You finish interviewing three candidates for the same opening and, when you sit down to decide, you realize you cannot compare them. You asked one about a conflict with a customer, another about where they see themselves in five years, and the third spent half the interview talking about their last boss. Every conversation was different, so what you are comparing is not candidates: it is impressions.
The structured interview exists to solve exactly that. At Voicit we work with recruiters and staffing agencies who interview by phone, over video calls and in person, and the same pattern shows up every time: when the script is the same for everyone, deciding is easier, fairer and much easier to defend. In this guide we explain what a structured interview is, how it differs from semi-structured and unstructured interviews, how to design one in six steps, and we give you a complete example with its rating scale and the interview script in Word and PDF. You can take the whole thing without signing up, whether you use Voicit or not.
What is a structured interview? A structured interview is a job interview in which every candidate for the same role answers the same questions, in the same order, and each answer is scored on a rating scale defined in advance and built from the competencies of the role. Its purpose is to make the hiring decision depend on what each candidate actually said, not on the impression they left.
What you will find in this guide
- What a structured interview is (and what it is not)
- Structured vs semi-structured vs unstructured: the differences
- How to design a structured interview in 6 steps
- Complete structured interview example (with rating scale)
- Downloadable interview script in Word and PDF + scorecard
- 7 mistakes that quietly unstructure your interview
- By video call, by phone and in person
- How to score without losing the thread: where AI fits
- Frequently Asked Questions
What a structured interview is (and what it is not)
In short: a structured interview standardizes three things: the questions (the same for everyone, derived from the competencies of the role), the procedure (same order, same length, same interviewers) and the evaluation (a rating scale with behavioral anchors, scored on the spot). If any of the three is missing, you are running a semi-structured interview.
The word “structured” gets used far too loosely. Having a list of questions on a sheet of paper does not make an interview structured. The most cited research on the topic, by Campion, Palmer and Campion (1997), identified fifteen components of structure and grouped them into two broad blocks: the ones that standardize the content and the ones that standardize the assessment. In practice, they boil down to four decisions you make before the first interview:
Questions derived from a job analysis
You do not ask “what usually gets asked”, you ask what predicts performance in that specific role: each question measures one concrete competency (customer focus, problem solving, teamwork and so on) and is worded identically for every candidate. They are usually behavioral questions (“tell me about a time when…”) or situational questions (“what would you do if…”).
Same order, same length, same procedure
Every candidate goes through the same sequence and gets an equivalent amount of time. Follow-ups are planned in advance (probes such as “what exactly did you do?”) and nobody improvises new questions for a candidate they happen to like.
A rating scale with behavioral anchors
Each question has a scoring guide, usually from 1 to 5, in which every level describes what kind of answer earns that score. That way, two interviewers who hear the same answer score it similarly. The scale is applied during or immediately after each answer, not at the end based on an overall impression.
Trained interviewers and a decision by adding scores, not by gut feel
Whoever interviews knows the script and the rating scale, takes notes of evidence (what the candidate said, not what the interviewer felt about it) and the final decision is built by adding up weighted scores. If there are several interviewers, each one scores independently before they talk to each other.
Why it works better than the “just a conversation” interview
The free-flowing interview is pleasant and gives you the feeling that you got to know the person, but it predicts poorly. The meta-analysis by Schmidt and Hunter (1998), which reviewed 85 years of studies, placed the predictive validity of the structured interview at 0.51, on par with cognitive ability tests and close to work sample tests (0.54), while the unstructured interview stayed at 0.38. The later review by Levashina, Hartwell, Morgeson and Campion (2014) confirmed the effect and added a finding that matters to any hiring team: structure reduces differences between demographic groups, which means it discriminates less.
| Selection method | Predictive validity (Schmidt and Hunter, 1998) | What it means in practice |
| Work sample test | 0,54 | Watching the candidate do the actual task |
| Structured interview | 0,51 | Almost as good as a work sample, and usable for any role |
| Cognitive ability test | 0,51 | Powerful, but measures neither attitude nor fit |
| Unstructured interview | 0,38 | Lots of interviewer confidence, little accuracy |
| Reference checks | 0,26 | Useful as a filter, weak as a predictor |
| Years of experience | 0,18 | What weighs most on a resume, what predicts least |
Beyond predicting better, the structured interview has three advantages you notice from the very first hiring process: you can compare candidates interviewed by different people in different weeks; you can explain the decision to a hiring manager or to a rejected candidate with evidence rather than adjectives; and you can calibrate the team, because when two interviewers score the same answer differently, the conversation is about the rating scale and not about who has the better eye for talent.
