7
Sep

Are Poor more rational than we think?

Poor-Man-Rich-Man

The wife and I have been debating on replacing an expensive mattress that we bought 3 years back. Over the years, we have stopped liking the mattress and it seems to be doing more harm than good. So now we both want to get rid of it. We already have a replacement mattress with us so the replacement cost is zero. The point of debate is – what should be the minimum price at which we should be willing to sell this expensive but useless mattress. In economic terms, what should be the WTA (willingness to accept) price? While I think any price is a good price, my wife is looking for a good deal. When I asked my friends, their responses also varied between 20 – 25% of the original cost. 

Is this response rational? As per classic economic theory, only incremental costs should impact decision. All historic costs are irrelevant since they are sunk costs. Considering there is no replacement cost of the mattress, my WTA price should actually be zero. So, why do we value these costs?

The idea of Loss Aversion which is discussed as part of Prospect Theory in Behaviour Economics literature explains this behaviour. We all operate from a reference point and evaluation happens based on the change from that reference point. Further, a negative change feels much larger than a positive change. For example, losing Rs. 1000 is much more painful than gaining the same amount.

 

loss_aversion

 

The same is applicable when we are replacing or giving up a product. When someone wants to replace a product, they evaluate not only the value of the replacement product (gain) but also the value of the product foregone (loss). If I believe that I have not completely utilised the value of my old product, I feel a sense of loss of giving it up. To cover up that loss, I want to recover at least part of my money. In the case of the mattress which is expected to be utilised for say 10 years, replacing after 3 years means a loss of 7 years of potential use.  This paper discusses Product Replacement decisions in detail. 

The theory of loss aversion has been used to explain multiple seemingly irrational decisions. In the book Thinking, Fast and Slow, Kahneman has described numerous decisions that people have made which are not in line with decisions predicted by classic economics. But there is also an interesting aberration – in case of poor or people with limited means, loss aversion may lead to  rational decisions. 

Being poor means living below the minimum level of income needed for adequate living. It means that one is always operating below the reference point that we discussed earlier. So, the individual is evaluating in the domain of losses. In this situation, money given to replace a good will not be interpreted as a gain but as a reduced loss. It will be considered as an aid to move towards the minimum reference point. Not taking money in this case would mean losing out an opportunity. Therefore, a person with limited means operating in the domain of losses is more likely to behave in accordance with the economic theory.

Can I use this learning to convince my wife? My colleague gave me an interesting argument. Instead of arguing based on replacement cost, I should argue based on the cost of not replacing the mattress. This could mean developing back pain and possibly paying for medical costs. In other words, I could use Loss aversion to my advantage. Whether this will be a convincing enough argument remains a matter of debate.

Image Sources: Here and here

8
Jul

Can Brazil still win the World Cup?

brazil-2014-world-cup-team

Brazil lost it’s star player in the game against Colombia. Some say they may have lost the world cup as well.  But does Brazil still have a upper hand over Germany today? Can playing at home give them the much needed edge? We believe it does. 

What constitutes this home advantage? Is it just the extra adrenalin flow when you the hear the roar of thousands supporting you? Or is there something sinister about it? Do referees who are suppose to be impartial tend to contribute to this home team advantage? 

Chicago Booth Business School has analysed several studies in this area and have come to some startling revelations. Yes, the referees have a bias towards the home team. The biases were clearly evident when it comes to taking that crucial decision as to how much injury time gets added to the game. 

Referees add extra time at the end of a match to make up for the time lost due injuries, substitutions or any other interruptions that could happen in a match. For the players, the coach and more so the fans the importance of the amount of time added to the game depends on the context at the scheduled end of the game. If  either of the teams are leading by a large number of goals at the regular 90 minutes of the game, nobody is really bothered about how many extra minutes get added to the game. But  if your team is trailing by a goal you will be happier if more injury time is added to the game. But if your team is leading by a slender margin of just one goal you will want the match to end as soon as possible. So the less the injury time gets added to the game the happier you are. 

Studies by Canice Prendenergast of Chicago Booth Business School and Luis Garicano and Ignacio Palacios Huerta of London School of Economics are showing that referees add on an average 2.93 minutes of injury time. But the crucial question is whether this average is affected by any bias. The surprising truth this study show is that the referees are biased while deciding the duration of time that gets added to a game. These researchers found that referees add 35% more than average time to a match when the crowd want the match to continue longer and add 29% less than average time when the crowd want the match to end earlier. If either of the teams are leading by a large number of goals, the added injury time is close to the average time. 

