How to Evaluate Interview Candidates: From Impressions to Evidence of Competence

The candidate speaks fluently, quotes impressive numbers and comes across well. Three months later, it turns out that sales “grew” because of a new product rather than their work, or that their deputy was the one actually preparing the reports. Mistakes like these rarely come down to bad candidates. More often, the problem lies in the interview itself: the interviewer assessed the conversation, not the evidence.

An unstructured conversation produces an impression: how likeable and confident the person is, and how much they seem like “one of us”. Impressions are useful, but they are poor predictors of performance. Different interviewers ask different questions, interpret a “strong answer” differently and then argue about candidates they effectively assessed by different rules.

This article sets out a process that fixes this. The logic is simple: define the role requirements, gather evidence of competence and assess every candidate on the same scale.

Impression vs. evidence: what’s the difference

An impression answers the question “Did I like the candidate?” Evidence answers the question “What did they do, under what conditions and with what result?” During the interview, it helps to mentally sort everything you hear into three groups.

 The interviewer’s job is to move the conversation from the first two rows to the third — and to base the assessment mainly on that.

Step 1. Competency profile: what to clarify about the role before the first interview

Assessment begins before you meet the candidate. If the requirements aren’t defined, each interviewer will substitute their own, and comparison becomes impossible.

Before interviews start, sit down with the hiring manager and answer five questions:

1. What results do we expect from the new hire at 3, 6 and 12 months?
2. Which 3–5 tasks take up most of their time and have the greatest impact on results?
3. Where have previous people in this role most often made mistakes, and what did those mistakes cost the company?
4. Which knowledge and skills are essential from day one, and what can they learn on the job?
5. Who will they work with — clients, direct reports, government agencies, foreign partners — and in which language?

Must-have and nice-to-have requirements

Split the requirements into two lists. Must-haves are requirements without which the work is impossible or the risk is too high, and which can’t be taught on the job within a reasonable time. Nice-to-haves give a candidate an advantage, but their absence can be compensated for.

A test question for each item: “If a candidate who is strong in every other respect doesn’t have this, would we reject them?” If the answer is no, the requirement is a nice-to-have. This split will later become the cut-off rule in your scorecard.

Why abstract competencies don’t work

“Communication skills, responsibility, ability to work under pressure” appear in almost every job ad and say almost nothing about the actual job. For an accountant, communication skills mean being able to explain a tax risk to a manager without jargon. For a sales manager, they mean uncovering the needs of a client who can’t articulate them clearly. These are different skills, and they are tested with different questions.

As a rule, 4–6 key competencies are enough for one role. With more than that, the interviewer won’t have time to gather evidence for each one.

How to turn a general requirement into a testable criterion

Each competency needs observable indicators: what a person does when they have it. Here is an example for an accountant:

 A criterion like this is clear to any interviewer and applies equally to every candidate.

One method, different roles

The principle stays the same, but the content of the profile changes. Compare four contrasting roles.

Language as a job requirement

In Uzbekistan, one role may require working in Uzbek, Russian and English, while another needs only one of them. Assess language as a separate requirement, and only to the level the job actually needs: correspondence with a foreign supplier, negotiations, reading regulatory documents, working with clients in the regions.

The best approach is a short practical task in the required language: writing an email, analyzing an extract from a contract, or conducting part of the interview in that language. If language isn’t important for the role, let the candidate answer in whichever language they are comfortable with. Otherwise, you will be assessing fluency rather than competence — and you will lose strong specialists.

Step 2. Behavioral and situational questions

The main tool of a competency-based interview is two types of questions. They serve different purposes and complement each other well.

 Neither type of question gives you a ready-made prediction. An account of past behavior is one source of evidence, not proof in itself: a candidate may exaggerate their role, retell a team success as their own or pick their one success story. An answer to a situational question shows how a person reasons and whether they know the rules, but not how they will act under pressure from deadlines and management.

So with an experienced candidate, start with behavioral questions, use situational ones to check whether their experience transfers to your tasks, and cross-check important conclusions against a practical task and fact verification.

Sample questions and what to assess in the answer

Handling objections (sales manager)

Question: “Tell me about the last deal a client wanted to cancel or postpone because of price. How did things unfold?”

What to assess: whether the candidate found out the real reason behind the objection; which arguments they used and why they chose those; what they offered besides a discount; how it ended.

