# Lesson 6 — Analysis Frameworks and Conceptual Models (v3 expanded) *Companion-podcast transcript • Sarah & Kiffer* *~5,800 words • ~30 min audio* --- **Sarah:** Welcome back to Office Hours. I'm Sarah. **Kiffer:** And I'm Kiffer. This is Lesson 6 of the qualitative methods arc. Analysis frameworks and conceptual models. And we have arrived at what the textbook calls the analytic fork. In Lesson 5 you built a codebook and applied it. You have a coded dataset. You have some emerging themes. You have, for the first time in the course, a stack of analytic material that's yours. **Sarah:** And now, predictably, the question hits. What next? **Kiffer:** Right. The textbook is emphatic that this is the moment of the analytic fork. You have to choose an analytic strategy, and the choice will determine what your eventual findings look like. The point of this lesson is to show you the whole map so that when you choose a path in your capstone, you're choosing it because it fits your research question and not because it's the only one you know. **Sarah:** And there's a second move the lesson makes that's harder to teach but more important. It explains why the choice of analytic strategy is downstream of the research question, not upstream. **Kiffer:** That's where most students stumble. They pick a method first and then decide what their question was. Usually because that method is what they have a software licence for, or what their committee chair publishes in. The textbook position, and the position of this course, is that the strategy is dictated by what you want to find out. If you want to know what people say, content analysis. If you want to know what people mean, thematic analysis or grounded theory. If you want to know how people think, schema analysis. If you want to know how people tell, narrative analysis. If you want to know how people do things with words, discourse or conversation analysis. **Sarah:** Why is the federation of qualitative methods so fragmented? Because the histories run separately. **Kiffer:** Yeah. Grounded theory emerged from medical sociology in the late nineteen sixties. Content analysis from communication research in the nineteen fifties. Ethnographic discourse analysis from linguistic anthropology in the seventies. Conversation analysis from a sociology that took ordinary talk to be the foundation of social order. Narrative analysis from a movement across psychology, sociology, and folklore studies. Schema and cognitive-anthropological methods from a small community working on how culture is mentally represented. These traditions developed in parallel and only recently began to talk to each other. The result is that a student new to qualitative analysis reads three books and gets three incompatible vocabularies. The textbook's contribution is to map the federation so that you can see which family you're in and what the alternatives are. **Sarah:** And the map isn't a hierarchy. **Kiffer:** Right. It's not the case that text to talk is more rigorous than text to themes, or that text to schemas is the destination once text to counts is exhausted. The map is a fork. Different paths for different kinds of questions. **Sarah:** Okay, the map. Five analytic paths. Each named by what the path turns text into. **Kiffer:** Right. Text to counts. Text to themes. Text to schemas. Text to narratives. Text to talk. **Sarah:** Walk through them. **Kiffer:** Path one. Text to counts. Treats the text as data that can be reduced to frequencies and analyzed quantitatively. Classical method is content analysis. Krippendorff defines it as a research technique for making replicable and valid inferences from texts to the contexts of their use. The procedure is develop a coding scheme, apply it, count, analyze the counts. Strengths are transparency, replicability, comparability across groups. Weaknesses are that counts cannot distinguish a thematic mention that's central to a participant's account from one that's incidental. **Sarah:** And in our corpus. **Kiffer:** A text-to-counts approach might code every transcript for presence or absence of, say, a structural attribution — "society is set up to make new mothers lonely," in Sarah's words — a phenomenological description of an embodied state, and a coping strategy. Then count prevalence across the twenty transcripts and compare across age and gender. Findings would be of the form "structural attributions appeared in fourteen of twenty transcripts and were significantly more common among women caregivers." **Sarah:** Path two. Text to themes. **Kiffer:** Treats the text as a field in which patterns of meaning are to be identified and elaborated. Dominant methods are thematic analysis — Braun and Clarke — and grounded theory — Glaser, Strauss, Charmaz. They differ in important ways. Module seven in the textbook