Hello Scott, I cannot see a file attached at all. You can copy and paste into excel, see the help file; Import data using copy and paste

Now that we’ve imported data from an Excel workbook, let’s import data from a table we find on a web page, or any other source from which we can copy and paste into Excel. In the following steps, you add the Olympic host cities from a table.

1.Insert a new Excel worksheet, and name it Hosts.

2.Select and copy the following table, including the table headers.

This is an example. I have no idea about the images. Thanks Debbie

All of the major qualitative data analysis (QDA) software packages can import and export data via Excel. One thing that you have to decide, however, is whether you want to most of your analysis in a QDA program, along with a few close-ended variables attached for sorting etc., or whether you want to use a quantitative package such as SPSS where your codes become variables.

I am guessing you want to do something like content analysis, where you would generate a list of variables that were either present or absent in any given open-ended question or image. In that case, your results would be largely in terms of percentages and cross-tabluations.

If so, with an N of 300, I would recommend using a quantitative software package. If you do want to stay within a QDA framework, you should look at Dedoose. It is an online, browser-based program that can produce all kinds of charts and graphs from your coded data.

With regard to the image data, most of the QDA packages (but not Dedoose) can handle these, often as the equivalent of pdfs. The idea is that you can mark a portion of an image and attach a code to it. For example, if you had a landscape that included both a tree and a man, you could outline the figure of the man and give that one code, and then do the same with the tree.

Note that this approach links specific visual segments with codes, but you may want something simpler, In particular, if all you want to know is that the image includes both a tree and a man, and you probably don't have to go to the lengths of scanning in a pdf Instead, you could just attach the codes to something with a label like Image #1 etc. This amounts to coding the images "by hand" and then entering the codes into each case.

Finally, NVivo might not be the best way to introduce yourself to this kind of work, since it has historically been heavily oriented toward interpretive qualitative work, without a link to survey data. So, if you want to get a working sense of how working with survey data works in the QDA world, I would look at the training materials for Dedoose -- with Max QDA as a second choice (rather than NVivo or Atlas).

Hello, i used R to analyse my survey results, its free, open-source and very powerful. You can find some tutorials for survey analysis on this website :

http://www.asdfree.com/

Thanks for the replies! I guess what I am really looking for is a way to avoid manually cutting-pasting, answer-by-answer, into a program like excel. 300 surveys X 13 questions = 3900 entries and 3900 opportunities to make errors.

David you are correct in that I'm doing a content analysis. As for the images, each one would need to be coded with the demographic survey data as well as the recognizable visual elements, like a tree or man being depicted.

Laurent, I took a look at R, but it's far too complex for me, I only have very basic HTML skills.

I just assumed (wrongfully) that there must be a program out there that could simply read the numbered answers from my text files and compile the stats, at least for the closed questions.

I don't see why you can't read your txt file into Excel directly. Presumably each case is on its own line, and the variables are separated by something like tabs or spaces or commas. If so, Excel will import the file automatically.

Alternatively, if the data are "solid" lines of text, then you will need to read that into something like SPSS using a format statement. Or, if you have a "tab-delimited" file, then SPSS or almost any other stats program will read that and you could do all your analyses there. In other words, you don't have to do your analyses in Excel, it is just a common format for getting data from here to there (and tab-delimited files are often just as good for importing and exporting).

As for the images, it sounds like you just need a codebook so you can keep track of what is present or absent in each image rather than "tagging" the exact location where each feature occurs in the image. In that case, you can avoid the hassle of scanning pdfs into a qualitative analysis program.

I believe that attaching at least a part of that file would be very helpful into determining the best way to import all those results. In that way you will get a very straight-forward response.

Here is what the files look like, with #13 removed as it contains identity info.

Ok. It's not that complicated, but requires some repetitive work.

If you have access to SPSS, you can import the txt file in spss (i'll provide a template for import that you can run). - importing being automated, possibility for error is very low.

You have two options:

- either do the import for each case, then save it as excel and add each new case (you'll get rid of the number of the question using excel formula)

- copy all cases into one big txt file (no lines between cases) and import all of them at once, then in excel remove the number of the question.

It's not that hard and it shouldn't take a long time.

(the files is set so that each case has 13 lines).

The underlying problem is the format in which you've entered the data. All of the programs that we've been talking about assume a "rectangular" format for the data, where each row is one case, and the columns are the variables. So, the first step you will need to do is convert the data into something that matches that format.

As Sergiu Chirilă suggests, do this case-by-case and use formulas in Excel, but I almost certain you could do the same thing with the Change/Replace command in Word. I have attached a revision of your sample file where I used Word to convert the original entry format into a single line with tabs between each of the variables (there are also tabs between the original question question numbers, but those would easy to clean out as well.

Oops, I hit the "Add button" before I did the attachment. Here it is.

