Submitted by Tracy Liu on Mon, 03/17/2008 - 21:56.
1
point
This paper investigates the effect of monetary incentive on individuals’ behaviors using laboratory experiments. The most interesting finding is that when the monetary incentive is small enough, it will be detrimental to subjects’ performance, especially when the task is intrinsic incentive oriented (Task2) Furthermore, individuals always fail to notice the existence of this trap(The test followed by these two experiment).
The results of these two experiments break the rule that “”A small payment is better than nothing", which has implication on experimental economic methodology. This interesting phenomenon could be explained by many arguments including “people are afraid of looking ‘cheap’” and “insulting compensations”, and the most promising reason is that the introduction of money changes individuals’ reference point under the incomplete contract.
Under these results, shall we claim that monetary incentive generally will not succeed in online communities as it seems that we could only implement small amount of monetary incentive in online communities? How could we give every online user $10 given the fact of more than thousands of online users?
This was an incredily interesting and counterintuitive study that tested the general beliefs that economic models predict that the more compesnation offered, the more perforamance will increase, while psychological studies predict the opposite. As Tracy already mentioned, these researchers only confirmed this when compensation was offered (as in, between a choice of compensation or no compesnation) but that small sums--especially in those involving intrinsic motivation like charity projects--can often be detrimental to performance and/or motivation.
The implications of this research, as least for me, tell us to value possible pscyhological effects over obvious economic models. Especially after the dot com bubble burst, and the recent web 2.0 craze, this makes sense. For example, on a community Wiki, if people are building knowledge and feel a part of their community then it makes sense to allow the intrinsic motivation of doing something for the good of the community to motivate them. A start-up may be tempted to offer a small sum for people to contribute to a new Wiki but this research shows that they may be better off not paying the contributors anything, or perhaps offering the possibility of earning a large sum of money that surpasses any ordinary amount.
I agree that in applying this to online communities it may be better not to offer any monetary incentive for doing a task that has high intrinsic motivation. But I want to clarify the framework that I believe is useful in making this claim.
The argument that I found most compelling in this paper is that these effects can be explained by imperfect information and incomplete contracts. People sign up for an experiment or to help raise money, and they have imperfect information about what is being asked of them. Further explanations upon arrival give them more information about their contract. If a new incentive is introduced at that point, then the participant's incentives change. They have completed the first part of the contract (and acted on whatever incentive was offered to begin with - intrinsic motivation to help raise money or baseline monetary incentive). Now they are entering a new contract with different incentives.
So I don't think it's necessarily wrong to offer monetary incentives to get participation in an online community. But I think it is important to think about what contractual obligation people may already feel under. If there is no contract, then a monetary incentive may help increase participation. If people are already invested in participating, but you still want to increase participation, then you may want to create a separate place in the community to give this monetary incentive.
For example, if you want people to answer more newcomer questions, don't just start paying 10 cents per question answered. This will disrupt the contract of oldtimers who answer questions for intrinsic motivation. Instead, create a separate forum where newcomers can ask questions for 1 cent and oldtimers can answer questions for 5 cents - that way the old contract is still in effect and a new contract involving monetary incentives can be put into place. [I'm not sure this set up would really work, but I hope it illustrates the idea...]
I'm glad you mentioned this concept of contractual obligations, and the extent to which they shift users' sense of obligation and perception of their relationship with the site.
I think this raises a sustainability issue: It doesn't make sense to offer a monetary incentive that you won't be able to maintain over the long haul. Setting a precedent of monetary incentives, may shape users' expectations around what the overall purpose of the site is and how they're framing their involvement in that community, beyond just shaping the the extent to which they're motivated to contribute.
I was wondering if providing monetary incentives at all might challenge the notion of "community" by identifying your site more as a source of income than a place where you feel a sense of connection with others? Do others have thoughts on this?
Submitted by Matt Adamo on Sun, 03/23/2008 - 15:23.
0
points
Challenging the notion of community is sort of another way of saying
that monetary incentives will displace intrinsic motivation. But I
think the extent to which "community" vs. "source of income" is a
problem depends on the goals of your site. If you're mostly concerned
with generating content (and thus ad revenue), do you care if your
users feel connected to each other? On the flipside, paying for contributions in a community meant to foster connections with other members would be counterproductive, and in some cases, just plain creepy.
Submitted by LizBlankenship on Fri, 03/21/2008 - 15:25.
2
points
I'm glad you brought up the idea of a start-up paying for contributions because I have a fun example to add to the mix. =)
Mahalo, a website I considered studying for this class, offers typically $10-$15 for anyone who writes a Search Engine Result Page (SERP). These SERPs are pages for popular search engine queries that are typically pretty great for a first stop on the web about a topic you don't know much about. I wrote a page for them on a band, and it includes all the great sites you *could* find through Google, but you don't have to sift through the really ad-infested or redundant pages to find what you want. They also have a great set of how to pages.
Anyhow, Mahalo seems to have been moderately successful so far, though there are still all sorts of questions about scalability. Will they continue to have enough funding after their seed funds run out? Google Ads may not suffice. And they have to pay a significant amount for their full time staff as well.
Anyway, I wonder if Mahalo's $10+ is significant enough for people to make them motivated to contribute a lot, or if it goes back to people simply being intrinsically motivated to contribute valuable SERPs. I really wonder if this will be able to last.
I like that you brought Mahalo up Liz, because it seems like it is the best example of an actual online community that offers monetary rewards for contribution. Google Knol, when and if it ever comes out, seems different in its premise but similar in execution to Mahalo.
I agree with you that Mahalo probably won't be able to last on this model. Though $10 per article is a sizeable amount, not the small amount shown in the article to reduce contibutions, I do think that for most people, doing edits on Mahalo for money would quickly start seeming like work instead of the intrinsic benefits of helping the community when it is unpaid. Perhaps this is why they have not met their goals of number of articles written - the money didn't end up being the draw that the community creator thought it would be. In my opinion, it seems like people need some sort of intrinsic motivation, even on a site with monetary rewards, because this is part of what forms their relationship and connection with the site. Money isn't always enough of a reason in and of itself.
