How long to allow for online survey data collection?

[futureapppsy2]

I'm heading up an exploratory survey and was wondering if anyone knew of any guidelines on how long to allow for data collection. Because there is very little empirical data on this specific group, I don't even have a good sense of how big the target population really is, so I can't really based it on a target n. In my ancedotal experience, people usually complete online surveys when they first get them or not at all, with a slight bump when reminders are sent. Tentatively, I'm thinking about sending a reminder at the two week mark and then allowing another two weeks before closing data collection, but does anyone know if there are general standards or even empirically- based time frames for this type of thing?

Thanks!

---

Members don't see this ad.

 

[Ollie123]

Nowhere near enough information to answer. Is this of undergrads in the psych department? Are you going through MTurk or something similar? Offering incentives? You may not know exact prevalence, but presumably you still have some sense of whether we're dealing with something at least somewhat common (e.g. undergrads who experienced a recent loss) or something wacky (e.g. left-handed, gay ornithologists with COPD).

We always just do a power calc and recruit til we get a sufficient sample for what we need, but this is meant to be population-based descriptive you probably just want as many as you can get. Depending on how you are going about this it might make sense to do one week or 3 years, it just depends on your goals, etc. I'm not sure how a general empirical guide for this sort of thing could exist.

 

[futureapppsy2]

Nowhere near enough information to answer. Is this of undergrads in the psych department? Are you going through MTurk or something similar? Offering incentives? You may not know exact prevalence, but presumably you still have some sense of whether we're dealing with something at least somewhat common (e.g. undergrads who experienced a recent loss) or something wacky (e.g. left-handed, gay ornithologists with COPD).

We always just do a power calc and recruit til we get a sufficient sample for what we need, but this is meant to be population-based descriptive you probably just want as many as you can get. Depending on how you are going about this it might make sense to do one week or 3 years, it just depends on your goals, etc. I'm not sure how a general empirical guide for this sort of thing could exist.

We're targeting a certain type of health care provider and professional students in that field, so we're going through profession-specific listservs. We're looking at a specific minority population within this profession and even the info we can kind of sort of get from the limited publicly available demographics is somewhat distorted by how the data presented and collected, so we don't have much of a guide. Our primary aims are descriptive (like I said, there's really nothing out there,) so a power analysis wouldn't be applicable.