---
title: Why Interview Scores Drift to the Middle, and How to Stop It
description: "Interview scores pile up at 3 because the middle is the cheapest score: never badly wrong, never explained. Four causes, and four ways to make it cost more."
image: https://nobusato.jp/hubfs/covers-2026-11/interview-scores-drift-to-the-middle-cover-en-2026-10-m.png
---

[Nobu Sato](https://nobusato.jp/en?hsLang=en)

[Profile](https://nobusato.jp/en/profile?hsLang=en) [Method](https://nobusato.jp/en/blog/method?hsLang=en) [Articles](https://nobusato.jp/en/blog?hsLang=en) [Glossary](https://nobusato.jp/en/blog/glossary?hsLang=en) [Contact](https://nobusato.jp/en/contact?hsLang=en)

[日本語](https://nobusato.jp/ja/blog/interview-scores-drift-to-the-middle)[English](https://nobusato.jp/en/blog/interview-scores-drift-to-the-middle)[中文](https://nobusato.jp/zh/blog/interview-scores-drift-to-the-middle)

Organizations & Hiring

In this article

1. [Why a stack of 3s is a problem](https://nobusato.jp/en/blog/interview-scores-drift-to-the-middle#s1)
2. [What pulls scores toward the middle?](https://nobusato.jp/en/blog/interview-scores-drift-to-the-middle#s2)
3. [Make the middle expensive](https://nobusato.jp/en/blog/interview-scores-drift-to-the-middle#s3)
4. [A Japanese interviewer's 3 needs a second reading](https://nobusato.jp/en/blog/interview-scores-drift-to-the-middle#s4)

1. [Home](https://nobusato.jp/en?hsLang=en)
2. [Articles](https://nobusato.jp/en/blog?hsLang=en)
3. [Organizations & Hiring](https://nobusato.jp/en/blog/tag/organizations-hiring)

# Why Interview Scores Drift to the Middle, and How to Stop It

Why interview ratings pile up in the middle

[Nobu Sato](https://nobusato.jp/en/profile?hsLang=en)8 min read

Contents4

1. [Why a stack of 3s is a problem](https://nobusato.jp/en/blog/interview-scores-drift-to-the-middle#s1)
2. [What pulls scores toward the middle?](https://nobusato.jp/en/blog/interview-scores-drift-to-the-middle#s2)
3. [Make the middle expensive](https://nobusato.jp/en/blog/interview-scores-drift-to-the-middle#s3)
4. [A Japanese interviewer's 3 needs a second reading](https://nobusato.jp/en/blog/interview-scores-drift-to-the-middle#s4)

Interview scores drift to the middle because the middle is the cheapest score to give. A 3 out of 5 can't be badly wrong, needs no defending in the debrief, and fits every interviewer's private idea of "average." The pull may be stronger in Japan: a classic 1995 study found Japanese students choosing the midpoint more often than North Americans. The fix is to make the middle cost as much as the ends.

## Why a stack of 3s is a problem

A stack of 3s can't separate candidates, and separating candidates is the whole job.

Structured interviews are one of the best selection tools we have. When Paul Sackett and colleagues redid the big meta-analysis of selection methods in the Journal of Applied Psychology in 2022, structured interviews came out on top with a mean validity of 0.42. Cognitive ability tests, long treated as the gold standard, fell to 0.31. But that number assumes the scores actually spread candidates apart. Ask the best questions in the world and record a 3 for everyone, and you have nothing left to compute with.

So central tendency, the rater's habit of piling scores in the middle, isn't a small statistical quirk. It quietly turns a good method back into a gut call.

## What pulls scores toward the middle?

Four things, usually at once: fear of being wrong, no anchor, a scale with a built-in resting place, and having to explain yourself afterwards.

Fear first. Give a 5 to someone who flops, and you misjudged. Give a 1 to someone who goes on to do well at a competitor, and you misjudged again. A 3 is never very wrong. When one interviewer decides alone and owns the outcome alone, 3 is cheap insurance.

Then the missing anchor. The form says "1 poor, 3 average, 5 excellent." Average compared with whom? The last candidate? The current team? The ideal hire in the interviewer's head? Everyone carries a different ruler, and when unsure they all move to the middle, which is never far from any of them.

