Quality & trust

Quality policy

How we check survey responses, what happens when one is held for review, and how to raise a concern. Our aim is fairness, not blame.

Three colleagues reviewing work together on a screen and tablets

Where a response can end up

Every completed response goes through the checks its study uses. These are the four ways it can end.

Response submittedThe study's checks run
  1. Passes the checks

    Accepted as it is.

    Participant paid: Yes
    Researcher charged: Yes
  2. Minor signs noticed

    Accepted. The researcher sees a note on what was noticed.

    Participant paid: Yes
    Researcher charged: Yes
  3. HeldA person reviewsApproved

    Paid in full once a person approves it.

    Participant paid: Yes
    Researcher charged: Yes
  4. HeldA person reviewsNot approved

    Not paid. The researcher still sees it, marked as held.

    Participant paid: No
    Researcher charged: No
No response is rejected by a script alone. Holding a response only means a person takes a look.

How much each signal counts

Researchers choose which checks their study uses. These are the signals the checks look for, sorted by how much each one counts.

  • Holds for review

    Enough on its own for a person to take a look.

  • Lowers the score

    Several together can hold a response.

  • Noted only

    Shown to the researcher. Never holds a response.

Pace
  • Holds for review: A section answered far too fast
  • Lowers the score: The whole survey much faster than reading time
  • Lowers the score: Most questions answered almost instantly
Attention
  • Holds for review: A wrong answer to an attention check
  • Holds for review: An answer to a hidden question only bots see
Answer patterns
  • Lowers the score: The same rating on every question
  • Lowers the score: The same answer across a whole scale
  • Lowers the score: Random letters in written answers
  • Lowers the score: Numbers far from everyone else's
Duplicates
  • Holds for review: Answers identical to another response
  • Holds for review: The same browser as another respondent
  • Lowers the score: The same network and browser within a day
  • Lowers the score: Several responses from one network
Location & time
  • Holds for review: Two of profile country, connection and device time zone outside the study's countries
  • Lowers the score: Profile country or connection outside, on its own
  • Noted only: Device time zone outside, on its own
  • Noted only: Answering late at night
Location checks run on every study that targets specific countries. We keep the exact thresholds private so the checks stay fair and hard to game.

For participants

Most people never notice these checks. They answer naturally and get paid. A few habits keep it that way:

  • Keep one account

    One account per person. If someone you live with also takes part on the same phone or computer, tell us so we don't mistake you for one person.

  • Answer screening questions honestly

    If you're screened out, changing your answer and trying again won't get you in. The second attempt is treated as a screen-out too.

  • Read every question

    Take the time the questions need. Rushing a section is the most common reason a response is held.

  • Write in your own words

    For open questions, a short honest sentence is better than random letters or copied text.

  • Don't hide your location

    Many studies are for specific countries. A VPN or proxy can make a genuine response look like it came from somewhere else.

  • Answer yourself

    Don't use bots, scripts or AI tools to answer for you. Researchers are paying for your own opinion.

If your response is held

Being held is not an accusation. It means something needs a second look before the reward is confirmed.

  1. 1

    You're told

    You get a notification that your response is under review, with the reason. Your reward is paused, not cancelled.

  2. 2

    A person reviews it

    Our team looks at the whole response, not just the signal that held it, together with your record on Tayqun.

  3. 3

    Decision

    If the response reflects genuine effort, it's approved and the full reward goes to your wallet. If not, it isn't paid.

  4. 4

    You can ask us to look again

    If you think we got it wrong, contact support. We'll review it personally and explain our decision.

Your trust level

Your dashboard shows your trust level (Bronze, Silver, Gold or Platinum), based on how many of your responses pass the checks. Our team sees it when reviewing held responses and withdrawals. One held response doesn't undo a good record; a pattern of low-quality responses does.

If we later find that a paid response was a duplicate or wasn't genuine, for example when a copy of it arrives afterwards, we may reverse that reward.

For researchers

You choose which checks your study uses and how strict they are, and you see exactly what they found.

Choose your checks

In survey settings, switch on the checks your study needs, set the score a response needs to be accepted, and set the reading pace for sections. In the builder, mark attention-check questions and the items that form one scale.

Pay only for accepted responses

Held responses aren't charged to your budget unless our team approves them. A screen-out costs nothing, or only the partial reward if you chose to pay one.

See what was found, in words

Every response has a quality score and a plain-language note on any issue. Each survey has a quality dashboard, and exports include score, flag and note columns with a column guide.

Re-checked as responses arrive

Some problems only show once a second response arrives, like a copy of an earlier one. We re-check the survey as a whole as it fills and alert you if the share of held responses rises.

Quality guarantee

5%

If more than 5% of a study's accepted responses don't meet the quality bar, we replace them at no extra cost. If you think a response slipped through, tell us and we'll review it.

The data these checks use

To run these checks we use your answers and how long they took, your network address, a random ID stored in your browser, your device's time zone, and the details in your profile. We use this only to check quality and prevent fraud, as our privacy policy describes.

Read our privacy policy

Think we got something wrong?

Contact our support team with the survey name. We aim to reply within a few business days, and we'll always explain our decision.

Last updated: October 2026

Custom quote

Have a study in mind?

Tell us your target audience and timeline - we'll send a tailored quote within one business day.

Get a quote

Or request a 20-minute walkthrough • No credit card needed