Error in time-based triggers?

Support
1

In my study (1280), I’ve run into a problem where participants are reporting that they’ve received triggers for surveys after the time period in which they’re supposed to be prompted. Participants complete an evening survey each day that is supposed to be triggered at a random time between 7:30 pm and 11:00 pm, but the surveys have been triggered later than 11:00 pm. For example, in survey 9515, for participant 36773, last night’s evening survey (March 20) triggered at 11:57 pm. And the error appears to affect coming surveys as well: the evening survey for March 22 (9517) is scheduled to trigger (according to the “sessions” page) at 11:29 pm, and the evening survey for March 26 (9521) is scheduled to trigger at 11:52 pm. Why are these surveys scheduled to appear after the end of the triggering period? Any help would be appreciated!

2

Hi @pluehrin

It seems there is a bug in our time calculation when the schedules cross the daylight saving dates, which is been the case here. My colleagues will look into this and will try to identify and fix it, in the meantime, I refreshed the sessions you mentioned above.

If you see any other discrepancy, please let us know.

Thanks a lot,
Mohammad

3

Hi Mohammad, thanks for your quick reply! I suspect that this bug is affecting all of the participants currently enrolled in the study (about 15 or so), so would it be helpful for me to give you their ID numbers so that you can fix the timings for their upcoming surveys? Or do you think your fix will be completed soon?

Thanks,
Peter

4

Hi Peter,

The issue should be resolved for your study participants. Please let me know if there is any other problem.

Thanks,
Mohammad

5

Hi Mohammad,

Thank you for your help!

All best,
Peter

6

Hi Mohammad,

I have one more question for you re: the time-based triggers. I was looking through the scheduled trigger times for my participants, and it seems that the vast majority of the triggers are skewed toward the later end of the time window (7:30 pm - 11:00 pm), with no triggers scheduled for many of the participants between 7:30 pm and 8:45 pm, and the most coming after 9:30 pm/10:00 pm. I know that it’s possible that this is just how the random assignment worked out, but across 50 or 60 triggers there should be more variety in the trigger times across the whole of the 7:30 - 11:00 pm time window. Can you check to see if there’s any reason why earlier triggers aren’t being assigned? It’s important for my study to have a good variety of trigger times, and right now everything is skewed toward later trigger times.

Thanks,
Peter

7

Hi @m.hashemian,

Unfortunately, it doesn’t seem like the daylight savings issue has been fixed. The same participant who I wrote about originally, 36773, received his survey after 11 pm last night, and looking at the future schedule, several of the surveys are scheduled to start after 11 pm. Can you take another look at the issue?

Thanks,
Peter

8

Hi @pluehrin

Just to give you an update, my colleagues are working on this, but a fix can take a bit time. As I noticed, most of your participants are almost done with their participation and will be done by the end of the March or early April, the fix might not arrive and get deployed in time.

I also tried to shift the times back by one hour for those that happened to fall after 11 pm, though that can also cause other problems, for example, if the participant reloads their settings, all my manual changes get discarded. So I guess the best is to resolve the main issue (daylight saving handling).

Thanks,
Mohammad