Third, the scale itself. An odd-numbered scale has an exact center, and the center is a legitimate way to say nothing. Researchers have measured how strong that pull is across cultures. In a 1995 study in Psychological Science, Chuansheng Chen, Shin-ying Lee and Harold Stevenson gave the same fifty-seven 7-point items to high school students in Sendai (944), Taipei (1,357), Canada (687) and the US (2,174). Japanese and Taiwanese students used the midpoint more than both North American groups. US students used the extremes most. Within every group, students who endorsed individualism chose the midpoint less. The differences were small but statistically significant.

I wrote about the Japanese side of this in [In Japan, "neither agree nor disagree" is an answer](https://nobusato.jp/en/blog/japan-neutral-survey-answers?hsLang=en). An interview scorecard is just a survey filled out by the interviewer.

The fourth force is the one people miss. You have to explain your score. In a 1995 experiment in the Journal of Applied Psychology, Neal Mero and Stephan Motowidlo had 247 undergraduates rate the performance of four simulated subordinates. Raters who had to justify their ratings in writing were more accurate. But accountable raters who were told the subordinates' earlier ratings had been too low gave more favorable scores.

Having to explain makes people careful. Who you explain to, and what they want to hear, pushes the score toward them. In a hiring debrief, a 5 has to be defended in front of the room and a 1 invites cross-examination. A 3 requires nothing. Part of the drift to the middle is people avoiding that conversation.

Scores drift to the middle because four forces stack, and each has its own fix

Left: thoughts interviewers commonly have while scoring (illustrative). Middle: the cause, based on Chen et al. (1995) and Mero and Motowidlo (1995). Right: the author's practice; apart from BARS (Smith and Kendall, 1963), not tested by research

What the interviewer thinksCauseFix

"If I score high and they flop, it's on me"

→

Fear of being wrong

→

Submit scores independently; the panel owns the decision

"Average compared with whom?"

→

No anchor

→

Write each level as observable behavior (BARS)

"Not sure, so the middle"

→

A scale with a center

→

Four levels; mark real uncertainty "to verify," no score

"A 3 won't need explaining"

→

Avoiding the explanation

→

Every score carries one observed fact, a 3 included

The middle score is the one nobody has to explain. Make every level need an explanation and the middle stops being cheap.

Figure 1. Interview scores pile up in the middle because of fear of being wrong, missing anchors, the center of an odd-numbered scale and the wish to avoid explaining; a 1995 study found Japanese and Taiwanese students chose the midpoint of 7-point scales more than North American students, and a separate experiment found raters who had to justify in writing were more accurate.

## Make the middle expensive

Four causes, four fixes, one idea. The middle is popular because it's cheap. Raise its price and the distribution spreads out by itself. To be clear, these fixes are my own practice. BARS has a research pedigree; I haven't found studies that test the rest directly.

No anchor? Provide one. Behaviorally anchored rating scales go back to a 1963 paper in the Journal of Applied Psychology by Patricia Cain Smith and Lorne Kendall. Each level is described as concrete behavior, and a second group sorts the behavior examples back into the dimensions; examples that don't land where they should are dropped. Take "drives a project." A 3 no longer says "average." It says "delivers to plan; problems outside the plan need the manager to step in." A 5 says "resolves problems outside the plan and tells the manager afterwards." To give a 3, the interviewer has to name the part of the candidate's answer that matches that sentence.

Avoiding explanation? Require it for every score. Next to each score, write one observed fact: something the candidate said or did, with no adjectives about how it felt. A 3 needs one too. Once that rule exists, a 3 costs as much effort as a 5, and interviewers become more willing to use the ends.

A scale with a center? I remove the center and use four points, which forces a direction. This is my habit; no research finding backs it. I haven't found a study showing that dropping the midpoint makes interview ratings more accurate, and a forced direction may just be noise. So I add one rule. If the interviewer genuinely can't tell, they write "to verify," give no score, and the gap becomes a question for the next round. It's the same rule as in [Does this person fit? Veto, weakest layer, five levels](https://nobusato.jp/en/blog/method-fit-check?hsLang=en): don't fill an unknown with a middle number.

Fear of being wrong? Spread the risk. Interviewers submit scores without seeing anyone else's. Then run a calibration session that starts with each interviewer's score distribution from the last quarter, before anyone discusses a candidate. If one interviewer gives 3s ninety percent of the time, everyone will read their 3s accordingly. The panel owns the decision. No single interviewer carries it.

