Jump to content
[[Template core/global/global/lkeyWarning is throwing an error. This theme may be out of date. Run the support tool in the AdminCP to restore the default theme.]]


  • Content count

  • Joined

  • Last visited

Everything posted by alexquse

  1. alexquse

    What is the "new nvivo"

    I receive advertisements for upgrading to the new nvivo. I bought nvivo 12 for less than 6 months and the upgrade price asked is USD 625... wow! It is very hard to find what is basicaly this "new nvivo" (nvivo 13?), and what are the improvements. Any information about this version?
  2. I have a rather difficult issue to explain. Let's take a survey with two open-ended questions. Both questions were coded using the same coding system: Code(A) and Code(b). I would like to analyse the differences in coding between the two questions for the same person: what the person expresses in question B according to what he/she expressed in question A. In other words, I am trying to create a matrix coding where only the coding concerning the same person appear: it is quite common with 1,2, ...5 cases, but with more than 100 cases it becomes tricky. Formally, if ID is the person's identifier, I want to creatre a matrix where: Code(A) x Code(B) for any ID(A) = ID(B). I thought of creating relationships manually between all cases to link ID(A) and ID(B) , but it's long (no way to creat relationship automatically from a table?) and finally I'm not sure what to do with all these relationships. I thought of creating as many cases as questions-ID. Then in the classification table, I indicate the ID, which allows to retrieve the couples. But after doing that, I am stucked. The workaround I found is to create this table of transformations on a statistical software to detect the most frequent movements from A to B. And then manually create relationships in Nvivo on these most frequent movements. But by doing this, I lose the interactive capabilities of nvivo. And if I get a set of relations between codes ("when there's that in A, there's that in B") but I still don't know how to display only the content where ID(A) = ID(B). Any idea, suggestions? (I'm not sure to be enough clear, both my english and my ideas are fuzzy )
  3. I am looking in filter of crosstab Query. There are a lot of options but can't find any option to filter when "in the same Set" or "in the same Document".
  4. Dear all, Sorry to ask a question that has been asked many times. One of the issue I have with nvivo is choosing how to structures the source files. Having one file per case has advantages and disadvantages. My experience is that nvivo has trouble when working with many separate sources. Merging the cases in the same file earns in processing speed (nvivo but also human processing). But I cannot access to some functions when all cases are merged in one source file. Nvivo is more powerfull with each case in a separate source document. I would have liked to have your point of view, the pros and the cons of having cases aggregate in one source or each one in separate sources. This project consist open ended questions of a survey. So I have about 3000 individuals. But we can discuss this issue more generally: at which point it is a good practice to separate the sources?
  5. Yes I will try this. I have not begin with my big project, still doing some tests. I have one more question about autocoding. Could this be a way to change the structure of project? I know that the autocoding (existing pattern) is not reliable. But if we have to restructure our project differently, one idea is to re-import a new source and autocode according existing pattern. Does anyone tested? (I'm going to try: I am working on an extract of 500 entries with two open-ended questions).
  6. alexquse

    Italian as content

    I'm interested to work with french, german and italian texts. Do you plan to implement italian as text content language? (or a way to implement custom lemmatization dictionnaries)
  7. alexquse

    Italian as content

    up! 😀
  8. Thank you. File is structured with individuals as Set (and Cases, which is straightforward). Then it is possible to do anything I want with node matrice. Another question: I tried the new function that allows to import data directly from a SPSS file. I couldn't find a way to get the open-ended questions on a "document page": there are embeded in a grid which is not a very practical layout for coding. Not a big deal. More bigger problem: the import stops at 100 cases, which makes the function useless: a "quantitative" study with open-ended questions has often a few hundred cases. Or did I miss something?
  9. My question is simple: does nvivo 9 work with windows 10?
  10. alexquse

    search folder

    search folder is a nice feature I've just discover now .... but the search form miss "not coded at"
  11. I notice an annoying problem with a big project of 1700 items: shifts appear in the coding! At first I thought it was accident, and I corrected them. But I'm beginning to realize that it is almost systematic. I do not understand what occasioned them, but it's pretty scary to see my coding work deteriorate ... Is there a solution? Uploaded with ImageShack.us
  12. alexquse

    Bug in nvivo 9

    I found a bug on Nvivo 9. I had noticed long ago, but since it is special circumstances and not a very important problem, I had forgotten it. It is impossible to copy more than one node in a node of the window "find result". Nvivo closes unexpectedly.
  13. alexquse

    Bug in nvivo 9

    I can't reproduce the problem anymore. I 've sent you anyway the errors log files. Thank you for reactivity Alex
  14. I worked for years with nvivo. My greatest disappointment is the slowness of the software (I've complained many times here). If you have a dataset of a some size, each operation takes several seconds. This does not seem serious, but when you repeat some coding operations 1000 times, it ended up taking time. The problem actually started with version 8. I went from version 8 to version 9, because the advertising said it would improve the speed. But in fact, the improvement may be 20%, which solves nothing since, for example, from 10 seconds to 8 seconds, it is still 8 times too long! Now we are told that version 10 is faster. But what is it really? is there people working with "normal" corpus of qualitative research (eg, 50 interviews is 1000 pages)?
  15. Here is a solution (by converting classification into nodes) http://forums.qsrint...?showtopic=3436
  16. Hi, I have many classifications of sources. So my question is what and how can I do with these classifications? I want to analyse the distributions of node by source attributes, but it seems very laborious. The only easy features I've found is to visualize sources by attributes. Does anyone have a solution to use these classifications. I should mention that I sought for a long time and the only option I've found is to create node. Does anyone know a way to pass from source classification to node classification. Thank you Alex oopss!! sorry for the multiple post : I received an error message and though that nothing was published
  17. alexquse

    Fix problem that causes corrupt files!

