Skip to content

Lincoln and Guba's Trustworthiness Criteria (1985)

  • by Ricardo De Castro King
By Ricardo De Castro King · University lecturer for 16 years · Psique Académica
Written by Ricardo De Castro King, university lecturer for 16 yearsPublished

Definition

Lincoln and Guba's criteria for rigor (1985) are the quality standards for qualitative research: credibility, transferability, dependability, and confirmability. In the qualitative approach, they correspond to internal validity, external validity, reliability, and objectivity.

Core thesis

In qualitative research, rigor is built through auditing the process and triangulating sources, not through statistical tests. In the 1980s, Guba and Lincoln proposed four criteria — credibility, transferability, dependability, and confirmability — that are still the reference framework for judging the quality of an undergraduate, master's, or doctoral study. The qualitative researcher's job is to produce a documentary trail (transcripts, memos, codebook, category versions, methodological decisions) that lets a peer reconstruct how they reached their interpretations and, on that basis, discuss whether the study opens a path in other contexts.

Mnemonic C-T-D-C (“Can The Data Corroborate?”): credibility (the functional equivalent of quantitative internal validity), transferability (external validity), dependability (reliability), and confirmability (objectivity). Each pillar demands a concrete practice in the field notebook, not just a statement in the final report. Whoever signs a thesis with no audit trail isn't defending a pillar: they're just declaring it.

Before getting into each pillar, it helps to separate two levels. The epistemological level asks about the truth of knowledge and has its own criteria, which are worked through in depth in the notes on epistemological foundations of qualitative analysis. The level of methodological rigor, which is our focus here, asks how a qualitative study shows its internal guarantees in an auditable way. The two levels overlap, but these notes concentrate on the second: the concrete practices that support (or erode) trust in an undergraduate, master's, or doctoral qualitative study. In the end, if you want to defend your study before a peer or a committee, this is the level that gets audited.

1. Credibility and member checking

In the qualitative paradigm, credibility plays the role that internal validity plays in the quantitative one: the degree to which findings faithfully represent the phenomenon studied from the perspective of those who live it. The core difference is how it's secured: quantitative internal validity comes from controlled design (randomization, placebo, blinding), while credibility is built through prolonged presence in the field, systematic cross-checking of sources, and returning findings to the people who shared their accounts. One path doesn't replace the other: they answer different questions about how we know.

In Naturalistic Inquiry (1985), Lincoln and Guba list five credibility strategies, used according to the design and the resources available. They're usually presented in increasing order of cost and depth, although in practice they're combined:

Prolonged engagement
→
Persistent observation
→
Triangulation
→
Peer debriefing (external peer review)
→
Member checking

Prolonged engagement means staying in the field long enough for participants to get used to the researcher's presence and show their everyday practices, not just the ones they put on display for a stranger. Persistent observation gives you time to tell what belongs to the phenomenon apart from momentary reactions, and to spot patterns that only show up when you stay and observe for several weeks. Triangulation is developed in section 5 of these notes. Peer debriefing (external peer review) means someone outside the study reviews the interpretations without having taken part in data collection, ideally with an anonymized data package. Member checking means taking preliminary results back to the people who shared their accounts so they can confirm, challenge, or expand on what the researcher heard.

Member checking as a conversation among equals. Morse, Barrett, and Mayan (2002) insist that member checking calls for genuine conversation, not a ritual where the person reads a summary and signs a form. You can tell the difference by whether the researcher is willing to change categories when a participant offers data that challenges them, or whether the exercise boils down to collecting a formal sign-off before the defense. A thesis without substantive member checking is a thesis missing the last link of credibility.

In the Colombian context, credibility also rests on compliance with Resolution 8430 of 1993 of the Ministry of Health, which regulates research with human subjects, and with Law 1581 of 2012 on personal data protection. A study without signed, safeguarded consent forms has no initial credibility, no matter how many years of fieldwork back it up. The regulations set the entry threshold for the rigor debate; rigor sets the ceiling the study can aspire to.

Exam trap 1 — “I trust my clinical experience.” Being familiar with the phenomenon from having lived it or treated it does not substitute for credibility when systematic fieldwork is missing. Credibility requires transcripts, analytic memos, and cross-checking against other sources, not a résumé that vouches for the researcher in front of the committee. Someone who has spent fifteen years treating patients has a valuable input, and should use it as an input, not as a shortcut.
What prolonged engagement looks like in the field notebook

A good indicator is a notebook with dated, progressive entries, descriptions of critical incidents, verbatim fragments of dialogue, and the researcher's reflections kept clearly separate from what was observed. A notebook with three entries a year, written retrospectively and with no separation between description and interpretation, doesn't support the credibility criterion, however much the author claims to have spent months on site. The discipline of dating every entry and keeping the descriptive and the interpretive in separate compartments of the notebook is, in itself, a credibility practice that an expert reader learns to spot within a few pages.

