Can docx and PDF documents be analysed in one project?

[Peng Hua]

Hi there.

I wish I can explain my problem clearly.

I have finished my analysis of data I collected in 2015-2016 in MAXQDA 12 and then I bought a new computer and upgraded to MAXQDA 2018. Unfortunately the data I collected in 2017 cannot be imported into MAXQDA in my new computer if they are in word file. So I decided to save 2017 data as PDF or rtf so that they can be imported. What worries me is if the coding of 2017 data in the form of PDF can be compatible with previous coding of 2015-2016 data in the form of word. For example, if I decide to view code segments, will segments from word files pop up with segments from PDF in the same matrix? If I want to count the number of a particular code that appears in three years from 2015 to 2017, will it work？

With thanks for your help.

Version: MAXQDA 2018
System: Windows 10

[Andreas II.]

Dear Peng Hua,

thank you for your question, and welcome to the MAXQDA forum!

I'm not entirely sure if I understand you correctly, but I believe the answer is yes: The file type is completely irrelevant in regard to counting codes etc. The only main difference between rtf/Word files and PDF (beside the fact that you can't edit the latter one) are that paragraphs are not recognized within PDF files. So you will not be able to use any functions that depend upon the recognition of paragraphs within PDF files (like the "near"-complex coding query).

Beside this, I don't think you will have any trouble. You will simply have to create a Document Variable "Year" and enter the values in the data editor for all of your documents. Then you can use Mixed Methods functions (like Crosstab) to see which codes occur in documents with "[variable]Year=[value]2017".

I hope this helps! If not, don't hesitate to ask again.

Kind regards,

Andreas

[Peng Hua]

Dear Andreas,

Thank you very much for your reply.

Since I need the function of recognition of paragraphs, I cannot save data files as PDF. I will use rtf instead.

Another question is why I cannot open MAXQDA 12 project in MAXQDA 2018? I once asked about it (reference number: 128624129) and I was told as such "You can open all your old projects with MAXQDA 2018.l The projects are saved on your device and not in the MAXQDA program“。

With great thanks and I am looking forward to your reply.

[Andreas II.]

Dear Peng Hua,

thanks for your feedback! Using .rtf sounds like a good idea. It gives you much more options to work with. And you do not have to worry about the quality of "Optical Character Recognition" (OCR), or about "DRM", so it is almost always preferable to use RTF instead of PDF, if there is a choice.

To open a MAXQDA 12 project-file (it has the ending .mx12) is very simple: just go to the tab "Home", and select "Open Projects". At the bottom right, you can choose .mx18-project files (MAXQDA 2018) or .mx12-project files (MAXQDA 12).

Kind regards,

Andreas

[Peng Hua]

Dear Andeas,

Thank you very much for your prompt reply. I saved the file as rtf and successfully imported into the new project and it worked!
I also learned to open my old MAXQDA files in the 2018 version of the software.
I wonder if I may ask a new question about intercoder reliability.
You know when we code text data, we have to report intercoder reliability if we want to get our findings published. The second coder usually code part rather than the whole set of the text data. I wonder if the autocoding function in MAXQDA could help in acting as a virtual coder when compared with a researcher who manually code the entire set. To what extent can we trust autocoding results? Can we establish an acceptable intercoder reliabilty by comparing manual coding and autocoding?
With great thanks again and I am looking forward to your reply.

[Andreas II.]

Dear Peng Hua,

Sure, I'm happy to help!

Automatic coding will enable you to automatically code the results of search queries. I think it's very useful if you have a large text corpus and you want to make sure that you find all the relevant passages concerning a specific topic. However, I would not compare the results of such a query to the coding of a human researcher, because at this level of complexity (is a word/word combination existing or not) the computer will always be superior and should simply be trusted.

Intercoder reliability is devised for coding that usually requires a more complex interpretation, factoring in different perspectives, authors, contexts, meanings, assumptions etc. Usually, coders get together before and during coding to define and consolidate codes, to have a shared understanding of what they mean and when they are to be applied. Intercoder reliablity will then show how well those definitions and consolidations have worked. In short: Intercoder reliablity is used for comparing a coding which most people (including myself) still consider to be a task that can only be performed by trained scientists.

Kind regards,

Andreas