Mixed methods research: a glossary of terms

1 + 1 = 3 integration challenge

Also called: 1+1=3, 1 + 1 = 3, one plus one equals three, 1+1=3 challenge

The 1 + 1 = 3 integration challenge is the call, made by Fetters and Freshwater in 2015, for mixed methods researchers to show what the extra work of collecting both kinds of data gained, so that the integrated whole says more than its parts. Onwuegbuzie contrasts it with his 1 + 1 = 1 approach, in which the data are transformed and fused into one whole.

A-paradigmatic stance

Also called: a-paradigmatic, aparadigmatic stance, a-paradigmatic approach

An a-paradigmatic stance is the position that a mixed methods study need not rest on any philosophical paradigm, so methods are chosen for practical reasons to suit the research question. Reviews of mixed methods philosophies associate it with Patton and with Reichardt and Cook, and it is one of several stances researchers take towards the paradigm question, alongside pragmatism and the dialectical stance.

Across-stage mixed-model design

Also called: across-stage mixed model design, across stage mixed-model design

An across-stage mixed-model design is a design that mixes qualitative and quantitative approaches across the stages of a single study, so that its research objective, its data and its analysis do not all come from the same tradition. Johnson and Onwuegbuzie set out six such designs in 2004, treating a study as having three stages: the objective, data collection, and analysis and interpretation.

Arrow (mixed methods notation)

Also called: →, arrow sign, arrow notation

The arrow, in mixed methods notation, shows that one component follows another, as in QUAN → qual, where a dominant quantitative phase is followed by a smaller qualitative phase that builds on it. It comes from Morse's 1991 notation, where the plus sign marks the contrasting case of concurrent components, and chains such as (QUAL + QUAN) → QUAN → QUAL can describe more complex designs.

Between-methods triangulation

Also called: between-method triangulation, across-method triangulation, between/across-method triangulation

Between-methods triangulation is the use of both qualitative and quantitative data collection methods in the same study to examine one phenomenon, so that the methods can check and fill out each other's findings. It is set against within-method triangulation, which stays inside one tradition, and some writers use methodological triangulation to mean this kind alone.

Bryman's rationales

Also called: Bryman's rationales for mixed methods, rationales for combining quantitative and qualitative research, Bryman 2006 rationales, reasons for mixing methods

Bryman's rationales are a list of sixteen reasons for combining quantitative and qualitative research that Alan Bryman drew in 2006 from what the authors of published mixed methods studies said they were doing, including triangulation, offset, completeness, explanation, illustration and instrument development. The list breaks Greene and colleagues' five purposes into more concrete parts and adds new ones, and one study can serve several at once.

Building

Also called: integration through building, building integration, build

Building, in mixed methods research, is integration in which the results of one strand shape how data are collected in the next, as when interview themes and participants' own phrases are turned into survey items. It is one of four methods-level approaches described by Fetters, Curry and Creswell, alongside connecting, merging and embedding, and is typical of exploratory sequential designs.

Commensurability legitimation

Commensurability legitimation is the extent to which the meta-inferences of a mixed methods study reflect a mixed worldview, reached by the researcher making repeated Gestalt switches between a qualitative and a quantitative lens until a third, integrated viewpoint emerges. Onwuegbuzie and Johnson base it on rejecting the idea that the two traditions are incommensurable, and accept that those who think such switching impossible may ignore it.

Compatibility thesis

The compatibility thesis is the view that qualitative and quantitative methods can legitimately be combined in one study, the counter-claim to the incompatibility thesis. Onwuegbuzie and Johnson trace it to Howe and to Reichardt and Rallis, and it underlies their idea of commensurability legitimation.

Complementarity

Also called: complementarity purpose

Complementarity is a purpose for mixing methods in which the results of one method are used to elaborate, enhance, illustrate or clarify those of the other, so that the two throw light on different facets of a question. It is one of Greene, Caracelli and Graham's five purposes of 1989, and Palinkas and colleagues use it for methods answering related questions, such as outcome and process.

Complementary strengths stance

Also called: complementary strengths

The complementary strengths stance is the view that qualitative and quantitative methods each give their own valid insight and can be combined, provided each keeps its integrity and is carried out to its own standards rather than blended with the other. Reviews link it with Brewer and Hunter and with Morse, and it differs from the dialectical stance, which seeks new insight in the tension between paradigms.

Completeness

Also called: completeness rationale

Completeness, as a rationale for mixed methods research, is the argument that a researcher can give a fuller account of the topic being studied by using both quantitative and qualitative research than by using either alone. It is one of Bryman's sixteen rationales, drawn from the reasons authors of published mixed methods studies gave for combining methods.

Complex mixed methods design

Also called: complex design, advanced mixed methods design, advanced design, advanced framework, hybrid design, complex mixed methods designs

A complex mixed methods design is a design that places one or more core designs within a larger framework or joins several of them, such as an intervention trial, a case study, a participatory project or a programme evaluation. Creswell and Plano Clark call these complex designs, Fetters, Curry and Creswell call them advanced frameworks, and Schoonenboom and Johnson propose a hybrid type for such combinations.

Component design

Also called: component mixed methods design, component designs

A component design, in Greene's classification, is a mixed methods design in which the qualitative and quantitative components are carried out independently of each other and kept separate until their results are compared or combined. It is set against the integrated design, in which the components depend on each other, and it is the simpler of the two.

Concurrent mixed methods sampling

Also called: concurrent sampling design, concurrent mixed methods sampling design

Concurrent mixed methods sampling is the selection of samples for the qualitative and quantitative strands at about the same time, so that neither sample depends on results from the other. In Onwuegbuzie and Collins's two-dimensional model it combines with four possible relationships between the samples, identical, parallel, nested or multilevel, to give four of their eight sampling designs.

Concurrent timing

Also called: concurrent, simultaneous timing, simultaneous, parallel timing, concurrent implementation, simultaneity

Concurrent timing is the carrying out of the qualitative and quantitative strands of a study at the same time, or nearly so, shown in Morse's notation by a plus sign, as in QUAN + qual. Schoonenboom and Johnson separate simultaneity from dependence, so strands collected at the same time can still be analysed in a way that lets one shape the other.

Confirm and discover

Confirm and discover is a rationale for mixed methods research in which qualitative data are used to generate hypotheses and quantitative research is then used to test them within the same project. It is one of Bryman's sixteen rationales and matches the logic of an exploratory sequential design that builds hypotheses before testing them.

Confirmation (fit of integration)

Also called: confirmation, agreement, convergence of findings, confirmed findings, confirmed metainference

Confirmation, in judging the fit of integrated data, is the outcome in which qualitative and quantitative findings reach similar conclusions, which gives the results greater credibility. It is one of the three outcomes named by Fetters, Curry and Creswell, with expansion and discordance, and corresponds to agreement in a triangulation protocol, whose other categories are partial agreement, silence and dissonance.

Connecting

Also called: integration through connecting, connecting integration, connect

Connecting, in mixed methods research, is integration in which one strand's data are linked to the other's through sampling, as when interviewees are chosen from survey respondents on the basis of the survey results. Fetters, Curry and Creswell describe it as one of four methods-level approaches, and NIH guidance adds that the first phase's results can also shape what the second phase asks.