I find Mahalo to be a very interesting concept. They do pay money for building the pages, but they also focus on the community aspect (They did not have it earlier, but added it few months ago). Maybe this concept might be OK for search (given that the company organizes vast amount of data on the internet in a very neat way), but might not be useful for other smaller communities.
Submitted by Daniel Zhou on Sat, 03/22/2008 - 19:22.
2
points
Instead of using several small amount of monetory rewards, an online community might introduce a fairly large prize/bonus to encourage contribution. I'd like to see how that could affect contribution.
Submitted by Rozaidi Rashid on Sun, 03/23/2008 - 17:56.
2
points
We had this Idea Management System at the Central Bank of Malaysia, where staff contribute ideas on anything that is useful to the organization. Staff gets points for contributing and rating ideas by others. There are threshold points, where staff can convert the points into inexpensive things like umbrellas or thumbdisks. Periodically, top ideas get to be presented to the management. Accepted ones are recognized in terms of end-of-year bonuses. Occasionally there are also “contests” that call for ideas to resolve certain problems or goals (e.g. save energy) that carries a cash prize ($500 for winner).
The observation is that more ideas are generated for the contests than the other one. Both have rewards, but one is large and uncertain, while the other is small and certain. But the small one is not too small to be insulting – the goods are useful – but there is also that very small chance of getting the big reward at end of year.
Online communities are able to provide something between non-monetary incentive and small monetary incentive. This is the point given on the communities. Point system is like something between encouraging altruism contribution and material incentive, in my opinion, it can be a matter to improve the adhesiveness in exchanging the effort with award. In other word, it makes it less transparent in comparing effort with award thus softening assaulting effects. In addition, the virtual sitting makes this exchanging system able to be easily adjusted and manipulated. For users, they can interpret the point award differently and for the administrators, they can manipulate the interpretation individually to encourage users' participation.
It would also be interesting to see whether similar happens in the online point system: like very few points have negative effect.
Since most online communities involves much volunteering contribution, thus adding point incentive is tricky. There might be something else that can affect monetary/point incentive. For example, on taskcn.com i observed, since people had very small chance to win a competition thus getting money (a task usually gets 65 participators but only one winner), there is no correlation between the real monetary award to the number of people's participation; while on Baidu.com, where people answer questions to earn virtual points, there is significant correlation between the number of answers and the number of points offered. A remarkable thing is there only 60% questions got at least one answer. This basically means when people had more choice, they follow more points, if not, and the intrinsic incentive is high, they still want to do it.
Submitted by John Blair on Thu, 03/20/2008 - 22:22.
0
points
John Blair
Many similarities between the Beck and Gneezy-Rustichini readings. I combine the postings here due to this similarity. I preferred the Gneezy article more, as it seemed that the explanations of the analysis were clearer and succinct. I also agree with Bem’s (Gneezy-Rustichini page 803) suggestion that people interpret their actions as any outsider does, by looking at the reasons and motives for what they do. I think this appealed to me because it focused more on the individual making a decision as opposed to group collective. One of my problems with the behavioral economics (meaning the combination of social psychology and economics) is that is seems to forego the individual being individuals. Obviously there is overwhelming evidence supporting group behavior (even when group is randomly assembled with individuals having no previous knowledge of each other). I just have a problem not having individual choice being represented.
I do wonder why the reward results in both articles identified that performance was better with no reward or a high reward. There seems to be no middle ground and the rationale for this I think may be due to individual perception of the reward – obviously they didn’t think it was enough to motivate them to be incented to perform well. But why would they not perform at least as well as the no reward group? Were they insulted (as suggested in Gneezy-Rustichini page 805)?
From the Beck reading, it’s hard to argue with Thorndike’s Law of Effect (effect of satisfying behavior is rewarded/strengthened, annoying behavior is punished – page 179) or Skinner’s reinforce concept (page 182).
Application to ecommunities is relatively straight forward. If you choose to offer an incentive for something make it worth their while – or don’t offer them anything, but this would require some sort of intrinsic motivator for them in order to be successful.
I think this reading was valuable, despite the fact that few e-communities offer monetary rewards. I would be curious to know if other types of incentives (non-monetary) also have the same displacing effect on intrinsic rewards. Would in-system rewards (privileges, "points", recognition) cause the same resentment and/or contractual reevaluation that resulted from the addition of monetary rewards? What if these non-monetary rewards still carry monetary value? I suspect even in the latter case that these rewards could be viewed in a different light than cold, hard cash.
Great questions Dustin. I am also wondering about the same stuff so I am going to take a small stab at offering my two cents. As several people pointed out above, it is definitely a little hard to compare the reading to online communities because the structure just seems a little distant. On one hand, we have the experimenters gathering a group of people, while for most of our online communities, people join on their own.
However, having said that users join online communities because they are already motivate to do so (for various of reasons depending on the community and the person's needs), I'm seem to be doomed and destined to say either:
1. introduction of monetary rewards will negatively impact activity in online communities, because people who were once intrinstically motivated, would find the monetary reward as an extrinsic motivation. Therefore, people who truely cared about the community may be turned off, while other users may become active simply because of the newly introduced monetary rewards. As a result, the community may be harmed as a whole.
2. introduction of monetary rewards will positively impact activity in online communities, because people who were once intrinsically motivated, will have even more incentives to do contribute. Similarly, users who were dormant or not as inactive, would be motivated to participate more. As a result, the community would seem to thrive.
