NVivo is a qualitative data analysis (QDA) software package used to organize, code, and interrogate unstructured data — interview transcripts, focus group recordings, field notes, open-ended survey responses, PDFs, images, and social-media extracts — in support of qualitative and mixed-methods research. It is developed by Lumivero (formerly QSR International, the Australian company that originally released NVivo in the late 1990s; QSR International merged with Palisade and Addinsoft to form Lumivero in October 2022). This page explains what NVivo actually does as a research tool, who uses it, and how it fits into a broader research data management workflow — not a product pitch, a research-methods and data-stewardship orientation for anyone deciding whether and how to use a QDA tool on a project.
What NVivo does
At its core, NVivo is a workbench for the same basic qualitative-analysis task researchers have always done with highlighters and index cards, computerized and made searchable at scale. The main operations are:
Coding
Coding is the process of tagging segments of source material — a sentence, a paragraph, a stretch of video — with labels (“codes” or “nodes” in NVivo’s terminology) that represent a concept, theme, or category relevant to the research question. NVivo supports this manually (a researcher reads and applies codes directly) and, in current versions, with AI-assisted coding suggestions that a researcher reviews and confirms rather than accepts automatically. Because every coded segment stays linked back to its exact location in the source, a researcher can always retrieve the full context behind a code, not just an isolated quote.
Organizing cases and attributes
Sources can be grouped into “cases” (e.g., one case per interview participant) and attached to demographic or categorical “attributes” (age, site, condition, role). This makes it possible to run comparative queries — for example, whether a theme appears differently across two participant groups — without leaving the coded structure.
Querying and visualization
Once material is coded, NVivo supports text-search queries, word-frequency queries, matrix-coding queries (cross-tabulating codes against cases or attributes), and coding comparison queries — the last of these compares how two or more coders applied codes to the same material and calculates an inter-coder agreement statistic, which is the standard way qualitative teams check and report coding reliability. Results can be exported as charts, cluster diagrams, or word clouds to support write-up.
Memos and annotations
Researchers can attach memos (analytic notes about emerging interpretations) and annotations (comments tied to a specific passage) throughout the coding process — this is how NVivo supports the audit trail that grounded theory and other reflexive methodologies expect: a visible record of how interpretation developed, not just the final code structure.
Methodologies NVivo supports
NVivo is methodology-agnostic software, not a single analytic method — it provides structure that researchers apply according to whichever approach governs their study, including:
- Thematic analysis — identifying, coding, and reporting patterns (themes) across a dataset.
- Content analysis — systematically categorizing and, often, quantifying the presence of concepts within text.
- Grounded theory — iterative coding and memo-writing aimed at building theory from the data itself rather than testing a pre-existing hypothesis.
- Discourse and narrative analysis — close reading of how language and structure construct meaning.
- Mixed-methods research — NVivo can import and cross-reference quantitative variables (e.g., a survey dataset) alongside qualitative material, which is one of its more distinctive capabilities relative to a pure QDA tool; see CASRAI’s Mixed Methods Research entry for the underlying methodological definition.
The software does not determine which of these is appropriate for a given study — that decision belongs to the research design, and using NVivo does not substitute for methodological rigor. A coding scheme applied without a defensible rationale is exactly as vulnerable to bias whether it is built in NVivo or on paper.
Who uses NVivo
NVivo (and QDA tools generally) is used across social science, health, education, business, and policy research: doctoral students coding interview or focus-group data for a dissertation, funded research teams conducting multi-site qualitative studies, program evaluators analyzing open-ended survey responses, and mixed-methods teams that need qualitative and quantitative findings to speak to each other in the same analysis. Universities commonly license NVivo institution-wide and make it available through library or research-computing services; students and researchers should check their institution’s software portal before purchasing an individual license, since site licenses are common and often cheaper than an individual subscription.