spends serious time on grounded theory specifically. But they share an analytic commitment. The unit of analysis is the theme, and the work is the identification, elaboration, and patterning of themes across the corpus. **Sarah:** And this is by far the dominant family in contemporary qualitative health research. **Kiffer:** It is. For both good and bad reasons. Good — it's intuitive, produces findings non-specialists can read, bridges easily into intervention design. Bad — "thematic analysis" has become a default residual category in journal articles, often invoked when no specific method was actually used, and the rigour varies enormously. The textbook is critical of this drift. If you say you did thematic analysis, you should say which version, in whose tradition, with what procedure. **Sarah:** Path three. Text to schemas. **Kiffer:** Treats the text as evidence about the underlying mental structures people use to organize their understanding of a domain. A schema is an abstract knowledge structure that lets a person recognize a situation, fill in defaults, and act appropriately. Schemas are not directly observable. They're inferred from speech, behavior, and consistency of response. Methods in this family include cultural-domain analysis, free-listing followed by pile sorts, cultural-consensus analysis, mental-model interviews. Less common in mainstream public-health qualitative work, but powerful when the question is about cultural cognition. **Sarah:** And in our corpus, a text-to-schemas approach might do what. **Kiffer:** Ask what kinds of loneliness participants distinguish in their own talk. Sarah at P-three distinguishes the loneliness of being awake while everyone sleeps, the loneliness of being in a marriage and unable to say the true thing, and the loneliness of mourning a previous self. These are not three instances of one schema. They're three distinct schemas she's articulating. The schema work is identifying these distinctions, formalizing them, asking whether they're shared. **Sarah:** Path four. Text to narratives. **Kiffer:** Treats each transcript as a story the participant is telling and asks what kind of story it is. Foundational text is Catherine Riessman's Narrative Methods for the Human Sciences, two thousand eight. Riessman distinguishes thematic narrative analysis — what is the story about — structural narrative analysis — how is the story built, orientation, complicating action, resolution — dialogic and performance analysis — what is the story doing for the teller — and visual narrative analysis. Narrative analysis is the right path when the unit of meaning is the story-as-told, rather than themes that can be abstracted from it. **Sarah:** And in our corpus. **Kiffer:** Kenji at P-fourteen is a good example. His entire interview is structured as a redemption-and-cost narrative. An old self that was inauthentic, a turning point — coming out at sixty — a present that contains specific lonelinesses, particularly his sons' refusal, and an emerging future the protagonist is working toward. "To be a man my sons might want to know someday." That narrative arc is itself the analytic unit. You don't decompose it into themes. You read the arc as the finding. **Sarah:** And you flagged earlier that many real studies use more than one path. **Kiffer:** They do. A grounded-theory study — text to themes — might include a content-analytic count of how often each theme appears across cases — that's text to counts. A narrative analysis — text to narratives — might include attention to discourse markers in the telling — text to talk. The paths aren't mutually exclusive. The point of distinguishing them is so you know which is your dominant path and which is supporting work. A methods section that says "we used thematic analysis with supplementary content-analytic counts" is doing a particular kind of methodological honesty. **Sarah:** And path five. Text to talk. **Kiffer:** Treats the transcript not as content to be interpreted, but as a record of action being performed through language. Classical method is conversation analysis, which emerged from Harvey Sacks's work at U C L A in the sixties. C A asks: at this turn in the conversation, what is the speaker doing? Repairing? Agreeing? Resisting? Hedging? Asking for affiliation? C A pays close attention to features of talk that other methods discard — pauses, overlaps, false starts, repair sequences — because those are the materials through which conversational action is built. **Sarah:** And the related methods. **Kiffer:** Discourse analysis and critical discourse analysis. Potter and Wetherell, Fairclough, Wodak. Less micro than C A, more attention to ideological and political work being done through language. **Sarah:** And in our corpus. **Kiffer:** A text-to-talk analysis might look at Sarah's