Methodology and question design
The Claform Research Team writes about survey methodology: question wording, scales, sampling and bias.
4 articles
Guides & Tutorials8 min read
Which survey elements are worth an A/B test, how many responses each variant needs, and a worked example of deciding whether the difference is real.
Claform Research Team
Guides & Tutorials9 min read
How to design a customer segmentation survey: the needs and attitude questions to ask, how many responses you need, and how to turn the answers into segments.
Claform Research Team
Data & Analytics10 min read
The biases that distort survey answers (acquiescence, social desirability, order effects, recall, halo, satisficing) with a wording or design fix for each one.
Claform Research Team
Best Practices11 min read
A step-by-step method for designing a survey: start from the decision, pick question types, write neutral wording, add logic, keep it short and pilot it first.
Claform Research Team8 guides & tutorials
Beginner36 exit interview questions grouped by reason for leaving, manager, role, culture, pay and would-return, plus a 12-question exit survey and anonymity rules.
BeginnerA questionnaire is the set of questions; a survey is the whole study that uses it. The difference in one table, plus how poll, form and assessment relate.
BeginnerSpot and fix leading, loaded, double-barreled and ambiguous questions: before-and-after rewrites for ten kinds of survey bias, plus a review checklist.
IntermediateExample engagement questions grouped by driver, the eNPS question, anonymity rules, a launch timeline and how to turn results into two or three actions.
IntermediateWhen a matrix (grid) question is the right format, how many rows and columns to use, what to do on mobile, how to catch straight-lining and analyze the rows.
IntermediateTwelve survey design rules with before-and-after examples: objectives, order, wording, scales, length, mobile layout and a pre-launch checklist to copy.
BeginnerThe main survey question families with an example question for each, when to use it, what the data looks like afterwards, and which types to avoid on phones.
IntermediateThe sample size formula with worked arithmetic, a reference table by population and margin of error, and how to turn a target into an invitation count.
3 survey templates
SurveyA multi-rater feedback survey: 24 behavior statements across six leadership competencies, an overall effectiveness rating and four open questions.
SurveyA 15-question brand awareness survey: unaided and aided recall, familiarity, perception on quality, value and trust, and consideration against competitors.
SurveyA 24-question market research survey: screening, demographics, current behavior, needs and importance ratings, competitor awareness and purchase intent.
20 definitions
A/B testing in surveys is a controlled experiment that shows two versions of a survey element to randomly split groups and measures which one performs better.
Anonymity in a survey means no one, including the organization running it, can link a response to the person who gave it.
An attention check is a survey question with a known correct answer, placed to detect respondents who are not reading the questions.
A double-barreled question is a survey question that asks about two things at once but allows only one answer, so the response cannot be interpreted.
A leading question is a survey question whose wording suggests the answer the asker expects, nudging respondents toward it and biasing the results.
Panel research is a survey method that draws respondents from a pre-recruited group who have agreed to take surveys repeatedly, your own panel or a vendor's.
A pilot survey is a small trial run of a questionnaire with real respondents, done before launch to find confusing questions, broken logic and timing problems.
A pulse survey is a short survey of 3 to 10 questions sent to the same group at a regular interval, weekly to quarterly, to track a few key measures over time.
Quota sampling is a non-random sampling method that sets a target number of respondents for each subgroup, then recruits until every quota is filled.
Random sampling is a method of selecting survey respondents in which every member of the population has a known, equal chance of being chosen.
A relationship survey is a periodic survey that measures how a customer feels about the company as a whole, rather than about one recent interaction.
Response bias is a family of survey errors in which respondents answer in a systematically inaccurate way, such as agreeing by habit or answering to look good.
Sample size is the number of completed survey responses needed to estimate a population value within a chosen margin of error and confidence level.
Straight-lining is a response pattern in which a respondent gives the same answer to every row of a grid or every question on a page without reading them.
Survey bias is a systematic error in how a survey is designed, sampled or answered that pushes results away from the true population value in one direction.
Survey design is deciding what a survey must find out and who should answer it, then writing, ordering and testing the questions to get accurate answers.
Survey fatigue is the drop in willingness and attention when people are surveyed too often or a survey runs too long, lowering response and data quality.
A survey incentive is a reward, such as a gift card, prize draw, discount or donation, offered to respondents to raise the response rate.
A transactional survey is a short survey triggered by one event (a purchase, a delivery, a support ticket) that measures that interaction while it is fresh.
Voice of customer (VoC) is the practice of collecting, combining and acting on what customers say across surveys, support, reviews and interviews.
Start free and turn feedback into decisions with Claform.