Setting Up a Qualitative or Mixed Methods Research Project in

Report
Innovative Media and Tools Track
7th Annual International Symposium
Emerging Technologies for Online Learning
(sponsored by Sloan-C and MERLOT)
Shalin Hai-Jew
April 2014
NVivo 10 is a widely used qualitative and mixed methods data analysis research tool used in
many domain fields. This presentation will showcase some of the capabilities of this software
by reviewing how to start a project in NVivo, how to manage complex data, how to begin
coding raw information, and some of the methods for making meaning “in vivo” or “within the
living” with various analyses and data visualizations. This presentation will include a
demonstration of NCapture, a feature for data extraction from some social media platform
APIs. For context, this session assumes some basic prior training in qualitative research, such as
ethnographies, surveys, interviews, qualitative meta-analyses, portraiture, grounded theory,
action research, Delphi studies, and / or others. The goals of this presentation are to
 introduce NVivo 10 as a data management and analysis tool for qualitative and mixed
methods research
 introduce the concept of multimodal data (from various types of multimedia sources)
 show data ingestion into the system
 establish how data may be coded for research value
 show how data structures may be created from the coded data and then analyzed
 demonstrate the extraction of a social media platform to open up another data channel for
research
Setting Up a Qualitative or Mixed Methods Research Project in NVivo 10 to Code and Analyze Data
2





Pre-story
Types of qualitative and mixed methods
research
General timeline of the research process
Quality standards
Ingestion of data
 Importation of data sets
 File types: multimedia

NVivo 10 (an overview)
Setting Up a Qualitative or Mixed Methods Research Project in NVivo 10 to Code and Analyze Data
3

Data coding
 Folder and node data structures
 Codebooks
 Manual coding
 Two types of auto-coding
 Flexible coding structures
 Coding for matrix analysis
 Human-machine analysis of data
Setting Up a Qualitative or Mixed Methods Research Project in NVivo 10 to Code and Analyze Data
4

Data coding (cont.)
 Visual models
 Text search query
▪ Text search query and resulting word tree
 Manual research journaling
 Automated ‘event log’ journaling




NCapture
Essential mechanics
My wants re: the software
Contact and conclusion
Setting Up a Qualitative or Mixed Methods Research Project in NVivo 10 to Code and Analyze Data
5

Multi-year site license at Kansas State University
 Sociology, anthropology, social work, education,
communications studies, mass media, business,
veterinary sciences, food sciences, family studies,
human ecology, and others; university research office;
multi-campus support
▪ Installed in student labs in multiple departments
 Available to faculty, staff, administrators, and
graduate students through protected website (with
software download and a university-wide license key)
 Y1 webinar trainings
Setting Up a Qualitative or Mixed Methods Research Project in NVivo 10 to Code and Analyze Data
6









Grounded theory
Heuristics / methods of
discovery
Hermeneutics
Life history
Participatory action
research
Phenomenology
Ethnography
Standard causal analysis
(SCA)
Narratology; serious
narration; storytelling;
historical narration









Social experimentation
Formalization; analysis
of “stylized facts”
Literary critique
Market research
Interviews
Focus groups
Surveys
Delphi study; modified eDelphi study
Actor-network analysis;
electronic social network
analysis; social network
analysis, and others
Setting Up a Qualitative or Mixed Methods Research Project in NVivo 10 to Code and Analyze Data
7

Mixed methods / multimethodology:
qualitative research methods with
quantitative research elements
Setting Up a Qualitative or Mixed Methods Research Project in NVivo 10 to Code and Analyze Data
8
QUALITATIVE METHODS






Relativism (with attendant
theories, tools, and research
practices); post-structuralism
and post-modernism
Non-measurability of
phenomena
Narration
Culturalism
Emergentism
Contextualization; social
constraints
QUANTITATIVE METHODS