NVivo and research data management
For CASRAI’s purposes, the more consequential question than “what does the interface do” is where a QDA project fits into a broader research data management (RDM) workflow — because an NVivo project file is, itself, a research data asset with the same stewardship obligations as any other dataset:
Data management planning
If a project’s underlying interview transcripts, recordings, or documents contain identifiable information, that has direct consequences for a data management plan (DMP) — storage, access controls, and retention need to be specified before data collection begins, not worked out after coding is underway. See CASRAI’s worked DMP examples and the taxonomy of research data types for how qualitative source material is typically classified and described in a plan.
Provenance and reproducibility
An NVivo project (the .nvp/.nvpx file, or a cloud project in NVivo Collaboration Cloud) bundles sources, codes, memos, and query results together, but it is a proprietary, software-specific container — a researcher who wants the coded dataset to remain usable independent of that software, or who is preparing material for archiving, generally needs to plan an explicit export step (e.g., a codebook, a coded-segment export, or a structured export of node hierarchies) rather than treat the native project file as the long-term preservation format. That distinction — a working analysis environment versus an archival, software-independent record — is the same one that applies to any proprietary analysis tool and is a standard consideration in reproducible qualitative research practice.
Data sharing and privacy
Qualitative source material is frequently the most sensitive kind of research data a project handles — verbatim interview transcripts are far harder to de-identify convincingly than a spreadsheet of coded variables, and participant consent for how the data may be reused or shared needs to be established at the point of collection, not retrofitted before deposit. Where funder or journal policy expects data sharing, qualitative researchers typically share de-identified transcripts, codebooks, and documentation of the analytic process rather than the full coded project file, and should confirm what a specific funder or repository expects before assuming full-project sharing is either required or appropriate.
Team coding and reliability
For multi-coder projects, NVivo’s coding comparison query is the practical mechanism for producing an inter-coder reliability statistic — a figure increasingly expected in methods sections and by reviewers as evidence that a coding scheme was applied consistently, not just by one analyst’s individual judgment.
NVivo relative to other QDA tools
NVivo is one of several established QDA packages; ATLAS.ti and MAXQDA are the two most frequently compared alternatives, along with the open-source option Taguette for simpler tagging-based workflows. The practical differences between them are mostly around interface, pricing model, platform support (native macOS support has historically differed across these tools), and specific query/visualization features rather than a fundamental difference in what qualitative coding accomplishes — an institution’s existing site license is often the deciding factor in practice. Because pricing, licensing terms, and exact current version numbers for NVivo (and its competitors) change over time and vary by institution, individual, and region, confirm current terms directly from Lumivero’s own NVivo product page or an institutional software portal rather than relying on a fixed figure.
Frequently asked questions
Is NVivo the same as statistical software like SPSS?
No. NVivo analyzes unstructured qualitative data (text, audio, video, images) through coding and thematic organization; it is not a statistical package for numeric data, though it can import and cross-reference quantitative attributes as part of a mixed-methods design. See CASRAI’s qualitative vs. quantitative research comparison for the underlying methodological distinction.
Do I need NVivo to do qualitative coding?
No — qualitative coding can be, and historically was, done manually or in general-purpose tools like spreadsheets or word processors. Dedicated QDA software becomes more valuable as a dataset or team grows, because it keeps coded segments linked to source context, supports systematic querying across a large corpus, and generates the inter-coder reliability statistics that manual coding cannot produce without separate calculation. See CASRAI’s overview of qualitative data collection techniques for the methods that typically feed into a QDA tool.
What file formats does NVivo accept?
NVivo imports common document formats (Word, PDF, plain text), audio and video files, images, spreadsheet/survey data, and content from certain social-media and web sources; exact supported formats and integrations change across versions, so check Lumivero’s current documentation for the specific version in use.
How does AI fit into current versions of NVivo?
Lumivero has added AI-assisted features to recent NVivo releases, including AI-suggested coding and summarization support, which a researcher reviews and confirms rather than treats as a final coding decision. CASRAI’s AI in qualitative coding entry covers the broader methodological and disclosure considerations that apply whenever an AI-assisted step is used in a published qualitative analysis.
Where should NVivo project files be stored during an active research project?
Institutional research-data storage with appropriate access controls, consistent with whatever the project’s data management plan specifies — particularly where source material includes identifiable participant information. See CASRAI’s research data management overview for the broader lifecycle this fits into.