Context (mixed methods rationale)

Also called: contextual understanding

Context, as a rationale for mixed methods research, is the justification that qualitative research supplies an understanding of the setting to go with the generalisable findings or broad relationships between variables found by a survey. It is one of Bryman's sixteen rationales, and Schoonenboom and Johnson use it to describe a study in which interviews showed how a measured pay gap arose in the workplace.

Contiguous approach

Also called: contiguous reporting, contiguous narrative, contiguous approach to integration

The contiguous approach is a way of integrating through narrative in which qualitative and quantitative findings appear in the same report but in separate sections, such as survey results in the first part of the results and interview findings in the next. Fetters, Curry and Creswell list it with the weaving and staged approaches, and either kind of finding may come first.

Convergence coding matrix

A convergence coding matrix is a table used in a triangulation protocol that lists the key findings from every data set on one page and records, for each pair of data sets, whether they agree, partly agree, disagree or say nothing about each finding. Tonkin-Crine and colleagues built one to compare interview and questionnaire findings from patients and doctors in a trial's process evaluation.

Convergent design

Also called: convergent parallel design, convergent mixed methods design, concurrent design, concurrent triangulation design, triangulation design, QUAN + QUAL design

A convergent design is a mixed methods design in which quantitative and qualitative data are collected and analysed in the same period, usually separately, and the two sets of results are then merged and compared. It is one of the three core designs, often written QUAN + QUAL, and has also been called the concurrent triangulation design, the triangulation design and the convergent parallel design.

Conversion legitimation

Conversion legitimation is the extent to which quantitising or qualitising data yields interpretable data and sound meta-inferences. It is one of Onwuegbuzie and Johnson's nine legitimation types, and it asks, for example, whether counting themes has stripped them of their context, or whether a narrative profile built from averages describes people who do not exist.

Conversion mixed design

Also called: conversion design, data transformation design, data-transformation design, transformation design

A conversion mixed design is a mixed methods design in which one type of data is transformed into the other, qualitative data into numbers or numbers into narrative, and then analysed again, with the extra findings added to the results. It is one of Teddlie and Tashakkori's five families of designs, and Warfa describes a similar data-transformation design as one of three basic approaches in biology education research.

Core component

Also called: core strand

The core component, in Morse and Niehaus's approach, is the part of a mixed methods study that matches its theoretical drive, inductive or deductive, and that must be complete and rigorous enough to stand on its own. It is written in capitals in Morse's notation, as QUAL in QUAL → quan, while the supplemental component is written in lower case.

Core mixed methods designs

Also called: core designs, basic mixed methods designs, basic designs, core design

The core mixed methods designs are the three designs that Creswell and Plano Clark treat as the building blocks of the field: the convergent, explanatory sequential and exploratory sequential designs. Fetters, Curry and Creswell call them basic designs, and more complex designs, such as an intervention trial or a case study, are built around one or more of them.

Credibility (mixed methods rationale)

Credibility, as a rationale for mixed methods research, is the claim that using both quantitative and qualitative approaches enhances the integrity of a study's findings. It is one of Bryman's sixteen rationales and is broader than the triangulation rationale, which specifically seeks to have the findings of each method corroborate the other.

Critical dialectical pluralism

Also called: CDP, critical dialectical pluralism 2.0, CDP 2.0

Critical dialectical pluralism is a research philosophy, associated with Onwuegbuzie and Frels, that gives dialectical pluralism an explicit social justice aim, treating participants as participant-researchers and researchers as facilitators. It favours culturally responsive, engaged mixed methods research that promotes inclusion, equity and social responsibility, and it uses knowledge to challenge power both within a study and in the wider population.

Cross-case comparison joint display

Also called: cross-case comparison display

A cross-case comparison joint display is a joint display that gives each participant or case its own column, with the quantitative scores and the qualitative summaries or quotations for each theme in the cells. Guetterman, Fetters and Creswell identified it as a type suited to case study designs, using an example that set heart failure patients' interview data beside their self-care and knowledge scores.

Crossover mixed analysis

Also called: crossover analysis, crossover mixed analyses

Crossover mixed analysis is analysis that crosses the boundary between the qualitative and quantitative traditions within a single phase, transforming, merging or interpreting data of one kind with the techniques or logic of the other. Onwuegbuzie and Combs coined the term in 2010, and quantitising and qualitising are its central processes.

Data transformation (mixed methods)

Also called: data conversion, conversion of data, integration through data transformation

Data transformation, in mixed methods research, is the conversion of one type of data into the other, qualitative into quantitative by quantitising or quantitative into qualitative by qualitising, so that the converted data can be integrated with data that were not converted. Fetters, Curry and Creswell describe it as a two-step approach to integration, and it differs from the statistical sense of changing a variable's scale.

Data-analysis triangulation

Also called: analysis triangulation, data analysis triangulation

Data-analysis triangulation is the use of more than one method of analysis, qualitative, quantitative or both, on the data about a phenomenon, so that the results of the different analyses can be compared. A 2024 scoping review of case studies found it used in several studies but found little detail on how results from the different analyses were compared or contrasted.

Dependence (mixed methods)

Also called: dependent components, independent components, dependency, sequential-dependent design, concurrent-dependent design

Dependence, in mixed methods design, is the degree to which carrying out one component relies on the results of analysing another, as when interview questions are written after the survey data have been analysed. Schoonenboom and Johnson treat it as separate from simultaneity, so a study can be concurrent yet dependent, or sequential yet independent.

Design quality

Design quality, in Teddlie and Tashakkori's framework, is the standard for judging the methodological rigour of a mixed methods study, that is, how well its design was chosen and carried out. Together with interpretive rigour it makes up inference quality, the term these authors proposed for validity in mixed methods research.

Development (mixed methods purpose)

Also called: development purpose

Development, as a purpose for mixed methods research, is the use of what one method finds to shape the other, whether in choosing a sample, carrying out the study or deciding how to measure. It is one of Greene, Caracelli and Graham's five purposes, and a qualitative phase that shapes a new questionnaire is the usual example.

Dialectical pluralism

Also called: dialectical pluralism 2.0

Dialectical pluralism is a metaparadigm for mixed methods research, set out by R. Burke Johnson, in which researchers expect and accept difference in paradigms, values and even views of reality, and work through dialogue towards new, workable wholes. It gives teams whose members hold different paradigms a framework for working together while treating the tension between their views as productive.

Dialectical stance

Also called: dialectic stance, dialectical approach, dialectic approach

The dialectical stance is the position, associated with Greene, Caracelli and Hall, that a mixed methods study should bring different paradigms or mental models into open, respectful conversation and treat the differences between them as a source of new understanding. It differs from pragmatism, which puts the research question before paradigm debates, and from the incompatibility thesis, which holds that paradigms cannot be mixed.

Discordance

Also called: discordant findings, discordant results, divergence, divergent findings, dissonance, conflicting findings, inconsistent findings

Discordance, in judging the fit of integrated data, is the outcome in which qualitative and quantitative findings are inconsistent, contradict or disagree with each other. Fetters, Curry and Creswell suggest looking for bias, re-examining methods, gathering more data or seeking an explanation in theory, and the triangulation protocol calls the same outcome dissonance, which can itself lead to deeper insight.

