ChapterPDF Available

Computer-Aided Qualitative Re- search: A Users Perspective



Content may be subject to copyright.
11 Computer-Aided Qualitative Re-
search: A Users Perspective
A. Lamont-Mills, Toowoomba, Australia
List of Contents
11.1 Introduction
11.2 Qualitative Data Analysis Software: A Variety of Programs
11.2.1 Text Retrievers
11.2.2 Text Base Managers
11.2.3 Code-and-Retrieve Programs
11.2.4 Code-Based Theory-Builders
11.2.5 Conceptual Network-Builders
11.3 A Word of Warning - Computers and Qualitative Data Anal-
11.4 NUD*IST 6: The Logic and the Structure
11.5 An Example. The Construction of Gender and Gender Iden-
tity in Elite Australian Sport
11.6 Limitations and Considerations
11.1 Introduction
Over the past decade there has been a rapid growth in the development and use
of computer-assisted qualitative data analysis software (CAQDAS) for use with
text-based data. Similarly there has been a growth in the number of qualitative
research articles being published in sport journals. Thus this chapter is timely as
an introduction and example of how CAQDAS can be used by qualitative sport
researchers. This chapter will briefly introduce CAQDAS and how different
software programs can be grouped based on function or assistance to the re-
After having studied this chapter, the reader is expected to:
1. Recognize the possibilities offered by the use of computers for the
analysis of data coded in words.
2. Understand the limitations associated with using computers in qualita-
tive research.
3. Identify which type of computer-assisted qualitative data analysis soft-
ware is best suited for different research purposes.
4. Understand the logic and structure of one particular computer-assisted
qualitative data analysis software program, NUD*IST 6.
searcher. A cautionary warning will then be sounded to the reader concerning
the use of CAQDAS before the chapter demonstrates the logic underpinning
one particular CAQDAS software program, a code-based theory-builder,
NUD*IST 6. NUD*IST 6 is produced by QSR International Pty Ltd and stands
for Non-numerical Unstructured Data Indexing Searching and Theorising. The
choice to include NUD*IST 6 in this chapter is driven by my familiarity with
the program. NUD*IST 6 was used in my doctoral dissertation (Lamont-Mills,
2001) and extracts from this research will be used to illustrate points made dur-
ing this chapter. The chapter concludes with some limitations of the NUD*IST
6 program along with wider CAQDAS limitations and concerns.
11.2 Qualitative Data Analysis Software: A Variety of Programs
In terms of what types of CAQDAS programs are being used by qualitative
sport researchers, a search of the literature reveals a number of different pro-
grams being utilised including ATLAS.ti (see Hoeber, 2008), HyperRE-
SEARCH (see Chu, Leberman, Howe, & Bachor, 2003), NUDIST (see Lane &
Matheson, 2005), NVivo (see Wright, Trudel, & Culver, 2007), and Leximancer
(Campbell, Gross, & Dodd, 2008). Different researchers use different
CAQDAS programs for different purposes. That is, CAQDAS programs can be
used to perform a variety of research tasks, from simple editing functions, data
management, right through to complex theory development (Bazeley, 2007;
Drisko, 2004; Gibbs, 2002; Krahn & Putnam, 2003; Lewins & Silver, 2007).
Given that different researchers use different CAQDAS for different purposes,
and the diversity of CAQDAS available, it is helpful to conceptualise CAQDAS
in terms of the type of analytical assistance each program affords the researcher.
This can be done by grouping this assistance into the five categories presented
below. The categories are listed in increasing levels of sophistication ranging
from text retrievers, text base managers, code-and-retrieve software, code-based
theory-builders, and conceptual network-builders1.
11.2.1 Text Retrievers
At the lowest level of analytical assistance sits text retrievers. Here focus is
on the management of data through the searching and retrieving of words,
phrases, and other string characters that have been determined by the researcher
to be of analytic interest. This can be inductively or deductively determined.
Nearly all CAQDAS available today allows for quick searching and retrieval of
data segments, albeit using different search strategies. To illustrate how text
retrieval would work, my PhD focused on gender and discourse. 75 elite ath-
letes and coaches were interviewed resulting in 75 transcribed interview word
files. From a text retrieval perspective, CAQDAS could be used to search and
retrieve all instances of the word feminine from each interview file as the re-
searcher would be interested in how this word is used within the sporting con-
text. In some older text retrieval CAQDAS programs, the raw data (e.g., inter-
view file) was not entered into the program but stood outside the program
where it was searched by the program. However more recent CAQDAS allows
for the raw data to be loaded into the program as a project. All text searches are
then done within the project (and thus the program). This allows for search out-
puts to be conveniently stored in the project for further consideration by the
researcher. What text retrievers allow is for all instances of a word or phrase to
be captured in one search. However, this does not mean that each identified
instance is of relevance to the research question. The researcher must still ex-
amine each instance and make a determination regarding if this will be used,
how will this be used and where will it be coded.
