Looks like it’s conference/workshop season for me. Last week was Educamp, today is a workshop on using NVIVO 8 qualitative data analysis software (and Friday I have a workshop on “Problem-Oriented Learning”).
So today’s blog posting is going to be my notes from the workshop… which will probably be interesting to about 0.01% of my readers (possibly fewer). I’ve taken a lot of notes, so I’m putting them after the jump so you don’t have to see the random thoughts and notes I wrote down during the workshop (unless you are part of that 0.01% – if you are, feel free to check them out – but I warned your of their boringness!)
- two programs from Microsoft download along NIVO 8 (they will update periodically)
- SQL server runs the database for NVIVO
- also .net framework
- Crystal reports – another third party piece of software (suggestion: .pdf any reports you want to print)
- learning NVIVO 8 is an iterative product – learn as you go – if you try to learn it all first, it would take you a year before you could start your work
Software works well if you have:
- multiple sources (vs. just text only)
- large data sets (vs. doing a Masters with just 10 participants, no plan to do more qualitative research)
- you are unsure what the data are saying (software really helps by allowing you to do many things to do with the data to figure out)
- need to classify, sort and arrange information as well as looking at trends, relationships and concepts)
- working on one project as a team
The software is difficult to learn, so you may not want to commit to learning the software (goes for all the qual data analysis programs – and they are all very different – so learning one will not help you learn the others). If you aren’t committed to one software package, you may want to check out some books/sources that list the strengths/limitations of each software package with respect to your particular project(s).
Tip: Save often. When NVIVO 8 crashes, you will lose your work (there is no recovery process).
Pros of NIVO 8:
- Coding systems, the node structure is easiest to use and most useful function
- organizing data: multiple ways to keep track of sources
- viewing relationships
- modeling (powerful, but difficult to use)
- query function (also quite powerful)
Qual data analysis is messy – you go through the process and nothing makes sense; then it gets worse! Have to learn to live with uncertainty for a while.
Limitations to NVIVO 8:
- importing files limited to .rtf or .txt and .docx
- web material imports only through a .doc save
- pdf available, but formats awkwardly
- third party add-ons may be troublesome for installation (instability, may need to reinstall occasionally)
- program looks & feels like MS Outlook, but doesn’t act like it!
- considerable loss of formating, especially paragraph sections and fonts
- large volumes of text do not cute & paste properly (you will need to subdivide at ~40 pgs)
- graphs from website are troublesome
- instability! (reboot!) more stable now than it used to be, but still crashes
- resource intensive, only the fastest processor and large amounts of memory will assure ease of use (a bit slow on older computers)
- you may need to quit other programs to free up memory, limits your ability to multitask (since Office is also resource intensive, you often can’t
- only 5 levels of undo, but no redo (and undo will often undo a lot of stuff in one go)
- saving needs to be done carefully (no auto-save, prompts only every 15 mins)
But, if you can take the time to learn it, it will allow you to do things that you can’t do without the software.
Originally was designed just for qual researchers, but now it’s also for marketers too (they pay more, so it’s more geared to them now)
- QSRinternational.com – lots of tutorials, a forum, etc.
- Lyn Richards (the developer) website with tutorials lynrichards.org
- books limited for version 8
- “Getting started in NVIVO” .pdf from QSRinternational.com
- “Volunteering” module – a supplied project to play with
- QSR forum is a good place to get help
- QSR tech help available too
- Internals: stuff you use only in NVIVO (e.g., transcripts of interviews)
- Externals: stuff that requires an external software program to open (e.g. sound & video files)
- includes both the external file itself as well as a place for you to put a description of what it is
- e.g., he gave an example of a Mind Manager map that he made
- if you transport your NVIVO file to another computer you won’t be able to open the externals if the new computer doesn’t have the necessary software
- Memos: your notes
Your entire project is stored in one file (helps you to transport from one place to another, but be sure to back up often in multiple locations!
