Month: November 2015

With all these talk in recent years about crowd sourcing and big data, I wanted to delve a little bit deeper on the costs of crowd wisdom. It is easy to see how crowd wisdom can be beneficial to society as a whole; for example, there are many crowd volunteering websites that aim to aggregate crowd decisions for social good. One of them is Tomnod which rose to fame when they used satellite imagery to crowd source data on where the missing MH370 plane is. However, in subtle ways, crowd wisdom might not be all that good.

In one of the more telling problems of crowd wisdom (herd mentality in particular), Sunil Tripathi was implicated as a suspect in the Boston marathon bombings because a Reddit user posted in the forums that they looked alike. Soon after, more people started agreeing with the first poster, and spread like wildfire over Twitter. It went into mainstream notoriety when mainstream media reported it over TV and newspaper. The lack of fact-checking by the latter was disturbing. However, this shows one thing – that crowd wisdom can turn mad. Critics of this point might point out that crowd wisdom works best if people are independent – they have no knowledge of other’s ideas. While this can be replicated in a laboratory setting, the fact is that it is difficult to ensure this online. People invariably will use search engines or share it on Facebook, which leads to a decoupling between reality and ‘reelity’. It is what makes crowd wisdom a dangerous idea to wholeheartedly believe in.

Secondly, the issue of privacy comes in. In recent years, due to the proliferation of smartphones with embedded sensors, app developers have been using the sensors to predict weather or traffic. Dark Sky, a weather app on the Apple App Store, uses pressure sensors on the latest iPhones to crowdsource data on whether it is going to rain in a particular region. Waze, a traffic directions app, uses your phone’s speed and location to crowdsource data on traffic conditions. If you are moving slower along a particular road, the app may intelligently tell other road users that a traffic jam might be imminent, and to reroute road users accordingly. It seems beneficial until you realized the same app can be used for nefarious purposes. Previously, Tinder, a dating app, uses one’s locations to tell potential dates where you are. Whilst other people see you as “3 miles away” in a vague sense of distance, some tech-savvy people figured out a way to perform deep-packet inspection of the traffic to find out your exact coordinates. It turns out that the actual coordinates are actually being sent to your phone, which then processes the data in conjunction with your location to provide you with the distance. Hence, it is possible to know where you potential ‘dates’ live, and possibly stalk them. This is a design flaw more than anything else though, but it still shows the flaw of crowd wisdom – when companies attempt to collect everything that it is possible to gather from a particular user, the possibility for irresponsible disclosure is heightened.

Thirdly, it might not be even accurate at all. As seen in class, the more variables there are in a particular prediction, the less accurate it becomes. For example, predicting how many leagues in the depth of the Marianas Trench requires people to have some notion of a league, in addition to predicting the depth of the trench. A more popular example which deals in big data was Google Flu Trends, which was recently shut down. Although it wowed researchers by predicting flu trends ahead of a full-scale epidemic, it constantly overstated the severity of it. People were too prepared, it seems. Whilst this might not be a bad thing, I can see the flaws of trying to extract information from crowd data – it might be fairly accurate now, but might not be a few years down the road.

Finally, I would like to end off with some food for thought. How do we ensure that crowd wisdom is indeed accurate? We might take a look at the intricate system Wikipedia has in place for its editing – talk pages, a system of editorial processes and its ilk. Is it possible to generalize for online communities and spaces?

Abstract

The idea of using social norms and behavioral science in environmental conservation is not new; however, the uniqueness of OPOWER’s approach is that they combined actionable strategies with subtle cues in their monthly energy reports. This being said, the paper surmises that there might be significant challenges in the future in trying to make OPOWER’s strategies more prevalent. In the short discussion piece that follows, this author attempts to elucidate these challenges, positing a nuanced basis for said issues and suggesting possible solutions.

