Skip to content
Back to search results

Face-to-face Survey — Interviewer-Administered Questionnaire

Introduced by
Established practice formalised in survey methodology (Gallup, Likert, and others from the 1930s–1940s)
Year
1930s–1940s (systematic survey era)
Type
Quantitative / mixed-mode data collection
Family
Data collection
Font size

What does it do?

A face-to-face survey is a structured data collection method in which a trained interviewer meets respondents in person and administers a standardised questionnaire.

The interviewer reads questions aloud, clarifies wording when permitted by protocol, and records answers — either on paper (PAPI) or a laptop/tablet (CAPI). This mode consistently achieves higher response rates and better data quality for complex or sensitive questionnaires than self-administered alternatives, and is the reference standard in large-scale population surveys.

The intuition behind the method

Think of a face-to-face survey as a structured conversation with a script. Unlike a self-administered form sent by post or online, a trained interviewer is physically present to guide the respondent through each question, maintain the intended sequence, and handle clarification requests in a controlled way. The human presence reduces item non-response and allows show-cards, physical stimuli, or complex branching that would be impractical on paper. The trade-off is cost and logistical complexity — every interview requires travel, scheduling, and interviewer time.

History

Systematic face-to-face surveying emerged in the 1930s with the rise of public opinion polling (Gallup, Roper) and government household surveys (U.S. Census Bureau). Rensis Likert's work in the 1930s–1940s formalised question formats for interviewer-administered instruments. The Interuniversity Consortium for Political and Social Research and the Survey Research Center at the University of Michigan codified best practices for large-scale face-to-face surveys through the mid-twentieth century. Computer-Assisted Personal Interviewing (CAPI) replaced paper-and-pencil forms from the 1990s onward, improving skip-logic fidelity and enabling real-time data transmission.

When to use it

Use a face-to-face survey when the target population has low literacy or limited technology access, when the questionnaire is long or cognitively demanding, when complex visual stimuli (show-cards, images, product samples) are needed, or when the research setting makes self-administration impractical (e.g., remote rural areas, household expenditure studies, health surveys). It is the preferred mode when response rates are critical and the research budget can support field operations. Avoid it when the budget is limited (it is the most expensive survey mode), when the topic is highly sensitive and anonymity is paramount (self-administered modes reduce social-desirability bias), or when a rapid turnaround is needed that field logistics cannot accommodate.

Limitations and considerations

  • The most expensive survey mode per completed interview, owing to interviewer salaries, training, travel, and supervision costs.
  • Interviewer effects — slight variations in vocal tone, appearance, or probing behaviour can introduce systematic bias across interviewers.
  • Social-desirability bias is elevated; respondents may give socially acceptable answers in the presence of an interviewer, particularly on sensitive topics.
  • Logistically complex to coordinate at scale: scheduling, geographic dispersion, and back-check procedures add project management burden.
  • Slower than online or telephone modes for data collection and entry, which limits feasibility for time-sensitive research.

Source conditions do not replace independent scientific review or validation by an analysis engine.

Strengths

  • Highest response rates of any survey mode — interviewer presence and rapport markedly reduce refusals.
  • Enables complex questionnaire designs: show-cards, physical stimuli, long batteries, and elaborate skip patterns.
  • Reduces item non-response because interviewers can prompt for a response without departing from protocol.
  • Reaches populations underserved by digital or mail modes — elderly, low-literacy, or rural respondents.
  • Real-time data quality control is possible with CAPI: range checks and skip-logic enforcement during the interview.

Applications

  • National household income and expenditure surveys requiring detailed financial recall from respondents with mixed literacy levels.
  • Health examination surveys combining interviewer-administered questions with physical measurements at the same visit.
  • Electoral and political opinion polls where high response rates and representative coverage of offline populations are essential.
  • Development economics studies in low-income or rural settings where internet penetration is insufficient for online surveys.
  • Marketing and consumer research involving physical product stimuli, taste tests, or packaging prototypes shown to respondents.
  • Crime victimisation and sensitive-topic surveys where interviewers use show-cards or ACASI handoff to protect respondent privacy while maintaining response rates.

How it works

How it works

1

The researcher defines the research objectives and translates them into a fully structured questionnaire with fixed question order, standardised wording, and explicit interviewer instructions. Skip patterns and show-cards are mapped out; the instrument is pilot-tested on a small convenience sample to catch wording problems, timing issues, and skip-logic errors before fieldwork begins.

