Jump to content

Most Liked Content

#12625 NVivo and Referencing - some possible strategies to add to your toolkit of tr...

Posted by Dr Jens Hansen on 21 March 2012 - 02:05 PM

The ability to access references and portions of references as well as notes is fundamental to the academic and research enterprise.  But the current process does not quite achieve that although it’s moving towards something that is very satisfactory.

Currently, when working with references and NVivo, we‘re able follow a routine that’s, as far as we can tell, something akin to this:
1.    We select needed citations from, e.g. EndNote, after having searched our library database for relevant materials.  That may mean invoking Boolean operators to find the appropriate citations from the academic library.  Typically, for most academics, such a library will include hundreds, or sometimes thousands, of entries.
2.    We export the selected citations from EndNote as XML files to a holding destination.
3.    We then import the XML files from that holding destination into NVivo by clicking on the ‘Bibliographic data’ option in External Data icon that’s a feature of the NVivo ribbon.
4.    From the pop-up table, we select from the available options, ensuring that requisite fields are imported into the resultant memos which NVivo creates.  The default for naming sources is ‘Title’ but the drop-down menu also allows for ‘Author and Year’.  It’s also possible from the ‘Assign sources to’ field, to generate either a series of single references, or to assign them to different classifications.
5.    We can then open the destination file/s and click on the link/s to the memo/s. created by NVivo.
6.    We should, though, make sure that the selected citation is the correct one [there can be occasions, as we’ve discovered, when more than one citation stems from the same author in the same year].
7.    Finally, we can code and node away to our brain’s content using both our references and primary data as required.
What we’d like to see built into the above is that the default option for step four has different consequences.  There are three outcomes we’d like to propose for inclusion:
1.    When materials from EndNote are migrated into NVivo, we’d like to see the full citation become presented as the default option for the ‘Name Sources’ field within the ‘Import from…’ table.  (Currently, as we noted above, we’re able to select either ‘Title’ or ‘Author and year’ and currently, ‘Title’ is the default.)  To our way of thinking, that does not align with the way in which academics tend to work.  We believe they’d prefer to work with the full citation as a default (and they’d like to work with more besides…).
2.    We’d like to see a capacity, therefore, of being able to import fuller details as directly as possible into NVivo.  This means we’d like to see an option in NVivo for being able to import selected EndNote materials directly into Sources.  We see this as being in the form of one or more documents (for Internals) or as inclusions within an external file that have been extracted from a linked file (NVivo to the EndNote library).
3.    We’d also like to open imported data directly when it is double clicked rather than having to click on the linked memo to access embedded details.  To our way of thinking, going directly to a source is especially important for beginning users of NVivo.

In justifying these three suggestions, we reason that most academics prefer to work with a full citation rather than just a title or an author’s name.  But in addition, our own experience tells us that academics want to be able to access a range of selected fields, e.g. key words/abstracts/notes and so on.  Quite simply, they want to be able to access, manage, search, and annotate material from fields they’ve selected whilst they are interrogating their primary data.  They invariably want to cross-reference their primary data with reference (literature) data.  Otherwise, what’s the point of introducing citations and associated fields?

Under the current set-up for NVivo, that can pretty much be achieved when memos are programmed to assume and display selected attributes.  But as noted above, full citation details cannot yet currently be displayed as a default for the ‘Name Sources’ option in the pop-up table.  Moving on…
Because we think the inbuilt processes of NVivo don’t quite achieve what we’d like to see, we’ve been using some work-around processes.  We believe these are quite useful and a tad quicker to use but in sharing them with forum members, we stress that they’re useful for us – that is – they fit with the ways in which we work.  Below, we’ve outlined the essentials of our work-around strategies:
  • If we’re working with a single reference:
    • we go to EndNote X4, click on the Hide Empty Fields button (unique to X4 and found in the top right hand corner);
    • select ‘Show All Fields’ from the ‘Output Styles’ option which resides within the ‘Edit’ drop-down menu (An alternative is to use the ‘Open Style Manager’ option within ‘Output Styles’ so that tailor-made options can be generated);
    • we then invoke the ‘Preview’ tab in the ‘Tab Pane’ and copy the contents directly into a ‘new document’ that has been created within NVivo or into a word document for importing as a standalone source.
  • If we’re working with multiple references (which might have emerged from Boolean searches):
    • we select citations and whatever fields we’ve opted to invoke after having used ‘Style Manager’;
    • we then use the ‘Format Bibliography’ button so that reference data we’ve selected will be produced as a separate document;
    • and at that point, we can either copy and paste the whole lot into a new document we’ve created within NVivo, or we can import the word document as a standalone source.  And, by the way, that procedure also works if you’re importing an important single reference and this can also be achieved with earlier versions of EndNote and with earlier versions of NVivo.