repeated rhetorical management of her own claim that her loneliness is structural. She says it, and then immediately performs evaluative work. "Which is — I don't know if that helps me feel better or worse. Worse maybe. Because I can't fix structural." That sequence isn't a content claim. It's a piece of self-management in talk. A discourse-analytic lens would treat the sequence itself as the unit. **Sarah:** Now, the methodological point. **Kiffer:** The choice is downstream of the question. Most students in this course will end up on the text-to-themes path with secondary content-analytic counts. That's the most common path in applied health qualitative research. But it should be your choice because it fits your question, not because it's the default. And what you owe the reader in your methods section is to name the path and the alternatives you considered and rejected. That naming is itself methodological transparency. **Sarah:** And the failure mode most students make. **Kiffer:** The failure to make the transition from codes to interpretation. The textbook returns to this repeatedly. Many qualitative studies stop at a list of codes. They report what was coded, how often, by whom, with what intercoder reliability. They're descriptive but not analytic. The promised explanation — why the patterns are there — is missing. **Sarah:** Codes are not findings. **Kiffer:** Codes are categories. Interpretation is the claim about what the categories tell us. A coded dataset with codes for structural attribution, embodied exhaustion, witness absence, maternal identity grief is a stack of labelled material. The interpretation is the claim. Postpartum maternal loneliness, in this dataset, is best understood as the intersection of an embodied state of depletion, a witness deficit, and a grief for a pre-maternal self that the structural arrangement of the nuclear family is no longer adequate to support. That's interpretation. The transition from codes to interpretation is the moment the analyst stops being a librarian and starts being a theorist. **Sarah:** Worth one more pass on the codes-to-interpretation gap. Because it's the chronic failure mode and I want listeners to walk away knowing how to recognize it. **Kiffer:** Yeah. Here's the diagnostic. Read the abstract or the discussion section of a qualitative paper. If the central claims are sentences of the form "four themes were identified — A, B, C, D — each discussed in turn" — and there's no claim about how the themes relate, why those themes and not others, what the analytic structure connecting them is — that paper has stopped at coding. It's descriptive but not analytic. The interpretive move has not happened. **Sarah:** And the good version reads differently. **Kiffer:** The good version makes a claim of the form "we propose that X is best understood as the intersection of A, B, and C, configured in a particular structural pattern that emerges across cases despite surface differences in life-stage." That's a claim about the phenomenon. The themes A, B, and C are scaffolding for that claim, not findings in themselves. The transition from codes to interpretation is the transition from list to argument. **Sarah:** And memos are the way you make that transition happen. **Kiffer:** Memos are the mechanism. Without them, you arrive at the writing-up phase with twenty codes and no claim. With them, you arrive with thirty pages of theoretical memos, each anchored in specific cases, and the writing-up phase becomes selection and assembly rather than invention. **Sarah:** And the lesson gives two mechanisms for making that transition. **Kiffer:** Memos and triangulation. Memos are the engine. Triangulation is the brake. **Sarah:** Memos first. Three kinds. **Kiffer:** Code memos, theoretical memos, operational memos. **Sarah:** Code memos. **Kiffer:** A running definition of a code as it develops. When you create a new code, you write a code memo. It records the working definition, what counts as an instance, the first instance you tagged, any boundary cases that pushed you to refine, revisions over time. The code memo serves the audit-trail function. Without it, your codebook entry is just a definition. With it, you can reconstruct that the code originally meant one thing but was expanded after Sarah at P-three described her husband as asleep, and now covers functional witness absence — someone present but not available as an adult witness. **Sarah:** Theoretical memos. **Kiffer:** Different beast. Not about a single code. About a relationship, a possibility, an explanatory claim. Where interpretation happens. They're the place where you write "the data suggest that..." or "what unites these three cases is..." or "this pattern cannot be explained without invoking..." Strauss describes theoretical memos as the form of writing