This chapter mainly discusses the process of manipulating animals and individuals’ behaviors by rewards.
The first part introduces the concept of reward as a instrument to reinforce some behavior. S-R Bond and Operant Conditioning Theory are also discussed in this section. The timing of reward is very important to the efficiency of reinforcement, which influences the design of reward timing on online communities. For instance, I suspect a leader boarder with high updating speed with motive more participation on online users compared to a slow updating one.

The second and the third sections discuss several theories and their motivation categories. Responses theories and Stimulus theories are representatives of intrinsic motivation and extrinsic rewards. The over justification effect which claims that “external rewards for behavior sometimes reduce intrinsic motivation to engage in those behaviors” is correspondent with the experimental results in  Gneezy&Rustichini.

In the end, behavior economics is briefly explained which opens a new window to study individuals’ behavior in open economics.

If we want to apply reward mechanism on online communities, monetary incentive is not a good idea due to the over justification effect. However, questions like “When to reward?” and “How frequent to rewards” are interesting to explore.  

I was interested in the idea of secondary reinforcers, brought up by this reading. A secondary reinforces is a "formerly neutral stimulus that have gained reinforces capacity of their own by being associated with primary reinforcers."

The author claims that secondary reinforcers explain a lot of human behavior - the most common one being money. I was trying to think of how secondary reinforcers are used in online communities and was having a hard time coming up with a good example. I wonder if the point system on our class website could be a secondary reinforcer because there is a (loose) connection between getting points and receiving a good participation grade. The problem is that points aren't necessarily formerly neutral - we use points so much as a status symbol in online communities, they could be motivating in their own right.

Any other ideas on how secondary reinforces are used in online communities?

"Organisms tend to respond to stimuli that rae just a little more complex than the optimum.  There are called pacer stimuli.  Once an individual's preferred complexity level is established, any change in stimulus preference will be in the direction of greater complexity, not less." -192

I found this quote interesting.  In some of the class presentations earlier in the semester, the mention of challenge as an incentives for deepening level of involvement in e-communities.  There are potentially some take-aways here about the benefits of creating tiered levels of complexity tied to user status as a potential motivating force. 

Lisa McLaughlin

Thinking back to the first paper we read about MovieLens, we saw that framing the "rewards" of participation in a certain way can change the amount of participation.  We may also ask, considering this article, whether the timing of giving rewards directly affects online community members' participation in the community.  Chris gave the example of feedback on wiki articles but also, we could consider other rewards as well, such as increases in status or secondary reinforcers such as points or site "currency" that can be used for various activities on the site.

One quirky example I can think of is the site iminlikewithyou.com.  Virtually every interaction on the site, every "game," involves bidding with points and one person coming out as "the winner" who receives recognition and the ability to contact the person who initiated the game.  This is an example of an obvious and immediate reward that seems to be effective in connecting people.

After reading the whole paper, it seems that reinforcement theory boils down to a Main Point: Consequences influence behavior. It means that people do things because they know other things will follow. Thus, depending upon the type of consequence that follows, people will produce some behaviors and avoid others. Pretty simple, realistic and makes sense.

The three Rules describe the logical outcomes which typically occur after consequences.

1. Consequences which give Rewards increase a behavior.
2. Consequences which give Punishments decrease a behavior.
3. Consequences which give neither Rewards nor Punishments extinguish a behavior.

Here is one of the best examples of reinforcement I've ever heard in terms of reinforcement. It came from an assistant football coach at a college. A little background: Some football players have trouble getting to team meetings. When this happens the coaches want to Punish the players so they will be on time. What to do?

The standard answer is extra exercise. When the team is in a workout, at the end of the session the coaches identify the tardy players and make them run extra laps or do more pushups, right? (When on this team, Do miss a team meeting, Get extra laps).

Well, this coach had a better idea. At the end of the workout he called everyone together, identified the tardy players who missed the team meeting. Then he made the rest of the team run extra laps while the tardy ones sat and watched. The coach claimed that this application had to be given only once a year. And I believe it is true.

But the reinforcement theory also seems to have limitations. For example, sometimes it is hard to identify rewards and punishment. You have to control all sources of reinforcement; and punishment might be difficult to do well. So it is really a tough question for community manager to think how to apply this theory really effectively in practice.

I think it is quite obvious that people need rewards and incentives to participate and contribute to online communities. Therefore, the key question is how to design the rewards. I think it would be useful to compile a list of basic rewards that people can get from online communities:

• fun
• recognition
• a sense of belongings, comradery
• knowledge (self-actualization)
• monetory reward
• a sense of duty fulfillment

I felt some concepts can be applied to online communities, such that “building behavior by shaping” (p183) may be applied to building newcomers’ behaviors, if rules of posting are very complicated. For example, a newcomer first will be brought to go through steps about posting messages on the forum along with a demonstration. And then, when the newcomer posts on the forum, existing members provide feedback immediately. If the post is not appropriate, she will receive feedback about this. However, if she does it right, she will receive some positive feedback to reinforce these behaviors.