Introduction

OPOWER is a publicly listed energy-efficiency software company that aims to reduce energy usage through social cues derived from behavioral science. Their results work on the premise that normative influence runs deep in people’s predilection to energy conservation.  Such normative beliefs mean that people tend to be more greatly affected by the conservation efforts of similar people around them. This was a milestone in energy conservation because energy reduction is the only method which provides a net positive value in benefits accrued to society, and OPOWER made it a cinch to integrate into existing utility billing systems. Contrast this to solutions like replacing incandescent bulbs with fluorescent lights, or installing photovoltaic panels – both require a significant sunk cost, the value of which can only be recouped over a long run. OPOWER’s solution, on the other hand, saw an almost immediate reduction in energy usage without any significant costs.

However, there are significant roadblocks to OPOWER’s widespread implementation. Firstly, there is the issue of boomerang effects (8) which might cause people who use less than average electricity to regress to the mean. Next, there is skepticism among utility companies that their methods will work on their customer base due to demographic differences. Finally, there have been questions on the sustainability and extensibility of OPOWER to other countries – would it work overseas without much modifications?

Mean Reversion

The boomerang effect describes a social phenomena where attempting to persuade people to do things in a certain way (such as conserving energy) unexpectedly results in them doing just the opposite. In this case, OPOWER knew its strategies would cause higher energy-usage households to consumer less energy, but it would also cause lower energy-usage households to use more energy. To counteract this, OPOWER used injunctive messaging such as smiley faces to signal approval and disapproval for their current energy usage. Subconsciously, the smiley faces serve as a reinforcement and approval indicator for people to continue with their above average performance. However, as seen in Exhibit 4, the boomerang effect still played its hand, albeit to a smaller extent.

One possible solution is to introduce subtle language cues into the energy report, as well as a revision of the presentation of the statistics. In figure B, we see that the Social Comparison Module (SCM) provides absolute statistics on energy usage comparison. (ie. I used 1033 kWh vs 784 kWh for my efficient neighbours) Instead of making absolute figures the central to the SCM, relative figures could be provided instead. For instance, one can place the user on a bar chart on a percentile basis. If I used 32% more energy than my efficient neighbors, then I would be placed in the 90th percentile, for example. Instead of saying “You used 32% more than your efficient neighbors”, the language could be revised to “90% of your efficient neighbours use less energy than you. What this serves is to provide reinforcement that a substantial number of people are ahead of the user when it comes to energy usage. Even if one is below average in energy usage, there will still be a percentage of households that are still using less than you. In this case, the optimal energy usage to strive for is the one that used the least. Hence, the boomerang effect can be mitigated to a certain extent.

Slow Adoption by Utilities

As mentioned in the case study, there has been less than favorable responses from utilities about the adoption of OPOWER strategies. One point that is noteworthy that most utility companies are for-profit. If they promote energy reduction, less profits accrue to them due to transmission and distribution revenue. To offset this, most companies depend on tax breaks or grants from governments in achieving green objectives. In Pittsburgh, Duquesne Light charges all customers a small sum of money in each monthly bill to recover the costs of implementing a program to educating customers on energy usage, in their WattChoices program. Without some kind of financial incentive by government stakeholders, there is a conflict of interest in power generation companies – both to their shareholders as well as energy conservation groups. Moreover, those companies believe that the demographics of each state mean that people would not be receptive to the idea at all.

It is difficult to tackle this problem due to said conflict of interest. Hence, an effective solution would call for some form of government intervention, one that does not shackle electricity generation companies, but rather rewards people for reduced energy usage. To take a leaf out of health care, some insurance companies pay people to keep fit – they provide complimentary gym passes  and even reward you based off the activity that you do [1]. The premise is that insurance companies would have a positive expectancy due to increased health that their customers have after keeping fit, even after the costs incurred in such perks. This can hopefully be applied to energy conservation, except that there is no clear stakeholder involved here – the environment benefits, but a clean environment is a public good. From economic theory, such a public good will not be funded at all. Hence a solution would have to involve the government. One way to go about this is to continue the healthcare approach – the government can offer tax breaks or refunds for reduction in energy usage, either in absolute or relative terms. Such a measure would be funded by discounting the fact that due to the reduces energy usage, less money would be needed to maintain existing power generators, or build new power plants.