11.2.2 Text Base Managers
The text base managers organise and store the raw data within the CAQDAS
program. These programs are able to systematically sort the database into mean-
ingful subsets for comparison and contrast by the researcher. For instance, one
could search and retrieve all instances of the word feminine but this time one
wishes to group the search output into two separate subsets, male athlete re-
sponses and female athlete responses. Thus the search output would be located
in two separate subsets. This would allow one to explore how the word femi-
nine is being used both within each subset and then allow one to compare across
each subset. Of course, before any exploration occurs the researcher still needs
to determine how relevant and meaningful each identified instance is in each
sub-set. Again as with text retrieval, most CAQDAS now allow for text based
11.2.3 Code-and-Retrieve Programs
The code-and-retrieve programs are an advance on the previous two catego-
ries in that they enable the researcher to not only retrieve and separate the data;
they also allow the researcher to code the data. That is, lines, sentences or para-
graphs can be coded on the basis of the research purpose with more advanced
programs being able to generate code frequency data as well. For example, one
could search and retrieve all instances of the keywords feminine, girly girl, girl
stuff, wearing make-up, and after careful consideration and assessment (that is
often ongoing) label all located items with the code feminine. The previous two
program categories do not necessarily allow coding to be conducted at the same
time as searching and retrieving nor do they necessarily allow for coding to be
stored within the program itself. The reader should note that there is considera-
ble difference amongst CAQDAS in terms of how each program codes and
marks the coded text. As with any program, a test trial to see how coding and
marking will allow the researcher to see which type or presentation and format
best suits his or her needs and preferences.
11.2.4 Code-Based Theory-Builders
The code-based theory-builders not only retrieve-and-code but also assist
the researcher to develop and test theory. Here categories can be developed
from the assigned codes, memos can be written and linked to these codes and
categories, and hypotheses that have been induced from the data can be formu-
lated and tested. For example, one could search and retrieve all instances for
keywords feminine, girly girl, girl stuff, wearing make-up, and label all located
items with the code feminine. In subsequent analysis the code feminine could
be grouped with the code masculine into the category “gender stereotypes”.
Further analysis of the gender stereotypes category could result in this being
subsumed under the higher order theme “gender”. The development and defin-
ing qualities of the two codes and the two categories would be reported in a
memo within the program that is attached to the category “gender stereotypes”.
Indeed memos can be written for and stored for each search and each code.
From this one may then hypothesise that the data appears to suggest that male
athletes construct gender stereotypes differently than female athletes. This hy-
pothesis could be tested via a search-retrieve-and-comparison of male and fe-
male responses. Again there is considerable difference amongst CAQDAS pro-
grams in how the researcher’s emerging theories are displayed and manipulated
by the researcher. A test run of different programs will allow the researcher to
best choose a program for their needs.
11.2.5 Conceptual Network-Builders
The last category of CAQDAS is conceptual network-builders. These pro-
grams use semantically meaningful networks to build and test theory. Further,
the researcher’s thinking and conceptualisation of the data can be represented
graphically in these programs. These programs allow for the researcher to de-
velop a visual representation of their theory but also allow the researcher to
refine and revises with relative ease. Unfortunately most CAQDAS programs
that are able to build theory networks are somewhat limited in how the network
images are represented and the images often appear somewhat crude when
compared to specific network building software.
11.3 A Word of Warning Computers and Qualitative Data Analysis
The phrase ‘qualitative data analysis software’ is somewhat of a misnomer in
that such software does not follow the same analytical principles or workings as
quantitative analysis programs (e.g., SPSS, SAS, AMOS, etc.). This is an im-
portant point for the reader to understand. Those readers who are familiar with
quantitative analysis programs would understand that after raw data is entered
into a program such as SPSS, the researcher proceeds with a particular analysis
of interest where the software program performs the actual analysis.
In this sense quantitative analytical programs compute the variance in the data
and then determine whether this variance is of some importance. That is,
whether there are differences in the variance between groups, between pre and
post treatment conditions, and so forth. Quantitative data analysis programs tell
the researcher whether the measured variance is important or of some value
based on probability levels, effect size, and power. For example, from the PhD
example one is able to say that elite male athletes and elite female athletes do
not differ in their responses to the Personal Attributes Questionnaire (PAQ:
Spence & Helmreich, 1978) because the F value was not significant. Probability
levels, effect size, power, and so forth are well-established parameters or stand-
ards of importance in quantitative statistics2.
CAQDAS, and in particular code-based theory-builders, do not analysethe
data as quantitative programs do. Indeed CAQDAS does not analyse the data at
all. Rather CAQDAS assists analysis by making information more accessible to
the researcher for assessment and judgement and ultimately analysis. Thus the
software does not perform the analysis itself. The researcher is and always will
be responsible for making analytic decisions when undertaking qualitative re-
search. CAQDAS can systematically and logically search and organise the data
for the researcher (remembering that these searches are driven by the research-
er), and this searching and organisation may lead to the highlighting of an area
of potential importance for the researcher. However CAQDAS will not tell the
researcher what to search nor will they tell the researcher whether the search
output is important or of some value. CAQDAS potentially enhances the re-
searcher’s thinking about complex data by logically and systematically structur-
ing and storing information in ways that can facilitate the discovery and explo-
ration of meaning in unstructured data. Here the use of the word potentially is a
deliberate choice. CAQDAS only potentially enhances the researcher’s thinking
and conceptualisation of the data. Any program is only as good as the research-
er using it, as is true of quantitative analytic programs. The qualitative research-
er, therefore, still needs to engage in reflexivity, critical thinking about the data,
rigorous methodological practices, and so forth. Any judgements or assessments
about the data in terms of relevance, meaning, and importance can only be
made by the researcher (Drisko, 2004). CAQDAS may make the data more
manageable but it does not automatically make sense of the data, it does not
automatically make the research more rigorous, and thus more methodological-
ly and theoretically sound.