If you want to see where it sits among the other formats (panel, BEI, technical, competency-based), you have the full map in types of job interviews and when to use each one. Here we focus on a single one: the format that improves the most decisions per hour invested.
Structured vs semi-structured vs unstructured interview: the differences
In short: in the unstructured the interviewer chats freely and decides by impression; in the semi-structured interview there is a guide of topics or questions, but the order, the follow-ups and the evaluation are left to each interviewer; in the structured the questions, the order and the rating scale are fixed for every candidate.
All three are used in hiring and all three have their place. The trouble starts when you use one believing it is another: most of the interviews that teams call “structured” are semi-structured, because they share common questions but every interviewer scores them their own way. This table helps you work out which one you are really running and when each one makes sense.
| Criterion | unstructured interview | semi-structured interview | Structured |
| Questions | Emerge from the conversation; change with every candidate | A shared list of topics or questions; others get added on the fly | The same for everyone, derived from the job analysis |
| Order and follow-ups | Free | Flexible order; follow-ups at the interviewer's discretion | Fixed order; probes planned in the script |
| Assessment | Overall impression at the end | Notes and a qualitative assessment; sometimes a final score | A scoring guide per question with behavioral anchors, scored on the spot |
| Comparability between candidates | Low | Medium | High |
| Predictive validity | Low (about 0.38) | In between | High (about 0.51) |
| Risk of bias | High (affinity, halo effect, first impression) | Medium | Low, and documented on top of that |
| Preparation effort | None | Moderate | High the first time; reused afterwards |
| When it makes sense | Informal closing chat, selling the role, culture | Highly unusual profiles, first exploratory calls, interviews with the hiring manager | Any opening with more than two candidates and, above all, high-volume processes or processes with several interviewers |
When to use each one
- Structured: whenever you are going to compare candidates. It is the default option for the competency interview, for high-volume hiring (customer service, sales, operations, entry-level roles) and for any role in which several interviewers take part in different stages.
- Semi-structured: when the role is so unusual that building a rating scale makes no sense (a single executive, a very technical profile interviewed by an expert) or in a first ten-minute exploratory call. Even then, it is worth closing it with a rating on the same scale for everyone.
- Unstructured: for everything that is not evaluation: answering the candidate's questions, introducing the team, “selling” the offer in the final stage. If something relevant comes up in that chat, write it down as evidence, but do not change the scores from the earlier stages.
How to design a structured interview in 6 steps
In short: to design a structured interview: define the 4 to 6 critical competencies of the role; write 1 or 2 behavioral or situational questions per competency; build a 1 to 5 scale with anchors for each question; assemble the script with timings, opening and close; train the interviewers; and calibrate after the first few interviews.
You design it once per role and reuse it in every hiring process. The first version takes between two and four hours of work, which you recover with the first decision you do not have to redo. These are the six steps, in the order it pays to take them.
Define the competencies that predict performance
Start from the job analysis, not from the job posting: what the person does on a normal day, what goes wrong when someone is not working out and what sets the best people on the team apart. That gives you 4 to 6 critical competencies. More than six stretch the interview and dilute the scores; fewer than four leave gaps. Ask the hiring manager to rank them and assign a weight to each one (for example, 30-25-20-15-10).
Write one or two questions per competency
Two types work: behavioral (“tell me about a situation where a customer was very upset and what you did”) and situational (“imagine you receive two urgent complaints at the same time; how do you handle them?”). The first capture past behavior; the second, judgment in scenarios from the role. Avoid opinion questions (“would you say you are customer-focused?”): everybody says yes. If you need a question bank to start from, job interview questions by competency has more than a hundred, already classified.
Build the rating scale: 1 to 5 with behavioral anchors
For each question, describe what kind of answer earns a 1, a 3 and a 5 (levels 2 and 4 are in between). Anchors describe observable behaviors, not adjectives: “describes a concrete action of their own and a measurable result” earns a 5; “speaks in the plural or in generalities, with no identifiable action of their own” earns a 1. This is the part that takes the most time and the part that most separates a structured interview from a list of questions. The STAR method gives you the frame for the anchors: situation, task, action and result.
Assemble the script: opening, blocks, timings and close
The script is the document the interviewer has in front of them: a two-minute opening (introductions, format, a heads-up that you take notes or record), the question blocks with their planned probes, the time allotted to each block and a close with room for the candidate's questions and the next steps. For a 45-minute interview, 6 to 8 main questions is the reasonable range. Further down you have a complete example and the downloadable script.
