ScienceVille: An Idea for a Citizen Science Gaming Community

ScienceVille

Why?

It’s said so often that it’s a cliche: “If he just put as much work into his life as he does into that video game, he’d be a millionaire.”

Though this notion contains some truth, we recognize it as something of a straw man. Video games aren’t really like work. To take a typical example – I used to play Everquest (don’t look at me like that) when I was a teenager. In this game, the way you learned language was by standing next to another player using the language in question. By “hearing” enough, you would eventually comprehend. But this is not how actual language comprehension works. In real life, you have to think a little. Learning is partially hacking away at something, but largely it’s about reorganizing how you think.

But still, there is something to that cliche. Individuals invest hundreds of hours every year performing complex tasks, some of which (like the language lessons in Everquest) they don’t even find interesting. Many popular facebook games involve building simple communities to no obvious end. There seems to be a quirk of human behavior that allows us to grow addicted to simple games that involve designing systems that grow ever larger.

I suspect this has something to do with a feeling of “progress.” As someone who’s been lucky enough to run a successful small business, I know what it’s like to experience the progress of oneself in the real world. When it works (if it works!), it feels great, but that working takes years of incremental improvement. In a video game, you can experience the entire rise and fall of human history under your control with the a priori knowledge that the universe centers around you. Despite being able to check on my real live business’ progress, I’m still tempted by the idea of a business video game, due to that progress-as-an-emotion feeling.

Video game designers have capitalized on this feel by adding arbitrary point and merit badge systems. Even though these things are essentially meaningless, to a human being they feel validating.

Now, what if we could harness that feeling? Well, as some of you are already aware, this is being done already. But, I believe there are some flaws in design and scale. So, I thought I’d lay out a cool idea I had while taking a walk a few days ago.

What?

I will elaborate in more detail on this concept during the next several sections, but the general idea is this: create a unified game for various types of crowdsource-style activity which benefit scientific research.

The game and activities don’t need to have anything in common. In the same way that WoW players don’t all have the same playing style, players of “ScienceVille,” could play a variety of types of games. The unifying factor is that the rewards are all given out in a single currency of points, merit badges, and virtual items.

How?

ScienceVille would be a game with a single website, but playable through phone apps and facebook.

For the research, we’d have to locate particular types of experiments that could benefit from the help of large groups of people. Here are some ideas:

A) Simple games

Many scientific activities, particularly in biology, involve looking over images to pick out details. For example, counting the number of parasite eggs in a sample, or clicking the location of an animal in a series of video frames in order to watch movement, or watching a video of animals interacting to analyze behaviors (such as number of aggressive acts). Games like these could be made fun be adding a racing quality. We could see who can count parasite eggs the fastest, or who could analyze the most slides in the shortest time, or whose counts are most accurate.

Another common activity in biology is going through images to analyze facts about organisms, such as surface area, length, volume, etc. This is sometimes done by taking an image, selecting the area, and having a computer determine how much area is selected. This could be played in a similar way.

These are some obvious examples, but there are probably many more ways simple human behaviors could save researchers a lot of time and money.

B) Puzzles

These would be more complex games. A number of problems in science involve having a computer repeatedly solve puzzles. I’m not in deep on the math enough, but my understanding is that there are still many processes at which humans are more adept than algorithms.

Some possibilities include protein folding, taxonomy trees, and simple open math theorems.

The ideal puzzle game is one that is relatively simple, can be learned easily, and whose outcome can quickly be measure.

C) Games that require training

Some games could also be designed for “advanced” players. That is, people who don’t have the time or desire to be inside academia, but who have some expertise (or are willing to learn) in a certain area.

Some examples could be taxonomizing simple samples, quantifying animal behaviors, and doing statistical analysis.

D) Distributed Computing

Okay, this isn’t a game, but it’s still good. And, since it’s part of the unified game system, you’re still doing it for points.

E) Bonus Activity

For sociological/psych/economics studies, filling out surveys could be of great use.

Additionally, there are a lot of old documents that really need their data brought online. For example, just yesterday, Kelly was at the library going through a parasite data set from 1924! Not only is the data hard to get, but it has to be transcribed by her onto excel. If this could be done (again, for digital rewards), it would be useful for all sorts of scientific meta-anlaysis.

Lastly, we could also set up a micropayments system for scientists. A lot of scientists doing fieldwork need simple, relatively inexpensive devices, such as tags, test tubes, and software. A few hundred dollars isn’t much for a big community, but it’s a lot for a starving grad student.

