The text unit You have to decide on the basic unit of text appropriate for the analysis of an imported document. The end of a unit and start of a new one is indicated by using a hard carriage return (producing the paragraph symbol in Word and other word-processing packages). A paragraph with one or more sentences (perhaps representing a reply to a question in an interview) is a common unit; however, you might choose to use a single sentence, or a line, or even an individual word as your unit; in each case, this is indicated by a hard carriage return at the end of a unit. If individual words are chosen as the unit the document, with a single word on each line, is unwieldy to deal with and has the effect of stripping a word from its context. Hence NUD*IST is not well adapted to this fine-grained analysis. You can have paragraphs in the document as default choice, but insert extra hard carriage return at points such as sentence ends, or phrase ends, if you wish to. ATLAS/ti and NVivo permit form coding’ (you simply select each piece of text you wish to act as a unit).
External document can be divided in a similar manner and can have text units of varying lengths and types (e.g. a page in a field notebook; or the tape counter number for an untranscribed audio or video-tape). You specify the number of text units in a document when integrating it into NUD*IST and can code these units as you would those in an imported document. While you can’t retrieve the text itself, the text unit number can be liked up in the documents themselves.