Diversity of views

Diversity of views is a rationale for mixed methods research that covers two linked aims: joining the researcher's viewpoint, often captured in quantitative measures, with participants' own viewpoints, often captured qualitatively, and showing both how variables relate and what those relationships mean to participants. It is one of Bryman's sixteen rationales, and Schoonenboom and Johnson list many forms it can take, from linking local and national knowledge to explaining complexity.

Embedded design

Also called: embedded mixed methods design, concurrent nested design, nested mixed methods design

An embedded design is a mixed methods design in which a smaller strand of one type is placed inside a larger study of the other type to answer a different or supporting question, as when interviews are nested within a randomised trial. It was one of Creswell and Plano Clark's six major designs, and Warfa describes the earlier concurrent nested design in the same terms.

Embedding

Also called: integration through embedding, embedding integration, embed

Embedding, in mixed methods research, is integration in which data collection and analysis are linked at several points, most often by placing qualitative work before, during or after a trial or other intervention study. Fetters, Curry and Creswell describe it as one of four methods-level approaches, which may combine connecting, building and merging, and warn that it must not threaten the validity of the trial.

Emergent mixed methods design

Also called: emergent design, emergent mixed design

An emergent mixed methods design is one in which a component is added or changed while the study is under way, rather than planned from the start, often because one component proves inadequate or produces unexpected results. Schoonenboom and Johnson set it against a planned design and advise planning for emergence where possible, since initiation by its nature follows the unforeseen.

Enhancement (mixed methods rationale)

Also called: building upon quantitative or qualitative findings

Enhancement, as a rationale for mixed methods research, is the aim of adding to or getting more out of the findings of one approach by gathering further data with the other, for instance following survey results with interviews. It is one of Bryman's sixteen rationales, where it is labelled enhancement or building upon quantitative or qualitative findings.

Equal-status design

Also called: equal status, equal-status mixed methods research, equal priority, interactive mixed methods research, QUAL + QUAN

An equal-status design is a mixed methods design in which the qualitative and quantitative components carry the same weight, take turns to lead the research and interact throughout, so that their results are integrated both along the way and at the end. Schoonenboom and Johnson also call it interactive mixed methods research and note that Morse and Niehaus's approach, which always has one core component, leaves no room for it.

Expansion (fit of integration)

Also called: expanded findings, expanded metainference

Expansion, in judging the fit of integrated data, is the outcome in which qualitative and quantitative findings diverge in a way that widens understanding, because each addresses a different aspect of the same phenomenon, such as the strength of an association and its nature. It is one of the three outcomes of fit named by Fetters, Curry and Creswell, alongside confirmation and discordance.

Expansion (mixed methods purpose)

Also called: expansion purpose

Expansion, as a purpose for mixed methods research, is the extension of a study's breadth and range by using different methods for different parts of the inquiry, such as a measure of an intervention's effect and a separate study of how it was delivered. It is one of Greene, Caracelli and Graham's five purposes, though Palinkas and colleagues use the word for qualitative data that explain quantitative results.

Explanation (mixed methods rationale)

Explanation, as a rationale for mixed methods research, is the use of one approach to help explain findings produced by the other, most often interviews that account for a pattern in survey or trial results. It is one of Bryman's sixteen rationales and the defining aim of the explanatory sequential design.

Explanatory sequential design

Also called: sequential explanatory design, explanatory design, explanatory sequential mixed methods design, explanatory mixed methods design, QUAN → qual design

An explanatory sequential design is a mixed methods design in which quantitative data are collected and analysed first and the results then shape a qualitative phase that helps explain them, for instance by deciding whom to interview and what to ask. It is one of the three core designs, usually written QUAN → qual, though qualitatively driven versions exist in which the interviews carry the main weight.

Exploratory sequential design

Also called: sequential exploratory design, exploratory design, exploratory sequential mixed methods design, exploratory mixed methods design, QUAL → quan design

An exploratory sequential design is a mixed methods design that begins with qualitative data collection and analysis and uses the findings to build a quantitative phase, often a new questionnaire, measure or intervention that is then tested. It is one of the three core designs, usually written QUAL → quan, and is sometimes described in three phases: exploration, development of a product, and testing of that product.

Fit of data integration

Also called: fit of integration, fit

The fit of data integration is how well the quantitative and qualitative findings hang together once they have been brought together, judged as confirmation, expansion or discordance. Fetters, Curry and Creswell describe the three outcomes, and researchers may label the rows of a joint display by their fit before drawing meta-inferences.

Following a thread

Also called: follow a thread

Following a thread is a technique for integrating data in which the analyst takes a question, theme or emerging theory from one component of a study and follows it across the other components, rather than comparing all findings at once. O'Cathain, Murphy and Nicholl described it as one of three integration techniques, alongside the triangulation protocol and the mixed methods matrix.

Fully mixed design

Also called: fully mixed methods design, fully mixed research design, fully integrated mixed design, fully integrated mixed methods design, fully integrated design

A fully mixed design is a mixed methods design in which qualitative and quantitative approaches are mixed interactively at every stage of the study, from the objective to the inferences, so that at each stage one shapes the other. Onwuegbuzie and Johnson treat Leech and Onwuegbuzie's term as the same as Teddlie and Tashakkori's fully integrated mixed design, and it contrasts with a partially mixed design.

Fundamental principle of mixed research

Also called: fundamental principle of mixed methods research, complementary strengths and non-overlapping weaknesses, complementary strengths and nonoverlapping weaknesses

The fundamental principle of mixed research is the rule that a study should combine methods, approaches and concepts whose strengths complement one another and whose weaknesses do not overlap, so that the mixture is stronger than any single method. Johnson and Turner named it in 2003, building on Brewer and Hunter, and Johnson and Onwuegbuzie treat it as a main justification for mixed methods research.

GRAMMS

Also called: Good Reporting of A Mixed Methods Study, GRAMMS guideline, GRAMMS 2.0

GRAMMS, Good Reporting of A Mixed Methods Study, is a six-item reporting guideline for mixed methods studies published by O'Cathain, Murphy and Nicholl in 2008 from their review of health services research. It asks authors to justify mixing, describe the design's purpose, priority and sequence, describe each method, explain where, how and by whom integration was done, note limits one method placed on the other, and report what mixing added.

Identical samples

Also called: identical sample relationship, identical sample

Identical samples, in mixed methods sampling, are the same participants taking part in both the qualitative and quantitative strands, as when one class answers a questionnaire that has both rating scales and open-ended questions. It is one of four sample relationships in Onwuegbuzie and Collins's typology, and it avoids the difficulty of combining conclusions drawn from different people.

Illustration (mixed methods rationale)

Illustration, as a rationale for mixed methods research, is the use of qualitative material, such as quotations or case descriptions, to bring quantitative findings to life, giving human detail and voice to results that would otherwise be only numbers. It is one of Bryman's sixteen rationales, and quotations set beside survey results in a table or joint display are a common form of it.