11.4 NUD*IST 6: The Logic and the Structure
What follows is not a users guide to the NUD*IST 6 program nor is it a detailed
technical description of the program. For those readers interested in these as-
pects the QSR web site and the User guide that accom-
panies NUD*IST 6 are the best places to find this information. Further, I will
not debate the advantages of the NUD*IST 6 program over other CAQDAS.
What will be done is a focus on the logical structure of the NUD*IST 6 pro-
gram and how this structure can potentially enhance the analysis of qualitative
data in the sport domain based on the experience with the program.
As mentioned previously, NUD*IST 6 is a code-based theory-builder pro-
gram that stores, searches, retrieves, codes, and aids in the analysis of qualita-
tive data. The program allows the researcher to manage and explore the data in
two interlocking systems, the document system and the index system. The doc-
ument system (called document explorer in Figure 11.4.1) is the holding sys-
tem for the raw data. Raw data or documents can be explored and coded within
this system. By doing this, the documents are linked to codes or categories that
reside in the index system. Thus no codes are stored in the document system.
The index system (called the node explorer in Figure 11.4.1) is, therefore,
where the codes, categories, or themes are organised. The reader should note
that NUD*IST 6 refers to nodes not codes but they are essentially the same
thing. Nodes contain the coded data as well as the researcher’s thinking about
the data as a memo attached to the node. In this way nodes can be explored and
further coding done on them. By doing this, nodes in the index system are au-
tomatically linked to the documents that reside in the document system (see
Figure 11.4.1 for the document system and index system as they appear in
NUD*IST 6.).
Figure 11.4.1 Index and document systems in NUD*IST 6.
11.4.1 The Document System: How to Manage Mountains of Raw Data
The document system allows the researcher to collect and organise the raw data,
study and explore the raw data, develop ideas about it, edit and annotate the raw
data, make notes and memos about raw data, and search and retrieve words and
phrases from the raw data. It is the researcher’s entry point to all the raw data in
the research project. In the research, the 75 transcribed interviews were held in
the document system. The document system thus allows the researcher to work
on the raw data, code it, and think about it. Exploration in the document system
organises the documents first, nodes second. The advantage of the document
system is that it codes, retrieves, and browses raw data more thoroughly, more
rigorously, and faster than can be done manually. Thus repetitious, factual, and
descriptive coding can be efficiently handled by the document system.
As previously mentioned, coding results are not stored in the document sys-
tem. They are stored in the index system. Through interlocking with the index
system, exploration and interpretation of the data becomes a continuous pro-
cess. Coding that is both inductive (codes that emerge from the data) and deduc-
tive (codes that are imposed on the data) can be efficiently handled by the sys-
Different researchers can and do use the document system in various ways. A
researcher can work predominantly within the document system where individ-
ual documents are explored, and the text contained in each document coded
with the results of each code being placed in the index system. Here the re-
searcher works on the raw data within the document system. Working in this
way, however, limits the researcher to basic word or phrase searches on the
documents. This is analogous to working with different coloured highlighter
pens when coding the original transcript. It does not allow for comparison be-
tween documents or between codes. For this the researcher needs to move into
the index system.
11.4.2 The Index System: Thinking about Mountains of Raw
Other researchers prefer to work primarily within the index system, and treat
the document system mainly as a raw data storage area. In this way the docu-
ment system is used for an initial search, code, and retrieval of factual data
(e.g., male athlete responses, male coach responses). The results of this initial
search are placed at the database node in the index system (see Figure 11.4.2 for
an example of a database node). This approach requires some forethought by
the researcher as to what should be included in the database node and is deter-
mined by the research purpose, the background, and interests of the researcher.
A node in NUD*IST 6 may be set up so that it has two or more sub nodes or
children that sit under the one parent node. Remember that nodes hold all the
researchersthinking about the node and can include lower order concepts with-
in the one node. To illustrate, I set up a database node that separated the raw
data into a sex node, which had two children nodes or sub nodes – ‘malesand
females. Further, I set up a status node, which had two children nodes
coachesand athletes. What this did was to store all the male interviews with-
in the database child node male, and all the coaches’ interviews were stored
within the database child node coaches. I also set up a separate responses node
that contained the responses to each question in the interview3. All the respons-
es to question one are contained in the child of this node. The advantage of this
is that it allows the researcher to automatically search and more easily compare
responses across different groups. For example, one could easily compare male
athlete responses to male coach responses at question 1. One could not do this
as readily in the document system.
By setting up the data in this way, coding can be done within the index system
by searching and retrieving raw data that is stored at one of the branches of the
database node or the response node. The result of this coding is then stored as a
node in the index system as can happen with coding in the document system.
However working within the index system allows a more flexible approach to
working with the data.
Figure 11.4.2 Part of the database node in the gender and gender identity pro-
ject. Sex and status children are displayed. The numbers refer to the position
that the node and children of the node have in the research projects overall in-
dex tree.
The index system forces the researcher to think about relationships between
nodes or concepts and is often thought of as the thinking system. It manages
ideas and exploration of the ideas that emerge from coding. Exploration organ-
ises codes first and documents second. It stores and locates codes, categories,
higher order concepts, and associated ideas all within a node. It aids the re-
searcher by helping to structure the codes and ideas that have emerged from, or
been imposed, on the data into a hierarchical tree structure. The index system
takes a top down approach to data organisation which contrasts with the tradi-
tionalist bottom up approach to qualitative coding.
