PURPOSIVE SAMPLING

SAMPLING KEY INFORMANTS

STRATEGIES
SNOWBALL SAMPLING
 INTERVIEWS
DATA
COLLECTION OBSERVATIONS
PROCEDURES DOCUMENT ANALYSIS
 STRUCTURED INTERVIEW
SEMI-STRUCTURED
RESEARCH INTERVIEW
INSTRUMENTS
UNSTRUCTURED
INTERVIEW
The learners will be able to:

•plan data analysis procedure

•identify the step-by-step process for
data analysis
•analyze data and infer patterns and
themes with the highest standards of
intellectual honesty
 GETTING STARTED?
Have you just conducted a qualitative study
involving…
• Interviews
• Focus Groups
• Observations
• Document or artifact analysis
• Journal notes or reflections?
 WHAT TO DO WITH ALL THIS DATA?

•Just as there are numerous

statistical tests to run for
quantitative data, there are just
as many options for qualitative
data analysis…
DATA ANALYSIS
Data Analysis

•According to Yin (1994:102) ―data

analysis consists of examining,
categorizing, tabulating, or
otherwise recombining the evidence
to address the initial prepositions of
a study.
LIKE CLEANING A CLOSET ???

Think of managing your qualitative

analysis process like cleaning your
closets – the same basic steps
apply!
It’s the same process…

• 1. Take everything out of the closet

• 2. Sort everything out – save or toss?
• 3. Look at what you have left and organize into sub-groupings
(chunking)
• 4. Organize sub-groups into clusters of similar things that
belong together (clusters, codes)
• 5. As you put things back, how would you group them to
maximize functionality? How do the groups make it work
together? (interpretation, presentation)
 FOUR BASIC STEPS
All qualitative data analysis involves the same four
essential steps:
1. Raw data management- ‘data cleaning’
2. Data reduction, I, II – ‘chunking’, ‘coding’
3. Data interpretation – ‘coding’, ‘clustering’
4. Data representation – ‘telling the story’,
‘making sense of the data for others’
 Step 1: Raw Data Management
•The process of preparing and organizing raw
data into meaningful units of analysis:
• Text or audio data transformed into transcripts
• Image data transformed into videos, photos, charts

As you review your data, you find that

some of it is not usable or relevant
to your study…
 Raw Data Sample
Transcript of Interview Data
• “Um,...finally starting the program Raw Data Overview
with others like me, it felt surreal.
Once you switch gears from being an •Are some portions of
established administrator at a college
to being a doc student, you realize this transcript
you lose control over your life. You are
not in charge in that classroom, like
unusable or
you are in your office. But also, once irrelevant? (purple)
you say you are a doc student, people
look at you differently.”
• “I can’t think of a particular example
right now…”
 Step II: Data Reduction I
• Get a sense of the data holistically, read several
times (immersion)
• Classify and categorize repeatedly, allowing for
deeper immersion
• Write notes in the margins (memoing)
• Preliminary classification schemes emerge,
categorize raw data into groupings (chunking)
 Winnowing

•Develop an initial sense of usable data and

the general categories you will create
•Preliminary set of codes developed, cluster
raw data into units that share similar
meanings or qualities
•Create initial code list or master code book
 Chunks-Clusters Sample
• I always wanted to get my doctorate but I never
felt I had the time; then I reached a point in my • Which sections of data are
career where I saw that without the credentials, I
would never advance to the types of positions I broadly similar? (red for
aspired to..but I doubted I could do the work. I credentials, blue for
wasn’t sure I could go back to school after so
much time. And did I have the time, with working personal struggles, green
and a family? These were the things I struggled for shift in identity)
with as I looked for the right program.
• finally starting the program with others like me, it • Which ‘chunks’ can be
felt surreal. Once you switch gears from being an clustered together to relate
established administrator at a college to being a
doc student, you realize you lose control over to a broad coding scheme?
your life.
Chunking? Clusters?
Transcript of Interview Data
 Step II: Data Reduction II
• The process of reducing data from chunks into
clusters and codes to make meaning of that data:

