Time and group size inconsistencies in preferences on environmental policy interventions

Last registered on August 21, 2020

Pre-Trial

Trial Information

General Information

Title
Time and group size inconsistencies in preferences on environmental policy interventions
RCT ID
AEARCTR-0006316
Initial registration date
August 20, 2020

Initial registration date is when the trial was registered.

It corresponds to when the registration was submitted to the Registry to be reviewed for publication.

First published
August 21, 2020, 10:17 AM EDT

First published corresponds to when the trial was first made public on the Registry after being reviewed.

Locations

Region

Primary Investigator

Affiliation
ZEW - Leibniz Centre for European Economic Research

Other Primary Investigator(s)

Additional Trial Information

Status
On going
Start date
2020-08-10
End date
2020-11-30
Secondary IDs
Abstract
Given subjects have preferences for environmental well-being, but fail to perform pro-environmental behaviors
due to a lack of self-control, they are assumed to have a positive willingness to pay for commitment devices
to enforce pro-environmental behavior. Therefore, I plan to investigate whether there is a positive demand for environmental
policy interventions as self-commitment devices on an individual level and whether the preference
for a certain environmental policy is dependent on the time of implementation and the amount of people affected
by it. In a laboratory experiment, I use a pro-environmental real effort task as an opportunity for subjects to behave
environmentally friendly. Prior to the task, subjects are asked to rank policy intervention according to their
preferences. The policy interventions vary in their degree of external pressure from lenient, in a social comparison
nudge, to strict, given a punishment scheme. Since I hypothesize that the intervention choice is dependent on
the time of implementation of the pro-environmental task and group size affected by the policy, the treatment
varies across two dimensions. Firstly, the time of implementation is varied as subjects must either perform the
pro-environmental task within the same day or with a delay of one week. Secondly, the regulation either affects
only the person choosing the intervention or it affects also others in the group. It is assumed that subjects with
lower self-control demand stricter policy interventions. In addition, the more the task is postponed into the future
and the greater the number of other people affected by the choice, the higher the likelihood of choosing more
compulsory interventions.
External Link(s)

Registration Citation

Citation
Alt, Marius. 2020. "Time and group size inconsistencies in preferences on environmental policy interventions." AEA RCT Registry. August 21. https://doi.org/10.1257/rct.6316-1.0
Experimental Details

Interventions

Intervention(s)
The design for the laboratory experiment consists of treatment variations across two dimension. The first dimension
is given by the timing of the implementation of the environmental policy regulation. The second dimension is
structured by the number of regulatees affected by a choice on an environmental policy. I implement a two times two between-subject full factorial design. In addition, the treatment variations contain a within-subject component, as on the regulatee-size dimension participants
who are choose an environmental regulation not only for themselves but also for other participants are
asked to make the same choices in case of being paired with one other person and four other individuals. The order
of these choices is randomized across subjects.
Intervention (Hidden)
3.1 Treatment "self x immediate"
The treatment "self x immediate" serves as a baseline comparison. As in every treatment, in "self x immediate" participants
must complete two sessions, which lie a week apart fromeach other. In the first session, participants have
the option to contribute to a pro-environmental cause by working on a real-effort task for eight minutes. The real
effort task represents a decoding task similar to Dorner (2019), in which an eight digit code must be translated into
letters with the help of a translation table. For three correctly decoded array of digits, a donation to an afforestation
project is generated, which suffices to finance the plantation of a tree. There exit four versions of the task, of which
three represent a policy intervention. Table 2 provides an overview of the versions. The first version, "NoIncentive",
has no policy intervention and participants can work on the decoding task without any consequences for their payoff
and no information on performance of others. In the second version, "Social Comparison", participants receive
real-time feedback on their own performance compared to the average performance of the whole sample of other
participants and the average performance of the best 25 percent of participants. These performance comparisons
are retrieved from a pilot of this study. Similar to the "NoIncentive" version, solving the decoding task has no consequences
for the payoff of participants. In the third version, "Monetary Incentive", participants refrain from a part
of the participation fee and get compensated for this by receiving a monetary reward for each correctly answered
question. The reward scheme is calibrated in a way that performing as good as the average in the pilot session leads
to the identical payment as in the "NoIncentive" version, while performing better leads to a larger payoff (capped
at 2 Euros). Performing worse than the average in the Pilot session leads to a lower payoff. In the fourth version,
"Punishment", participants must solve at least ten decoding tasks to avoid a reduction in payoff of a share of 5/17 compared to the "NoIncentive" version. Prior to the decoding task participants are asked to rank these versions
given their preferences for it. It is also conveyed to them that the preference ranking also influences the implementation
probability, favoring the versions that are further up in the ranking. Thereafter, participants willingness to
pay for implementing the first preference instead of another version is elicited to obtain a measure on how strong
their preferences for a certain version is. This is carried out by asking participants to make binary choices concerning
their first choice and each of the three other choices. In these ten binary choices for each of the three pairs (first
preference vs second preference, first preference vs third preference, first preference vs fourth preference), a monetary
compensation is offered to participants ranging from 0eto 5efor not implementing the first preference. Once
all of these 30 decisions, one of these decisions is randomly selected, favoring the binary comparison between first
and second preference in probability. If the first preference is chosen in the selected choice, the respective version
choice will be implement in the decoding task. If another preference is selected in the choice, the respective version
is implemented in the decoding task and the compensation fee is added to the overall payoff. In the second session,
participants are asked to solve two decoding task as a trial and are then redirected to a Questionnaire containing a
range of incentivized and non-incentivized questions (see Table 3).

