The 1–5 interview rating scale, with anchors written for every point

A 1–5 interview rating scale only works if every one of its five points means the same thing to every interviewer. Most teams that adopt "1 to 5" stop there and never write down what a 2 looks like versus a 3, so the scale becomes five ways of saying "I liked them" or "I didn't." Below is the full scale: five points, each with a written anchor, applied to five common competencies, plus the decision rules that keep a 3 from becoming the default answer.

This page assumes you have already decided to use five points. If you are still choosing between four, five, seven or ten, see interview rating scale for that decision; this page goes one level deeper into the scale most teams land on.

What each of the five points means, in general

Before anchoring individual competencies, interviewers need a shared definition of the five levels themselves. Write this once and put it at the top of every scorecard.

Two rules make this table usable instead of decorative. First, 3 is a genuine pass, not a midpoint interviewers pick when they are unsure — unsure evidence is a 2. Second, 5 should be rare; if a third of your scores are 5s, the anchor for 5 has drifted down to what 4 used to mean.

The scale applied to five competencies

A shared definition of the five levels is necessary but not sufficient. Interviewers also need the levels translated into what they would actually hear a candidate say, for each competency on the scorecard. Copy the rows below for the competencies your loop already uses, and write new ones the same way for anything not covered here.

Ownership

Communication

Problem solving

Collaboration

Domain or role-specific skill

Replace "domain or role-specific skill" with the actual skill your loop tests — a SQL query, a sales role-play, a med-surg scenario. The five-level shape stays the same: absent, thin, solid, solid-plus-something, and rare.

Writing an anchor for a competency not listed above

Use this order and each anchor writes itself in a few minutes:

1. Write the 3 first. Describe the example that would make you comfortable moving the candidate forward on this competency alone. This is your bar.
2. Write the 1. Describe what a candidate says when they clearly do not have this. Usually it is the opposite behavior, not just "less of the 3."
3. Write the 2. Describe the version of the 3's example that is thinner: fewer details, weaker outcome, or the candidate needed the interviewer to supply half the reasoning.
4. Write the 4. Take the 3's example and add one specific thing: harder circumstances, self-correction, or a second example that confirms the first.
5. Write the 5 last, and try to avoid needing it. If you cannot describe what would be rarer than the 4, your 4 was actually a 5, and you should renumber rather than force a 5 into existence.

Decision rules that go with a 1–5 scale

The scale only settles disputes if everyone agreed on the rules before the debrief, not during it.

• No half points. A 3.5 means the interviewer could not decide between the 3 anchor and the 4 anchor. Send them back to which one the evidence actually matches.
• Set the pass line per competency, not as a single overall number. A common rule: no score below 2 on any must-have competency, and an average of 3 or better across the rest. Averaging alone lets a 5 on a nice-to-have cancel out a 1 on something the role cannot do without.
• A 1 on a must-have ends the loop. Later interviewers should be told, so they do not spend a session re-testing something already settled.
• Ties are broken by re-reading evidence, not by re-scoring from memory. If two candidates land on the same total, compare the actual notes for the competency that matters most for this role, not the numbers.
• "Not assessed" is not a 3. If a competency was not actually covered in the interview, mark it as not assessed and adjust the loop next time rather than filling in a guess.

Converting scores from a different scale

If you are migrating from a 1–4 or 1–10 scale, do not attempt a mathematical conversion — the meanings do not map cleanly and a formula hides that. Instead, re-score a handful of past candidates' written notes against the new five anchors and see where they land. The table below is a rough starting point for how the ranges tend to relate, useful for talking to a team mid-migration, not for converting historical data automatically.

Notice the 1–4 scale has nowhere to put a genuine 5. That gap is exactly why some teams move to five points in the first place: a 1–4 scale forces "strong" and "exceptional" into the same box.

Keeping a 3 from becoming the default

The most common failure of any 1–5 scale is not a bad anchor, it is interviewers defaulting to the middle because it feels safe. Three things reduce it:

• Require a one-sentence justification with every score. "3, because they described debugging a production issue independently and the fix held" is checkable. "3" alone is not.
• Review the score distribution monthly, not just individual scores. If one interviewer's scores cluster at 3 across every candidate and every competency, that is a training signal, not a reflection of the candidate pool.
• Run a short calibration exercise where several interviewers score the same recorded answer independently and compare. See interview calibration for the full format.

Putting the scale on a scorecard

The scale above is the rating logic; it still needs a place to live. Pair it with an evidence column so the score and the reason sit next to each other, the way interview scorecard template lays out. A score with no evidence column next to it is the single most common reason calibration sessions run long: nobody can remember, three weeks later, what a given 3 was actually based on.

Competency: [name]
Score (1-5): [ ]
Evidence: [what the candidate said or did, specific enough to check against the anchor]
Anchor matched: [copy the sentence from the anchor table above that matches]

Common mistakes with a 1–5 scale

When five points are the wrong choice

A 1–5 scale is not the right default for every situation. If your interview loop is short (a 15-minute phone screen) or the interviewers are new and not yet trained on anchors, a 1–4 scale that removes the middle box altogether is usually more reliable, because it forces a lean one way or the other instead of hiding uncertainty in a 3. Five points earn their keep once interviewers have enough practice to reliably tell a 3 from a 4, which usually means a handful of calibrated loops, not the first one.