Candidate Evaluation Form: What to Include (Plus Template)
Candidate Evaluation Form: What to Include (Plus Template)
Most candidate evaluation forms fail at the one job they exist to do. They collect opinions. An interviewer types "seemed strong technically" or "great communicator," picks a 4 out of 5, and moves on. Six weeks later, when two interviewers disagree or a rejected candidate's lawyer asks how the decision was made, that form says nothing. A score with no evidence behind it is just a feeling with a number stapled to it.
The fix is a single design choice we call the evidence box: every criterion on the form gets a required field that holds one quote or one observed behavior, not an adjective. No quote, no score. That rule is the difference between a form that records the interviewer's mood and a form that records what the candidate actually said. This guide gives you the field-by-field rules and a usable template you can paste into your ATS today. The frame throughout is simple: score the signal, not the vibe.
Why the evidence box changes the math
The evidence box works because it forces structure, and structure is what makes interviews predict anything. A 2021 meta-analysis by Sackett and colleagues in the Journal of Applied Psychology found that structured interviews are roughly twice as effective at predicting job performance as unstructured ones. On their own, structured interviews predict performance with a validity of about .55 to .70. The form is where that structure lives or dies. A form that lets interviewers write "good vibe" quietly converts a structured interview back into an unstructured one.
The evidence box also carries the legal weight, which is the part most teams underrate until they need it. The EEOC's Uniform Guidelines on Employee Selection Procedures judge a hiring process by its disparate impact, meaning whether it screens out a protected group at a meaningfully higher rate. A score alone cannot defend itself. A score plus "candidate described migrating a 12-service monolith and named the rollback step" can. The contrast shows up in outcomes: one analysis found almost 60% of discrimination lawsuits based on unstructured interviews were judged discriminatory, while 100% of cases built on structured interviews were found not discriminatory. The evidence box is what turns a disputed gut call into a documented, job-related rating that a reviewer can audit line by line.
The fields every candidate evaluation form needs
A candidate evaluation form should contain seven parts. Skip any of them and you reopen a gap that bias or a challenge can walk through.
| Field | What it captures | Why it matters |
|---|---|---|
| Role and req ID | Which opening this rates | Ties the form to a job analysis (the legal anchor) |
| Interviewer and date | Who scored, when | Audit trail and accountability |
| Criteria (3 to 6) | Job-related attributes only | Derived from job duties, not personality |
| Rating scale | A fixed scale per criterion | Forces a directional call |
| Evidence box | One quote or behavior per criterion | The defensible core; no opinion allowed |
| Overall recommendation | A single directional verdict | The decision, separated from the criteria |
| Risk flag | Anything off-limits or unresolved | Catches bias and open questions early |
Criteria come first because everything downstream depends on them. Greenhouse recommends three to four attribute categories per role and no more than five or six attributes per category, and that ceiling matters: a form with twelve criteria gets skimmed, and a skimmed form is an opinion form again. Pick the few attributes that actually separate strong hires from weak ones in this role, write them down before the interview, and use the same set for every candidate. The same bar for every candidate is the whole point.
The template
Here is the form. Copy it, swap in your role's criteria, and put it in front of every interviewer for this req. If you would rather start from a weighted grid with named anchors at each level, the Interview Scorecard Template: Free Download and How to Use It ships one ready to paste.
