Lecture 9 (Ethics) Flashcards
If we should rank the strength of the policy instruments in directly changing behavior, how would it look like?
- Rules and regulations
- Economic incentives
- Nudge
- Information
When should we try to change people’s behaviour?
If we see a market failure, where the free market is inefficient in distributing goods and services.
If we are faced with externalities
- But it can be hard to prove that there is an externality, when we do not have full information about the. market.
Intention to action gap.
What is the difference of a nudge and sludge/dark patterns. Define them.
Nudge is a way of changing behavior without restriction to obtain good behavior.
SLUDGE:
Unnecessary friction that
prevents a good
behavior
‘Dark patterns’:
‘user interface design choices that
benefit an online service by coercing,
steering, or deceiving users into
making unintended and potentially
harmful decisions’
What types of dark patterns exist?
Confirm shaming - is when designers use manipulative language and emotionally charged design to make you feel really guilty about canceling a service.
Eg. Duolingo redesigned their owl to start crying when you wanted to leave. Do you wanna make the owl cry.
Misdirection - is when designers hide the things they do not want you to see and draw lots of attention to the things that you do in order to help you behave the way that they want.
Eg. Flight companies move extra cost to the end of your transactions. Only showing the min price without any add ons at all. By moving the extra necesary price to the end increase sales with 21%. People might be too lazy to not buying after using time to sign in.
Eg. Instagram have an algorithm with beholds likes, so you feel a bid discourage, but then bam they let all the likes stream in even though some was liked hours ago. That will give you a huge dopamine rush and keep you addicted.
Trick question - is when designers use complicated language, which would keep you on their platform
Eg. “Are you sure you want to cancel your account? Cancel. Continue.”
Privacy Zuckering - is when designers combine all of these techniques to reveal much more information about themselves then they intend to.
- Eg, If you agree to the terms. Facebook is able to do digital graberobbering, where they still collect data on you even though your account is closed.
Why is dark patterns hurtful to our behavior?
The problem with this is, that big tech companies get so much information on you. They sell their data through data brokers, who in average had around 3000 pieces of personal information on almost every Americans in the US. They can then group this information together in maybe discriminating or dangerous. Eg. they group people together into a group called “Elderly and easily persuaded”, and then sell the data to highest binding gamble market, because theses markets can increase their income on bad behavior. Another example could be health apps, which has been documented in selling your data to platforms like Facebook, Pinterest, Google and Snapchat, so they can use it to affect you in different ways they fell to, eg, personalized experiences = advertisements. When you are your lowest and see a therapist or a doctor they are by law not allowed to give information further, but because tech moves so fast the law can not catch up with it. They are using information about your health to sell your advertisement.
Companies like Facebook, make money on our attention. If they can keep you hooked, they earn more money. So they use every trick in the book.
How can we spot a if the toll used is a sludge and not a nudge?
Sludge Audits
- First, institutions should undertake a periodic ‘lookback’ at existing
burdens to see if the current ‘stock’ can be justified and to eliminate
those that seem outmoded, pointless or too costly. - Second, institutions should choose the least burdensome method for
achieving their goals. This is essentially a requirement of cost–
effectiveness. - Third, institutions should ensure that the benefits of administrative
burdens justify the costs. - Time is the most valuable commodity that human beings have. We
should find ways to give them more of it.
What is FORGOOD
Framework for ethical nudging
“F”airness - Does the behavioral policy have undesired redistributive effects?
“O”penness - Is the behavioral policy open or hidden and manipulative?
“R”espect - Does the policy respect people´s autonomy, dignity, freedom of choice and privacy.
“G”oals - Does the behavioral policy serve good and legitimate goals
“O”pinions - Do people accept the means and the ends of behavioral policy?
“O”ptions - Do better policies exist and are they warranted?
“D”elegation - Do the policy-maker have the right and the ability to nudge using the power delegated to them.
Real experiment with danish unemployed. Outline the experiment and use the FORGOOD framework on it.
The employment ministry
noticed that many
unemployed people did not
know when their benefits
were running out and when
they needed to get a new
job or work part time to
prolong the benefits.
They created a nudge with a timeglas there where counting down.
Fairness
* Does the behavioural policy focus too much on one
group and neglect another group that is in more need
of an intervention?
Openness
- Is it possible for the person under the influence of the
behavioural policy to identify the policy and its
influence and impact?
Respect
* Does the behavioural policy respect people’s freedom
to choose?
Goals
* Does the behavioral intervention serve goals that are
ethically acceptable?
* For behavioural interventions that aim to improve people’s
lives, do these interventions really make people better off
and how is this ‘better off’ defined?
