Aggregate codes and then recode

09.10.2018, 12:47

[follow-up of ]

I figured out the fastest way to aggregate frequencies:

1) Collapse the parent code
2) Doubleclick on it
3) Click on an entry and press CTRL+A on your keyboard
4) Autocode the results with a new code

Then you can right-click on this new code to transform the frequencies into document variables.


I tried this. See this image:

What I don't understand, is why there is a difference in the frequencies between the 'old' code (=collapsed subcodes) and the new code. For instance, after collapsing, the parent code ' Meta' has 157 occurences.
After following your instructions, and recoding these segments with the new code 'META' (purple), this has only 122 occurences instead of 157.

I don't understand where this differences comes from, since I selected all 157 segments and assigned the new code to them.

Version: MAXQDA 2018
System: Windows 10
Posts: 45
Joined: 25.06.2018, 12:39

Re: Aggregate codes and then recode

09.10.2018, 14:30

Aahr yes. Sorry, my mistake! I didn't think long enough. Most likely this is because there are different subcodes that are assigned to the same segment, or coded segments of different subcodes that overlap. In this case, transfering them all to the same code will merge those overlaps to being one coded segment...

This again could also be the desired result in many cases, maybe it also makes sense for your project? For example, if the parent code would be "emotional outbursts" and the subcodes "laughing" and "crying", and one instance is coded with both, it would make sense that it gets only counted once (one emotional outburst where someone is both laughing and crying).

Hm but if this logic doesn't apply for your project, then I'm also a bit lost.. but could it also be the other way around entirely? In the way that it makes sense to always aggregate them/count one instance of a phenomonen two or three times? In other words, would 157 be the correct number? Maybe it would help if you could describe a bit this category "Meta".. ?

If you should see the need to revise them manually, you could use the "Smart Coding Tool" to do so relatively quickly:

If it would make sense to aggregate them completely, I'll try to think of another workaround (can't see one right now though). I'll also ask the developers when I see them about how they want to tackle the whole aggregating-subcodes-features, I find it a bit hard to wrap my head around it right now..
MAXQDA Support Team
Andreas V.
Posts: 272
Joined: 13.04.2017, 16:23

Re: Aggregate codes and then recode

10.10.2018, 11:48

Hi Andreas,

Just thinking with you, I also don't have the answer.

In the function Subcode statistics, I can chose between 3 options for the unit of analysis:
1) documents - count all subcodes
2) documents - count most frequent subcode only
3) coded segments

I think there is a reason why this difference was made, so why 'coded segments' is another option than 'documents'.

After recoding on (only) parent level, the number (in this case 122) refers to coded segments. There are no subcodes anymore.

NB Feature request: an easier way to see how many documents these are, by using the right mouse click.

In my case, these 122 coded segments of META are from 108 documents.

In the old code 'Meta' using subcode statistics, I also can see that this is on 108 documents. In this case: 157 segments on 108 documents.

For my report I'm interested in the number of documents, so for me it's interesting that 108 documents (=respondents) mention the theme 'Meta'.
Posts: 45
Joined: 25.06.2018, 12:39

Return to Technical Questions

Who is online

Users browsing this forum: No registered users and 2 guests

We use cookies to improve your experience on our website. By clicking OK or by continuing to browse the website, we’ll assume that you are happy with their use. Click here to review our Cookie Policy. OK