Positivism (with attendant
theories, tools, and
research practices)
Measurability of
phenomena
Analysis
Behavioralism
Individualism
Non-contextualization;
independent agentry; open
choices
Setting Up a Qualitative or Mixed Methods Research Project in NVivo 10 to Code and Analyze Data
9
QUALITATIVE METHODS
QUANTITATIVE METHODS
Situated knowledge
 Subjectivity
 Constructionism
 Experiential and detailed
 Low n, case-based
 in vivo: “within the living”
(studying empirically inworld) (in Latin)







Transcendent or universal
knowledge
Objectivity
Realism
Statistical and numerical
“Statistical significance” /
relatively high n (and even
“big data”)
in vitro: “within the glass”
(experimental model) (in
Latin)
Setting Up a Qualitative or Mixed Methods Research Project in NVivo 10 to Code and Analyze Data
10
QUALITATIVE METHODS




Rigor, saturation,
immersion, triangulation
Criticality of identification
(and methodological
mitigation) of researcher
biases
Centrality of language
Unclear end point
QUANTITATIVE METHODS
Rejection of the null
hypothesis
 Criticality of nullifying
researcher biases through
strict methodologies
 Centrality of numbers
 Generally defined end
point

Setting Up a Qualitative or Mixed Methods Research Project in NVivo 10 to Code and Analyze Data
11


Single approach designs, with augmentation
of either qualitative or quantitative strategies
to enhance research richness and outcomes
(with a predominant qualitative or
quantitative approach)
Mixed approach designs, to enhance
construct validity for multiple analytical
interests
Setting Up a Qualitative or Mixed Methods Research Project in NVivo 10 to Code and Analyze Data
12
13

General steps:
1. Initial secondary and primary research
2. Immersive research and documentation
3. Coding and annotation
4. Data analysis
5. Write-up
6. Dissemination

Flexible semi-recursive process
Setting Up a Qualitative or Mixed Methods Research Project in NVivo 10 to Code and Analyze Data
14

Various quality standards at every step (depending
also on ontology and research methods)
 Protection of human participants (informed consent;






protection from harm; voluntary participation throughout;
anonymity and confidentiality; beneficence)
Data saturation (comprehensive literature review)
Data triangulation (multi-sourcing of data)
Clear definitions of terms
Address of researcher bias
Inter-rater reliability; rater norming to standards
Transparency in research methods
Setting Up a Qualitative or Mixed Methods Research Project in NVivo 10 to Code and Analyze Data
15
 Internal logic and non-contradiction; plausibility
 Falsifiability of hypotheses and assertions
 Logical analysis
 Bounded and delimited assertions
 Sufficient detail and depth
 Consistent and appropriate unit(s) of analyses
 Explanatory power
 Clear language
 Intellectual elegance and originality
Setting Up a Qualitative or Mixed Methods Research Project in NVivo 10 to Code and Analyze Data
16
 Puzzle solving
 Problem-solving
 Informative for decision-making
 Practical applications (decision-making; policies)
 Data protection against leakage and compromise
Setting Up a Qualitative or Mixed Methods Research Project in NVivo 10 to Code and Analyze Data
17
PRIMARY








Interviews
Surveys (.xls, electronic
platform files)
Focus groups
Cases
Observations
Oral histories
Text, photos, videos, and
audio files
Social media platform
contents (NCapture)
SECONDARY





Literature reviews
YouTube videos (links only)
Web pages
Social media contents
Grey literature
Setting Up a Qualitative or Mixed Methods Research Project in NVivo 10 to Code and Analyze Data
18

AUTOMATED (data import wizard, thirdparty tools like Survey Monkey)
 From electronic survey sites and tools
▪ Cells whited out for open-ended questions; grayed-out
cells for fixed response (attribution details)
 From social media platforms and sites (NCapture
in IE and Chrome and extracted into NVivo)
 From content providers
 From Excel (codebooks); SPSS; SAS, etc.
 From bibliographic data managers / citation tools
Setting Up a Qualitative or Mixed Methods Research Project in NVivo 10 to Code and Analyze Data
19