It turns out that both perspectives are really quite flawed. Obviously, the authors dinstinguish between monetary "size" so that itself might change how a community reacts to monetary incentives. Thats why I think you raised an excellent idea, about the case where non-monetary rewards still might lead to some sort of monetary value or gain. I think along with the "size" of the monetary reward, whether or not the non-monetary rewards offered by an online community under analysis, would at some point translate into monetary value, is another good way to measure how a community might react. For example, on a community that has e-commerce functionality, highly active people may be awarded with non-monetary reards such as getting a higher "reputation" or rank higher on a leaderboard, will probably also get more monetary gains.
Agree! The monetary incentives seem to have an immediate effect on motivation, Nevertheless, the quality of the knowledge shared can be inferior, and the attitude that knowledge is a private and non collective good is enforced. Once knowledge is shared only because monetary rewards are obtained, knowledge sharing will decrease when these rewards are withdrawn. Additionally to these consequences, high costs led us to focus on non monetary rewards that may not have an immediate, but a long-term impact on motivation. Therefore the use of monetary incentives should only be used with cautious.
Overall, monetary rewards may have an immediate effect on motivation to share knowledge but at the same bear the risk of spoiling users. However, monetary incentives can be used to start a knowledge management system and to incentivize users from time to time. Yet, in the long term users should be incentivized non monetarily for sharing their knowledge.
One of the design claims I leant from this reading is that: “Positive but small enough compensations, there is a reduction in performance as compared with the zero compensation.”
I just wonder how small a small amount is in an online community setting, since “small” is a concept that you need to compare it with others. Also, I am confused about the sequential vs. single-stage set up. Authors explained there is a shift in perception of the contract, after the introduction of additional compensation for correct answers. (p 804) The perception about the game starts from “Sixty NIS was paid for showing up,” and then changes to “The activity of answering the questions is now paid by rate”. And the rate (10 cents) is low if it’s compared with 60 NIS. Since the conception has been changed, can this example be seen as a sequential game? If an amount of money is offered by participating in an online community once the site has been created, and there is no other values for participants to compare with, will participants be motivated to contribute more by offering monitory incentives?
I was curious about the initial payment and how it might have framed the rest of the experiment. 60 NIS at the exchange rate given in the footnotes is about $17 which is the average hourly wage for economic experiments. That was just for showing up.
How might the results been different if there was a much smaller show-up payment? That is, how much did the 60 NIS show-up fee frame the marginal payment of 10 cents for each correct answer in a negative light? I would be interested to know what the results would look like with a much smaller show-up fee, where the incentive is almost completely on performance of the task.
I think a good way to combat paying too little would be to have some kind of lottery system where a user might be paid a large sum with a level of uncertainty. So instead of getting 10 cents each time you do something, you might have a chance to win $100 or something. Amortized, it could end up that you really got 5 cents for each "action," but the perceived payoff is what counts.
This is sort of what Blingo does. It's basically a frontend for Google, but each time you do a search, you have a chance of winning a cool prize.
I think Amazon.com's mechanical turk suffers the too small problem. I tried doing some tasks but felt the super small sum of doing them was not worth my time. If the tasks were more fun to do, I might have just done them for that.
Submitted by Satyendra on Sun, 03/23/2008 - 03:01.
1
point
Although the authors say that the explanation based on attribution
theory is does not really explain the behavior here I think that it plays a
vital part. If a subject observes
himself doing a task for which no exogenous incentive is provided then he
perceives his motives as being driven by intrinsic motivation. When he is
offered a monetary incentive, irrespective of whether the activity is
sequential or not, it directly associates his effort with a price. Now when he
evaluates his opportunities he realizes that that he will get the original
amount of payment regardless of the questions he answers. Regardless of what he
thought the original 60 INS were for (showing up or answering all questions) he
will evaluate the effort he needs to put in and decide whether the extra effort
is worth the extra money he will be paid.
An interesting experiment would be to see whether how extra
information can change the behavior. For example if an agent was ‘told’ that
his principal was one of the very few principals who actually chose to pay him 10
cents compared to others, then the motivation to do well might again be
intrinsic irrespective of whether the amount paid was 10 cents or 40 cents –
maybe the degree of effort would change but the result for 10 cents might still
be an improvement – because now the agent values his outcome as a intrinsic duty,
and not only in terms of the 10 cents.
So can information about the motivations actually make a
difference to the results despite the money offered ?
I think that while small monetary compensations are detrimental as a motivating factor, small non-monetary rewards can still have some motivating effect. Antecdotally, I have seen this behavior in particular at trade shows and conferences, where booth operators can make attendees jump through hoops and give out all sorts of information for the "prize" of a t-shirt or promotional USB thumb drive. In this case, I doubt that conference attendees would sit thru a half hour presentation for the cash value of these products.
While classic economic theory would prodict that we would naturally abstract the cash equivalent value for rewards we receive, I think we have a hard time decoupling objects and their value when they are right in front of us. Consider the difference, say, between playing pool in a bar for a $6 wager, versus playing pool for who buys the next round of drinks. The $6 is an abstract figure, and your utility of it is probably subject to a discount factor which the beer is probably not. You may consider $6 "beer tomorrow" whereas the alternative is "beer today".
I would be interested to see the effect on these experiments of offering low-value non-monetary compensation ie offering candy to kids in proportion to how much money they raise and see if that has any impact on effort outcomes. My guess is that the more immediate the enjoyment of the compensation is is an additional term to the magnitude of compensation.
Behavioral Economics studies are always interesting. Sometimes these studies quantity human values like: compassion, love, racial discrimination, etc.
The study involving day care center was mentioned in Freakanomics. It shows how money can change the behavior of a parent picking their child from day care.
I don't recall any online community being studied in our class provides any explicit monetary incentive. But I think a new wave of services that rely on monetary services are coming. Example - the Google Knoll a.k.a. the Wikipedia killer. According to this post by a Google VP, Knol will allow users to maintain articles on topics in which they are experts. Unlike Wikipedia, the articles will promote the authors name (like the cover of a book) and also provide a share of the ad revenue from that page. When this service is introduced it will be interesting to compare the success of Knol and Wikipedia. It will be another behavioral economics study, to show the effect of money on contributing to public goods. From this paper, my prediction is that the monetary reward for most users will be too small thereby making the motivation extrinsic. So many users will find the monetary reward insufficient and would switch back to Wikipedia.
