“Short note : Limitations of questionnaire as a technique of data collection.” (1999)
- A questionnaire is a self-administered set of written questions the respondent fills in without an investigator present — its limitations follow directly from that single design choice.
- Every weakness below traces back to the fact that no one is there to check who is answering, to clarify a confusing question, or to reach the sizeable part of the population that cannot fill in a written form at all.
- This is precisely why the schedule — the same set of questions, filled in by a trained enumerator in the respondent’s presence — is the instrument used wherever a questionnaire’s limitations would be fatal to the study.
The Literacy Barrier — Why India’s Large Surveys Use Enumerators, Not Questionnaires
- A questionnaire demands that the respondent can read the questions and write legible answers, which immediately disqualifies every respondent who cannot do both.
- Even with India’s literacy rate now estimated at roughly four in five adults, a substantial share of the population — concentrated among older adults, women, and rural residents — remains outside the reach of a self-administered instrument.
- This is the specific reason the Census of India and the National Sample Survey rely on enumerator-administered schedules rather than mailed or self-filled questionnaires: only a trained investigator physically present with the respondent can collect standardised, large-scale data without excluding the illiterate by design.
Low and Biased Response Rates
- Return rates for mailed or distributed questionnaires are typically low, and — more damaging than the low rate itself — non-random: those who do respond skew toward the more educated, more interested, and more literate segment of the sample.
- This quietly substitutes a self-selected group for the sample the researcher’s design actually intended, biasing findings in a direction the original sampling procedure cannot correct.
No Probing, No Verification, No Follow-Up
- There is no opportunity to probe an ambiguous answer or clarify a question the respondent has misunderstood — the misunderstanding simply becomes part of the data.
- The researcher has no way to verify who actually completed the form — a family member, a more literate neighbour, or anyone else could have filled it in on the respondent’s behalf without this ever being detectable.
- Unanswered or incomplete responses cannot be chased up, since there is no interviewer present to notice a skipped item and request an answer before leaving.
Fixed Categories Impose the Researcher’s Own Framework
- A questionnaire’s pre-set, closed response categories force every answer into boxes decided before the study ever met a single respondent — a structural problem Herbert Blumer identified in his critique of how sociological variables are often defined by researcher convenience rather than derived from respondents’ own lived meanings.
- Aaron Cicourel’s related charge of “measurement by fiat” applies directly here: a respondent whose actual view falls between or outside the offered categories is simply coded into whichever box is closest, silently converting a real complexity into a false precision.
- The result is that a rigidly structured questionnaire can systematically miss what a respondent actually means, even while producing tidy, easily tabulated data.
Wording Effects Are Not a Minor Risk
- Even small changes in a question’s wording can shift results dramatically, independent of any real change in what respondents actually believe.
- A well-documented American survey from 1939 illustrates this sharply: when respondents were asked to place themselves in the “upper,” “middle,” or “lower” class, the overwhelming majority selected “middle” — but when the same question replaced “lower” with “working” class, self-identification shifted dramatically, with far more respondents now willing to place themselves there.
- The lesson generalises well beyond class identification: a questionnaire’s fixed wording, precisely because no interviewer is present to explain or rephrase a term, locks in whatever bias that specific phrasing happens to carry.
- Every one of these limitations is a direct consequence of removing the interviewer from the data-collection process.
- Literacy exclusion, biased response rates, the impossibility of probing or verifying, imposed categories, and wording sensitivity together mean a questionnaire trades reach and low cost for a real loss of data quality and representativeness.
- Where any of these costs would be fatal to a study’s validity — as in most large-scale Indian social surveys — the enumerator-administered schedule remains the preferred alternative.
- The questionnaire’s limitations are therefore not a reason to discard it, but a reason to match it carefully to populations and questions where its weaknesses matter least.