in which the analyst thinks on paper. Voice is exploratory but the content is empirical. Every theoretical claim cites the cases that support it and — importantly — the cases that don't. **Sarah:** I want to underline the worked example for theoretical memos, because it does a lot of teaching work. **Kiffer:** The worked memo in the lesson is about "loneliness inside companionship." The form is what to notice. Claim under development. Cases that fit. Cases that complicate. What I think this means. Next step. That's the architecture of a serious theoretical memo. Notice that the memo names not just the supporting cases — Sarah postpartum, Diana caregiving, Jacob in a hollowed partnership — but the complicating cases. Kenji, who has companionship but locates his loneliness elsewhere. Daniel, who has existential loneliness with no relational locus at all. The memo doesn't bury the disconfirming evidence. It works with it. That's what disciplined memo-writing looks like. **Sarah:** And the final move — "next step" — is what keeps the memo alive. **Kiffer:** Right. A memo without a next step is closure. A memo with a next step is investigation. Most strong theoretical memos end with "go read P-seventeen again with this claim in mind." That's the analytic loop. **Kiffer:** Different beast. Not about a single code. About a relationship, a possibility, an explanatory claim. Where interpretation happens. They're the place where you write "the data suggest that..." or "what unites these three cases is..." or "this pattern cannot be explained without invoking..." Strauss describes theoretical memos as the form of writing in which the analyst thinks on paper. Voice is exploratory but the content is empirical. Every theoretical claim cites the cases that support it and — importantly — the cases that don't. **Sarah:** That last bit is the discipline. **Kiffer:** Right. A theoretical memo that names only supporting cases is a hypothesis-confirmation exercise, not analysis. The work is to look for the disconfirming case and either reconcile it or revise the claim. **Sarah:** Operational memos. **Kiffer:** About the analytic process itself rather than the substantive content. Methodological decisions, problems, revisions. "I'm going to stop double-coding every transcript and shift to double-coding every fifth, because we've reached saturation on the codebook." "I noticed I've been giving more analytic weight to longer transcripts because there's more material. I need to check this." Operational memos are the most underused of the three. But they're what makes the methods section precise rather than retrospective fiction. **Sarah:** And there's a beautiful claim the lesson makes about the trajectory of memos. **Kiffer:** That by the end of a disciplined project, you have effectively drafted your findings section in the form of theoretical memos. The work of writing the paper is partly the work of assembling memos into argument. A memo that has been revised, cross-checked against new transcripts, and survived the discovery of disconfirming cases is no longer a tentative musing. It's a finding. Strauss is direct about this. A polished theoretical memo is "ready to be inserted with little revision into a publication." Most papers that feel laboured to write feel that way because the memo discipline was skipped. **Sarah:** Okay, triangulation. Denzin's four kinds. **Kiffer:** Right. The metaphor is from surveying. To locate a point precisely, take bearings from three positions. Denzin imported it into qualitative methodology and named four kinds. **Sarah:** Data triangulation. **Kiffer:** Checking the emerging interpretation against multiple sources of data on the same phenomenon. Interviews plus participant-generated diaries. Interviews from participants plus from family members. Interviews plus direct observation. Logic is that an interpretation supported across multiple kinds of source is more robust than one supported by a single source. For the capstone, full data triangulation may not be possible — you're working with interview transcripts alone. But internal data triangulation is available. An interpretation that holds across very different life-stages — Maya at twenty-two, Margaret at seventy-three, Kenji at sixty — is more robust than one that holds only in a single subgroup. **Sarah:** Investigator triangulation. **Kiffer:** Multiple analysts on the same data. Most familiar to public-health audiences because it overlaps with intercoder reliability. But Denzin meant something broader. Not just agreement on a codebook. Multiple analytic eyes on the same material, ideally bringing different perspectives. The disagreements are often the most productive analytic material. **Sarah:** Theory triangulation. **Kiffer:** Interpreting the data through multiple theoretical frames and seeing