There are two types of nodes that can be used in NUD*IST 6. Free nodes con-
tain information pertaining to a specific subject (e.g., physical free node - con-
tains all references to physical attributes). The researcher may begin with free
nodes and then structure them into a hierarchical index tree. The index tree con-
tains information from numerous sub nodes in separate sections of the tree
(Figure 11.4.3 is an example of an index tree). The index tree allows for catego-
ries, themes, and higher order themes to be developed and organised, and thus
relationships to be represented. The index tree is a hierarchical coding structure
that helps the researcher develop thinking about relationships between catego-
ries. For example, in the research example one of the index tree nodes was Cat-
egorisation (see Figure 11.4.3) and contained three child nodes: Particularisa-
tion, Particularisation and Categorisation, and Categorisation.
Database Node 1
Sex 1 1
Age 1 2
Status 1 3
Education 1 4
1 1 1
1 1 2
1 3 1
1 3 2
Figure 11.4.3 Part of the index tree from the author’s PhD. General Self
branch including Categorisation sub branch and children is displayed (2 3 1 =
Particularisation, 2 3 2 = Categorisation and Particularisation, 2 3 3 = Categori-
sation). The numbers refer to the position that the node and children of the node
have in the research projects overall index tree.
All coding related to the differing categorisation strategies was contained at
the Categorisation child node, all coding related to differing particularisation
strategies was contained at the Particularisation child node, and so forth. Par-
ticularisation, Particularisation and Categorisation, and Categorisation are
all examples of the categorisation process (Billig, 1996) and thus were orga-
nized within the Categorisation node.
General Self
2 1
2 2
2 3
2 5
Database 1
2 3 1
2 3 2
2 3 3
Inductive and deductive types of analysis are also possible in the index sys-
tem. I used both an inductive coding scheme with the raw data and a deductive
coding scheme where the items from the PAQ (Spence & Helmreich, 1978)
were imposed upon the raw data.
As mentioned in the document system section, results of text searches within
the document system are held in the index system, in the text searches area.
These can become nodes in their own right within the index system. Conversely
searches of coding held at nodes are held in the index search area, and can also
become nodes, as with the text searches. The index system, therefore, allows for
greater types of searching to be conducted. Boolean searches such as collation
searches, contextual, negation, restriction, and tree-structured can be easily and
efficiently conducted (see Weitzman & Miles, 1995 for a discussion of each
type of search operator).
Hence the index system in NUD*IST 6 allows for the researcher to logically
and coherently structure his or her thinking so that theory building is enhanced.
Memoing allows for the researcher to track thinking associated with nodes and
child nodes such as how these where developed, defined, decided upon, and
other related information. The index system itself allows the researcher to
quickly and easily edit, modify, and build an index tree that represents emerg-
ing relationships amongst nodes. Further, the ability of the index system to use
searches as data sources allows for system closure. All of these can potentially
enhance theory development and building by organising data logically and co-
Therefore, both the document system and the index system have a logical and
coherent structure for organising, managing, and thinking about the data. Both
allow for flexibility within their structure. To illustrate this point further the
following section is a brief overview of how the PhD data and subsequent anal-
ysis was set up within NUD*IST 6.
11.5 An Example The Construction of Gender and Gender Identity
in Elite Australian Sport.
The given PhD explored gender and gender identity construction in elite Aus-
tralian sport using an inductive discursive psychology theoretical and methodo-
logical framework and a deductive content analysis approach. Of analytic inter-
est were how elite sportsmen and elite sportswomen did being male and female,
and what it meant to be male and female in elite Australian sport. Figure 11.4.1
shows the document system and index system as they appeared in NUD*IST 6.
In the index system there were 223 nodes in total. These nodes included induc-
tive and deductive coding as discussed above.
Thirty-seven elite Australian athletes and 38 elite Australian coaches were
interviewed using a semi-structured format. Each transcribed interview, that is
the raw data, was stored within the document system in NUD*IST 6. Thus Raw
Data 1 (see Figure 11.4.1) held the transcribed interviews from participant 1
through to participant 75. Each document also had attached to it a memo. I set
up a memo for each transcribed interview that held the field notes associated
with each interview.
As mentioned previously I like to work predominantly within the index sys-
tem. Thus in this case the database node was the first section of the research
project that was set up (see Figure 11.4.2). All basic demographic information
was organised as a child node4. All the participants’ responses to question 1
were put into the response node in the child node question 1. This allowed to
more easily search just responses to question 1 both inductively and deductive-
ly. All of these nodes were housed within the index system.