Follow-up questions: “What exactly did you say to the client?”, “What information about the client did you gather before the conversation?”, “What would you do differently?”

Sign of a generic answer: “I always show the value of the product” without describing a specific conversation. This is a cue to ask for a concrete example, not a reason to lower the score straight away.

Decision-making and accountability (manager)

Question: “Describe a decision you made without complete information that turned out to be wrong or debatable.”

What to assess: how the candidate weighed the risks; who they consulted; how quickly they noticed the problem; what they did to deal with the consequences; whether they changed their approach.

Follow-up questions: “What options did you consider?”, “Who else was involved in the decision?”, “How did your own managers find out about the problem?”

Red flag: the candidate can’t recall a single mistake, or shifts responsibility onto the team and circumstances.

Accuracy and knowledge of the rules (accountant)

Behavioral question: “Tell me about the most serious accounting or reporting error you have discovered. How did you find it?”

What to assess: whether the candidate knows the applicable requirements; whether they distinguish between what they can correct themselves and what requires a manager’s decision; whether they name specific steps and timelines.

Assessing candidates (recruiter)

Question: “Think of a candidate you recommended to a hiring manager who then didn’t pass probation or left soon after. What did you miss?”

What to assess: whether the recruiter forms testable hypotheses about candidates; which facts they used to support their recommendation; what they changed in their process after the mistake.

How to use STAR without turning the interview into a questionnaire

STAR stands for Situation, Task, Action, Result. It isn’t a sequence of questions but a checklist for the interviewer. Listen to the answer, note which elements are missing, and then ask a follow-up about that specific gap.

Most often, candidates describe the situation in detail and say almost nothing about their own actions. So spend most of your time on the A: “What did you personally do?”, “What did you say?”, “Which decision was yours?” If the candidate keeps saying “we”, politely ask about their own role.

How to critically examine a candidate’s story

Even a detailed and logical account should be tested against five questions. They help you separate real experience from a well-constructed story.

 A first answer is often generic: the person is nervous, isn’t sure how much detail to give or is answering in a language that isn’t their first. This isn’t a lack of competence. Give them time, rephrase the question, ask for another example — and only then make your assessment.

Step 3. Verifying specific results and personal contribution

A claimed result is not yet proof of competence. Revenue may have grown along with the market, a project may have been carried by the manager, and “30% savings” may have been calculated from an inflated baseline. To separate the candidate’s contribution from that of the team and circumstances, break down one or two key achievements using seven points.

 If the candidate answers the first three questions confidently but gets confused on the later ones, their role was probably smaller than they describe. If there is plenty of detail but the figures don’t add up, that’s a reason to verify, not an automatic rejection: people often don’t remember exact data well.

When the result can’t be expressed in numbers

For a lawyer, an internal control specialist, an HR manager or an administrator, key results often aren’t measured in revenue. In these cases, assess the quality of their work through other evidence:

• Task complexity: which non-standard situations the person handled on their own.
• Consequences of mistakes that were avoided: a risk they spotted and prevented.
• Solutions that outlasted them: a policy, template or procedure that is still in use.
• Trust and scope of responsibility: which matters they were entrusted with without supervision, and how their authority expanded.
• Work samples: an anonymized document, report or procedure — without disclosing a former employer’s confidential information.

Step 4. A single scale, fact verification and the scorecard

The evidence you’ve gathered needs to be turned into scores that can be compared. The scale below is a practical working tool, not a validated psychological test. Its purpose is to make sure that different interviewers understand what each score is awarded for in the same way. Before using it, adapt the level descriptions to the role and test them in your first few interviews.

Two separate ratings: competency and verification

For each competency, the interviewer records two things separately:

1. Competency score (1–5) — how well the actions and results the candidate described or demonstrated match the role’s criteria.
2. Verification status — how thoroughly the key facts behind the score have been checked.

 Two rules that are easy to break: the absence of external confirmation doesn’t mean the absence of competence, and a convincing story isn’t necessarily a true one.

Lack of competence and lack of information are different things

A generic answer, nerves or an incomplete first account don’t yet indicate low competence. Ask follow-up questions first:

• “Can you think of a specific time when this happened?”
• “Where did you start? What did you do next?”
• “How did you know whether it had worked?”
• “Was there a similar case that was more difficult?”