To a Sustainable Future

With all these strategies being bandied about, one thing is almost certain if OPOWER sticks to its current method. It might be useful to consider the fact that people will tend to be numbed to such messages after some time. Energy conservation is not a Boolean entity. Consider the fact that some people might use more energy than other households, simply because there are more people living in it. How would we prevent such users from being discouraged and not care about energy conservation? To take a leaf out of obesity studies, most dieters successfully lose weight. However, when their weight loss flattens out, they inevitably get discouraged and start to eat more. To bring in the energy usage perspective, the most avid energy reduction users would, in the near future, start using more energy. What is needed then is a way to engage them over a long run.

One possible idea is the use of dynamic messaging. Instead of saying the same “You use xx% less energy than other households”, it might be possible to present the information in a different layout every time, like in infographics. Appealing to people’s emotions is also a good way. For example, “From the electricity you reduced this month compared to last month, you have provided enough electricity to power a school in less developed countries for 24 days.” Instead of bandering about with terms like kWh which most users do not understand, this strategy actively engages people with the real-world consequences of their actions.

Conclusion

In this short write-up, I considered the future of OPOWER as the authors of the Harvard Business School case study raised in the prologue and epilogue of their case study. I argued the need for the government to intervene because of a conflict of interest in energy conservation, as well as suggested possible amendments in data representation to avoid the dreaded boomerang effect. Finally, I also explored the possibility of changing how data is represented – not only to keep people engaged, but to allow them to see the real-world effects of their energy reduction (or increase).

Auspicium Melioris Aevi

Future Discussion

• Do you think OPOWER should consider telling everyone that they are above average energy usage, even though they are not, thus doing away with the boomerang effect? The advantages of such a strategy is obvious, but is the key issue here ethics?
• For those statistically inclined, would a percentile based approach really do away with the boomerang effect.

References

[1] http://www.wired.com/2014/12/oscar-misfit/

[2] http://www.nytimes.com/2011/09/20/health/20brody.html

My interest was piqued on the concept of nudge theory when I read about a study which successfully got high school students to make healthier food choices in their school canteen, simply by making the choice to choose fruits and vegetables more convenient, attractive and normative [1]. It got me thinking about the potential of nudging in public governance and regulation, a field that I am particularly interested in. It is plausible to extend the application of nudge theory from a hyperlocal space (like a school tuck shop) to a neighbourhood or even the entire country. Indeed, this concept was broached albeit briefly in class when we discussed how we can use nudge effects to persuade people to increase voter participation in the recent judicial elections. Amongst the points brought up were getting people to like the candidates’ Facebook pages, or telling people the importance of the elections. In this post, I explore the use of nudge theory as a substitute for public regulation in the health domain. I will describe each strategy in turn and end off with a nuanced evaluation of their effectiveness.

On Government Regulation

There is no lack of literature in the use of nudge effects in government regulation. In fact, Cass Sunstein who posited the idea of nudges used to head the Office of Information and Regulatory Affairs. However, to the casual observer, it seems that most government interventions work via explicit law and regulation, rather than subtle nudges. There are a few reasons why this is so. Taking a leaf out of economic theory, it is known that government regulation has both demand-side and supply-side portions. An example of the former is an outright ban on smoking in public spaces, hence making consumers smoke less. In contrary, a supply-side decision would entail imposing a hard limit on the number of cigarettes all producers can produce, hence reducing the amount in the market. Both serve the same purpose – to reduce public smoking. To this end, they work very well. Every 10% increase in tobacco prices result in a 4% fall in demand from consumers (2). Further, in America, such policies have been so effective that tobacco companies have looked to other developing markets to increase their revenue since growth locally have been stagnating. As can be seen, government regulation seems to be an effective measure. This arises from the irrationality of most people. Given perfect information about the health risks of smoking, people still tend not to reduce consumption. Advertising is one possible cause of this irrationality – smoking is portrayed to be hip or cool. In this respect, the heavy hand of the law such as adopting public smoking bans is required. This moves people away from accepting public smoking as a social norm. Hence, government regulation is seen as effective by many.