3.2 Treatment "self x delay"
The treatment "self x delay" is constructed similarly to the treatment "self x immediate". However, it distinguishes
in the implementation time of the selected version of the decoding task, e.g., the point in time when participants
will have the option to solve the decoding tasks for eight minutes, in order to contribute to a pro-environmental
cause. Figure ?? displays the difference in the structure of the experimental design between both treatments. Instead
of taking place directly after having made the decisions on the preferences for a certain version of the decoding
task, in "self x delay" it is postponed by a week to session 2. Hence, in "self x delay participants make decisions,
which will become relevant a week after. By comparing the preferences for the versions in the decoding task in "self
x immediate" and "self x delay", I can identify whether the preferences for regulation to induce pro-environmental
behavior varies depending on the implementation time of this regulation, e.g., whether the preferences are timeconsistent.
3.3 Treatment "group x immediate"
The treatment "group x immediate" distinguishes from the baseline treatment ("self x immediate") as the policy
regulation of pro-environmental behavior in the decoding task will not exclusively become relevant for themselves,
but also for other participants in the group. These groups consist either of two or four individuals. After all participants
have made their decision it is randomly selected, whose preferences will become relevant for the whole
group. The decisions in the case of one other or three other group members are inquired within subjects by asking
for regulative preferences in the case of one other person in the group and three other players in the group
prior to revealing the actual group size. The order, in which participants are asked to make a decision for the depending
group size is randomized for each individual. Hence, by the means of this design I am able to compare
preferences for regulative measures concerning the pro-environmental real-effort task given variations in the size
of other regulatees affected by the decision. This size varies from zero in "self x immediate" to one and three in
"group x immediate". This controlled variation provides insights on whether and to which degree individual preferences
vary with the size of other individuals affected by the particular regulation.

3.4 Treatment "group x immediate"
In the treatment "group x delay", the treatment variations on the time and the regulatee-size dimension are interacted
to analyze possible interaction effects of these variations. This is implemented in the treatment by asking
participants to make a decision on preferred regulative measures, which affects themselves and other players in
the group (in case of a group size of two and four participants), in the first session while informing them that the
actual pro-environmental effort-task including the corresponding regulation will take place a week after in session
two. I assume that a substantial share of decisions on choosing self-regulation devices are inconsistent with respect
to time. However, this behavior might differ if asked to choose a regulation that will also affect others. Given
that preferences for regulation that exclusively affects others involves less time-inconsistent decisions, we should
be able to observe less time-inconsistent behavior, given that participants include that they make choices also on
behalf of others into their decisions. The experimental design is able control for this by comparing differences in
decisions along the time dimension of treatment variations in the case of participants deciding only for themselves
and the case in which participants decide for themselves and others in the group.
Intervention Start Date
2020-08-10
Intervention End Date
2020-11-30

Primary Outcomes

Primary Outcomes (end points)
In the regulatee-size dimension, the matter of investigation lies in the difference of this willingness to accept depending onwhether participants decide only for themselves or whether the regulation will also affect others in the group.
In the case of the treatment variations across the time dimension treatment, the effects are slightly weaker. Figure 2
shows the amount of observations per treatment, which would be necessary to detect a significant effect in the
time difference of the willingness to accept to change from no regulation to monetary incentive
Primary Outcomes (explanation)
Difference in willingness to accept to change from the first preference to the second third and fourth preferences will be used to contruct the willingness to accept to not conduct the "NoIcentive" version and conduct either the "Nudge" verison, the "Monetary incentive" version and the "Punishment" version instead.

Secondary Outcomes

Secondary Outcomes (end points)
Secondary Outcomes (explanation)

Experimental Design

Experimental Design
The design for the laboratory experiment consists of treatment variations across two dimension. The first dimension
is given by the timing of the implementation of the environmental policy regulation. The second dimension is
structured by the number of regulatees affected by a choice on an environmental policy. I implement a two times two between-subject full factorial design.In addition, the treatment variations contain a within-subject component, as on the regulatee-size dimension participants
who are choose an environmental regulation not only for themselves but also for other participants are
asked to make the same choices in case of being paired with one other person and four other individuals. The order
of these choices is randomized across subjects.
Experimental Design Details
Randomization Method
Within session randomization of treatment assignment.
Randomization Unit
Individual
Was the treatment clustered?
No

Experiment Characteristics

Sample size: planned number of clusters
400 individuals
Sample size: planned number of observations
400 individuals
Sample size (or number of clusters) by treatment arms
400 individuals
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
71 observations
Supporting Documents and Materials

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information
IRB

Institutional Review Boards (IRBs)

Post-Trial

Post Trial Information

Study Withdrawal

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information

Intervention

Is the intervention completed?
No
Data Collection Complete
Data Publication

Data Publication

Is public data available?
No

Program Files

Program Files
Reports, Papers & Other Materials

Relevant Paper(s)

Reports & Other Materials