An alternative work-around we’ve devised involves creating External sources.  Here, our goal has been to generate, as simply and as rapidly as possible, an easily accessible repository for reference data.  We achieve this by following these steps:

  • we create a new External File and, when the External Properties box shows, we complete the fields so that:
  • the ‘Type’ becomes a ‘File link’ that links to the EndNote library (and that’s what creates speedy access);
  • the contents and units become references (because that’s the unit we’ve created) and the range is however large we, as researchers determine it should be;
  • relevant reference information is then pasted into each external unit as desired.
The advantages of this are twofold: first, we’ve found that we’ve been able to very quickly go to our source library, invoke Boolean searches (or go directly to known references), so that we can select what’s wanted.  We’re then able to copy and paste references into the external document divisions or units (which we’ve called references).  We could, if we want to, also apply attributes from the Attributes Tab but in practice, we’ve not really found that to be useful for the approach we’re using.  We want to underscore that this is really is a very simple work-around that enables us to assemble reference data for coding into a convenient place within NVivo.  It’s not perfect but it works for us and it pretty much enables us to show full reference material with which we can work robustly.

We apologise to those who already know about such strategies because the last thing we want to do is teach anybody’s grandparents how to suck the proverbial egg!  Indeed, we recognise that it’s quite likely that others amongst the NVivo community may throw their hands up in horror and will have found much simpler solutions.  But we have to say that in all honesty, we’ve found these relatively simple approaches have enabled us to quickly assemble what we need by way of reference materials.  Note that we are stressing, yet again, that they are useful for our approaches to undertaking qualitative research.

Our hope is that academics and researchers will not be bound by the determinism of the technology, and if they are, we hope they’ll recognise that and speak up so that better ways can be collaboratively developed.  We know from conversations with QSR staff that those designing NVivo features already seek to mirror the ways in which academics and researchers actually do their work.  We also know that QSR has moved a long, long way towards creating expanded synergies and capabilities within the software.  Congratulations are, therefore, in order.  But there is still a little way to go.  Listening to the whispers and voices of academic and research communities of practice is now necessary and that’s what this forum enables.  What do you think?

Dr Jens J. Hansen,
Anna Jo Perry,  
Woodhill Park Research Retreat, Aotearoa New Zealand

#17699 Multiple Users/Coders and Coding Comparison Query

Posted by Kari Struemph on 06 May 2016 - 10:31 PM

I'm trying to learn how to do the Coding Comparison Query - I've figured out how to change users, however it is not intuitive to me that the document stays coded when you change users instead of giving you a clean version of the document for the second coder to code so a comparison can be done.  Can anyone provide me a bit of guidance on how this process works? 


I have documents that will be coded separately by two coders and then would like them to be compared.




#17669 Timeout for matrix queries

Posted by j.innes on 26 April 2016 - 10:54 PM

Hi all,


I have revived an old thread in Nvivo9/10 forum about query timeout. 


I have gratefully received a response from QSR helpdesk about this topic and so am sharing this in case anyone else finds it useful:


Firstly, please follow the steps below to perform a Compact and Repair of your project:

1. If your project is located on a network drive, copy it to your local machine.

2. If the project is currently open, close the project via File > Close.

3. Click the File tab > Help Compact and Repair Project.

4. At the Select Project dialog box, locate and select the project - select NVivo Projects (*.nvp) to repair a standalone project.

5. Click Open.

If the issue persists, you may need to change the SQL Time out period manually by editing your configuration file. This may be a little technical. You can ask your IT support personnel to help you with this if you are unable to do this yourself. The steps are below:

o    Ensure NVivo 11 is closed.

o    Go to Start -> All Programs -> Accessories -> Wordpad

o    Click on File -> Open. Ensure "All Documents (*.*) is selected in the 'Files of Type' dropdown menu in the Open dialog box.


1. With NVivo 11 closed, open Windows Explorer

a)    Right-click on the Windows Start circle -> Explore.

