Master qualitative research methods including interviews, ethnography, case studies, grounded theory, and thematic analysis
Use this skill when you need to:
Purpose: Explore individual perspectives and experiences
Interview Protocol Template:
from dataclasses import dataclass
from typing import List, Optional
from datetime import datetime
@dataclass
class InterviewProtocol:
"""Structure for interview guide"""
research_question: str
introduction: str
opening_questions: List[str]
main_questions: List[str]
probes: dict # question -> list of probes
closing_questions: List[str]
estimated_duration: int # minutes
def generate_guide(self):
"""Generate formatted interview guide"""
guide = f"""
# Interview Guide
## Research Question
{self.research_question}
## Introduction ({estimated_duration} minutes)
{self.introduction}
## Opening Questions (5-10 minutes)
"""
for i, q in enumerate(self.opening_questions, 1):
guide += f"{i}. {q}\n"
guide += "\n## Main Questions (30-40 minutes)\n"
for i, q in enumerate(self.main_questions, 1):
guide += f"{i}. {q}\n"
if q in self.probes:
for probe in self.probes[q]:
guide += f" - Probe: {probe}\n"
guide += "\n## Closing Questions (5-10 minutes)\n"
for i, q in enumerate(self.closing_questions, 1):
guide += f"{i}. {q}\n"
return guide
# Example protocol
protocol = InterviewProtocol(
research_question="How do remote workers maintain work-life balance?",
introduction="""
Thank you for participating. This interview will take about 60 minutes.
I'm interested in understanding your experiences with remote work.
There are no right or wrong answers. Everything you share will be kept
confidential. Do you have any questions before we begin?
""",
opening_questions=[
"Can you tell me about your current remote work situation?",
"How long have you been working remotely?"
],
main_questions=[
"Walk me through a typical workday. What does it look like?",
"How do you separate work time from personal time?",
"What challenges have you faced with work-life balance?",
"What strategies have you found helpful?"
],
probes={
"Walk me through a typical workday. What does it look like?": [
"What time do you typically start?",
"How do you structure your day?",
"Where do you work from?"
],
"What challenges have you faced with work-life balance?": [
"Can you give me a specific example?",
"How did that make you feel?",
"What did you do about it?"
]
},
closing_questions=[
"Is there anything else you'd like to share?",
"What advice would you give to new remote workers?"
],
estimated_duration=60
)
print(protocol.generate_guide())Purpose: Identify patterns and themes in qualitative data
Six-Phase Process:
import pandas as pd
from collections import defaultdict
from typing import List, Dict, Set
class ThematicAnalysis:
"""Conduct rigorous thematic analysis"""
def __init__(self):
self.transcripts = {}
self.codes = defaultdict(list)
self.themes = {}
self.codebook = {}
# Phase 1: Familiarization
def add_transcript(self, participant_id: str, text: str):
"""Add transcript and initial notes"""
self.transcripts[participant_id] = {
'text': text,
'initial_notes': []
}
def add_initial_note(self, participant_id: str, note: str):
"""Record initial observations"""
self.transcripts[participant_id]['initial_notes'].append(note)
# Phase 2: Generate initial codes
def code_segment(self, participant_id: str, segment: str,