MANUAL (drag-and-drop)
 From online libraries and repositories
 From Excel
Setting Up a Qualitative or Mixed Methods Research Project in NVivo 10 to Code and Analyze Data
20
Text files (.docx, .doc, .rtf, .txt, .pdf, and others)
Slideshows (.pptx, .ppt)
Datasets (.xls, .rtf, .txt, .doc, and others)
Websites captured as .pdfs (using NCapture or
other tools)
 Video files (and link-outs to video files on
content-sharing platforms)
 Audio files (and link-outs to audio files)
 Image files (virtually all digital image files)




 and others…
Setting Up a Qualitative or Mixed Methods Research Project in NVivo 10 to Code and Analyze Data
21


Requires manual transcription of audio and
video files
Requires annotation
 Annotation of audio
 Annotation of video stills and video segments
Setting Up a Qualitative or Mixed Methods Research Project in NVivo 10 to Code and Analyze Data
22
An overview
Setting Up a Qualitative or Mixed Methods Research Project in NVivo 10 to Code and Analyze Data
23




Formerly NUDIST (non-numerical
unstructured data indexing, searching, and
theorizing)
Runs on PCs and parallels / virtual machines
on Macs; professional Mac version
forthcoming June 2014; mobile app access in
development
Includes add-ins that enable extraction of
social media platform data
Created by QSR International
Setting Up a Qualitative or Mixed Methods Research Project in NVivo 10 to Code and Analyze Data
24

Tool as theory and research methodology “agnostic”
and flexible
 May be a longitudinal many-year project or a short-term
one
 May be individual researcher-based or team-based
(without requirement for co-location, even with nonserver based licensure because of data integration
options)
Live support webinars with responsive presenters (for
both broad audiences and organizations with site
licenses, at additional cost)
 Pre-recorded help webinars
 Textual help supports well documented

Setting Up a Qualitative or Mixed Methods Research Project in NVivo 10 to Code and Analyze Data
25
Setting Up a Qualitative or Mixed Methods Research Project in NVivo 10 to Code and Analyze Data
26

Project files saved as .nvp (QSR
NVivo Projects) files
 Back up to server, cloud service,
external hard drive, or other device
beyond main work device
▪ Coding is time-sensitive and
irrecoverable and impossible to
recreate if there is data loss
 May hover over file names at
bottom left of opening screen to
see the whole path of the files
Setting Up a Qualitative or Mixed Methods Research Project in NVivo 10 to Code and Analyze Data
27
Setting Up a Qualitative or Mixed Methods Research Project in NVivo 10 to Code and Analyze Data
28
A PRIORI DATA STRUCTURE

Pre-defined with general
constructs, data
categories, and ideas
(based on deductive logic
and prior research)
MIXED DATA STRUCTURE

A priori and emergent data
structure
EMERGENT DATA STRUCTURE

Initially undefined
structures with information
extracted from the
research documentation
and researcher journal and
other sources (based on
inductive logic and
interpretation of collected
information); aka
freecoding
Setting Up a Qualitative or Mixed Methods Research Project in NVivo 10 to Code and Analyze Data
29





Units of analysis
Data manageability
Relevance
Ways to flexibly query the data using the
software
Semantics and terminology
 Should document and describe all nodes for
coding clarity (particularly if there are teams
working together on one project); codebook
definitions, too
Setting Up a Qualitative or Mixed Methods Research Project in NVivo 10 to Code and Analyze Data
30
31

Definitions of codes and their meanings (often
with examples)
 May be static (and predefined, partially defined based
on hypothesis and secondary research)
 May be evolving (and emergent and recorded in
codebooks along the way…but initially unstructured)


Digital or on paper
May be used with one coder or a group
 Inter-rater reliability (Kappa coefficient in coding
comparisons between coders) and norming
Setting Up a Qualitative or Mixed Methods Research Project in NVivo 10 to Code and Analyze Data
32