2. Display all hidden files and folders:

a)    Click on Organize -> Folder and Search Options

B)    On the 'Folder Options' dialog, click on the View tab and check 'Show hidden files and folders' option

c)    Click on the Apply button and click OK.

2. Navigate to the following folders:



o    You should see few folders that start with NVivo.exe..... Go into the first folder and see if it has subfolder named ""2. If not, go back one level and go into the second NVivo.exe.. folder to see if it has a subfolder One of the two folders will definitely have a subfolder named ""2. Go into this subfolder ( You should see a file called user.config, double click on it to open it.(Open it with Notepad)

o    The 19th3 line from top in this file (or one or two lines above or below) should have something that says SqlCommandTimeout setting. The next line should have a value of 60. Change this value to a higher number like 120 or 180.

o    Open NVivo 11 and re-run your query. If you still get the same error. Try increasing the timeout value to a higher number.

Please note changing the sqltimeout value may result in the Cancel button (when you run a query) to become unresponsive.

Lastly, if you are using 32-bit NVivo 11, it will also help to switch to 64-bit version. Please refer to the link below to find whether you can install 64-bit,



1. Rather than going through all those steps, I found it easier to type %appdata% into my Windows Explorer and choose the right version of Nvivo from there.

2. Just to point out that this number is the version of Nvivo you're running. Your version may be different, so check this first. You can do that by going to 'File', 'Help', 'About Nvivo'. 

3. It was not the 19th line for me. I just used CTRL+F to find the SqlCommandTimeout setting.

#17608 Drag and drop coding

Posted by JulianYates on 06 April 2016 - 09:19 AM

I find NVivo's software quirk on drag and drop coding extremely frustrating. This quirk relates to text selection: once a section of text is selected it is intuitive to drag it quickly to the desired code, but in NVivo this doesn't work; instead the selection is dropped and a process of selecting new text begins. It seems that NVivo requires the user to click on the selected text, pause for over 1 second, and then begin the drag motion. This is painful - it doesn't seem like much but it interrupts fast working, especially when selections are dropped.


Moreover, it diverges from standard logics in other software. For most software, to drop a selection and begin a new selection, the user must first click and release to reset the selection. Why doesn't NVivo follow this logic, rather than abandoning the selection as soon as the cursor is moved while still holding the click?


Please make this very simple change; it would make coding so much faster and less painful! 

#16991 Coding query

Posted by Orionspants on 13 June 2015 - 12:06 AM

Does your source show zero nodes and zero references? If so you may be experiencing the same issue I am, likely a misstep in project step-up. QSR advised zipping the file and submitting a support ticket: http://assets.qsrint...m/support-form/



However if just trying to run a standard query I've also had no luck with that, but matrix coding queries work fairly well. 


Good luck!

#16817 Program Design Adds Time to Coding Process

Posted by Boston2014 on 11 March 2015 - 10:26 PM

I switched a few months ago from NVIVO 10 for PC (running on a Mac through Parallels) to NVIVO for Mac. Although I appreciate not using Parallels, which makes most computers run incredibly slow, I have found that NVIVO for Mac takes much more time to code than when I used NVIVO 10 for PC. I know NVIVO for Mac is still in development and I am hoping you consider a few improvements in future updates/versions.


1. It would be helpful if NVIVO for Mac had a “quick coding bar” and allowed code nicknames like NVIVO for PC does. The drag-and-drop coding (node) is a nice feature. However, it can be difficult to use and it increases inaccuracy in coding. Often when a user clicks a highlighted area to drag it, the item will be mistakenly be dropped by the user in the wrong place and/or the user will mistaken click another area and then the text becomes un-highlighted and not coded. The user then needs to go back re-highlight and drag and drop again. It also seems that the drag and drop feature does not allow the user to code a single piece of text with multiple codes at once and/or quickly. I have also tried using the right-click process and code the selection that way using the drop down list, which is also incredibly time consuming (and the codes always default to collapsed view here as well, taking more time to “expand all”). Coding is already a time consuming process and without a quick coding bar/nicknames, using NVIVO for Mac adds significant amount of time to your coding. Users could still use drag-and-drop if they find it useful, while others could use the more reliable and quicker “quick coding bar.”


