Science Can Be a Game with Citizen Sort

Citizen Science is a great way to learn about the world.  It's also a stimulating intellectual pursuit and fantastic learning opportunity.  It's even a great way to connect with nature all around you.  But most importantly, citizen science is fun.

Today's featured project uses the fun of citizen science to it's advantage.  Citizen Sort uses gaming technology to separate different types of plants and animals using just a photograph.  This can be tricky for even experienced biologists who have to rely on subtle characteristics of each animal to classify it.  Citizen Sort turns this into a game the everyday public can participate in.  As the researchers themselves state:

To aid in identification of the species of specimens, biologists have developed taxonomic keys, which identify species from their particular combinations of characteristics, known as character-state combinations (i.e., attributes and values). The specific characters and states vary by taxon, but are broadly similar in structure. For example, a moth character might be its "orbicular spot," with states including, "absent," "dark," "light," etc. Given sufficient characters and states, it is possible to identify a photographed specimen to a specific family, genus, species, or even sub-species.  To support the biological science goal of image classification, we have developed several games and tools that let ordinary members of the public undertake to classify various photos of living things. Over time, if enough people help out, this project will produce a very large data set of classified photos that will be very useful to natural and biological scientists in a variety of fields!

The Citizen Sort game is actually a collection of games.  As the research team has developed their idea the games have gotten increasingly entertaining and scientifically important.  Their initial effort, Happy Match, identified moths in a very simple game environment.  This led to Happy Rays (a similar program) and then a leap to Forgotten Island.  This is a much more involved game with excellent graphics and a storyline to engage players.  Finally, all this will lead to eventual launching of Hunt and Gather.  It's not ready yet but promises to build on everything from the previous games and take the gaming of citizen science to the next level.

Obviously this helps classify existing images collected by researchers.  But it's also a great tool for computer science researchers looking to improve computer processes that classify animals as well.  So don't just help one project...help them all!

Getting Started is Easy:

• Visit the Citizen Sort web page to learn more about the project and the various games waiting for you.
• Register on the front page with your proposed username, e-mail, and date of birth.  That's all the information they need.
• Once registered,  confirm your account and answer a few simple questions about your interests in science and gaming.  Just to help the designers understand their audience.
• Click on the Games and Tools tab to see the available games to play.  Pick the most interesting one.
• That's all there is to it!  Sure, there are instructions on how to play each game, but that's all part of the game itself.  No use for writing up my dry directions...enjoy it by clicking over yourself.

Hopefully I've done justice to the project.  I actually had the good fortune of meeting the research team last year in Portland and am excited to see the progress they've made since then.  So I don't want to let them down. 

Tell them OpenScientist sent 'ya!

Gregor Mendel - Father of Genetics and Son of Citizen Science

I love Citizen Science.  But I REALLY love Citizen Scientists.  Their passion for knowledge and ability to advance fields outside their own professions should inspire everyone.  So I've tried to combine this blog with both stories of citizen science research projects as well as discussions of the regular people behind those discoveries.  So it was with great interest that I found John Malone's book "It Doesn't Take a Rocket Scientist", highlighting the stories of the many great amateurs of science.

The first person I'd like to highlight is Gregor Mendel, the father of modern genetics.  This unassuming monk spent years working in his garden breeding pea plants; centuries later we still use his Laws of Inheritance are a pillar of modern biology and studied in high schools across the world.  Pretty impressive for a humble man who failed all his high school teaching certification tests.

Gregor Mendel was born in 1822 on a farm in Moravia, a part of Austria that is now part of the Czech Republic.  His father Anton was extremely hard-working and serious-minded while his mother and the youngest sister, Theresia, had much lighter personalities.  Even at an early age he was very bright with an inquisitive mind.  One of his idols was the father of the printing press, Johann Gutenberg, whom Mendel even wrote a poem about.

As a farm boy with little family fortune, higher education was not affordable for young Gregor.  Fortunately his kind-hearted sister Theresia knew something the rest of the world didn't, and helped Mendel pay for continued education by donating her share of the family fortune.  Money that should have been her marriage dowry.  This helped him get through the Philosophical Institute, a two-year program required of all students who wished to study at a university, but it was not enough money to attend university.  Even adding both a small scholarship grant and the extra money earned from tutoring on the side came up short.  For a young man in the 1800s there was only one choice left if he wanted to keep learning.  He became monk.

