Tuesday, October 22, 2019

Job Browser Pro 1.7 -- the Good, the Bad, and the Ugly

I installed Job Browser Pro, ver. 1.7, yesterday and took it out for a test drive.  If anyone represents claimants and listen to vocational experts testify, then JBP is a mandatory part of the representative's library, period.  Vocational experts will cite to it, rely on it, and the ALJ corps will accept it.  If the representative wins one case every 20 years because the representative owns and uses JBP, the program pays for itself.

1.  JBP Methodology

JBP uses the matrix approach found in the Occupational Employment Statistics and the Employment Projections published by the Bureau of Labor Statistics.  Occupations exist in industries.  The DOT recognizes this with the parenthetical industry designation that is party of every occupational description.  According to the DOT, the industry designation is an integral part of every description.  When a vocational expert identifies an occupation and then uses the entire occupational group number for that occupation (and it isn't one of occupations that is the sole inhabitant of that group), the vocational expert ignores and conflicts with the industry designation contained in the DOT, usually without an explanation.  More about the methodology in the Bad section.

2. The Good

Users can no longer add and delete industries.  Vocational experts dissatisfied with the low job numbers reported could and would add industries.  This was a particular problem in prior versions of JBP because the vocational experts would add inappropriate industries or add industries where JBP had already listed the two, three, four, or five digit designation resulting in double or triple counting.

 JBP has added the CBP button to the occupation-industry matrix inside of the box.  County Business Patterns is a matter of administrative notice by regulation, listed second after the DOT.  The link sends uses to a SkillTran page for that industry.  The web page lists the total number of establishments in the industry (sector (two-digit), sub-sector (three-digits), industry group (four-digits), or industry (five- and six-digits)), number of employees, and other statistical data that we do not use.

Jeff Truthan tells me that if the user looks up all the DOT codes within a group and count all the reported job numbers, user will get roughly the number of jobs reported within the OES group.  Critics will no longer be able to complain that JBP has under-counted the number of jobs.

3.  The Bad

JBP continues to use equal distribution  to distribute the number of jobs at the occupation-industry intersection.  When I looked up production workers, all other (SOC 51-9199) at the intersection with animal food manufacturing (NAICS 311100), I found that JBP lists five occupations, each representing 20% of the total jobs, four requiring no skills, and one semi-skilled occupation.  All of them are light.  We know from the Occupational Requirements Survey that most of the production worker jobs are medium and 45% are semi-skilled or skilled.  The equal distribution methodology is not good statistics at the occupational group level and is not good statistics at the occupation-industry intersection level.

While JBP continues to allow users to see the list of DOT codes at the occupation-industry intersection, it still does not permit users to print out the Estimated National Distribution of DOT Employment of that occupation-industry intersection.  This has left me with the tedious chore of taking a screen shot of that popup and if the screen lists more than 13 DOT codes, multiple screen shots.

4. The Ugly

Some of the industry choices are just wrong.  I looked up the oft-cited small products assembler I (DOT 706.584-022) to check.  SPA work on assembly lines to mass product small products.  That is the DOT narrative.  Because the occupation exists in more than four DOT industries, it carries the designation of "any industry."  But that does not mean every industry, it means more than four.

JBP lists eight industry groups within the food manufacturing sub-sector and 22 other industries that have nothing to do with the functions to "mass produce small products, such as ball bearings, automobile door locking units, speedometers, condensers, distributors, ignition coils, drafting table subassemblies, or carburetors."

I checked lens inserter (DOT 713.587-026).  JBP lists this occupation in the jewelry and silverware manufacturing industry.  The DOT puts lens inserter in the optical goods industry.  JBP puts the other 19 production worker occupations that are designated in the optical goods industry into medical equipment and supplies manufacturing, which contains the ophthalmic goods manufacturing industry (with some of the 20 adding to that list).  I understand why JBP puts lens inserter in clearly the wrong industry -- it is the only DOT code assigned.

The production workers that do carry the jewelry and silverware industry designation in the DOT have no industries assigned by JBP.  Stringer (DOT 509.587-018) is "N/A."  Stamper (DOT 734.685-010) in the button and notions industry is given the other miscellaneous manufacturing industry designation.  The problem is that jewelry and silverware is part of other miscellaneous manufacturing, resulting in double counting or over-estimating the number of jobs across DOT codes.

The occupation-industry intersection approach to job number estimates is still valuable for finding those occupations that have rare characteristics -- reasoning level 1, occasional use of the hands, etc. -- but JBP still has holes.  This creates opportunities for the representative to dampen the number of jobs through a reliable methodology, don't double count or explain why the assembler of small products is making dog food.  Proper cross-examination will require that vocational experts come to the hearing with the occupation report from JBP and allow the representative to see it, or the ALJ will have to accept the report from the representative post-hearing.

The job of the representative just got easier and harder at the same time.

No comments:

Post a Comment