Incompatibility thesis

Also called: incommensurability thesis, incompatibility, paradigm purism, purist position

The incompatibility thesis is the claim that quantitative and qualitative research belong to rival paradigms whose methods may not be combined in one study, since each rests on assumptions about reality and knowledge that the other denies. The label is credited to Howe in 1988, Johnson and Onwuegbuzie describe purists on both sides of the paradigm wars as holding it, and mixed methods research rejects it.

Inference quality

Inference quality is the term Teddlie and Tashakkori proposed for validity in mixed methods research, on the grounds that every study draws inferences whether its reasoning is inductive or deductive. It has two parts, design quality and interpretive rigour, and Onwuegbuzie and Johnson built their typology of legitimation on it.

Inference transferability

Inference transferability is Teddlie and Tashakkori's term for how far the conclusions of a mixed methods study apply beyond it, covering both generalisation in quantitative research and transferability in qualitative research. It has four kinds: to other people or groups (population), to other settings (ecological), to other times (temporal) and to other ways of measuring (operational).

Initiation

Also called: initiation purpose

Initiation, as a purpose for mixed methods research, is the search for paradox and contradiction between the results of different methods, so that questions or findings are recast and new perspectives emerge. It is one of Greene, Caracelli and Graham's five purposes, and because it follows up the unexpected it usually comes after the first results and often makes a design emergent.

Inside-outside legitimation

Also called: insider-outsider legitimation

Inside-outside legitimation is the extent to which a mixed methods researcher represents fairly and uses well both the insider's, or emic, view and the outside observer's, or etic, view. It is one of Onwuegbuzie and Johnson's nine legitimation types, supported by member checking for the insider view and by review from a disinterested researcher for the outsider view.

Instrument development (mixed methods rationale)

Instrument development, as a rationale for mixed methods research, is the use of qualitative research to write or refine the items of a questionnaire or scale, for instance to improve their wording or to offer a fuller range of closed answers. It is one of Bryman's sixteen rationales and a particular case of Greene and colleagues' development purpose.

Instrument development joint display

Also called: instrument development display

An instrument development joint display is a joint display that maps qualitatively derived themes or dimensions onto the items of a new or adapted questionnaire, showing where each item came from. Guetterman, Fetters and Creswell found it in exploratory sequential studies, such as one that matched patients' accounts of continuity of care to existing and new survey items.

Instrument fidelity

Instrument fidelity is a rationale for mixing qualitative and quantitative techniques to judge whether existing instruments are appropriate and useful, to create new ones, or to check how well researchers perform as human instruments. It is one of four rationales set out by Collins, Onwuegbuzie and Sutton, with participant enrichment, treatment integrity and significance enhancement.

Integrated design (mixed methods)

Also called: integrated mixed methods design, integrated designs

An integrated design, in Greene's classification, is a mixed methods design in which the components are interdependent, so that one shapes the other during the study rather than meeting only when results are compared. Schoonenboom and Johnson note that integrated designs are the more complex, since complexity grows with how far components depend on each other.

Integration (mixed methods)

Also called: mixing, data integration, mixed methods integration, integration of qualitative and quantitative data

Integration, in mixed methods research, is the deliberate linking of quantitative and qualitative approaches, data or findings so that they depend on each other in answering a shared question. Most methodologists treat it as what separates mixed methods from running two studies side by side, and it can happen at the level of design, methods, and interpretation and reporting. Schoonenboom and Johnson find the word mixing misleading for this reason.

Integration at the design level

Also called: design-level integration, integration through design, design level integration

Integration at the design level is integration built into how a mixed methods study is conceived, through the choice of a core design, such as explanatory sequential, exploratory sequential or convergent, or of a more complex framework built around one. It is the first of three levels of integration described by Fetters, Curry and Creswell, followed by the methods level and the interpretation and reporting level.

Integration at the interpretation and reporting level

Also called: interpretation and reporting level integration, integration at the interpretation level, integration at the reporting level

Integration at the interpretation and reporting level is the bringing together of qualitative and quantitative findings as they are interpreted and written up, through narrative, through data transformation or through joint displays. It is the third of Fetters, Curry and Creswell's levels of integration, and it is where meta-inferences are usually drawn.

Integration at the methods level

Also called: methods-level integration, integration through methods, methods level integration

Integration at the methods level is the linking of the data collection and analysis of the qualitative and quantitative strands, which Fetters, Curry and Creswell describe as happening through connecting, building, merging and embedding. The design limits the choices: connecting follows naturally in sequential designs, merging can occur in any design and embedding usually belongs to intervention designs.

Integration through narrative

Also called: narrative integration, integrating through narrative

Integration through narrative is the reporting of qualitative and quantitative findings together in words, in one report or a series of reports, and it takes three forms: the weaving, contiguous and staged approaches. It is one of three approaches to integration at the interpretation and reporting level described by Fetters, Curry and Creswell, alongside data transformation and joint displays.

Interactive approach (mixed methods design)

Also called: interactive approach, dynamic approach, interactive model of research design, dynamic design approach

The interactive approach to mixed methods design treats design as a continuing process in which a study's goals, conceptual framework, research questions, methods and validity are repeatedly checked and adjusted so that they keep fitting together. The most often mentioned version is Maxwell and Loomis's, Creswell and Plano Clark call it a dynamic approach, and it contrasts with the typological approach.

Interpretive rigour

Also called: interpretive rigor, interpretative rigour, interpretative rigor

Interpretive rigour, in Teddlie and Tashakkori's framework, is the standard for judging whether the conclusions and meta-inferences of a mixed methods study follow credibly from the data and their analysis. Its five components are interpretive consistency, theoretical consistency, interpretive agreement, interpretive distinctiveness and integrative efficacy, and with design quality it makes up inference quality.

Iterative sequential mixed design

Also called: iterative sequential mixed methods design, iterative sequential design

An iterative sequential mixed design is a sequential mixed design with more than two phases, in which qualitative and quantitative strands alternate, as in QUAN → QUAL → QUAN. It appears in Teddlie and Tashakkori's typology as a more complicated form of the sequential mixed design, and Onwuegbuzie and Johnson suggest that alternating phases can show whether the order of phases shaped the conclusions.

JARS-Mixed

Also called: MMARS, Mixed Methods Article Reporting Standards, JARS–Mixed, APA mixed methods reporting standards, journal article reporting standards for mixed methods research

JARS-Mixed is the American Psychological Association's set of journal article reporting standards for mixed methods research, also cited as the Mixed Methods Article Reporting Standards, or MMARS. It was set out in 2018 in a task force report by Levitt and colleagues, alongside standards for qualitative research, and asks authors among other things to describe their philosophical approach and how the data were integrated.

How to write an APA research paper or lab report

Joint display

Also called: joint displays, joint display table, mixed methods joint display

A joint display is a table, matrix, graph or figure that brings quantitative and qualitative data or results together visually so that insights can be drawn beyond those from either set alone. It is one of three ways of integrating at the interpretation and reporting level, and Guetterman, Fetters and Creswell advise labelling each kind of result, matching the design and integration approach, and stating the inferences drawn.

Legitimation

Also called: legitimation in mixed methods research, mixed methods validity, validity in mixed methods research, legitimation types, legitimation typology

Legitimation is the term Onwuegbuzie and Johnson proposed in 2006 for validity in mixed methods research, chosen as a bilingual word that both quantitative and qualitative researchers could use. They treat it as a process to be checked at every stage rather than an outcome, and set out nine types, from sample integration to political legitimation, for problems that arise only when methods are combined.