Train whoever is going to interview
A one-hour session is enough to get started: how to read the questions without sounding like a questionnaire, how to use the probes (“and what exactly did you do?”, “how did it end?”), how to note down verbatim evidence and when to score (right after each answer). If several people interview, they should do it separately and score before comparing notes. The basic rules of a good interviewer still apply, with or without a script.
Calibrate and improve after the first interviews
After the first five or six candidates, get the team together and compare scores question by question. If everyone gets a 4 on one question, that question does not discriminate and needs to change; if two interviewers score the same answer two points apart, the anchor is badly written. Also check the correlation with what happens next: six months in, whoever performs best should match whoever scored best. If they do not match, the problem is in the competencies you chose.
Complete structured interview example (with rating scale)
In short: this structured interview example is for a customer service representative role: 45 minutes, five competencies, eight questions (six behavioral and two situational), each with its 1 to 5 rating scale and space to note evidence. It is the same script you can download in the next section.
Nothing explains a structured interview better than seeing one in full. This is the script a hiring team would use for a customer service representative role in a contact center (support by phone, chat and email), with the scores of a fictional candidate so you can see how it gets filled in. The competencies and the weights were set by the hiring manager in step 1; the questions and the anchors came out of steps 2 and 3.
The interviewer introduces herself and the format: “I am going to ask you the same eight questions I ask every other candidate, in the same order, and I will take notes while you answer. Some of them will ask for concrete examples of situations you have been through; the more specific you are, the better. At the end you will have ten minutes for your own questions.” If the interview is being recorded, this is where the candidate is told and asked for consent.
What to listen for: whether they separate the customer's emotion from the problem, whether they take ownership even though the cause was elsewhere and whether they close with a result. Probes: “what exactly did you say?”, “what happened with that customer afterwards?”.
Score: 12345 · Evidence: “I told her I understood why she was angry and that, even though the delay came from logistics, I would own the case; I called her the next day to confirm the delivery.”
What to listen for: whether they explain the reason behind the rule instead of hiding behind it, whether they offer alternatives and whether they know when to escalate. Probe: “and if they keep insisting?”.
Score: 12345 · Evidence: “I would explain that I cannot do it and pass it on to my supervisor”; no alternatives even though asked twice.
What to listen for: whether they identified the cause and not just the symptom, whether the initiative was theirs and whether they measured the effect. Probe: “how did you know it was recurring?”.
Score: 12345 · Evidence: “30% of our tickets were the same billing question; I proposed an automatic message in the customer portal and they dropped by half within a month.”
What to listen for: whether they look for the information actively, whether they manage the customer's wait and whether they avoid making up answers.
Score: 12345 · Evidence: “I told him I did not want to give him the wrong figure, that I would email him before 6 p.m., and I did.”
What to listen for: whether they adapt their language to the listener, whether they check for understanding and whether they structure the explanation. The answer itself is also a sample of verbal communication.
Score: 12345 · Evidence: a well-ordered answer in three steps; mentioned that “I always send a recap email so they have it in writing.”
What to listen for: a professional tone without coldness, focus on the problem, clear boundaries if the behavior turns abusive.
Score: 12345 · Evidence: “I understand this has been frustrating. To sort it out I need your order number; as soon as you send it over I will look into it right away.”
What to listen for: an explicit prioritization rule, management of their own emotions and recovery afterwards. Probe: “what did you leave undone and why?”.
Score: 12345 · Evidence: prioritized “whatever was due that day”, but could not say what was left pending or how it was communicated.
What to listen for: whether they fix first and discuss later, whether they speak to the colleague directly and whether they avoid assigning blame in front of the customer.
Score: 12345 · Evidence: “First I fixed the shipment with the customer; then I asked my colleague if we could review the address template together.”
Candidate's questions, explanation of the next steps and timelines. The interviewer completes the scoring before opening the next interview or discussing it with anyone.
| Competency | Weight | Average score | Weighted |
| Customer focus | 30 % | 3,5 | 1,05 |
| Problem solving | 25 % | 4,5 | 1,13 |
| Communication | 20 % | 4,0 | 0,80 |
| Working under pressure | 15 % | 3,0 | 0,45 |
| Teamwork | 10 % | 4,0 | 0,40 |
| Total out of 5 | 3,83 |
Recommendation: advance to the next stage (process threshold: 3.5). Point to explore in the interview with the hiring manager: working under pressure (question 7), the only competency below 3.5.
Notice three details in the example. First, the evidence is verbatim: you write down what the candidate said, not “good attitude”. Second, the score is recorded on the spot, question by question, and the total comes from a weighted sum, not from a final gut feeling. Third, the recommendation flags what to explore in the next stage, so the interview with the hiring manager does not repeat the eight questions but digs deeper where it matters. If you want to go further on the behavioral side, the behavioral event interview (BEI) based on critical incidents is the most demanding version of this same principle.