whether the conclusions converge. A loneliness study could read its data through social-isolation theory, attachment theory, existentialist phenomenology, feminist theories of relational care. Each frame highlights different aspects. If your central claim survives translation across theoretical frames, it has earned theoretical robustness. **Sarah:** And methodological triangulation. **Kiffer:** Checking the qualitative findings against findings from a different methodology, usually quantitative. Your interview-based finding that postpartum maternal loneliness has a strong structural dimension is methodologically triangulated against survey literature showing that countries with longer paid parental leave have lower postpartum-depression prevalence. Mixed-methods designs engineer this explicitly. **Sarah:** And there's a critique of triangulation worth knowing. **Kiffer:** Laurel Richardson's crystallization. The argument is that triangulation sits inside a postpositivist assumption. There's a phenomenon out there, and multiple readings should converge. Richardson argues this is the wrong metaphor for much qualitative work. Triangles are flat. They have three sides. Real social phenomena are multifaceted, refracted differently from every angle. The goal of an interpretive analysis is not to nail down one point. It's to show the phenomenon's multiple shapes. A crystal — a prism — refracts light in different directions depending on the angle. Different methods, theories, analysts, writing forms reveal different facets. **Sarah:** And the methodological stakes. **Kiffer:** Real. Triangulation language fits convergent-validity audiences trained in quantitative public health. Crystallization fits interpretive and critical traditions where the point of qualitative work is to expand what is sayable about a phenomenon. Most contemporary qualitative health research lives somewhere in between. The discipline is to be honest about which one your study actually does. Don't invoke triangulation as a label when you have no actual multiplicity of source, analyst, or method. **Sarah:** Okay, third big chunk. Conceptual models. The thing the analysis ultimately builds toward. **Kiffer:** This is the move that turns coded data into a finding. Bernard, Wutich, and Ryan are explicit. The purpose of qualitative analysis is, in the end, to produce a model of the phenomenon. Not a regression equation. A structured visual or verbal representation of how the parts of the phenomenon hang together. What the kinds are, how they relate, in what order they unfold, under what conditions one leads to another. **Sarah:** Four kinds of conceptual model to know. **Kiffer:** Taxonomies. Typologies. Concept maps. Process models. **Sarah:** Taxonomies first. **Kiffer:** A hierarchical organization of categories. Top-level category divided into subordinate categories, which may be further divided. Biological taxonomy of life is the classical example. In qualitative analysis, taxonomies usually have two or three levels. The goal is interpretive structure, not exhaustive Linnaean classification. Taxonomies answer the question: what kinds of X are there? **Sarah:** And the worked example for loneliness. **Kiffer:** First-pass taxonomy might have four top-level kinds. Existential, relational, structural, situational. Each with sub-kinds. Existential includes mortal-finitude loneliness — Daniel at P-ten — identity-disruption — Kenji at P-fourteen after coming out at sixty — existential-developmental — Marcus at P-eight in mid-life burnout. Relational includes loss loneliness — Linda after widowhood, Robert after divorce — and companioned-but-witnessless — Sarah postpartum, Diana caregiving for her mother with dementia, Jacob in a hollowed partnership. Structural includes migration loneliness — Amira's wahda — and institutional loneliness — Margaret in long-term care. Situational includes acute-event loneliness — Maya post-breakup, Tyler post-job-loss. **Sarah:** And the standards for a defensible taxonomy. **Kiffer:** Five things. Grounded in data — every category has named exemplars. Inclusion criteria explicit. Categories at the same level of abstraction within a tier. Empirical claims testable — every transcript can be assigned somewhere. And revisable — disconfirming evidence prompts revision, not denial. **Sarah:** Typologies. Different beast. **Kiffer:** Hierarchical versus dimensional. Where a taxonomy is sub-kinds of a kind, a typology is cases located in a space defined by two or more variables. Classical sociological example is Max Weber's ideal-types — rational/traditional crossed with instrumental/value-oriented. Typologies are powerful when you have two dimensions that crosscut each other in a way no single hierarchical sort can capture. **Sarah:** And in the loneliness work. **Kiffer:** Two