Merging

Also called: integration through merging, data merging, merge

Merging, in mixed methods research, is integration in which the qualitative and quantitative databases are brought together for joint analysis and comparison, usually after each has been analysed separately. It is one of Fetters, Curry and Creswell's four methods-level approaches, typical of convergent designs, and works best when both kinds of data were collected on parallel questions, such as scale items matched by open questions.

Meta-inference

Also called: metainference, meta inference, meta-inferences, metainferences

A meta-inference is a conclusion reached by combining what the qualitative and quantitative strands of a mixed methods study each showed, one that goes beyond what either strand could show by itself. Meta-inferences can be written as statements, narratives or theory, and Younas and colleagues distinguish global ones, which extend beyond the people studied, from specific ones, which stay with them.

Methodological eclecticism

Also called: methodological pluralism, eclecticism

Methodological eclecticism is the practice of choosing and combining whatever methods best answer a research question, rather than keeping to one tradition's set of tools. Johnson and Onwuegbuzie call methodological pluralism or eclecticism a defining feature of mixed methods research and argue that it often produces better research than a monomethod study.

Mixed analysis

Also called: mixed methods analysis, mixed methods data analysis, mixed data analysis

Mixed analysis is the analysis of data in a mixed methods study with both quantitative and qualitative techniques in a way that brings their results together, rather than analysing each set in isolation. Onwuegbuzie and Combs list nine purposes it can serve: to reduce, display, transform, correlate, consolidate, compare, integrate, assert and import data.

Mixed methods action research

Also called: MMAR, mixed-methods action research

Mixed methods action research is action research that collects, analyses and integrates both quantitative and qualitative data within its cycles of planning, acting and reflecting, often with participants as co-researchers. O'Sullivan and colleagues, drawing on Ivankova, used it with young co-researchers who helped design the survey and interviews in a study of cold housing and fuel poverty.

Mixed methods case study

Also called: mixed methods case study design, case study mixed methods framework, mixed methods case study research

A mixed methods case study is a case study, of one case or several, in which both qualitative and quantitative data are collected and integrated to build a detailed, contextual understanding of each case. Creswell and Plano Clark count it among complex designs, often built around a convergent design, and comparison across cases is frequently added.

Mixed methods design

Also called: mixed methods research design, MMR design, mixed methods study design, mixed-methods design

A mixed methods design is the plan of a study that combines at least one qualitative and one quantitative component, setting out their purpose, timing, priority and where they will be integrated. Methodologists offer competing typologies of named designs, and Schoonenboom and Johnson argue that most real studies need a design built for their questions rather than one taken whole from a list.

Mixed methods evaluation design

Also called: mixed methods program evaluation design, mixed methods programme evaluation design, mixed methods evaluation, program evaluation design

A mixed methods evaluation design is a complex design in which one or more core mixed methods designs are used, often over several phases, to evaluate a programme or intervention and guide its development, improvement or adaptation. Creswell and Plano Clark list it among their complex designs, and its strengths include publishing each phase separately while contributing to one overall evaluation.

Mixed methods grounded theory

Also called: MMGT, mixed-methods grounded theory

Mixed methods grounded theory is grounded theory research that combines qualitative and quantitative data and analysis to build or refine a theory, keeping the inductive aim of grounded theory while drawing on measurement as well. Anttila and colleagues used it to deepen an earlier qualitative grouping of cardiac rehabilitation patients to the level of a core category.

Mixed methods intervention design

Also called: mixed methods experimental design, intervention mixed methods framework, mixed methods intervention framework, experimental mixed methods design, intervention design

A mixed methods intervention design is a complex design in which qualitative data are collected before, during or after a trial or other intervention study, to shape the intervention, understand how context affects delivery, or explain the results. Creswell and Plano Clark call it the mixed methods experimental design, and Medical Research Council guidance notes that process evaluations of complex interventions usually need both kinds of method.

Mixed methods matrix

Also called: mixed-methods matrix

A mixed methods matrix is a table that sets each participant's quantitative and qualitative data side by side, so that the two can be compared person by person rather than only in aggregate. O'Cathain, Murphy and Nicholl described it as one of three integration techniques, and Etkind and colleagues used one to explore how stable frail older people's care preferences were after acute illness.

Mixed methods notation

Also called: Morse notation, Morse's notation, Morse notation system, mixed methods notation system

Mixed methods notation is the shorthand, introduced by J. M. Morse in 1991, for writing a design as a formula: QUAL and QUAN for the two approaches, capitals for the dominant component and lower case for the supplemental one, a plus sign for concurrent work and an arrow for sequential work. Creswell and Plano Clark later extended it to show embedded strands and recursive processes.

Mixed methods research

Also called: mixed methods, mixed-methods research, mixed method research, mixed-method research, mixed methodology, mixed research, MMR, mixed methods approach, mixed methods study

Mixed methods research is research that combines qualitative and quantitative approaches, data collection, analysis or inference in a single study or programme of studies, and deliberately integrates them to understand a problem more broadly and deeply and to check findings against each other. Definitions vary, but most now require integration, so a study that simply reports a survey and interviews side by side may not qualify.

Mixed methods research question

Also called: mixed methods question, MMR question

A mixed methods research question is a question that can be answered only by bringing together the qualitative and quantitative strands of a study, stated alongside the separate qualitative and quantitative questions. Wu and colleagues, citing the APA reporting standards and Tashakkori and Creswell, recommend asking all three, with the answers to the strand questions feeding the mixed one.

Mixed methods sampling

Also called: mixed methods sampling design, mixed sampling, mixed methods sampling designs

Mixed methods sampling is the selection of participants or cases for the qualitative and quantitative strands of a study, including decisions about whether the samples are drawn at the same time or one after the other and how they relate. Onwuegbuzie and Collins classify designs by timing, concurrent or sequential, and by relationship, identical, parallel, nested or multilevel, giving eight basic designs.

Mixed-model design

Also called: mixed model design, mixed-model research, mixed model research

A mixed-model design, in Johnson and Onwuegbuzie's 2004 typology, is a design that mixes qualitative and quantitative approaches inside one stage of a study or between its stages, such as its objective, data collection and analysis. They set it against a mixed-method design, which runs a separate quantitative phase and qualitative phase and integrates their findings at some point.

MMR-RHS

Also called: MMR-RHS checklist, standards for mixed methods reporting in rehabilitation health sciences research

MMR-RHS is a checklist of standards for reports of mixed methods studies in rehabilitation and the health sciences, developed systematically by Tovin and Wormley and published in Physical Therapy in 2023. The EQUATOR Network lists it with GRAMMS and the APA's reporting standards among the reporting guidelines for mixed methods studies.

Monomethod research

Also called: mono-method research, monomethod design, monomethod study, single-method research, mono-method design

Monomethod research is research that uses a single method or approach, only quantitative or only qualitative, throughout a study. Johnson and Onwuegbuzie place it at one end of a continuum that ends in fully mixed designs, with partially mixed designs between, and they suggest the word probably goes back to Campbell and Fiske's 1959 work on measuring traits by several methods.