2. It would be helpful if the Detail View (where nodes are displayed), which can be Expanded/Collapsed, had a default view to whatever the user last used. For instance, when I toggle between free nodes and tree nodes, the program usually returns to a collapsed view. I almost always want to see an expanded view (all my codes), especially if I have to drag and drop text into each one. It takes a good amount of time (and is an annoyance) that when I switch between free nodes and tree nodes that I have to go back and expand the nodes again.

#12901 Naming sources imported from EndNote to author-date-title or as per original...

Posted by david004 on 06 August 2012 - 09:47 AM

Importing references from EndNote. The only option for naming is either Author-date (which comes in with superfluous semi-colons and EndNote record numbers) or Title. Is there some way to name the sources Author-date-title? If not, is there some way to import the file with its original name? If not, can you apply the classification and attributes of an external to an existing internal?

#12806 Node structure and relevant coded text

Posted by Jason Flett on 18 June 2012 - 10:36 AM

Hi all

In response to the previous query that you have raised with us, we have now included two new reports/extracts in NVivo 10. Coding Summary by Source, and Coding Summary by Node, which include the text of the coding reference.

For information on purchasing or upgrading to NVivo 10, or to download a free trial, go to http://www.qsrintern...ucts_nvivo.aspx


#12793 Node structure and relevant coded text

Posted by Marcus Ogden on 07 June 2012 - 03:26 PM

Hi all,

I'm pleased to advise that NVivo 10 (to be released later this month) will include new "Coding Summary by Node" and "Coding Summary by Source" reports, which will include the coded text (organised either by node then source, or vice versa). You can choose which nodes and/or sources to include in the report, and you can save the results to a range of file formats.

@MJ Drennan - What you're asking for ("content with the codes as the secondary pieces of data") is somewhat different to what's being discussed in this thread, but it's also been requested before. One possibility you could consider (depending on the number of nodes involved) is to print with Coding Stripes turned on. We are looking into other ways we could better meet this need.


#18250 Spread Coding with Broad Context Crash

Posted by aliciap on 14 March 2017 - 06:57 PM

When trying to set spread coding to broad context for all references on a single node, NVivo crashes. Stack trace error reads "Object reference not set to an instance of an object". Seems to work fine if I set spread coding on a single reference.


Similar problem when trying to set spread coding to broad when running a query, this time I just get an error box saying "Object reference not set to an instance of an object" and NVivo doesn't crash.

#18066 Word frequency on word frequency results

Posted by QSR Support on 17 November 2016 - 10:30 AM

Hi Mitja,


One you have done the word frequency query:


1. Right-click the word help and select Create As Node.


2. Open the 'help' node created at step 1 and click in the detail view and then press CTRL + A so the whole content of the node is selected. 


3. Right-click in the detail view of the node and select Spread Coding and then Broad Context.


4. Do a Text Search query and for 'Selected Items' choose the node 'help' and then run a query for the word administrator. 


5. The references that appear are those where the word administrator is within 20 words of the word help. You can increase/decrease the spread if need be. For example, at step 3, select Spread Coding and then Custom and you can adjust the number of words for the spread.


For more details on spread coding, check out the following link:







#17553 Export Matrix Coding query with all node references

Posted by ahmadasemoni on 05 March 2016 - 05:01 PM

Its not currently possible to export the table and references as one in the way you describe. You could export the table by itself, and then export the references for each cell as a separate file. I will suggest this to our product team so that they can consider it for a future release.  

 Thank you, if you can add this possibility to Nvivo, it could be more useful.

#17309 Linking Imported Node Classification Sheet to Text in Sources

Posted by karinb15 on 22 October 2015 - 09:19 AM

Hi there,


I have been working on the solution to linking data like this for two straight days as I knew it could be done and just could not put all the steps together in the right order. I finally

got it! After lots and lots of trial and error and searching through the available documentation/support, I found these are the steps that worked for my data, which is from online surveys. I'm posting here just in case this works for others too, but I'm sure it will need to be altered for different data types.


