Jobs by Holland type: five decades of US census data on the RIASEC areas

Holland’s theory sorts people’s interests into six types, Realistic, Investigative, Artistic, Social, Enterprising and Conventional, and sorts occupations into the same six (Holland, 1997). That makes it possible to ask, in the same terms, how much work there is that fits your interest type.

One study coded the occupations in the US census by Holland type and added up the people employed in each, over five decades from 1960 to 2010 (McClain and Reardon, 2015). In short, the area where the most Americans worked in 2010 was Realistic (27%), followed by Social (24%) and Enterprising (20%), while Artistic was the smallest at 2%.

This guide uses that study’s figures to set out the share of jobs in each type, how it changed over 50 years and the mean income in each, then explains what these numbers do and do not say about your own interest type, and how to read them next to a result on the mind.io RIASEC interest test.

For the 494 detailed occupations in the 2010 census (military occupations excluded), the study took the first letter of each occupation’s Holland code, took employment for each occupation from O*NET and added it up by area. Of 119.8 million people employed, Realistic jobs were the largest share at 27% (31.9 million), followed by Social at 24%, Enterprising at 20%, Conventional at 17%, Investigative at 10% and Artistic at 2% (2.0 million).

The number of occupation titles was uneven too. Of the 494 detailed occupations in the 2010 census, 211 (43%) were Realistic, while only 19 (4%) were Artistic. Artistic work accounts for few jobs and is also spread over only a few named occupations. Averaged over the six censuses since 1960, the titles split much as in 2010, 46% Realistic, 18% Enterprising, 12% each Investigative and Social, 9% Conventional and 3% Artistic, so the scheme for naming work changed little in 50 years.

In 1960, 55% of US jobs were Realistic. By 2010 the share had fallen to 27%, yet Realistic remained the largest area in all six censuses. Over the same period Social rose from 9% to 24% (from 5.6 million to 29.6 million people) and Investigative from 3% to 10% (from 2.0 million to 11.5 million). Artistic was the smallest area in every census, at 1% to 2%.

Total employment grew from 64.1 million to 119.8 million. Enterprising climbed to 30% in 2000 (level with Realistic at 30%, but with slightly fewer people) and stood at 20% in 2010, while Conventional stayed almost unchanged between 15% and 17%. The authors suggest that the much smaller gap between Realistic and Artistic in 2010 than in 1960 may be a sign of the decline in manufacturing.

The study also compared the mean annual income of occupations by area in 2010. Investigative was highest at $73,670, followed by Enterprising at $53,880, Artistic at $52,380, Social at $41,060, Realistic at $35,200 and Conventional at $34,740. The Investigative mean was slightly more than double the Conventional one.

Investigative also had the highest mean income in 1990 and 2000. The authors read this as evidence that education pays, since Investigative work requires more education and training than the other areas. They add, though, that they used means rather than medians and that the way income was calculated differed across the decades, so comparisons between years call for caution.

These are shares of jobs, not of people’s interest types. Each occupation was given only the first letter of its Holland code, so work that mixes, say, Social and Artistic was counted under one of them only. The authors also note that sorting hundreds of thousands of jobs into 350 to 500 occupational categories takes considerable judgment by occupational analysts, so some may be misclassified.

The data are also American, and Korea’s mix of industries may differ. We have not found a Korean analysis of the same kind whose source and method can be checked. So the table is better used for the broad picture, that the amount of work is uneven across the types, than as a guide to which kinds of work are common in Korea.

Holland advised remembering, when interpreting a result, that types and jobs are not evenly spread across the six areas. This study bears that out: an area with few jobs may also have fewer positions and fewer openings (McClain and Reardon, 2015).

That is no reason to drop an interest. Holland codes are usually read as two or three letters, and work with the same first letter differs with the second and third. If your first letter is Artistic, look through the two-letter codes for work that combines it with another area, and think about ways to use the interest outside work as well.

The mind.io RIASEC interest test scores the six interest areas from 42 questions and shows a three-letter code in order of your scores. The study above describes jobs while the test describes the order of your interests, so the two cannot be set side by side in one table.

The Holland interest type ranking anonymously counts only the first letter of the code in the results of people who tap the join button on their own result page. All that is sent is the test, the result and the page language; no answers or scores. The numbers count entries, not people, and the ranking stays hidden until 30 entries are in.

Q. Which Holland type has the most jobs? A. In the analysis of the 2010 US census, Realistic was the largest at 27%, followed by Social at 24% and Enterprising at 20% (McClain and Reardon, 2015). Artistic was the smallest at 2%.

Q. Is it hard to find work as an Artistic type? A. Artistic was the smallest area in this analysis, at 2% of jobs. These figures count only the first letter of each code, though, so work that uses Artistic interests alongside another area may have been counted under the other area. Look at your second and third letters as well when exploring work.

Q. Which type earns the most? A. In 2010, the mean annual income was highest for Investigative ($73,670) and lowest for Conventional ($34,740). These are averages from US data, so they cannot predict any one person’s pay.

Q. Why does the mind.io ranking differ? A. This study counted jobs, while the mind.io ranking counts the first letter of the interest codes of people who chose to join. They count different things to begin with.

References: Holland (1997), Making Vocational Choices, Psychological Assessment Resources; McClain & Reardon (2015), The Professional Counselor 5(1), 1-14, https://doi.org/10.15241/rcr.5.1.1.