## A Japanese interviewer's 3 needs a second reading

What follows is my own observation. I have no data for it.

Japanese interviewers tend to avoid committing to strong written judgments, negative ones especially. The scorecard will be read by HR and maybe by the candidate's future manager. A 1 on paper is an assertion you have to own. So a Japanese 3 holds a mix: real "average," "I have concerns I'd rather not write," and "I didn't get to that."

Foreign managers often misread this. Used to wider scores at headquarters, they see a column of 3s from the Japan team and conclude the team has no judgment. Usually the judgment is there and just isn't in the number. Pushing people to give 1s rarely works. The observed-fact rule does. Interviewers who won't assert will describe, and they remember exactly what the candidate said and how.

Read those descriptions before you read the scores. How I pull evidence out of them is in [After the resume screen: what I look for in an interview](https://nobusato.jp/en/blog/what-i-look-for-in-interviews?hsLang=en).

A small thing to try before your next calibration session: pull each interviewer's scores from last quarter and work out what share were 3s. Don't publish it. Just look. The person with the most 3s probably has the most interesting notes.

Sources (as of November 2026)

- [Chen, Lee & Stevenson, "Response Style and Cross-Cultural Comparisons of Rating Scales Among East Asian and North American Students," Psychological Science 6(3), 1995](https://escholarship.org/uc/item/1b00n182)
- [Mero & Motowidlo, "Effects of Rater Accountability on the Accuracy and the Favorability of Performance Ratings," Journal of Applied Psychology 80(4), 1995](https://doi.org/10.1037/0021-9010.80.4.517)
- [Smith & Kendall, "Retranslation of Expectations: An Approach to the Construction of Unambiguous Anchors for Rating Scales," Journal of Applied Psychology 47, 1963 (Bowling Green State University bibliography)](https://www.bgsu.edu/arts-and-sciences/psychology/services/job-descriptive-index/patricia-cain-smith.html)
- [Sackett, Zhang, Berry & Lievens, "Revisiting the design of selection systems in light of new findings regarding the validity of widely used predictors," Industrial and Organizational Psychology 16(3), 2023 (reporting their 2022 meta-analysis)](https://www.cambridge.org/core/journals/industrial-and-organizational-psychology/article/revisiting-the-design-of-selection-systems-in-light-of-new-findings-regarding-the-validity-of-widely-used-predictors/A20984B138319E3D432E643978BF026D)

---

Author

[Nobu Sato](https://nobusato.jp/en/profile?hsLang=en)

Launching in Japan, starting with people and systems. I write about building organizations, sales teams and technology organizations in Japan, in English, Japanese and Chinese.

[Profile](https://nobusato.jp/en/profile?hsLang=en)[Contact](https://nobusato.jp/en/contact?hsLang=en)

## Read next

1. 01[Reverse Talent Maps: Read a Company by Where Leavers Go](https://nobusato.jp/en/blog/reverse-talent-map?hsLang=en)7 min
   
   Where a company's former employees went, whether they moved up and when they left says more than exit interviews. How ...
2. 02[Five Reasonable Hiring Requirements Left Ten People in Japan](https://nobusato.jp/en/blog/hiring-requirements-multiply?hsLang=en)7 min
   
   Run one job description through Japan's official statistics and five reasonable requirements leave about ten likely ...
3. 03[Why Recruiters in Japan Keep Sending the Same Candidates](https://nobusato.jp/en/blog/recruiters-send-the-same-candidates?hsLang=en)7 min
   
   Of 10.23 million people in Japan who want a new job, only 3.6 million are looking. That is why shortlists overlap, and ...

[Column guide: Organizations & Hiring](https://nobusato.jp/en/blog/guide-organizations-and-hiring?hsLang=en) · [Glossary](https://nobusato.jp/en/blog/glossary?hsLang=en) · [The method behind the articles](https://nobusato.jp/en/blog/method?hsLang=en)

[← All articles](https://nobusato.jp/en/blog?hsLang=en)

## Writing about launching in Japan, in three languages.

[Profile](https://nobusato.jp/en/profile?hsLang=en) [Get in touch](https://nobusato.jp/en/contact?hsLang=en)

Nobu Sato

Writing about launching in Japan, in three languages.