2. Transferability and thick description

Transferability plays the role of external validity in the qualitative paradigm: the possibility that a study's findings can inform a context other than the one studied. The operational difference runs deep: in quantitative research, the researcher generalizes with probabilities derived from a representative sample, whereas in qualitative research the reader generalizes, not the author. This shift in responsibility changes how the report is written at the root, because it obliges you to hand the reader the clues that, in another paradigm, were hidden behind the sampling design.

To make that transfer workable, Lincoln and Guba adopted thick description from Clifford Geertz: texts so contextualized that a colleague in another setting can recognize which conditions they share and which differ from their own, and decide for themselves whether the finding helps them think about their case. Transferring is not the same as comparing probability samples: transferring means giving the reader enough contextual elements to judge relevance for themselves, with the detail that only qualitative immersion makes possible.

Sandoval Casilimas, in his adaptation of the framework for Latin America, written while at Colombia's ICFES (the national education assessment institute), expanded thick description to include, where relevant, dialects, forms of sociability, references to local historical memory, and specific political circumstances. In a thesis on community work in rural areas of the Pacific coast or the southern Andes, those elements are not contextual ornaments but a constitutive part of the data: without them, the finding loses the texture that lets a reader recognize the phenomenon.

Qualitative logic — the reader judges

  • Transferability: thick description of the context (Geertz, Sandoval Casilimas).
  • Dependability: process audit and an accessible documentary trail.
  • Confirmability: explicit reflexivity and traceability of every interpretation.
  • Credibility: triangulation, persistent observation, and substantive member checking.

Quantitative logic — the author demonstrates

  • External validity: probability sampling and statistical inference.
  • Reliability: internal consistency of the instrument (Cronbach's alpha, test-retest).
  • Objectivity: standardized instruments and blinding procedures.
  • Internal validity: controlled design, placebo, and randomization.

The parallel in the table above isn't meant to translate one criterion into another, but to remind you that rigor lives in different practices. A rigorous qualitative study documents its context; a rigorous quantitative study controls its conditions. Forcing the second onto the first, or vice versa, erodes rigor instead of strengthening it, and exposes paradigm errors that a seasoned committee spots quickly.

Data collection
Descriptive notes
Analytic memos
Coding
Categories
Auditable report
Key elements of a thick description

A thick description usually includes: precise physical location (neighborhood, rural district, institution, city), dates and duration of the fieldwork, the sociodemographic and educational profile of the participants, regulatory frameworks relevant to the phenomenon, relevant language or dialect variants, historical events that shape how the phenomenon is read (peace processes, armed conflicts, election cycles, pandemics, natural disasters), and references to local theorists who have worked on the question. How much of each element is enough depends on the study, its research question, and the level of transfer it aims to enable.

3. Dependability and the process audit

Dependability plays the role of reliability in the qualitative paradigm: the possibility that another researcher, starting from the same data and the same context, reaches similar findings. In naturalistic logic, exact repetition isn't expected, because the phenomenon changes over time and with the observer. What's required is that an auditor can reconstruct the analytic path from dated, preserved materials, not that they obtain the same numbers as the original author.

The practical operationalization of dependability is the audit trail (the trail an auditor follows): an ordered set of evidence supporting each interpretive step. Its usual components, ordered by the point in the study at which they're produced, are:

  1. Verbatim transcripts of interviews and focus groups, accompanied by the audio file and the participant's signed authorization for its use.
  2. Dated analytic memos recording emerging hypotheses, surprises, and coding decisions as they happen.
  3. Codebook showing how categories evolved across versions (v1, v2, v3…) and a written justification for each merge, split, or discard.
  4. Descriptive and reflective field notes clearly separated typographically (for example, in columns or with labels) so the auditor can tell what was seen from what was interpreted.
  5. Diagrams of relationships between categories, with an explicit timestamp for each revision.
  6. Report drafts with visible marks of the changes made after peer review and after member checking.

Morse, Barrett, and Mayan (2002) stress that qualitative rigor must be thought through proactively from the design of the study: how many cases to include, which inclusion criteria to use, and how the evidence will be stored are decided before entering the field, not when you're already writing the results chapter. If the audit trail is reconstructed at the end, it tends to be incomplete and biased by what the researcher already believes they found, and the committee notices when it asks for the supporting documentation.