However, social psychologists have pointed out that government regulation may not be effective. For example, the problem of boomerang effects. Boomerang effects occur when a measure is used to bring about a specific intent, but the opposite of that actually happens. In essence, government regulation can backfire and worsen the situation. Continuing the case on smoking, in Asia, there have been massive spikes in the consumption of illegal cigarettes due to tax duties imposed by the government (3). Illegal cigarettes do not have the same quality control as legitimate ones and could contain more poisonous chemicals in their manufacture, hence it brings more public harm. Whilst government regulation is intended to reduce the consumption of cigarettes, the nett effect was to steer consumers to illicit sources where the government can control less effectively. Secondly, such regulation may not be cost-effective. It takes effort and money to properly enforce policies. Policy without enforcement is ineffective. With these disadvantages, it is not surprising that scholars have been considering nudge theory in getting people to make personally beneficial decisions.

On Nudges

Nudges were first mooted by Sunstein. He defined it as ‘any aspect of the choice architecture that alters people’s behaviour in a predictable way without forbidding any options or significantly changing their economic incentives’ (4). By choice architecture, he meant the environment in which allows people to more easily make decisions that maximise their well being. In his research paper, Marteau notes that nudges are comparatively lower in cost, and appeals to governments who do not subscribe to the in loco parentis doctrine. For example, we can make smoking less appealing by reducing smoking cues – these can include removing public cigarette stubs and ashtrays in dustbins, or getting people to sign a form acknowledging health risks each time they want to buy a pack of tobacco. By doing so, these measures attempt to nudge people into accepting non public smoking as a social norm. This can lead to cascade effects as well – if there is no second hand smoke in the surroundings, smokers will be less inclined to start smoking as well, leading to positive externalities. Moreover, by constantly reminding them of the health risks, we might possibly get them to be constantly aware of its dangers. Of course, there is the possibility that people are numbed to such a measure. This can be seen by a law in some countries forcing cigarette manufacturers to put images of people having diseases arising from smoking. Most smokers ignore the photos on the pack.

However, there are detractors who, while conceding that nudging can be effective, they can only be truly effective when combined in concert with some form of legislation. For example, how do we get shopkeepers to make buyers to sign a health risk form? It is indubitable that it needs some kind of local law to enforce it. Secondly, nudging can also be done by the private sector to further their interests as well, if we further the definition of a nudge originally proposed by Sunstein. Their strategies can be inimical to the effectiveness of government nudging; this possibly calls for some kind of intervention to render their strategies moot.

One thing is certain is that nudging has sparked interest in the wider field of behavioral theory that can ameliorate public health problems. Obesity alone costs $200 billion in health costs a year. The UK Cabinet has a Behavioral Insights Team dedicated to such issues and published a very interesting paper on the various issues. While no concrete government policy arose out of this, the paper shows how nudging is designed to still improve public health while respecting consumer sovereignty.

 

In conclusion, I believe the public policy challenge is clear. While nudging can be very effective, two forms of government intervention is still required – legislation to implement the nudging strategy nationwide, as well as regulation to prevent other parties from counteracting said strategy. It is also clear that government regulation is effective in itself; however, because it disrupts free market dynamics, governments have to take note on where to draw the line.

Possible Discussion Questions

• In government regulation, there are limits as to what the government can or cannot regulate, such as free speech. Do you think the same limits apply to nudging?
• I mainly discussed about nudging on the consumer-side in this post. Is it possible to extend nudge theory to the producer sector as well or can only government regulation help? For example, how would we nudge food companies to not use hormones and fake additives in their food, other than outright banning them?
• What are your viewpoints on this issue? Do you think nudging or regulation is more effective?

