Leonie Daws

The Slider which allows the inclusion of stemmed words and synonyms in Text Search and Word Frequencies is constructed as a continuum. There needs to be an option to include synonyms and generalisations for the narrow term alone.The current Slider function enforces the inclusion of all stemmed words in the search for synonyms, and returns synonyms for each stemmed term as well as for the original stem. Including synonyms for all stemmed words frequently returns many irrelevant finds. To take an example, if you searched for synonyms for "protect" you would get a particular set of finds. Once words like "protection" and "protective" are included, the range of finds blows out to something quite unworkable. See the attached screen shots for examples.

Just in case this has disconcerted other users, the process is identical in both versions but the location of the Classify nodes from a Dataset command has changed from the Analyse Tab in Version 9 to the Create Tab in Version 10. Nicole's original confusion about the option to "choose column that contains node names" relates to informing NVivo which column contains the names of the nodes you wish to classify. This is usually the Respondent ID or Case Name column. Cheers, Leonie

Hi Mikael, You can use Source Classifications with Matrix Coding Queries. They are not available in a Coding Query as the Attribute option there can only relate to the node you are querying - i.e. an attribute of that node. Cheers, Leonie

Well done, Annabel. That was the point at which I stumbled the first few times I tried to work with classifying nodes from a dataset. I'm glad it's working now. For others who may have been following this discussion, the attachments I've posted form part of a set of tutorials that are available for purchase in pdf format. Details are available from http://www.kihi.com.au/mod/resource/view.php?id=68 Cheers, Leonie

Hi Annabel, I've attached a second document which hopefully contains the steps you wish to check. Having your quantitative data as integers should not present a problem. You can, if you wish, instruct NVivo to read these items as text rather than as numbers at the time that you import the data. Good luck, LeonieWorking with Datasets and Classifications.pdf

Hi, I'm not exactly sure at what point the process has gone wrong for you but the attached instructions may assist. The instructions are an extract from a set of tutorials which I have prepared, based on those originally prepared by Lyn Richards. They are available for purchase if you are interested - details from my website which should appear in the signature below this message. Regards, Leonie Classify nodes from values in a dataset.pdf

Hi Margherita, Have you checked the Help file? In the section on Visualizations, there's a full explanation of how similarities are calculated, including cross references to articles explaining each of the similarity metrics. Regards, Leonie

Peter, NVivo does not have a function to do this automatically. Someone else may have a suggestion for you. Regards, Leonie

This is something you do as you import the dataset. At Step 4 of the Dataset Import WIzard, check each column by scrolling across. If it is recorded as Classifying Field and you want to be able to code it, select the column, then click on the button beside Codable Field under Analysis Type above the list of columns. Do this for each column you need to change. Hope this helps, Leonie

Hi Anita, How long are the documents you are working with? If they are long - more than about 10 pages each - split them into shorter parts. You can do this in NVivo without losing any of the work you have already done. Create a new document, select a section of an existing document, copy it, and paste it into the new document. Your coding will be copied across. Then delete that section from the original document. You will find the program will work more quickly with short documents. Hope this helps, Leonie

Hi Liz, If you want Section 1 to contain all the responses to parts A, B, C, etc for Section 1, and Question A to contain all the question A answers from each section, you need to use Heading 1 for the Sections and Heading 2 for the Questions, then autocode twice, first using paragraph style Heading 1 then again using paragraph style Heading 2. Jennifer is correct that each question would need to be named identically to be coded into a single node. This should produce the results you want. Regards, Leonie

Hi Peter, I'm not sure I understand exactly what you are looking for. Looking at the example you have given, you would need to create subnodes under weather for each possible weather description. To use your original statement, you would need to have a node representing each stage in the development of the concept. If you then use a Matrix coding query with the dates as rows and the phases in the concept as columns you should see if there is a pattern in its development over time. Regards, Leonie

Hi Susan, The ability to store files externally and still continue to code them as 'internals' applies only to audio and video files that exceed the size limit for embedding files. If your sources are Word documents they can only be coded directly when imported as internals. The number of sources is not such an issue as the size of individual sources. You can get around this by breaking large documents down into a number of smaller files. If you have already commenced coding, you can create a new document, then copy a section of the larger source to the new document. This will take your coding with it. The copied section can now be deleted from the original large document. I hope this helps. Regards, Leonie

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. Regards, Leonie

Hi Simon, You are correct that NVivo does not offer the level of automated analysis you are looking for. There is also no equivalent of the SPSS syntax files that would enable you to program NVivo to produce the calculation you require. Regards, Leonie