Survey Research
The most common survey mistake is asking for the answer you want. This skill fixes the decision first, designs non-leading questions and uses cross-tabs to find differences.
Workflow
- 2
Design non-leading questions
Ask one thing per question, make options mutually exclusive and exhaustive, and avoid “do you agree our product is excellent?” Ask behavior before attitude and put sensitive questions last.
- 3
Pilot and collect
Pilot with 5-10 people and note where they hesitate or misinterpret. In the live survey, state purpose, duration and privacy, and avoid incentivizing biased sharing.
- 4
Cross-tabulate and extract insights
Do not stop at overall percentages; cross by user type, usage frequency and source. Separate “what most think” from “where the biggest differences are,” which is more actionable.
Recommended prompts
AI Evaluation Rubric Prompt
Design scorable evaluation dimensions, weights and decision rules for model output.
Data Analysis Prompt
Start from a business question and get a framework, metrics and validation steps.
Review Analysis Prompt
Turn a batch of reviews into themes, sentiment, evidence and prioritized fixes.
FAQ
How many responses are enough?
100-300 responses usually reveal trends; segmented analysis needs more. Representativeness matters more than raw count.
Should surveys offer incentives?
Yes, but avoid attracting reward-only respondents. Use screening questions to remove invalid responses and explain data use.
Related skills
Data Analysis Workflow
Start from a business question and move through querying, cleaning, analysis, visualization and conclusion instead of charting first and finding a story later.
Dashboard Design
Design a dashboard that supports one decision and can be read at a glance without misleading.
Long-Form Blog Production
A repeatable pipeline from topic selection and research to a 3,000-word draft and fact-check.