Prof. Kimberly Goyetteired.edu.vn/vn/Upload/files/lt.trang/researchmethods_ired.pdf · August 5,...
Transcript of Prof. Kimberly Goyetteired.edu.vn/vn/Upload/files/lt.trang/researchmethods_ired.pdf · August 5,...
August 5, 2014
“Qualitative and Quantitative Research: A comparative view
Prof. Kimberly Goyette
Inductive vs. Deductive Approaches
Comparing Qualitative and Quantitative Research
Collecting and Analyzing Qualitative Data
Collecting and Analyzing Quantitative Data
Overview
From the top-down
Goes from the general to the specific
Theory to hypotheses to observation to confirmation
Is most typical of quantitative research, but can also be found in much qualitative research
Deductive Reasoning
To test an already existing theory
To see if a “hunch” or preconceived notion is indeed supported
When you have a clear, answerable question
Deductive Reasoning: When is it useful?
Named and “systematized” by Nathan Glaser and Anselm Strauss
Start with data/observations
Record everything
Code right away: Begin to group similar observations
From these groupings themes emerge
Once themes emerge, go back into the field and seek confirmation
Inductive Approach: Grounded Theory
Identify ‘critical instances’ -highlight key passages of transcripts.
‘Open coding’ - assign passages to categories (i.e. abstract conceptual labels). Work through all transcripts and collect numerous illustrative quotes to ‘saturate’ categories.
‘Axial coding’ - refine initial list of categories. Delete and amalgamate some. Make connections between the categories and define their properties e.g. context, pre-conditions. These are sub-categories.
‘Selective coding’ - identify a core category and themes from which theory will derive.
In Glazer and Strauss’ words…
When is this useful?
No one has approached this topic/phenomenon before
Old theories do not seem to explain the topic well
Lots of disconfirming cases
The Grounded Approach
The theory is not well formulated
You don’t know how to measure concepts well
Your question makes assumptions about the phenomenon
Problems when using the Deductive Approach
Practical problems
Literature review?
Observer “positioning”
Problems with Grounded Theory
The Goal: Different Approaches to Theory Building
Using the deductive method (often appropriate for quantitative approaches):
Assess previous literature
What have other researchers found?
What theories have their results supported or refuted?
Building on analysis of previous research and your own ideas:
How are the concepts of your research question related?
Theory Building
Using the inductive method (best approached qualitatively):
Observe phenomena with as open a mind as possible
Try to record everything
Come up with descriptions for reappearing observations
Look for how phenomena are related
Building concepts to build theory
The Process of Theory-Building: Conceptualization
What are the underlying concepts in your research question/observations? (Sometimes you can get clues about these if you ask yourself: Why did you ask this question? What are similar question you might ask? Why are they similar?)
List those concepts
Validity Does the measure reflect what we are truly
trying to capture?
Face validity: Does the measure reflect a common understanding of the concept? ◦ How do you measure a family?
Criterion validity: Predictive validity. ◦ If something is designed to measure a concept,
how well does it predict that concept?
◦ University examination scores
Inductive vs. Deductive
Inductive ◦ Build a concept using grounded theory ◦ “Conceptualilzation: arises from observations ◦ Reliability? ◦ Generalizability?
◦ Deductive Concepts are derived from previous theory
Reliability can be strong because measures have been used previously
Generalizability can be strong
Some Motivations for Qualitative Research
Exploratory projects that focus on:
– Describing
– Understanding
– Explaining
Small in scale, but in-depth studies
Questions ◦ What?
◦ Why? (But not to determine causality, more to determine meaning)
◦ How?
(not: How many? How frequently?)
Qualitative Methods Are Used To gain more in-depth information that my be
difficult to convey quantitatively
To better understand any phenomenon about which little is yet known or knowledge is incomplete
To gain new perspectives on things about which we might already know a lot
May be used to generate theories and hypotheses
Very good for exploring surprising/unexpected findings
Good for exploring complexity
Good for exploring the effects of context
Some Characteristics of Qualitative Research
Phenomenon of interest should be minimally disrupted.
Researcher subjectivity and participation is acknowledged.
Often inductive in theory-building.
Often emphasizes the voices/perspectives of participants.
Acknowledges that participants construct their own “realities” and “stories,” and that events and “facts” can have multiple interpretations.
