- Posts: 5
- Thank you received: 0
Welcome to the LimeSurvey Community Forum
Ask the community, share ideas, and connect with other LimeSurvey users!
Statistics multiple Selectors
- georgfischer
-
Topic Author
- Offline
- New Member
-
I'm new to limesurvey and I have a hopefully simple question:
I've created a survey and now I want to see the results. If I mark 1 possible answer everything works but if I choose 2 then the response is empty. So 1 item marked in G01Q01 works but another one in addition in G01Q02 doesnt.
What can I do?
Thanks for your help!
Georg
your first information always has to be: the version of LimeSurvey you are using.
Seeing your question codes I suppose it is 4.x., but which exactly.
Well, I tested in 4.4.16. and I do not see an issue.
We have only two questions with three answer options.
Without filter the statistic is like this:
If I set a first filter
only the five responses are displayed with Q1==1
Now I set the second filter
And as expected only these responses are displayed (Q1==1 AND Q2==2)
Of course if I select the combination (Q1==1 AND Q2=

This is not existing in the dataset.
So you may explain your issue a bit more detailled, show some screenshots, or send the lss export of this your test survey.
Joffm
Volunteers are not paid.
Not because they are worthless, but because they are priceless
- georgfischer
-
Topic Author
- Offline
- New Member
-
- Posts: 5
- Thank you received: 0
my version is 4.5.2.
Im sorry I can't post screenshots as there might be sensible data in it.
If I filter all questions with one answer or some answers in the same question it works. But I need all results for 2 or more filters in 2 or more questions.
For example I try to explain the logic:
q1 = are you m,f or d
q2 = age 0-18, 18-25, 25-40
if I filter Q1 m and f its ok
if I filter Q2 0-18 its ok
if I filter Q1 m and Q2 18-25 its empty
How can I send the los to you?
Joffm
By the way:
Nobody uses this built in "statistic" to analyse data. This is only to get a rough overview during fieldwork.
Data analysis you'd better do in a statistic tool.
Volunteers are not paid.
Not because they are worthless, but because they are priceless
- georgfischer
-
Topic Author
- Offline
- New Member
-
- Posts: 5
- Thank you received: 0
Which statistic tool do you suggest? The one from IBM is way to expensive. We are OK with the buildin statistics.. except for the one thing..
So here we need all answers regarding one location "Standort" so we can make a printout per location and not per group
in which way is this related to your first question?or example I try to explain the logic:
q1 = are you m,f or d
q2 = age 0-18, 18-25, 25-40
if I filter Q1 m and f its ok
if I filter Q2 0-18 its ok
if I filter Q1 m and Q2 18-25 its empty
So here we need all answers regarding one location "Standort"
Do you try to display all responses which have
"Geretsried Gartenberg" either in G01Q10 or in G01Q11.
This obviously is not possible. No child is in "Krippe Geretsried Gartenberg" and in "Kindergarten Geretsried Gartenberg" at the same time.
This filter is an AND filter not OR.
To be able to do this you'd have structured the questions differently.
Like
G01Q09
Hidden Multiple question with all locations.
Equation that sets the valid locations depending on the selection of G01Q09
G01Q10 with all locations as answer options and this hidden multiple as array filter .
Or do you want to display all responses of "Kindergarten Geretsried Gartenberg", in total, not "Hamster", "Maulwürfe", ... separately?
This should be possible by "G01Q11==AO02"
But if you want to display all reponses of Kindergarten Geretsried Gartenberg " or "Geretsried Künnekeweg" you may select both.
In this case it is OR
As you see, I am not quite sure what exactly you want to display.
Joffm
Volunteers are not paid.
Not because they are worthless, but because they are priceless
- georgfischer
-
Topic Author
- Offline
- New Member
-
- Posts: 5
- Thank you received: 0
so I need a or and not a and..
no chance to get it?
the survey is done and I can’t change the questions anymore
Well, there is the free SPSS clone called PSPP. A bit outdated, but usually sufficient.Which statistic tool do you suggest? The one from IBM is way to expensive
Then there are "Jasp" and "Jamovi", "R" , ... (all free)
But if you do not expect more than this built in statistics, you may use EXCEL or OpenOffice.
As you have scales in your survey, you will calculate some differences between groups (t-Test, ANOVA)
Unfortunately you coded the answer options with the alphanumerical codes "AO01", "AO02".
So before you are able to calculate a mean you have to recode all.
The mean of "AO01" and "AO04" is not "AO02,5"
Conclusion: You did not test your survey completely.
Joffm
Volunteers are not paid.
Not because they are worthless, but because they are priceless
I would be surprised if 50% of the LimeSurvey userbase is using an external statistic tool. I see research projects with closed aource saas tools which do everything inside the platform. PowerPoint and Excel are available, but no stat tool. LimeSurvey GmbH is aiming for these kind of users.Nobody uses this built in "statistic" to analyse data. This is only to get a rough overview during fieldwork.Data analysis you'd better do in a statistic tool.
The meaning of the word "stable" for users
www.limesurvey.org/forum/development/117...ord-stable-for-users
- georgfischer
-
Topic Author
- Offline
- New Member
-
- Posts: 5
- Thank you received: 0
Is there a way to make a copy of the survey and change the answers via database. So for example move everything related to the location "geretsried gartenberg" to the first question and I select that and can at least check the answers for all childgroups?
if I filter Q1 m and Q2 18-25 its empty
This might sound obvious, but do you have any respondents in your sample that meet both criteria?
So do you have any male respondents that are between 18-25 years old?
I answer at the LimeSurvey forum in my spare time, I'm not a LimeSurvey GmbH employee.
No support via private message.
If you add
Bad Tölz Peter-Freisl-Straße
Lenggries
Pullach
to Krippenstandort, all possible location are there in this question.
Then you may run some queries in phpmyadmin or whatever you use.
Something like
"UPDATE `lime_survey_469178` SET `469178X83X1193`="AO05" WHERE `469178X83X1194`="AO01"
to set "Geretsried Künnekeweg" to this now general question.
"UPDATE `lime_survey_469178` SET `469178X83X1193`="AO06" WHERE `469178X83X1194`="AO02"
for Geretsried Gartenberg
and so on.
Of course you have to adapt the SQQ-Code and the surveyID.
But in my opinion it is much easier to export to EXCEL and do everything there
As PSPP is a clone of SPSS 14 (I think; have a look at their website) you have to use this LS-Export (SPSS before 16)OK. I tried to figure out how pspp works but I dont get it done.
Joffm
@holch,
That was only a - bad - example.if I filter Q1 m and Q2 18-25 its empty
But the real situation in the survey was indeed, that this condition never could be TRUE.
No child is in "Krippe Geretsried Gartenberg" and in "Kindergarten Geretsried Gartenberg" at the same time.
Volunteers are not paid.
Not because they are worthless, but because they are priceless