Qual Research Process:
research question –> data –> coding –> interpreting(concepts) –> theory
- interpreting is key! the research doesn’t end at coding
- you need to be writing about it the codes you are finding to figure out what it means
- memos are the place where you write down developing interpretations –> working towards concepts
- memos are from grounded theory (NVIVO has a bias towards grounded theory; also works well with content analysis)
- NVIVO only allows you to attach memos to one other thing in your NVIVO file
- can have folders for different types of memos (e.g., code memos [describing each code], memos on each interview, etc.)
- codes are descriptions of what’s going on; concepts are a higher order
- e.g., codes of “weight loss” and “stress relief” may go into a category of “benefits of exercise”
- should use plain language for codes
- NVIVO may make you think that something that happens often (lots of instances of codes) are important
- a “node” is a container: for your codes… or also dates (everything that happened on a date) or a book (all the things you wrote about a certain book) or a location (everything that happened in a given location), etc. – this makes me think of tags in social media
- free nodes: nodes are first made as free nodes, but as you go through the process, you’ll start to create tree nodes (where you start to link up the free nodes in a tree structure)
- when you create “tree nodes,” the top level categories should have no individual things in those categories (just placeholders or folders in which you place your codes) – e.g., say you have a tree node of “assumption” and you have a datum that does fit into a specific code, you can create a “assumptions not otherwise specified” code to put them into
- case node: a case is “anything that has attributes”
- e.g., case = person, attribute = gender, age, socioeconomic status, ethnicity
- this allows you to view your information specifically for an attribute (e.g., personal goals by age; benefits of exercise by gender) – this is very powerful! much easier to do with software than to try to do this by hand
- case may also be a location or something else – it depends on your work
- need to thing about what a case is before you start (easier to create a case at the beginning than to try to do it later)
- if you interview individual and in focus group, you can take everything that person 1 said both their individual interview and in the focus group and put it all into their case
- should also keep a project log (what happened when) – just create a “source” file as your log
- Relationships: how two things are related (e.g., Anna lives with Sunil; Bernadette is friends with Ken); can also relate Nodes to each other – some people never use the relationships feature
- Sets: allow you to temporarily move things around without changing the underlying data
- e.g., keeping track of transcripts that haven’t been coded yet; coding by multiple researchers that need to be compared; stage of researcher a given participant is (e.g., made contact –> interviewed –> transcribed; lost to follow up group)
- need to keep track of order that you do things (when analysis of a given transcript happened, when codes are created – because the process is iterative) – don’t rely on software to do this for you (because it won’t do it thoroughly)
- Queries: text search (like a “find”), word frequency (important if language is important to you), coding (allows you to look for codes), matrix coding (to do with attributes; complicated), compound (very, very complicated), coding comparison,
- can limit query to just interviews, or just focus groups or both or etc.
- Models: complicated to learn to use in NVIVO!
- the model function will bring in all of your nodes and then you play with them – you can colour similar things, start to see what goes together, move things around – this is how he figured out tree nodes (looking at all the free nodes and tried to determine what went together)
- as your nodes change, your model will change (won’t add new nodes (you have to do that yourself), but will, for example, put an “x” when you turn a free node to a tree node, if you’ve imported them as free nodes)
- static model: to capture a stage in the process (so in this case, does not change as codes change)
- Window –> Undock all – this makes all the pieces into their own windows so you can make things bigger
- Classifications: where you put attributes of your cases
Tip: “New” button in upper left – unlike other Microsoft programs – allows you to do “new” things in whatever section you are in.
Note: Can always drill right down to the original data from any point in the analysis.
One workshop participant informed the group that CIHR is now taking feedback on their updated ethics statement, including a specific section on qual research.
Speakers pet peeve: methods sections that say “we used NVIVO to analyze our data” as if “NVIVO” is research method. It’s not. A method is “grounded theory” or “content analysis” or etc. That would be like saying I used “Statistica” or “SPSS” to analyze my quantitative data (without saying I used an ANOVA or a MANCOVA or etc.)