    For a long time I wish that qsr make a more stable, lighter, more practical in terms of usal practice of research (efficiency, flexibility, speed). I lose a lot of hours because nvivo often freezes. But I had only once a corrupted file: am I lucky?
  18. I've found my previous issue. It was just the reverse problem (and with nvivo 8) http://forums.qsrint...?showtopic=3436 I'm still looking for a solution to automatically transform attributes which are pointed with a series of nodes to a classification edit: I've solve my problem in ten minutes of manual coding ;-°
  19. Hello, I 've searched for hours, maybe it's easy !!?? I have a series of nodes that I want to turn into classifications. In other words, attribute values are stored in nodes, and I want to stored them in classification. It seems to me that in NVivo 8 it was possible to transform nodes into classification by the way of the collection (set) and then export in an excel sheet and re-import as classification sheet. But I do don't see how to achieve this with nvivo9. Any idea? Thanks Alex
  20. I'm not sure to fully understand your method, but it should work (do you select "automatic select hierarchy"?). Otherwise I have a method which is a bit "prehistoric". I usually often create an "allnode" (with aggregate fonction (nv9) or merge (nv8)). Then with a simple coding query = everything that is not coded at allnode. It's pretty simple, very robust and the "allnode" code is useful for other "administrative operations".
  21. Not an answer but add: it would be great if there was a simple procedure to export in pajek (a powerfull and free tool to visualise network).
  22. Hello, I don't know if your idea really fit to a qualitative perspective,but... I made a proposal some time ago, that goes in your direction. I can developp it here: The principle is to include a timeline, for example to detect the first appearance of a concept (in different interviews/sources/cases), to identify what precedes/accompanies/follows this emergence. This tool could trace the evolution of the concept: the first appearance is accompanied by such a node, second is accompanied by such a node, third... I think you could then realize the kind of analysis you're talking about. But I do not know exactly how to realise this feature. just a first idea: - the user builds a timeline with categories (1-2-3-4-5) , selects a node (N)and select a filter (source xx) - the fonction creates a "time-node" as a sequence of nodes (N1xx - N2xx - N3xx - N4xx - N5xx....) by the timeline and related to the filter. - the user can test what is link to N1, or the difference between different filter N1xx and N1yy... I hope to be not too obscure...
  23. I often read this. I confess that I am not very confident with my technique, but I do precisely the opposite: I distinguish very early between rough concepts and I refine later. I understand that this way is "hazardous", but it seems to me to have heuristics virtues (otherwise, I'm quickly jammed). I practice at the extreme this way of doing but sometimes I wonder if I do not getting lost in such an analytical decomposition when it goes too far (for example: I code the modalities separately, so the negation "a" and "non-a" shares one code: "a"; and are distinguish with another one: positiv/negativ) @Pablo: We are indeed quite agree. I also implicitly raised the idea to discuss structures of nodes here. Once I posted one of mine. But I realized that, if it takes an half day to anyone to understand the study of someone and make relevant remarks about the working technics... there won't be a lot of dicussion. (Another problem is that my studies are in French)
  24. I can not completely agree with that. Yes, strictly spoken, each topic, each theoritical framework involves different concepts. But first, there is not an infinite number of approaches and theoretical models. Second, there are similarities, logic, how to structure a tree which can be shared. We may discuss tree structures because we can have meta-comprehensions (otherwise, it would mean that only those who have the same theoretical model and the same object could discuss a study?) To take an example of meta-structure: in a study based on observations, I code types of actions and speeches, transformations (such as affirmative, critical, ironical, etc. ..), roles or characters, used objects or accessories, etc. Then, by crossing these dimensions, I can test my assumptions. In another study, I distinguished types of argument, as they are based on analogy, "rationality", persuasion (aggression, humor),etc. and on the other hand, objects mobilized (general principles, examples, etc.) and finally actors.
  25. I fully agree that NVivo is a very good software well manufactured. But about the specifications of the computer I can not agree: my notebook is good and recent: do I need a computer that has less than six months and costs $ 4000 in order nvivo works well ? if so, I would stop immediately using such a software! But fortunately this is not true. In fact, the slowness problems exist even with a powerfull computer. Then we must know more about what we mean when we talk about slowness: with large projects, each operation takes time. So nvivo works well for small projects (20 nodes, sources of 10,000 characters). But nvivo don't work well with large projects: try to encode for 30 hours with a delay of 5 seconds per operation: you'll realize that's unbearable. This is not a major defect. And finally it seems that nvivo9 is faster than nvivo8 (I had sql problems). But a nvivo software with smaller latency (such as a keyboard: you cannot tolerate to wait one second for every keystroke) would be so great. But, however, I appreciate a lot Nvivo even with its smoothness. Unfortunately I am currently in a period of theory and in other quantitative projets, and I am pleased to resume my analysis with NVivo. Alex