(1) Smarter Lunchrooms Can Address New School Lunchroom Guidelines and Childhood Obesity (http://www.jpeds.com/article/S0022-3476(12)01478-3/abstract?cc=y=)

(2) http://www.paristechreview.com/2011/10/26/nudging-smokers-new-frontiers-fight-against-tobacco/

(3) http://www.straitstimes.com/singapore/singapore-is-no-4-for-illicit-tobacco-use-1-in-4-cigarettes-here-illegal

(4) Nudge: Improving Decisions About Health, Wealth, and Happiness Thaler and Sunstein

(5) Judging nudging: can nudging improve population health?  http://www.bmj.com/content/342/bmj.d228

(6) The Healthcare Costs of Obesity http://stateofobesity.org/healthcare-costs-obesity/

Penned with ❤️ for 08-624 and 17-704

Just some of the opinions I have regarding the readings this week

Before the year 2000, organizations like Apple, Pepsi and Coke used television and newspaper as the primary medium for advertising their products. They relied heavily on the 30 second commercial for creating the brand awareness. Ever since the advent of Internet, these companies have migrated to different channels for advertising. They have started employing nudges to attract more customers online. In this blog, I would be focusing on how nudges created and used by companies like Amazon, Google, and Facebook to generate revenue are invading the privacy of the users.

In 2004, Google launched its email service called ‘Gmail’. Gmail became one of the most popular and widely used email services across the world. Though it is a free service, Gmail became one of the most important apps for Google as it generated a lot of revenue through ads. This email service gathers data regarding the sites that you have visited. They also display ads that are extremely relevant to your current needs. For instance, if you just now received a mail from your friend regarding vacation in California, they display hotel’s ad as a text. This text will be placed right on the top of your inbox so that it is visible to you when you are checking your emails. By placing a relevant ad on the top, you are most likely to click on the link [1]. How do they display relevant ads in Gmail? Google reads all the emails present in your inbox and then displays an ad based on the email’s content. When you click on the ad, Google will receive the money from the advertiser.

Apart from Google, e-commerce websites like Amazon track the websites and products that you have searched so far. When a person logs into Amazon, it will recommend products based on the data they have collected relevant to search preferences. Apart from the shopping history, these companies also recommend products that are specific to your location [2]. Social networking websites like Facebook employ a different type of tactic. Apart from gathering location details, they also keep track of the pages that your friends like. If your friends like a particular product on Facebook say a restaurant in your location, Facebook will display ‘ABC and 5 others like this’ on news feed where most of your attention would be. By displaying that, you are most likely to check out what your friends have bought (social proof). If they have liked a restaurant, you are most likely to try it out.

After collecting the data, these organizations are not going to delete them. For instance, Google or Amazon will not delete your history from their servers. They are going to build on top of that and learn more about you using latest technologies. Is it safe? I think the data will be stored safely. However, there have been many instances of security breach. For instance, TJX, a shopping store, had all of the data about their users stored insecurely in their servers. Anonymous hackers hacked their store’s account and made them public [3]. Is it ethical to collect information about a user? I personally feel that it’s not ethical to collect data about you without your permission and use them for creating nudges. As discussed in the ‘Privacy Nudging’ lecture, a person can use the App Permission manager in Android to prevent apps from collecting information about you. The users can also make use of ad blockers or incognito mode for browsing using desktop. So what do you guys think? Do you think this is an ethical way of creating nudges? Please feel free to comment on this topic.

========

You mentioned that “nudges created and used by companies like Amazon, Google, and Facebook to generate revenue are invading the privacy of the users”. Can you elaborate on the nudges used by said companies? In the case of Gmail, you mentioned Google “places a relevant on the top”, but I think that would be tough to qualify as nudging since it lacks a user objective.

What I think would be more evident is Facebook. Every so often, Facebook will notify you on the site to do a Privacy Checkup[1] which prevents users from sharing information that they don’t intend to. Before, the privacy settings were deeply embedded in the bowels of Facebook that people do not make sure they are correctly set the way they want it to be. With this tool, more Facebook users are nudged into adjusting their privacy settings, which is a step in the right direction privacy-wise after their privacy brouhaha a year ago.

