Glossary
The language of market research.
Plain-English definitions of the terms you'll hear when you brief a study — from incidence and quotas to conjoint, trackers and data-quality screening.
- Attention check
- A question or trap embedded in a survey to confirm a respondent is reading and answering carefully, used to screen out inattentive or fraudulent responses.
- B2B research
- Market research conducted among businesses and professional decision-makers rather than general consumers — often low-incidence and screened by role, seniority and firmographics.
- CAWI / CATI / CAPI
- The three main modes of data collection: Computer-Assisted Web Interviewing (online), Telephone Interviewing (phone), and Personal Interviewing (in-person).
- Completes
- The number of respondents who fully finished a survey and passed quality checks — the count that actually counts toward your sample size.
- Conjoint analysis
- A technique that estimates how people value different features of a product by asking them to choose between bundles, revealing trade-offs and optimal configurations.
- Deduplication
- Removing duplicate respondents across sources or waves so no one is counted more than once, protecting data integrity.
- Digital fingerprinting
- Identifying a respondent's device and browser signals to detect duplicates, bots and fraud before they enter a survey.
- ESOMAR
- The global association for market, opinion and social research. ESOMAR guidelines (such as the ESOMAR 28/37) set standards for quality and transparency in sampling.
- Feasibility
- An estimate of whether a study's target audience, sample size and quotas can be delivered in the required markets and timeframe — and at what cost.
- Fielding / fieldwork
- The period during which a survey is live and collecting responses from respondents.
- In-survey checks
- Quality controls that run while a respondent takes the survey — speeding, straight-lining, open-end coherence and consistency logic.
- Incidence rate (IR)
- The percentage of the general population that qualifies for a study. Lower incidence means a harder-to-reach audience and higher cost.
- ISO 20252
- The international quality-management standard specifically for market, opinion and social research organisations and their processes.
- MaxDiff
- Maximum Difference Scaling — a method that asks respondents to pick the most and least important items from sets, producing a clear preference ranking.
- Nat-rep (nationally representative)
- A sample structured (usually via quotas on age, gender and region) to mirror a country's population so results can be generalised.
- Open-end
- A free-text survey question with no preset answers. Responses are later coded into themes; coherence checks help catch low-quality or AI-generated answers.
- Panel
- A managed, profiled group of people who have agreed to take part in research. An owned panel (like Bigfoot Surveys) is recruited and governed by the research provider itself.
- Programmatic sample / API
- Automated sourcing of respondents by connecting platforms directly through an API, enabling always-on sample and live quota management.
- Quota
- A target number of completes for a defined subgroup (e.g. 500 women aged 25–34), used to ensure the sample has the right composition.
- Segmentation
- Dividing a market into distinct groups of people with shared needs or behaviours, so strategy and messaging can be tailored to each.
- Straight-lining
- A low-quality response pattern where a respondent selects the same answer down a grid without reading — flagged and removed in quality screening.
- Tracker
- A study repeated at regular intervals (waves) with a consistent design, so changes in brand health or attitudes can be measured over time.
- U&A (Usage & Attitudes)
- A foundational study mapping how people use a category and what they think about it — behaviours, attitudes, needs and brand perceptions.
- Weighting
- Statistically adjusting a dataset so its composition matches known population figures, correcting for over- or under-representation.