Although a mediocre student, One of Mendel's physics professors recommended him to the head of a local Abbey named Abbot Napp.  Napp was an old friend of the professor's and headed the monastery of St. Thomas; he was also a former president of the Royal and Imperial Moravian Society for the Improvement of Agriculture, Natural Science, and Knowledge of the Country (Agricultural Society).  Once he joined the abbey at age 25 Mendel began teaching math and Greek at the local elementary, and he was quite successful in his first year of teaching was quite successful.  Although he was often uncomfortable speaking to adults, working with children was not a problem.  However, although well-regarded by his  contemporaries, he was unable to pass the teaching certification tests.  He even took additional math courses at the University of Vienna but still could not get certified.  So he continued teaching at the elementary level while looking for another use for his skills.

Malone specifically describes that time in detail, and sees it as highly symbolic of not just Mendel's efforts, bot those of citizen scientists throughout the ages:

Great amateur scientists have often received a considerable amount of education, but it tends to be spotty and sometimes lacking, ironically, in the very area in which the scientists ultimately makes his or her mark.  In Mendel's case, he received more mathematics training than anything else, and that would make it possible for him to apply a mathematical rigor to his experiments with pea plants that was highly unusual for the period.  He also studied some botany, but this was a subject that caused him particular problems.  Because he was a farm boy, he had an ingrained knowledge of plants that caused him to balk at various academic formulations.  When a student refuses to give the answer he or she has been taught, academicians inevitably conclude that the student is stupid rather than reassess their own beliefs.  Brilliant amateurs have always been prone to question the questioner, and that usually gets them into deep trouble.  The end result is often a young person of great talent who has not attained the kind of academic degree or standing that would serve as protection when he or she puts together an unorthodox idea...There is another side to this coin, however.  If Gregor Mendel had in fact passed the tests what would have made him a full-time high school teacher, it is unlikely that he would have had the time to devote to the experiments that would eventually make his name immortal.

After his first failed certification Mendel's first set of experiments included breeding mice in his room to understand how fur color was transmitted.  Animal husbandry was a long-standing art practiced by farmers and approved by the Catholic Church, but was very poorly understood. at the time  Mendel hoped to shed some light on the practice with his mouse-breeding experiments.  Unfortunately the local Bishop did not appreciate the sexual connotations of animal breeding.  Although Napp was a long-standing supporter of scientific experimentation, Mendel was eventually forced to stop the mouse experiments and turned to a more favorable test subject, peas (species Pisum).

Breeding peas requires much painstaking work.  Since they are hermaphroditic each plant contains both a male sex organ (stamen) that releases pollen (sperm), and female organs (pistil and style) that contain eggs.  For Mendel's experiments, he removed the stamen from all flowers so they wouldn't pollinate each other, and instead hand-pollinated the female organs with pollen from a known source.  He then covered each plant to prevent stray pollen coming in through the air or on the legs of an insect.  That way he could confirm the identity of both parent plants and characterize the offspring.  What Mendel looked for was:

• Seed Shape - Smooth and Wrinkled
• Seed Coat Color - White and Grey
• Seed Color - Green and Yellow
• Color of Unripe Pods - Green and Yellow
• Shape of Ripe Pods - Inflated and Constricted
• Stem Length - Tall and Short
• Position of Flowers - All Along Stem and Single Flower at Top of Stem

This wasn't as easy as it sounds.  Before any experiments could begin, Mendel needed to make sure his peas were pure and that a plant showing certain characteristics would only show that characteristic and not any others.  So he spent two years cultivating the peas and making them "True" before ever attempting to cross-breed them.  In Malone's words,

"The fact that he spent so much time laying a rigorous foundation for his experiments is one of the many reasons his work has come to be so highly regarded.  Many people think this was a boring prelude to the experiments to come, but Mendel took so much pleasure in gardening for its own sake that even this preliminary stage must have brought its satisfactions."

Unfortunately none of Mendel's papers or lab notes remain.  He was eventually elected the new Abbot of St. Thomas after Napp's death in 1868, and quickly had too many other duties occupying his time and distracting from the pea work.  A few years later a tornado destroyed his greenhouse, and he passed away a few years later on January 6, 1884. He continued to serve as Abbot until his death on January 6, 1884.  At this point history kicks him while he's down with the burning of his papers in the monastery courtyard by a narrow-minded successor.  

All we have left is a two-part paper presented to the Agricultural Society in 1865.  Interestingly he even sent a copy to Charles Darwin.  But despite the support it would provide to evolutionary theory in later years, the math was too dense for many and his findings languished in obscurity for years.  Again we hear from Malone:

It would be another half century before the technical language would be developed to explain these results.  But Mendel had clearly demonstrated the difference between a phenotype (in which the physical traits are visibly displayed) and a genotype (in which the gene variants are present, and still capable of being passed on to another generation, but are not necessarily visible).  He had started with a theory and ended with confirmation of what eventually came to be called Mendel's laws.  He went on to experiment for a couple of years with a variety of other plants, including snapdragons and maize, which appeared to show that the results he had achieved with his peas would hold true for any plant.