Opinions
* How does the public view the means used by the
behavioural intervention?
Options
* Is the behavioural policy more or less cost–effective
than other policies?
Delegation
* Does the policy-maker have the competency to design,
administer and evaluate the behavioural policy?
Summary
* A framework cannot decide whether an intervention/nudge
is good or bad.
* Using the framework can help to think through an
intervention or compare different interventions to each
other to choose the ones that have the highest scores in
most dimensions.
* Not all questions will be relevant for all interventions.
How do we measure welfare effects of a nudge? What is welfare and how do we maximize it?
There are many different ways of measuring welfare dependent on the context and opinions.
Eg.
Education, real income, job satisfaction, employment prospects, leisure time, environment/air pollution, happiness, life expectancy/health care, cost of living etc.
Sunstein et al. (2017, 2018)
a) Outline the context and the design:
b) outline the results from the survey. Discuss them.
Question a)
Cass Sunstein and co-authors have run surveys in 15 countries to ask a
representative sample of 1000 citizens what they think about different interventions.
design:
eg. comparing the same policy using different nudges:
1) The federal government encourages (without requiring) electricity providers to
adopt a system in which consumers would be automatically enrolled in a “green”
(environmentally friendly) energy supplier, but could opt out if they wished.
2) The federal government requires large electricity providers to adopt a system
in which consumers would be automatically enrolled in a “green”
(environmentally friendly) energy supplier, but could opt out if they wished.
Question b)
results
see figure 26
We do not have big differences between the two types of interventions achieving which tries to achieve the same policy. For example has UK approval in percentage only decreased with 0.1 percentage point going from one nudge to the other. This tells us individuals might be more concerned about the what the nudge tries to achieve then what the type of nudge have been used to get their.
Discussion of results on results:
The type of nudge matters much less than what the nudge is trying to
achieve.
Green party voters are more in favor of green nudges.
There are no cost-benefit calculations taken into consideration. Some
nudges will be much cheaper to implement than others.
Is this really showing approval of nudges or just approval of policy making in general? Survey evidence gives us easy to interpret and cite answers – however, people do not always know what is best for them in. They might support the policy, but do not know the effect of the nudge used, which if they knew would argue was to manipulative and controlling of their or the markets freedom.
What are dark patterns in relation to Mills et al. 2024. Dark patterns audit framework (DPAF).
Detour (resampling “confirm shaming”) - A dark pattern designed to delay or distract a decision-maker.
eg.
Up-selling at an online retailer checkout.
Roundabout - A dark pattern designed to tire or bore a decision-maker when they are trying to achieve an outcome.
eg. A long series of ‘are-you-sure’ decision-maker, or otherwise redirect a checks.
Shortcut - A dark pattern designed to exploit feelings encourage decision-makers to take ‘easier,’ but ultimately more costly, decisions.
eg. Salient ‘Accept All’ online created by detours and roundabouts to tracking (cookie) prompts.
Forced Action (resampling privacy Zuckering) - dark pattern which forces a decision-maker to incur an additional, unexpected or undesired cost, in order to achieve their objective.
eg. Mandatory terms and conditions requirements for new users of an online service.
When we talk about the welfare effects of nudges, what have Allcot and Kessler on their mind?
They argue that their extensive model predicts that 41 percent of HER receivers would want
to opt-in to the reports.
However, in the real world only 1,5 percent opt-in in a similar setting.
Eg. their extensive model, which do not include:
1. Installation costs of more efficient appliances
2. Experience costs of setting the thermostat
lower/higher
3. Psychological (dis-)utility from receiving the
reports
Nevertheless, the experiment shows that we can increase welfare by targeting better who receives the nudge. But we should be aware of the difference the cost of using nudge in a model and in the real world. The real welfare effect might be too small to use the money to implement it, even though it has a positive effect off welfare.
Damgaard and Gravert (2018) The hidden costs of nudging.
A) What is the hidden cost?
b) What is the result of the study?
A)
Context:
In a field experiment with a charity, we find reminders increasing intended behavior (donations), but also increasing avoidance behavior (unsubscriptions from the mailing list).
Costs of sending reminders
1.Direct costs – low due to technology
2.Hidden Costs
- .Annoyance or guilt costs to
recipients
- .Loss of future donors
We will use unsubscriptions as a revealed preference
measure!
B)
Results:
In line with Prediction 4, the Low Frequency treatment reduces the unsubscription
rate from 0.49% to 0.30% (see Figure 26). That is a reduction of 39%. When we lower the frequency of reminders, the amount of people who unsubscribe changes. The level of donations also slightly increases.
