What Is Pattern Coding?
Pattern coding is a second-cycle qualitative coding method that groups first-cycle codes into a smaller number of themes, categories, constructs, or meta-codes. After your initial coding pass produces dozens or hundreds of codes, pattern coding consolidates them into higher-order patterns that explain the data more parsimoniously. Miles, Huberman, and Saldana identify four types of pattern codes: categories or themes, causes or explanations, relationships among people, and theoretical constructs. The method is the workhorse of applied qualitative analysis, it's how you move from a flat list of codes to structured findings that stakeholders can act on.
Why Pattern Coding Matters
First-cycle codes are building blocks, not findings. A study with 200 individual codes doesn't tell a coherent story. Pattern coding organizes those codes into a manageable number of patterns, typically 5-15, that capture the major themes in your data. It's the step that transforms a coded dataset into a set of findings you can present, discuss, and use for decision-making. Without it, qualitative analysis ends at description and never reaches explanation.
How Pattern Coding Works
The Process
Step 1: Lay out your first-cycle codes. Review your complete set of first-cycle codes, from open coding, descriptive coding, process coding, or whatever methods you used. List them, along with frequency counts and representative data excerpts.
Step 2: Look for clusters. Which codes seem to go together? Grouping can be based on:
- Thematic similarity: codes that address the same underlying topic or concept.
- Causal relationships: codes that describe conditions, actions, and outcomes related to the same phenomenon.
- Temporal sequences: codes that represent stages in a process.
- Dimensional variation: codes that represent different degrees or types of the same construct.
Step 3: Name the pattern. Each cluster gets a pattern code, a label that captures what the grouped codes share. Pattern codes should be more abstract than first-cycle codes but grounded enough to be clearly connected to the data.
First-cycle codes: reading online reviews, asking friends for recommendations, testing free trials, comparing feature lists, watching demo videos
Pattern code: "Building a decision safety net", the pattern captures the underlying behavior of reducing risk before committing.
Step 4: Test pattern codes against the data. Go back to the raw data and check whether your pattern codes hold up. Can every first-cycle code be meaningfully assigned to a pattern? Are there data segments that don't fit any pattern? Outliers might indicate a pattern you haven't identified yet, or they might be genuinely peripheral.
Step 5: Write analytical memos. For each pattern code, write a memo explaining what it captures, how it relates to other patterns, and what it means for the research question. These memos often become the backbone of your findings section.
Pattern Coding vs. Other Second-Cycle Methods
Pattern coding shares territory with axial coding and focused coding, but the emphasis differs:
- Axial coding maps formal relationships (conditions, strategies, consequences) between categories using a coding paradigm. It's theory-building oriented.
- Focused coding selects the most analytically productive first-cycle codes and tests them across the full dataset. It's about identifying the most important codes.
- Pattern coding groups codes into clusters based on shared meaning or function. It's about consolidation and synthesis.
In practice, these methods overlap. Many researchers use pattern coding as their primary consolidation strategy and borrow elements of axial or focused coding when relationships between patterns need specification.
Pattern Coding in Applied Research
In market research, UX research, and evaluation studies, pattern coding is often the final analytic step. The patterns become your findings, the 5-8 themes that answer the research question and structure the report. Each pattern is supported by representative quotes and connected to practical recommendations.
When to Use Pattern Coding
- After any first-cycle coding method: pattern coding works with codes from open coding, descriptive coding, in vivo coding, process coding, or any combination.
- Cross-case analysis: when you need to identify patterns that hold across multiple interviews, focus groups, or case studies.
- Report preparation: when you need to organize coded data into a presentable structure with 5-15 major themes.
- Mixed-methods research: pattern codes from qualitative data can be mapped onto quantitative variables for integration.
Common Mistakes
- Creating patterns that are too broad. A pattern code like "user experience" is so broad it doesn't tell you anything specific. Aim for patterns that are specific enough to suggest action: "trust-building through transparency before purchase" is more useful than "trust."
- Forcing every code into a pattern. Some first-cycle codes may be genuine outliers, interesting but not part of a broader pattern. It's better to have a small "miscellaneous" category than to distort a pattern by including codes that don't belong.
- Stopping at pattern labels without analysis. A list of pattern codes isn't findings, it's an organized index. Each pattern needs a narrative explanation: what it means, how it works, what evidence supports it, and what implications it has.
Quali-Fi Support
Quali-Fi's AI-powered qualitative analysis can cluster initial codes into suggested patterns, giving researchers a head start on second-cycle coding. The platform's thematic coding interface supports drag-and-drop code organization across focus group transcripts, discussion board data, and open-ended survey responses, making the pattern coding process visual and collaborative.
Discover patterns in your qualitative data with Quali-Fi{:.cta-button }
FAQs
How many pattern codes should I have?
Most studies consolidate into 5-15 pattern codes. Fewer than 5 often means your patterns are too broad to be useful. More than 15 means you haven't consolidated enough, you're still at a level of detail that's hard for audiences to absorb. The right number depends on data complexity and your research question's scope.
Is pattern coding the same as thematic analysis?
Pattern coding is one step in the broader process of thematic analysis. Braun and Clarke's thematic analysis framework includes six phases, and pattern coding aligns with phases 3-5 (searching for themes, reviewing themes, defining and naming themes). You can do pattern coding without doing full thematic analysis, but thematic analysis always involves something like pattern coding.
Can AI do pattern coding?
AI tools can suggest code clusters based on co-occurrence patterns and semantic similarity, which provides a useful starting point. But deciding whether a suggested cluster represents a meaningful pattern, one that's analytically coherent and relevant to the research question, requires human judgment. The best workflow is AI-suggested clusters refined by researcher interpretation.