Multilevel mixed design

Also called: multilevel mixed methods design, multilevel mixed methods

A multilevel mixed design is a mixed methods design in which qualitative and quantitative data are collected at different levels of a system, such as schools and pupils or hospital units and patients, and integrated to answer related questions. It is one of Teddlie and Tashakkori's five families, and Schoonenboom and Johnson note that little has been written on how to integrate data across levels.

Multilevel samples

Also called: multilevel sample relationship, multilevel sampling, multilevel mixed methods sampling

Multilevel samples, in mixed methods sampling, are samples for the qualitative and quantitative strands drawn from different levels or populations of a study, such as pupils surveyed in one strand and their teachers or parents interviewed in the other. It is one of four sample relationships in Onwuegbuzie and Collins's typology, and each level may be sampled randomly or purposively.

Multimethod research

Also called: multi-method research, multimethod design, multi-method design, multiple methods research, multiple-method research

Multimethod research is research that uses more than one method, and writers disagree on how it differs from mixed methods. For Morse and Niehaus and for Schoonenboom and Johnson it combines methods from one tradition only, such as two qualitative components, whereas others use it for studies that use qualitative and quantitative methods but report them separately without integrating them.

Multiphase design

Also called: multiphase mixed methods design, multi-phase design, multistage mixed methods framework, multistage framework, multiphase mixed methods

A multiphase design is a mixed methods design that joins several sequential or concurrent phases over time within a programme of studies aimed at one overall objective, such as developing, testing and evaluating an intervention. Fetters, Curry and Creswell call it a multistage framework and require at least three stages when the phases are sequential, or two when one is convergent.

Multiple validities legitimation

Multiple validities legitimation is the extent to which a mixed methods study meets the relevant quantitative validities for its quantitative component, the relevant qualitative criteria for its qualitative component and the mixed legitimation types for its integration, so that its meta-inferences are sound. Onwuegbuzie and Johnson consider it relevant to almost every mixed study, and Schoonenboom and Johnson treat it as a goal of design.

Narrative profile

Also called: narrative profile formation, narrative profiling

A narrative profile is a written description built from quantitative data, such as a modal, average, holistic, comparative or normative profile of a group, and is a common way of qualitising. Onwuegbuzie and Johnson warn that such profiles can over-generalise the numbers, and that a profile built from averages may describe people who do not really exist.

Nested samples

Also called: nested sample relationship, nested sample, nested mixed methods sample

Nested samples, in mixed methods sampling, are samples in which the participants of one strand are a subset of those in the other, as when the highest and lowest scorers on a class survey are chosen for interviews. It is one of four sample relationships in Onwuegbuzie and Collins's typology, and Hamilton and colleagues used it to connect the two phases of an explanatory sequential design.

Offset

Also called: offset rationale

Offset, as a rationale for mixed methods research, is the argument that quantitative and qualitative methods each have strengths and weaknesses, so combining them lets a researcher counter the weaknesses of each with the strengths of the other. It is one of Bryman's sixteen rationales and restates the idea behind the fundamental principle of mixed research.

Paradigm wars

Also called: paradigm war, quantitative-qualitative debate, qualitative-quantitative debate, quantitative versus qualitative debate, qual-quant debate

The paradigm wars were the long-running dispute between advocates of quantitative and qualitative research over which paradigm was superior and whether the two could ever be combined, sharpest in education and the social sciences around the 1980s. Purists on both sides held to the incompatibility thesis, and mixed methods research presented itself, through pragmatism, as a practical way past the conflict.

Paradigmatic mixing legitimation

Paradigmatic mixing legitimation is the extent to which the epistemological, ontological, axiological, methodological and rhetorical beliefs behind a study's quantitative and qualitative components are successfully combined or blended into a usable whole. Onwuegbuzie and Johnson hold that it comes from making paradigm assumptions explicit and doing research that fits them, whether the two sets are kept separate or held in moderate forms.

Parallel mixed design

Also called: parallel mixed methods design

A parallel mixed design is a mixed methods design with qualitative and quantitative strands run at the same time or with a short gap, each analysed separately, answering related aspects of one question. Teddlie and Tashakkori expect the strands to be joined in meta-inferences, whereas Onwuegbuzie and Johnson describe parallel designs as leaving each strand with its own inferences, which some call quasi-mixed.

Parallel samples

Also called: parallel sample relationship, parallel sample

Parallel samples, in mixed methods sampling, are different groups of people drawn from the same population for the qualitative and quantitative strands, such as one class surveyed and pupils from another class in the same school interviewed. It is one of four sample relationships in Onwuegbuzie and Collins's typology, and it raises the question whether conclusions from the two groups can fairly be combined.

Partial agreement

Partial agreement, in a triangulation protocol, is the category recorded when two data sets support a finding in part, which is read as complementarity between them rather than full convergence. It sits between agreement and dissonance, and silence is recorded instead when only one of the two data sets speaks to the finding at all.

Partially mixed design

Also called: partially mixed methods design, partially mixed methods

A partially mixed design is a mixed methods design in which the qualitative and quantitative components are carried out in full, concurrently or sequentially, and are mixed only when the data are interpreted. In Leech and Onwuegbuzie's typology it is set against a fully mixed design, and crossing level of mixing with timing and emphasis gives eight design types.

Participant enrichment

Participant enrichment is a rationale for mixing qualitative and quantitative techniques to improve a study's sample, for example in recruiting participants and in checking that each person selected is suitable to include. It is one of four rationales set out by Collins, Onwuegbuzie and Sutton, alongside instrument fidelity, treatment integrity and significance enhancement.

Participant selection joint display

Also called: participant selection display

A participant selection joint display is a joint display used in explanatory sequential studies to show how the quantitative results led to the choice of participants for the qualitative phase. Guetterman and colleagues described it, and James, DeJonckheere and Guetterman extended it to record community advisers' priorities as well, so that transformative concerns also shape the qualitative sample.

Participatory-social justice design

Also called: mixed methods participatory-social justice design, participatory social justice design, participatory framework, mixed methods participatory design, social justice design

A participatory-social justice design is a complex mixed methods design that places one or more core designs within a participatory or social justice framework, such as feminist or participatory theory, so that participants help shape the research and the findings support change. Creswell and Plano Clark list it among their complex designs, and Fetters, Curry and Creswell treat community-based participatory research as one such framework.

Plus sign (mixed methods notation)

Also called: plus sign, + sign

The plus sign, in mixed methods notation, shows that two components are carried out concurrently, as in QUAN + qual, a design in which a smaller qualitative component runs alongside a dominant quantitative one. It comes from Morse's 1991 notation, in which an arrow marks the contrasting case of components carried out one after the other.

Point of interface

Also called: point of integration, point of mixing, mixing point

The point of interface is any point in a mixed methods study where the qualitative and quantitative components are brought together or connected, such as in data collection, analysis or interpretation. Morse and Niehaus and Guest use this term, Schoonenboom and Johnson call it the point of integration, and Guest proposed describing studies by the timing and purpose of integration at each such point rather than by design type.

Political legitimation

Political legitimation is the extent to which the users of mixed methods research, such as stakeholders and policymakers, value the meta-inferences drawn from both its quantitative and qualitative components. Onwuegbuzie and Johnson tie it to tensions of power and values within mixed teams and suggest pursuing practical results that people will value because they answer important questions.

