Validity Criteria for Qualitative Research
Three slides of the supplied source set out a four-for-four substitution: credibility for internal validity, transferability for external validity, dependability for reliability, confirmability for objectivity. Three slides earlier, the same deck says the pursuit is pointless.
The substitution table
The deck presents the criteria as a straight replacement scheme. Each quantitative criterion has one qualitative counterpart, and the table is offered as a substitution rather than as a second set of tests to pass alongside the first.
THE FOUR PAIRS, EXACTLY AS THE SLIDE SETS THEM OUT
| Traditional criteria for judging quantitative research | Alternative criteria for judging qualitative research |
|---|---|
| internal validity | credibility |
| external validity | transferability |
| reliability | dependability |
| objectivity | confirmability |
Four pairs. The source states no count, and does not require you to satisfy both columns. The slide attributes the table to an online research-methods knowledge base page on qualitative validity; the address was scrubbed from the extract and the page is not otherwise named.
The four criteria as the source defines them
All four definitions are reproduced here in full. This is worth stating plainly: it is the only place in either of these two weeks where a named set is completely defined, and the count of the definitions matches the count in the table.
The definitions, quoted
- Credibility
- "the research results are believable from the participant's perspective."
- Transferability
- "the degree to which the results of qualitative research can be generalised or transferred to other contexts or settings."
- Dependability
- "Emphasizes the need for the researcher to account for the changing context within which research occurs and how these changes affect the way the researcher approaches the study."
- Confirmability
- "refers to the degree to which the results can be confirmed or substantiated by others."
What evidence the source says satisfies each
The definitions are even; the procedures behind them are not. Two criteria carry elaborations describing what the researcher does, and two carry nothing at all. The table below separates the definition from the evidence so you can see where the instruction stops.
CRITERION AGAINST STATED EVIDENCE
| Criterion | What the source says demonstrates it | Status |
|---|---|---|
| Credibility | Nothing. Twelve words of definition and no procedure. No member checking, respondent validation or participant review is described. | No procedure given |
| Transferability | "Researcher does a thorough job of describing the research context and assumptions central to the research." And: "Generalising the results to a different context is the decision of the person wishing to do so." | Two elaborations |
| Dependability | Nothing. The definition states what the researcher must account for and no method for accounting for it. | No procedure given |
| Confirmability | Three options, all modal: the researcher "can" document the procedures for checking and rechecking the data throughout the study; "can" actively search for and describe negative instances that disagree with prior observations; and after the study "can" conduct a data audit examining the data collection and analysis procedures and make judgements about the potential for bias or distortion. | Three optional procedures |
Preserve the modal in all three confirmability items. The source says "can", not "must" — none is stated as a requirement. The second elaboration as printed reads "search for and describe and negative instances", where "any" is presumably intended.
The deck denies validity and then teaches it
This is not a subtlety. It is the single largest conflict in the week, and both positions are stated flatly, twelve slides apart, in the same deck.
Slide 6 — what qualitative researchers believe
- "There can be no objectivity: even the analysing researcher has a lived reality"
- "There is no point in trying to establish 'validity' in any external or objective sense"
- "All that we can hope to do as researchers is interpret our view of the world"
Slides 18-20 — validity in qualitative research
- A four-row table of criteria, headed as criteria for judging qualitative research
- Transferability supplied as the analogue of external validity
- Confirmability defined as substantiation "by others"
- A post-study data audit judging "the potential for bias or distortion"
Three further conflicts inside the criteria themselves
Each of these sits within the four slides, not between the week and something else
Conferred by participants, secured by the researcher
Slide 18 says only participants can claim the credibility of the research. Slides 19 and 20 then set out procedures by which the researcher establishes credibility, transferability, dependability and confirmability. Sole authority is assigned to one party and the instruction is given to the other.
A property of the results, and someone else's decision
The definition makes it a degree — something the results have. The elaboration makes generalising "the decision of the person wishing to do so" — something a reader does. The study notes' comparison table adds a third position, that qualitative research "applies ideas across contexts" as an activity of the research itself.
Two incompatible definitions across the two weeks
This week defines it as accounting for the changing context and its effect on the researcher's approach. The earlier research design week invokes dependability as part of rigour — "the soundness and the dependability" — offered as the qualitative substitute for consistent results. A criterion satisfied by documenting change cannot also be the analogue of one satisfied by repeating a result.
Substantiation by others, on a deck that denies objectivity
Confirmation or substantiation by others is objectivity's function under another name, and the data-audit procedure presupposes a standpoint from which bias and distortion can be identified. Slide 6 states that no such standpoint exists.
The left-hand column is never defined
The whole basis of the comparison is a set of four quantitative criteria, and the supplied material treats them very unevenly.