Working rule for dependability

Thinking about dependability means asking yourself one concrete question at every field decision: what would I put in writing today so that a colleague could audit my decision six months from now? The answer defines the content of the audit trail and, in practice, separates defensible theses from those that rest on the author's word. This working rule applies to big decisions (inclusion criteria) as well as small ones (the label on a code).

Warning signs in the audit trail

The audit trail is fragile when: the final categories appear in the report with no earlier versions; transcripts are summarized or paraphrased instead of transcribed; memos are written only once at the end of the project, in a single block, losing the time sequence; the codebook doesn't record which codes were merged with which and why; sampling decisions aren't justified with an explicit date; or the files backing each claim aren't mentioned in the body of the report. Each of these signs indicates that the study's dependability rests more on the author's rhetoric than on the available material, and it's exactly the kind of point a committee tends to catch first.

4. Confirmability and reflexivity

Confirmability is the qualitative equivalent of objectivity: the guarantee that it's the findings, and not the researcher's biases, that speak in the report. As with dependability, the naturalistic strategy doesn't pursue an impossible neutrality, but rather traceability of interpretation combined with an explicit practice of reflexivity. The neutrality of the external observer is replaced by the traceability of the process and the declared honesty of the involved observer.

Reflexivity has at least two levels. The epistemic level spells out the theoretical assumptions the researcher brings to the phenomenon (which authors guide them, which categories they bring from the literature, which debates they set aside when defining the problem). The personal level spells out the relevant biography: professional training, political activism, migration experience, gender, social class, anything that might shape how they listen. Sarah Tracy (2010) captures this practice under the criterion of sincerity: the reader should be able to infer why the researcher arrived at the categories they propose, without having to rely solely on their good faith or reconstruct the process outside the text.

Galeano and Aristizábal, in the Antioquian tradition of social research, call collective reflexivity the practice of a team that meets periodically to confront its blind spots: categories that one team member doesn't see because of their background may be visible to another. It's a concrete way to operationalize confirmability when the study is carried out by a group rather than one person, and it replaces the romantic figure of the solitary researcher with that of a team that audits itself. Collective reflexivity works especially well when the team includes varied educational profiles (clinical, social, artistic) that see different dimensions of the same data.

Exam trap 2 — reanalyzing qualitative data with quantitative software to “validate” it. Running a cloud of qualitative codes through a hierarchical cluster analysis or computing a Cronbach's alpha from emergent categories mixes two paradigms and produces a false sense of rigor. Statistical software measures numerical regularities; the qualitative criteria for rigor live in another epistemological register and operate on qualitatively different materials. Replacing the verbatim with a numerical matrix and running tests on it is abandoning qualitative rigor, not strengthening it.
How to write the reflexivity section

A reflexivity section is usually written in the first person singular even if the study is a team effort, because it spells out the position of the person who signs as the study's author. It includes, at least: educational background (disciplinary training, years of practice if applicable), epistemological stance (constructivist, interpretive, critical, among others), prior experience with the phenomenon, conflicts of interest, and the concrete ways in which their effect on the fieldwork was minimized. It's also worth including specific examples: a quote from the field notebook showing how a moment of epistemological discomfort with a participant was handled, or a methodological decision that changed as a result of the team's collective reflexivity.

5. Triangulation as a rigor strategy

Triangulation was born in quantitative research as a technique for cross-checking measurements, and it was Norman Denzin who extended it to the qualitative field. His central contribution was to distinguish four vectors of cross-checking that a qualitative study can combine according to its research questions, without having to use all four in every design. Each vector addresses a different source of potential bias.

  • Data triangulation: compare sources of different kinds (interviews, observations, official documents, personal memoirs, digital archives, photographs).
  • Investigator triangulation: have several people with different backgrounds code the same material and discuss divergences with the data in hand.
  • Method triangulation: approach the same phenomenon with different techniques (interviews + focus groups + participant observation + document analysis).
  • Theory triangulation: test interpretations against alternative theoretical frameworks, not just the researcher's favorite one.

The COREQ (Consolidated Criteria for Reporting Qualitative Research, specific to interviews and focus groups) and SRQR (Standards for Reporting Qualitative Research, for general use) checklists capture the elements a qualitative study must report so that an outside team can judge it. They are instruments of transparency, not of credibility: they help the reader assess whether the report has the minimum elements to discuss its rigor, and help the team catch omissions before the defense.