dimensions that emerged in the data. The kind of trigger that brings loneliness on — existential, relational, structural, situational. And the kind of response the participant has developed — reframing, behavioural change, withdrawal, no consistent strategy. Crossing those produces a four-by-four typology with sixteen cells. Not every cell will be populated, and that itself is a finding. Reading by row reveals that existential triggers produce predominantly reframing responses. Reading by column reveals that "no consistent strategy" clusters with relational and structural triggers — a finding with intervention implications. **Sarah:** The taxonomy told you what kinds exist. The typology tells you about relationships between dimensions. **Kiffer:** That's the analytic difference. And the methodological point. The dimensions you choose are themselves an interpretation. Choosing "existential, relational, structural, situational" as the trigger dimension reflects a particular reading of the data. Another analyst might use "internal versus external" or "acute versus chronic." The typology renders this interpretive choice visible. Your methods section justifies the dimensions you used the same way it justifies your codebook. **Sarah:** Third kind. Concept maps. **Kiffer:** Relational diagrams. Nodes are concepts, edges are relations. Concept maps are most useful when the analytic claim is about how multiple concepts relate to each other, and where the relations are not strictly hierarchical or dimensional. A concept map of loneliness might show, say, witness-absence as a node connected by an edge labelled "causes" to depleted-self, with reframing-response as another node, and arrows showing that the structural-attribution finding sits between the experience and the response. The map renders interpretive claims in a form a reader can evaluate at a glance. **Sarah:** And fourth. Process models. **Kiffer:** Temporal or sequential. They render the phenomenon as something that unfolds in time. Stages, transitions, branching paths. A process model of postpartum loneliness might show an early phase characterized by intense embodied depletion, a middle phase in which the witness-deficit becomes visible to the participant, and a later phase that branches depending on whether the participant develops a structural reframing or remains in self-blame. Process models are particularly powerful for implementation-science work, where the question is how a process unfolds and where the intervention should sit. **Sarah:** Let me push on the boundaries between these four model types, because they shade into each other. **Kiffer:** They do. A taxonomy with two cross-cutting dimensions is becoming a typology. A typology with arrows showing temporal flow between cells is becoming a process model. A process model that gets re-rendered as a network of relations between phases is becoming a concept map. The four types are useful shorthands, not rigid containers. What matters is that whichever form your model takes, it makes a structured interpretive claim that another analyst could engage with. **Sarah:** And the choice is partly aesthetic, partly substantive. **Kiffer:** Both. Substantively, you choose the form that best renders the relationships you found. Aesthetically, you choose what your audience will read. Public-health-policy audiences often prefer process models because they translate easily into intervention design. Sociology audiences often prefer concept maps because they accommodate theoretical claims. Anthropology audiences often prefer taxonomies because the categorical work is the disciplinary tradition. Know your audience. **Sarah:** And the tools mentioned in the lesson. **Kiffer:** Miro and draw dot i o for collaborative diagramming. R's DiagrammeR package for code-driven diagrams that can sit inside an R Markdown document. ggplot two for heatmaps of code-by-participant intensity. And hand-drawing, which is honestly often the right starting point. The point isn't the tool. The point is to make the interpretive structure visible to yourself and to your reader. **Sarah:** And there's one thing worth saying about the residual category problem. When you build a taxonomy, there's almost always a case that doesn't quite fit. **Kiffer:** Yeah. And the temptation is to either force the misfit case into an existing category, which weakens the category, or to dismiss the misfit case as anomalous, which weakens the taxonomy. The disciplined move is a third option. Either expand the taxonomy with a new category — which usually means revising the dimensions of existing categories — or honestly acknowledge the misfit as a boundary case the taxonomy doesn't yet cover. Boundary cases are findings, not failures. A taxonomy of loneliness that explicitly names which transcripts don't fit and why is