Categories or classifications as nodes
Textual / auditory / visual (still) / video
information highlighted and marked as
belonging to a certain node
 Highlighting and dragging text (of text or .pdf
OCR-text) to node or right clicking and coding to
node
 Highlighting an area (of image or sound or video)
and dragging to node or right clicking and coding
to node
Setting Up a Qualitative or Mixed Methods Research Project in NVivo 10 to Code and Analyze Data
33
“STYLES” IN WORD

Ingestion and
simultaneous autocoding
of MS Word files with
questions coded as
headings and answers as
paragraph styles (styles
used to code a mass of
data)
ALGORITHM-BASED
Machine-based emulation
of qualitative / mixed
methods researcher
highlighting of contents
into certain nodes
 Machine-coded
information clearly
identified as such (with
coding easily reversible)

Setting Up a Qualitative or Mixed Methods Research Project in NVivo 10 to Code and Analyze Data
34

Nodes used for
 data structuring (and data visualizations)
 data clustering for analysis (for patterns and
researcher sense-making and understanding)
 further data querying
 pattern identification
Folders and nodes may be re-named, moved,
annotated, regrouped, and changed at any time
 Coding stripes to indicate density of coding and
identify patterns
 Can always uncode, revise, move nodes around,
merge nodes, and even delete them

Setting Up a Qualitative or Mixed Methods Research Project in NVivo 10 to Code and Analyze Data
35
Variable
Variable
Variable
Variable
Variable
Variable
Variable
Individual
or Group
(based on
attributes)
”
Variable
Variable
Variable

May export to Microsoft Excel for larger
range of data visualizations
Setting Up a Qualitative or Mixed Methods Research Project in NVivo 10 to Code and Analyze Data
36

Matrix analysis enabling ability to slice and
dice data for findings and insights
 Within and between-group analysis
 Enabling comparisons of variables from
categories as rows and columns
 Enabling simple cell counts for comparisons,
anomaly detection, others (if relevant)
Setting Up a Qualitative or Mixed Methods Research Project in NVivo 10 to Code and Analyze Data
37

Code…
 Participants / interview subjects as own nodes
 Participants with demographic data (in attributes
and categorizing / classifying data fields such as
age group, sex, and others)
 Participants with close-ended query responses as
attributes (such as Likert scale responses, Y/N
responses)
 Questions as own nodes (variables)

Matrix analyses results saved in new node(s)
Setting Up a Qualitative or Mixed Methods Research Project in NVivo 10 to Code and Analyze Data
38

Text frequency count (with a built-in
and modifiable stop-words list)
 Word clouds (word size and placement
centrality = frequency)
 Treemap (size = frequency)
 Cluster analysis (branch = coappearance)
39
 Human observations and readings /




interpretations of the research
Text searches
Text frequency searches
Machine-enhanced data queries
Data visualizations and structures
Setting Up a Qualitative or Mixed Methods Research Project in NVivo 10 to Code and Analyze Data
40
41


Node hierarchies in any direction (parent,
child, etc.)
Types of layouts
 Circular
 Directed
 Hierarchical
 Orthogonal


Revealing of relationships between entities
More effective with sparse contents
Setting Up a Qualitative or Mixed Methods Research Project in NVivo 10 to Code and Analyze Data
42


Words used in a group with “and” to indicate
co-location in the same document (not in the
same sentence or near-proximity)
Data management
 Find a document or identify an author
 Identify information for use in a quote

Use “summary” view to look at types of
sources
Setting Up a Qualitative or Mixed Methods Research Project in NVivo 10 to Code and Analyze Data
43
Setting Up a Qualitative or Mixed Methods Research Project in NVivo 10 to Code and Analyze Data
44
Enables intuitions and leads; provides a gist of
the data
 May click over a part of the branch to get its full
phrasing
 May request broader captures (than the narrow
5 words on either side of the focus term), termed
“spread coding”
 May put cursor over phrases in a tree to see the
source (which is linked and may be accessed for
further exploration)
 Words in red are ones in initial text search query