If there is still no evidence after these follow-ups, mark the competency “N/A — insufficient evidence” rather than giving a 1 or 2. A low score is given only when there is evidence of weak behavior: for example, the candidate described their actions in detail, and those actions don’t meet the criterion. “N/A” on a must-have competency is a signal for further verification, not grounds for rejection.

The 1–5 scale: observable indicators

The score primarily reflects the quality of the candidate’s actions against the requirements of your role. The complexity of the situation is context that helps you understand the answer — not a ceiling on the score, and not a reason to raise it automatically.

 How to combine four aspects into one score. The score is based on the candidate’s actions and decisions: how well they match the criterion and how independently they were carried out. The other aspects refine the picture:

• Complexity of the situation. Excellent work in typical situations can merit a 4 or a 5 if those situations are the core of the role. A complex or unusual project doesn’t earn a high score by itself: what counts is what the candidate actually did in it.
• Independence. Compare it against the level of the role. For a junior specialist, working to instructions and making the right decisions is a normal level 3; for a manager, it isn’t.
• Result. It confirms or calls into question your assessment of the actions, but doesn’t replace it. If the candidate acted correctly but the result wasn’t achieved for reasons beyond their control, score the actions and note this in a comment. Check whether they saw the risk and how they responded when the situation changed.
• Result without explanation. If there is a result but the candidate can’t explain their actions, first ask for clarification or another example. If the logic behind their decisions still isn’t apparent, that’s level 2 evidence: the result may have been down to other people or circumstances. If there simply isn’t enough data, mark it “N/A”.

The levels describe the quality of behavior. How thoroughly the facts have been checked is recorded separately, through the verification status.

Example level descriptions for the “Handling objections” competency:

• 1 — in the conversations described, immediately offers a discount without finding out the reason for the objection, and sees this as the only option.
• 2 — describes the conversation, but responds to the objection with a stock phrase and doesn’t explore what lies behind it.
• 3 — finds out the reason for the objection and offers a solution that addresses it; the deal is saved, or the loss can be explained.
• 4 — adapts their arguments to the different decision-makers on the client side, saving not just the deal but also the margin.
• 5 — notices recurring objections and changes the offer or the sales script for the whole team; the result is visible across several deals.

Interviewer errors that distort the assessment

A structured interview reduces the influence of subjectivity, but doesn’t eliminate it. Below are the errors that most often undermine even a well-prepared interview, and ways to reduce their impact.

Step 5. Checking references, facts and practical skills

The interview shows how the candidate talks about their experience. Verification shows how closely that account matches reality. You don’t need to check everything: focus on the facts the decision depends on and the skills that are critical to the role.

What applies to almost any vacancy — and what only in specific cases

How to check references

• Get the candidate’s consent and ask them to name referees: a direct manager, a colleague, a direct report or a client, depending on the role.
• Ask about facts and behavior, not character traits. Not “Is she a good specialist?”, but “What tasks was she responsible for? How was the work divided on project X? What did she do when a client threatened to leave?”
• Check the reference against your hypotheses. If doubts about the candidate’s personal share of sales remained after the interview, ask the referee exactly that.
• Don’t draw conclusions from a single reference. A sharply negative or perfectly glowing reference needs a second source.
• Don’t contact the candidate’s current employer without their explicit consent. For someone who is still employed, this can cause real problems.

Foreign specialists and experience abroad

Experience gained in another country is verified using the same methods, with some adjustments. References may only be available in English or another language, and job titles may not match local ones. Compare the scope of responsibilities, not titles: how many direct reports the person had, what budget they managed, which decisions they made independently.

An employment contract with a foreign national coming to Uzbekistan to work is concluded after confirmation of their right to work has been obtained (Article 30 of the Labour Code). This should be taken into account when planning the start date, but not when assessing competencies.

Legal framework in Uzbekistan

Selection criteria must be strictly job-related. The Labour Code (Article 4) prohibits restrictions or preferences based on sex, age, race, nationality, language, social origin, place of residence, attitude to religion, beliefs and other circumstances unrelated to a person’s professional qualities. Distinctions that arise from the requirements of a specific type of work are considered justified. The Code also explicitly links the interview to selecting an employee who meets the qualification requirements in terms of professional and business qualities. The practical takeaway: every requirement in the profile, including language, must follow from the tasks of the role.

