Online Computer Graphics/Psychology Experiment

ennisca

Apologies, permission for such experiments need to be sought prior to posting on this forum. Please contact one of the Mods if you wish to post.

ennisca

Hi,

My name is Cathy Ennis and I am a post-doctoral researcher in the Computer Science department in Utrecht University. We are currently running a perceptual experiment on the animation of a virtual character (to be used in a serious game) and would be grateful of any participants we could get.

Details of the experiment:
You will be asked to view a series of short animations of a 3D Character (60 in total). After viewing each animation, you will be asked to rate the motions based on four parameters, which will be explained at the start of the experiment.

Duration: It will depend on how long it takes you to answer each question, approx. 1 hour. I have included a feature that allows you to save your progress and continue later if needed.

Deadline: This experiment will be online for a number of weeks

Experiment link: http://questionpro.com/t/AJTkYZOwUq


Thanks very much,
Cathy

*** Mod permission has been given for this post***

deegs

Have you have many respondants in the 2 months?

ennisca

I got a good few (I presume you mean from my last post, as this just went up this week), but I don't think I got any from this forum

deegs

It sounds interesting, I read your piece above and saw that it would take me to answer about 240 questions. I'm not sure if I am motivated to spend an hour doing that...
I clicked through anyway as it wouldn't be too bad if it was well laid out... but I closed it down after 3 or 4 pages of demographic info. Sorry.

I hope you get the numbers you want but I can understand how it will be hard to get even a few.

Do you have a budget for it? You could always use mechanical turk and pay a few cent to people to complete it?

ennisca

Yeah the demographic info is a bit of a pain - but really important for validating the experiment or examining effects. Once you get past that, it's actually not too bad. There are 5 videos on each page, so it's a bit faster. People seem to be getting through it in 30mins or so. But you're right, it's a lot of time to expect people to spend without an incentive. If I get a few more and the results look promising, we'll look into setting up some sort of prize raffle among participants which might help things.

Thanks for the input!

deegs

I do a lot of user testing with one of my jobs and you really need to put a value on the participants time.
For example if my company pays me 25Euro an hour they value my efforts at that.
If you do not incentivise me and expect me to give you one hour for free you do not value my time.

Generally people will give 2 or 3 minutes for a novelty factor if it’s well laid out and appeals to them.
Any more than that up to 10minutes and you should include a raffle.
Anything over 15minutes really needs an exceptional raffle or payment/voucher.
30mins + really needs a payment I think.

That said I ran a limited experiment a while back that lasted 25minutes. I had about 80 participants I think and I gave away 2 ipod touches.

You really need to rethink how valuable your participants time is... remember for a single or comment, they are entered into some fab draws on facebook for example.

I am planning an experiment that will run soon, it will take 6 minutes and I am considering mechanical turk and paying users for completion.

Just some thoughts, good luck!

ennisca

Thanks for the input - I know how important it is to recognise participants' time in recruiting them for these types of studies. Unfortunately, the university's funding doesn't stretch to compensating each participant for their time. I guess that's why I get more success in recruiting participants from education focussed forums, where participants take part out of interest to support academic research. Of course, if we can get some experiment reward funding down the line, it'll make things much easier for me, and more rewarding for participants.

deegs

Well, how about just designing the experiment so that rather than x number of participants reviewing 60 video’s you have (60)x number of participants reviewing one video. At the end of them reviewing one video you can thank them and ask they would they please review another.
I would probably do several there on the spot, and it may be easier to recruit?

Velvety

Because the actual content of the videos probably isn't very important to the research question.

deegs

Velvety wrote: »

Because the actual content of the videos probably isn't very important to the research question.

I don't follow? The video content is what the response is based on?

I watched the first five, and despite the instructions being confusing, I could rate any of the 5 vids any different, maybe differences become apparant later?

ennisca

Unfortunately, the experiment is a repeated-measures design, which means that each participant needs to see the same stimuli. So asking them to view one video at a time until they tire of it won't work. I realise it's too much to expect participants to take part in something for such a long time, which is why there is a "save and come back later" option.

The videos are all different, and the aim of the experiment is to assess how different the participants find each one to be, which is why we need everyone to view and respond to each. As you know if you read the intro, I am conducting research on the perception of emotional body language, which is incredibly subjective and varies hugely from person to person. So, for the purpose of this experiment, it's important that each participant responds to the same stimuli.

deegs

I do know, I did read it

I suppose I am interested as I am researching self reported states of cognitive load which is also highly subjective. My first experiment was repeated measures, and likewise I felt that it was asking a lot of the participant. However, next time around I won't be using repeated measures.

I'll try and get through it all (can't give the time as I'm at work) but it may take a few days (or weeks as the holiday season and all ).

Thanks for the clarification.