“Do you think ethnomethodology helps us in getting reliable and valid data? Justify your answer.” (2021)
- Harold Garfinkel founded ethnomethodology to study not the causes of behaviour but the taken-for-granted methods ordinary members use, moment to moment, to produce and recognise social order as sensible, orderly, and accountable.
- The question forces a direct technical judgement — reliability (consistency on repetition) versus validity (correspondence between what is measured and the concept intended) — rather than a general endorsement or rejection of the approach.
- The considered answer is neither a flat “yes” nor “no”: ethnomethodology scores strongly on validity but poorly on conventional reliability, and this asymmetry itself has to be explained, not just asserted.
What Ethnomethodology Actually Studies
- Garfinkel’s break from his own teacher, Talcott Parsons, is the starting point: for Parsons, actors are “cultural dopes” who simply enact internalised norms; for Garfinkel, members are skilled, practical reasoners who interpret, improvise, and negotiate order in situ.
- His key concepts describe how this happens: indexicality — the sense of any account or utterance depends irreducibly on the specific context of its use, so no account is ever complete in itself; reflexivity — accounts do not merely describe a situation but help constitute it; and the documentary method — members read particular appearances as evidence of an underlying pattern while using the presumed pattern to make sense of the appearances.
- The signature method, breaching experiments, deliberately violates ordinary, unstated norms — for instance, treating a casual “how are you?” as a request for a literal, detailed medical account — to make the normally invisible background expectancies sustaining order visible through the bewilderment, then anger, that the violation provokes.
The Case for High Validity
- Validity asks whether an inquiry captures what it claims to be studying, and on this test ethnomethodology performs well: it does not impose the researcher’s categories on interaction before understanding it — the failure Aaron Cicourel called measurement by fiat — but instead reveals how order is actually, locally produced by the people living it.
- Because the method studies members’ own practices in their own settings rather than substituting a survey category or a laboratory proxy, its findings have high fidelity to lived social reality — a closeness to the phenomenon that structured, pre-coded instruments typically sacrifice.
- Conversation analysis, ethnomethodology’s most technically developed offshoot (Harvey Sacks, Emanuel Schegloff), sharpens this further: by working from verbatim, repeatedly reviewable recordings and transcripts of talk, it demonstrates that turn-taking follows a describable, empirically verifiable machinery — evidence that validity here is not merely impressionistic.
The Case Against Reliability
- Reliability asks whether the same procedure yields the same result on repetition, and ethnomethodological findings are difficult to standardise in this sense: a breaching experiment’s effect depends on a highly specific context — the relationship between participants, the setting, cultural norms in play — none of which can be exactly reproduced in a different time or place.
- Findings are small-scale and interpretation-dependent; a different researcher immersed in a different but superficially similar setting may notice and report a differently inflected account of “how order was accomplished,” since the researcher’s own interpretive skill is part of the instrument.
- This is not incidental sloppiness — it follows directly from indexicality itself: if meaning is genuinely tied to context, then demanding context-free, repeatable “reliability” is asking ethnomethodology to violate its own founding premise.
A Deeper Complication: The Criteria Are Themselves Members’ Accomplishments
- Ethnomethodology pushes the question one level further: reliability and validity are not neutral, external yardsticks but are themselves practical accomplishments — produced by researchers’ own coding conventions, disciplinary training, and background expectancies about what counts as a “good” finding.
- Applying positivist criteria unreflectively to evaluate ethnomethodology therefore risks a category error, judging a study of locally produced order by a standard of universal, portable law that the approach was built specifically to contest.
- A related, current illustration: as ethnomethodology and conversation analysis are increasingly applied to digital interaction — recorded chat logs, social media exchanges, and human interaction with recommendation systems and chatbots — the same tension resurfaces in a new setting, with scholars debating how far conversational order documented on one platform travels to another, underscoring that the reliability problem is structural to the approach, not a defect of its early, face-to-face studies.
- Weighed together, ethnomethodology is best read as trading generalisable reliability for high, situated validity: it is a poor tool for producing standardised, replicable “data” in the positivist sense, but a strong tool for showing, with unusual fidelity, how order is actually accomplished by the people who live it.
- The honest verdict is not that ethnomethodology fails the reliable-and-valid test, but that the test itself is calibrated for a different kind of knowledge claim than the one ethnomethodology sets out to make.
- Its enduring value lies less in generating portable “data” than in demonstrating a method — that even the most mundane, unremarked interaction is held together by rules that become visible only through disciplined disruption — a genuine contribution that a strict reliability standard is simply the wrong instrument to measure.