Analytic work was then carried out by browsing the question 1 child node in
the index system and applying codes to the data. Post-doctoral research has also
subjected the data to a modified version of a grounded theory. This will be used
to illustrate the workings of the coding and thinking as it is perhaps more famil-
iar to the reader than discursive psychology. In the grounded theory open codes
were allowed to emerge from the data. These codes became free nodes in the
first instance. After numerous passes through the data it was felt that some free
nodes appeared to be related to a common theme. Participants described them-
selves in reference to social or vocation roles etic free node, and also with
reference to informal or culturally specific roles emic free nodes. It was felt
that the underlying theme associated with these two free codes was reference to
a role. Hence the two free codes were re-organised under a newly created role
node in the grounded theory node section of the index tree. This cannot be seen
in Figure 1 due to the large number of nodes in the project. It was also felt that
some nodes were being discussed in particular relationships with other nodes. I
hypothesised that when participants described themselves in terms of attributes,
traits, behaviours they often used roles to exemplify these traits, attributes,
and/or behaviours. For example “I think I’m very understanding, like when I’m
with my kids I really listen to what they’re saying”. Hence a Boolean collation
search was conducted. Thus it can be seen how theory building can be aided by
the logical structure of NUD*IST 6 where free nodes can be moved into index
trees and sub divided into child nodes.
The above is but a brief overview of what was done within the PhD data. It is
hoped that the reader takes away from this small illustration is that NUD*IST 6
helped to organise, manage, and search this data. Notwithstanding I still had to
think, code, and form theoretical propositions about the data.
11.6 Limitations and Concerns: NUD*IST 6 and
CAQDAS Software in General
NUD*IST 6 is a logically structured program that coherently organises qualita-
tive data in a manner that facilitates the researchers’ thinking about the data. Its
logical structure helps the researcher make sense of the data. There are some
limitations with the program, which are primarily technical in nature, such as
the inability to switch between text units of a line to text units of a paragraph. In
addition, one of the most detracting features of the program is that it is not able
to graphically represent relationships between concepts. However QSR has de-
veloped a program that accounts for NUD*IST 6’s shortcomings, this program
is called NVivo. NVivo allows the researcher to work with text units of differ-
ing sizes such as individual letters to whole pages, to single words all within the
same node and project. Further, NVivo allows the researcher to graphically
represent his or her thinking through a multi-layer graphical modeller. This
modeller is linked to the data so that the researcher is easily able to explore re-
lationships in the data and add any items from the data directly to the model.
For further information about NVivo an overview can be found at the following
website One of the benefits that the most recent ver-
sion of NVivo has over NUD*IST 6 is that audio and video data can also be
loaded into NVivo and analysed within the same system. This is not possible
with NUD*IST 6. NVivo also allows for the linking of audio and video files to
the transcripts of this data, thus both textual and sound based data is able to be
analysed within the one software program.
Regardless of which type of CAQDAS that a researcher uses, there are some
broader limitations or concerns that need to be recognised. The raw data in the
first instance must be rigorously checked for errors and omissions (Drisko,
2004). It does not matter how ‘good’ a CAQDAS software program is, data that
is not accurate and complete will result in flawed analytical outcomes. This
checking process also allows the researcher to engage with the data at an in-
depth level that may not be possible otherwise. In this way the researcher is
exposed to nuances in the data that often leads to questions and new ways of
considering the data.
As mentioned previously CAQDAS does not replace the researcher. It is the
researcher who analyses the data not the software (Rennie, 2006). Any
CAQDAS outputs are reflective of the skills and knowledge of the researcher
using the program. Using CAQDAS does not automatically increase the rigour
of your findings, your assessment, judgement, coding, member checks, etc.
You, the researcher, are responsible for these. There is also some question sur-
rounding whether the use of CAQDAS actually results in any time saving bene-
fits for the researcher (Krahn & Putnam, 2003). However Rennie argues that
when used appropriately, CAQDAS has the potential to save the research time
that can be better used in thinking about the data.
Two other common concerns is the potential for the researcher to become
disengaged or distant from the data and the potential for CAQDAS to decontex-
tualise the data. Auerbach and Silverstein (2003) argue that by using CAQDAS
the researcher becomes distant from the text and as a result the total immersion
in the data that often leads to analytical breakthroughs is not possible. However
Drisko (2004) suggests caution over such criticism as currently there is no evi-
dence that a researcher’s analytical judgement is impaired or that immersion in
the data is not possible when using CAQDAS. There is the real potential for the
researcher to use identified words or phrases out of context (Agar, 1991). Any
search conducted by CAQDAS identifies all instances of the word and does not
consider the context that surrounds the word. However that is the job of the
search function, to search and identify. It is then up to the researcher to ensure
that what is identified is appropriate to the research purpose. After all,
CAQDAS is only a software program and should be treated as such.
1. See Weitzman & Miles (1995) for an overview of each grouping.
2. The author notes the current debate between significance levels and
effect size.
3. Note that this could have been included under the database node.
4. I could have subdivided the data differently, so that under sex the
breakdown was female athlete, female coach, male athlete, and male
coach. However this would not have enabled me to the compare ath-
letes to coaches’ responses as a whole or female to male responses as a
whole. The former division is more flexible in terms of greater search
Acknowledgments: The author would like to thank Steven Christensen for his
helpful comments on earlier drafts of this chapter.
Auerbach, C. F., & Silverstein, L. B. (2003).Qualitative data. New York: New
York University Press.
This chapter has aimed to provide the reader with a brief overview of
CAQDAS and expose the reader to one particular program NUD*IST 6.
What stands NUD*IST 6 apart from other programs and a manual ap-
proach to qualitative data analysis is the logical, coherent, and flexible na-
ture of the NUD*IST 6 framework.
