Article: Microcopy That Reduces Hesitation
Reviewed: November 2026 · Method version 2 · How these reviews work
Each factual claim in the article was checked against the source it cites, then challenged by reviewers whose job was to find the strongest evidence against it: replication attempts, meta-analyses, later papers that cite the original work, and research from other fields that study the same mechanism. A separate reviewer weighed both sides, and a script re-checked every quoted source. Where a claim was overstated or misattributed, the article was corrected. The research below supports the article as it now reads.
Hesitation often works like a small cost-benefit check.
Hedged ‘is’ to ‘often works like’.
In 1978 Langer, Blank and Chanowitz ran three field experiments on everyday requests, comparing requests with a real reason against requests with ‘placebic’ information that had the form of a reason but added nothing.
No change needed.
When the request was small, people tended to comply without attending to what the reason actually said. Later work is mixed on how far that goes: a 1985 replication found people did respond to what the reason said, so treat it as a tendency, not a law.
Added the 1985 Folkes replication, which found people attended to the content of reasons.
When the request demanded real effort, or was phrased in an unfamiliar way, people paid attention to the content of the reason.
No change needed.
For interface copy, that suggests two things.
Reframed ‘that cuts two ways’ as a suggestion, since the study did not test interfaces.
Marreiros and colleagues, in an online survey experiment, highlighted either positive or negative aspects of companies’ privacy policies.
No change needed.
Participants who saw a privacy message became more conservative about disclosing sensitive and identifiable information than those who saw no mention of privacy, even when the message was positive.
No change needed.
In that study, participants’ stated attitudes toward privacy did not change; their disclosures did.
No change needed.
The authors conclude that people’s ‘dormant privacy concerns may manifest only when consumers are asked to think about privacy.’
No change needed.
A banner promising “We take your privacy seriously” at the top of a form can wake a worry in people who didn’t have one; one study of privacy notices found they lowered trust and purchase interest even when they stressed protection.
Added Kim, Barasz and John (2021), which tested privacy notices directly, and softened ‘wakes’ to ‘can wake’.
Read together, the studies point to a rule of thumb.
Reframed as a rule of thumb and added Tan et al. (2014) on explanations in permission requests.
The GOV.UK Service Manual says to know why you are asking every question and what you will do with the answer, and to drop the question if you can’t say.
No change needed.
Labeling buttons with their result (‘Delete report’, ‘Keep it’) rather than ‘OK’/’Cancel’ makes destructive confirmations clearer.
Support is practitioner guidance (Yifrah), not experimental.
A date, an amount or a named use is reassuring, whereas ‘secure’ and ‘safe’ are just words (specific facts reassure more than adjectives).
Supporting research on concrete language (Hansen and Wanke, 2010) noted.
Kinneret Yifrah’s ‘Microcopy: The Complete Guide’ (Nemala, 2017) is a practitioner’s guide devoted entirely to the small text in interfaces, from form labels to error messages.
No change needed.
These edits were made on November 12, 2026 after the review.
Claim reviewed: Field-level form analytics show a common shape: progress is steady, then drop-off concentrates at one field, such as phone number, card details, date of birth or a Delete button.
Now reads: Field-level form analytics often show the same shape.
Why: This is uncited common knowledge from form analytics. Survey research confirms that drop-off clusters at certain questions, but it also tracks question format and who the respondent is, so ‘one place’ is a simplification.
Claim reviewed: Users who stop at a form field have not run out of motivation; they have run into a question the screen does not answer.
Now reads: Often the user hasn’t run out of motivation. They have run into a question the screen doesn’t answer.
Why: Unresolved uncertainty does make people defer: a classic experiment found people put off a choice until an open question was settled, explanations beside a data request raise approval, and survey respondents who quit did not look inattentive. But the best evidence on form completion also points elsewhere: in a randomized trial, giving families the information they lacked did nothing for a hard form while hands-on help did, breakoff rises with length and awkward formats, and much cart abandonment comes from people who were never going to buy. Unanswered questions are one common cause among several, so the flat denial of motivation overstates it.
Claim reviewed: Hesitation is a small cost-benefit check: momentum carries people through easy steps, while a step that seems to carry risk makes them stop and think.