What  causes the referees to make these biased decisions? Blame it on the social pressure exerted by the home fans. The size and noise generated by the home crowd is a  key determinant that influences referees decisions.  It is also interesting to note that when match has few spectators or when the visiting fans are large in number, the tendency of the referee to take these biased decisions decrease. This was confirmed by a study by Per Pettersson-Lidbom and Mikael Priks of 21 Italian League matches which were played with no spectators. The authorities had asked the teams to play in front of empty stadiums due to the previous unruly behaviours of the spectators of those clubs. The studies confirmed that when there are no fans in the stadium, referees don’t display any bias.  

Now, we cannot play the game without fans. There may be less extreme solutions too. Studies in Germany show that when the football field is surrounded by a running track, there by  creating larger physical distance between the fans and the referee, the tendency of the referee to take biased decisions decrease.     

A soccer referee is assumed to be an epitome of unbiased behaviour. But the above studies are showing that even they are susceptible to behaviour biases. So this soccer world cup is not just a showpiece of human athleticism, but also a stage that displays the frailties of human behaviour, of the fans, players, coaches  and surprisingly that of the referees too. Lets see if Brazil is able to exploit these in the next game.

 

22
Apr

Nudging to Safety

 

diagonal

 

 

 

 

 

 

 

A recent article in The Economist paints a rather grim picture of Road Safety. Fatalities due to road accidents could surpass HIV related deaths by 2030. Bulk of these fatalities happen in developing and underdeveloped nations.

BBC Future carried a story that features Final Mile’s work in using Behavioral Sciences to improve Road Safety. Human error is the overwhelming cause of accidents and fatalities. Nudging or Behavioral design interventions aimed at reducing driver errors is one effective way to deal with the human errors. There is an urgent need to go beyond plain awareness campaigns. Driving is a highly over-learned and non-conscious activity, one that is very complex involving many senses and skills. Interventions that impact at a non-conscious level, therefore are likely to make a better impact.

The story by BBC Future collates some of the interesting examples of such interventions. Read more here

16
Apr

Noun or Verb

We are not referring to the grammatical usage here.

Consider a research context where participants are placed in a situation where they are tempted to cheat for personal gain. However, subtle change in instructions impacted their likelihood of cheating. When the instruction was in verb form (Please don’t cheat), participants cheated more than when the instruction was given in a noun form (Please don’t be a cheater).

Our earlier blogpost explains this behaviour. The primary idea is that the noun form invokes group identity while the verb form only refers to the action or the effort. People may downplay the action but that becomes very difficult when it comes a self-relevant noun.

So framing a praise in noun form may have a much more sustainable behaviour impact.  A recent New York Times article  also corroborated this idea.

But can we apply this rule universally? The book Social gives an interesting counter perspective. In his research on Altruistic behaviour, Dale Miller – a social psychologist at Stanford University, consistently found that people prefer not to accept an altruistic identify of self. So, when someone asks us why we decided to help, more often we tend to ascribe it as a selfish behaviour. Miller explains this as our tendency to conform to the cultural norm that human beings are self-interested.

Thus, sometimes we prefer focusing on the action while in other cases the identity works stronger. And these contradicting studies highlight the most important perspective of human behaviour – Context is critical.

Context-is-king-1024x768

Explanation of human decision making will always be incomplete without considering the influence of the contextual elements. The context of Cheating is different to the context of Altruism and as a result our appraisal of the stimulus also varies.

Impact of behaviour change communication will always be impacted by the context - be it using noun or verb.

 

Image Source: Here

3
Jan

Rise in ski helmet use, but no reduction in fatalities. Why?

 

UPDATE: Schumacher still in coma following skiing accident

The recent skiing accident of Michael Schumacher has shook the sporting world. The F1 world and many more people are praying for him. We also heard the impact was so severe that the helmet broke in to two. The doctors were sure that if it weren’t for the helmet, Schumi wouldn’t have even made it to the hospital. In this case, the helmet has played a big part in reducing the impact of the crash.

