Forums >Health and Nutrition>UNC - Chapel Hill Male Exercise and Libido Study
ultramarathon/triathlete
There are things such as scientific surveys, which are imperfect of course especially when you are dealing with sensitive aspects human behavior. A good approach would be to design a stratified random sampling survey. (this can be done after the fact too--stratify according to response to the question of how much mileage do you do). To make it valid the investigator should aim for a 70 percent response rate (60+/- is acceptable, lower than that the data would be suspect). And they would need to do a follow-up survey of non-respondents to determine if there is a non-response bias. How many do you select for survey--that depends on variance--but 500 to 1000 questionnaires with a 10% to 20% follow up on non-respondents would get you some statistically valid numbers.
There are things such as scientific surveys, which are imperfect of course especially when you are dealing with sensitive aspects human behavior.
A good approach would be to design a stratified random sampling survey. (this can be done after the fact too--stratify according to response to the question of how much mileage do you do). To make it valid the investigator should aim for a 70 percent response rate (60+/- is acceptable, lower than that the data would be suspect). And they would need to do a follow-up survey of non-respondents to determine if there is a non-response bias.
How many do you select for survey--that depends on variance--but 500 to 1000 questionnaires with a 10% to 20% follow up on non-respondents would get you some statistically valid numbers.
70% response rate. Wowzers. If only.
But yeah, "academic" and "marketing" are two different worlds of research. This is why nothing really gets done in academic research (beyond other academics reading it, lol). Unrealistic expectations for perfection.
Having said that, some of my research gets published each year and is used to change government policy, which is beyond marketing. So THERE lol. Sampling bias is always an issue but not as large of one as people once believed. Recently we conducted research using an RDD platford, email panel, lage snowball sample, and a customer list. Same questions were asked across the board. No statistical differences of any significance.
It's always going to be people who want to take surveys that take the survey unless you FORCE random people to. And even then, many of them would probably have just taken it in the first place. People can't resist.
HTFU? Why not!
USATF Coach
Empire Tri Club CoachGatorade Endurance Team
#artbydmcbride
Why is the female libido after exercise not being surveyed?
Runners run
Not now I have a headache. Or maybe just wash my hair.
a comprehensive study appears to have been completed in the medical journal "Ask Men"
http://www.askmen.com/dating/vanessa_100/129_love_secrets.html
All,
I'm extremely (and pleasantly) surprised about the amount of discussion my survey has generated. As many of you have suggested, doing research using surveys is a very inexact way of answering a question. But in many cases, this type of research is done to assess if an actual in lab research study (that costs significant time and money) is needed. As of right now, there are no studies that have examined the effects of longterm (intense) exercise on libido in males (to my and my advisors knowledge). So, we decided that a survey (while inexact) would allow us to begin to examine this question, which would hopefully lead to a second study that actually brought males into our lab at UNC. I think many of you would be surprised by our preliminary results.
I'm sure that many of you have friends that would also be interested in this survey, please send it to them. As of right now we are at about 950 people and only need about 50 more (https://unc.az1.qualtrics.com/SE/?SID=SV_eWESbBdaONSuUPb). Sorry for all of you women, but I think our lab is planning on doing this same type of study using women once we've completed this survey.
Please continue the discussion. I will be checking periodically to answer any questions. Thank you all for your time.
Sincerely,
Colin
I read (but rarely post!) on the RA message boards but as a social scientist (in public health) who is teaching undergrad research methods this semester, I couldn't help chiming in! I think it is so easy for people to completely dismiss a study because the research design has some flaws, but rarely is a study designed in a haphazard way. Most scientists (I hope!!) consider the benefits and drawbacks of a variety of research designs and select the best method given the resources available and the aim of the study (in this case, just collecting preliminary data to make a case for future study).
Any study will have some level of error. For example, there will always be some level of self-selection bias, since we can't compel people to participate in research (as much as we researchers dealing with low response rates would like to, lol). Different survey modes, sampling strategies and different measures will also produce error. Fortunately there is a considerable amount of literature specifically examining how these things affect the data, so that researchers can be aware of and attempt to account for any bias their chosen design might introduce.
Internet studies tend to be one of the more feasible approaches, especially for graduate students who typically do not have lots of funding for their independent research. Obviously we wouldn't want to make policy decisions or develop treatment protocols or anything serious based on these results but for preliminary data gathering this seems like a perfectly reasonable approach. Though maybe I am biased...I have 3 grad students with online surveys in the field right now (two of them are with undergrad samples...oy). Also want to commend the researcher for checking back in and responding to concerns...so often researchers think of their participants as data points as opposed to humans with questions and concerns about their participation!