The design claim, I infer from this paper is that – when offering monetary reward, it should be sufficient to provide extrinsic motivation greater than the intrinsic motivation provided without any rewards.
As I tried to break down this article into takeaway points, I started to understand why changing or proposing rewards usually has a negative effect on intrinsic motivation. (for most of the article, I had a problem with the author's argument that introducing a reward will have a negative effect on internal motivation, rather than just an impact that could be positive or negative. Why must internal and external rewards conflict?) If people interpret their actions from an outsider's perspective (p13), the internal reward won't be considered if an external reward exists. People would base their interpretation for another person's motive on an external reward - "that's why he's doing it" - and wouldn't need to consider alternatives. From an outsider's perspective, the external and internal rewards match.
Takeaway points:
1) People will be less likely to expend effort toward achieving a goal if an external reward is offered and the compensation is small. They might be a) insulted (their effort isn't valued) or b) ashamed (they don't want to look cheap)
2) Appealing to a person's internal motives is difficult. If a person values an internal reward more than another person makes it out to be valued, or if another person offers someone possible internal rewards as incentives and leaves out the particular internal reward that would motivate them, then they will be less likely to expend effort toward achieving a goal.
3) Small compensation changes the way people interpret rewards (internal --> external), so the reward becomes small.
Submitted by Sean Munson on Sun, 03/23/2008 - 14:34.
1
point
One challenge for designers looking to give rewards for tasks that are intrinsically rewarding will be to figure out what additional rewards they can build in that won't trigger this perceptual shift in whether or not a task is intrinsically or extrinsically rewarding.
It also seems that the answer will vary basd on the task and user. Awarding points may work well to support intrinsic rewards in many systems; providing users who are intrinsically motivated to contribute to a community with feedback that they are indeed contributing. There may be other situations where they cause problems. To illustrate, I'll use an example from one of the CHI teams that presented at FIRST this past Friday (since it's fresh on my mind and many people in class where there). The team proposed a system (pdf) that allowed experts to answer questions from case managers serving the homeless. The team had designed in a reputation system (based on point assignment) that was intended (I think) to help case managers identify and reward trusted sources. A number of people in the panel felt that the idea of rewards or treating the system as a game with points would diminish the volunteer experts' satisfaction with their volunteerism. Unfortunately, the theory and experimental results that could guide the team as they weigh the value of having a point system haven't been developed yet.
Figuring out when and which rewards can work to reinforce intrinsic motivations vs. triggering perceptual shifts to seeing a task as extrinsically motivating seems like an interesting design challenge for further experimentation and theory development.
This article was very relevant to Newsvine’s community model. Newsvine tries to motivate users by giving them 90% of the ad revenue that their articles generate. After about 6 months of use this had netted me a whopping $0.06. This amount of money is more insulting than motivating and I often forget that there are earnings to be had at all. The main motivator for me is submitting content that will get interesting discussions started.
At this point though Newsvine is obligated to continue with their financial incentives. If they were to quit their earnings programs the top users--who are probably making some money from it--would be alienated and might defect to another website community. Fortunately this aspect of Newsvine is not a major portion of the site or else it may face more significant criticism from users.
I have always imagined that Paradise will be a kind of library.
This is not a very deep insight, but in this case I would argue that paying the money is not so much a reward of contribution, but a greenwashing message demonstrating that Newsvine is not exploiting it's users' contributions.
For some top users this might be different, but for the most part I would argue that what I mentioned is the case.
The authors premise:
"According to standard economic reasoning, an increase in the financial incentives provided for an activity will improve performance."
They conclude:
"usual prediction of higher performance with higher compensation, when one is offered, has been confirmed: but the performance may be lower because of the introduction of the compensation."
Many of these kinds of studies have been performed worldwide with very interesting results that pretty much demolish the idea of homo economicus. Bowles, Boyd, Henrich, Gintis, Camerer, Gachter, Fehr, etc. mostly in Foundations of Human Sociality but also key papers like:
In Search of Homo Economicus: Behavioral Experminets in 15 Small-Scale Societies
Henrich, Boyd, Bowles, Camerer, Fehr, Gintis, McElreath
American Economic Review, 91, 2, May 2001, p 73-78
Camerer and Fehr are two of the reviewers mentioned in the footnote at the beginning of this paper.
One of the more interesting things (to me) about the emerging p2p economy (noted by some of my colleagues who study p2p economics) is how the introduction of financial compensation often lowers participation (both amount of effort and volume of participants).
-------------------------------------------------------- PHartzog@umich.edu
--------------------------------------------------------
The Universe is made up of stories, not atoms.
--Muriel Rukeyser
Both
Geoff and Dustin brought up some good points on the viability of non-monetary
rewards leading to some sort of monetary value or gain. And I thought of Northwest Airlines, who like
many other airlines has an Frequent Flier Miles Club, where given enough points
you are allowed to stay in their Executive quarters in any airport throughout
the world given that NWA has a presence within that region. I recently read of an individual whose
accumulated miles were going to expire unless he met a contractual obligation
to fly 10K miles by X date. Realizing
that the date was shortly upcoming he flew from NY-LAX twice in a row. This is literally 20+ hours on planes back to
back. Money certainly wasn’t the
incentive, it was the amenities that go along with flying. I wonder how NWA could parlay that enthusiasm
that exists for accruing miles onto an online community where money isn’t necessarily of vital interest but where money
factors into it (i.e. you accrue miles by spending money but you don’t
necessarily fly aimlessly just to accrue miles)
Monetary incentive in online communities?
