Factor Degree Definitions: How to Write Scales That Score Consistently

Factor Degree Definitions: How to Write Scales That Score Consistently

Most point-factor systems fail in the same quiet way. The factors are sensible, the weights were argued over for weeks, and the math is clean — but two evaluators score the same job three degrees apart on "Problem Solving," and nobody can say who is right. That is not a factor problem or a weighting problem. It is a degree definition problem.

Degree definitions are the written descriptions of each level inside a compensable factor. They are the only part of your system that does actual measuring. Everything else just processes the number a degree definition produced. Write them loosely and you have built a very precise machine on top of a guess. Here is how to write degree definitions that hold up in a committee, in an appeal, and in front of a regulator.

TL;DR

• Degree definitions are the level descriptions inside each sub-factor. They carry all the measurement work in a point-factor system.
• Use 4–6 degrees for most sub-factors. Fewer than 4 can't separate jobs; more than 6 invites false precision and rater drift.
• Anchor every degree in observable job content — what the job requires, not who holds it or how well they do it.
• Give each degree a real-job example from your own organization. Examples resolve more disputes than definitions do.
• Space points geometrically, not linearly, when the underlying difference between levels grows (OPM's federal system runs 50 points at level 1 to 1,850 at level 9 on one factor).
• Test the scale before you deploy it: two evaluators, ten jobs, blind. If agreement is under 80%, the definitions are too vague.

What a degree definition actually is

The terms get muddled, so start with the hierarchy. A compensable factor is a broad dimension of job worth — skill, effort, responsibility, working conditions. Each factor breaks into sub-factors like "Formal Education," "Problem Solving," or "Budget Accountability." Each sub-factor is scored on a scale of degrees (also called levels), and each degree has a definition and a point value.

A single scoring decision looks like this:

The evaluator's whole job is to read that definition and decide whether the job matches it. If the sentence is fuzzy, the score is fuzzy. If you want more background on choosing the dimensions themselves, start with compensable factors; this article assumes you have already picked them.

Rule 1: Pick the right number of degrees

Four to six degrees works for most sub-factors. With three degrees you get pile-up: most jobs land on the middle level and the sub-factor stops discriminating — it adds points to everything and separates nothing. With eight or nine, evaluators cannot reliably tell degree 6 from degree 7, so scores wander based on who is in the room that day.

One legitimate exception: sub-factors that span an enormous real range deserve more levels. Formal education genuinely runs from "basic literacy" to "doctorate plus board certification" — 6 to 8 defensible steps. The U.S. Office of Personnel Management's Factor Evaluation System uses nine levels on its "Knowledge Required by the Position" factor for exactly this reason, and that factor carries far more point spread than any other in the system (OPM Classifier's Handbook).

Rule of thumb: add a degree only when you can name two real jobs in your organization that clearly belong on different sides of the new boundary. If you cannot, the degree is decorative.

Rule 2: Describe the job, never the person

This is the single most common defect, and it is the one that creates legal exposure. Compare:

Weak: "Requires a highly experienced professional who is a strong communicator."

Strong: "Requires explaining technical findings to non-technical audiences and negotiating scope with internal stakeholders. Errors are typically caught within the same work cycle."

The first sentence scores a person. The second scores work content. Point-factor evaluation measures the job as designed, not the incumbent's performance or tenure — that is what separates job evaluation from performance evaluation, and conflating them is how "we've always paid Dana more" gets laundered into a point score.

It also matters under equal pay law, which compares jobs on equal skill, effort, responsibility, and working conditions performed under similar conditions (EEOC compensation discrimination guidance). A degree definition anchored in incumbent traits gives you nothing to point at when someone asks why two jobs of equal value scored differently.

Rule 3: Every degree needs an observable anchor

A good degree definition contains at least one thing an evaluator can verify from a job description or a job analysis interview. Useful anchor types:

• Scope: number of direct reports, headcount influenced, size of budget owned, geography covered.
• Autonomy: who reviews the work, how often, and at what stage.
• Consequence of error: what breaks, who notices, and how long it takes to correct.
• Nature of the problem: routine and precedented, adapted from precedent, or genuinely novel.
• Audience: who the job persuades, informs, or negotiates with.

Here is a five-degree "Decision-Making Authority" scale built on autonomy and consequence:

Notice that each degree changes on the same three dimensions, in the same order, every time. That parallel structure is what makes a scale scoreable. When degree 2 talks about review frequency and degree 3 talks about job titles, evaluators lose the thread and start scoring on vibes.

Rule 4: Space the points to match the real gap

Degree points rarely should be evenly spaced. In the table above, the gaps run 25, 35, 50, 65 — widening as you climb. That reflects reality: the difference between a job that follows procedures and one that chooses among options is smaller than the difference between running a function and setting policy for a business unit.

Two common approaches:

• Arithmetic (equal steps): 20, 40, 60, 80, 100. Fine where each level is a genuinely equal increment, such as physical effort or a simple education ladder.
• Geometric (widening steps): 20, 35, 60, 105, 180. Better for responsibility, problem solving, and impact — anything where the top of the scale is qualitatively different from the middle, not just "more of it."

