By Rovaryn Digital · May 14, 2026 · 7 min read
The Report That Fell Apart on Cross-Examination
A claimant's attorney once asked a vocational expert, on the record, why a job on her occupation list required lifting fifty pounds when the treating physician had capped the claimant at twenty. The expert had built the list by keyword-matching job titles against O*NET descriptions, then eyeballed the physical demands. She hadn't filtered systematically by exertion level, and she had no documented threshold to point to. The list didn't survive. The report went back for revision, the deposition took twice as long as scheduled, and the referral source noticed.
The fix isn't more software or more O*NET tabs open at once. It's sequence and documentation. Exertion level, SVP, and skill level are three separate filters, each drawn from a different part of the claimant's file, and each one has to be applied — and recorded — in a way that another reader can reconstruct without asking you why an occupation dropped off the list. By the end of this piece, you'll be able to apply all three filters in the right order and document the thresholds so the resulting list holds up when someone else reviews it.
Why Filtering Order Matters Before You Ever Open O*NET
It's tempting to start broad — pull every occupation that shares a skill or two with the claimant's work history — and then whittle down. That approach produces long lists full of occupations that never had a chance, because they fail on physical demands or training time before skill relevance even matters. Working in the wrong order also makes the documentation harder to defend: if a reviewer can't see which restriction eliminated which occupation, the list looks arbitrary even when it isn't.
The more defensible sequence starts with the hardest, most objective constraint and narrows from there:
- Exertion level — eliminated by physical restrictions in the medical record.
- SVP — eliminated by the claimant's realistic training/learning capacity and time horizon.
- Skill level — narrowed by transferable skills, education, and aptitude.
Each filter should be applied to what survived the previous one, not to the full universe of occupations each time. That's also the order a transferable skills analysis generally follows once the base occupation set is built from the O*NET-to-DOT crosswalk.
The Exertion Filter: Matching Physical Demand Levels to Documented Restrictions
The Dictionary of Occupational Titles classifies every occupation into one of five exertion categories — Sedentary, Light, Medium, Heavy, and Very Heavy — based on the strength demands of lifting, carrying, pushing, and pulling, along with standing and walking requirements. This is the filter with the clearest paper trail, because it maps directly to language that already exists in the medical file: a treating physician's functional capacity evaluation, an independent medical exam, or a permanent work restriction typically states a lifting limit and a postural tolerance in terms that translate cleanly onto the DOT scale.
The discipline here is refusing to round in the claimant's disfavor. If the medical record caps lifting at twenty pounds occasionally, that claimant does not tolerate Medium work, even though Medium's lower bound sits close to that number — Medium is defined by exertion up to fifty pounds. Document the specific restriction language you used and the exertion category it excludes, so the exclusion is traceable back to a specific page in the file rather than to your judgment alone.
Where the file has conflicting restrictions — a treating physician and an IME reaching different conclusions — the filter itself doesn't resolve the conflict. That's a case-management decision, and it belongs in the report narrative, not buried in a spreadsheet column.
The SVP Filter: How Long Can This Claimant Realistically Retrain
Specific Vocational Preparation measures the amount of time typically required to learn the techniques, acquire the information, and develop the facility needed for average performance in a job — not the time to become an expert, and not formal education alone. It's expressed on a scale from SVP 1 (short demonstration only) up through SVP 9 (over ten years). Two occupations can carry the same physical demand level and still sit far apart on SVP, and that gap is often what separates a realistic reemployment target from an aspirational one.
The SVP filter draws on a different part of the file than the exertion filter does: education records, work history duration in prior roles, and any vocational testing on the claimant's learning aptitude. A claimant who spent fifteen years in a single SVP 2 role has a documented pattern that argues against assuming rapid mastery of an SVP 6 occupation, regardless of what the exertion and skill filters would otherwise allow through. For the full mechanics of how SVP is scored and where it appears on an occupation's crosswalked profile, see our explainer on SVP levels.
Jurisdictions vary in how much weight they place on SVP thresholds when defining "reasonable" retraining time in a wage-earning-capacity or vocational feasibility opinion — confirm the standard your forum applies rather than assuming a fixed cutoff carries over from another state or province.
The Skill Level Filter: Reasoning, Math, and Language Requirements
The skill level filter is where the transferable skills analysis actually earns its name. Once exertion and SVP have trimmed the occupation set to jobs the claimant can physically perform and plausibly learn in a realistic time frame, the remaining question is whether the claimant's existing work-acquired skills — not just credentials — carry over. This is measured through the General Educational Development scale embedded in each DOT occupation: separate ratings for reasoning development, mathematical development, and language development, each running from level 1 through 6.
This filter is the most interpretive of the three, because it requires mapping the claimant's actual job duties — often described loosely in a work history interview — onto the more structured skill and task language used in an O*NET-to-DOT crosswalk. A claimant who "kept the books" for a small business may have functioned at a higher mathematical development level than her job title suggests, or lower, depending on what she actually did day to day. Document the specific duties that justify the skill level you assigned — this is the part of the analysis most likely to be challenged, and the one most dependent on interview quality rather than database lookup.
Stacking the Three Filters Without Losing the Trail
Applied correctly, the three filters produce a shrinking, documented sequence rather than a single unexplained final list. A clean working structure looks like this:
- Start with the full crosswalked occupation set relevant to the claimant's industry and geography.
- Apply the exertion filter; record the restriction and the exclusion count.
- Apply the SVP filter to the survivors; record the training-time rationale.
- Apply the skill level filter to what remains; record the duty-to-skill mapping.
The habit that separates a defensible list from a vulnerable one isn't the filtering logic itself — it's keeping the intermediate list at each stage instead of only the final output. When a report is challenged months or years later, the ability to show why an occupation dropped at the exertion stage rather than the skill stage is often the difference between a five-minute clarification and a lost afternoon in deposition.
Documenting Thresholds So the List Survives Scrutiny
Every threshold used in the filtering process should trace back to a source in the claimant's file or a named classification standard — never to an unstated assumption. That means citing the specific restriction language for the exertion cutoff, the specific work-history or testing basis for the SVP cutoff, and the specific duties that support each skill level assignment. A reviewer, opposing counsel, or a hearing officer should be able to reconstruct the filter without asking you to explain your reasoning from memory.
This is also where a standardized crosswalk reference pays for itself: rather than re-deriving exertion, SVP, and GED ratings from scratch for every occupation on a growing list, working from a single consistent reference reduces the chance that two occupations get scored on slightly different assumptions. Our O*NET/DOT Occupation Crosswalk Reference Workbook is built for exactly this — a working reference you can pull from case to case so the exertion, SVP, and skill level values you cite stay consistent across a caseload, not just within a single report.
For the full walkthrough of where this filtering step fits inside a complete transferable skills analysis, start with our transferable skills analysis resource hub and work through the step-by-step TSA methodology alongside it.
Download the reference workbook, apply it to your next occupation list, and you'll have a filtering trail that doesn't fall apart the first time someone asks why an occupation made the cut.