code: str, line_numbers: tuple):
"""Apply code to text segment"""
self.codes[code].append({
'participant': participant_id,
'segment': segment,
'lines': line_numbers
})
def get_code_frequency(self):
"""Count code applications"""
return {code: len(instances)
for code, instances in self.codes.items()}
# Phase 3: Search for themes
def create_theme(self, theme_name: str, codes: List[str],
description: str):
"""Group codes into themes"""
self.themes[theme_name] = {
'codes': codes,
'description': description,
'subthemes': []
}
def add_subtheme(self, theme_name: str, subtheme_name: str,
codes: List[str]):
"""Add subtheme to existing theme"""
self.themes[theme_name]['subthemes'].append({
'name': subtheme_name,
'codes': codes
})
# Phase 4: Review themes
def get_theme_excerpts(self, theme_name: str):
"""Extract all coded excerpts for a theme"""
excerpts = []
theme = self.themes[theme_name]
for code in theme['codes']:
if code in self.codes:
excerpts.extend(self.codes[code])
return excerpts
def check_theme_coherence(self, theme_name: str):
"""Assess internal homogeneity of theme"""
excerpts = self.get_theme_excerpts(theme_name)
report = {
'theme': theme_name,
'n_excerpts': len(excerpts),
'n_participants': len(set(e['participant'] for e in excerpts)),
'codes': self.themes[theme_name]['codes'],
'excerpts_sample': excerpts[:5] # First 5 for review
}
return report
# Phase 5: Define and name themes
def define_theme(self, theme_name: str, definition: str,
essence: str):
"""Provide clear theme definition"""
self.themes[theme_name]['definition'] = definition
self.themes[theme_name]['essence'] = essence
# Phase 6: Produce report
def generate_codebook(self):
"""Create detailed codebook"""
codebook = []
for theme_name, theme_data in self.themes.items():
entry = {
'Theme': theme_name,
'Definition': theme_data.get('definition', ''),
'Codes': ', '.join(theme_data['codes']),
'N_Excerpts': len(self.get_theme_excerpts(theme_name))
}
codebook.append(entry)
return pd.DataFrame(codebook)
def generate_theme_report(self):
"""Generate thematic analysis report"""
report = "# Thematic Analysis Report\n\n"
report += f"Total Participants: {len(self.transcripts)}\n"
report += f"Total Codes: {len(self.codes)}\n"
report += f"Total Themes: {len(self.themes)}\n\n"
for theme_name, theme_data in self.themes.items():
report += f"## Theme: {theme_name}\n\n"
report += f"**Definition**: {theme_data.get('definition', 'N/A')}\n\n"
report += f"**Essence**: {theme_data.get('essence', 'N/A')}\n\n"
report += f"**Codes**: {', '.join(theme_data['codes'])}\n\n"
excerpts = self.get_theme_excerpts(theme_name)
report += f"**Prevalence**: {len(excerpts)} coded segments "
report += f"across {len(set(e['participant'] for e in excerpts))} participants\n\n"
# Sample excerpts
report += "**Representative Excerpts**:\n\n"
for i, excerpt in enumerate(excerpts[:3], 1):
report += f"{i}. *{excerpt['participant']}*: "
report += f'"{excerpt["segment"]}"\n\n'
# Subthemes
if theme_data.get('subthemes'):
report += "**Subthemes**:\n\n"
for subtheme in theme_data['subthemes']:
report += f"- *{subtheme['name']}*: "
report += f"{', '.join(subtheme['codes'])}\n"
report += "\n"
return report
# Example usage
analysis = ThematicAnalysis()
# Add data
analysis.add_transcript('P001', "I find it hard to stop working...")