This paper investigates the effect of monetary incentive on individuals’ behaviors using laboratory experiments. The most interesting finding is that when the monetary incentive is small enough, it will be detrimental to subjects’ performance, especially when the task is intrinsic incentive oriented (Task2) Furthermore, individuals always fail to notice the existence of this trap(The test followed by these two experiment).
The results of these two experiments break the rule that “”A small payment is better than nothing", which has implication on experimental economic methodology. This interesting phenomenon could be explained by many arguments including “people are afraid of looking ‘cheap’” and “insulting compensations”, and the most promising reason is that the introduction of money changes individuals’ reference point under the incomplete contract.
Under these results, shall we claim that monetary incentive generally will not succeed in online communities as it seems that we could only implement small amount of monetary incentive in online communities? How could we give every online user $10 given the fact of more than thousands of online users?
This was an incredily
This was an incredily interesting and counterintuitive study that tested the general beliefs that economic models predict that the more compesnation offered, the more perforamance will increase, while psychological studies predict the opposite. As Tracy already mentioned, these researchers only confirmed this when compensation was offered (as in, between a choice of compensation or no compesnation) but that small sums--especially in those involving intrinsic motivation like charity projects--can often be detrimental to performance and/or motivation.
The implications of this research, as least for me, tell us to value possible pscyhological effects over obvious economic models. Especially after the dot com bubble burst, and the recent web 2.0 craze, this makes sense. For example, on a community Wiki, if people are building knowledge and feel a part of their community then it makes sense to allow the intrinsic motivation of doing something for the good of the community to motivate them. A start-up may be tempted to offer a small sum for people to contribute to a new Wiki but this research shows that they may be better off not paying the contributors anything, or perhaps offering the possibility of earning a large sum of money that surpasses any ordinary amount.
Agree with Claim, but Clarifying Reasoning
I agree that in applying this to online communities it may be better not to offer any monetary incentive for doing a task that has high intrinsic motivation. But I want to clarify the framework that I believe is useful in making this claim.
The argument that I found most compelling in this paper is that these effects can be explained by imperfect information and incomplete contracts. People sign up for an experiment or to help raise money, and they have imperfect information about what is being asked of them. Further explanations upon arrival give them more information about their contract. If a new incentive is introduced at that point, then the participant's incentives change. They have completed the first part of the contract (and acted on whatever incentive was offered to begin with - intrinsic motivation to help raise money or baseline monetary incentive). Now they are entering a new contract with different incentives.
So I don't think it's necessarily wrong to offer monetary incentives to get participation in an online community. But I think it is important to think about what contractual obligation people may already feel under. If there is no contract, then a monetary incentive may help increase participation. If people are already invested in participating, but you still want to increase participation, then you may want to create a separate place in the community to give this monetary incentive.
For example, if you want people to answer more newcomer questions, don't just start paying 10 cents per question answered. This will disrupt the contract of oldtimers who answer questions for intrinsic motivation. Instead, create a separate forum where newcomers can ask questions for 1 cent and oldtimers can answer questions for 5 cents - that way the old contract is still in effect and a new contract involving monetary incentives can be put into place. [I'm not sure this set up would really work, but I hope it illustrates the idea...]
The drawbacks to "creating contracts"
I'm glad you mentioned this concept of contractual obligations, and the extent to which they shift users' sense of obligation and perception of their relationship with the site.
I think this raises a sustainability issue: It doesn't make sense to offer a monetary incentive that you won't be able to maintain over the long haul. Setting a precedent of monetary incentives, may shape users' expectations around what the overall purpose of the site is and how they're framing their involvement in that community, beyond just shaping the the extent to which they're motivated to contribute.
I was wondering if providing monetary incentives at all might challenge the notion of "community" by identifying your site more as a source of income than a place where you feel a sense of connection with others? Do others have thoughts on this?
Lisa McLaughlin
Challenging the notion of
Challenging the notion of community is sort of another way of saying
that monetary incentives will displace intrinsic motivation. But I
think the extent to which "community" vs. "source of income" is a
problem depends on the goals of your site. If you're mostly concerned
with generating content (and thus ad revenue), do you care if your
users feel connected to each other? On the flipside, paying for contributions in a community meant to foster connections with other members would be counterproductive, and in some cases, just plain creepy.
paying for contributions
I'm glad you brought up the idea of a start-up paying for contributions because I have a fun example to add to the mix. =)
Mahalo, a website I considered studying for this class, offers typically $10-$15 for anyone who writes a Search Engine Result Page (SERP). These SERPs are pages for popular search engine queries that are typically pretty great for a first stop on the web about a topic you don't know much about. I wrote a page for them on a band, and it includes all the great sites you *could* find through Google, but you don't have to sift through the really ad-infested or redundant pages to find what you want. They also have a great set of how to pages.
Anyhow, Mahalo seems to have been moderately successful so far, though there are still all sorts of questions about scalability. Will they continue to have enough funding after their seed funds run out? Google Ads may not suffice. And they have to pay a significant amount for their full time staff as well.
Anyway, I wonder if Mahalo's $10+ is significant enough for people to make them motivated to contribute a lot, or if it goes back to people simply being intrinsically motivated to contribute valuable SERPs. I really wonder if this will be able to last.
Thoughts
I like that you brought Mahalo up Liz, because it seems like it is the best example of an actual online community that offers monetary rewards for contribution. Google Knol, when and if it ever comes out, seems different in its premise but similar in execution to Mahalo.
I agree with you that Mahalo probably won't be able to last on this model. Though $10 per article is a sizeable amount, not the small amount shown in the article to reduce contibutions, I do think that for most people, doing edits on Mahalo for money would quickly start seeming like work instead of the intrinsic benefits of helping the community when it is unpaid. Perhaps this is why they have not met their goals of number of articles written - the money didn't end up being the draw that the community creator thought it would be. In my opinion, it seems like people need some sort of intrinsic motivation, even on a site with monetary rewards, because this is part of what forms their relationship and connection with the site. Money isn't always enough of a reason in and of itself.