Full recognition of Mendel's' work would wait until 1900, when three separate biologists would rediscover it at nearly the same time.  One of these was William Bateson.  He would eventually coin the word genetics and helped tie Mendel's work to evolutionary theory.  It was controversial for a long time, but "{U]ltimately one major scientist after another came over to the side of the Mendelians, for many particular reasons and one general one: Gregor Mendel's laws proved to be the most useful and logical approach to the new science of genetics."

In the end this poor farm boy with only some academic training and who was denied even the ability to teach science, would end up creating an entire field of biology and teaching scientists of the day that everything they understood was wrong.  Supposedly, in Mendel's later years, when people asked about his lack of recognition, he would say "My day will come".

Fortunately for scientists and citizen scientists everywhere, that day did indeed come.
 

Continuing to Help Hurricane Sandy

Photo Courtesy: SUDS and RocketHub

Hurricane Sandy struck the US nearly three months ago.  While many of us have moved on with our daily lives, the people impacted by the Storm have not.  Their lives were permanently disrupted; they are still working through the storm's effects and will be for years to come.  Much of this is from the financial and psychological impact of the storm.  But there is an ongoing environmental and medical impact that must be dealt with.

The Send Us Your Dirt from Sandy (SUDS) project addresses these very issues.  Researchers are collecting samples from citizen scientists across the region and testing the dirt for cancer-causing carcinogens and other dangerous chemicals  This can be used to map out continuing danger areas and pinpoint areas requiring continued clean-up.  Researchers are already lining up the samples and the testing equipment.  All they need now is financial help.

In the previous SUDS post I talked a lot about KickStarter (which SUDS was using) along with other crowd-funding web sites that let the public donate to important causes.  I am a supporter of SUDS and made a donation myself...unfortunately it was not enough.  The Kickstarter campaign ended and they did not meet their initial goal, meaning all the funds were sent back to donors.  But now they are back with RocketHub.  It's a very similar site but without project minimums, so even if the fall a bit short of the goal their research can continue.

Right now things are looking great for SUDS!  They have just over week left in the campaign but have already raised their minimum amount.  But we all know that's not always enough.  Let's help push them up even higher and increase the scope of their research. It's also a great chance to show the power of citizen science funding it's own projects, and will encourage other researchers to involve the public in funding decisions.  I made my contribution, won't you help too?

Finally, this is a great example of the differences between crowd-funding web sites and picking the best one for your project.  In this case if you don't need the entire amount to get started but can continue with only partial funding, RocketHub may be better than KickStarter.  On the other hand, KickStarter is more well-known and might draw more people to see your project.  There are also differences in how the project is highlighted, the demographics of each site, how much donors on each site contribute, how sites encourage continued donations, etc.  So site selection is critical as project designers look to the public as a source of funding.

Just one more important tip for the next generation of citizen scientists.

The Levels of Citizen Science Involvement - Part 2

Monday night I posted a well-received article on the types and levels of citizen science projects.  It my highest rated post ever (thank you everyone!).  It also let me combine much of the citizen science data I've collected over the last few years.  So I want to go another level today and talk about how this information can help us.  We talked about theory...now let's talk practice.

After all, expanding and improving the practice  of citizen science is what the California Academy of Sciences meeting was all about.

One of the questions I'm most often asked is how scientists can get more people more involved in their projects.  While this can have many answers, much comes down to Motivation.  So this is the perfect test case to see what the activity level models can tell us about ways to improve citizen science motivation and recruitment.  For example:

• Challenge projects are often the most difficult sort of citizen science project to participate in, and can be considered Expert or Co-created depending on the model.  For these, due to the high barrier to entry, large amount of training research required, and time/effort required, tangible rewards (such as cash prizes) seem to be required.  Obviously intellectual curiosity  and civic-mindedness play a role, but keeping people involved requires the promise of more tangible rewards.  Thus the challenge leading to a prize at the end.  Examples of this include the Ansari X-Prizes and various Innocentive initiatives.  Both require expert levels of participation and involve much creation (co-creation) by participants.  In return cash rewards (or the potential for such rewards) keep people motivated.
• At the Medium or Collaborative Level much time and effort is required, so designers should put much time and effort into keeping participants motivated.  Tangible rewards often work for these groups and may be very useful.  But other less tangible rewards can be successfully used too.  One of the most powerful are intellectual rewards (such as extensive training in the science or public recognition of important discoveries)  These are highly prized and can be given often without significant cost, and not only keep the person motivated but also makes them a more productive citizen scientist.  They understand the project more, provide higher-quality data, have more insightful analysis, and provide larger data sets.  Good examples of this are the Sungrazer comet detection project and the Citizen Weather Observing Program.
• At the Novice and Contributory Level people don't require much initial training (in most cases) or require extensive time commitments.  For example, people only need enough training to collect water samples from local streams, or count wildlife in their backyards.  Projects are certainly improved when people stay for a long time or receive extensive training that allows high-quality data, but it's usually not required.  A project can still be successful without large motivation levels.  So successful participation requires stoking immediate interest for recruitment purposes and then just enough other benefits (usually intangible)  for them to continue participating.  This can include civic-mindedness (wanting to help the world) or fostering a sense of dependence (showing how the science will benefit them in their real lives). 
• Distributed computing takes the least amount of effort for citizen scientists to participate in and it doesn't even register on the CAISE model.  But these projects also require a different type of  motivation...it seems just providing feedback and keeping the project in the public's eye is good enough.  So focusing on initial recruitment and getting the program on a person's computer is a project designers most important taks.  Once that is done the program can just keep running.  There isn't much need to continue adding rewards or other motivational techniques to keep the research moving forward.

Now let's dive a bit deeper.  Pulling from my previous research on keys to successful citizen science projects, a few important factors stand out for engaging users.

First is the need to focus on "creating a community" and "educating participants" about the science as keys to success in ecology-based projects.   These are not monetary rewards, but are non-monetary incentives to recruit and motivate participants.  People want to learn more about their projects and the world around them, and appreciate the education provided by joining.  They also find a sense of community and camaraderie that both provides training and makes the work fun.

All this fits perfectly with the Contributory type of participation common with most ecology-based projects.  Grand monetary rewards or extensive public recognition are not required...instead personal recognition (sincere thank yous) and basic science training just as powerful.  Understanding this helps designers tailor projects to their audience.

Second, we also see this with distributed computing projects as well.  The most important success factor is providing feedback.  Not even public recognition, just feedback about the project's accomplishments and confirmation that each person's involvement is worth while.  This is a very low-level of engagement but is enough to meet the needs of people participating with a minimal level of ongoing activity.

Next, let's go back to the highest level of participation, Co-created or Expert levels of activity.  At this level we ask for large commitments of time/energy and project designers must work that much harder to keep people involved.  These are the big projects that require big solutions.  They also seem to require large benefits and tangible rewards...thus the need for cash-based challenges to promote participation.  We've seen this is not required for other levels of citizen science activity, but at the Co-created level we need a greater push to keep large numbers of people involved.  Also, remember that they are the creators and want to receive the credit, both tangible and intangible.  Just like the project designers do.  So it's just fair to provide it.  After all, these biggest challenges also provide the best opportunity for huge advances to the scientific field.  So it's often worth it.

Finally, while we've talked a good game above, can we really be sure that this model works?  Are those success factors and motivation techniques really correlated to the various types/levels of citizen science activity?  Let's find out.  I have a lot of data on existing projects and their levels of activity, so let's run some numbers and see if the model holds up.  So come back soon and we'll walk through everything together.



PREVIOUS POSTS IN THIS SERIES:

• The Levels of Citizen Science Involvement - Part I (Comparing the Models)
• The Levels of Citizen Science Involvement - Part II (Implementing the Models) - Today
• The Levels of Citizen Science Involvement - Part III (Testing the Models) - Coming Soon!

The Levels of Citizen Science Involvement - Part 1

A few months ago I wrote about a California Academy of Sciences meeting on incorporating citizen science into museums and science/technology center activities.  There was a ton of good information provided and a lot of best practices exchanged by the participants.  It was a great resource for project designers and researchers looking to increase their involvement with citizen scientists.  I still highly recommend reviewing the proceedings for tips you can use in your own programs.

An important item mentioned by during her presentation, and an item of particular interest for us, is a proposed series of participation categories for people amateurs involved in scientific research.  It's mentioned briefly but deserves a much lengthier discussion.

Dr. Bonney, the Center for Advancement of Informal Science Education (CAISE), and the team at Cornell's Ornithology Lab, offered these categories back in 2009.  In a recent Ecological Society of America journal article, they are described as:

• Contributory: Generally designed by scientists and for which members of the public primarily contribute data; also includes studies in which scientists analyze citizens' observations, such as those in journal and other records, whether or not those citizens are still alive.
• Collaborative: Generally designed by scientists for which members of the public contribute data but may also help to refine project design, analyze data, or disseminate findings.
• Co-created: Designed by scientists and members of the public working together and for which at least some of the public participants are actively involved in most or all steps of the scientific process; also includes research wholly conceived and implemented by amateur (non-professional) scientists.

