Naked Science Forum
General Discussion & Feedback => Just Chat! => Topic started by: AntonMaeso on 04/07/2013 11:18:38
-
Please Read this page after filling in the survey. Go to Science Experiments then "Please help with an online Psychology experiment. Takes less than 10 mins" (if you want to do the survey at all)
You may of noticed that I have posted an online study up on the experiment form. However I am finding it challenging to get the participants. I have tried using social media (e.g. facebook), and to make my naked scientist post as inviting as possible with no funds to offer (see experiment form).
Are there any other places on the internet that you would recommend posting it?
Do you have any other ideas for attracting participants?
Advise would be greatly appreciated.
-
It's very hard to get anyone involved in any survey, so you need to guard against the few who do start filling it in bailing out like I did half way through. Question 10 is impossible for me as I have never been through customs, so any answer I try to give will be a lie. Question 13 is possible for me to answer with a 0, but as I've never driven a car you will likely misinterpret my answer. Most importantly though, almost all of the answers I would give would have to be either 0 or 7 because there is no logical possibility of an in-between answer, so this keeps shouting at me that the survey is badly designed and that I'm wasting my time on something sloppy.
-
I do say in the post that "Any questions you cant answer or don't want to answer just leave blank".
The the thing is it is not my questionnaire and conveniently for me it has been standardized. So it has been tested in large populations and replicated a number of times. More importantly the people who made it went to some effort validating (making sure it measured what they think it measures). These are quite impressive feats for what I'am trying to measure (If you are interested I can send a private message about what that is).
Assuming I could do better, I don't have the resources or time to make my own standardized questionnaire. So I am kind of stuck with questionnaire (unless some kind soul would like to pay for my PhD).
However saying that i appreciate the feedback. It may end up my sample is not representative and I am missing the interesting participants (one to discuss).
-
I do say in the post that "Any questions you cant answer or don't want to answer just leave blank".
Everyone's in a rush to get on with their lives, so they don't read the small print. As soon as they hit a problem, they're gone. I tend to have a go at filling in surveys just because I know how hard it is to get anyone to fill them in, but most of them I end up having to bail out of because they're badly designed.
The the thing is it is not my questionnaire and conveniently for me it has been standardized. So it has been tested in large populations and replicated a number of times.
That's awkward then, but the number of times it's been tested and replicated only tells you that it's "confirming" itself by repeating the same errors.
More importantly the people who made it went to some effort validating (making sure it measured what they think it measures). These are quite impressive feats for what I'am trying to measure (If you are interested I can send a private message about what that is).
If you provide that message in a PM I will look through the questions again and comment on them in a PM in the light of that knowledge, but it won't change the problem that the questions are asking how true or false something is where the answers should in most (or possibly all) cases be either true or false with no possibility of compromise. That instantly led me to think that it was unscientific and poorly designed, making me think I was wasting my time filling it in. If the wording was more along the lines of how often is it true versus how often it is not true, then answering with numbers 1-6 would have been valid.
However saying that i appreciate the feedback. It may end up my sample is not representative and I am missing the interesting participants (one to discuss).
Ah, but maybe you want to put the "interesting" participants off as they are likely to be overrepresented in that the people who try to fill in surveys are already a self-selected group of freaks, perhaps dominated by people of the kind that will try to pick apart the survey as they go. With a lot of the questions, I also kept thinking that my answers would not be believed because so many people would lie where I was telling the truth and that my answers would therefore just add to the belief that people lie about certain things - if you're trying to measure liars, it's better if the people who are telling the truth don't get involved because then you have a better chance of weighing up the liars correctly.
-
I personally don't just blanket open JAVA to unknown users based on a redirect post on the WWW.
Thus, I get stuck on the second screen on your survey.
Look at Server Side Scripting (http://en.wikipedia.org/wiki/Server-side_scripting)
Here are your screens.
[ Invalid Attachment ]
[ Invalid Attachment ]
-
Here is a link to a chapter on the topic with the scale I am using.
http://fermat.unh.edu/~mas2/Chapter2-Paulhus.pdf
I am interested in how the presence of eyes affect the social desirability bias (a tricky topic). I am using the impression management of the scale as a control. Trials are also counter balanced with eyes and lips.
I agree. I have made it bigger in the hope that will help.
By testing the reliability it means that the scale works in different populations and you get the same results if you repeat the test. An important asset to any measure.
Its interesting because most of the scale measuring social desirability are true or false. However its not good if people are quitting because they think the scale is bad.
I am interested in traits that effect all humans. So I am interested in everybody is welcome to fill it in. In theory some people will massage the truth and give slightly overly positive results.
-
I am interested in how the presence of eyes affect the social desirability bias (a tricky topic). I am using the impression management of the scale as a control. Trials are also counter balanced with eyes and lips.
Is that some weird kind of jargon they use? It's beginning to look as if it's a branch of science that needs a big shake up. Don't let them suck you in too deep - do whatever you have to do to pass the exams, but don't let them dictate how you think.
By testing the reliability it means that the scale works in different populations and you get the same results if you repeat the test. An important asset to any measure.
What that really means is that it generates the same good/bad results in different populations. It certainly doesn't tell you that the results are right or useful.
-
Is that some weird kind of jargon they use?
It is jargon but that don't mean it is not a useful concept. I guess at the root of it I am investigating why people lie in questionnaires. Whether they lie to mange the impressions they make on others or whether they are unconsciously overly positive.
What that really means is that it generates the same good/bad results in different populations. It certainly doesn't tell you that the results are right or useful.
That would be validity. However reliability of a scale is still important. If results cant be replicated across a wide range of situations and populations what use is the scale.
-
So it sounds as if the "lips" and "eyes" stuff is a disguised way of saying that you're doing the survey in two different ways: speaking to someone vs. having them read and answering it much more anonymously, at which point you can see if they lie to a lesser, greater or the same degree than in the former case. Now that I understand what the jargon means, I can see that you could indeed get a useful result from it. You'd just have an easier time of it if more people filled it in, and that's where the design of the questions is important if you want to enocurage people who take it on not to drop out on the way through it.
What surveys really need to do to attract participants is offer a reward of some kind, and perhaps this could be done with a couple of jokes. Give them a good joke up front, and then tell them they'll be shown an even better one at the end once they've completed the survey. It would be vital to deliver on that promise, otherwise it would harm future surveys.