Mahalo
I find Mahalo to be a very interesting concept. They do pay money for building the pages, but they also focus on the community aspect (They did not have it earlier, but added it few months ago). Maybe this concept might be OK for search (given that the company organizes vast amount of data on the internet in a very neat way), but might not be useful for other smaller communities.
prize/bonus
Instead of using several small amount of monetory rewards, an online community might introduce a fairly large prize/bonus to encourage contribution. I'd like to see how that could affect contribution.
Idea Management System
We had this Idea Management System at the Central Bank of Malaysia, where staff contribute ideas on anything that is useful to the organization. Staff gets points for contributing and rating ideas by others. There are threshold points, where staff can convert the points into inexpensive things like umbrellas or thumbdisks. Periodically, top ideas get to be presented to the management. Accepted ones are recognized in terms of end-of-year bonuses. Occasionally there are also “contests” that call for ideas to resolve certain problems or goals (e.g. save energy) that carries a cash prize ($500 for winner).
The observation is that more ideas are generated for the contests than the other one. Both have rewards, but one is large and uncertain, while the other is small and certain. But the small one is not too small to be insulting – the goods are useful – but there is also that very small chance of getting the big reward at end of year.
I agee!
Most marketing companies already do that. We see coupons, raffles, prizes, gifts etc. Same thing for the online communities!
Online communities are able
Online communities are able to provide something between non-monetary incentive and small monetary incentive. This is the point given on the communities. Point system is like something between encouraging altruism contribution and material incentive, in my opinion, it can be a matter to improve the adhesiveness in exchanging the effort with award. In other word, it makes it less transparent in comparing effort with award thus softening assaulting effects. In addition, the virtual sitting makes this exchanging system able to be easily adjusted and manipulated. For users, they can interpret the point award differently and for the administrators, they can manipulate the interpretation individually to encourage users' participation.
It would also be interesting to see whether similar happens in the online point system: like very few points have negative effect.
Since most online communities involves much volunteering contribution, thus adding point incentive is tricky. There might be something else that can affect monetary/point incentive. For example, on taskcn.com i observed, since people had very small chance to win a competition thus getting money (a task usually gets 65 participators but only one winner), there is no correlation between the real monetary award to the number of people's participation; while on Baidu.com, where people answer questions to earn virtual points, there is significant correlation between the number of answers and the number of points offered. A remarkable thing is there only 60% questions got at least one answer. This basically means when people had more choice, they follow more points, if not, and the intrinsic incentive is high, they still want to do it.
Pay enough or don't bother
John Blair
Many similarities between the Beck and Gneezy-Rustichini readings. I combine the postings here due to this similarity. I preferred the Gneezy article more, as it seemed that the explanations of the analysis were clearer and succinct. I also agree with Bem’s (Gneezy-Rustichini page 803) suggestion that people interpret their actions as any outsider does, by looking at the reasons and motives for what they do. I think this appealed to me because it focused more on the individual making a decision as opposed to group collective. One of my problems with the behavioral economics (meaning the combination of social psychology and economics) is that is seems to forego the individual being individuals. Obviously there is overwhelming evidence supporting group behavior (even when group is randomly assembled with individuals having no previous knowledge of each other). I just have a problem not having individual choice being represented.
I do wonder why the reward results in both articles identified that performance was better with no reward or a high reward. There seems to be no middle ground and the rationale for this I think may be due to individual perception of the reward – obviously they didn’t think it was enough to motivate them to be incented to perform well. But why would they not perform at least as well as the no reward group? Were they insulted (as suggested in Gneezy-Rustichini page 805)?
From the Beck reading, it’s hard to argue with Thorndike’s Law of Effect (effect of satisfying behavior is rewarded/strengthened, annoying behavior is punished – page 179) or Skinner’s reinforce concept (page 182).
Application to ecommunities is relatively straight forward. If you choose to offer an incentive for something make it worth their while – or don’t offer them anything, but this would require some sort of intrinsic motivator for them in order to be successful.
Non-monetary Incentives
I think this reading was valuable, despite the fact that few e-communities offer monetary rewards. I would be curious to know if other types of incentives (non-monetary) also have the same displacing effect on intrinsic rewards. Would in-system rewards (privileges, "points", recognition) cause the same resentment and/or contractual reevaluation that resulted from the addition of monetary rewards? What if these non-monetary rewards still carry monetary value? I suspect even in the latter case that these rewards could be viewed in a different light than cold, hard cash.
good point, we should consider if non-monetary becomes monetary
Great questions Dustin. I am also wondering about the same stuff so I am going to take a small stab at offering my two cents. As several people pointed out above, it is definitely a little hard to compare the reading to online communities because the structure just seems a little distant. On one hand, we have the experimenters gathering a group of people, while for most of our online communities, people join on their own.
However, having said that users join online communities because they are already motivate to do so (for various of reasons depending on the community and the person's needs), I'm seem to be doomed and destined to say either:
1. introduction of monetary rewards will negatively impact activity in online communities, because people who were once intrinstically motivated, would find the monetary reward as an extrinsic motivation. Therefore, people who truely cared about the community may be turned off, while other users may become active simply because of the newly introduced monetary rewards. As a result, the community may be harmed as a whole.
2. introduction of monetary rewards will positively impact activity in online communities, because people who were once intrinsically motivated, will have even more incentives to do contribute. Similarly, users who were dormant or not as inactive, would be motivated to participate more. As a result, the community would seem to thrive.
It turns out that both perspectives are really quite flawed. Obviously, the authors dinstinguish between monetary "size" so that itself might change how a community reacts to monetary incentives. Thats why I think you raised an excellent idea, about the case where non-monetary rewards still might lead to some sort of monetary value or gain. I think along with the "size" of the monetary reward, whether or not the non-monetary rewards offered by an online community under analysis, would at some point translate into monetary value, is another good way to measure how a community might react. For example, on a community that has e-commerce functionality, highly active people may be awarded with non-monetary reards such as getting a higher "reputation" or rank higher on a leaderboard, will probably also get more monetary gains.