In general, CAQDAS allows researchers to more easily manage, organ-
ise, and share their analytical insights. The last point in the previous sen-
tence is worth emphasising. Collaboration on qualitative research projects
is made easier through the use of CAQDAS. Leaving aside the difficulties
associated with working collaboratively with other people on any type of
data, CAQDAS allows for an electronic sharing of data, analytical insights,
and assessments that is immediate. Depending on the specific CAQDAS,
different researchers can be working on difference aspects of the same pro-
ject at different locations and these different workings can be merged back
into the original project. Thus working with qualitative data in a CAQDAS
program allows for splitting up of the project and then a re-merging of the
project. This splitting and merging is able to be electronically tracked so
who did what, when, and any analytical insights are better managed. How-
ever like all programs, be they qualitative or quantitative, CAQDAS will
only be as good as the researcher who uses them.
Bazeley, P. (2007). Qualitative data analysis with NVivo. London: SAGE.
Billig, M. (1996). Arguing and thinking (2nd ed.). Cambridge, UK: Cambridge
University Press.
Campbell, H., Gross, J, & Dodd, J. (2008). Our motivations are important too!
A qualitative analysis of masters squash players in Australia (Abstract). Aus-
tralian Journal of Psychology, 60 supplement, 125.
Chu, M. M. L., Leberman, S. I., Howe, B. L., & Bachor, D. G. (2003). The
Black Ferns: The experiences of New Zealand’s elite women rugby players.
Journal of Sport Behavior, 26, 109-120.
Drisko, J. (2004). Qualitative data analysis software: A user's appraisal. In D.
K. Padgett (Ed.), The qualitative research experience (pp. 193-214). Toronto,
Canada: Thomson Brooks/Cole.
Gibbs, G. (2002). Qualitative data analysis: Explorations in NVivo. Bucking-
ham, UK: Open University.
Hoeber, L. (2008). Gender equity for athletes: Multiple understandings of an
organizational value. Sex Roles, 58, 58-71.
Krahn, G., & Putnam, M. (2003). Qualitative methods in psychological re-
search. In M. C. Roberts & S. S. Ilardi (Eds.), Handbook of research methods in
clinical psychology (pp. 177-195). Malden, MA: Blackwell.
Lamont-Mills, A. (2001). The construction of gender identity in elite sport: A
mixed methods exploration. Unpublished doctoral dissertation, University of
Southern Queensland, Toowoomba, Queensland, Australia.
Lane, H. & Matheson, H. (2005). A qualitative study of sport and exercise par-
ticipants eating attitudes and eating behaviour (Abstract). Journal of Sports
Sciences, 23 176.
Lewins, A., & Silver, C. (2007). Using software in qualitative research: A step-
by-step guide. London: SAGE.
Rennie, D. L. (2006). The grounded theory method: Application of a variant of
its procedure of constant comparative analysis to psychotherapy research. In C.
T. Fischer (Ed.), Qualitative resaerch methods for psychologists (pp. 59-78).
Burlington, MA: Elsevier.
Spence, J. T., & Helmreich, R. L. (1978). Masculinity and femininity: Their
psychological dimensions, correlates, and antecedents. Austin: University of
Texas Press.
Weitzman, E., & Miles, M. B. (1995). Computer programs for qualitative data
analysis. Thousand Oaks, CA: Sage Publications.
Wright, T., Trudel, P., & Culver, D. (2007). Learning how to coach: The differ-
ent learning situations reported by youth ice hockey coaches. Physical Educa-
tion and Sport Pedagogy, 12, 127-144.
ResearchGate has not been able to resolve any citations for this publication.
Full-text available
This dissertation explored how gender identity is constructed in elite sport. I argued that sport is a unique socio-cultural context where gender category membership, may be enacted both the same and differently than in other contexts. Historically, most gender stereotyping, gender trait, and gender identity research in sport (e.g., Andre & Holland, 1995; Csizma, Wittig, & Schurr, 1988; Harris & Griffin, 1997) has employed researcher-generated constructions of masculinity and femininity, or non-sporting constructions of masculinity and femininity. By failing to define and construct gender from the participants' perspective, researchers have imposed their own preconceived cultural standards of gender upon participants (Doyle & Paludi, 1995). To generalise these preconceptions to other groups is to do so without consideration of cultural diversity and possible difference (Doyle & Paludi). Therefore, previous sport gender studies that have used these methodologies are tenuous as contemporary and future models upon which to base gender work. Further, gender identity research that has utilised a discursive psychological theoretical and methodological framework has produced findings that question the empirical validity of current models of gender in sport and exercise psychology (see Wetherell & Edley, 1999). These discursive results suggest that gender is a multifaceted, multidimensional, multifactorial, negotiated, dynamic, and variable concept (Wetherell & Edley, 1999). Therefore, two research questions were addressed by this dissertation: 1) How do participants perceive themselves in terms of gender-related characteristics?; and 2) How do elite sportswomen and sportsmen enact and negotiate membership of idiosyncratic, gender, and gender identity in sport categories? In order to address these research questions two self-report measures were utilised, the 24-item Personal Attributes Questionnaire (PAQ) (Spence & Helmreich, 1978) and a semi-structured interview concerning identity prescription. Thirty-eight elite level coaches (19 women, 19 men) and 37 elite level athletes (19 women, 18 men) voluntarily participated in this study. The interview data were analysed using two divergent theoretical and analytical frameworks, an a-priori content analysis (imposition of the PAQ items on interview responses) and a discursive psychological framework. The results of the PAQ analysis suggest that sportswomen and sportsmen perceive themselves differently in relation to gender-related characteristics. Differences which did not reach statistical significance, were found between male and female responses on the PAQ Masculine (M), Feminine (F), and