Recruiting for research studies online

[psychanator]

Hi all,

As I am sure many who are on this site have had to recruit for studies, I am looking for suggestions. Our cohort is trying to brainstorm new ways to recruit for research participants - free or paid (free is ideal, but we would consider paid). Does anyone have suggestions of sites or services they would recommend?

Thanks!

---

Members don't see this ad.

 

[deleted343839]

Have you considered Amazon MTurk? I haven't used it myself but the consensus seems to be cautiously optimistic.

http://pps.sagepub.com/content/6/1/3.short

 

[MCParent]

What kinds of participants?

 

[Justanothergrad]

Have you considered Amazon MTurk? I haven't used it myself but the consensus seems to be cautiously optimistic.

http://pps.sagepub.com/content/6/1/3.short

+1, much cheaper option than qualtrics.

However, it definitely gets the recruitment done quickly (I usually get a few hundred done in about an hour) but can be a bit of a pain with high drop out if you want more than a single shot. Its tough to do follow-up as there isn't anyway to 'call back' participants. It all depends on what your needs are. Also, people should be warned that there is no way to get in contact with customer support (they seldom answer emails and there is no phone contact) so if you run into an issue, it can be very frustrating. And the system itself seems purposefully designed to be obtuse.

http://research.microsoft.com/en-us/um/people/nath/infographics-study/articles/mturkmethods.pdf

 

[WisNeuro]

Make sure you put in some validity checks in them there online questionnaires. I've done this in the long, long ago. Had to throw out almost 10% of the data due to validity concerns.

 

[bmedclinic]

I bet a lot of us get these- but just about weekly I get some "brief survey for psychologists in the field" to fill out and it says "5 minutes" and takes about 30. And I'm a pretty darn efficient person. If you want 30 minutes of my time, a) I dont have that to give and b) you should at least say that up front. That's all I got.

Also what I wrote above should be standard, so I dont know why recently I'm getting 5 minute surveys that clearly take much more time than that.

 

[Ollie123]

I bet a lot of us get these- but just about weekly I get some "brief survey for psychologists in the field" to fill out and it says "5 minutes" and takes about 30. Also what I wrote above should be standard, so I dont know why recently I'm getting 5 minute surveys that clearly take much more time than that.

Near the tail end of graduate school I got tons of these. Most were for student theses/dissertations that almost invariably will not make it into the literature, are comprised of poorly written questionnaires that likely haven't been validated and are usually being done for some vague purpose related to either multiculturalism or supervision experiences despite having few actionable and well-written items on those topics.

I will take a look if they are providing compensation of some form (even a drawing) since that shows they are at least somewhat invested or if they have established investigators in training or there is evidence to suggest it is an actual worthwhile study. Most are not and it just seems a way to run cheap studies and push through students who probably shouldn't be getting degrees.

 

[WisNeuro]

I will take a look if they are providing compensation of some form (even a drawing) since that shows they are at least somewhat invested or if they have established investigators in training or there is evidence to suggest it is an actual worthwhile study. Most are not and it just seems a way to run cheap studies and push through students who probably shouldn't be getting degrees.

Ding ding ding. Exactly. As soon as I get to terribly worded questions that an even an undergrad in stats would know better than to write, I leave the survey early.

 

[MCParent]

Ding ding ding. Exactly. As soon as I get to terribly worded questions that an even an undergrad in stats would know better than to write, I leave the survey early.

I usually do the list serve ones that are for psychologists. I'll even give a shot to schools that I think are not especially reputable. But I also drop out the second I see stupid items, "transgender" listed as an option for sexual orientation in the demographics, or whatever.

Sometimes on psych listserves I see requests for participation in studies on random, general topics. I literally cannot think of a worse methodology short of "I got some of my friends to do it" (a.k.a. snowballing where the snowball is made of **** and is growing into a larger ball of ****).

 

[WisNeuro]

Sometimes on psych listserves I see requests for participation in studies on random, general topics. I literally cannot think of a worse methodology short of "I got some of my friends to do it" (a.k.a. snowballing where the snowball is made of **** and is growing into a larger ball of ****).

Yeah, listserv studies are terrible, unless the listserv population is specifically what you need. Such as our salary and practice surveys for neuropsych.

 

[EmotRegulation]

+1, much cheaper option than qualtrics.

However, it definitely gets the recruitment done quickly (I usually get a few hundred done in about an hour) but can be a bit of a pain with high drop out if you want more than a single shot. Its tough to do follow-up as there isn't anyway to 'call back' participants. It all depends on what your needs are. Also, people should be warned that there is no way to get in contact with customer support (they seldom answer emails and there is no phone contact) so if you run into an issue, it can be very frustrating. And the system itself seems purposefully designed to be obtuse.

I've used mTurk quite a bit, and there IS a way to do followup. You can contact people via email, it's a bit of a pain in the ass but absolutely doable, and now waaaaay easier using R. I've done small big surveys and then brought people back who meet my inclusion criteria.

I too get a few hundred people within a few hours, but there are a lot of turkers out there who are (rightfully) demanding reasonable pay (like equal to minimum wage).

Also, the fees are ridiculous now. Supposedly there is a way around that one too, but I haven't explored that much yet.

 

[Justanothergrad]

I've used mTurk quite a bit, and there IS a way to do followup. You can contact people via email, it's a bit of a pain in the ass but absolutely doable, and now waaaaay easier using R. I've done small big surveys and then brought people back who meet my inclusion criteria.

I too get a few hundred people within a few hours, but there are a lot of turkers out there who are (rightfully) demanding reasonable pay (like equal to minimum wage).

Also, the fees are ridiculous now. Supposedly there is a way around that one too, but I haven't explored that much yet.

You're right, there is a way but its a workaround and its not built into the system as a natural part of the research process there. As a result, it loses a lot of participants fairly quickly- particularly if you're looking to do any form of longitudinal study versus using the first study to screen. I want to say that my follow-up after 3 months was something 5% of participants, making longitudinal study effectively impossible.

Like you said, the costs are exceeding the usefulness and validity imho. I'm fairly sure the 'reading speeds' of participants is FAR faster than one could argue is evidence of full information processing.

 

[MCParent]

You're right, there is a way but its a workaround and its not built into the system as a natural part of the research process there. As a result, it loses a lot of participants fairly quickly- particularly if you're looking to do any form of longitudinal study versus using the first study to screen. I want to say that my follow-up after 3 months was something 5% of participants, making longitudinal study effectively impossible.

Like you said, the costs are exceeding the usefulness and validity imho. I'm fairly sure the 'reading speeds' of participants is FAR faster than one could argue is evidence of full information processing.

I will say that, while I was not a fan of mturk the one or two times I used it before, I used it this month for some really specific preliminary qualitative data and got amazingly good data. It's definitely more amenable to some things than others. e.g., I needed written responses to prompts, so it would have been easy to identify nonsense answers (though, amazingly, I got 0 nonsense answers and 50 super useful responses). But I think using it for a long survey would be a pain in the butt.