Agree! The monetary
Agree! The monetary incentives seem to have an immediate effect on motivation, Nevertheless, the quality of the knowledge shared can be inferior, and the attitude that knowledge is a private and non collective good is enforced. Once knowledge is shared only because monetary rewards are obtained, knowledge sharing will decrease when these rewards are withdrawn. Additionally to these consequences, high costs led us to focus on non monetary rewards that may not have an immediate, but a long-term impact on motivation. Therefore the use of monetary incentives should only be used with cautious.
Overall, monetary rewards may have an immediate effect on motivation to share knowledge but at the same bear the risk of spoiling users. However, monetary incentives can be used to start a knowledge management system and to incentivize users from time to time. Yet, in the long term users should be incentivized non monetarily for sharing their knowledge.
Some Thoughts
One of the design claims I leant from this reading is that: “Positive but small enough compensations, there is a reduction in performance as compared with the zero compensation.”
I just wonder how small a small amount is in an online community setting, since “small” is a concept that you need to compare it with others. Also, I am confused about the sequential vs. single-stage set up. Authors explained there is a shift in perception of the contract, after the introduction of additional compensation for correct answers. (p 804) The perception about the game starts from “Sixty NIS was paid for showing up,” and then changes to “The activity of answering the questions is now paid by rate”. And the rate (10 cents) is low if it’s compared with 60 NIS. Since the conception has been changed, can this example be seen as a sequential game? If an amount of money is offered by participating in an online community once the site has been created, and there is no other values for participants to compare with, will participants be motivated to contribute more by offering monitory incentives?
Related question regarding initial motivations
I was curious about the initial payment and how it might have framed the rest of the experiment. 60 NIS at the exchange rate given in the footnotes is about $17 which is the average hourly wage for economic experiments. That was just for showing up.
How might the results been different if there was a much smaller show-up payment? That is, how much did the 60 NIS show-up fee frame the marginal payment of 10 cents for each correct answer in a negative light? I would be interested to know what the results would look like with a much smaller show-up fee, where the incentive is almost completely on performance of the task.
Lottery Vs. Micropayments
I think a good way to combat paying too little would be to have some kind of lottery system where a user might be paid a large sum with a level of uncertainty. So instead of getting 10 cents each time you do something, you might have a chance to win $100 or something. Amortized, it could end up that you really got 5 cents for each "action," but the perceived payoff is what counts.
This is sort of what Blingo does. It's basically a frontend for Google, but each time you do a search, you have a chance of winning a cool prize.
I think Amazon.com's mechanical turk suffers the too small problem. I tried doing some tasks but felt the super small sum of doing them was not worth my time. If the tasks were more fun to do, I might have just done them for that.
Motivating Intrinsic Motivations
Although the authors say that the explanation based on attribution
theory is does not really explain the behavior here I think that it plays a
vital part. If a subject observes
himself doing a task for which no exogenous incentive is provided then he
perceives his motives as being driven by intrinsic motivation. When he is
offered a monetary incentive, irrespective of whether the activity is
sequential or not, it directly associates his effort with a price. Now when he
evaluates his opportunities he realizes that that he will get the original
amount of payment regardless of the questions he answers. Regardless of what he
thought the original 60 INS were for (showing up or answering all questions) he
will evaluate the effort he needs to put in and decide whether the extra effort
is worth the extra money he will be paid.
An interesting experiment would be to see whether how extra
information can change the behavior. For example if an agent was ‘told’ that
his principal was one of the very few principals who actually chose to pay him 10
cents compared to others, then the motivation to do well might again be
intrinsic irrespective of whether the amount paid was 10 cents or 40 cents –
maybe the degree of effort would change but the result for 10 cents might still
be an improvement – because now the agent values his outcome as a intrinsic duty,
and not only in terms of the 10 cents.
So can information about the motivations actually make a
difference to the results despite the money offered ?
small non-monetary compensations vs small monetary compensations
I think that while small monetary compensations are detrimental as a motivating factor, small non-monetary rewards can still have some motivating effect. Antecdotally, I have seen this behavior in particular at trade shows and conferences, where booth operators can make attendees jump through hoops and give out all sorts of information for the "prize" of a t-shirt or promotional USB thumb drive. In this case, I doubt that conference attendees would sit thru a half hour presentation for the cash value of these products.
While classic economic theory would prodict that we would naturally abstract the cash equivalent value for rewards we receive, I think we have a hard time decoupling objects and their value when they are right in front of us. Consider the difference, say, between playing pool in a bar for a $6 wager, versus playing pool for who buys the next round of drinks. The $6 is an abstract figure, and your utility of it is probably subject to a discount factor which the beer is probably not. You may consider $6 "beer tomorrow" whereas the alternative is "beer today".
I would be interested to see the effect on these experiments of offering low-value non-monetary compensation ie offering candy to kids in proportion to how much money they raise and see if that has any impact on effort outcomes. My guess is that the more immediate the enjoyment of the compensation is is an additional term to the magnitude of compensation.
--------------------------------------------------------------
oostendo@umich.edu
--------------------------------------------------------------
Behavioral Economics and Online communities
Behavioral Economics studies are always interesting. Sometimes these studies quantity human values like: compassion, love, racial discrimination, etc.
The study involving day care center was mentioned in Freakanomics. It shows how money can change the behavior of a parent picking their child from day care.
I don't recall any online community being studied in our class provides any explicit monetary incentive. But I think a new wave of services that rely on monetary services are coming. Example - the Google Knoll a.k.a. the Wikipedia killer. According to this post by a Google VP, Knol will allow users to maintain articles on topics in which they are experts. Unlike Wikipedia, the articles will promote the authors name (like the cover of a book) and also provide a share of the ad revenue from that page. When this service is introduced it will be interesting to compare the success of Knol and Wikipedia. It will be another behavioral economics study, to show the effect of money on contributing to public goods. From this paper, my prediction is that the monetary reward for most users will be too small thereby making the motivation extrinsic. So many users will find the monetary reward insufficient and would switch back to Wikipedia.