But, if we look beyond this individual case and examine the overall statistics on skiing accidents, helmet usage, head injuries and fatalities, there is a rather disturbing trend. There are news reports that suggest that while ski helmet usage has improved significantly, the fatalities due to head injuries have remained pretty much the same over the last 10 years. Helmet usage has tripled as compared to 2003. While there is a reduction in head injuries, the fatalities have remained the same. Ten year average fatalities from 2000 has been 41.5 fatalities per year, 54 fatalities having occurred in 2011/12. (NSSA facts). Of these 54, 36 were wearing helmets. (Skiing now might suddenly seem like a dangerous sport, but at 54 fatalities a year, its lot less dangerous than riding bicycles which lead to about 800 fatalities a year.)

Also, most of the fatalities involved Men, especially in the age group of late teens to 30′s. And a majority of these accident victims are above average at skiing, they’re not amateurs.

What would explain this ?

Overconfidence bias / Private optimism could be a credible explanation as far as the demographic of fatalities is concerned. There is research that supports a hypotheses that men tend to be more overconfident than women. An amateur would be cautious and hence ski at low speeds. An above average would be more susceptible to private optimism. Some 90% of motorists think that they are excellent/above average at driving. The believe that they are better than other drivers and that their chance of meeting an accident is lower than that of others. This could lead to people taking more risk.

The most intriguing observation though is this lack of correlation between increase in helmet usage and fatalities.  Here’s where John Adams’ take on individual risk management – the ‘Risk Thermostat’ could helpin making sense of this.  According to Mr. Adams, each individual has a specific level of risk-taking with which they are comfortable. If their sense of safety is increased, say by protective gear like seatbelts, helmets or systemic changes like ABS, their behaviour becomes riskier – they compensate for this increase in safety till the set-level is reached again. The safer we feel, therefore, the more risky our behaviour. There are studies that indicate that drivers who wear seat belts tend to drive faster. The net impact of ABS on car collisions has been negligible because drivers have driving faster and braking late. We have compensated for the safety feature. This is often referred to as “Offset hypothesis”

So, its likely here that those wearing helmets are skiing faster, taking more risks and becoming more adventurous, simply because there is an increase in overall sense of safety. So, in a strange way, the more safer we feel, the more risky our behaviour.

How can we solve this ?

While we improve the safety features, its important to induce a sense of vulnerability, a feeling of being unsafe. This could be through better design of safety systems and signage or by having interventions at point of action that introduce an element of vulnerability. The idea is not to take regressive steps like reducing safety gear or features. Safety standards should be high, but the place should feel a little unsafe.

For now, lets hope that Schumi pulls through.

30
Dec

IbIn progress report 2013

The India Backbone Implementation Network (IbIn) was launched in April 2013 by The Planning Commission of India, with the primary objective of promoting widespread capabilities in the country to systematically convert the manifest confusion to coordination, and rampant contention to collaboration, so that intentions can be converted into implementation. Final Mile is one of the partners in IbIn.

As 2013 draws to a close, Planning Commission has put out a 2013 progress report on the status of the Initiative. IbIn has made remarkable progress in spite of lack of any seed capital or a formal institutional structure.

The report also features 2 projects of Final Mile. One is a project to improve usage of toilets built under the Nirmal Bharat Abhiyaan aimed at making our villages free of open defecation. This project is currently in progress in Karnataka and is sponsored by Arghyam

The other project is involves changing the behaviour of government agencies on ground with a view to ensure better adoption of internet based government services. The report can be downloaded here

IBIN_Progress Report_Dec2013

2
Dec

First Impressions Last

Ever wonder what is the maximum number of items a drop down list can have? Does the sequence of the items in the list affect the interaction of the viewer?

The Serial Position Effect explains the phenomenon of memory where items presented at the beginning and ends of lists are more likely to be remembered than those in the middle of the list. The term ‘Serial Position Effect’ was first coined by Hermann Ebbinghaus based on studies on himself and others, where he best recalled items from the end of a list (the recency effect) and more frequently among the first items than among the middle ones (the primacy effect).

For example, consider the following list and try to remember the items in the list.

List image

 

It will be observed that you will more reliably recall items in position 1, 2, 6 and 7 than those in the position 3, 4 and 5.

