It's time form me to stop being a furtive lurker and to become a
contributor. I do so because there seems to have been a fair bit of
discussion, naturally enough, about the relative virtues of and
differences between, QSR NUD*IST and NVivo.
As it happens, I've been a user of NUD*IST since version 2.1, and
recently, when the Association for Qualitative Research Conference offered
a pre-conference workshop that sought to demonstrate NVivo, I went along.
Lyn Richards was the demonstator. In my view, Lyn has been an utterly
indespensible player in the development of both the NUD*IST and the NVivo
packages. Indeed, I'd even go so far as to venture that Lyn has been a
leading light in developing the argument that qualitative software,
appropriately used, can and should be used to augment fundamental
qualitative research processes irrespective of the package being used.
She reasons, correctly in my view, that computer aided analysis only
assists the qualitative research process if researchers have a sound
understanding of the theory underlying qualitative inquiry for these, and
not the application are what should inform the qualitative researcher's
practices.
But let me not digress. As it happens, I took quite extensive notes at
that workshop/demonstration and I've reproduced them herewith in the hope
that they may be useful. Bear in mind when reading these notes that they
were derived from a dmonstration by the company that developed both
programmes, but bear in mind also that the notes were made by somebody who
has used NUD*IST quite extensively and who was seeing the NVivo programme
for the first time. I'm bound to say that I took a demo copy of the
application with me and played with it (until around 3.30 am.) a few days
later and I found it very easy to use but even now, some six to eight
weeks later, I'm still learning. So here then, are my original notes in
unexpergated form:
****************************
AN INTRODUCTION TO NVIVO
Jens J. Hansen, Melbourne, 7 July, 1999
INTRODUCTORY COMMENTS;
Package is brand new (six weeks in public arena at time of writing this,
announced in February and launched in May) and comes with two books, a
manual and methods resource book.
Tutorial System is different to the N4.
Both are for self-teaching. In short, it is highly documented and helps
the process of self-teaching.
Differences between N4 and NVivo
N4 will:
1. Run on old machines, NVivo won't; (Pity because I hate being encouraged
to purchase new stuff when the old stuff still works fine.)
2. Won't run on Macintoshes but OK on Power PC's; (Probably will be slow
if using emulation systems.)
3. Merge projects but NVivo won't; (This has implications for large
projects where multiple resarchers are engaged in analyses.)
4. N4 enables command files whereas NVivo automates it. (That will tend, I
suspect, to make NVivo more user friendly to the average scholar.)
Incidental comment from Lyn: NB: Look at CAQDAS URL for info on
Qualitative Research. Especially have a look at the links.
Best info on qualitative research currently, in her view, tends to emerge
from the Health field.
Historical Background
First (Ethnograph) developed in 1985. Thereafter, code and retrieve
framework. After 1990's need to link questions and data was recognised
and iterative nature of the work, i.e. coding on.
DEMONSTARTION (Jens' View and comments)
There is no doubt that this is much quicker than the N4 programme insofar
as the coding/noding process is concerned. With NVivo, this is done in
rich text format and that means that individual letters of elements can,
if desired, be coded and edited. It also means that styles and colours
can be introduced and these in turn can become markers between sections if
necessary. (Perhaps one could, for example, use traffic light colours to
indicate the potential strength of data????)
Moreover, links can be shown pictorially and the use can access the data
(text and/or nodes) from the pictorial representation. (The likeness to
Inspiration is quite remarkable. Will have to explore if Inspiration
mind-maps can be imported.)
Free Nodes and Tree structure is retained as per N4, and as well,
attributes, i.e., things that are known about the data, can be retrieved
in matrix forms (just as per excel). Importation and exporting can occur.
(Including Inspiration generated mind-maps????)
A new feature is sets, i.e. what Jens always did as Data management nodes.
However, this application allows drag and drop and this will tidy up the
management of data.
Project pad is a strong feature as it speeds up access to segments of the
programme.
The rich text characteristic is noteworthy because:
a) it enables editing and addition,
b) it allows direct importing from ms MS Word.
Projects that have been processed in N4, can be shifted directly and
seamlessly into NVivo. Can't go the other way though.
Can then continue with data as per N4, or can develop it as NVivo data
enrichment or can export and then examine attributes. To take it from N4
must use Merge application.
Can generate live matrixes (as per the form frequently advocated by
Weitzman and by Miles & Huberman, for instance.)
LYN'S HIGFHLIGHTING OF KEY FEATURES OF THE APPLICATION:
a) capacity to have rich text;
b) capacity, therefore, to be able to colour code;
c) ability to code as you transcribe and nodes are colour striped in
separate margin.
