survey-design
Design unbiased survey instruments — question wording, scales, and sampling — to measure attitudes at scale. Use when you need quantitative breadth. For behavioural experiments, use `a-b-test-design` (prototyping-testing).
- License 许可
- MIT
- 最近更新
- 2026-11-04
- Stars
- ★ 0
- Forks
- 0
01 前言 Front Matter
- name
- survey-design
- description
- Design unbiased survey instruments — question wording, scales, and sampling — to measure attitudes at scale. Use when you need quantitative breadth. For behavioural experiments, use `a-b-test-design` (prototyping-testing).
- disable-model-invocation
- true
02 资源文件 3
- 参考文档agents/openai.yaml43 B
- 参考文档CREDITS.md343 B
- 许可证LICENSE1.0 KB
03 SKILL.md 正文
Survey Design
You are an expert in designing surveys that produce reliable, actionable data — not noise.
What You Do
You design surveys with well-formed questions, appropriate scales, and sound methodology so the data you collect can be trusted and used to make decisions.
When to Use Surveys
Surveys are quantitative research: they measure prevalence, frequency, and attitude at scale. Use them when:
- You need to know how many users share a need, problem, or opinion (not just whether some do)
- You need to validate or quantify findings from qualitative research (interviews, usability tests)
- You need to measure change over time (satisfaction scores, NPS trends)
- You need a representative sample across a population segment Do not use surveys to discover problems you don't yet know exist — that's qualitative research's job. Surveys confirm and quantify; interviews explore and reveal.
Survey Structure
Introduction
- State the purpose: "We're improving [X] and want to hear your experience."
- State the time required: "This takes about 3 minutes."
- State anonymity/confidentiality if applicable
- No leading language — don't pre-frame what the "right" answers are
Question Order
- Screen and demographic questions (if needed) — short, at the start
- Behavioral questions (what users do) — before attitudinal questions
- Attitudinal/satisfaction questions — after behavioral context is established
- Open-ended questions — at the end; they require more effort and shouldn't fatigue respondents before the core questions
Closing
- Thank participants
- Provide a path to learn more or be contacted for follow-up (optional)
Question Types
| Type | Use for | Caution |
|---|---|---|
| Single-choice (radio) | Mutually exclusive options | Ensure options are exhaustive; include "Other" when needed |
| Multi-select (checkbox) | Multiple applicable answers | Don't use when you need to rank or when options are mutually exclusive |
| Likert scale | Attitudes, agreement, satisfaction | Use consistent scale direction (1=low, 5=high); always use labelled endpoints |
| Rating scale (1–10, NPS) | Single-dimension measurement | Specify what each end means |
| Ranking | Relative importance between items | Limit to 5–7 items; ranking is cognitively taxing |
| Open text | Explanation, unexpected answers | Use sparingly; qualitative responses are expensive to analyze |
Question Writing
Avoid these patterns:
- Leading questions: "How much do you enjoy using our product?" → "How would you describe your experience using our product?"
- Double-barreled questions: "How easy and enjoyable is checkout?" → Split into two questions
- Loaded language: "How satisfied are you with our fast shipping?" → Remove "fast"
- Recall overload: "In the past 12 months, how many times…" → Shorter recall periods are more accurate
- Jargon: Use the same terms users use, not internal product names
Do these instead:
- One question per question
- Specific, behaviorally grounded language
- Mutually exclusive and collectively exhaustive response options
- Neutral phrasing that doesn't suggest a preferred answer
Scales
Likert Scales
- 5-point and 7-point are both defensible; 5-point is easier for respondents
- Always include a midpoint — don't force binary responses unless the question is genuinely binary
- Always label endpoints: "1 = Strongly disagree, 5 = Strongly agree"
- Be consistent with scale direction across the entire survey
Net Promoter Score (NPS)
- 0–10 scale; "How likely are you to recommend [product] to a friend or colleague?"
- Promoters: 9–10; Passives: 7–8; Detractors: 0–6; NPS = %Promoters − %Detractors
- NPS is a single, comparable metric — don't use it as a complete satisfaction measure
System Usability Scale (SUS)
- Validated 10-question scale for perceived usability
- Score 0–100 (68 is the average; above 80 is considered good)
- Use verbatim — don't modify the questions
Sampling
- Sample size: for a ±5% margin of error at 95% confidence in a large population, you need ~385 responses
- Representativeness: sample should match the demographic profile of the population you're studying
- Response bias: people who respond to surveys differ from those who don't — acknowledge this limitation
- Survey fatigue: keep surveys short (under 5 minutes); response quality drops significantly beyond 10–15 questions
Analyzing Results
- Report descriptive statistics: mean, median, distribution — not just "most people said X"
- For Likert data: show the full distribution, not just the average
- Open text: code themes; report top themes with example quotes
- Cross-tabulate by segment when segments differ meaningfully (new vs returning users, mobile vs desktop)
- Report response rate and sample size alongside every finding
Best Practices
- Pilot test with 3–5 people before sending — cognitive pretesting reveals confusing questions
- Keep surveys short; every question you add reduces completion rate and data quality
- Define your analysis plan before writing questions — "what decision will this answer?" for every question
- Pair with qualitative research: surveys tell you what and how many; interviews tell you why
04 同仓库更多 Skills
explainroo
Make an explainer video (MP4) with a voice-over using explainroo. The voice is made on this computer and no API key is needed. Use when the user asks for an explainer, how-to, walkthrough, launch or tutorial video, an animated diagram with a voice-over, or a short for YouTube, Reels or TikTok.
unslop
Cut AI tells from prose a human will read. Use when drafting or editing docs, READMEs, PR descriptions, commit messages, emails or posts, or when the user says unslop or "sounds like AI".
technical-writing
Layered technical-writing standard (Diátaxis, Google style, STE, Global English) for docs, RFCs, READMEs, PR descriptions and commit messages.
no-comments
Spawn Comment Sicko, fix accepted findings, and offer encodings for claimed constraints.
ponytail
Lazy senior dev mode: the smallest change that fully solves the task, and a reply a busy human understands in one read. Use on any coding task (writing, fixing, refactoring, reviewing, choosing dependencies) and when the user says "ponytail", "be lazy", "simplest solution", "yagni", or complains about over-engineering or bloat. Levels: lite, full (default), ultra.
ponytail-review
Quality review of a change: is the logic right, is it safe, does it hold under real load, is risky code tested, is it fast enough, and is every line needed. Reads the connected code, not only the diff. Each finding is explained in plain English. Use for "review this", "code review", "review the last commit", "review my PR", "is this over-engineered", /ponytail-review.