This is an interesting paper on human and animal behavior. Two concepts I found interesting are secondary reinforces and delay between response and reinforcement.

In some communities the primary objective can be hard to reach or it can take long time. In such cases, providing secondary reinforces will ensure that the members get a positive feedback for their behavior.

The delay between response and reinforcement should be considered when there is a manual process involved like moderation, voting, etc. Additionally I think the duration of the reenforcement is also important. If the entries in a leader board in a community like Yahoo Answers keeps changing every 10 minutes then users may not get as good reinforcement as when their names stay in the leader board for a day or week.

I was intrigued by the spontaneous alternation behavior mentioned by the authors. Brief recap: when there is no reward in a T-maze, rats tend to alternate running different routes on successive trials. The two proposed explanations are response inhibition and stimulus satiation. Quoting from page 192, the authors write: "... the animal becomes satiated for the stimulus just experienced and so goes to the alternate stimulus on the next trial because it is either more novel or more informational" (italics mine).

The key to this theory seems to be the absence of reward. Obviously, a reward would reinforce the previous traveled route in the T-maze experiment. Without rewards or incentives, the goals of the subject seem to change.

I wonder how this could be applied to online communities. If there are no apparent rewards or incentives (at least initially), does this cause visitors/members of the site to explore more and seek paths that are "more informational"? To me, it seems like an "invisible carrot" – people expect there will be some reward somewhere but exploring must continue until it's found. It seems like there might be some interesting avenues to explore in this direction. Perhaps a socialization period with no rewards could be put in place to encourage new members to explore before "throwing them the carrot on the stick."

I guess we shouldn't take this reading too literally and assume that feeding community members sugar cubes as positive reinforcement is a good idea.

But the idea of rewarding users is a good idea, and the reading goes over the basics on how to do it: Make sure the user knows why he/she is being rewarded, give the reward immediately after the desired behavior, etc.

For humans, the reward could be in the form of points, money, fun, or whatever. I think "fun" would be the most challenging, though also the most effective in the long run. Not to keep referencing previous speakers, but Luis von Ahn gave some good examples on how to keep people doing stuff for you by making the tasks fun.

So many takeaway points from this article, I don't know where to begin summing up.  Without a major psychological or economic background, most of the in-depth writing went over my head.  Here's what confused me - (btw guys, if I use a term incorrectly please don't hesitate to point it out - you'll help me learn [thanks] and your smartness status will increase at the same time - so it's a win win win) <-- insert your own third win here so I don't leave out your own internal motivator

Figuring out which reinforcer schedule or type of reinforcer applies to a given reward seems like a matter of context.  Quick refresher:

Types of Reinforcers:

    Positive Reinforcer - presentation increases likelihood of response

    Negative Reinforcer - removal decreases likelihood of response

    Punisher 1 - presentation decreases likelihood of response

    Punisher 2 - removal increases likelihood of response

Getting to the point - So I'm having trouble figuring how Positive Reinforcer and Punisher 1 are different from one another except by wording.  Same for Negative Reinforcer and Punisher 2.

    Jimmy is more likely to do his homework when he gets a cookie (Positive Reinforcer)

    Jimmy is less likely to not do his homework when he gets a cookie (Punisher 1)  

and:

    Jimmy is less likely to do his homework when his cookie is removed (Negative Reinforcer)

    Jimmy is more likely to not do his homework when his cookie is removed (Punisher 2)  

One concept I either missed or would have liked to seen discussed is what effect the presentation or removal of a reinforcer will have if it's available before.  For example, if Jimmy has plenty of cookies, and he's offered more cookies, he'll be more likely to respond to an alternative reinforcement - but how many cookies does it take to get to a tipping point?  

One last thing - from the chart on page 186 that displays the relationship between time elapsed, reinforcer presentations, and cumulative responses, it appears that if the graph were extended to display more cumulative responses, the variable ratio schedule will eventually exceed the fixed ratio schedule.  This makes sense because if the likelihood of a reinforcement being presented isn't set in stone, but is guaranteed to happen sometimes (for example, 15% of the time), people may translate their uncertainty into increased effort.  

Beck indicates that variable ratio and interval rewards are good ways to motivate behavior. Variable interval will encourage a steady response from the user. With variable ratio the amount of use increase. The best example of variable ratio is slot machines, because people playing the slot machines don’t know how many times they will need to pull the lever to win and this is what encourages use.