2

A probability or purposive sampling frame is constructed from a population list, map, or household register. Individual respondents or addresses are selected according to the chosen design — simple random, stratified, cluster, or multi-stage. Target sample size is calculated from the desired margin of error, expected response rate, and design effect introduced by clustering.

3

Field interviewers receive standardised training covering verbatim question reading, probing neutral responses, handling refusals, recording answers accurately, and following skip logic. Training typically includes role-play, mock interviews, and a supervised practice session to ensure protocol fidelity before live fieldwork begins.

4

Interviewers visit each sampled address or pre-scheduled location and administer the questionnaire in person. Questions are read exactly as written; permissible probes are scripted. Responses are recorded on a paper form (PAPI) or entered directly into a laptop or tablet (CAPI), which can enforce skip logic automatically and flag out-of-range entries in real time.

5

Completed interviews undergo quality checks: back-checks (re-contacting a random subset of respondents to verify the interview took place and key answers match), logic editing (checking for internal consistency), and missing-data review. Interviewers with unusually high refusal rates or anomalous data patterns are flagged for supervisory follow-up.

6

Cleaned data are weighted to correct for unequal selection probabilities, non-response, and post-stratification benchmarks. Descriptive statistics, regression models, or other planned analyses are applied to the weighted dataset. Design effects from complex sampling must be accounted for using survey-specific software routines (e.g., SPSS Complex Samples, Stata svy, R survey package).

Common pitfalls

  • Undertrained interviewers — departing from verbatim question reading or using non-neutral probes introduces measurement error that is difficult to detect after fieldwork ends.
  • Skipping the pilot test — questionnaire problems (confusing wording, ambiguous response categories, broken skip logic) discovered during fieldwork are costly to fix and may compromise data quality for already-completed interviews.
  • Ignoring design effects in analysis — treating clustered or stratified probability samples as simple random samples leads to underestimated standard errors and inflated statistical significance.
  • No back-checks — without re-contact verification, fabricated interviews (interviewer fraud) may go undetected and contaminate the dataset.

Frequently asked questions

What is the difference between PAPI and CAPI face-to-face surveys?

PAPI (Paper-and-Pencil Interviewing) means the interviewer records answers on a printed form; CAPI (Computer-Assisted Personal Interviewing) means answers are entered directly into a laptop or tablet during the interview. CAPI enforces skip logic automatically, flags out-of-range responses in real time, and transmits data electronically — reducing transcription errors and speeding up data delivery. PAPI is cheaper when technology is unavailable or unreliable in the field.

How do face-to-face surveys compare with online surveys for response rates?

Face-to-face surveys consistently achieve higher response rates — often 70–90% in well-designed studies — compared with typical online survey rates of 10–40%. The interviewer's physical presence and social interaction reduce refusals. However, the coverage advantage depends on the target population: for digitally connected populations, online surveys can reach comparable response rates at far lower cost.

How should I account for the complex sample design when analysing face-to-face survey data?

You must specify the full design — stratification variables, cluster identifiers, and sampling weights — to your statistical software before running any analysis. Using ordinary regression or chi-square tests on weighted data without accounting for clustering underestimates standard errors. Software routines such as Stata's svy prefix, R's survey package, or SPSS Complex Samples handle design-corrected estimation correctly.

When is a face-to-face survey not the right choice?

Avoid face-to-face surveys when the budget is tight, the questionnaire is short and non-complex, the population is digitally accessible, the topic is highly sensitive and anonymity is critical (self-administered ACASI or online modes reduce interviewer presence effects), or when rapid turnaround is required that field logistics cannot meet.

How many interviewers do I need, and how do I control interviewer effects?

The number of interviewers depends on fieldwork duration, geographic spread, and target interview count. To control interviewer effects: standardise training and monitor protocol fidelity, randomly assign interviewers to respondents where possible, include interviewer ID as a random effect in multi-level models, and conduct back-checks on at least 10% of completed interviews to detect fabrication or deviation.

Related methods

Sources

  1. Fowler, F. J. (2014). Survey Research Methods (5th ed.). Sage Publications.

    ISBN: 978-1452259000
    Copied.
  2. Groves, R. M., Fowler, F. J., Couper, M. P., Lepkowski, J. M., Singer, E., & Tourangeau, R. (2009). Survey Methodology (2nd ed.). Wiley.

    ISBN: 978-0470465462
    Copied.