When we recall the first few items on the list, its called the Primary Effect. Primacy effects happen because there is more chance of earlier items being stored in long-term memory (long term memorisation requires rehearsal of the list with less likelihood of remembering the further along the list you go). When lists are rapidly presented, the effect is weaker because there is no time for long-term memory to work. When lists are presented slowly, the effect is stronger as there is more time to store in long-term memory.

Recency effects occur as the last few items in a list are still in working memory and therefore readily available. The strength of this effect is not impacted by the speed of delivery. However, passage of time has a huge effect on the recall of information, weakening the recency effect. If you are distracted by other matters for even 30 seconds, then this effect completely disappears. This is not true on the primacy effect which relies on long-term memory.

What does this have to do with making great content, you ask?

Well, great content starts with great copy, and the arrangement of that copy can dramatically impact how much or how little your viewer will remember. Your content may relay several messages in a list, or you might ask viewers to remember a coupon code, phone number or SMS code. In each of these cases, even a simple modification to how the information is presented can have a significant impact on how much the viewer will remember later.

You can take advantage of this knowledge when presenting information in lists (be it a set of links, your sales pitch, a feature list, client list, etc.)

- Place the least important items in the middle of your lists because these items tend to be stored less frequently in long-term memory and working memory.

- If the viewer’s decision is to be taken long after exposure (> 30 seconds), then place the most important items first. If the decision is to be taken immediately after reading the list, then place the most important item last on the list.

If all the information is equally important, then the best thing to do is group the items in chunks. Presenting long lists of information places significant strain on limited attentional resources and restricted memory systems, especially short-term memory, where it appears only three or four items ‘chunks’ can be maintained at one time. Therefore, you should reduce the strain on viewers by presenting information in small pockets, or chunks, and limit the amount of distraction between presentation of items and recall.

Quite simply, Chunking is a way of arranging information so that your memory has to recall fewer items later. Chunking is the most effective when all of the items in the list are roughly the same “type” and “size” (e.g. numbers versus words versus phrases). That’s where coding comes in. Simply put, Coding is how our brains make things easier to remember by arranging them into groups of like items. By “like items,” I mean practically any grouping that makes the list’s elements seem more similar to each other.

Our brains do chunking and coding automatically as we make our way through the world. However, given how much competition and visual clutter your content may face, you might want to take these few steps to make sure the process is as easy as possible. After all, viewers may only be gracing you with a few seconds of their attention, so you need to make that exposure count.

30
Oct

Andhra Bus Accident: A tale of different narratives

Blog Andhra Bus Accident

A tragic accident involving a Volvo bus happened on the Bangalore – Hyderabad highway today. 44 passengers are feared killed. The bus driver, the helper and five passengers managed to escape death.

So, how did this accident happen? Since morning, we have heard multiple narratives of the accident across different media channels. Initially, it seemed like the bus hit an Oil tanker that resulted in the bus catching fire. Later, a channel reported the bus driver’s statement that the primary cause was a tyre bust which resulted in the bus skidding onto the side of the road.

After a few hours most channels reached a consistent narrative. The bus was being driven at high speeds and while overtaking another car it hit a culvert that impacted the fuel tank. The fuel tank burst into flames which soon swept over the entire bus.

Was the press irresponsible or misleading? Or is there are a larger issue at play?

As humans, we have a high degree of curiosity. We want to know why and how the accident happened when in the immediate term we should be worried about rescue and relief. This curiosity means, we are always searching for a story that makes sense, one thats logical and mostly, one that is a good story. This pushes us towards a narrative which is a significant problem with most investigations. People look for a narrative first rather than going about it objectively and this narrative kills objectivity and truth.

Once people get into a story telling mode, their reporting is influenced by a number of biases. The narratives also tend to be swayed by the stories mentioned by others around. The driver’s themselves tend to explain traffic accidents by reporting circumstances of lowest culpability with credibility (Baker’s Law). In addition, our own interpretation adds another layer of subjectivity. Remember Chinese Whispers?

Using a scientific approach to investigations can help us overcome this issue of subjectivity and reach an accurate diagnosis of the incident. Paul Meehl, an American psychology professor, in his 1954 book “Clinical vs. Statistical Prediction: A Theoretical Analysis and a Review of the Evidence” argued that mechanical methods of data combinations can make more efficient decisions about patients’ prognosis and treatment as compared to clinical or subjective methods. These mechanical methods can use a combination of data, checklists and even clinical judgments to predict the outcome.

