Help me graduate!

July 28, 2006 3:06 PM

So I'm an undergrad working on my honors thesis, and the deadline is disconcertingly soon. (less than a month!) and I still can't get my sample size large enough! How can I recruit children ages 8-18 with ADHD with school out? I've already asked all the kids at the practice I'm associated with. Any suggestions, especially ones that would work quickly are much appreciated. I'd like to begin data analysis within a week to be able to start writing. Target number of kids: 10-15 more

The study itself is about how parent beliefs (attributions) about ADHD symptoms and their cause affects the quality of the relationship between the parent and child.

Its about 20 minutes worth of questionnaires for the child and parent each, and while I had no problem having people say yes to letting me send them a packet, I can't seem to get them back. I know they're doing me a favor, but if I dont get more people, I don't graduate from undergrad, and grad school starts a week after my paper deadline!

Any suggestions on how to recruit kids without being the scary person on the playground would be most appreciated!

posted by gilsonal to education (18 comments total)

As I understand it, using children in studies is highly regulated, so you likely have to do this through your university.

I'd strongly suggest asking your department.
posted by k8t at 3:13 PM on July 28, 2006

The study has been approved by the research review board at my school, but they don't really have any ideas on where else to recruit
posted by gilsonal at 3:16 PM on July 28, 2006

Try putting ads on Craig's List and offering some sort of incentive (cash?).
posted by k8t at 3:19 PM on July 28, 2006

You could try contacting groups like CHADD. There may be a local chapter that meets frequently, or they may have a mailing/email list that they'd let you spam advertise to.
posted by needs more cowbell at 3:24 PM on July 28, 2006

Summer schools are filled with ADHD kids forced to retake subjects. Try there.
posted by Neiltupper at 3:49 PM on July 28, 2006

Are you near a medical school? Go to their Psychiatry dept. and ask if you can recruit their patients or even just hang around their Child Psych waiting room and give their patients and their parents who are usually always there your survey. Good luck!
posted by skepticallypleased at 4:12 PM on July 28, 2006

For recruitment go of course to the pediatricians, PCPs, child/adolescent psychiatrists in your area. But the bigger problem sounds like your subjects don't want to respond, for whatever good or bad reason. Maybe you can refine the pitch for your project that is more pro-parent, emphasizing how this will help their child, etc., and not just be a perceived criticism of the parents, which they will reject or toss out.
posted by madstop1 at 4:53 PM on July 28, 2006

Have you discussed this with the faculty member(s) supervising your thesis so that you know for a fact that they will fail you if your sample is too small?

If not, discuss it with them.

I would not be at all surprised for them to very reasonable about it and expect you to have a good theory, lit review, and research design, and to make what conclusions you can with your data. And, if you couldn't get a good sample, to discuss some possible reasons why that is and what steps could be taken to obtain a larger sample. Unless maybe you're at Uber-Harvard, I don't they're actually expecting a publishable piece of original research from you. It might not be what you want it to be, but in 5 years nobody on Earth but you will care about that.

How have you been dealing with your non-respondents? The most common way to deal with them in "real" survey research is to HOUND THEM MERCILESSLY DAY AND FUCKING NIGHT AGAIN AND AGAIN AND AGAIN, so that filling it out and making you go away is a Very Good Thing. If you've been being polite about it, you might well find that you'll get more completed questionnaires by being a bothersome jerk than you would by recruiting more people who will go home and not fill it in.
posted by ROU_Xenophobe at 5:22 PM on July 28, 2006

How large is your sample size now, and what kinds of statistical tests do you want to use?
posted by jtfowl0 at 5:25 PM on July 28, 2006

You need to deal with your own ADD. Why have you waited until the very last minute to collect data? You know that you should have completed this task many months ago so that at this point all you are doing is editing the finished thesis. Agghhh. If I weren't worse, I would criticize you.

I think you are screwed. Rather than rush through a piss poor product with faulty data, get yourself an extension. No real human being graduates any serious school without one, only robots do. Bite the bullet. Talk to the dean. Point out all the reasons why you have had trouble collecting data, except your own sloth, and beg for mercy. It will likely be granted, although not without some price. Pay it.
posted by caddis at 6:16 PM on July 28, 2006

Xenophobe- I'm calling people more than once a day, with and without call blocking, and leaving a voicemail once a day, and even then, apparently they like the sound of my voice, since they aren't answering. They're all really nice people, but I sometimes wonder if there isnt truth to the office joke that all the ADHD parents have ADHD themselves.