Downloadable structured interview script (Word and PDF) + scorecard
In short: download the structured interview script from the example above in Word or PDF, with the eight questions, the rating scale anchors and space for evidence, plus the scorecard in Excel to calculate the weighted score. No sign-up. Swap the competencies and questions for the ones in your opening.
The script is designed so you can adapt it in half an hour: replace the five competencies with the ones for your role, adjust the weights, rewrite the questions and, above all, the anchors. The structure (opening, blocks, probes, scoring on the spot, close with the weighted table) is already solved. It also includes a bank of twenty extra questions sorted by competency, so you do not have to start from scratch when you move on to the next opening.
The scorecard is the same one that comes with our guide to interview questions: one sheet per candidate with the competencies, the weights and the automatic calculation of the weighted total.
How to fill it in during the interview
- Before: print or open the script with the candidate's name, the date and the format (phone, video call or in person). Reread the anchors of the rating scale: scoring well requires having them fresh in your mind.
- During: read the question exactly as written, allow silence, use the planned probes if the answer stays generic. Note verbatim phrases in the evidence space and mark the score as soon as the answer ends, not at the end of the interview.
- After: transfer the scores to the scorecard, write the recommendation and the point to explore in the next stage. Do it before opening the next resume: the memory of an interview fades within minutes.
Ideally, you would not fill it in by hand at all. Listening, probing, noting verbatim evidence and scoring all at the same time is hard, and it is the reason so many structured interviews fall apart halfway through. Voicit records the interview (by phone, over a video call or in person), transcribes it and fills in the rating scale with the evidence for each competency, so you only have to review and score. See how it works for hiring teams
7 mistakes that quietly unstructure your interview
In short: the most common mistakes in a structured interview are reading the questions like a questionnaire, skipping questions with the candidate who “already convinced you”, scoring at the end by impression, not using the probes, letting one interviewer influence another before scoring, not writing down evidence and never revisiting the rating scale.
- Reading the questions like a robot. Structured does not mean cold. You can keep the exact wording of the question and still introduce it naturally (“let's move on to the next one, which is about how you handle pressure”). What does not change is the question; the tone can.
- Skipping questions with the candidate who has already won you over. This is the most expensive mistake: precisely when the first impression is good is when you need the script most, because the halo effect will make you hear confirmations.
- Scoring at the end “as a whole”. By the time you finish, the last answer and the general rapport weigh more than the eight answers. Scoring happens question by question, on the spot.
- Not using the probes. Most answers start out generic (“we usually…”, “the normal thing is…”). Without the “and what exactly did you do?” you are scoring the candidate's team, not the candidate.
- Comparing notes with other interviewers before scoring. One “I liked her” in the hallway is enough to pull the second score towards the first. First everyone scores; then you talk.
- Writing down impressions instead of evidence. “Good attitude” cannot be compared or defended. “Said she called the customer the next day to confirm the delivery” can.
- Never revisiting the rating scale. A scale that is never calibrated decays: each interviewer interprets it their own way and within six months you are back to the impression-based interview, just with more paperwork.
Structured interviews by video call, by phone and in person
In short: the structure does not change with the format, but the execution does: over a video call it helps to share the opening of the script and keep an eye on timings; over the phone, probes and pauses matter more because there is no body language; in person, the risk is loosening the script because of the closeness. In all three cases, taking notes or recording requires telling the candidate first.
A structured script works the same over the phone, online or face to face, and in fact it is the best guarantee that the first screening call and the final interview with the hiring manager measure the same thing. What changes are the details of execution:
- Over a video call (Meet, Teams, Zoom): this is the most common format in hiring today. Share the format and the length at the start, keep the script on a second screen so you do not lose eye contact and watch the time per block: in remote interviews answers tend to run long. You will find more detail in how to run effective virtual interviews.
- Over the phone: perfect for a first structured screening of 15 to 20 minutes with three or four questions from the script. With no body language, probes and silences are your main tool; take notes more literally than ever. If you record so you do not lose evidence, in recording phone interviews we explain how to do it with notice and consent.
- In person: closeness invites conversation and tempts you to skip the script. Keep the document in view, score on paper or a tablet between questions and save the free-flowing chat for the end, outside the evaluation.
If you combine the structured interview with psychometric tests, keep the same logic: each tool contributes a score to the same scorecard, with its own weight. And if you are reviewing the whole process, from posting the opening to onboarding, the place of the structured interview within the stages of the recruitment process is the evaluation stage, never the first one and never the last one.