Setting Up a Qualitative or Mixed Methods Research Project in NVivo 10 to Code and Analyze Data
45


Via built-in memo and Memo Links
Via unique folder or node (and variety of
multimedia files)
46

Apply real-time event logging to project
Setting Up a Qualitative or Mixed Methods Research Project in NVivo 10 to Code and Analyze Data
47

May highlight text for annotation
(highlighted in blue and viewable from
various approaches)
 Analyze -> Annotations -> New Annotations


Limit annotations because too much blue
highlighting will be ignored
If too many annotations are needed, that
may indicate the need for a coding node
instead
Setting Up a Qualitative or Mixed Methods Research Project in NVivo 10 to Code and Analyze Data
48

Per Dr. Cynthia Jacobs (of QSR International)
 Keep a memo for justifying methods for the coding
 Set up a node for “I am struggling with how to code




this” issues for later reference
Maintain a node for “good quotes”
Keep a “to do” node list
Do not delete nodes which you don’t plan to use, but
save them in a folder for possible future reference if
needed
Use a “left off” node to remember where the coding
was stopped, so you can find your place later
Setting Up a Qualitative or Mixed Methods Research Project in NVivo 10 to Code and Analyze Data
49
An add-on to Chrome or Internet Explorer browsers
Setting Up a Qualitative or Mixed Methods Research Project in NVivo 10 to Code and Analyze Data
50





Websites (with or without ads)…as .pdfs
Twitter microblogs
Facebook text and images
LinkedIn text and images
YouTube links
 May require log-ins on the various platforms to
access the social media platform APIs (application
programming interfaces)

A sociometric online ethnography
Setting Up a Qualitative or Mixed Methods Research Project in NVivo 10 to Code and Analyze Data
51
Setting Up a Qualitative or Mixed Methods Research Project in NVivo 10 to Code and Analyze Data
52




Collection
Coding
Analysis
Output
Setting Up a Qualitative or Mixed Methods Research Project in NVivo 10 to Code and Analyze Data
53


Collection of data in folder and file structures
Manual coding and analysis of data
electronically (or on paper) (cognitively taxing
no matter what form)
 Wordle (for word clouds), Freemind (spatial
outlining), and others
 Notetaking in MS Word
 Manual modeling in MS Visio

Outputs for dissemination in MS Office Suite
and other tools
Setting Up a Qualitative or Mixed Methods Research Project in NVivo 10 to Code and Analyze Data
54
Setting Up a Qualitative or Mixed Methods Research Project in NVivo 10 to Code and Analyze Data
55




Wizarding and templating: Clearer setup for
different types of qualitative and mixed
methods research
Quant analysis features: A wider range of basic
quant and statistical analysis packages for more
of an all-in-one feel
A friendlier learning curve: Simpler semantics;
more aggressive and interventionist help
supports
A built-in source bibliographic data manager /
citation tool (or an integrated freeware one):
Not just EndNote, which is a commercial tool
Setting Up a Qualitative or Mixed Methods Research Project in NVivo 10 to Code and Analyze Data
56

A public repository for NVivo data sets (to be
released with formal publications)
 A repository of research journals linked to NVivo-
managed data sets and archived research
 Agreements to enable the release of copyrighted
materials in NVivo-based data sets (for this
limited purpose)
Setting Up a Qualitative or Mixed Methods Research Project in NVivo 10 to Code and Analyze Data
57

Dr. Shalin Hai-Jew, Instructional Designer







Information Technology Assistance Center (iTAC)
212 Hale Library
Kansas State University
Manhattan, KS
785-532-5262
[email protected]
The presenter has no tie with QSR
International.
Setting Up a Qualitative or Mixed Methods Research Project in NVivo 10 to Code and Analyze Data
58

similar documents