There is a lot of value in these categorizations and while I use slightly different ones, the concepts are solid.  Not only do they accurately describe each person, but they are a useful model for different ways to engage and work with each type of citizen scientist. 

Instead of categories I looked at activity levels, and chose Minimal-Novice-Medium-Expert...this helped differentiate distributed computing (with very minimal involvement) from contributory projects that don't require much individual planning, but may require much work from volunteers.  I also think making the highest point based on "Creation" of projects overemphasizes design aspects of citizen science and discounts high-level analysis performed at an expert level, but not "Created".  So combining these models we see the connections with the following graphic:



Citizen Science Categories and Participation Levels
Image Courtesy: OpenScientist

 

This sets everything up, but why a pyramid?  Well, my previous research has shown that as the levels of complexity (or activity) increase, the number of available projects and number of people participating decreases.  So the number of amateur scientists with expert, co-crated levels of activity are extremely rare...and they are shown at the pinnacle.  Meanwhile the large number of other citizen scientists have lower levels of activity but are much more common...they make up a solid base.  I also think it shows how everything still relies on basic contribution levels of activity.  Even if someone wants to participate at a high level, they need many people collecting data and performing the initial analyses that they can build on for their expert level analysis.

Can we take this model any further?  Next let's look at the types of activities performed by citizen scientists at these activity levels (read more about each of these types in my previous postings).  I believe this is the best way to approach the problem since the amount of activity should be correlated to the type of activity people perform.  But it does fit somewhat with CAISE model:

• Distributed Computing: For the vast majority of users this just involves downloading onto a computer and pressing "Run".  So I'd have to put it at the lowest level of involvement...Contributory.  And it does contribute data (in the form of work-hours) to the project.  But the person's involvement is extremely small leading me to categorize it in the Minimal category, below the Contributory level already at the base of the pyramid.
• Observational Measurement: In these projects scientists provide a collection of images or specimens for citizen scientists to measure, or provide specific tools for amateurs to use for measuring things they encounter on their own.  So it can fit at the Contributory level for this first set of activities, up to the Collaborative where citizen scientists actively seek out their own data/specimens to measure.
• Observational Analysis:  These are often identification projects that provide a set of data or images to identify.  So it goes beyond mere measurement by having citizen scientists propose actual conclusions.  In most cases these projects have all been Contributory in nature but future projects could extend into the Collaborative level as well.
• Transcription: This is very similar to observation measurement in that a project designer provides the raw material and the citizen scientists transcribes (or translates) the data.  So this is a Contributory level of activity.
• Challenge: This one can run the gamut of involvement levels.  It can involve designing a new rocket ship (Co-created), spotting a hard-to-find bird species (Contributory), or having your computer find a new prime number through distributed computing (Minimal).  So we must place it in all levels of the pyramid.
• Game: Projects in this category often start at the Contributory level where people add little pieces to a larger puzzle in a game-like format.  But many also involve teams and higher-level problems requiring extensive research, analysis, and effort.  So it can also fit in the Collaborative pyramid level as well.
• Research Analysis: There is a fine line between this category, Observational Analysis, and Observational Measurement.  But this is normally a more involved analysis that requires independent knowledge or research performed by the citizen scientist (Observational Measurement often requires comparison with known data provided by the project designer).  So I consider this a Collaborative level of activity, though admittedly some could argue for higher or lower placement on the pyramid.
• Collection: Most examples of this have all been in the Contributory level, where the collection is the contribution making up a new data set.  While this could be expanded and become a citizen scientists own project that asks others to collect large amounts of information for them, I don't know of any good examples yet.  So it will remain in the first pyramid level for now.

It's important to note that items in the "Expert" or "Co-created" categories are admittedly less defined.  So far there are fewer projects and fewer people involved at those levels to fully describe them.  I also don't think this area has "matured" enough to really pin things down.  A lot can still develop with plenty of room for new innovations and developments.  So it's an exciting field to watch and develop some early hypotheses, but precise definitions won't be available for some time.

Combining our various models the graphic now looks like this starred items may appear at multiple activity levels):



Types of Citizen Science Activities by Category and Participation Level
Image Courtesy: OpenScientist

 

This is very useful information...but how can it help us?  My next post will talk all about that and more...so Stay Tuned!