Write the questions
Choose what you need to learn, then adapt the questions and answer choices to your message.
Use the survey template See 12 question rewritesFor marketers preparing a message test
A confusing question can give you a confident answer to the wrong problem. Check your draft for leading wording, combined questions and answer choices that overlap.
The checker flags wording patterns. It cannot tell you how a customer will interpret a question; pretest the survey with people from your intended audience.
Edit this example or paste your own. Separate questions with a blank line; put answer choices below each question. Limit: 3,000 characters here, 12,000 in the full checker.
In the example, someone could be happy with the speed but unhappy with reliability. One answer would hide that difference.
English wording rules run in your browser. Your draft is not uploaded. See the checker’s tested limits.
Choose what you need to learn, then adapt the questions and answer choices to your message.
Use the survey template See 12 question rewritesRecord what would change your decision. Estimate the number of people and the recruitment cost before starting.
Build a decision brief Estimate fieldworkEnter the raw counts from two independent groups. Check the size of the difference and its uncertainty.
Compare two response rates Score rating-scale answersWorking on the copy itself? Count characters, clean pasted text or adapt a survey template.
Examples you can inspect
The checker note records actual software outputs. The survey calculations use labeled fictional data so you can reproduce each step.
12 software checks
We ran twelve selected questions through our actual wording rules. Inspect the inputs, alerts, missed problems and downloadable results.
Worked scoring example
Use 5-point and 7-point Likert scale examples, download a free 1–5 survey template, and learn how to handle neutral and unsure answers.
Worked test decision
Write down the decision, what you will compare, and what would change your mind before seeing the results.
No. It uses a published set of English wording rules. The full checker lets you copy a review or a prompt to use with an assistant of your choice. You decide whether to share that draft elsewhere.
No. The rules miss some assumptions and context-dependent problems. Our test note shows specific misses. Read the question and answer choices together, then ask someone from your intended audience to explain what they mean.
The numerical examples are fictional teaching cases. The checker test note records actual software behavior on constructed inputs. We have not published findings from a consumer survey.
Found a misleading example or a tool problem? Send a correction.
Want to check our examples? Read the methods.