The consequence for a reader is specific. You can reproduce the substitution table accurately and still not be able to say what two of the four things being substituted for are. If you need those definitions, take them from a methods text and attribute them there — not to this material. The generic treatment of the pair is on Validity and Reliability, and published practice on the same question is on Rigour, Triangulation and Validity in Practice.
Two defects in the slides' own scholarship
Both are worth knowing before you cite anything from these three slides.
Running the four criteria on a live project
The four criteria are usable as a structure for a methods section even where the source leaves the procedures blank, provided you are honest about which column of the table above your evidence comes from. The sequencing below is this handbook's, not the source's.
A workable order
Describe the context and your assumptions in full
This is the one transferability procedure the source actually states. It is also the cheapest to do and the easiest to omit. Write it while the context is in front of you, not at the end.
Document your checking and rechecking as you go
The first confirmability option. It has to be contemporaneous — it cannot be reconstructed after the fact without becoming something else.
Record the observations that disagreed with you
The second confirmability option: actively search for and describe negative instances that disagree with prior observations. The source names the technique in one clause and never develops it; how hard you search is your decision to state.
Decide before you finish whether an audit will happen
The third confirmability option is a post-study data audit of collection and analysis procedures, judging the potential for bias or distortion. It only works if the records it needs were kept, so the decision belongs at design time, not at the end.
State plainly what you did not do
The source supplies no procedure for credibility or dependability. If you did nothing to establish them, say so; if you did something, say where the method came from. Neither answer is a weakness. Attributing an invented procedure to this material would be.
What to carry forward
- Credibility, transferability, dependability and confirmability replace internal validity, external validity, reliability and objectivity one for one; all four are defined, which is rare in this material.
- Evidence is unevenly supplied: three optional procedures for confirmability, two elaborations for transferability, nothing for credibility or dependability.
- The deck denies that validity is worth pursuing and then teaches four criteria for it; the study notes make validity a component of every design. Give both, and pick a side explicitly in your own write-up.
- Internal validity, external validity, objectivity and trustworthiness are named and never defined anywhere in the supplied material.
- The credibility citation runs backwards, has no year and no title, and cannot be followed. Do not reproduce it as a working reference.
Frequently asked questions
Do I have to satisfy the quantitative criteria as well as the qualitative ones?
No. The table is presented as a substitution scheme — four alternatives offered in place of four traditional criteria, not in addition to them. The source does not require both columns and does not describe any project that satisfies both.
How do I demonstrate credibility?
The supplied material does not say. It defines credibility as results that are believable from the participant's perspective, states that only participants can claim it, and describes no procedure for obtaining that judgement. Any method you use — member checking or otherwise — comes from elsewhere and should be attributed elsewhere.
Is transferability the same as generalisability?
The source is inconsistent on this. It defines transferability as the degree to which results can be generalised or transferred to other contexts, then says generalising to a different context is the decision of the person wishing to do so. Generalisability itself is named in the earlier research design week and never defined.
What is a data audit, in this material?
One of three optional confirmability procedures: after the study, the researcher can conduct an audit examining the data collection and analysis procedures and make judgements about the potential for bias or distortion. Who conducts it, against what standard and with what output are not stated.
Why does the deck say validity is pointless and then spend three slides on it?
The material does not explain the shift and does not acknowledge it. Slide 6 states a strong relativist position; slides 18 to 20 supply criteria, definitions and procedures. Both are in the same deck, and the study notes take the third position that validity is a component of every qualitative design.
References and source attribution
- Supplied teaching source, Week 9: The nature of qualitative research — slide deck, slides 18-20 (the criteria table and the four definitions) and slide 6 (the denial of validity).
- Supplied teaching source, Week 7: Research Design — slide 9 (validity and reliability) and slide 10 (context of data gathering, rigour, soundness and dependability).
- Bryman, A. 2016, Social Research Methods, 5th ed., Oxford University Press. Listed in an earlier week's references of the supplied source. The Week 9 slide's credibility citation gives only this surname with no year, no title and a reversed page range, and cannot be resolved from what the slide supplies.
- Trochim, W. M. K. 2006, Research Methods Knowledge Base. An online research-methods knowledge base is cited elsewhere in the supplied source; the Week 9 slide attributes its criteria table to such a page without naming it, so the attribution cannot be confirmed from the material.
- Maxwell 2009 — cited in the Week 9 study notes for the five components of qualitative research design, one of which is validity. No title or publisher is given and the work does not appear in the week's reference list.
Suggested questions for Ask KEVOS
- What procedures does this material actually give for confirmability?
- How should I write a methods section using credibility and transferability?
- Where does the supplied source contradict itself on qualitative validity?
- What is the difference between dependability in the design week and in the qualitative week?
- Which validity terms in this material are named but never defined?