Thanks again for the helpful responses. I have gotten a hold of SPSS 20, will try to figure it out and let you know!

you may try Epinfo 07 to code and also do the analysis. with visual dashboard of Epinfo, you just see everything happening. i found it useful when i was extraction information from a set of studies. Alternatively, try using this command read.table(file.choose(), header=T, sep="\t") in R

Scott,

What puzzles me a bit is why you would want to analyze quantitative survey responses in NVivo? I'm guessing that the survey data must be linked to qualitative interview transcripts or the like i.e. that you're doing a Mixed Method study.

NVivo does have support for doing (some of) that kind of analysis, but (as David Morgan hints) Dedoose and especially MaxQDA are more specifically designed for that sort of work. Dedoose has a one-month trial and you pay by the month, so it can be an affordable option*. MaxQDA has generous student licensing.

But assuming you're stuck with NVivo, or already know your way around it, then what you need to use are what are now (NVivo 10) called "Classification Sheets" (previously, as best I recall, it was called the "Casebook"). You can directly import Excel or tab/comma delimited TXT files. There is a pretty thorough help file at

http://help-nv10.qsrinternational.com/desktop/procedures/import_data_from_spreadsheets_and_text_files.htm

It's organized very much the same as the Import Wizard in Excel, except for the part where you mark fields as "classifying" (i.e. strictly quantitative, e.g. attributes such as age, gender) or "codable" (e.g. open-ended categories or responses).

It can be puzzling IMHO whether you represent your respondents in NVivo as "Sources", "Nodes" or "Cases". I use "Cases", and I think that's the "correct" choice of course, but I've seen others do it differently.

* There's been some publicity recently about researchers losing data in a system crash at Dedoose

http://chronicle.com/blogs/wiredcampus/hazards-of-the-cloud-data-storage-services-crash-sets-back-researchers/52571

Personally I think it should be a general warning to all of us about keeping secure and local backups. I've lost work through NVivo crashes - as well as through old-fashioned recklessness and foolishness ;-(

Having looked at the sample data, it is a combination of responses to closed-ended and open-ended questions that have specific responses. The way I personally handle this kind of data is to create a set of 1/0 variables for each OE questions, with each variable capturing whether a specific response was mentioned or not.

For example, lets say you ask respondents what their main sources of news are, and there are 4 common responses (television, internet, etc.), plus a scattering of less common responses (friends, etc.). Then you would create 5 variables -- one for each of the most common responses plus an "Other" variable. So, you would have a variable with a name such as TV and if the respondent mentions television, then this would be scored 1 (or 0 if television is not mentioned).

Overall, if the qualitative data were richer or not essentially "pre-coded", then you might consider a qualitative data analysis package, but with these rather straightforward responses (and the large N), I would definitely recommend SPSS.

OK, So I've decided to go with Dedoose as thus far it has the most straight forward tutorials and easily allows me to work on the data in multiple places since it's housed online.

As suggested in the resources: http://www.dedoose.com/resources/#, my survey data does need to be converted into an Excel file and as mentioned above the format needs to be horizontal not vertical.

I've tried importing a single .txt survey into SPSS and I'm seeing the option where I can transform the file to tab delimited, but was unsuccessful in doing so. I've also tried the .tpf file/template Sergiu provided and the attached screen shot is what I ended up with but am not sure what the next step is?

I've also determined that question 13 is irrelevant to the analysis as it is an open comment section where few participants bothered to write anything besides "no" or a simple comment that they enjoyed the survey, nothing that will really influence the data.

My advisor has suggested I simply bite the bullet and manually enter all the surveys into excel (follow the Dedoose examples) however I'm still open to giving SPSS a try if it can speed the process and help eliminate the chances of data entry errors.

The transpose feature in Excel is the most promising (in terms of manually cut-pasting): https://www.youtube.com/watch?v=37kwwEeFKvY but the numbers need to be deleted in each row/column. I've attached my preliminary spreadsheet format (similar to what dedoose needs) again if someone can suggest how I may speed the process or automate the survey input, that would be great.

Otherwise I'm simply going to go ahead with the manual input and use transpose, then find-replace the numbers with nothing. Once again I really appreciate the helpful responses here, thanks!

RG only allows one file attachment at a time, here is the shot of SPSS

SPSS:

Launch app > Open File>(on file type select All Files)>select your txt file>in the window that appears select YES (Does your text file match a predefined format?)>Browse>select te file i attached earlier>Finish.

After the data is loaded in SPSS you go to File>Save as> chose excel format

After you open the exl file:

You copy one column at a time. You select from A1 to A301 (if you have 300 cases), you select includin the header row and copy.

Then create a new sheet, go there and paste special and use transpose (there is a sign).

Repeat for all columns and transform them into rows.

That should take like 5 minutes for all variables and that's it.

Sorry Sergiu, I'm not understtanding your instructions, " A1 to A301. . ." if there is no data there what will this do? I still see no solution to avoid manual entry survey-by-survey and using the transpose for each file.