By Daniel Backman
You asked, we delivered.
The American Time Use Survey (ATUS) was first fielded in 2003 and data have been collected and released annually since. Over that period, ATUS occupations have been coded using three distinct coding schemes. IPUMS ATUS users have requested that we extend our occupation harmonization work to IPUMS ATUS to bridge changes in occupation coding schemes. The wish of those users has been granted! OCC2010 is now available via IPUMS ATUS. As always, the original, un-recoded occupation categories are also available in OCC.
Why do we need to harmonize occupations?
The Census Bureau regularly updates occupation and industry codes to reflect the rapidly changing nature of employment in the United States. Changes to occupation codes are important because they provide the most up to date information about jobs and the nature of work. Updates to occupation codes reflect the evolution of jobs and work and allow for the incorporation of new occupations. While updating occupation classification schemes improves the quality of occupation data, navigating these changes can be cumbersome for researchers.
What is OCC2010?
OCC2010 is created by IPUMS to implement consistent occupation codes across all years of currently available ATUS data (2003 to 2024). This harmonized occupation variable assigns all original occupation codes in the ATUS (i.e., OCC) to a single Census Bureau coding scheme. We chose the 2010 occupation coding scheme (well, actually the 2012 coding scheme which is very similar to the 2010 scheme1) because then we only have to bridge one change in occupation codes forward in time and backwards in time. OCC2010 was developed to enhance the comparability of occupational data by providing a consistent set of occupational codes for IPUMS ATUS from 2003 forward. OCC2010 is also available in IPUMS CPS and IPUMS USA.
How was OCC2010 created?
Census Bureau Crosswalks
When occupation coding schemes change, the Census Bureau publishes crosswalks describing how the old and new occupation coding schemes relate to one another. Sometimes the occupation title remains unchanged, but the numeric code assigned to that occupation changes. Other times, an occupation is split into several new occupations, or multiple occupations in the old scheme are combined into a single occupation in the new coding scheme. Here are the crosswalks we use for the creation of OCC2010.
- 2003-2010 ATUS: 2002 to 2010 Census occupation code mapping
- 2011-2019 ATUS: 2010 (some 2010 Census occupations are combined into 2012 Census occupation codes).
- 2020-forward ATUS: 2018 to 2010 Census occupation code mapping
The 2002 (derived from the 2000 SOC) occupation coding scheme is used for the 2003 to 2010 ATUS data, so we use the 2002 to 2010 crosswalk to identify the appropriate 2010 occupation code, reported in OCC2010, for these years. For ATUS samples that natively use the 2010 occupation codes (2011-2019), the unharmonized OCC codes will be nearly identical to OCC2010 in these years. However, we used an adapted version of the 2010 occupation codes (implemented in the 2012 ACS) to improve comparability across IPUMS projects (OCC2010 is available in IPUMS CPS and IPUMS USA). For these years, any differences between OCC and OCC2010 can be mapped using the 2010 to 2012 occupation code map. Occupations in the 2020 to 2024 ATUS data are coded using the 2018 occupation coding scheme, so we use the 2018 to 2010 occupation crosswalk to create OCC2010.
Modal Assignment
Our approach to creating OCC2010 entails using the published Census occupation crosswalks to map occupation codes from surrounding occupation coding schemes into the base 2010 coding scheme. We largely do this using modal assignment. Census Bureau crosswalks describe the percentage of a given occupation that splits into more specific occupations in the adjacent coding scheme and those that combine into a single, more general occupation. We describe this “modal assignment” for mapping the 2018 Census Bureau coding scheme back to the 2010 coding scheme (Williams and Flood, 2020).
For example, the 2002 Census occupation code 0620 (Human Resources, Training, and Labor Relations Specialists) splits into three 2012 Census occupation codes 0630 (Human Resources Workers), 0640 (Compensation, Benefits, and Job Analysis Specialists), and 0650 (Training and Development Specialists). According to the Census Bureau crosswalk, 73% of the 2002 occupation category Human Resources, Training, and Labor Relations Specialists would be classified as Human Resource workers in the 2012 Census occupation coding scheme. Thus, IPUMS recodes the 2002 code 0620 to the 2012 code 0630 in the variable OCC2010. In other words, all cases assigned as 0620 under in the OCC variable in 2003-2010 will be coded as 0630 in OCC2010.
| Human Resources, Training and Labor Relations Specialists | ||||||
|---|---|---|---|---|---|---|
| 2002 OCC Code | 2010 OCC Code | Conversion Rate | OCC2010 Assignment Rate | |||
| 0620 | Human Resources, Training, and Labor Relations Specialists | → | 0630 | Human resources workers | 0.7385 | 100% |
| 0640 | Compensation, benefits, and job analysis specialists | 0.1244 | 0% | |||
| 0650 | Training and development specialists | 0.1371 | 0% | |||
Limitations of OCC2010
OCC2010 isn’t perfect. In an ideal world, we would be able to identify the individuals who were coded as Human Resources, Training, and Labor Relations Specialists in the 2002 census occupation codes and precisely assign them one of the corresponding 2010 codes (0630, 0640, or 0650). Instead, the best we can do is assign everyone to 0630. OCC2010 represents the maximum level of comparability while still preserving as much of the detail in occupation codes over the 2003 to 2024 time period as possible. Further collapsing occupation categories is one strategy you can use to further increase the comparability of occupation codes over time.
Footnotes:
- In the 2012 ACS, the Census Bureau introduced an adapted occupational coding scheme which simply grouped a few small occupations in the 2010 occupational coding scheme together. The CDC provides a crosswalk for occupation codes between 2010 and 2012. For ATUS samples that natively use the 2010 occupation codes (2011-2019), the unharmonized OCC codes will be nearly identical to OCC2010 in these years. However, we used an adapted version of the 2010 occupation codes (implemented in the 2012 ACS). For these years, any differences between OCC and OCC2010 can be mapped using the 2010 to 2012 occupation code map.