analysis.add_initial_note('P001', "Struggles with boundaries")
# Code data
analysis.code_segment('P001',
"I find it hard to stop working at 5pm",
'boundary_difficulty',
(23, 25))
analysis.code_segment('P001',
"My laptop is always open on the kitchen table",
'physical_workspace_integration',
(45, 47))
# Create themes
analysis.create_theme(
'Blurred Boundaries',
codes=['boundary_difficulty', 'work_intrusion', 'always_on'],
description='Difficulty maintaining clear work-life boundaries'
)
analysis.define_theme(
'Blurred Boundaries',
definition='The challenge of creating and maintaining separation between work and personal life in remote settings',
essence='Work and life become entangled'
)
# Generate report
print(analysis.generate_theme_report())Purpose: Generate theory from data inductively
Coding Approach:
from dataclasses import dataclass
from typing import List, Optional, Dict
from enum import Enum
class CodingLevel(Enum):
OPEN = "open"
AXIAL = "axial"
SELECTIVE = "selective"
@dataclass
class GroundedTheoryCode:
"""Represent a grounded theory code"""
code: str
level: CodingLevel
definition: str
properties: List[str]
dimensions: Dict[str, tuple] # property -> (min, max) dimensions
memos: List[str]
examples: List[str]
class GroundedTheoryAnalysis:
"""Conduct grounded theory analysis"""
def __init__(self):
self.codes = {}
self.categories = {}
self.core_category = None
self.theoretical_memos = []
def open_coding(self, code_name: str, definition: str):
"""Initial open coding"""
self.codes[code_name] = GroundedTheoryCode(
code=code_name,
level=CodingLevel.OPEN,
definition=definition,
properties=[],
dimensions={},
memos=[],
examples=[]
)
def axial_coding(self, category_name: str, codes: List[str],
conditions: List[str], actions: List[str],
consequences: List[str]):
"""Relate categories to subcategories (paradigm model)"""
self.categories[category_name] = {
'codes': codes,
'conditions': conditions, # When/why
'actions_interactions': actions, # What strategies
'consequences': consequences, # With what results
'level': CodingLevel.AXIAL
}
def selective_coding(self, core_category: str, storyline: str):
"""Identify core category and integrate theory"""
self.core_category = {
'category': core_category,
'storyline': storyline,
'level': CodingLevel.SELECTIVE
}
def add_memo(self, memo: str, code: Optional[str] = None):
"""Write theoretical memo"""
if code and code in self.codes:
self.codes[code].memos.append(memo)
else:
self.theoretical_memos.append(memo)
def constant_comparison(self, code1: str, code2: str):
"""Compare two codes"""
c1 = self.codes.get(code1)
c2 = self.codes.get(code2)
if not c1 or not c2:
return "Code(s) not found"
comparison = f"## Constant Comparison: {code1} vs {code2}\n\n"
comparison += f"**{code1}**: {c1.definition}\n"
comparison += f"**{code2}**: {c2.definition}\n\n"
comparison += "**Similarities**:\n- [Identify similarities]\n\n"
comparison += "**Differences**:\n- [Identify differences]\n\n"
comparison += "**Theoretical Insight**:\n[What does this tell us?]\n"
return comparison
def theoretical_sampling_guide(self):
"""Generate guide for next theoretical sampling"""
guide = "# Theoretical Sampling Guide\n\n"
guide += "## Current Theory State\n"
if self.core_category:
guide += f"Core Category: {self.core_category['category']}\n"
guide += f"Storyline: {self.core_category['storyline']}\n\n"
guide += "## Gaps to Address\n"
guide += "1. [Identify underdeveloped categories]\n"
guide += "2. [Identify missing relationships]\n"
guide += "3. [Identify negative cases needed]\n\n"
guide += "## Next Sampling Criteria\n"
guide += "- Participants who... [specific characteristics]\n"
guide += "- Settings that... [specific conditions]\n"
guide += "- Events that... [specific situations]\n"
return guide
# Example
gt = GroundedTheoryAnalysis()
# Open coding
gt.open_coding('seeking_flexibility',
'Actions to gain control over work schedule and location')
gt.open_coding('managing_expectations',
'Negotiating others\' expectations about availability')
# Axial coding
gt.axial_coding(
category_name='Boundary Work',
codes=['seeking_flexibility', 'managing_expectations', 'creating_rituals'],
conditions=['Work from home', 'High autonomy', 'Family present'],
actions=['Set physical boundaries', 'Communicate availability', 'Use transitional rituals'],
consequences=['Reduced conflict', 'Better focus', 'Improved wellbeing']
)
# Selective coding
gt.selective_coding(
core_category='Continuous Boundary Negotiation',
storyline='Remote workers engage in ongoing negotiation of boundaries '
'between work and life, using spatial, temporal, and communicative '
'strategies to manage competing demands and maintain wellbeing.'