Priority (mixed methods)

Also called: priority, weighting of methods, method weighting, dominance, paradigm emphasis, dominant status

Priority, in mixed methods design, is the relative weight given to the quantitative and qualitative components, which may be equal or may favour one. It is written with capitals for the dominant component and lower case for the other, as in QUAN → qual, and Hamilton and colleagues distinguish the volume or order of data, often called dominance, from the lens that drives interpretation.

Process (mixed methods rationale)

Process, as a rationale for mixed methods research, is the argument that quantitative research describes structures in social life while qualitative research conveys a sense of process, so combining them shows both. It is one of Bryman's sixteen rationales and fits studies that pair a measured effect with a qualitative account of how it came about.

Purposes of mixed methods research

Also called: purposes of mixing, mixing purposes, purpose of mixing, five purposes of mixed methods

The purposes of mixed methods research are the reasons for combining methods in a study, most often classified by Greene, Caracelli and Graham's five: triangulation, complementarity, development, initiation and expansion. Schoonenboom and Johnson advise settling the purpose before choosing a design, while noting that one study often serves several and that the list of possible purposes keeps growing.

QUAL (mixed methods notation)

Also called: qual, QUAL notation

QUAL, in mixed methods notation, stands for the qualitative component of a study, written in capitals when it is the dominant or core component and in lower case, qual, when it is supplemental. Morse shortened both approaches to four letters, which Schoonenboom and Johnson read as a gesture of equity between the traditions, as in QUAL → quan.

Qualitative comparative analysis

Also called: QCA

Qualitative comparative analysis is a case-based method, developed by Charles Ragin, that uses set theory to find which combinations of conditions are necessary or sufficient for an outcome across a small to medium number of cases. Conditions are coded as crisp, fuzzy or multi-value sets, and some mixed methods writers list it among advanced mixed analysis techniques because it turns case knowledge into set membership.

Qualitatively driven mixed methods

Also called: qualitative dominant mixed methods research, qualitative-dominant mixed methods, qualitatively driven design, QUAL-dominant mixed methods, qualitatively driven mixed methods research

Qualitatively driven mixed methods research is mixed methods research that rests on a qualitative, constructivist or critical view of inquiry while adding quantitative data and approaches because they are expected to benefit the study. Hamilton and colleagues argue that the drive lies in the lens used to make meaning, not in the order or amount of data, so even an explanatory sequential study can be qualitatively driven.

Qualitising

Also called: qualitizing, qualitisation, qualitization, qualitize, qualitise

Qualitising is the conversion of quantitative data into a qualitative form, such as narrative profiles, themes or typologies, so that they can be analysed qualitatively or compared with qualitative findings. Tashakkori and Teddlie introduced the term in 1998 as the reverse of quantitising, and Onwuegbuzie has since recast it as an interpretive process of meaning-making rather than a simple conversion.

QUAN (mixed methods notation)

Also called: quan, quant, QUAN notation

QUAN, in mixed methods notation, stands for the quantitative component of a study, written in capitals when it has priority and in lower case, quan, when it plays a supporting role. So QUAN + qual describes a mainly quantitative study with a smaller qualitative strand run at the same time, and qual → QUAN a qualitative phase that prepares a larger quantitative one.

Quantitatively driven mixed methods

Also called: quantitative dominant mixed methods research, quantitative-dominant mixed methods, quantitatively driven design, QUAN-dominant mixed methods, quantitatively driven mixed methods research

Quantitatively driven mixed methods research is mixed methods research that rests on a quantitative, postpositivist view of inquiry while adding qualitative data and approaches because they are expected to benefit the study. Reviews in mental health and nursing cited by Hamilton and colleagues found that about three-quarters of published mixed methods studies gave priority to the quantitative strand.

Quantitising

Also called: quantitizing, quantitisation, quantitization, quantitating, quantitize, quantitise

Quantitising is the conversion of qualitative data into numbers, usually by coding text and then counting or scoring the codes, so that the results can be analysed statistically or merged with quantitative data. Sandelowski, Voils and Knafl stress that it involves judgements about what and how to count, and a trade-off between numerical precision and narrative complexity.

Quasi-mixed design

Also called: quasi-mixed methods, quasi-mixed study, quasi-mixed methods design

A quasi-mixed design, in Teddlie and Tashakkori's terms, is a study that collects both qualitative and quantitative data but does not genuinely integrate them, so each set leads to its own conclusions. Accounts of which designs the label covers vary, and methodologists disagree over whether such studies count as mixed methods at all, though many current definitions would exclude them.

Sample integration legitimation

Sample integration legitimation is the extent to which the relationship between a study's quantitative and qualitative samples supports sound meta-inferences. Onwuegbuzie and Johnson warn that when conclusions from a small or different qualitative sample are combined with those from a large random sample, generalising the combined conclusion to the whole population may not be justified.

Sampling (mixed methods rationale)

Sampling, as a rationale for mixed methods research, is the use of one approach to help select the respondents or cases for the other, as when survey scores identify people to interview. It is one of Bryman's sixteen rationales and corresponds to connecting at the methods level of integration.

Sequential legitimation

Sequential legitimation is the extent to which a sequential mixed methods study has ruled out the possibility that its meta-inferences are an artefact of the order in which the quantitative and qualitative phases were carried out. Onwuegbuzie and Johnson suggest checking it with a multiple-wave design in which the two kinds of data collection and analysis alternate several times.

Sequential mixed design

Also called: sequential mixed methods design

A sequential mixed design is a mixed methods design in which qualitative and quantitative strands follow one another in chronological phases, with the questions or procedures of each later strand growing out of the earlier one. It is one of Teddlie and Tashakkori's five families of designs, and explanatory and exploratory sequential designs are its best-known forms.

Sequential mixed methods sampling

Also called: sequential mixed methods sampling design

Sequential mixed methods sampling is the selection of the sample for a later strand after an earlier strand has been carried out, often using its results, as when interviewees are chosen according to their questionnaire scores. It is one of the two time orientations in Onwuegbuzie and Collins's typology, combined with identical, parallel, nested or multilevel sample relationships.

Sequential timing

Also called: sequential, sequential implementation, sequential data collection

Sequential timing is the carrying out of the qualitative and quantitative strands of a study one after the other, so that the second can build on the first, shown in Morse's notation by an arrow, as in QUAL → quan. It usually brings dependence, since the later phase is designed from the earlier one's results, and Onwuegbuzie and Collins judge it unsuited to triangulation for that reason.

Side-by-side joint display

Also called: side-by-side display, side-by-side comparison joint display, side by side joint display

A side-by-side joint display is a joint display that places qualitative results next to the related quantitative results, usually in adjacent columns, often with a further column for the meta-inferences drawn from each pairing. Guetterman, Fetters and Creswell found it, with the statistics-by-themes display, to be one of the two most common types.

Significance enhancement

Significance enhancement is a rationale for mixing qualitative and quantitative techniques to make the data thicker and richer and to strengthen the interpretation and usefulness of the findings. It is one of four rationales set out by Collins, Onwuegbuzie and Sutton, alongside participant enrichment, instrument fidelity and treatment integrity.