F) Hands-on work

This is a little more pie in the sky, but there’s no reason some of the less skilled hands on work couldn’t be done by players for big points. For example, a trained fisherman might be very useful to a lake ecology sampling trip. Or, someone with minimal microscope training, might be able to go through images and take photos for a scientist.

Avoiding/Ameliorating Cheating and Abuse

In order for this to work, we must be assured that the data generated is valid. For example, in the counting games above, it’s easy to imagine someone clicking a thousand times randomly, declaring himself ┬áthe winner, then collecting points.

Obviously, this is a problem. In order to be valid, the data needs to be at least as good as what a professional would’ve collected.

For puzzle games, this might not be a problem – the players are creating solutions which are either good or bad. But, for the aforementioned clicking games, we could use redundancy and statistical analysis.

For example, say we have 500 images of microscope slides for parasite egg counting. We could have professional scientists analyze 10 random slides. Then, start the game. Each slide would get a certain number of games played (say 20) before dumping in the pile. If the 20 are in rough agreement, we take a weighted average to determine what the actual number is.

For a particular slide that was also analyzed by a scientist, we do a comparison to see how big the error bars need to be. Presumably an experiment could be done to determine how accurate a particular game setup is compared to a real scientist. With enough redundancy, it seems reasonable to suppose the accuracy would be pretty solid.

Additionally, players caught cheating could be permanently banned.

Pre-existing examples

A lot of the above might sound fanciful. In particular, like many ideas, it may seem to fail to bridge the gap between a cool idea and actual implementation. However, games like those mentioned above are already being done.

Check out Phylo, or Foldit. In fact, Wikipedia has a whole section for so-called GWAP.

These games are reasonably fun, but I believe what they lack is a harnessing of the addictive property of progress-as-emotion.

Harnessing Addictiveness in a Broader Game

It is reasonable to suppose that these games and activities will not be as fun as the equivalent versions which are designed purely for amusement. However, there’s an addictiveness trade-off. If you play Angry Birds for 12 hours, at the end you probably feel like you’ve wasted time. If you play a science GWAP for 12 hours, you at least have the consolation of having supported something good.

In addition, you get points which can be spent on virtual goods, unlocking new game aspects, and achieving status in the virtual ScienceVille community.

Here are some ways the game might be made more addictive:

-Every day, you get SOME baseline amount of points. You get more for playing, but the baseline level will keep you coming back.

-Unify all games under one “currency,” so as to encourage the broadest range of activity

-Provide games which are just games, and not scientific puzzles. If they’re addictive enough, they can lure in players who wouldn’t normally be interested in the science games. Counterbalance their non-GWAP quality by making them provide far fewer points than other games.

Going Beyond

Because this game would be philanthropic in nature, one can imagine reaching out to other more traditional games, like the various MMOs, to agreed to a shared digital currency market.

Scienceville could also provide basic educational services, such as access to material, discussion forums, and simple tests of understanding. These could be used as barriers to the “advanced” games in order to give them more mystique and fun.

Conclusion

In the last 20 years, video game designers have practically created a science of virtual incentives, with the result that simple games can be designed that keep humans engaged for hours. This is a powerful thing when you consider it. Imagine watching a sped-up video of a person playing Farmville. You’ll see someone looking at a sequence of simple images, moving his cursor, and tapping the occasional button. And yet people will play for hours and hours this way without growing bored. If we can use this quirk of human behavior to for better purposes, we could get access to literally millions of hours of scientific work, which would benefit professional scientists while promoting citizen science.

What do you think?

This entry was posted in Games, Ideas. Bookmark the permalink.

25 Responses to ScienceVille: An Idea for a Citizen Science Gaming Community

  1. davejac says:

    Damn good idea. I’m currently undergoing training to be a high school chemistry teacher and have been trying to think of a way to do something along these lines.

  2. JK says:

    Interesting. As a biology academic I can think of a lot of pros and cons for this kind of project… Its certainly worth a try, although I wonder how easy it would be to find scientists who would be willing to trust the internet with images of all of their data before its published…

  3. Steven says:

    I’ve had similar thoughts – what if you could take all those World of Warcraft man hours and turn it into something productive. In my own case I’ve made a control panel for my web business that attempts to recreate the addictive qualities of an MMORPG. So instead of trying to kill goblins for XP I’m trying to gain customers for real money.

