Response quality & fair review
Real opinions. Clear standards. Fair outcomes.
Buyers pay for genuine human responses to the question they publish. Participants earn the displayed reward for authentic work that meets the requirements shown before they answer. Payment does not depend on giving a positive answer or agreeing with the buyer.
1. What a valid response looks like
- Read the question, relevant background, options, and task requirements before responding. Consider the material that is actually provided.
- Give your own honest answer. Select the option or scale value that reflects your view; for multi-select questions, follow the selection limit; for rankings, rank every option once in your genuine order of preference.
- When a written response is requested, address the question directly and provide the required level of detail. Explanations should identify a relevant reason, observation, example, or improvement. Meeting the character limit with repeated text is not sufficient.
- A quick vote, rating, or scale with no explanation requested does not require a written justification. Choosing any available option can be valid.
- Submit one response per person per study. Use your own account and represent your age, eligibility, and claimed experience honestly. Participants must be 18 or older.
2. What can be rejected and left unpaid
Rejection requires a specific, documented reason considered in the context of the study. Examples include:
- Random clicking or button-mashing: evidence that selections were made without reading or considering the task. Repeatedly selecting the same option or responding quickly is a signal to investigate, not proof by itself.
- Unusable answers: gibberish, irrelevant text, repeated characters, filler, materially missing required content, or generic comments that do not address a requested explanation. For example, padding “looks good” to meet a length requirement is not meaningful feedback.
- Copying or duplicate participation: copying another person’s response, reusing unrelated canned answers, submitting more than once, or using additional accounts to evade the one-person rule.
- Automation or manipulation: bots, scripts, auto-clickers, AI-generated responses presented as personal human feedback, or coordinated voting intended to manufacture a result. Accessibility and assistive tools are allowed when the person still considers and supplies their own answer.
- Material deception: knowingly inventing required experience or eligibility, impersonating someone, or fabricating observations about material the participant did not review.
- Failure to follow the task: materially ignoring clear, relevant requirements that were available before participation. Buyers cannot add undisclosed requirements after submission.
- Harmful submissions: unlawful content, targeted abuse, threats, or deliberate disclosure of someone’s private information. Relevant, respectfully expressed criticism is allowed.
A low-quality response is not automatically proof that someone intended fraud. The reviewer should describe the actual defect rather than make an unsupported accusation.
3. What is not a valid reason to reject
- Disagreeing with the buyer, choosing a less popular option, giving a low rating, expressing a genuine negative view, or failing to support a preferred political position.
- An honestly selected neutral, “not sure,” “none,” or “prefer not to say” option when the study offers it.
- Minor spelling or grammar mistakes, a different writing style, or accessibility-related differences when the response remains understandable and meets the task.
- Speed, answer patterns, or a suspicion of AI use without sufficient supporting evidence and contextual review.
- A technical problem or unclear question caused by the site or study material, the buyer changing their mind, or the buyer finding an authentic response commercially unhelpful.
Feedback is paid for authenticity and task compliance, not for its sentiment, usefulness to a marketing claim, or agreement with other participants.
4. How review works
Submitting a response does not immediately create withdrawable earnings. The operator reviews pending responses for validity. Study validation and account controls support that review, but do not prove every response is authentic. We do not promise perfect fraud detection or publish a guaranteed review turnaround.
For a rejection, the operator must select a quality category and record a specific explanation. Participants can see the decision, reason, and response reference in Help & support. Buyers can report a concern through support; they do not directly approve, reject, or withhold another person’s reward.
Published rewards apply to approved responses. Quality review is separate from the settlement hold, payout minimum, recipient verification, and bank processing described in the payment terms. A pending or held balance is not a bank payment.
5. Rewards and creator commissions
A rejected response earns $0 in participant rewards and $0 in respondent-referral commission. It is excluded from the buyer’s approved results and delivered-response count. Creator earnings arise from eligible approved activity, never from a signup alone or manufactured submissions.
Buyer-referral commissions are based on eligible service revenue and adjust when that revenue is refunded or disputed. An unrelated person’s valid, approved work is not made unpaid because another response was rejected.
This policy does not automatically confiscate an entire account balance or authorize an additional bank debit. A concern discovered after approval or payout requires separate case review through support. Repeated or deliberate abuse may lead to restrictions on future participation after review.
6. What happens to the buyer’s money
Rejected responses are not treated as purchased work delivered. While collection remains open, the remaining budget can fund valid replacement responses. The buyer can stop collection early through My tests. Once pending responses have been reviewed and the study closes, any undelivered response budget is refunded to the original payment method.
The refund covers the unused portion of the purchased response price, not just the participant’s reward. If zero responses were approved, the entire study price, including setup, is refundable. If valid responses were approved, their costs and the $1 setup fee are retained. For example, if 100 new-price quick responses cost $51 and only 80 are approved at closure, the unused 20 × $0.50 is refunded: $10.00.
A rejection is not an immediate separate cash refund: replacements may still be collected, pending submissions must be reviewed, and the processor takes time to return funds. The same missing response is not refunded twice. Report concerns about an already approved response to support; a complaint does not automatically void valid work or guarantee a full-study refund.
7. Question a decision or report a problem
Use Help & support and include the response or study reference, what you believe is wrong, and a concise explanation. Participants can challenge a rejection; buyers can report a suspected invalid response or a refund issue. Keep passwords, bank details, identity documents, and other people’s private information out of the message.
The operator reviews the concern and records a reply in your support inbox. An appeal does not automatically approve a response, change an earnings balance, or stop a buyer refund that is otherwise due. Any necessary correction is handled as a separate reviewed action. Contact us promptly so the context is available; no fixed response-time guarantee is made.
8. Buyer responsibilities and research limits
Ask one clear, appropriately scoped question. Supply readable material, publish relevant requirements in advance, and do not request fabricated reviews, false endorsements, coordinated political manipulation, or deception. Do not reject or pressure participants because you dislike an honest opinion.
Quality checks reduce unusable data; they do not make an opt-in panel representative, verify voter status, prove purchase intent becomes a sale, or guarantee an accurate prediction. Apply the methodology and sampling limitations shown with study results.