- you aren’t going to get it right the first time through
- e.g., he read through the first few transcripts and made notes in the margins (using “add comments” feature in MS Word – kept these “annotated documents’ as External Files so that they are embedded in NVIVO file); helps you to get a sense of what’s going on!
- then read again and added codes
Process for Coding in NVIVO 8: (many possible ways. find one you like)
- open a transcript
- open your tree nodes (or free nodes if you don’t have tree nodes yet) – allows you to see available nodes/codes as you code (if you have a lot, you may want to print the list out – right click over the list of tree nodes and you will get a print option)
- view –> detail view right (to put tree nodes on the side
- view –> turn off navigation view (to give yourself more space)
- highlight the words in the text that you wish code – right click –> code selection – existing, new or current (also keyboard shortcuts – e.g., CTRL + F3 = code at new node)
- new nodes: give the node a name and a description (description allows you to specify what is included in, or excluded from, a given node)
- icon for “View Coding Stripes” will show you highlights along the right side of the screen to show you what has been coded with what. Note you cannot edit while coding stripes are being described. You can add more codes, but not edit the text.
- “View Coding Stripes” –> selected items will allow you to select which codes you want to see
- hover over the coding density bar – will show you all the codes at a given spot
- left click on the coloured stripe in the coding section on the right side of the screen and it will highlight the specific text that you highlighted with a given code
- right click on the coloured stripe –> “open node” will open everything that’s been included in that node (helps you to be consistent)
- using an existing node (instead of a new node) will bring up a list for you to chose from (you can choose more than one if you want to code on piece of text at multiple nodes); you can also drag and drop the text to the node list on the left
- can also use the coding tool bar (select from the drop down or type in a new node)
- “code in vivo” will use the exact text you select as the code name
- Coding a Current Node – be careful using it, it codes at the last node you used, so need to make sure you know which node it is using!
Starting a New Project:
- file –> new project – give it a title & descriptioin
- project –> new internal (e.g., document)
- format –> insert –> date & time (or CTRL + Shift + T) (do this everything! (memos, project logs, etc.)) important for your audit trail )
- file –> properties (to change some stuff)
- more commonly, you’ll import something
- project –> import internals – allows you to import a file, will give you an option to include a description
- you can highlight text, right click –> links (a memo link, or a “see also link” – which links from one thing to another thing… it’s just a hyperlink (to either a whole document or just a piece of document)… e.g., if someone says something that reminds you of something that someone else said, you may want to link them to each other)
- right click –> links –> new annotation (like creating a “comment” or a footnote”) – e.g., when you want to keep track of something, but it isn’t exactly a code (e.g., if someone used jargon, you could explain what it means). Annotations stay highlighted in blue, can see them by going to the icon at the top (or use View menu) to see what the annotation says (it’s kind of like a post-it note)
- can also hyperlink text to a URL (e.g., if they talk about a website in their interview, you can link to that website)
Turning Free Nodes into Tree Nodes
- go to the Node section
- cut the free node, go to the heading you want to put it into in the tree node and paste (can even make a subcategory of the other nodes)
- to create a top-level node, make sure you are in the white space at the bottom (not on any of the existing nodes)
Places To Get Help:
- the help file (of course)
- e.g., Coding –> Working With Your Data – written by other researchers, helps you learn how to do things in NVIVO
- NVIVO tutorials – worth going through (not that long; about 20 minutes to go through all four)
- go to main page, then File –>Copy Project
- find it by Browse (usually goes to “My Documents”, then pick a place to back up (and give it a different name, like “Copy of Beth’s First Project”)
- be VERY CAREFUL! it’s very easy to overwrite your file! best to have a slightly different name for your back up than the main file
Remember, qualitative research is fluid – play around with things, e.g. create codes, but can change, etc. But also remember to keep track of all the things you are doing!