  1. Prepare and load your database.
  2. Make sure your participant nodes have NOT been created yet and make really sure there are NO classification sheets loaded into the Node Classification folder yet.
  3. Autocode the database to create CASE NODES. Use the “source structure and style” option and then the  “nodes for each value in a column” option.
  4. If you have already coded some content in this database, check to see that these new case nodes have coded material in them. If not, try coding some to make sure it registers in these case nodes.
  5. Import your classification sheet. Make sure to use the same identifier that you used when creating the case nodes.
  6. Highlight all the case nodes (use the scroll and shift key method if there are many) and classify them under the classification sheet you just imported.
  7. Now you should be able to run matrix queries that use your Node Classification attribute values for one axis (so for example, English, French, and Spanish speakers or ages 20-30, ages 30-40, ages 40-50, etc.) and coded material to any node(s) you choose on the other axis.

Hope this helps - good luck!

#17025 Coding paragraphs by a word search

Posted by TanjaC on 26 June 2015 - 03:33 PM

Hi there,


I have done a substantial amount of coding so far as part of a very large project. Half way through my coding I realised that I should have created a node for a specific topic which is reoccurring frequently. Currently, most of this topic is coded under a more broader node which includes other things as well. Now I want to separate those.

What would be the best way to go back and select a paragraph that I want to code into a more specific node? I know I can search for a word, and I could do this because it is the word that will define the node, like an acronym. However, when I search for a word I only get a snippet of the sentence in which this word is mentioned. Is there a way that I can see the whole paragraphs without double clicking on each sentence where the word is mentioned and opening this section in another window and selecting a paragraph from there.


I hope this makes sense.


Thanks for your help,



#16524 coding and node classification questions

Posted by QSR Support on 02 January 2015 - 03:22 PM



You cannot assign more than one node classification to a node. You could, however, do your second idea and create regular nodes for “experiences of discrimination” and then use a matrix-coding query to compare your results based off of other factors. The query would depend greatly on the type of differences that you want to see, however.


Please see the following for more information about matrix-coding queries: http://help-nv10.qsr...oding_query.htm


I hope this helps.


Kind regards,


#14927 Linking Typed Transcript (PDF) with Original Audio File

Posted by createpac on 06 June 2014 - 10:02 PM

In response to your original question that for housekeeping purposes you want a link from a pdf to a audio and vice versa - a 'see also link'  from Links and use  the ''xisting item option will enable this - so you can use a see also link from a piece of the audio timeline to the pdf and from a piece of text in the  pdf to the audio.  This works well in version 10 and Iam sure I used it in some earlier version as well.

#12844 Coding web page

Posted by Robin Gauld on 03 July 2012 - 03:26 PM

Thanks folks, yes cut and paste as usual. I watched the webinar the other day and watched the cool stuff 10 can do with online content. I want it.

#11163 Open project in same view as when closed

Posted by Karma on 30 September 2011 - 04:53 AM

I agree, this is an absolute necessity for future updates. I spend far too much time resetting source and nodal tree views when I first open a project file and when Nvivo crashes after failed queries.

#10982 Sort Sources by Date?

Posted by Leonie Daws on 22 August 2011 - 08:56 AM

Hi Troy,

Sources appear in List View and in Nodes in alphabetical order. To be able to view the sources by date, I'd suggest including the date at the beginning of the source title. If you use YYYYMMDD format then sources should appear in chronological sequence sequence. You can use attributes in matrix coding queries to compare, for example, all sources published in a given year with those in other years. You can use attributes and the Find tool to create sets of documents with similar dates. You can also sort sources by date Attribute values in the Casebook (Version 8) or Classification Sheet (Version 9). I'm not sure if any of these options achieves exactly what you want but they may go some way towards it.



#10374 Node structure and relevant coded text

Posted by Marcus Ogden on 10 February 2011 - 01:44 PM

we used to be able to export to Word but now the only options look like doing one file at a time in HTML or printing.

You can export nodes to Word in NVivo 9 (albeit to one file per node). Select the nodes in the list view, choose Export > Export Node from the context menu, and in the Export Options dialog, select 'Reference View' from the Export picklist.

Your last option of selecting all nodes from the Collections/search folder only allows printing/exporting of this list itself and not coded content. Am I missing something?

Again, you need to run Export Nodes here, not Export List.

KMcLilli and Laurajd, could I ask you a few questions to help us improve this area of NVivo?

1. If you were able to export the content of multiple nodes to a single Word file, what would you use this for – i.e how would having all this data in one file help you do your research?
2. For each node you want to export, would you want all of the node’s coded content, or would you want the node’s coded content for specific sources?
3. Do you use audio, video or picture sources in your NVivo project? If so, what would you expect the exported file to include regarding coded content of these sources?