On November 19, 2007, the Assistant Commissioner of the Office of Occupational Statistics and Employment Projections (Dixie Sommers) wrote a response to an attorney about several issues:
- Utility of the Dictionary of Occupational Titles (DOT)
- Data source or method to estimate the numbers of jobs by DOT code.
In her reply (paragraph 2), she stated that "The DOT is no longer in use by the Bureau of Labor Statistics, and we do regard it as obsolete since much of the information contained in the most recent version is based on research conducted at least two decades ago." In paragraph 5, she also stated "We are not aware of any data source or methodology for reliably determining the number of jobs by DOT code." This letter has been cited in a considerable number of Social Security disability claims in attempts to discredit both the use of the DOT in the claims decision process and the use of Job Browser Pro methodology for estimating job numbers to the DOT level.
On the first point, note that the the BLS comment is only in regards to its use for the purposes of the BLS (implied). However, the Social Security Administation (SSA) recognizes and continues to use the DOT as an authoritative resource in the disability claims process. In fact, SSA has expressly rejected O*NET as a suitable replacement for the DOT for a number of important reasons. SSA is currently funding an effort to collect information at the SOC level (the Occupational Requirement Survey), but useful data from this effort was not available until 2019. Many issues remain to be resolved in the process of migrating to admittedly more current, but far more highly aggregated data. SkillTRAN is carefully monitoring this process and has integrated both the first wave of ORS data (2016-2018) and the second wave (2019-2023) into its products.
Lending further support to continued rejection of the DOT, the O*NET program in early 2012 rejected SkillTRAN's suggestions for a substantial number of important changes in the SOC 2000 crosswalk from the DOT to the newer SOC 2010 occupational classifications. In their rejection of our suggestions, they stated: "Developers are free to develop their own crosswalks using other processes." Here is a link to the text of the email of suggestions by SkillTRAN and O*NET rejection of same.
On the second point, Ms. Sommers overlooked the fact that for 182 of the 820 civilian SOC groups (17.4%), there is but one DOT in each of those 182 SOC groups. So in effect, the OES program does in fact collect data at the DOT level for those 182 DOT occupations. Click here for a list of them. Of course, that leaves 677 SOC/OES groups with 12,618 DOT occupations (average is 18.6 DOT codes per SOC group)! This is a perplexing issue, for which there was no suitable solution in 2007. SkillTRAN introduced its industry-context methodology for estimating employment numbers at the DOT level in June, 2008.