Documentation

Note: Updates to this data product are discontinued.

Table of Contents

• Background
• Identifying Creative Occupations Using 2000 Census Data
• Refining the Occupation Classification
• Updating the Creative Class Codes with 2007-11 American Community Survey Data
• Related Resources

Background

Richard Florida's Rise of the Creative Class makes a compelling argument that urban development now depends on novel combinations of knowledge and ideas, that certain occupations specialize in this task, that people in these occupations are drawn to areas providing a high quality of life, and that the essential development strategy is to create an environment that attracts and retains these workers. While developed with urban areas in mind, this thesis may be particularly relevant in rural areas, which lose much of their young talent as high school graduates leave for college, the armed forces, or "city lights."

Our analysis of recent development in rural U.S. counties, which focuses on natural amenities (for which ERS has also computed county-level scores) as quality-of-life indicators, supports the creative class thesis (see Recasting the Creative Class to Examine Growth Processes in Rural and Urban Counties and "The Creative Class: A Key to Rural Growth," Amber Waves, April 2007). A repetition for urban counties also shows a strong relationship between creative class presence and growth, although creative class employment plays a smaller role. However, our results depend on a recast creative class measure, which excludes from the original Florida measure many occupations with low creativity requirements and those involved primarily in services to the residential community (i.e., with numbers roughly proportional to population). Our measure conforms more closely to the concept of creative class and proves to be more highly associated with regional development than the original Florida measure.

Other work by Florida has demonstrated that a critical subset of the creative class is that comprised of fine, performing, and applied artists. His "Bohemian index"—the share of employment in arts occupations—is strongly associated with new firm formation and high-tech specialization in metropolitan areas. ERS research (see Emoting with their feet: Bohemian attraction to creative milieu, Journal of Economic Geography, August 2007) confirms that nonmetro counties with a surplus of artists tend to have higher rates of employment growth and new firm formation. The creative class codes data file also breaks out the share of employment in the arts.

Identifying Creative Occupations Using 2000 Census Data

O*NET, a Bureau of Labor Statistics data set that describes the skills generally used in occupations, was used to identify occupations that involve a high level of "thinking creatively." This skill element is defined as "developing, designing, or creating new applications, ideas, relationships, systems, or products, including artistic contributions." 

The O*NET compendium, previously known as the Dictionary of Occupational Titles, is produced by the Employment and Training Administration, U.S. Department of Labor, and provides comprehensive information on the functional requirements of more than 1,000 detailed occupations. The creativity measure provides a quantitative, though arguably imperfect, reference for assessing the creativity requirements among summary occupations that typically require a high degree of education.

Refining the Occupation Classification

Occupations were removed from consideration in the creative class measure—even if they typically required high levels of creativity—when their numbers are generally proportional to the residential population they serve (such as schoolteachers, judges, and medical doctors). Specifically, the measure excludes the summary "health care practitioners and technical occupations" group and schoolteachers and aides in the "education, training, and library" occupational group. We argue that such economic reproduction character does not disqualify college professors and "librarians, curators, and archivists" as their services are often provided to a nonresident population. Purging legal support occupations and judges while retaining lawyers might be questioned. However, the important role that lawyers play in devising solutions to new problems created by economic development is a compelling argument for their inclusion.

"Life, physical, and social science technicians" are excluded from the recast classification due to generally low requirements for creative thinking, although technicians in "architecture and engineering occupations" are retained due to higher requirements for creative thinking. This same justification for exclusion applies to "business operations specialists" and "other financial specialists" within the "business and financial operations" occupational group. Within "management occupations," "farmers and farm managers" are excluded due to low creativity requirements of farmers as reported in O*NET. However, management positions in public administration that would be appropriately excluded given the economic reproduction criterion are not separated from other management positions in the classification and so are retained. "Supervisory sales" creates a problem as many small business owners fall in this category, yet in the 2000 Census of Population, the category is mixed with other sales occupations (although not retail sales and cashiers). We have kept this larger category in the recast creative class as we are uncomfortable with excluding small business owners.

Using the occupations as detailed in the table, data from the 2000 Census of Population, summary file 4 was compiled for 3,139 counties (due to small sample size there are no data for Kalawao, HI, and Loving, TX). The share of the employed population 16 years and older in these occupations represents the ERS measure of creative class.

The creative class measure was also constructed for 1990. Creative class occupations from the finest level of detail in the 2000 Census occupational data (93 occupations) were used to define the 1990 creative class, constructed from 1990 Equal Employment Opportunity special tabulation files, much more disaggregate data (512 occupations). A major change in the Standard Occupational Classification codes between 1990 and 2000 complicated the construction of comparable measures, as the 2000 occupations did not correspond to summary 1990 occupations. The majority of the codes in 2000 could be constructed from the disaggregate 1990 data, but a number of detailed 1990 occupations were distributed across creative class and non-creative class occupational codes in 2000. Conversion factors that partitioned the 1990 data to correspond with the 2000 codes were computed from the U.S. Census Bureau's Technical Paper #65 (The Relationship Between The 1990 Census and Census 2000 Industry and Occupation Classification Systems) for "managers and administrators, n.e.c., salaried;" "inspectors and compliance officers, except construction;" and "sales representatives, mining, manufacturing, and wholesale."