In online communities both of these mechanism are built into the design somewhat fundamentally. In online forums it is never clear when you will get a response. With social book marking sites it is never entirely clear which submission will net the most positive attention. Because these two features are built into many websites now, a baseline of reward structures can be assumed.
 

I have always imagined that Paradise will be a kind of library.

-Jorge Luis Borges

I found the concept of Variable Ratio (VR) schedule very interesting,
particularly the kind of incentives it can induce “…the more they gamble, the
more reinforcers they collect so they keep gambling even more” – all the while
not really sure when they are going to get the next reinforcer.

It’s an interesting mechanism that can be used for a site
like PagalGuy – the community I’m studying. Currently they do not have any
place where people can take test questions online. They could have a test where
the student doesn’t know the how many questions he must answer correctly in a
set amount of time. He does know that the system will let him know (reinforce)
if he reaches a standard that the system dynamically computes from other
students test to let him know if he improves his percentile by a certain
number.

The process might get demanding and a little unpleasant as
the authors state but so is the real test!

Beck's discussion of variable reward schedules seems to connect to the organizational socialization literature. With a variable time schedule, you can (it seems) get consistent performance. (Though I do not understand why the subjects don't take a break after receiving a reward on the variable schedule, like the fixed reward schedule subjects.) The downside, as Beck says, is that it is "demanding and often considered unpleasant." This sounds a lot like the variable socialization tactic.

I'm also wondering whether some ideas from prospect theory can be applied to variable vs. fixed reward schedules, in that people would prefer fixed (certain) reward amounts and frequency over variable (uncertain), even if the variable reward amount were lower. This might particularly matter in settings where people are choosing between spaces/communities with different rewards structures (rather than just measuring how subjects perform based on the reward structure).

Interesting overview of a wide variety of behavioral psychology concepts.

The key challenge with behavioral psych (as they have since figured out) is that it really requires  highly controlled conditions with specific reinforcing or punishing stimuli (i.e. pain or pleasure stimuli, often direct brain electrification).

Otherwise, as in the case out in the real world, you can never quite be sure if you are reinforcing or punishing or both, or worse yet what exactly ARE you reinforcing or punishing?

For example, we receive so many stimuli on a regular basis, that it is often the case that a reward or punishment becomes associated with one or more actions which may or may not be the action that the reward/punishment issuer is trying to affect.

Online communities may provide some environmental constraints very similar to laboratories in that we know that the user is using his/her computer, acting in a designed and highly controlled space (like a maze), etc.  What we do not know (and which may or may not be relevant and/or confounding) is what the user's environment is providing as stimuli while the user is participating in the online community.

--------------------------------------------------------
PHartzog@umich.edu
--------------------------------------------------------
The Universe is made up of stories, not atoms.
--Muriel Rukeyser

The theory focuses on the degree to which people
endorse their actions at the highest level of reflection and engage in
the actions with a full sense of choice.

Beck’s section on Self Determination Theory seemed to me to
be a great baseline for how to better build equilibrium within an online community:

1)      Provide
users with autonomy; within cgsociety.org you can elect to become a paying
member or not.  They both have
similarities up to a point and then there is a vast differentiation among
both.  At this point the site gives you
the feeling that you are free to choose the best possible alternative to meet
your needs as a digital artist. 

2)      Reinforce
the feeling of a users sense of competence; 
this is where it is vital to know your audience very well so as to
anticipate what they can and can’t do. 
Polling members of Epicurious.com on whether or not a bot should be used
to make necessary edits/updates to the hundreds of thousands of recipes that
exist on the site would probably be counter productive as it would more than
likely intimidate and confuse members who are typically not technically
oriented .

3)      Relatedness,  cgsociety.org does an extremely good job of
promoting a true sense of connectivity and even a level of interdependence that
works to the advantage of the community by holding different levels of
competitions.  It is not uncommon for the
Forum leaders to hold 48 hour group competitions where they will ask members to
consolidate their talents and skills and share amongst each other to reach a
competitive goal (ex. Create a 3d model of the Louvre grounds in Paris).  The group must then coordinate itself and
divide the tasks necessary while cooperating to reach their goal.

Incentive:  In case you were wondering what the incentive
is for the above competition, to enhance individual skills through group
work.  Not unlike the foundations course SI501!

Most communities promote rewards over punishments. There are several
versions of rewards, several levels of achievements, but there is one
way of punishment: Simply kick the member out of the community.

I think that if there is a layered structure of punishments
(negatives), then this would be a big deterrent for the growth of the
community. Members usually do not like to display -ve thing on their
profile/comment.

eBay is an exception to ths rule, maybe because people come there not for the community but for the business.