Why Doesn't Your Self-Rating Match the Evidence of Your Progress?

Why Doesn't Your Self-Rating Match the Evidence of Your Progress?

They disagree because they measure different things. A self-rating captures how a skill felt to use — effort, confidence, strain. Measured evidence captures what happened — a recording, a result, a number someone else can check. The two drift apart for ordinary reasons, and most of the gap is not a mystery to solve but a pattern to expect.

What counts as "measured evidence"?

Evidence is anything a third party, a recording, or a stable instrument could confirm without asking you how you felt. A call recording scored against a rubric. A split time. A client's written feedback. A reviewer's rating of a piece of code. A self-rating, by contrast, is your own number on your own scale, filled in from memory.

Neither is wrong. They just answer different questions: "how did this feel" versus "what can be shown".

Why do self-ratings and evidence disagree?

A few recurring reasons:

1. Hard days feel like bad days. Diary research on workplace progress found that people's sense of a good day tracked small wins and setbacks closely, sometimes more than the actual scale of what was achieved (Amabile and Kramer, 2011). A difficult but productive week can produce a lower self-rating than an easy, unproductive one.
2. Memory is short and recent. A self-rating filled in at a check-in leans on the last few days, not the whole period.
3. The scale shifts as you learn. Early in learning a skill, "good" means something different than it does three months later. This is sometimes called response shift: the yardstick itself moves, so a flat self-rating can hide real gains, and a dropping one does not always mean decline.
4. Evidence has its own noise. A single recorded call, a single race, a single client email is one data point, not a trend.
5. Specific, trackable goals reduce (but don't remove) the gap. Goal-setting research has long shown that specific, measurable targets produce better performance and clearer feedback than vague ones (Locke and Latham, 2002) — which narrows the gap between felt and measured progress, but does not close it.

How much of the difference is real?

This is the part most development plans skip. Any measurement — a rubric score, a self-rating, a lap time — has a margin of error. In rehabilitation and clinical research this is formalised as the "minimal detectable change": the smallest change that is larger than the measurement's own noise and can be called real rather than random (see the discussion of minimal detectable change in outcome measures). Workplace self-assessment rarely has that level of rigour, but the principle transfers directly: a one-point shift on a five-point self-rating scale, filled in from memory three minutes before a meeting, is close to meaningless. A two-point shift sustained across three check-ins is worth attention.

A worked comparison

Three examples make the pattern concrete. A founder negotiating term sheets might rate their own negotiating skill as flat for two quarters, while the actual term sheets show steadily better terms — the skill moved, the self-rating lagged because the work still felt hard. A professional athlete's self-rated recovery can dip during a heavy training block even as objective markers hold steady, because fatigue and progress often arrive together. A performing artist reviewing their own recorded sets may rate a strong night as average, simply because the hardest part of the set is what stays in memory — a pattern covered in our piece on development plans for performing artists.

What should a check-in actually record?

Record the self-rating and the evidence separately, without forcing them to agree. Note the gap itself as a data point. If the gap persists over two or three cycles, use it to recalibrate the scale rather than the goal. This fits a broader point worth a longer read: in measuring professional growth honestly, "held steady" is treated as a legitimate result rather than a failure, and the same discipline applies to reconciling what you felt against what you can show. For guidance on how often to even ask the question, see how often you should reassess a development goal.

Where this fits in a development plan

This is one reason Jobulary separates reflection from measurement in its check-ins, which take about three minutes: growth is only reported when it is larger than the measurement error on that particular goal, otherwise the plan records "held steady" — not as a verdict, but as an honest reading.

A plan built this way also respects who owns the judgement. On Jobulary, the plan belongs to the person, not the employer; sponsors never see reflections or check-in notes, only the goals a person chooses to make visible. That matters here specifically, because self-ratings are the part most likely to feel exposing, and the part least useful if softened for an audience.