The change in the Standard Occupational Classification also affected the compilation of arts occupations as a subset of the creative class. Within the 93 detailed occupations included in the 2000 Census STF4 file, "Art and design workers" and "Entertainers and performers, sports, and related workers" are the only two categories that are not substantially co-mingled with non-arts occupations. The corresponding 1990 occupational categories are "Designers," "Painters, sculptors, craft-artists, and artist printmakers," "Photographers," and "Artists, performers, and related workers, n.e.c.," "Musicians and composers," "Actors and directors," "Dancers" and "Athletes." The 2000 aggregation does not allow purging athletes from the data series, though they comprise a minimal share of the total, nor does it allow the inclusion of authors who are co-mingled with the considerably larger number of technical writers.

These refinements resulted in an estimated creative class share of the workforce of 21 percent in 1990 (23 percent in metro areas and 14 percent in nonmetro areas) and 25 percent in 2000 (27 percent in metro and 17 percent in nonmetro).

Updating the Creative Class Codes with 2007-11 American Community Survey Data

Updating the 2000 Census creative class data with 2010 Census creative class data is not possible because the long form sent to 1 in 6 households to collect detailed data such as occupation is no longer used. Instead, the American Community Survey (ACS)—a continuous survey that samples 1 in 40 households each year—now provides the detailed data that had been collected through the long form. While this major change in data collection strategy greatly increases the timeliness of these detailed data for the nation, States, and large cities, generating reliable estimates for smaller settlements and nonmetropolitan counties requires pooling 5 years of ACS data. By pooling 5 years of data, taking into account nonresponse, we can achieve an effective sampling rate of about 1 in 11 households. Thus, the latest update provided is for pooled data collected for 2007 through 2011.

The change from the decennial Census summary file 4 to the 5-year pooled ACS file also affected the level of detail of occupational categories that are published and made available to the public. The 93 occupations in the 2000 summary file 4 have been collapsed into 26 occupations in the publicly available ACS file which is too coarse an aggregation to construct a creative class measure that is comparable. ERS ordered a special tabulation from the Census Bureau that uses unpublished disaggregated occupational categories to replicate the 2000 creative class measure. The specification by detailed occupational category from the latest 2010 Standard Occupation Classification used in the 2007-11 ACS (SOC Code to Census Code Crosswalk) is available on the Census website, and also provided below:

The major benefit accompanying the special tabulation is the provision of standard errors accompanying each creative occupation and arts occupation estimate. The standard errors provide information on the reliability of the estimate. Since the estimate is derived from a sample of the population, it might vary substantially if the size of the total sample is relatively small and if the phenomenon of interest is relatively rare. For example, suppose the true share of artists is 0.01 or 1 in 100 employees in a county. If a county has 1100 total employees, then the estimate will be based on about 100 respondents. By the luck of the draw, the sample chosen for the county could contain no artists, 1 or more artists, or (improbably) all 11 artists. The standard error provides an exact statistical measure of the reliability of the estimate. This is very valuable in comparing creative or arts occupation shares between 2007-11 and 2000, as a nominal increase or decrease in share may be due solely to sampling error. Formulas for comparing estimates across time periods incorporating standard error are provided in Appendix 4 of A Compass for Understanding and Using American Community Survey Data: What Researchers Need to Know.

For a complete discussion of ERS research on the creative class, see Recasting the Creative Class To Examine Growth Processes in Rural and Urban Counties, by David A. McGranahan and Timothy R. Wojan, Regional Studies Vol. 41 No. 2, pp. 197-216, April 2007.

Related Resources

• Recasting the Creative Class To Examine Growth Processes in Rural and Urban Counties, by David A. McGranahan and Timothy R. Wojan, Regional Studies Vol. 41 No. 2, pp. 197-216, April 2007.
• "Arts Employment Is Burgeoning in Some Rural Areas" and "The Creative Class: A Key to Rural Growth," in Amber Waves magazine, November and April 2007.
• Emoting with Their Feet: Bohemian Attraction to Creative Milieu, by Timothy R. Wojan, D. M. Lambert, and D. A. McGranahan, Journal of Economic Geography 2007; Digital Object Identifier (DOI): 10.1093/jeg/lbm029.
• The Rural Growth Trifecta: Outdoor Amenities, Creative Class and Entrepreneurial Context, by David A. McGranahan, Timothy R. Wojan and Dayton M. Lambert Journal of Economic Geography, May 2011. Digital Object Identifier (DOI): 093/jeg/lbq007.