Survey methodology turns questions into population inferences
Survey methodology is the applied statistics of human-research surveys—designing samples, building questionnaires, collecting responses, and carefully adjusting results so a chosen subset can speak reliably for an entire population of interest that cannot be measured directly.
As a branch of applied statistics, survey methodology studies how to sample individuals from a population and how to construct instruments—chiefly questionnaires—that elicit accurate answers. Polls, public-health studies, market research, government censuses, and social surveys all depend on it. A survey combines at minimum a sample (or full population in a census), a data-collection method, and individual questions whose answers become analysable data.
Success hinges on representativeness: the sample must mirror the target population—whether a nation's adults, a profession's membership roll, or a school system's enrollees. Each respondent may answer for themselves, their household, or an organisation. Methodologists balance systematic and random errors against cost, deciding how to select members, reach reluctant participants, choose administration modes, train interviewers, verify data files, and adjust estimates.
Selection bias arises when sampling procedures over- or under-represent key groups—say 60 percent males in a sample from a 75-percent-female population. Stratified random sampling divides the population into strata and draws proportional random subsamples to reduce this risk. Design types include cross-sectional snapshots, successive independent samples tracking change without revisiting individuals, and longitudinal panels that remeasure the same people over months or years.
Questionnaires dominate data collection but succeed only when items are valid and reliable. Reliability means consistent results on retest; validity means measuring the construct originally intended. Six construction steps guide designers: decide what to collect, choose administration mode, draft questions, revise, pretest, and specify procedures. Self-report scales measuring attitudes must be carefully worded with clear instructions and minimal distraction.
Source: Survey methodology
