Preparing textual data for QSR NVivo versions 1, 2, 7, 8 and 9
Date written: July 2005
Updated 3rd April 2012
Documents can be typed directly into the rich text editor in NVivo (i.e. in a normal Document Browser window - v. 1 & 2 - or the Document View Pane - v. 7 & 8)
OR they can be imported in rich text (.rtf) format. In MS Word: type it up and Save As Rich Text Format. (If you don't use Word, you can import in plain text from any word processor or in rich text from another word processor via Word or Word pad).
OR in Versions 7 and 8 they can be imported as Word files (.doc only, not the new .docx format). However, there are certain contents that NVivo 8 cannot import. See notes on trouble shooting below.
OR in Version 9 Word documents can be imported in both doc and docx formats. Also pdf documents can be imported.
Caution: When saved to rich text format the document will lose embedded items like illustrations, tables. BUT before you save, you MUST remove the tables and graphics or the file will not be recognizable for NVivo. If important save such objects separately, and you can place links (databites) to them after importing the text document. You can copy text directly into an NVivo document from any source such as word processor, web site, email etc.: it will copy as plain text (or as formatted text in v. 7 and 8). You can edit any document freely (font, colour, bold, italics, heading levels etc.) once it is in NVivo, without invalidating coding.
There is no spell check in NVivo. Do as much cleaning up as possible using edit tools e.g. in Word before importing into NVivo. See General transcription guidelines
Consistent Heading levels will also provide ready-made structure which can be utilised to great extent in NVivo. See below.
Sections and Paragraphs
The ability of the software to recognise either of these levels of unit of text may be useful when autocoding structured data: see in software the Section Coder, and Paragraph coder (Auto Code and Paragraph code in v. 7 and 8).
PARAGRAPHS are recognised.
..by at least one hard return at the end of paragraphs. Because sentences are not recognised during autocoding/text searching, paragraphs become the most useful unit of context in Nvivo, Make use of them whenever you can. At transcription. put paragraphs in as often as possible if appropriate. Multiple paragraphs may fit within one section.
SECTIONS are recognised.
.by creating consistent heading levels (if appropriate e.g. with structured data-see notes
..EITHER Create Heading levels in Word file before saving in Rich Text format ..
.. OR Perform similar tasks using NVivo edit tools in the Document Browser/Document View Pane after the data have been added to the project:
The use of sections will be relevant if you have structures which provide useful contextual chunks of data, which group several paragraphs within them. For example if you have used structured interviews and everyone has answers to Question 1, Question 2 and so on, or if you have undertaken focus group interviews and want to automatically code each speaker using their name which appears before what each says in the transcript.
YOU DO NOT HAVE TO HAVE SECTIONS OR SECTION HEADERS IN NVIVO
Files can be changed and edited flexibly after import - formatting and annotating can continue during analysis in NVivo.
Create up to 9 nested levels: e.g. by using corresponding Header level numbers
REASONS TO USE SECTIONS (i.e. denoted by heading levels)
You would use Sections in your data, especially if you plan to use 'Section Coder (autocode or paragraph code) in NVivo in such data as :-
Structured data: i.e. Open ended questions, to denote question number or Respondent ID, or when using very structured interview framework
Focus groups / group interviews: to make the autocoding of speaker sections easy within (this will impact on the ease with which you can assign sociodemographic variables or 'attributes' to these speakers later - and then interrogate data by those attributes or combinations of them.
You may use sections or paragraphs to spread the coding to a larger contextual chunk or to view more data around a previously coded segment.
Trouble shooting in versions 1 and 2
RTF and heading levels are a bit problematic. If you only need one level of heading -choose Option 1, on import, then use Heading level 1 for section headers..it just works better. Be warned ! if you use the Section coder , you may have hidden heading level codes embedded on empty lines which will upset the tidy auto coding of sections when using the Section coder tool - (you do not have to use the section coder)
See other problems with import of transcribed documents below
Rich text structures which will be modified or ignored include :
footnotes, tabbed tables of contents, page numbers .. and many more:
Structures which will STOP the import of data include :
Tables, Embedded objects (pictures, sounds, OLE objects, text boxes etc) headers, footers (inc page numbering)
RTF documents created with *Corel Word Perfect
If you are having your files transcribed - you may not know if Corel Word Perfect has been used - you may have to copy and paste text from existing RTF files into new Word files - in order to make format acceptable
Trouble shooting in versions 7 and 8
Charts and objects created using the Microsoft Word drawing features cannot be imported.
Documents containing many images can significantly increase the size of your project file. If you are concerned about this, limit the number of images in imported sources.
If a document contains styles not supported by NVivo (such as styles you created for yourself in Microsoft Word), the content is converted to the Normal style and formatting is lost.
Document footers are not imported.