anniston-Oxford-Jacksonville, AL (formerly Anniston-Oxford, AL); Atlanta-Sandy Springs-Roswell, GA (formerly Atlanta-Sandy Springs-Marietta, GA); Austin-Round Rock, TX (formerly Austin-Round Rock-San Marcos, TX); Bakersfield, CA (formerly Bakersfield-Delano, CA); Baltimore-Columbia-Towson, MD (formerly Baltimore-Towson, MD); Bend-Redmond, OR (formerly Bend, OR); Boise City, ID (formerly Boise City-Nampa, ID); Boston-Cambridge-Nashua, MA-NH (formerly Boston-Cambridge-Quincy, MA-NH); Buffalo-Cheektowaga-Niagara Falls, NY (formerly Buffalo-Niagara Falls, NY); Cape Girardeau, MO-IL (formerly Cape Girardeau-Jackson, MO-IL); Charleston-North Charleston, SC (formerly Charleston-North Charleston-Summerville, SC); Charlotte-Concord-Gastonia, NC-SC (formerly Charlotte-Gastonia-Rock Hill, NC-SC); Chicago-Naperville-Elgin, IL-IN-WI (formerly Chicago-Joliet-Naperville, IL-IN-WI); Cincinnati, OH-KY-IN (formerly Cincinnati-Middletown, OH-KY-IN); Cleveland-Elyria, OH (formerly Cleveland-Elyria-Mentor, OH); Denver-Aurora-Lakewood, CO (formerly Denver-Aurora-Broomfield, CO); Detroit-Warren-Dearborn, MI (formerly Detroit-Warren-Livonia, MI); Elizabethtown-Fort Knox, KY (formerly Elizabethtown, KY); Eugene, OR (formerly Eugene-Springfield, OR); Fort Collins, CO (formerly Fort Collins-Loveland, CO); Greenville-Anderson-Mauldin, SC (formerly Greenville-Mauldin-Easley, SC); Gulfport-Biloxi-Pascagoula, MS (formerly Gulfport-Biloxi, MS); Hinesville, GA (formerly Hinesville-Fort Stewart, GA); Houma-Thibodaux, LA (formerly Houma-Bayou Cane-Thibodaux, LA); Houston-The Woodlands-Sugar Land, TX (formerly Houston-Sugar Land-Baytown, TX); Indianapolis-Carmel-Anderson, IN (formerly Indianapolis-Carmel, IN); Janesville-Beloit, WI (formerly Janesville, WI); Kankakee, IL (formerly Kankakee-Bradley, IL); Kennewick-Richland, WA (formerly Kennewick-Pasco-Richland, WA); Killeen-Temple, TX (formerly Killeen-Temple-Fort Hood, TX); La Crosse-Onalaska, WI-MN (formerly La Crosse, WI-MN); Las Vegas-Henderson-Paradise, NV (formerly Las Vegas-Paradise, NV); Leominster-Gardner, MA (formerly Leominster-Fitchburg-Gardner, MA); Madera, CA (formerly Madera-Chowchilla, CA); Miami-Fort Lauderdale-West Palm Beach, FL (formerly Miami-Fort Lauderdale-Pompano Beach, FL); Muskegon, MI (formerly Muskegon-Norton Shores, MI); Myrtle Beach-Conway-North Myrtle Beach, SC-NC (formerly Myrtle Beach-North Myrtle Beach-Conway, SC); Naples-Immokalee-Marco Island, FL (formerly Naples-Marco Island, FL); New Orleans-Metairie, LA (formerly New Orleans-Metairie-Kenner, LA); New York-Newark-Jersey City, NY-NJ-PA (formerly New York-Northern New Jersey-Long Island, NY-NJ-PA); North Port-Sarasota-Bradenton, FL (formerly North Port-Bradenton-Sarasota, FL); Norwich-New London-Westerly, CT-RI (formerly Norwich-New London, CT-RI); Olympia-Tumwater, WA (formerly Olympia, WA); Panama City, FL (formerly Panama City-Lynn Haven-Panama City Beach, FL); Parkersburg-Vienna, WV (formerly Parkersburg-Marietta-Vienna, WV-OH); Phoenix-Mesa-Scottsdale, AZ (formerly Phoenix-Mesa-Glendale, AZ); Portland-South Portland, ME (formerly Portland-South Portland-Biddeford, ME); Providence-Warwick, RI-MA (formerly Providence-Fall River-Warwick, RI-MA); Reno, NV (formerly Reno-Sparks, NV); Sacramento-Roseville-Arden-Arcade, CA (formerly Sacramento--Arden-Arcade--Roseville, CA); Saginaw, MI (formerly Saginaw-Saginaw Township North, MI); Salisbury, MD-DE (formerly Salisbury, MD); San Diego-Carlsbad, CA (formerly San Diego-Carlsbad-San Marcos, CA); San Francisco-Oakland-Hayward, CA (formerly San Francisco-Oakland-Fremont, CA); San Luis Obispo-Paso Robles-Arroyo Grande, CA (formerly San Luis Obispo-Paso Robles, CA); Santa Rosa, CA (formerly Santa Rosa-Petaluma, CA); Scranton-Wilkes-Barre-Hazleton, PA (formerly Scranton-Wilkes-Barre, PA); Spokane-Spokane Valley, WA (formerly Spokane, WA); Stockton-Lodi, CA (formerly Stockton, CA); Texarkana, TX-AR (formerly Texarkana-Texarkana, TX-AR); Trenton, NJ (formerly Trenton-Ewing, NJ); Vineland-Bridgeton, NJ (formerly Vineland-Millville-Bridgeton, NJ); and Wenatchee, WA (formerly Wenatchee-East Wenatchee, WA).
U.S. Bureau of Labor Statistics (BLS) Employment Projections (2018–28): The team retrieved information on entry-level education requirements and the projected change in employment by occupation. Employment projections are updated every two years. The 2018–28 projections are located on the BLS website (see link above). For 63 occupations that required a "doctoral or professional degree," we assumed this educational requirement superseded the data obtained from Burning Glass Technologies' job ads.
U.S. Bureau of Economic Analysis (BEA): The team used the regional price parity for each metro area and state between 2012 and 2018 in order to adjust the OES wages for cost of living differences. When the estimates were created, regional price parity data were available only through 2017, so 2017 regional price parity figures were applied to 2018 OES data.
U.S. Census Bureau American Community Survey (ACS) Five-Year Estimates from IPUMS: The team retrieved American Community Survey data from IPUMS (2008–12, 2009–13, 2010–14, 2011–15, 2012–16, 2013–17, and 2014–18 five-year estimates) in order to estimate the median weekly hours worked for each occupation in each year. We then used these estimates to calculate annual median wages using the OES data, where these data indicated the use of annual wages was not appropriate.