)
# Memos
gt.add_memo('Flexibility appears to be double-edged sword - enables control '
'but also creates expectation of constant availability',
code='seeking_flexibility')Purpose: In-depth examination of bounded system
Case Study Protocol:
from dataclasses import dataclass
from typing import List, Dict
from enum import Enum
class CaseType(Enum):
SINGLE = "single"
MULTIPLE = "multiple"
EMBEDDED = "embedded"
class DataSource(Enum):
INTERVIEW = "interview"
OBSERVATION = "observation"
DOCUMENT = "document"
ARTIFACT = "artifact"
ARCHIVAL = "archival"
@dataclass
class CaseStudyDesign:
"""Define case study parameters"""
case_type: CaseType
research_questions: List[str]
propositions: List[str]
units_of_analysis: str
case_selection_logic: str
data_sources: List[DataSource]
def generate_protocol(self):
"""Generate case study protocol"""
protocol = f"""
# Case Study Protocol
## Overview
Type: {self.case_type.value}
Unit of Analysis: {self.units_of_analysis}
## Research Questions
"""
for i, q in enumerate(self.research_questions, 1):
protocol += f"{i}. {q}\n"
protocol += "\n## Propositions\n"
for i, p in enumerate(self.propositions, 1):
protocol += f"{i}. {p}\n"
protocol += f"\n## Case Selection\n{self.case_selection_logic}\n"
protocol += "\n## Data Collection\n"
for source in self.data_sources:
protocol += f"- {source.value}\n"
return protocol
class CaseStudyDatabase:
"""Organize case study evidence"""
def __init__(self):
self.evidence = defaultdict(list)
self.chain_of_evidence = []
def add_evidence(self, source: DataSource, content: str,
date: str, related_rq: str):
"""Add piece of evidence"""
evidence_id = f"{source.value}_{len(self.evidence[source])}"
self.evidence[source].append({
'id': evidence_id,
'content': content,
'date': date,
'research_question': related_rq
})
return evidence_id
def link_evidence(self, evidence_ids: List[str], finding: str):
"""Create chain of evidence"""
self.chain_of_evidence.append({
'finding': finding,
'evidence': evidence_ids
})
def generate_evidence_table(self):
"""Create evidence summary table"""
rows = []
for source, items in self.evidence.items():
for item in items:
rows.append({
'ID': item['id'],
'Source': source.value,
'Date': item['date'],
'Research Question': item['research_question']
})
return pd.DataFrame(rows)
# Example
design = CaseStudyDesign(
case_type=CaseType.SINGLE,
research_questions=[
'How does the organization implement remote work policy?',
'What challenges emerge in the transition?'
],
propositions=[
'Remote work adoption requires cultural change',
'Technology alone is insufficient'
],
units_of_analysis='Organization-level remote work transition',
case_selection_logic='Selected for being early adopter with documented transition',
data_sources=[DataSource.INTERVIEW, DataSource.DOCUMENT, DataSource.OBSERVATION]
)
print(design.generate_protocol())✓ Reflexivity acknowledged
✓ Thick description provided
✓ Triangulation of data sources
✓ Member checking conducted
✓ Negative cases sought
✓ Audit trail maintained
✓ Saturation documented
✓ Context richly described✗ Researcher bias unacknowledged
✗ Thin description (anecdotes only)
✗ Single data source
✗ Cherry-picking quotes
✗ Ignoring disconfirming evidence
✗ Unclear analysis process
✗ Sample too small for claims
✗ Decontextualized findingsInterviews: 5-25 for phenomenology
20-30 for grounded theory
1-3 for case study
Focus Groups: 3-5 groups of 6-10 participants
Ethnography: Extended engagement (months to years)✓ Theoretical saturation reached
✓ No new themes emerging
✓ Negative cases explored
✓ All research questions addressed
✓ Rich description achieved