With the great additions to make computations based on previous responses, the question may arise: when is a survey completed?

For example: when answering quiz-like questions and reaching a last group (or groups...) where the results are being shown in text display questions, you may consider a survey to be completely filled in.

Can we imagine to add a feature that gives the survey designer the possibility to state at what point a survey is completed?

And: when the users reaches this featured end point, the date & time is written in the database.

tammo wrote:

With the great additions to make computations based on previous responses, the question may arise: when is a survey completed?
 

What addtions do you mean?   Just to understand the motivation.

Even under older LimeSurvey versions there is a gap which could be filled by a terminate question type.
This questiontype could be placed anywhere and would come with a few options.

First, the questiontype would write a Statuscode/Depositioncode into the response set.
Typically you numbers with certain meanings in survey tools.

Complete/Qualified
Disqualifed
QuotaFull
etc.

That way you could easily see, why a response set it like it is.
Beside the status code, you would allow to define a another code, which would be saved into the response set as the answer to the termination question.
Another option needed, you be the destination of the termination question.  Endpage of survey or external URL
 

I meant the additions like:
www.limesurvey.org/manual/StatFunctions
 

I use functions like these to make up a last page like here:
research.respondage.com/index.php/774648...t=Y&lang=nl-informal
But when the user is here, she does not proceed, thinks this is ready now.

I really think LimeSurvey core must allow submitted survey when "Submit" is done …

Maybe we can add a new function
Some think like this

statCountIfCondition(QuestionCode.sgqa, value[, 'submitdate is not null']);

The case Tammo described seems not limited to statCount usage.The case that respondents leave before that last page is common in many survey situations.
To extend the statCountIfCondition with a way to set a submit-date (if that is what the draft should tell us), seems to be too narrow.

The purpose to end a survey at more than one position in the survey is valid. That's why other survey tools offer a question type for that.
Another generic way would be to have an advanced option in every questiontype to be the endpage itself (with or without displaying the content of the endpage below the question content.
A special question type is easier to understand, cause it can be seen in the structure list. And it would replace the missing jumpto functionality in many situations, where you want to route respondents directly to the end of the survey.

@Tammo: Why not open a feature request based on what you got?

@Jelo: you are right, it is not limited to the statCount functions, that was an example.

Your idea to have a special question type, something like: Survey End question which writes in the database the submitdate and defines the survey as finished, even when you are not on the end page (group) would be something great. After which more groups & questions can be defined, but these question are "bonus".

I can make this into a feature request, but I started the discussion here, because I wanted to hear multiple voices. I leave this open for now, will consider writing a feature request in a week or so, because the discussion may not be over yet. I think it is better that real users discuss this feature before to throw it in the developers arena... And yes, some real users are also developers...

tammo wrote:

Your idea to have a special question type, something like: Survey End question which writes in the database the submitdate and defines the survey as finished, even when you are not on the end page (group) would be something great. After which more groups & questions can be defined, but these question are "bonus".

 

Warning : for limesurvey (without allow edit) : this disable save response if you use submitdate.
A survey completed can not be edited again.

Question : why not use an equation question type ? And the feature are more : export filtered by + admin statistics filtered by + public statistics filtered by etc … ?
 

jelo wrote:

The purpose to end a survey at more than one position in the survey is valid. That's why other survey tools offer a question type for that.
 

How it's used ? In GUI ? When it's used exactly ?
 

DenisChenu wrote:

Warning : for limesurvey (without allow edit) : this disable save response if you use submitdate.
A survey completed can not be edited again.

Question : why not use an equation question type ? And the feature are more : export filtered by + admin statistics filtered by + public statistics filtered by etc … ?

I had not thought of that warning. I never use "save the response and resume later" functionality, since my experience is that most (all?) people that use this option never come back to finish the survey. But that would indeed be a consequence: once it is submitted ( = once you have passed the submitted point), you will not be able to save and resume.

DenisChenu wrote:

jelo wrote:

The purpose to end a survey at more than one position in the survey is valid. That's why other survey tools offer a question type for that.

 

How it's used ? In GUI ? When it's used exactly ?

 

No design has been made yet, I guess.. But I see the point: submitdate for completely finishing the survey and a submitdate2 for a finish somewhere at a point defined by the survey designer. Plus a way of indicating when a survey is considered to be completed so it will be counted as completed.

About GUI : i ask it to @jelo , i really don't have idea on how it work for Survey manager .

About column name : 'completed'

Tammo "only" wants to ensure that LimeSurvey marks a completed survey, even tough the respondent hasn't hit the submit button on the last page.

If we want to just solve that issue, which many of us have experienced, the most direct solution would be to allow every question type to set a termination code in the advanced settings.

Better seems to be to place that in a separate question type (make more questionstypes with less options instead of less question types with more options). The equation questiontype might be not overloaded with options.

If you place the termination/complete question in a question group (BTW: Wasn't LS5 using page instead of group in the GUI. When was that reverted?), the survey response get's the completion code of that termination/complete question.

The advanced version of that termination/completion question type would contain endpage text and would allow to redirect to a different URL. The question type could be used to only mark complete status and to mark complete status and allow showing an endpage or a autourl redirection.

That way you could save a lot of time, when you have different respondent groups in your survey which only have to fill out parts of the survey, but should not be screened out via the quota workaround.

Sometimes you see surveytools using not "complete", but qualified/disqualified status codes in the response set. E.g. if a part of the survey is rushed (timer), the coding will set the responseset to disqualified even though the responseset is completed.

If we consider the equation question as a way to start and don't want to invest time in GUI work, wouldn't be the first stept to have ExpressionScript functions which allow to set the complete status. Wouldn't that be the minimal project, which still is huge.

In a nutshell: Place a equation question in a question group which is considered as an end of the survey. The ExpressionScript function to set the status of the survey to complete is used in the equation question. The big unknown is, how the core functions will react and what routines this new ExpressionScript function has to trigger to really end the survey. It doesn't look like an low hanging fruit.




an status field (perhaps an additional field, till we can be sure it has no negative impact on the basic functions) to qualified/disqualified

> In a nutshell: Place a equation question in a question group

Alternative :
Place a short text hidden question in 1st group (qualified)
Place some equation to set qualified according to expression