The idea is to move from a pure intuition based judgement towards a more objective and systematic way of diagnosing the incident. The system needs to ensure that we come to the narrative at the end of accident investigation rather than lead it. In one of our recent work on Road Safety, we have worked on designing an accident investigation system that avoids bias reporting. The new process captures all the accident elements objectively before arriving to the final prognosis. This system is now being used for future investigations.

We know that media’s job may not be to conduct accident investigations. However, they can also follow simple rules such as not releasing driver’s version immediately and avoid such varied narratives of the same tale.

The key issue is that if we get lost in the narrative and end up drawing wrong inferences, we will be learning wrong things and addressing wrong problems.

Image Source: Here

15
Oct

Chicago transit authority – Right diagnosis. Wrong Prescription

 

 

CTA

 

 

 

 

 

 

 

 

 

 

We caught this piece of communication developed by Chicago Transit Authority(CTA). The basic point of this is to tell people that the train is faster than you think. There is a scientific backing to this misjudgment of speed. Our brain underestimates the speed of large objects, including trains. Often, while crossing tracks, even after spotting a train, we tend to attempt to cross because the train appears to be moving slower. (Leibowitz hypothesis). Look at this amateur video to see this in action.

 

http://www.youtube.com/watch?v=3taglTyvlxc

Now, how do we solve this problem ? The approach used by CTA is based on the assumption that making people aware of this shortcomings is good enough to solve this problem. The belief is that people will take in this information, process it and put it to use when they are in such situation. The same kind of thinking that automobile companies adopt when it comes to warning about the distance of objects in the rear view mirrors.

This approach in our view is flawed.

For starters, the speed perception or underestimation is a non-conscious activity. It happens through automatic processing. We don’t actually stand next to tracks and carry out an accurate estimation of speed. We are not equipped with such capabilities, which is why we use speed guns and other measurement tools to judge speed accurately. We cannot presume that we can suddenly make this non-conscious process of speed judgement in to a conscious one. And that people have the time, intention and cognitive ability to judge the speed accurately. This is expecting too much. In fact it is likely that most people will not even remember this message while crossing tracks. They are likely to be pre-occupied with many other things and are likely to be in ‘Auto’ mode. This seems like a classic case of right diagnosis, but wrong prescription. Is there a better way to deal with this problem?

If the problem, fundamentally is at a nonconscious level, the solutions should work at this nonconscious level for it to make a definite and quick impact. The solution should make the brain recalibrate the speed of the train in an ‘Auto’ mode where it doesn’t need to deliberate and expend energy. These interventions have to be at the point of action. While it might appear that inside a train is close to being on tracks, mentally these are very different contexts. Being in the train and crossing the tracks on foot are very different contexts.

How can we get the brain to recalibrate the speed and get the judgement right. We can do this by providing stationary reference points. The highly successful ‘Yellow Lines’ intervention is one where the yellow lines act as speed references.

YL

 

 

 

 

 

 

 

 

 

These are lines painted across the railway tracks either side of crossings. As these yellow lines disappear under the train, the brain can instantly get the speed judgement right and take a decision not to cross the tracks. The beauty with this intervention is that it works at a nonconscious level, has an instant impact and is low cost. Most importantly, it is at the point of action. And this has worked in reducing fatalities significantly in Mumbai Suburban railway network where hundreds of thousands of people trespass across thousands of crossings in Mumbai. At an average of 10 fatalities a day, it is the largest cause of unnatural death in Mumbai city. Yellow lines, coupled with other interventions have reduced the fatalities significantly and this case is well documented. Read the story in The Boston Globe and BusinessWeek
We often have this temptation to believe that making people aware of a problem will solve it. This seems to be the thinking at CTA. However, for us to see impact, we need to make it easier for people to correctly judge the speed of trains. Not by telling them that the train is faster than you think, but by helping them take a right decision quickly and easily when it matters.

26
Sep

Which iPhone are you buying?

Blog iPhone 5s vs 5c pic 2

Yesterday my colleague and I were debating which iPhone version he should buy since he has finally decided to give up his old blackberry. With the launch of the 2 new models, his choice set included the earlier model iPhone 4S, the step-down model iPhone 5C and the most high end version iPhone 5S. While we were comparing the three models, the iPhone 5c quickly went out of favor and the debate centered around the other 2 models.

