Interview Completion Rates by Mode in 2026

Last updated: 9 September 2026

Quick Answer: Completion rates by mode in 2026 span roughly 6 percent for cold phone polls, 30 to 46 percent for web, and 76 percent interviewer-administered. No single 2026 benchmark exists, because published figures are response, cooperation or completion rates on different bases. One 2025 study lifted IVR completion from 1.9 to 44.9 percent by adding a human call first.

Ask a search engine for a completion-rate benchmark and the honest answer is that nobody publishes one for 2026. What exists is a set of well-documented studies, each reporting a rate on its own base, for its own population, after its own number of contact attempts. Put side by side, they are more useful than a benchmark, because they show which lever moved the number.

Every figure carries its source, its year and the thing it actually measured. Nothing is rounded to a benchmark.

What Are Interview Completion Rates by Mode in 2026?

The table below is the current evidence, ordered from the modes that complete least often to the ones that complete most. Read the third column before the second: a response rate, a cooperation rate and a completion rate are different fractions, and the section after the table explains how.

Two rows to read with care

The WhatsApp figure is a vendor's published average, not a peer-reviewed rate, and it is listed because it is the only published completion figure for that mode. The Census row is a mandatory survey with field follow-up, which is why it sits at the top and why no commercial study should be measured against it.

Why Is There No Single Benchmark by Mode?

There is no single benchmark because the published rates are calculated on different bases, for different populations, with different effort behind them. The same mode produces a 6 percent rate in one study and a 76 percent rate in another. Both are correct.

Start with the base. In the South Korean randomized trial published in BMJ Open in March 2026, contact rates for telephone and mobile web were comparable. The gap came from cooperation: 73.2 percent of contacted patients completed the mobile web survey against 52.2 percent by telephone. A page that reports only the headline response rate hides where the mode actually helped.

Population comes next. A systematic review of surgical surveys covering 811 patient surveys and 1,746 doctor surveys found patients responded at 70 percent on average and doctors at 53 percent, in the same modes over the same years. Doctors are harder to reach. The channel does not change that.

Effort is the third lever. The Census Bureau's American Community Survey reached 82.9 percent of housing units in 2024 because response is legally required and non-responders get an in-person visit. Its own table shows what effort costs when it is cut: the 2020 rate fell to 71.2 percent in the pandemic year.

The American Association for Public Opinion Research's Standard Definitions, first published in 1998, exist so that these rates can be compared at all. The association's own caveat is that response rate information alone is not sufficient to determine how much nonresponse error a survey carries, or whether it carries any.

How Do Response Rate and Completion Rate Differ?

Three fractions share the word rate, and each drops a group from the denominator:

• Response rate: completed interviews divided by every eligible unit that was sampled.
• Cooperation rate: completes divided by the eligible people who were actually reached.
• Completion rate, as most platforms use the term: completes divided by the people who started.

Same fieldwork, three rising numbers.

One study shows all three at once. In the South Korean trial, 1,600 patients were assigned to each mode and given five contact attempts. Telephone reached about the same share as mobile web.

Cooperation is where they split. Of those reached, 52.2 percent finished by phone and 73.2 percent on mobile web. Divide by the full 1,600 instead and the rates fall to 22.4 and 32.5 percent.

Qualitative interviews add a fourth fraction that surveys rarely report. The PLOS One comparison of online interview modes tracked an eligibility rate, the share of screened people willing to do a given kind of interview, before any completion rate applied.

Only 49 percent of respondents who had a webcam and a computer were willing to do a video interview. A completion rate calculated after that filter looks healthy. It describes a narrower group.

When a vendor quotes a completion figure, three questions make it comparable:

1. Which fraction is it, and what is the denominator?
2. Who was invited, and from what list?
3. How many were screened out before the count began?

What Do Web, Telephone and Face-to-Face Modes Complete At?

Web modes cluster between 30 and 46 percent on a response basis. Telephone ranges from single digits for cold polls to the low 20s with repeated attempts. Face-to-face runs from the low 20s to the low 70s, depending on country and effort.

Web and mobile web

The European Social Survey is moving every participating country to web and paper self-completion by Round 13 in 2027 to 2028. Under that design most countries achieve 30 to 40 percent, with more consistency across countries than face-to-face produced.