The design claim, I infer from this paper is that – when offering monetary reward, it should be sufficient to provide extrinsic motivation greater than the intrinsic motivation provided without any rewards.
Looking Cheap
As I tried to break down this article into takeaway points, I started to understand why changing or proposing rewards usually has a negative effect on intrinsic motivation. (for most of the article, I had a problem with the author's argument that introducing a reward will have a negative effect on internal motivation, rather than just an impact that could be positive or negative. Why must internal and external rewards conflict?) If people interpret their actions from an outsider's perspective (p13), the internal reward won't be considered if an external reward exists. People would base their interpretation for another person's motive on an external reward - "that's why he's doing it" - and wouldn't need to consider alternatives. From an outsider's perspective, the external and internal rewards match.
Takeaway points:
1) People will be less likely to expend effort toward achieving a goal if an external reward is offered and the compensation is small. They might be a) insulted (their effort isn't valued) or b) ashamed (they don't want to look cheap)
2) Appealing to a person's internal motives is difficult. If a person values an internal reward more than another person makes it out to be valued, or if another person offers someone possible internal rewards as incentives and leaves out the particular internal reward that would motivate them, then they will be less likely to expend effort toward achieving a goal.
3) Small compensation changes the way people interpret rewards (internal --> external), so the reward becomes small.
which incentives cause perceptual shifts?
One challenge for designers looking to give rewards for tasks that are intrinsically rewarding will be to figure out what additional rewards they can build in that won't trigger this perceptual shift in whether or not a task is intrinsically or extrinsically rewarding.
It also seems that the answer will vary basd on the task and user. Awarding points may work well to support intrinsic rewards in many systems; providing users who are intrinsically motivated to contribute to a community with feedback that they are indeed contributing. There may be other situations where they cause problems. To illustrate, I'll use an example from one of the CHI teams that presented at FIRST this past Friday (since it's fresh on my mind and many people in class where there). The team proposed a system (pdf) that allowed experts to answer questions from case managers serving the homeless. The team had designed in a reputation system (based on point assignment) that was intended (I think) to help case managers identify and reward trusted sources. A number of people in the panel felt that the idea of rewards or treating the system as a game with points would diminish the volunteer experts' satisfaction with their volunteerism. Unfortunately, the theory and experimental results that could guide the team as they weigh the value of having a point system haven't been developed yet.
Figuring out when and which rewards can work to reinforce intrinsic motivations vs. triggering perceptual shifts to seeing a task as extrinsically motivating seems like an interesting design challenge for further experimentation and theory development.
My 6 cents
This article was very relevant to Newsvine’s community model. Newsvine tries to motivate users by giving them 90% of the ad revenue that their articles generate. After about 6 months of use this had netted me a whopping $0.06. This amount of money is more insulting than motivating and I often forget that there are earnings to be had at all. The main motivator for me is submitting content that will get interesting discussions started.
At this point though Newsvine is obligated to continue with their financial incentives. If they were to quit their earnings programs the top users--who are probably making some money from it--would be alienated and might defect to another website community. Fortunately this aspect of Newsvine is not a major portion of the site or else it may face more significant criticism from users.
I have always imagined that Paradise will be a kind of library.
-Jorge Luis Borges
greenwashing
This is not a very deep insight, but in this case I would argue that paying the money is not so much a reward of contribution, but a greenwashing message demonstrating that Newsvine is not exploiting it's users' contributions.
For some top users this might be different, but for the most part I would argue that what I mentioned is the case.
Human Sociality
The authors premise:
"According to standard economic reasoning, an increase in the financial incentives provided for an activity will improve performance."
They conclude:
"usual prediction of higher performance with higher compensation, when one is offered, has been confirmed: but the performance may be lower because of the introduction of the compensation."
Many of these kinds of studies have been performed worldwide with very interesting results that pretty much demolish the idea of homo economicus. Bowles, Boyd, Henrich, Gintis, Camerer, Gachter, Fehr, etc. mostly in Foundations of Human Sociality but also key papers like:
In Search of Homo Economicus: Behavioral Experminets in 15 Small-Scale Societies
Henrich, Boyd, Bowles, Camerer, Fehr, Gintis, McElreath
American Economic Review, 91, 2, May 2001, p 73-78
Camerer and Fehr are two of the reviewers mentioned in the footnote at the beginning of this paper.
One of the more interesting things (to me) about the emerging p2p economy (noted by some of my colleagues who study p2p economics) is how the introduction of financial compensation often lowers participation (both amount of effort and volume of participants).
--------------------------------------------------------
PHartzog@umich.edu
--------------------------------------------------------
The Universe is made up of stories, not atoms.
--Muriel Rukeyser
2 cents are worth 10K miles
Both
Geoff and Dustin brought up some good points on the viability of non-monetary
rewards leading to some sort of monetary value or gain. And I thought of Northwest Airlines, who like
many other airlines has an Frequent Flier Miles Club, where given enough points
you are allowed to stay in their Executive quarters in any airport throughout
the world given that NWA has a presence within that region. I recently read of an individual whose
accumulated miles were going to expire unless he met a contractual obligation
to fly 10K miles by X date. Realizing
that the date was shortly upcoming he flew from NY-LAX twice in a row. This is literally 20+ hours on planes back to
back. Money certainly wasn’t the
incentive, it was the amenities that go along with flying. I wonder how NWA could parlay that enthusiasm
that exists for accruing miles onto an online community where money isn’t necessarily of vital interest but where money
factors into it (i.e. you accrue miles by spending money but you don’t
necessarily fly aimlessly just to accrue miles)