Masculine-Feminine (M-F) sub-scales. Statistically significant differences were found with reference to PAQ classification, with women more likely to be classified as Androgynous and men as Masculine. There were no statistically significant occupational differences on either PAQ sub-scale responding or PAQ classification. The above results call into question the underlying assumptions and theoretical foundations of the PAQ. The a-priori content analysis also revealed a number of contradictory findings with reference to the assumptions and foundations of the PAQ scale. For example, women were more likely to utilise the PAQ M item Self-confident to describe themselves as gendered individuals than men. Whereas men were more likely to use the PAQ F item Gentle than women in the same identity category. Further, Feminine classified people were more likely to use the PAQ M-F item Very Dominant when describing themselves as women/men in elite sport. Therefore, the PAQ and a priori results cast doubt on the empirical utility of two factor models of gender to understand gender as a complex and dynamic construct. The results suggest that elite sport might be a context where gender is distinctively enacted and constituted. In order to determine how gender identity is enacted and negotiated in competitive sport, the interview data were analysed using a discursive psychological approach. Discursive psychology focuses upon how representations are constructed within, and constitutive of, the social practices that are found in language. In this respect, gender is conceptualised as being negotiated within the local interactive context where culture, history, and social contexts are reflected within discursive practices. In Research Question Two, interest centred on the interpretative repertoires and reflexive positions that participants used to prescribe themselves as idiosyncratic, gendered, and gendered individuals in sport. Interpretative repertoires are recurrent, culturally familiar global discursive patterns that individuals use to make sense of themselves in conversations (Wetherell, 1998; Wetherell & Potter, 1988). Reflexive positions are offered as an alternative discursive notion to the social psychological concept of role (Davies & Harré, 1990). A person is not considered as an individual free agent, but rather as the subject of the interaction, where the individual takes up or is placed in various subject positions depending upon the discourse and the particular social context in which the individual interacts. Thus we make sense of ourselves, or position ourselves, within social interactions through the cultural and personal resources (interpretative repertoires) that are made available to us in our discourse. Overall, the results of the discursive analysis suggest that participants enacted something gender scholars would call Masculinity, Femininity, and Androgyny when prescribing themselves across the three identity categories. That is, participants used gendered, culturally familiar discursive patterns (interpretative repertoires) to make sense of themselves across identity categories. However, participants were also able to draw upon non gender-related discourses during this process. Thus, identity work was characterised by variability, inconsistency, and contradiction. Different interpretative repertoires and reflexive positions were used by participants both within and across identity categories. Therefore, the use of gender-related interpretative repertoires differed according to the identity that was being scripted up. Thus participants were able to be Masculine, Androgynous, and Feminine, and position themselves differently depending upon the identity that was being prescribed and the local interaction context. That is, participants used interpretative repertoires to talk one way, but walk another (e.g., Androgynous interpretative repertoire, Hegemonic Masculine reflexive position) that was specific to the social, historical, and cultural context, and the local interactional context. The above results call into question Spence and Helmreich’s (1978) postulation that there is one Masculine and one Feminine identity. Indeed the results are suggestive of many Masculinities and many Femininities. Participants also deployed specific discursive strategies that incorporated the action and epistemological orientation of their talk when constituting their identities. That is, they worked to increase the facticity of their talk and worked to align themselves with certain positions (e.g., Hegemonic Masculine man) and not others (Feminine man) through their discourse. Thus gendered talk carried with it gendered ideological practices that participants used to reproduce, reinforce, and challenge the current gender order. The above results, combined with the disparity between the PAQ results and the a-priori content analysis, suggest that earlier and current models of gender that conceptualise gender as a multifaceted, multidimensional, bi-directional but static concept are probably not representative of how people do gender in everyday talk. The results support extant theory that gender identities might exist rather than a single gender identity. Overall, the results of this dissertation suggest that elite sportswomen and sportsmen enact and negotiate membership of identity categories that is specific to the local interactional context, as well as the cultural, social (i.e., sport), and historical context. I infer, therefore, that current static gender models in sport and exercise psychology may not fully capture the complexity of gender in everyday talk and that alternative ways of understanding gender in sport are needed.
Full-text available
This paper provides an introduction to computer programs for qualitative data analysis [QDA]. Four software packages are examined in detail: ATLAS/ti, The Ethnograph, HyperRESEARCH and NUD·IST. These packages are applicable to a variety of qualitative research approaches used in the human service fields. The merits of QDA software are described, as are some hazards. The use of QDA software described and the strengths and limitations of each software package are identified and discussed. Issues relevant to deciding among them are detailed.
