People answer every survey question that can't surprise you

A score tells you how much. Only the words people write can tell you something you didn't think to ask. In 63 responses to 24 surveys made with Enform, people answered every optional list and score question, picked "Other" once in 18 chances, and left the open box empty about four times in ten. The questions people always answer are the ones that can't surprise you.

That sounds like a case against the open box. Here's the case for it, using those counts, decades of survey research, and what gets people to type.

What we counted

These numbers were worked out with read-only counts in Enform's database. Nobody read any answers. The sample is small, and the last section says exactly what was counted.

Read the first two rows together. People never skipped an optional list or score question, and they almost never stepped outside the list. On a dashboard that looks like a survey that worked.

A full list of answers proves people can click

A list can only return the answers you thought of. Its numbers can't show you what you left out, so they always look complete.

Pew Research Center tested this after the 2008 US election. It asked voters which one issue mattered most to their vote, in two versions. With a list of five, "fewer than one-in-ten (8%) provided a response other than the five they were read." With an open box, "fully 43% of those asked the open-ended version provided a response not listed in the closed-ended version" (Pew).

Same question, two versions. One said the list covered more than nine answers in ten. The other said it missed more than four in ten.

That's an old result. Howard Schuman and Stanley Presser ran a similar test in 1979 on what people want most from a job, and "nearly 60% of the responses to the open question are not included in the five researcher-created responses" (Bit By Bit). Matthew Salganik, who describes the study, says closed questions came to dominate surveys "not because closed questions have been proven to provide better measurement," but because "they are much easier to use."

Our own row, "Other" picked once in 18 chances, sits right next to Pew's 8%. That doesn't mean the lists on those surveys were good. It means "Other" isn't an open question. Our guess at why: it sits at the bottom of a list, and picking it means stepping outside the answers you were offered. Most people pick the closest thing that's there and move on.

Scores have the same problem. A 6 out of 10 is always a valid answer. Two people can give a 6 for opposite reasons, and the score keeps no record of which reason was which.

The open box looks like the weak question

On a dashboard, the open box loses. People skip it, and when they don't, they write very little.

That matches the research, though our skip rate is higher. On Pew's panel, people skip closed questions 1% to 2% of the time on average, while "double-digit item nonresponse rates are common for open-ended questions," at about 18% on average (Pew). Our sample is tiny, so read the gap loosely. The direction is the same.

Then there's length. Three words feels like nothing, but it can be enough to say what happened. Here's an invented example, not from a real survey: a score of 4 out of 10 with the words "renewal price jumped" next to it. The score says something is wrong. The three words say what, and point to a fix that no score could. A list could only have caught that answer if you had already put "price went up at renewal" on it, and then you would have already known.

So the dashboard comparison is unfair. List and score questions ask less and can only say what you expected. The open box is the one place people can tell you something new.

What makes people write

The useful question is how to ask so that more people fill the box in.

Ask about one thing. In Pew's analysis, how hard a question is to answer mattered more than anything else. Medium and high effort questions had median skip rates two and three times those of low effort ones (Pew). "Anything else?" is about as hard as a question gets, because people have to decide what the question is before they can answer it. In our data it got 2 answers in 7 tries.

Ask for a sentence, not an essay. Pew found that questions asking for several sentences had a median skip rate of 18%, against 13% for questions asking for a word or a short phrase (Pew). "A sentence is plenty" under the box costs nothing.

Make the box look like it wants an answer. Jolene Smyth and colleagues tested bigger answer boxes in web surveys. A bigger box "has little effect on early responders to the survey" but "substantially improved response quality among late responders," and any line that made length and quality clear helped both groups (Smyth et al.).

Put it right after the score. "What's the main reason for your score?" asks about something the person thought about a second ago. It turns a number you can't explain into a number with a reason next to it. One box placed there beats five scattered ones.

Reading what comes back

Short answers are easy to label and count. The method is in how to analyze open-ended survey responses, and open-ended vs closed-ended questions covers when to use each format.

Enform writes the survey from a description, you share one link, and it reads every reply and counts what people said. Every count opens to the replies behind it, so you can check the words before you act on a number. Pricing is here.

About this data

• Source: responses stored in Enform's production database, made between 7 January and 6 October 2026, on surveys people built with Enform.
• Excluded: every survey owned by the Enform team's own account, responses with no answers at all, and responses whose browser identified itself as a bot or crawler.
• Left: 63 responses with at least one answer, 41 of them finished, across 24 surveys made by 19 accounts.
• Question types came from each survey's own definition: list (single or multiple choice), score (ratings, scales, 0 to 10) and open (short or long text).
• Open questions asking for a name, company, role, website or location were left out. "Why, what and how" and "Anything else?" questions were picked by matching question wording inside the query.
• Words were counted by splitting each answer on spaces. A skip is a question the survey recorded as reached with an empty answer. Questions never recorded are in neither count, so real skip rates may be higher.
• Privacy: all of this ran as read-only counts. No answer text, survey title or name left the database.

Limits. One survey holds 45% of the optional why, what and how questions, and the 11 answers of five words or more came from just 2 surveys. Some responses may be owners testing their own surveys. We wanted to compare what low scorers and high scorers wrote, but only 7 responses had a low score next to an open question, which is too few to say anything. Treat these counts as a sketch that agrees with the published research, not as a benchmark.

FAQ

Why do people rarely pick "Other" on a survey?

Picking it means stepping outside the list, so most people choose the closest option. Pew found 8% went outside a five-item list, while 43% given an open box wrote an answer that wasn't on it.

Do people skip open-ended questions?

Yes, more than list questions: about 18% of the time on Pew's panel, against 1% to 2% for closed ones. In our small sample, about four times in ten.

How do I get more people to answer an open question?

Ask about one specific thing, ask for a sentence, and put the box right after the score it explains. Broad questions like "Anything else?" get skipped most.

Sources

1. Writing Survey Questions, Pew Research Center
2. Why do some open-ended survey questions result in higher item nonresponse rates than others?, Pew Research Center (2021)
3. Bit By Bit: Social Research in the Digital Age, 3.5.2 Wiki surveys, Matthew Salganik
4. Open-Ended Questions in Web Surveys, Smyth, Dillman, Christian and McBride, Public Opinion Quarterly (2009)