    The best example I’ve seen of something like this was the collaborative aerial photo search for Jim Gray and Steve Fossett when they went missing. Users were repeatedly presented with an image and asked if they saw anything unusual. If enough people said yes then it was worth looking into. As you say, humans are still better at some processing than computers, and it’s just a case of somehow integrating the human into the process.

    I’m also reminded of a video I watched recently of a talk by Robert Sapolsky. He explained about the effect of dopamine, work, and reward. The most interesting point he made was that dopamine is highest when there’s a chance of no reward and how casinos use this fact.

    http://www.youtube.com/watch?v=hrCVu25wQ5s#t=26m30s

  4. Watashiii!!1 says:

    I recently watched a similar idea on ted:
    http://www.ted.com/talks/seth_priebatsch_the_game_layer_on_top_of_the_world.html

    Now, all we need is a superhero greedy enough to gather talented people and make an addicting game which actually helps the world

  5. Jonathan says:

    That is a good idea. I was going to bring up FoldIt, but then you went ahead and mentioned it at the end! I hope this model can find wider application….

  6. Nathan says:

    I know that you intended this article to put forth a new idea, but I just wanted to say that it’s gotten me very intrigued by FoldIt.

  7. Darbus says:

    At the moment, there’s a game called Improbable Island (improbableisland.com) which, if you participate in World Community Grid’s (http://www.worldcommunitygrid.org/) distributed computing (which is explained briefly as it relates to the game here: http://www.improbableisland.com/grid.html), you get in-game benefits. I don’t know how many other games there are like that out there, but it’s the only example I’ve run across.

  8. Cat says:

    It’s an interesting idea, that’s for sure. I think that you’d have a bit of trouble translating the scientific processes into games, but I can definitely think of a few things that could mold well.

    One of the particulars that springs to mind is the formation of complex organic molecules. I can imagine something of a race/puzzle game featuring a tiered gaming system, where a person is taught the basics with simple alkanes and functional groups, then moving up to try their hand at creating a difficult structure. Because they really are nothing more than very complex puzzles.

    Simple distributive computing might be more efficient than a bunch of people doing them by hand, but you might be surprised at what you get.

  9. Nick says:

    In many ways, this idea reminds me of reCAPTCHA [http://www.google.com/recaptcha].

    In theory, your idea could work. However, in practice I suspect that only a small subset of these games will be compelling enough to keep people coming back for more.

  10. Chaz says:

    For a supremely successful example of games that double as croudsourced manual labour, look no further than Google Image Labeler.

    http://images.google.com/imagelabeler/

    Much as you suggested, it uses COMPETITION (you’re always playing against someone else) and POINTS to reward speed and accuracy.

    The competitive aspect adds the all-important human element to the gameplay. It also serves as a way of checking the viability of the answers (you can’t move on until both partners agree on a tag).

    The countdown timer encourages you to do it as fast as you can (good in itself), but also keeps you coming back long after you might have normally got bored, just so you can try to do it faster and faster.

    Essentially, you’re helping a big corporation organise its image database — but it’s fun and addictive, and you’re wiling to spend more time of it because (at some level) you don’t see it as completely wasted time.

    I’d encourage anyone creating these sorts of games to follow Google’s example: make them competitive, make them time-sensitive, and make sure there’s always the possibility of DOING IT BETTER NEXT TIME — that’s what keeps people coming back for more.

  11. Pingback: Tweets that mention The Weinerworks ┬╗ ScienceVille: An Idea for a Citizen Science Gaming Community -- Topsy.com

  12. Here’s another one:

    http://www.newscientist.com/article/dn19966-online-game-helps-predict-how-rna-folds.html

    A fun little flash game, not a megalomaniac project, playable from work

  13. Jeremy says:

    This is a really cool idea, and one I’d be interested in.

    As for preventing cheating/trolling, I have an idea that comes from The Prefect by Alastair Reynolds. The society the book is about is truly democratic-everyone votes about everything. But the people who vote correctly in retrospect have their votes weighted.

    So going with the parasite counting example, the more times your count lands within some range of the final average, the more heavily your count is weighted in future counts.

  14. Grae says:

    It sounds to me like a beneficent version of Neopets, with its flash games to provide neopoints to spend on virtual crap for virtual pets. Except where some of their games are advertisements for commercial enterprise, some of these games provide some gain to science. Such a site could also extend beyond science — as mentioned above, the aerial search is one example of how such a large community could otherwise benefit the world.

  15. Chris says:

    There’s nothing wrong with a social platform for science games, but I think they inherently resurrect bad memories of science class for most casual “FarmVille” gamers.