[info@nobusato.jp](mailto:info@nobusato.jp) [Substack](https://nobusato.substack.com)[Newsletter](https://www.linkedin.com/newsletters/the-japan-entry-memo-7519750783555518464/)[Medium](https://medium.com/@nobusato5888)[LinkedIn](https://www.linkedin.com/in/realnobusato/)[X](https://x.com/nobusato5888) [RSS](https://nobusato.jp/en/blog/rss.xml?hsLang=en)

[Profile](https://nobusato.jp/en/profile?hsLang=en)[Method](https://nobusato.jp/en/blog/method?hsLang=en)[Articles](https://nobusato.jp/en/blog?hsLang=en)[Glossary](https://nobusato.jp/en/blog/glossary?hsLang=en)[Contact](https://nobusato.jp/en/contact?hsLang=en)

Views here are my own and don’t represent any organization I work or have worked for.

© 2026 Nobu Sato. All rights reserved.·[Privacy and disclaimer](https://nobusato.jp/en/privacy?hsLang=en)

```json
{
  "@context" : "https://schema.org",
  "@type" : "BreadcrumbList",
  "itemListElement" : [ {
    "@type" : "ListItem",
    "item" : "https://nobusato.jp/en",
    "name" : "Home",
    "position" : 1
  }, {
    "@type" : "ListItem",
    "item" : "https://nobusato.jp/en/blog",
    "name" : "Articles",
    "position" : 2
  }, {
    "@type" : "ListItem",
    "item" : "https://nobusato.jp/en/blog/tag/organizations-hiring",
    "name" : "Organizations & Hiring",
    "position" : 3
  }, {
    "@type" : "ListItem",
    "item" : "https://nobusato.jp/en/blog/interview-scores-drift-to-the-middle",
    "name" : "Why Interview Scores Drift to the Middle, and How to Stop It",
    "position" : 4
  } ]
}
```

```json
{
  "@context" : "https://schema.org",
  "@id" : "https://nobusato.jp/en/blog/interview-scores-drift-to-the-middle",
  "@type" : "WebPage",
  "about" : {
    "@type" : "Thing",
    "name" : "Organizations & Hiring"
  },
  "author" : {
    "@id" : "https://nobusato.jp/#person"
  },
  "description" : "Interview scores pile up at 3 because the middle is the cheapest score: never badly wrong, never explained. Four causes, and four ways to make it cost more.",
  "inLanguage" : "en",
  "isPartOf" : {
    "@id" : "https://nobusato.jp/en/blog",
    "@type" : "Blog",
    "name" : "Articles"
  },
  "keywords" : "Organizations & Hiring",
  "name" : "Why Interview Scores Drift to the Middle, and How to Stop It",
  "url" : "https://nobusato.jp/en/blog/interview-scores-drift-to-the-middle"
}
```

```json
{
  "@context" : "https://schema.org",
  "@id" : "https://nobusato.jp/#person",
  "@type" : "Person",
  "alternateName" : [ "佐藤伸明", "Nobu Sato", "左滕" ],
  "description" : "Japan market launches: organization, sales and technology",
  "knowsAbout" : [ "Japan market entry", "Organization design", "Talent acquisition", "Sales organization building", "Technology organization building", "Japanese labor law" ],
  "name" : "Nobu Sato",
  "sameAs" : [ "https://www.linkedin.com/in/realnobusato/", "https://x.com/nobusato5888" ],
  "url" : "https://nobusato.jp/en/profile"
}
```

```json
{
  "@context" : "https://schema.org",
  "@type" : "BlogPosting",
  "author" : {
    "@type" : "Person",
    "name" : "Nobu Sato",
    "url" : "https://nobusato.jp/en/blog/author/nobu-sato"
  },
  "dateModified" : "2026-11-04T17:38:50.059Z",
  "datePublished" : "2026-11-04T17:41:02.000Z",
  "headline" : "Why Interview Scores Drift to the Middle, and How to Stop It",
  "image" : [ "https://nobusato.jp/hubfs/covers-2026-11/interview-scores-drift-to-the-middle-cover-en-2026-10-m.png" ],
  "mainEntityOfPage" : {
    "@id" : "https://nobusato.jp/en/blog/interview-scores-drift-to-the-middle",
    "@type" : "WebPage"
  },
  "publisher" : {
    "@type" : "Organization",
    "logo" : {
      "@type" : "ImageObject",
      "url" : "https://nobusato.jp/hubfs/seo/logo-nobusato-2026-10.png"
    },
    "name" : "nobusato.jp"
  }
}
```