Be prepared to justify a rejection. If a rejected candidate requests it, the employer must provide a written justification of the reasons for rejection within three days, signed by the person authorized to hire (Article 119 of the Labour Code). Unlawful rejections include, in particular, those that breach the prohibition on discrimination and those based on pregnancy or having children. In a dispute, the burden of proving that the rejection was lawful lies with the employer (Article 120 of the Labour Code), which is why scorecards with recorded evidence help you prepare such a justification.

Candidates’ personal data. The Law “On Personal Data” allows data to be processed not only with consent but also on other grounds, so it’s worth determining the legal basis for each step of the hiring process with a lawyer. Whatever the basis, general rules apply: use the data only for the stated purpose, collect only as much as that purpose requires, don’t disclose it without consent or another legal basis, and destroy it once the purpose has been achieved. Checking references and keeping CVs in a talent pool for future vacancies go beyond selection for the current vacancy, so it makes sense to obtain separate consent from the candidate for these, in a form that lets you prove it was obtained. The law classifies information about health, private life, criminal record, and religious and political beliefs as special categories of personal data subject to a stricter processing regime — there is no need to collect it during an interview.

This is an overview of general requirements, not legal advice. Before introducing the procedure, check the current version of the relevant legislation.

Step 6. Making the hiring decision based on all the evidence

By the end of the hiring process, you have scorecards from several interviewers, task results and verified facts. The decision is made at a meeting where all this information is on the table. Discuss it in this order:

1. Check the must-have requirements. If a low score on a must-have competency is confirmed by several pieces of evidence, the candidate is not considered further. If it rests on a single ambiguous answer or on one interviewer’s assessment, or is marked “N/A”, carry out additional verification first, using the same rules for everyone.
2. Compare candidates competency by competency, not just by their total score. A “candidates × competencies” table immediately shows each person’s strengths and weak spots.
3. Take verification into account. A high score that rests on an unverified key fact carries less weight than the same score confirmed by a practical task.
4. Work through disagreements between interviewers. Each interviewer explains their score with reference to the evidence they recorded.
5. Name the uncertainty. What don’t we know about the candidate? Which facts are unverified or contradict each other? How much does this matter for the job?
6. Treat nice-to-have competencies as an advantage, not as a substitute for must-haves.
7. Record the grounds for the decision: why this candidate was chosen, which risks were accepted and how they will be tested during the probationary period.

Why the candidate with the highest total doesn’t always win

The total score is a tool for comparison, not an automatic decision. Sometimes it’s better to choose a candidate with a lower total if:

• the top scorer has a critical gap in a must-have competency;
• their strengths lie in nice-to-have competencies rather than in those that drive results;
• key facts about them weren’t confirmed or contradict their references;
• the difference in total scores is smaller than the precision of the assessment itself: 3.3 and 3.4 on a five-point scale are effectively equal, and other evidence should decide.

When to verify further rather than choose

Additional verification is needed if interviewers’ scores on a must-have competency differ by two points or more, if the candidate’s main achievement hasn’t been confirmed, or if all the finalists are borderline on a key requirement. In such cases, it’s better to run a short practical task or have another conversation with a referee than to choose on the basis of incomplete data.

Where to start: the basic and extended formats

Not every vacancy needs the full set of tools. The complexity of the procedure should match the complexity of the role and the cost of a mistake.

 You can start with the basic format on your next vacancy and add elements of the extended format wherever decisions remain contentious without them.

Conclusion

The quality of an assessment depends not on the number of questions but on three things: how precisely the role requirements are defined, how thoroughly evidence of competence is gathered and how fairly the comparison of candidates is organized. A structured interview doesn’t make a decision error-free, but it does make it explainable: you can see what it was based on, where there were doubts and which risks the company accepted knowingly.

You can start with the basic format: one competency profile, the same questions and a simple scale for your next vacancy. That alone noticeably reduces “liked them / didn’t like them” debates.

When it’s worth bringing in outside expertise

If hiring runs without consistent rules, if managers assess candidates against different criteria, or if mistakes in hiring key specialists keep recurring, it helps to look at the process from the outside. SPECTRE HR Consulting provides recruitment of specialists and executives using competency-based interviews, case studies and reference checks. Through an HR audit, the team shows where HR processes are losing money and people, and proposes a prioritized plan for change.