I hope to god they let me at least pass even if the sample size is too small. I dont think they would fail me, but the fact that I don't know is keeping me up nights.

jtfowl0-My sample is currently at 17, and I'm honestly not exactly sure how we're going about it, but I'm testing which of two models (moderating vs mediating) explains the relationship better. And a little bit of factor analysis to make up the relationship variable. My advisor wants to work through it all with me tomorrow and doesnt seem to understand why I can't get the packets back

They wanted 50 originally to get a moderate effect size, but when you start with a pool of 60 possible participants, and 30 say they'll do it, and only 10 send them back, things dwindle.
posted by gilsonal at 6:21 PM on July 28, 2006

caddis, gilsonal says quite clearly that the problem has not been in getting people to agree, it's been in getting the respondents to turn in their information. This can happen even when you aren't procrastinating. And you are criticizing gilsonal.

An undergraduate thesis that does as best it can with the data it collected, and discusses the issues surrounding the data collection, would not be a "piss poor" product.

Talking to the dean is very premature, since the dean is not going to be grading it. The only reason to talk to the dean would be if the supervisor(s) had adopted the (imho unreasonable) position that without a p-value of .05 or less, gilsonal will fail.
posted by ROU_Xenophobe at 6:25 PM on July 28, 2006

Are you in a large enough city that there might be some kind of support groups for ADHD or special schools for learning disabilities (there really are some for milder problems like ADHD and dyslexia)? Could you talk to someone in the local school system to see if they have any ideas?
posted by dilettante at 9:06 PM on July 28, 2006

needs more cowbell: "You could try contacting groups like CHADD. There may be a local chapter that meets frequently, or they may have a mailing/email list that they'd let you.. advertise to."

I'm with Cowbell. Also - is there a reason why you can't contact online ADHD groups and see if some of them would be willing to participate? Perhaps you could set up a website with your questionnaire, and give out invitation codes to some of the online groups who could give them to willing participants. I know I'm always more likely to fill out something online, simply because it won't then sit around in my house waiting for me to buy stamps (or some other equally ridiculous excuse) until it's so late that I throw it away.
posted by Meep! Eek! at 9:09 PM on July 28, 2006

A quick psych trick may help here:

Leave a more detailed message and/or send a real letter -- whatever will be more convincing and will get there in a reasonable amount of time -- and establish that

(a) $50 if they return the form within 3 days
(b) you will not be calling or writing again, and any reward will not be awarded if it's not turned in in three days

You can always go back on (b) if this fails, but giving them a hard deadline and a cash incentive may do the trick here (with people who already have the packet). $50 is enough to motivate most parents, if this was kids only $20 would be more reasonable.
posted by little miss manners at 9:11 PM on July 28, 2006

I'm sure there must be a Boys and Girls Club or a Family Y in or near where you live. Most of them have special needs day camps. Contact the people in charge. Good luck.
posted by Gungho at 5:53 AM on July 29, 2006

Seventeen might be a little too few for a factor analysis, but there are certainly other statistical avenues available. What are your expectations regarding the distributions? Can you assume normality? If so, you could create an ratio/interval index measurements of your dependent and independent variables and do some regression.

While 17 seems like not very many, there are plenty of drug trials, etc. that deal with smaller sample sizes than this and still produce accepted results. Just remember that your effect is going to have to be pretty large to be noticed--a small effect is just going to get lost in the noise of the wide confidence intervals generated by the small sample.

Additionally, you need to think about the reason your response rate is so low. Do you think there are non-random factors influencing the decision to respond to your survey? If so, you need to take this into account as well--logit/probit models, for example, are designed to deal with self-selecting samples and might be appropriate, depending on how you are planning to use the data.

Overall, I think you should be ok with your sample size as is. Sure, ideally, you should have collected the data months ago. But that being said, there is no reason that your project should be failed--your work and knowledge indicate that you are pretty far above the typical undergrad, at least at every school I've ever been associated with. Your advisor(s) would have to be a huge jerk to fail you, IMO.

Seeing the dean, applying for an extension, etc. seem a little bit extreme. Do the best you can. Real world research is never ideal--something always comes up like this. Part of being a good researcher is having the ability to get the most out of what you have, and it sounds like you are right there.
posted by jtfowl0 at 8:40 AM on July 29, 2006

Many special ed programs through school districts have summer programs. Call a local school district and ask for a contact.
posted by SuperSquirrel at 4:08 PM on July 29, 2006