You mentioned in your views that “it’s not ethical to collect data about you without your permission and use them for creating nudges”. You gave the example of Android apps collecting unnecessary information. At this point, I am pretty confused about what your a nudge entails. Does showing an advertisement to a user constitutes a nudge? Personally I don’t think so. Sunstein defines it as anything that influences our choices that leads to *improvements* in the rationality of decisions people make. Such nudges are informed by the findings of behavioural economics.

[1] http://arstechnica.com/business/2014/09/facebook-to-prompt-users-with-a-privacy-checkup/
===================================================

“There is an app for everything”, so much so that Apple has trademarked the quote for their advertising campaigns. If one looks at majority of the apps developed, especially the ones using the open data from the US government, one can see that they are variations of a standard set of applications that help with credit processing, medical aid, tax filing, bill payments and so forth. This is understandable as these applications are used a lot by the public and there is the incentive for a third party to monetize through ads. Seeing the success story of one such company motivates other players to enter the domain. But how does one persuade someone to use your data for the public good?

The flaw in this model is that it takes a very long time for some player to enter into the niche markets (by niche, I mean the agencies that are not targeted by most of the third parties – Center for Dietary Studies for example), and more often than not, it does not address what the niche community might want. To encourage diversity of applications for government agencies, the US government spawned the idea of Challenge.gov. Challenge.gov is a technical platform and list of challenge and prize competitions, all of which are run by more than 80 agencies across federal government. These include technical, scientific, ideation, and creative competitions where the U.S. government seeks innovative solutions from the public, bringing the best ideas and talent together to solve mission-centric problems. This platform is available at no cost to all federal agencies to help them list their challenge and prize competitions and learn how to engage the public through this innovative approach. To read more on this – https://www.challenge.gov/about/

This approach succeeded in persuading and motivating people to create solutions that leveraged the organization’s data and create solutions specific to the target audience. Crowdsourcing the solution helped in alleviating costs for the organization and incentives persuaded the public to respond. Since this is backed by the US government, it prompted a lot of response to every government agency posting a challenge on the platform – https://www.challenge.gov/success-stories/

We can see some of the techniques we learnt in class at play. We can see Authority at play here. Since this backed by the US Government and not just any other organization, people tend to take notice and participate in the challenges more than what a standalone organization could have achieved. Also we see the role of Reciprocity at play. Incentives for creating the solution outweigh the efforts needed to put in and thus people are motivated to contribute.

I believe that Challenge.gov is a great example of influencing and persuading third party to contribute more to projects without looking for monetization through the market. A non – profit organization for example can tie up with a well- known commercial company and organize hackathons to get the solution that they want – check out – http://www.hackingeatingtracking.org/, where this organization partnered with Harvard Med School and other big players to come up with innovative solutions for their problems.

So what do you guys think? Is latching on to big fry to promote our interests a good way to persuade someone? Looking forward to seeing your comments.

=====

You mentioned that the name cachet afforded by the US Government led to more active participation. Perhaps the monetary prizes have a significant role in motivating involvement? Just something to think about.

As to your question if getting ‘big fry’ to promote your interests is effective or not, it all depends on what you mean by effective. If you mean performance-per-cost effectiveness, there have been studies that show otherwise. [1] This paper can be distilled into a few main points: 1) influence to difficult to measure and predict; 2) ordinary influencers might be more cost-effective than ‘big’ influencers; 3) influence works best via quantity (number of influencers) rather than quality (fame of influencers) For example, in a case study, Lady Gaga and Michelle Phan were given brand endorsements for a cosmetic product on Twitter. Even though Lady Gaga had almost 100 times more followers than Michelle Phan, it was the latter that got more clicks on the Twitter link. A possible explanation is that when your social network is smaller, it becomes more personable and hence followers are more likely to engage with your content more. As can be seen, how popular the person or company is might not be a determinant of effectiveness in influence and persuasion.

[1] http://winteram.com/papers/wsdm333w-bakshy.pdf

===================================================

Penned with ❤️ for 08-624 and 17-704