Online questionnaires in the surgical review averaged 46 percent for web links and 51 percent for email. Mobile web with five contact attempts reached 32.5 percent of discharged patients in South Korea. Adding a phone follow-up for non-responders lifted that to 39.3 percent.

Telephone

Pew Research Center's telephone polls fell from about 9 percent to 7 percent in 2017 and 6 percent in 2018. Pew attributes part of the decline to an estimated 3.4 billion robocalls a month and to carrier spam flags on unfamiliar numbers. With a known institution calling and five attempts, the South Korean trial reached 22.4 percent. The earlier piece on why telephone response rates collapsed covers the mechanics.

A Uganda study published in Oxford Open Digital Health in 2025 shows the size of the trust effect on phone modes. A cold IVR survey to 23,288 randomly dialed numbers completed at 1.9 percent; the same IVR survey sent after a human interviewer had called completed at 44.9 percent. The three recorded reasons for hanging up were being busy, not understanding the instructions, and not trusting the caller.

Face-to-face

The European Social Survey's face-to-face specification sets a 70 percent response rate target and a 3 percent non-contact ceiling over a fieldwork period of at least six weeks. Achieved rates have ranged from the low 20s to the low 70s across countries. That spread is one reason the survey is switching modes. In the surgical review, interviewer-administered surveys averaged 76 percent, the highest of any mode, and the review groups in-person and telephone interviewing together in that figure.

What Do WhatsApp and AI Phone Modes Add to the Table?

WhatsApp and AI-moderated phone calls add two rows that the academic literature does not yet report. One is an interview that arrives inside a channel the respondent already uses. The other is a phone interview that can call back at scale without a call center behind it.

WhatsApp interviews

The published figure for the first is Alchemic's 78 percent average completion rate on WhatsApp interviews. It is the brand's own number. Read it as one, with no independent benchmark alongside it.

The mechanism is checkable even if the rate is not. There is no link to open and no app to install. Respondents answer by text or voice note in the language they speak, including code-mixed Hindi and English, and the interview sits in the same thread as their family messages. For a comparison of what a survey and an interview each do on that channel, see WhatsApp survey versus WhatsApp interview.

AI phone calls

The second row is AI phone research, which runs hundreds of simultaneous calls, captures consent on the first turn, dials inside permitted calling windows and scrubs Do Not Disturb registries before dialing. The Uganda result is the relevant evidence. Comprehension and trust, not device ownership, decided whether a phone interview completed. A call that explains itself, in the respondent's language, on a number that can be called back, is working on the two reasons people hung up.

Where the fieldwork runs

Alchemic runs both modes as managed fieldwork or on the client's own lists, publishes 57+ languages including Hindi, Tamil and Telugu, and has fielded in the USA and the UK as well as across India, the Gulf and Southeast Asia. Its sample norms run 200 or more for a concept test and 40 to 80 per wave for a deeper brand track, so a completion figure can be read against a known target.

What a High Completion Rate Can Hide

A high completion rate can hide who was never invited, who was screened out, and who the mode quietly excluded. Three published examples show each one.

Common Mistakes When Reading a Completion Figure

The eligibility filter. In the PLOS One study, audio and video interviews completed at under 40 percent among people who had already agreed to that mode, and scheduling engagement fell 20 percent from email to video. Email interviews had the highest eligibility and completion rates and the lowest no-show rate, at a total cost of 199 dollars against 638 for video.

A video-only study that reports 90 percent completion is reporting on the half who said yes to video. The other half never entered the count.

The composition shift. The European Social Survey notes that its self-completion design draws in younger and more educated respondents than face-to-face did. A web mode can raise the completion rate and change who is in the sample at the same time, which is why the piece on sample validity in AI-moderated research treats completion as a diagnostic rather than a target.

The accuracy question. Pew's studies from 1997 through 2016 found little relationship between telephone response rates and poll accuracy, and 2018 midterm polls with 6 percent response performed well by historical standards. The Center's own caveat is the right one: a low rate raises the risk of error rather than guaranteeing it, and standard weighting corrects some of that risk but not all of it.

Every number in the table above is best used as a check on a study design, not as a score. The question a completion rate answers is whether the mode reached the people the study needed, and the comparison of AI phone interviews, IVR and voice bots covers what each phone mode does to that answer.