CANDIDATE EVALUATION FORM
Role: ______________________ Req ID: __________
Candidate: __________________ Stage: __________
Interviewer: ________________ Date: __________
--- CRITERIA (rate each, evidence required) ---
1. [Criterion, e.g. "System design depth"]
Rating: 1 2 3 4 (1 = no signal, 4 = strong signal)
Evidence (a quote or observed behavior, REQUIRED):
____________________________________________
2. [Criterion, e.g. "Handles ambiguity"]
Rating: 1 2 3 4
Evidence (REQUIRED):
____________________________________________
3. [Criterion, e.g. "Stakeholder communication"]
Rating: 1 2 3 4
Evidence (REQUIRED):
____________________________________________
--- OVERALL ---
Recommendation: Strong No No Yes Strong Yes
One-line reason (cite the strongest evidence above):
____________________________________________
Risk flag (anything off-limits, unresolved, or to retest):
____________________________________________
Notice the rating scale has no middle. That is deliberate. Lever's scorecard uses a four-point scale that removes the neutral option for a reason: a 3-out-of-5 is where indecision hides. For the fuller case on why even-numbered scales beat the bare 1-5, and when a behaviorally anchored scale earns its build cost, see Interview Rating Scales: 1-5 vs Behaviorally Anchored. Greenhouse does the same with its overall recommendation, offering only Definitely Not, No, Yes, and Strong Yes. Force the directional call, then make the evidence box justify it.
How to fill it out so it holds up
A good form is only as good as the discipline around it. Run it in this order.
- Write the criteria before the interview, from the job's actual duties. If you cannot tie a criterion to something the person will do on the job, cut it. This is your job-analysis anchor, and it is what makes the rating job-related rather than personal.
- During the interview, capture verbatim phrases in the evidence box as they happen, not from memory afterward. "Said he'd ship the smallest reversible change first" beats "good judgment" every time.
- Score immediately, before you talk to anyone. As TA experts Atta Tarki, Tyler Cowen, and Alexandra Ham put it, "Having interviewers submit their ratings before getting input from their colleagues will have the further benefit of reducing the chance of groupthink in your evaluations" (Marginal Revolution, 2022).
- In the one-line reason, point at your own strongest evidence box. If you cannot, your verdict is running ahead of your evidence, and that is exactly the gap a hiring committee or a reviewer needs to see.
We build the interview agent at Asked, and across the live transcripts we have processed, the single most common failure we see is the empty evidence box: a confident 4 with a comment that just repeats the criterion back. When the form makes that quote mandatory, interviewers either find the evidence or lower the score honestly. Both outcomes are better than the gut call you were about to write down.
Evidence-box readiness checklist
Before you ship the form to your interviewers, run it against this list.
- Every criterion is tied to a real job duty, not a personality trait.
- The evidence box is a required field, not optional.
- The rating scale has no neutral middle.
- Interviewers score before any group discussion.
- The same criteria are used for every candidate in the role.
- Forms are stored and retrievable per candidate, not in email threads.
If any box is unchecked, you have a form that records opinions. Fix it before the next interview, not after a dispute.
Frequently asked questions
What is the difference between a candidate evaluation form and an interview scorecard?
None worth arguing about. The terms are used interchangeably. Both name the structured form interviewers use to rate a candidate against fixed, job-related criteria. "Scorecard" tends to emphasize the rating; "evaluation form" tends to emphasize the criteria and notes. The version that holds up is the one with a required evidence box either way.
How many criteria should a candidate evaluation form have?
Three to six per role. Fewer than three and you are not measuring enough to separate candidates; more than six and interviewers skim. Greenhouse's guidance of three to four categories with five or six attributes max is a sane ceiling. Pick the attributes that genuinely predict success in this specific role and drop the rest.
Does a candidate evaluation form actually help with legal defensibility?
Yes, when it records evidence and the criteria come from a job analysis. A score with no supporting behavior is hard to defend if a decision is challenged. A score paired with a quoted, job-related observation, applied identically to every candidate, is the kind of documentation the EEOC's Uniform Guidelines and structured-hiring research both point toward. The evidence box is the part that does the defending.
Do This Next
Pick one open role you are actively interviewing for this week. Build the three-to-six criteria for it straight from the job's real duties, and add a required evidence box under each one. Use the four-point scale above with no neutral middle, and make every interviewer score before the debrief. Score your next two candidates with it, then compare evidence boxes side by side and watch how fast the disagreements resolve. For the bigger picture on tying these forms into a process your team will actually adopt, read Interview Scorecard: How to Build One That Actually Gets Used. Start today: try Asked free and let it draft the scorecard, with the supporting quotes already pulled from the transcript, before you even open the form.