Expert-witness analogy. A forensic report is considered solid when the expert presents their methodology step by step, lays out the documents consulted, records who was interviewed, and leaves in plain view the theoretical models used to interpret the signs. Triangulation serves a similar function in qualitative research: it doesn't seek a single irrefutable conclusion, but a set of fronts that prop each other up and make the author's reasoning visible. A good forensic expert, like a good qualitative researcher, leaves their methodological footprint as legible as their conclusion.

An illustrative nearby case is the work of Esteban Valderrama on depression in 22 women from Manizales, Pereira, and Armenia. When the Chilean committee asked how he responded to the objection that his categories might be personal projection, he presented the codebook with its versions, the memos from the joint coding sessions, and the verbatim fragments backing each category. That documentation was the audit trail that made the interpretation auditable, without claiming an impossible neutrality. The case shows that the answer to a confirmability objection is not a rhetorical statement but a package of archived evidence.

Tracy (2010), in her “big tent” proposal for qualitative quality, offers eight concurrent criteria that interact with one another: worthy topic, rich rigor, sincerity, credibility, resonance, significant contribution, ethics, and meaningful coherence. Triangulation touches several of these criteria at once, so it's best understood as a cross-cutting strategy of the design and not as an isolated validation tool. A thesis can be credible, sincere, and resonant without having exhausted Denzin's four vectors; what it can't do is claim credibility without having made explicit, at least, which vectors it used and why.

Denzin vectorWhat is cross-checkedIndicator in the reportRisk if omitted
DataTypes of source (interviews, documents, observation, digital archives).Table of provenance and detailed description of each corpus.Bias from a single type of source.
InvestigatorsIndependent coding and review of agreement between coders.Qualitative inter-rater kappa or documented consensus meetings.Single voice of the coding team.
MethodsDifferent techniques applied to the same phenomenon.Methods section with combined techniques.Partial reading of the phenomenon.
TheoriesAlternative frameworks for interpreting the same data.Discussion with at least two opposing theoretical frameworks.Confirmation of the author's favorite framework.
Exam trap 3 — transferability as a probability sample. Assuming a qualitative study “generalizes” because it interviewed thirty people is a category error: what more cases buy you, in qualitative terms, is depth and theoretical saturation, not statistical power. Transferability is built with thick description of the context, not with sample size.
Minimum checklist at the close of the report
  • Transcripts archived with identification codes kept separate from personal data.
  • Codebook with at least two dated, justified versions.
  • Signed and dated analytic memos.
  • Sampling decisions justified and explicitly dated in the methods section.
  • Signed informed consent forms in line with Resolution 8430 of 1993.
  • Handling of personal data in line with Law 1581 of 2012.
  • Declaration of conflicts of interest.
  • Documented return of findings to participants, with their reactions and the resulting adjustments.
  • Use of an external checklist such as COREQ or SRQR.
  • Audit log accessible to anyone who requests it within the deadlines the university sets.
Common mistakes when defending rigor before a committee

A demanding committee tends to spot three weak patterns quickly: (i) categories not supported by the quoted verbatim, which force the committee to rebuild the argument on the fly; (ii) member checks formalized as a signature but with no substantive dialogue, where the participant had no real chance to challenge the categories; (iii) rhetorical reflexivity sections, with no anchor in specific methodological decisions, that work more as a disclaimer than as an audit of one's own positioning. These patterns are fought with an audit trail filled in from start to finish and a field notebook kept up to date, not with extra paragraphs in the report's introduction or with appeals to authority that stand in for fieldwork.

These notes work best when read alongside Group criteria and typologies, Rigor in qualitative research, and Attitudes and attitude change: from ABC to ELM, from dissonance to persuasion.

Frequently asked questions

What are the criteria for rigor?

They're the standards for judging the quality of a study. In qualitative research, Lincoln and Guba (1985) proposed four: credibility, transferability, dependability, and confirmability.

What are the criteria for rigor in qualitative research?

Credibility (equivalent to internal validity), transferability (external validity), dependability (reliability), and confirmability (objectivity). They're secured through strategies such as triangulation, auditing, and member checking.

What are the 4 main methods of qualitative research?

The most widely used designs are phenomenology, grounded theory, ethnography, and the case study; many textbooks add narrative research.

Keep going on PsiqueAcadémica

  • Resources to practice on your own: tests and practice exams section.
  • Epistemological framework of qualitative analysis: epistemological foundations of qualitative analysis.
  • How to write the final report: qualitative research report.
  • Data collection tools: qualitative data collection techniques.
  • Common bioethical decisions: ethics applied to qualitative analysis.

PsiqueAcadémica · study material for qualitative research with a regional focus.