more methodologically defensible than one that pretends every case fits cleanly. **Sarah:** And the same logic applies to the typology cells. **Kiffer:** It does. Empty cells in a typology are findings. They might mean the combination doesn't exist in the data — which is informative. They might mean the dimensions you chose don't intersect the way you assumed — which is also informative, and prompts revision. Either way, the empty cells say something. Don't hide them. **Sarah:** Let me try the synthesis. First takeaway. The choice of analytic strategy is downstream of the research question. Five paths — text to counts, text to themes, text to schemas, text to narratives, text to talk. Each answers a different kind of question. The methods section should name the path and the alternatives considered. **Kiffer:** Second. Codes are not findings. The transition from codes to interpretation is the move from being a librarian to being a theorist. It's the failure mode that most often distinguishes a published qualitative paper from a manuscript that gets rejected. **Sarah:** Third. Memos are the engine of interpretation. Three kinds — code memos, theoretical memos, operational memos. Each becomes part of a different section of the eventual paper. Theoretical memos especially are the substrate of the findings section. **Kiffer:** Fourth. Triangulation is the brake. Denzin's four kinds — data, investigator, theory, methodological — each addresses a different threat to interpretive validity. Don't invoke triangulation as a label when you have no actual multiplicity. **Sarah:** Fifth. Richardson's crystallization is the contemporary alternative. The phenomenon's multiple shapes are the analytic target, not convergence on a single point. Which language you use depends on which audience you're writing for and what your study actually does. **Kiffer:** Sixth. The end product of disciplined qualitative analysis is a conceptual model. Four kinds — taxonomy, typology, concept map, process model. Each makes interpretation visible in a way a list of codes cannot. **Sarah:** Seventh. A defensible model is grounded in data, transparent in construction, and revisable in the face of disconfirming evidence. The model is a hypothesis about the data's structure, not a finished claim. **Kiffer:** And eighth. Visualization isn't decoration. The cognitive limits on holding twenty interconnected qualitative codes in your head at once are real. Most students try, fail, and end up with flat findings sections that report codes one after another. A conceptual model is the analytic move that turns a list into a structure. Miles, Huberman, and Saldaña have been making this point throughout the field for decades. **Sarah:** Callback to earlier lessons. **Kiffer:** The three commitments from Lesson 1 are doing serious work in this lesson. Systematic — your codes were applied consistently. Transparent — your memos make the analytic moves visible. Replicable — your conceptual model is testable by another analyst. The paradigm declaration from Lesson 2 also lands here. A postpositivist study reads its conceptual model as a model of the underlying phenomenon. An interpretivist study reads it as a model of how participants construct the phenomenon in their accounts. A critical study reads it as a model of how structural conditions produce the phenomenon. Same dataset, three different model readings. The paradigm shapes how you defend the model. **Sarah:** Anything to carry forward? **Kiffer:** One thing. The work this week is the conceptual move that distinguishes a qualitative analysis from a coding exercise. Coded passages by themselves are not findings. The model — the taxonomy, typology, concept map, or process model that organizes the coded material into a structured claim — is the finding. Build the model. Memo why each piece is there. Be explicit about which transcripts don't fit. That last move is the hardest and the most important. **Sarah:** And this is where the first arc of the course ends. The upstream design and the central analytic moves. From here, the rest of the course goes deeper into specific methods — comparing variables and grounded theory, content analysis, schema and narrative analysis, discourse analysis, qualitative comparative analysis, and finally computational text and large-language-model analysis. **Kiffer:** And the capstone keeps building toward the journal-article-format paper. From here you have a research question, a sampling defense, a codebook, a first conceptual model. The remaining weeks deepen the analysis and write the paper. **Sarah:** Thanks for listening. This is the end of the foundations arc. We'll see you in the next module. **Kiffer:** Take care of yourselves. And remember — codes are categories. Interpretation is the claim. Build the model. See you in class.