Full-text available
Background: Large-scale coach education programs have been developed in many countries around the world to help prepare coaches for their important role. Coaches have said that they also learn to coach from experience, starting from when they were young athletes until their current coaching positions. Finally, in the last decade, Internet resources have begun to be promoted as valuable tools for learning. Most of the studies on coaches' development have focused on only one of these three ways of learning how to coach. Purpose: To explore the different learning situations in which youth ice hockey coaches learn to coach. Participants: 35 volunteer youth ice hockey coaches from five minor hockey associations in the province of Ontario, Canada. Data collection: Coaches were interviewed individually using a semi-structured interview guide. The questions asked to coaches were about their learning through formal large-scale coach education programs, their learning experiences outside of these programs starting when they were young athletes until their actual head coaching positions, and their use of the Internet. Data analysis: The first part of the interview consisted of specific questions regarding the number of years coaches had played and/or coached hockey and their level of coaching certification. The answers to these questions were entered directly on a form and entered later into Microsoft Excel to perform descriptive statistical tests. The second part of the interview involved more in-depth questions regarding what learning opportunities contributed to their development as a coach. Finally, questions were asked regarding how they use the Internet in their coaching. The content of this second part of the interview was transcribed verbatim into Microsoft Word rich text format for further data analysis using Nvivo software (Qualitative Solution Research, 2002). Findings: The results revealed seven different learning situations including (a) large-scale coach education programs, (b) coaching clinics/seminars, (c) formal mentoring, (d) books/videotapes, (e) personal experiences related to sport, family, and work, (f) face-to-face interactions with other coaches, and (g) the Internet. Conclusion: Considering that coaches learn to coach through many learning situations, it would be inappropriate to discriminate against any of these situations, since each situation seems to have a unique role in the development of a coach. Therefore, it may be concluded that coach education should include a combination of all seven learning situations, instead of focusing on one. Future research should concentrate on investigating the complementary potential of these situations and what can be done to make each of these situations more appealing to coaches.
In illustrating examples of software use in different contexts through three distinct case-study examples we hope to paint a picture of some common aspects of analysis in the context of software tools so that you can draw out ideas about what might be useful in your own particular research. We understand that your choice of software may be limited within the constraints of local provision, but our purpose is to enable ambitious yet secure use of any CAQDAS package and the moulding of its functions to your needs, while also adding to your awareness of what other tools work well for particular contexts. We believe that a broad understanding of software packages other than the one you happen to be using will open up your thinking about your own work. Above all, we see ourselves as ‘facilitators’ rather than ‘instructors’. The way we teach is informed by the belief that you are the expert about your project and your needs. We can show you tools, illustrate their benefits and caution against their potential limitations. We can make suggestions about their suitability (or not) for different approaches to data analysis. But you need to decide whether to use software at all – and if so, then which package. If you decide not to use software then you need to be able to justify this. If you decide to use software, you need to design a strategy for doing so within the parameters of your broader methodological context, specific analytic needs and any practical constraints within which you are working. We hope this book will provide you with the context you need to frame your thinking about software, to give you insights into the way particular tools might be useful at various moments, and to heighten your reflection about the relationship between technology and methodology. More than anything else, we hope this book will inspire you to explore your data to greater depths, to experiment with software tools and to develop systematic and creative ways of conducting robust and well-evidenced analysis.
This collection of 14 original articles teaches readers how to conduct qualitative research. Instead of characterizing and justifying certain methods, the contributors show by means of actual research studies what assumptions, procedures, and dilemmas they encountered. Fischer's introduction, which emphasizes the practical nature of qualitative research and the closing chapter, which uses a question-and-answer format to investigate, among other subjects, what is scientific about qualitative research, are complemented by a glossary and other features that increase the book's utility and value.
Dr. Rennie has conducted a large study of 14 psychotherapy clients' moment-by-moment experiences of the entire hour of a therapy session. His participants were engaged in therapy at the time of the study. He interviewed them about their recall of moments in a therapy session from which they had just emerged, with the recollections having been stimulated by a replay of a tape of the session. He was thus given access to their inner worlds of what it was like to be a client. In this chapter he focuses on the grounded theory method (GTM) and his modifications of it. This method was originated by sociologists Glaser and Strauss and is among the first rigorous qualitative research methods to be widely practiced. The name refers to the grounding of research in life-world data and to the development of a theory of the subject matter from analysis of those data. In the traditional method, the theory arises from developing codes for the data through a "constant comparative" procedure in which each line of data is compared with identified (coded) content, and either receives an earlier code or is assigned a new one. Categories are similarly developed from the codes, and then a core category is named to hold all the others. Rennie describes the constant comparative procedure originally suggested by Glaser and Strauss, and contrasts it with a variation that he and his students developed. This variation involves breaking transcripts into units of meaning and interpreting the meanings of each meaning unit. Categorizing is done immediately, from one meaning unit to the next, rather than through the intervening step of developing codes and then categorizing them. He illustrates the variation with a case drawn from his study and gives an overview of the returns from his study as a whole. The chapter closes with a constructive view of the implications, for those wishing to adopt the method at the current time, of a rift that developed between Glaser and Strauss on how to do constant comparative analysis. (PsycINFO Database Record (c) 2012 APA, all rights reserved)
Although gender equity for athletes is a frequently researched topic, it is often assumed that understandings of gender equity are unitary and shared, which may complicate the implementation of it. The purpose of this study was to understand and critique the meanings of gender equity for athletes through in-depth interviews with 5 administrators, 6 coaches, and 17 athletes at 1 Canadian university athletic department. These data were coded and categorized using Atlas.ti. The findings revealed multiple but narrow meanings of gender equity: (a) equality, (b) conditional equality, and (c) a women’s only issue. None of these challenged the taken for granted assumptions associated with university athletics; however, they illustrated the complexities and struggles involved in understanding this organizational value.