How to link respondent ID to specific codes

[JSP]

Hello,

I am trying to help a colleague who came to me with a rather daunting problem.

She recently purchased MAXQDA and has no prior experience with the software (i.e., on the steep end of the learning curve—and I am in the same boat, btw).

She is analyzing comment questions from a survey. For each code she assigns to an idea within a passage she wants to also link that code / passage to the survey ID so she can ultimately group frequency of occurrence by demographic variables.

Can anyone provide advice on how to go about this? She has tried several things, to include creating a code for the ID that is linked that to the same passage as the idea code (which is too cumbersome to be of use).

Here is a simple example of the final product:

Male Female

Theme 1 A. List of ID’s who A. List of ID’s who
mention this theme mention this theme
B. Count of ID’s that B. Count of ID’s that
mention this theme mention this theme


Part B is easy, it is part A she is struggling with.


Thanks,

J

[Stefan Rädiker]

Hi J,

the answer to your question depends on how the project has been build up and in which format the survey data is available. Anyhow, we try to answer it in a way, that is independent from this factors. This makes the explanation a little bit longer.

To get the results you and your colleague are looking for, you could use the "Attributes" in MAXQDA: For each document you can save additional quantitative information in variables, which are managed like in statistical programmes as SPSS. Each row in the Attributes table represents a document/case and each column represents a variable.

To import the information male/female and the id for each answer (=document) into this table follow these steps:

• Choose "Attributes > Export data matrix" from the menue and open the TXT file with Excel.
• Now you have the required structure for importing variables into MAXQDA and you can add two columns: one is for the ID (if you don't use it as the documents name already) and one is for gender with categories "female/male" or "1/2".
• Save the file in the original TXT format, close EXCEL and choose "Attributes > Import" from the menue to import the information.

A) List of ID's who mention this theme
Afer importing the There are more than one possible solutions. How about this one, which can handle lots of codes/themes at once, if you like:

• Click on the fifth symbol "Activation by Attributes"
• Click on "Reset" and then choose "New"
• Choose the Attribute containing the information female/male and click okay. Add a formula like "gender = female" using the buttons and drop down menues.
• Activate one or more codes by holding STRG and clicking on them
• MAXQDA presents all retrieved segments of the female persons who wrote something to the selected themes.
• Choose "Analysis > Overview of retrieved segments" and you get a table which you can sort by id and which you can export as you like.

B) Count of ID's who mention this theme

• Activate one or more codes by holding STRG and clicking on them
• Choose "Visual Tools > Crosstabs"
• Click on "Reset" and then choose "New"
• Choose the Attribute containing the information female/male and click the option "Insert all values into the table"
• After clicking "OK" MAXQDA presents a cross table with the code frequencies differentiated for female/male

Please refer to the manual for additional informations to the mentioned functions.

Hope, this helps.
Stefan

[JSP]

Thank you!

[JSP]

I am working through your instructions and have a quesiton. What is the best way to include the attribute I want to summarize in the original data file? Should I preceed it with the attribute's name? For example, Gender: Female?

Thanks,
John

[Stefan Rädiker]

John,

MAXQDA doesn't use labels like statisic-programs normally do. If in the original file the variable was called gender with the values 1=female and 2=male and you import only the numbers in MAXQDA, this could be confusing, because you have to remind what 1 and 2 stand for.

Thus the best way normally is to use variable names and value labels which explain theirselves.

You could do it like you suggest:
Variable name: Gender: female
Variable type: Boolean (means yes or no)
Usage: yes for female

Alternatively, this should work to, and could be a little bit easier to work with, especially if you a variable with more values:
Variable name: Gender
Variable type: String
Using: m for male f for female

Hope, I have understood your question correctly
Stefan

[JSP]

Stefan,

I believe I am making this harder than it needs to be. Let me ask a basic question: Is it possible to assign an different attributes to elements within a document?

The data we are working with are comments from a survey. Thus most passages are short (longest is about 10 lines) and I have all the responses in a single .rtf file. I can include the demographic variables in this file, too. So I can create a dat afile that looks like this:

SurveyID: 1 Gender:M response to comment question 1

SurveyID: 2 Gender:F response to comment question 1

etc.

Is this the right way to prepare the data file?

Thank you,
John

[Stefan Rädiker]

John,

is it possible, that you send me the status of your actual project and the rtf/dat file with the comments?

I think it could be helpful, if we have a look on it, because otherwise we give you tipps and hints, that don't fit to your situation.

stefan.raediker (at) maxqda . de

Stefan