Hello,

I have been completing a large-scale data collection project over the past few months. I have several versions of a similar survey that have been running beautifully. There are three parts to each survey version, so Part 1 is directed via end URL to Part 2, and Part 2 is directed via end URL to Part 3.

After hundreds of successfully completed surveys and no changes in settings on my end, Part 2 is no longer linking to Part 3. Instead, participants are caught in an endless Part 2 loop, and they report that their answers are deleted each time. This report is confirmed because there are no answers showing in their response table. This issue is happening in multiple versions of the survey, so I know it is not linked to one particular version's settings. I have also checked the end URL to make sure that it is accurate. It does contain the suffix &var1={SAVEDID}&var2={TOKEN}, but that has been part of the end URL all along through hundreds of successful redirects. There appear to be no issues with Part 1 successfully linking to Part 2.

Does anyone have any idea why this issue would surface after functioning successfully for so long?

Thank you.

Hmm, if there has been no change on your side, usually software doesn't just change things.

When you test it, what happens?

Difficult to help without any further information.

Any updates lately? (server or Limesurvey)

Thank you for answering. When participants on various computers in various locations complete their responses and click "submit," the survey is cycling back to the first question and deleting their responses. When I test the survey, the same is happening.

These surveys have been working well for months with no changes to the settings. I am not sure about recent LimeSurvey updates...

To try and isolate the problem, I exported the survey and created a new one, removing all connecting URLs. The survey will still not submit and move on to the end message. The same happens when I try to recreate the survey by importing question by question instead of the whole survey. I should note that this is particular survey is an array question with conditions.

The same happens when I try to recreate the survey by importing question by question instead of the whole survey. I should note that this is particular survey is an array question with conditions.

What version of LimeSurvey is used?  What was the result of importing question by question? If you create a new survey with one textdisplay question only, can you reproduce the issue?

Another source of issue are ongoing browser updates. Using older or buggy LimeSurvey versions, you sometimes see such issues occurring after browser updates. So is Firefox and Chrome making a difference?

 

All great questions. The version is 3.25.21, which I believe has remained the same throughout the time that I have been running the survey. If I create a new survey and import the questions from the original version, the problem persists, but I have many other surveys that are currently running just fine. This was a particularly complicated one in that the answer to a dropdown question (1-20) can produce one of 20 different arrays, depending on the answer given. The problem exists in all browsers on multiple devices, so I suspect some sort of recent change on LimeSurvey's end might be at work. If that is the case, and I spend the time needed to try and recreate the survey from scratch using the same questions and conditions, is it likely to work? Do you have any other suggestions?

If I create a new survey and import the questions from the original version, the problem persists, but I have many other surveys that are currently running just fine.

It still can be an issue, which only occurs with a certain combination of survey, theme, JavaScript and browser versions.
Instead of recreating the survey from scratch, it's often easier to copy the survey and delete questions till the issue vanishes. Depending on the survey structure this can be an easy or an hard task.
I recommend to use an unmodified theme for the testing.If you have isolated the problematic question, you might recreate that question.