    My suggestion would be to convolute a science game so abstractly that the gamer doesn’t even realize they are doing science, embed it deep within some un-related platformer and reap the benefits.

    In short, raping the intelligence of gamers for science.

    • NP says:

      I think you are exactly right, Chris. My mom spent hours learning about breeding dragons for some DS game. If amusement is the primary goal in the game design, the actual content doesn’t matter as much. Substitute in actual natural laws for artificially constructed game mechanics and you get more active participants in the scientific process.

      I would love to see a person or group actively pursuing this goal.

  16. Rhiannon says:

    At the Cornell Lab of Ornithology we are currently working on our first game-style/ virtual world citizen science project. While the emphasis is not on winning the game, we intend to use a complex badging system to encourage data-sharing and behavioral change. We hope that the community of folks who end-up using YardMap develop their own patterns of altruism and cheat-ID. We are also just starting a partnership with the GWAP folks to develop a truly crowd-sourced, smart ID tool for birds. And, we’ve got an image labeling game, that while isolated from other citizen science projects, really tries-out the game-feel: http://www.camclickr.org and then, of course there is the wildly popular galaxy zoo. Its good stuff out there and if we could eventually collapse into one community with multiple content-foci. . . all the better!

  17. Andrew says:

    Another example is Galaxy Zoo: http://www.galaxyzoo.org/

    I think the same strategy could/should be used for education. People playing WoW have spent considerable time calculating the optimal strategies to maximize loot or honor, because they need that info in game. Imagine if the game required you to have knowledge of, say, biology. This game has the right idea, but could use better execution: http://www.kongregate.com/games/CellCraft/cellcraft

  18. One thing I was very excited about was websites that are funny, I find myself spending hours and hours surfing through each one. So it was at that time that I decided to make my own, it’s called IronMuffin and we’d like to invite you all to come and visit. Thank you so much for allowing me to do the invite. See you there.

  19. Samantha Vimes says:

    I think it’s an amazing idea. I know there are many people (myself included), who would love to do something like that, both the amusement of the game structure, and the knowlege that you’ve actually done something to help a scientist somewhere!

  20. Aditya says:

    In the past few years, some people in the NLP (natural language processing, a.k.a. computational linguistics, NOT the bunk that is neuro-linguistic programming) community have started using Amazon Mechanical Turk to crowdsource tasks like annotating corpora. The idea is that you pay a small amount (cents) per datum; each datum takes a small amount of time to be annotated by a user. Many users doing this gives you many labels per datum, and there’s been some work in looking at how to clean the annotations and remove the bad ones (simple voting, etc.). We had a whole workshop last year for people using Amazon Mechanical Turk:
    http://www.wikicfp.com/cfp/servlet/event.showcfp?eventid=7979&copyownerid=320

  21. Abraham Feinberg says:

    I think this is an excellent idea. I agree about the need for such proactive games to be “addictive.” I have a background in biology and psychology (currently a grad student in psych), so I’ve been interested in how we might apply psychological principles to helping people learn difficult (especially scientific) topics using computer software. My focus has been on how these principles could be used to help improve education, but obviously a person can be learning and contributing to scientific knowledge at the same time (indeed, this is perhaps the ideal format). Psychological research has established quite a few principles and techniques that I think would be put to good use in developing the type of project you describe. In particular, the area of applied behavior analysis contains some very useful ideas about how to get someone doing something redundant (e.g., playing slot machines) for hours on end. It would be nice to apply these principles to something that helps the person get a better understanding of a given topic and, hopefully, contribute to scientific research at the same time.

    I’m not sure if your post was simply to throw an interesting idea onto the web or if you might consider actually developing something along these lines at some point. If the latter, I’d love to be involved or offer any assistance I can! Feel free to shoot me an e-mail if you want to chat about it.

  22. Finn says:

    Another place this is being used is http://www.FreeRice.com

    Addicting little word game that helps you learn words (really good for kids SATs) and while you do so, the money from small banners on the page buys rice to donate to aid groups.
    Its also smartly designed, repeating words you mistaked earlier to help learning and scales difficulty automatically based on your score.

    I always enjoy these posts, keep writing!

  23. Devon says:

    You could use such a project to do psychological studies related to game theory and similar fields. I’d imagine you could collect tons of data related to the manner in which people play these games.

    I’m not psychologist or computer scientist, but I would imagine there could be some use for recordings of say 1,000,000 games of Chess or Go, or even simple betting games. If you could introduce more complicated strategy games I would guess you could learn a lot about human behavior.

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>