How to score without losing the thread: where AI fits
In short: AI does not design the structured interview for you and does not decide whom to hire: it records the interview with consent, transcribes it and fills in the rating scale with the verbatim evidence for each competency. The interviewer keeps the listening and the scoring; they delegate the note-taking and the report.
The main practical difficulty of the structured interview is not designing it, it is running it well: truly listening, probing, writing down verbatim evidence and scoring on the spot, all at once. When the interviewer writes, they stop listening; when they listen, the evidence slips away. That is where an AI tool built specifically for interviews adds value, and where it pays to be honest about what it does and what it does not.
Record the interview with consent
You tell the candidate in the opening and get their consent. With Voicit, a phone interview is recorded from your phone, a video call in Meet, Teams or Zoom from the Chrome extension, and an in-person interview with the Voicit Go recorder or from your laptop.
The AI transcribes and fills in your template
When you finish, Voicit transcribes the conversation, telling interviewer and candidate apart, and completes the report using the template you defined: for each competency, the verbatim evidence of what the candidate answered, ordered according to your script.
You review, score and decide
The score on the rating scale is still yours: with the evidence in front of you, applying the anchors takes a couple of minutes and is more consistent than doing it from memory. The report is ready to share with the hiring manager, in the format you always use.
For hiring teams that interview every day
Voicit records the interview, transcribes it and fills in your structured script with the evidence for each competency. You keep the listening and the decision.
Frequently asked questions about the structured interview
What is a structured interview?
It is a job interview in which every candidate for the same role answers the same questions, in the same order, and each answer is scored on a rating scale defined in advance (usually 1 to 5, with a description of what earns each score). The questions are derived from the competencies of the role and are usually behavioral (“tell me about a time when…”) or situational (“what would you do if…”).
What is the difference between a structured and an unstructured interview?
In an unstructured interview the interviewer chats freely, adapts the questions to each candidate and decides on overall impression. In a structured interview the questions, the order and the way of scoring are the same for everyone. The structured interview predicts job performance better (validity of 0.51 versus 0.38 in Schmidt and Hunter's meta-analysis), makes candidates comparable and reduces bias. The semi-structured interview sits in between: a shared guide, but order and evaluation left to the interviewer.
Can you give an example of a structured interview question?
For the competency “customer focus”: “Tell me about a time when a customer was very upset about something that was not your fault. What did you do and how did it end?” It is behavioral, asks for a concrete case and comes with a rating scale: a 1 if they blame the customer and describe no action of their own, a 3 if they describe a reasonable action with no clear result, a 5 if they take ownership, propose a solution, follow up and report the outcome. In the article you have the full script with eight questions of this kind.
How many questions should a structured interview have?
Between 6 and 8 main questions for a 45 to 60 minute interview: one or two for each of the 4 to 6 critical competencies of the role, plus the planned probes to dig deeper. For a 15 to 20 minute phone screen, 3 or 4 questions from the same script. More questions do not improve prediction and do tire out both the candidate and the interviewer.
Is it legal to record a structured interview?
Yes, as long as you tell the candidate before you start and get their consent. In the United States, federal law and most states require one-party consent, but around a dozen states, including California, Florida, Illinois and Washington, require every participant to agree, so the safe rule is to ask every candidate regardless of where they are. The EEOC treats interviews as a selection procedure under the Uniform Guidelines, and a structured, job-related interview with documented scores is far easier to defend if a hiring decision is ever challenged. In the United Kingdom a recording is personal data under the UK GDPR and the Data Protection Act 2018: you need a lawful basis, you must tell candidates why you record, who will access the recording and how long you keep it, and the ICO's employment practices guidance on recruitment expects exactly that. In both countries, record only for selection purposes and delete the recording when the process ends, unless the candidate agrees otherwise.
Does the structured interview work for operational or high-volume roles?
That is where it pays off most. In processes with many candidates for the same role (customer service, sales, operations, entry-level roles) the structure lets several interviewers evaluate the same way, speeds up the decision and leaves a comparable record. The script is designed once per role and reused in every process; you only need to calibrate the rating scale from time to time.
Actionable summary
If you take only three ideas from this guide:
- Structuring means standardizing three things: questions, procedure and evaluation. A list of questions without a rating scale is a semi-structured interview, whatever you choose to call it.
- The rating scale with behavioral anchors is the heart of it. Describe observable behaviors for the 1, the 3 and the 5 of every question, score on the spot and note verbatim evidence. That is what makes the interview comparable and defensible.
- Design once, calibrate always. The script for a role is reused in every process; what you cannot skip is the periodic review of which questions discriminate and whether the scores match later performance. Download the script and the scorecard and start with a single opening.
