(SOLVED) / Request: Sort of nominal data in frequency tables

[stats2019]

Hi,

I notice, that data in frequency tables is not sorted. It would be great, if the user can check another box to sort those values.
When there are many values the display can get quite confusing and individual values hard to find.
Example:



Olaf

[jonathon]

so the frequencies table honours the order of the levels in the factor. simply reorder the levels of the factor, and all analyses will follow suit. of course i'm a little puzzled as to why these levels aren't in the correct order in the first place, but that will depend on the import procedure.

(if you double-click on the column header, you'll be able to reorder the levels there ... alternatively, if you make the data type of the column integer, i think the order will fix itself too ... but not 100% sure.)

jonathon

[stats2019]

Hi Jonathon,

thank you for the explanation how the order falls initially into place.
Also thank you for your second tip: I was actually able to leave the level at nominal but change the type to integer. This made the descriptives order ASC. Yout hint was gold anyways, as I know, my case (only numbers) is special, and people might have text that they want to order.
Result:


To solve the "puzzle": In this case the value was received automatically by the invitation link. Thus - if people come from place A they might have 8888, after that someone might come from place B and have 2222. If I understand you correct, as 8888 would be in a row above 2222 on import, it shows up above.

Olaf

[stats2019]

Hi Jonathan,

I have to add to this thread again. Sorting of values would be a great thing.
Example: I requested participants to choose from a text list. Now, that I get some descriptives of that field, the values (again) show up in a kind of random order.
I am attaching another image to demonstrate.


Olaf

[jonathon]

where did this data come from? i.e. what file did it come from?

if it came from a .csv file (which doesn't provide a level order) then they would be sorted alphabetically. if it came from a .sav file, then the order of levels provided by the .sav file would be honoured. i suspect the .sav file is at fault here?

jonathon

[jonathon]

actually, i'm not sure, we may be at fault here.

jonathon

[stats2019]

Hi Jonathon,

the behaviour is a bit strange.
My setting: I am asking what people do (whyHagen01). Depending on their answer to this, they get asked either whyHagen02_BA or whyHagen02_MA or even something else (whyHagen_else).
This results in different variables: whyHagen01, whyHagen02_BA, whyHagen02_MA and whyHagen_else.
I want to put those 02_ variables into a single column whyHagen_gemeinsam which then can be used for a descriptive for example.
Now data in whyHagen02_BA, whyHagen02_MA and whyHagen_else is scales nominal (but some have levels ... don't know why, see image), but my aggregation whyHagen_gemeinsam of course is nominal without any levels.
In the image below the descriptives of whyHagen_gemeinsam can be seen. I would assume, that as this is nominal, it should be ordered alphabetically.



Olaf

[jonathon]

a) so if there are level labels, we use the order provided by the .sav file -- i think this has to be the behaviour, otherwise all your agree, strongly agree, etc. wouldn't come through in the correct order.

b) however, if there's no level labels, and the variable is text, then i'd agree that's our responsibility to sort them alphabetically.

if it's b), can you send through a .sav file?

with thanks

[stats2019]

Hi Jonathon,

I think I found a bug (which shows itself e.g. in this case). I am emailing you two files with description.
Olaf

[stats2019]

Hi Jonathon,

I decided to post in the thread again, as I think it's still somehow the same issue.
In an analysis I notice, that the order if actually in the order of the occurence in the (calculated) column.
I suggest, that it is possible for the user to choose the sort order of the analysis:
alphabetcial, by count by percent, by cumulative value/percent
And ASC or DESC



Just taking the values as they come does not seem very useful.

Olaf