This anecdotal experience, however, seems to be a representative of the market. While Apple continues to break all the records in iPhone sales, an interesting pattern has emerged. Analytics research suggest that the 5S model is outselling the 5C model by a margin of more than three to one in US. In Australia, this seems to be as high as eleven to one.

Why would a higher price model outshine a cheaper version by such a large extent? Lets try to explain using the concept of Context Effects. People’s preferences are often uncertain and are constructed dynamically in response to a choice set. This dynamic nature can thus be influenced by the format and the manner of the presentation.

There are multiple Context effects that influence consumer’s choice behavior. Attractiveness effect is a key one that is relevant in this situation. In an expanded set of choice, the similarity between certain options can make one of them appear more attractive. This will happen if one option makes the other one look superior. For example, in a choice set with two options X and Z, introducing a third option Y which is similar but inferior to option Z, leads to an increase in the preference of the superior option Z.

The iPhone 5C and 5S models have many technical similarities in terms of display, memory options, ports, thickness etc. iPhone 4S seems more different. In fact, introducing both the new models as part of the iPhone 5 series, ensures that iPhone 4S stands out as a separate evaluation option. The low price difference between 5C and 5S coupled with the faster processor and a slightly better camera in 5S, establishes the superiority of 5S over 5C. As consumers, when we compare the 3 models – 4S, 5C and 5S, 5S comes out as the most favorable option.

iPhone 5C works like a decoy by letting 5S come out as a clear winner. In other words, iPhone 5C may be key to the success of iPhone 5S.

So, which iPhone are you planning to buy?

 

Image Source: Here

19
Sep

Why didn’t the driver stop?

Blog Level Crossing Image

In a a tragic accident between a bus and train in Ottawa, Canada, 6 people died, 8 were critically injured and 30 were hospitalized. The crash happened in the morning rush hour when the bus ran through the guard rail and collided with a train. Passengers in the bus reported screaming at the bus driver while on course but even that did not prevent the accident.

Investigations by the Transportation Safety board to determine the cause of the accident have begun but an obvious question is – Why didn’t the driver see the oncoming train even in broad day light? Did he fall asleep? Were there issues with the brakes? Is this a localized driving problem?

Accidents at level crossing is a universal issue. At least 6000 people die at level crossings every year and there is even an International Level Crossing Awareness Day (ILCAD) dedicated to this cause. Canada alone has seen 257 accidents in the past decade. In India, 63% of railway related fatalities are attributed to accidents at unmanned level crossings.

So what explains this behavior that seems to be prevalent globally? One possible explanation is the negative expectancy of the train by the bus driver. The way we see things is based on how we look for them. It is influenced by what we expect to see. So we are perceptually restricted and tend to pay attention to things that we expect. Psychologists term this phenomena as “Inattentional blindness”. This famous Selective Attention Test cleverly demonstrates our propensity to miss things.

Why do we suffer from this problem? It is actually an evolutionary mechanism to help us cope with the sensory chaos in the world. It helps us makes sense of the infinite visual stimuli that is always present around us and act accordingly. However, in certain cases this mechanism may backfire.

The attention spans of experienced drivers who have driven through a place multiple times tend to be very low due to the similarity and monotony of the environment. The low probability of finding a train also influences their expectation. Thus, breaking this monotony and changing their expectation is key to managing this behavior. This is the focus of our solutions that have been implemented at certain unmanned level crossings in India. Refer this blog post for more details.

Could these solutions prevent this tragedy? We can’t be certain but our research shows they can.

Image Source: Here

16
Sep

Turning cities into living labs. 10 examples from around the world

Rapid urbanisation is resulting many anticipated and unanticipated problems. The nature and intensity of such problems varies rapidly across geographies. There is a need to be creative to solve some of these problems as some of the problems are entirely new and some unresolved in spite of a lot of attempts.

Edward Gardiner, who leads the Behavioural Design Lab wrote a piece in the Guardian recently on turning cities into living labs. 10 really interesting examples cited range from City Science at MIT Media lab, JPAL, London bridge redevelopment. It also features a project executed by Final Mile in minimising collisions at Rail road crossings in India. Read more about the 10 examples here

Next Page »