Now reads: Hesitation often works like a small cost-benefit check.
Why: This restates the idea from Langer’s study that routine requests get little scrutiny and effortful ones get attention. It is reasonable, but the original finding has a mixed replication record and repeated warnings lose their pull, so ‘a risky step makes them stop’ is a tendency, not a rule.
Claim reviewed: Most of the risks that cause hesitation are about money, data or permanence.
Now reads: Many of those risks are about money, data or permanence, and most can be answered in a sentence if the sentence is in the right place.
Why: ‘Most’ is a numeric claim with no data behind it. No source was found that ranks the causes of hesitation, and survey research adds embarrassment and social risk, which the list leaves out.
Claim reviewed: When the request was small, people tended to comply without attending to what the reason actually said.
Now reads: When the request was small, people tended to comply without attending to what the reason actually said. Later work is mixed on how far that goes: a 1985 replication found people did respond to what the reason said, so treat it as a tendency, not a law.
Why: This is what the 1978 study reported. But a 1985 replication using the same photocopier setup found that the content of the reason did change compliance, and its author concluded people were thinking. Newer field work finds general reasons still help with small asks, so the effect is real but weaker and less settled than the famous version. Now cites Folkes, Mindlessness or mindfulness, JPSP 1985.
Claim reviewed: For interface copy, small steps flow when they have the familiar shape of a sensible request, while big asks get read, so the reason beside a card number or passport upload has to hold up under scrutiny.
Now reads: For interface copy, that suggests two things.
Why: This is the author’s extension of a 1978 study of in-person requests and memos to screen copy, which no one has tested directly. Studies of software explanations are mixed: placebic explanations earned similar trust ratings, but people acted on real explanations more often.
Claim reviewed: A banner promising ‘We take your privacy seriously’ at the top of a form wakes a worry in people who did not have one; reassurance is not free.
Now reads: A banner promising “We take your privacy seriously” at the top of a form can wake a worry in people who didn’t have one; one study of privacy notices found they lowered trust and purchase interest even when they stressed protection.
Why: The cited study tested survey prompts about other companies, not banners, but a 2021 study of real privacy notices found the same backfire, so the point holds. It has limits: warm, benevolent wording removed the effect, and survey research finds assurances help when the data really is sensitive. Now cites Kim, Barasz & John, The Bulletproof Glass Effect, Journal of Marketing Research 2021.
Claim reviewed: A specific answer at the one field where the worry already arises meets it where it lives (contextual, just-in-time explanation works better than general reassurance).
Now reads: Read together, the studies point to a rule of thumb.
Why: Neither cited study compared field-level answers with general reassurance, so ‘the studies suggest’ overstates. Other work does support it: explanations attached to a data request raise approval, and confidentiality assurances help mainly where the data is sensitive. Now cites Tan et al., CHI 2014.
Claim reviewed: Stating rules (e.g. password requirements) before input removes the pause an error message would create.
Now reads: Rules stated up front can prevent the pause an error message would create.
Why: Stating password rules up front is common advice, but no study found tests whether it removes hesitation. ‘Remove the pause’ is stronger than anything the evidence shows.
Claim reviewed: Privacy banners and trust badges scattered across a form raise the concern for everyone.
Now reads: Blanket reassurance. A generic privacy banner across a form can raise a worry in people who had none. Trust seals are a different case: in field tests they have raised purchases or made no difference, so test before adding or removing one.
Why: The privacy-banner half has real support: formal privacy notices lowered trust and purchase interest in one set of studies, and in the cited source a mention of privacy made people more cautious about disclosing. The trust-badge half is contradicted by the best evidence: a randomized field test found a seal raised purchases, a field experiment found a seal made no difference while a privacy statement increased disclosure, and a 2026 study found a privacy label lowered felt vulnerability and raised purchase intent. ‘For everyone’ also fails: the notice backfire did not appear in already distrustful people, badge effects depend on brand trust and the person’s privacy attitudes, and in a shopping task privacy cues usually do not trigger privacy thoughts at all. Now cites Ozpolat, Gao, Jank & Viswanathan, The Value of Third-Party Assurance Seals in Online Retailing, Information Systems Research 2013.