The tool bar at the foot of the window is useful and if coding option is
selected, the code shows on the menu bar as bold.
If details about a node are incomplete, type in what is known and then hit
Tab -- package will then find the node.
Can also see what goes on by looking at selection option re Scope of
Coding Menu.
Tutorials included with package
One is on Road Rage, One on Disability Needs, one on Bush Schooling
(analyses the use of email to connect remote women in Oz); American
project also on violence prevention.
PROCESSING TEXT:
Can either:
a) Imbedding Data Bites - i.e. pictures, sound; the software selects the
links which have been previously linked to either sound or picture. This
means that the hot link becomes a compound of the text and the data-bite
become three dimensionally rich. This means that focus groups can have
transcript plus annotations plus video clip. The move, therefore, is a
move away from mono media data to multi media data. You can also add
annotations (which N4 can also do).
b) Link can also occur to the transcript directly which is the same
feature as browse. This is how memos are written and they are full on
documents, i.e. a document in it's own right. This means that webs of
data can be generated - that is - more than one document will become
linked at that point.
c) Node links are also possible and this enables you to link to data that
have been already coded and assembled.
SOME MORE OF JENS' THOUGHTS;
a) This has potential if used with EndNote;
b) The overall impression that emerges here is that NVivo linkages enable
a more potent range of retrievals and linkages to the exact data. A new
node (hyperlink) can be generated with specific text and or sentence and
or references.
c) Can also go to specific segment and then generate a spread to see the
embedded context.
d) Key point is that RETREIVAL becomes facilitated.
Each window in NVivo is independent and has own icons. Still has either
pathway (icons or pull down screens).
No Undo feature but can (like N4) save as which gives versioning option.
NVivo does not have an automatic save but doesn't overwrite the previously
saved version.
Attributes can be imposed just as characteristics can emerge from data.
But three attribute options are presented as standard and they are:
Unassigned,
Unknown,
Not Applicable.
In the end, it is the matrix that is very useful here. That is, a set of
attributes can be exported into SPSS or Access and can also import from
there into NVivo matrix.
Sets of Data can also be created which makes data management rather
easier. Can, therefore, take an attribute and filter it (that is the
option to choose) so that the documents with a particular attribute
emerge.
Search and scoping, therefore, are concurrent (as per N4 via
restrictions).
NVivo packages searches together so that it looks up either Nodes, or
Attributes or Text search, or Boolean or Proximity. These latter two
search options are the equivalent of the seventeen N4 search options.
So enables features to be searched quite specifically using programmed
criteria (as per Statview or SPSS) so that these are selected one at a
time and then run as a search equation or as a search operation.
Note the assay feature which is to have a quick look at what the data say.
When a qualitative matrix is generated, can select option to have a look
to see what has been coded and these occurrences are shown as quartiles.
Can then inspect the data attributed to a cell of intersection.
(Something similar can be don in N4 if look at statistics of finds but not
really as fancy.)
The matrix can be exported and Excel can process but remember that this is
a qualitative matrix. Can then proceed to the model mode and show
visually.
Documents don't need to be introduced in the same fashion as N4. With
NVivo, can import directly as RTF and then go into programmes.
Fini.
*******************
So there you have it folks. I hope that these notes are of some use. I
want to reiterate that the comments above are really akin to a set of
intitial impressions. I also want to reinforce the point that I have only
played with NVivo for a few weeks so that means that I have still to get
my head around this new application completely and certainly, I don't have
any clear idea about which of the two I think that I'm going to prefer to
use. In fact, what I'm finding is that I have, at once, a degree of
reluctance to move away from the familiar, as well as a degree of
excitment that here is a new tool that might enable me to generate a fresh
level of research rigour, a rigour that will enable me to retain many of
the techniques that were completed manually in the past and that were
underpinned by sound methodlogical understandings. My advice to anyone
that is looking to conduct qualitative research via computing, is to
explore as many programmes as possible and these are but two.
cheers,
Jens
___________________________________
Jens J. Hansen, Ph.D.(New England)
Programme Leader, Master of Educational Management,
UNITEC Institute of Technology, Te Kura Puukenga o Wairaka,
Private Bag 92025, Auckland, New Zealand
UNITEC Phone: 64 9 815 4321 Ext. 8797 UNITEC Fax: 64 9 815 4310
UNITEC email: [log in to unmask]
91 Domain Cresent, Muriwai Beach, RD 1 Waimauku, West Auckland, New Zealand
Home Phone: 64 9 411 7703
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|