BitShares Forum

Main => General Discussion => Topic started by: inarizushi on February 03, 2015, 05:22:10 pm

Title: Paying subjects of Scientific Experiments
Post by: inarizushi on February 03, 2015, 05:22:10 pm
Hello all,

I may have a great possibility of BitAssets use : remunerating time spent on scientific experiments/surveys.

I am a cognitive psychology PhD student, and as such I need to find subjects to pass experiments. It is extremely time consuming to be present during the experiment, so setting up an experiment on internet is much easier. But it's basically impossible to both 1) pay subjects and 2) make sure a person passes the experience only once. Amazon mechanical turk can do the trick, but it's unavailable outside the US, and most mturk workers do taht for a living and don't try to comply with the requirements of the experiments, for example they listen to music or do something else simultaneously (yup, some research has been done on that). Moreover passing experiments is not at all in the spirit of amazon mechanical turk, which is designed for brainless activity (well, a brain requiring brainless activity !) like classifiying images or typing relevant text from images.

I don't know how marketers and survey makes find their prospects, but for scientists conducing experiments in front a computer, be it psychological experiments or surveys, there is NO solution to the problem of finding subjects on internet. Which is crazy, when we think about it : this is remunerated work, generally not interesting, but at least uncommon and varied.

So, I would love to have that service on a blockchain, paying with BitAssets, and, the icing on the cake, the possibility to uniquely identify subjects (when the Vote identification functionnality is there...), to cross-check data between different experiments, etc. That would change everything in the way we psychologists work.

It probably cannot be decentralized : as I see it, the blockchain would track who did what experiment, so that the subjects can be paid. The experiments themselves would be passed, and data collected on an external website. (I would LOVE to see scientific data on a blockchain, but that's probably very inconvenient).


For adoption of BitAssets, it would be great : with so many people looking for ways to earn money, remunerated work markets itself.
The amazon mechanical turk market and the survey market are, I guess, huge compared with what we little scientists do, but I can't see why they couldn't be implemented on a blockchain. However, if I were to provide this service, I would probably want to start with my little niche scientist market.


Probably, it would be quite costly to make this a third party DAC. It should clearly use the BTS DAC, with its BitAssets and (present and future) identification features. As I see it now (not much reflexion yet), the subject connects himself to the website and chose a proposed experiment. He would then be asked to identify himself by sending a transaction to a specified account (maybe with a specified memo), then pass the experiment, and receive his payment at the condition he respected some specified rules of the experiment, automatically verifiable (obviously, he must go the end of the experiment). It would be so much better with unique identification, but it could still be great without. I think that I could convince some colleagues to use that service to run their experiments, and that it would be quite easy to find subjects, so it might be a great way to find new users and drive adoption of BitAssets (well, I know, mini market, but the concept seems good to me :D ).

What are your impressions ? Do you see any reasons that would make it hard to implement this service ?

Also, has someone thought of/is someone putting amazon-mechanical-turk-like service on a blockchain ?

Title: Re: Paying subjects of Scientific Experiments
Post by: Ander on February 03, 2015, 11:25:02 pm
The only issue I see is that a person could simply make multiple different BTS accounts and use each one to repeat the test.
Title: Re: Paying subjects of Scientific Experiments
Post by: yellowecho on February 04, 2015, 06:10:00 am
The only issue I see is that a person could simply make multiple different BTS accounts and use each one to repeat the test.

I think there are a few ways this could be prevented.  Each experiment could have an associated access code that's generated prior to the start and expires within a certain time period.
Title: Re: Paying subjects of Scientific Experiments
Post by: inarizushi on February 05, 2015, 11:00:18 pm
The only issue I see is that a person could simply make multiple different BTS accounts and use each one to repeat the test.

I think there are a few ways this could be prevented.  Each experiment could have an associated access code that's generated prior to the start and expires within a certain time period.

I'd rather start by asking for identification through facebook/social network during the creation of the account on the third party website, I guess that would do the job. On mechanical turk, payments are not automatic, so the experimenter can have the possibility not to pay if he fears that someone is doing the experiment wrongly (and if someone re-does the experiment, I bet he will be bored and do it in a way that is detectable).

Not a real market, but I definitely like that project, and it can be my little platform for making new people discover cryptocurrencies by putting them to good use :). I'll work on it