Research is often “emergent.”
Comparing Qualitative and Quantitative Methods
Quantitative
– Objectivity valued
– Social facts
– Reduction, control, and
Prediction
--Focus on causality
– Concepts are
measurable/quantifiable
– Report statistical analyses
– Researcher separate
-- Context is not the focus
Qualitative
– Subjectivity valued
– Multiple realities
– Discovery, description,
and
Understanding
-- Focus on interpretation
– Report rich narrative
– Researcher part of the
process
– Context dependent
Qualitative Research Choosing Who To “Research” -- Sampling
Purposive sampling: choosing respondents/groups to get the most complete understanding of a phenomenon
Quota:
“Enough” people from all salient categories
Snowball:
Friends of friends
Deviant cases:
What can we learn from those that do not conform to our expectations?
Balancing general with range of possible
Ethnography
Participant/Non-participant Observation
Interviews
Focus Groups
Content Analysis
Archival Research
Types of Qualitative Research
Describe everything
Pay attention to environment
Participants
As much detail as possible
Write reactions
Write as soon as possible
Draw maps
Note emotions
Memo to self at the end
Field Notes
Observing: What do you look for?
The short answer: everything
Initial impressions ◦ details about the physical setting including size,
space, noise, colors, equipment and movement, about people in the setting, such as number, gender and race, appearance, dress, movement, comportment, feeling and tone.
Interactional detail
Key events
The routine and mundane
Looking for categories: beginning to code
Beginning to build concepts:
Labeling observations
Grouping into categories
Naming categories (try not to borrow others’, at least not to start)
Finding properties of categories
e.g.
Frequency
Duration
Intensity
Open Coding can proceed
Line-by-line
By paragraph
By document
Axial Coding
Moving back from the data
Seeing how categories and sub-categories relate to each other
Create outlines, pictures, maps, etc. of the categories and sub-categories
(You may be going back and forth from data to analysis at this point)
Selective Coding
Recognizing a main theme you want focus on
Organizing data according to the categories and sub-categories of that theme
Building a narrative/story/theory from that theme
Sequential analysis suggest that you return to the data to check that interpretation and modify as necessary
Consideration of disconfirming cases
Enough is enough?
When do you know or feel you are done? When can/should you stop collecting data?
Test of congruence/verifiability Could you explain the rules, patterns, norms
of a setting to an outsider? Does what seemed strange at first seem natural or normal now from the perspective of a group member? Can you take the position of understanding the world form your group members’ eyes?
Sometimes you can show coding and analysis to a member of the group to see if they concur with your interpretation.
Process
Start with codes
Put codes into categories
How do categories relate to each other?
This is the theme.
Build to a theory, comparing categories under a theme.
Illustrate the categories using direct quotes from your field notes.
Collect own data from surveys, experiments
Use secondary data
Assuming you have ◦ Asked a research question
◦ Checked relevant literature
◦ Built theory
◦ Operationalized concepts
◦ Measured concepts validly and reliably
Quantitative Analyses
Descriptive Statistics
You want to tell me what that picture says using some summary numbers (statistics).
What characteristics do you want to capture? What will help me make a picture?
How to Summarize Data: Looking at a Distribution/Shape
Where is the center? ◦ What does the average or typical case look like?
How spread out is the distribution? ◦ How much variation is there among cases? (next
week)
Frequency and Relative Frequency Tables
Show the numbers or count in each category.
Measures of Center for Categorical Variables Mode: The single value (or values) that
appear most often.
Bar Chart of Marital Status
Show the percentage in each category.
Pie Chart of Marital Status
Bivariate Relationships: What do we want to know?
Sample:
Describe the association (pattern/direction)
Population:
Is this relationship likely to exist in the population to which we want to generalize?
(Hypothesis testing)
Bivariate Relationship with Two Categorical Variables
Describe: Crosstabulation Tables or Crosstabs, Side by Side bar charts, pie charts
Hypothesis testing: Chi-square
Crosstabs
Describe the association
Side-by-Side Bar Chart
Hypothesis testing The Problem: Are married people happier (in the
population)?
Step 1: Specify the H0 and HA.
H0: Happiness is independent of marital status.
HA: Happiness is not independent of marital status.
Questions?