Study freezes upon branching

[jwright8]

I have branching in my study. Roughly 5-10% of the time, during a preview or when a participant is doing the study, when either one of the branching choices is chosen, the study freezes and does not continue at all. This is fairly problematic. I suppose it is happening for other researchers as well.

[Ting]

Thank you for the note. We are looking into this issue.

[CarsonGMR]

I thought I'd mention that I've noticed the same issue occurring as well. Freezing seems to occur at a similar rate, and occurs just prior to when it should advance to the first slide of a branching path (you click the continue button but it freezes on the screen just before branching). In my case, the same branch point splits off based on counterbalancing group and a participant response. When checking in preview mode using one device, it seems like freezing occurs at a similar rate for either group and either response option.

Any suggestions on possible work arounds would be great!

[Ting]

Hi Folks,

We have an update on this. We have recently (made effective this past Saturday) improved the mechanisms of conditional branching. First of all, please allow me to explain why conditional branching was a bit unreliable.

Conditional branching, by definition, means that upcoming study content needs to be retrieved dynamically based on participants' past responses. This means that at the point of branching, the browser needs to make a trip back to the server to retrieve additional content. This is the point of failure, as a tiny network hiccup would interfere with the retrieval process.

We have now since made two improvements. The first one is visual, in the area of user experience: when the retrieval process is happening, FF now displays a message "Retrieving additional study content..." with a spinning wheel. This will hopefully make participants wait patiently for the upcoming study content. Second is the underlying retrieval mechanism: the retrieval of study content now retries up to 10 times. Before this update, the retrieval happened exactly once, and thus had no tolerance of network failure at all. Under the new setup, it'd be much more robust, although occasionally, it'll take a little bit of waiting on the part of participants.

Please keep us updated on whether this makes a difference in your studies. We are eager to hear more feedback. Thanks!

[CarsonGMR]

[quote pid="237" dateline="1613658697"]
Hi Ting,

Excellent -- thanks for figuring that out!

I'll update if I hear issues come up when the study is released to a larger participant pool, but at least in provisionally testing, it seems to have worked perfectly. I see clear occasions where the 'Retrieving additional content...' message is up maybe 10s before continuing on, and those seem to occur at the same frequency as I would previously have freezing on this system. A good sign!

[/quote]

[Ting]

This is fantastic to hear! Thanks! We are also upgrading our network infrastructure this weekend so hopefully that'll address the root cause of these network hiccups.