Silence (triangulation protocol)

Also called: silence

Silence, in a triangulation protocol, is the category recorded when only one of two data sets being compared contains data on a particular finding, so the other neither confirms nor contradicts it. Tonkin-Crine and colleagues found that most pairwise comparisons in their trial's process evaluation ended in silence, often because patients and doctors had been asked about different things.

Staged approach

Also called: staged narrative, staged reporting, staged approach to integration

The staged approach is a way of integrating through narrative in which the results of each stage of a mixed methods study are analysed and published separately as they become available, with later papers referring back to earlier ones. It is common in multistage studies, and Fetters, Curry and Creswell list it with the weaving and contiguous approaches.

Statistics-by-themes joint display

Also called: statistics-by-themes display, themes-by-statistics display, theme-by-statistics joint display, themes-by-statistics joint display

A statistics-by-themes joint display is a joint display that arranges qualitative themes or quotations against quantitative categories, such as high, medium and low scores, so the reader can see how accounts differ across groups. Guetterman, Fetters and Creswell found it the most frequently used type across designs, often in convergent designs that merge the two kinds of data.

Strand (mixed methods)

Also called: strand, qualitative strand, quantitative strand, research strand

A strand, in mixed methods research, is one of the qualitative or quantitative components of a study, with its own questions, data collection, analysis and inferences. Typologies describe designs by how their strands are timed, weighted and brought together, and meta-inferences are drawn by integrating the inferences from each strand.

Substantive theory stance

The substantive theory stance is the position that the theory of the topic being studied, rather than a philosophical paradigm, should guide how methods are chosen and combined in a mixed methods study. Reviews associate it with Chen's theory-driven evaluation, and its advocates mix methods whenever the theory calls for it.

Supplemental component

Also called: supplementary component, supplemental strand, secondary component

The supplemental component, in Morse and Niehaus's approach, is the part of a mixed methods study that supports the core component and is interpreted within its theoretical drive, and it need not stand on its own as a complete study. It is written in lower case in Morse's notation, and Schoonenboom and Johnson question the suggestion that it may be done less rigorously.

Theoretical drive

Also called: inductive drive, deductive drive, inductive theoretical drive, deductive theoretical drive

Theoretical drive, in Morse and Niehaus's approach, is the overall direction of a mixed methods study, inductive when it aims mainly to explore and describe, and deductive when it aims mainly to test and predict. The component matching the drive is the core component, and Schoonenboom and Johnson argue that drive belongs to each research question rather than to a whole study.

Third research paradigm

Also called: third paradigm, third research movement, third wave, third research community

The third research paradigm is a name for mixed methods research as a tradition in its own right, beside quantitative and qualitative research, put forward by Johnson and Onwuegbuzie in 2004. They picture it as the broad middle of a continuum between purely qualitative and purely quantitative research, and Teddlie and Tashakkori speak in a similar way of a third research community.

Timing (mixed methods)

Also called: time orientation, sequence of methods, implementation of data collection

Timing, in mixed methods design, is the relationship in time between the qualitative and quantitative strands: concurrent, when they run together, or sequential, when one follows the other. Schoonenboom and Johnson split it into simultaneity and dependence, and it is one of the main decisions, with priority and the point of integration, by which typologies classify designs.

Transformative design (mixed methods)

Also called: transformative mixed methods design, transformative design

A transformative design is a mixed methods design shaped throughout by a transformative theoretical framework, such as feminism or critical race theory, which guides the priority, timing and mixing of its strands towards social change. It was one of Creswell and Plano Clark's six major designs, and their later complex designs include a participatory-social justice design with a similar aim.

Treatment integrity (mixed methods rationale)

Treatment integrity, as a rationale for mixed methods research, is the use of qualitative and quantitative techniques together to assess the fidelity of an intervention, that is, whether it was delivered as intended. It is one of four rationales set out by Collins, Onwuegbuzie and Sutton, alongside participant enrichment, instrument fidelity and significance enhancement.

Triangulation (mixed methods)

Also called: triangulation purpose, corroboration, triangulation or greater validity

Triangulation, as a purpose for mixed methods research, is the use of different methods on the same question to see whether their results agree and corroborate one another. It heads Greene, Caracelli and Graham's purposes and Bryman's rationales, and is narrower than triangulation in qualitative research. Onwuegbuzie and Collins tie it to concurrent designs, since a later phase shaped by an earlier one is no independent check.

Triangulation protocol

A triangulation protocol is a technique for integrating the findings of data sets that have already been analysed separately, by listing the key findings in a convergence coding matrix and judging, for each pair of data sets, whether they agree, partly agree, disagree or are silent. O'Cathain, Murphy and Nicholl describe it as one of three integration techniques, drawing on a protocol Farmer and colleagues developed for qualitative health research.

Typological approach (mixed methods design)

Also called: typological approach, typology-based approach, taxonomic approach, design typology

The typological approach to mixed methods design is the choice of a named design from an existing typology, such as convergent or explanatory sequential, to serve as a template for a study. Schoonenboom and Johnson value typologies for teaching and for naming common designs, but note that none is exhaustive and that most real studies combine several designs.

Unexpected results

Also called: unexpected results rationale

Unexpected results, as a rationale for mixed methods research, is the case for combining quantitative and qualitative research when one produces surprising findings that the other can help to understand. It is one of Bryman's sixteen rationales, and when the second method is added only after the surprise the design becomes emergent.

Utility (mixed methods rationale)

Also called: improving the usefulness of findings, utility or improving the usefulness of findings

Utility, as a rationale for mixed methods research, is the claim that combining quantitative and qualitative approaches makes findings more useful to practitioners and other users. It is one of Bryman's sixteen rationales and is most prominent in articles with an applied focus, where findings are meant to inform practice or policy.

Weakness minimisation legitimation

Also called: weakness minimization legitimation

Weakness minimisation legitimation is the extent to which the weaknesses of one approach in a mixed methods study are compensated for by the strengths of the other. Onwuegbuzie and Johnson argue that researchers must assess this deliberately when designing the study and use it when weighting and interpreting results, since the better the compensation, the stronger the meta-inferences.

Weaving approach

Also called: weaving, narrative weaving, weaving narrative

The weaving approach is a way of integrating through narrative in which qualitative and quantitative findings are written up together and organised by theme or concept, so that numbers and quotations are discussed side by side. Fetters, Curry and Creswell set it against the contiguous and staged approaches, and Hawkins and colleagues pair it with joint displays in their guide for novice researchers.

Within-method triangulation

Also called: within-methods triangulation, within method triangulation

Within-method triangulation is the use of at least two data collection procedures from the same approach, only qualitative or only quantitative, to study one phenomenon, such as interviews combined with observation. It is set against between-methods triangulation, which combines qualitative and quantitative methods, and some writers treat it as a form of data source triangulation.

Within-stage mixed-model design

Also called: within-stage mixed model design

A within-stage mixed-model design is a design in which qualitative and quantitative approaches are mixed inside a single stage of a study, the standard example being a questionnaire that combines a rating scale with open-ended questions. Johnson and Onwuegbuzie set it against across-stage mixed-model designs, in which the mixing runs between stages.