• Chunks of data that are similar begin to lead to initial

clusters and coding
• Clusters – assigning chunks of similarly labeled data into
clusters and assigning preliminary codes
• Codes – refining, developing code books, labeling codes,
creating codes through 2-3 cycles
 Coding Process

•Initial coding may include as many as 30

categories
•Reduce codes once, probably twice
•Reduce again to and refine to codes
that are mutually exclusive and include
all raw data that was identified as
usable
 A Priori or In Vivo Codes
• A Priori
• Codes derived from literature, theoretical frames

• In Vivo (inductive or grounded)

• Codes derived from the data by using code names
drawn from participant quotes or interpretation of the
data
• “Its like magic” is a phrase that could form the basis for a
code category
 Coding Levels
• Descriptive to Interpretative to Pattern Coding
• Moves from summary to meaning to explanation

• Open to Axial to Selective Coding

• Moves from initial theory to developing relationships between
codes for emerging theory

• First cycle to second cycle coding

• Moving from describing the data units to inferring meaning
 Coding Sample
• I always wanted to get my doctorate but I
never felt I had the time; then I reached a
point in my career where I saw that without • Chunking to coding:
the credentials, I would never advance to • Red for credentials – codes
the types of positions I aspired to..but I include career goals CG, career
doubted I could do the work. I wasn’t sure I advancement CA
could go back to school after so much time. • Blue for personal struggles- codes
And did I have the time, with working and a include self-doubt SD, time
family? These were the things I struggled management TM
with as I looked for the right program.
• Green for shift in identity – codes
• -finally starting the program with others like include student role SR, identity
me, it felt surreal. Once you switch gears at work IW, shift in control SC
from being an established administrator at a
college to being a doc student, you realize
you lose control over your life. Its like Chunking? Clusters? Coding?
magic!

Transcript of Interview Data

 Step III:
Data Interpretation & Themes
• ‘Chunks’ of related data that have similar meaning
are coded in several cycles
• Once coded, those ‘chunks’ become clustered in
similar theme categories
• Create meaning for those clusters with labels
• Themes emerge from those clusters
• Interpret themes to answer research questions
 Themes Sample
• I always wanted to get my doctorate but I
never felt I had the time; then I reached a • How do you compile the clusters
point in my career where I saw that without into emerging themes? (red for
the credentials, I would never advance to the credentials, blue for personal
types of positions I aspired to..but I doubted I struggles, green for shift in
could do the work. I wasn’t sure I could go identity)
back to school after so much time. And did I
have the time, with working and a family? • Begin to see themes emerge:
These were the things I struggled with as I Getting the degree, becoming a
looked for the right program. new person, personal
achievement…
• -finally starting the program with others like
me, it felt surreal. Once you switch gears
from being an established administrator at a How do broad sections emerge
college to being a doc student, you realize into thematic groupings?
you lose control over your life.

Transcript of Interview Data

 Step IV: Data Representation
• Interpretation or analysis of qualitative data
simultaneously occurs
• Researchers interpret the data as they read and
re-read the data, categorize and code the data
and inductively develop a thematic analysis
• Themes become the story or the narrative
 Data Representation Types
• Telling the story with the data
• Storytelling, Narrative
• Chronological
• Flashback
• Critical Incidents
• Theater
• Thematic
• Visual representation
• Figures, tables, charts
DATA ANALYSIS TYPES
CONTENT ANALYSIS
•Used to analyze and interpret
verbal data, or behavioral data
•Content can be analyzed
descriptively or interpretatively
DISCOURSE ANALYSIS
•a method of analyzing naturally
occurring spoken interactions and
written text and is concerned with the
social context in which the
communication occurred
•focuses on how language is used in
everyday life and looks at how people
express themselves.
GROUNDED THEORY
•also called analytic induction
•a method that attempts to develop
causal explanations of a phenomenon
from one or more cases being studied
•explanations are altered as additional
cases are studied until the researcher
arrives at a statement that fits all
cases.
CONVERSATION ANALYSIS
•examines the use of language by people
as a type of action or skilled
accomplishment
•A key concept in this analysis is the
principle of people taking turns in
conversation
•meanings are usually shaped in the
context of the exchange itself.
How does research design, data
collection, sampling,
procedures, methodologies and
analysis affect the over-all
research output of a proposed
study?