OPM's federal system is unapologetically geometric: 50 points at the first knowledge level, 1,250 at level 7, 1,850 at level 9. If your scale runs 20/40/60/80/100 on every factor, your executive jobs and your senior analyst jobs will land closer together than the market says they should, and you will spend the next year explaining why.

Spacing interacts with weighting, so do them together rather than in sequence — the mechanics are in how to weight compensable factors.

Building this from scratch? PointFactors ships with pre-written degree definitions across a full factor set, so you can start from tested language and adapt it instead of drafting 40 level descriptions on a blank page. See how it works.

Rule 5: Attach a real job to every degree

This is the highest-leverage step and the one teams skip. For each degree, name one job in your own organization that sits squarely on that level. Not a hypothetical — an actual job with an actual job description.

Benchmark examples do three things. They make the definition concrete for evaluators who are new to the system. They force you to discover empty degrees (if no job in a 900-person company sits at degree 4, that degree probably does not belong). And they give your job evaluation committee a shared reference point, so arguments become "is this job above or below Payroll Specialist?" instead of "what does 'moderate complexity' mean to you?"

Pick examples from different job families so the scale does not quietly become an engineering ladder.

Rule 6: Write for gender neutrality on purpose

Degree definitions are where gender bias enters a job evaluation system, and it usually enters through omission rather than intent. The classic pattern: physical effort gets a detailed 5-degree scale because someone remembered warehouse work, while emotional effort, caregiving demands, and sustained concentration get no scale at all. Female-dominated roles then score systematically low on "effort" — not because the work is easier, but because the demands were never written down.

The EU's guidance is explicit on this. The European Institute for Gender Equality's step-by-step toolkit on gender-neutral job evaluation breaks skill, responsibility, effort, and working conditions into sub-factors and asks employers to define levels for each in plain, observable language — including the sub-factors traditional systems leave out. If you operate in the EU under the Pay Transparency Directive, that toolkit is close to the expected standard; see our walkthrough of gender-neutral job evaluation.

Two checks before you sign off a scale:

1. Read every degree definition and ask whether the language leans toward one type of work. "Manages a team" and "coordinates across teams without authority" are both real responsibility — make sure both are scoreable.
1. Score a set of jobs that includes your most male-dominated and most female-dominated roles. If the pattern is stark, find which sub-factors are producing it before you touch the weights.

Rule 7: Test the scale before you deploy it

Draft definitions are hypotheses. Pick 10 jobs spanning your full range and at least three job families, have two trained evaluators score them independently, then compare sub-factor by sub-factor.

Target agreement: same degree on 80%+ of scores, within one degree on 95%+. Anything below that points at a specific sub-factor, not the system as a whole. Find the sub-factor with the worst agreement, read its definitions out loud, and you will usually spot the vague word in about thirty seconds. "Complex," "significant," "advanced," and "as needed" are the usual suspects.

Then rewrite and retest. Two rounds is normal. Three means the sub-factor is measuring something you have not defined clearly enough to measure.

Frequently asked questions

How many degrees should each sub-factor have? Four to six for most. Six to eight is defensible for formal education or knowledge, where the real range is genuinely wide. Fewer than four rarely discriminates between jobs.

Do all sub-factors need the same number of degrees? No. Uniformity looks tidy but forces you to invent levels that do not exist. Match the number of degrees to the real spread of the work, and let the weights and point ranges handle comparability.

Should degree definitions mention job titles? Not inside the definition itself. Keep the definition about work content, then list benchmark job examples alongside it. That way, when a job changes, you update the example without reopening the scale.

What is the difference between a degree and a grade? A degree is a level within one sub-factor. A grade is a pay band that a job lands in after all its factor scores are totaled and converted. One job has many degree scores and exactly one grade. The conversion is covered in converting job evaluation points into pay grades.

How often should we rewrite degree definitions? Rarely. Rewriting the scale invalidates every score produced under the old version, which means re-evaluating your whole job catalog. Review every three to five years, or when the nature of work has changed enough that jobs are consistently landing at the top or bottom of a scale.

What if evaluators keep disagreeing on one sub-factor? That sub-factor is probably measuring two things at once. "Complexity" often turns out to be problem novelty plus stakeholder difficulty. Split it, define both, and the disagreement usually disappears.

The bottom line

Degree definitions are unglamorous work. They are also the only part of a point-factor system that touches reality — every other component is arithmetic performed on the numbers your definitions produce. A well-weighted system with sloppy degree definitions gives you precise, repeatable, defensible-looking answers to a question nobody actually measured.

Spend the extra week. Write in observable terms, anchor each level to a real job, space your points to match real differences, and test with two evaluators before you go near your job catalog. Then the rest of the point-factor method does what it promises. If you want a working starting point, our free point-factor scorecard includes degree definitions you can adapt.

Ready to stop drafting scales from scratch? PointFactors gives you a validated factor set with pre-written degree definitions, consistency checks that flag scores drifting from your benchmarks, and an audit trail for every evaluation decision. Book a demo and see your own jobs scored in under an hour.

Justin Hampton is founder and CEO of PointFactors.