Measuring attention to sales taxes

Last registered on September 19, 2016

Pre-Trial

Trial Information

General Information

Title
Measuring attention to sales taxes
RCT ID
AEARCTR-0001597
Initial registration date
September 19, 2016

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
September 19, 2016, 2:55 AM EDT

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

Last updated
September 19, 2016, 3:56 AM EDT

Last updated is the most recent time when changes to the trial's registration were published.

Locations

Region

Primary Investigator

Affiliation
UC Berkeley Economics

Other Primary Investigator(s)

Additional Trial Information

Status
In development
Start date
2016-09-19
End date
2016-09-30
Secondary IDs
Abstract
I introduce a new experimental design that 1) enables a direct estimate of the variance of underreaction to not-fully-salient-taxes, 2) makes it possible to test for “overreaction” as well as “underreaction” 3) enables many novel tests of a model of “deliberate inattention” to taxes and 4) from a statistical standpoint is generally a significantly more efficient design for estimating underreaction than existing laboratory and field experiments.
External Link(s)

Registration Citation

Citation
Taubinsky, Dmitry. 2016. "Measuring attention to sales taxes." AEA RCT Registry. September 19. https://doi.org/10.1257/rct.1597-2.0
Former Citation
Taubinsky, Dmitry. 2016. "Measuring attention to sales taxes." AEA RCT Registry. September 19. https://www.socialscienceregistry.org/trials/1597/history/10673
Experimental Details

Interventions

Intervention(s)
Intervention Start Date
2016-09-19
Intervention End Date
2016-09-30

Primary Outcomes

Primary Outcomes (end points)
The likelihood of buying a household product, across prices and tax rates of different sizes. These shopping decisions will be used to construct measures of underreaction to taxes. In addition to estimating average underreaction in the different "tax regimes," I will also estimate the variance of underreaction in each regime, as well as the variance of changes in underreaction across the tax regimes. (Full details of all of the hypotheses to be tested, as well as the empirical model and the estimation approach are in the pre-analysis plan.)
Primary Outcomes (explanation)

Secondary Outcomes

Secondary Outcomes (end points)
Secondary Outcomes (explanation)

Experimental Design

Experimental Design
Subjects will make decisions about their willingness to buy household products at various prices and in three different tax regimes 1) no sales tax, 2) the standard sales tax in the subject's city of residence and 3) triple the standard sales tax rate in the subject's city of residence. This is a within-subject design that enables estimates of individual differences in underreaction, as well as how underreaction changes from the "standard tax" regime to the "triple tax" regime.
Experimental Design Details
The experiment will be implemented through ClearVoice Research, a market research firm that maintains a large and demographically diverse panel of participants over the age of 18. This platform is frequently used by firms who ship products to consumers to elicit product ratings, but is additionally available to researchers for academic use.

Because the experimental design involves language about the sales tax rate that study participants pay in their city of residence, to avoid confusion I asked ClearVoice to only recruit panel members from states that have a positive sales tax. This excludes panel members from Alaska, Montana, Delaware, New Hampshire, and Oregon. The remaining 45 states should all be represented in the final sample.

After agreeing to participate, and before learning anything else about the design, consumers will be asked what state/county/city they live in. (To correctly determine the money spent in the experiment, the study software will match this information to a dataset of tax rates in all cities in the United States.)

Next, subjects will read the instructions. They will learn that they will see three different products (drawn from a set of 9 possible products) in three stores, A, B, C, for a total of 9 sets of decisions about the 9 store-product pairs. In store A, subjects will face no sales tax. In the initial instructions, subjects will be told that “When you purchase an item in Store A, you will pay no sales tax in addition to the prices.” In store B, subjects will face the standard sales tax in their city of residence. In the initial instructions, subjects will be told that “The sales tax rate in Store B is the standard sales tax rate that applies in your city of residence.” In store C, subjects will face triple the sales tax rate in their city of residence. In the initial instructions, subjects will be told that “The sales tax rate in Store C is triple the standard sales tax rate that applies in your city of residence.” Subjects will never be told what their actual tax rate is.

After reading the instructions, subjects will have to answer six quiz questions correctly before they are allowed to move on. Subjects will be able to use back buttons to navigate back to the instructions should they need to do so.

Subjects will be given a $16 budget which they will be able to use to buy one of the products at one of the prices encountered in the study. For each-product store pair, subjects will see a multiple price list. The set of displayed prices starts $4 and goes up in multiples of $1.15 to a maximum of $14.07 for stores A and B, and to a maximum of $12.34 in store C (to prevent subjects from unknowingly violating their budget constraint). If a subject does not answer consistently---i.e., indicates that he is not willing to buy at some price p but is willing to buy at some price p'>p---he will be notified of this inconsistency and given the option (but not required to) change his answers. The prices will be displayed in random order every time. Also, store-product pairs will be presented in random order as well.

To make decisions incentive compatible, subjects will have a one-in-four chance of receiving the $16 budget. In this case, one of their decisions will be randomly (and with equal probability) selected to determine their outcomes. Otherwise subjects just receive a "show-up fee" of $2. For the question that is randomly selected to be implemented, if the subject decides not to buy in that decision, then he simply keeps the $16 shopping budget and does not receive the item. If the subject decides to buy, then he pays the price of the item in that question, plus the applicable sales tax, out of his budget. The item is then mailed to him by ClearVoice.

To not artificially draw too much attention to the taxes, subjects will not be reminded about taxes over the course of their shopping decisions. Before beginning decisions for each store-product pair, subjects will see a screen with the following language: “You are now entering Store [ ] to shop for: [product]”. After they click “continue,” they will see the screen with the questions about about prices for the items. Once subjects reach the screen with the multiple price list, they will always have the option to click “back” to remind themselves what type of store they are shopping in.

After subjects complete all of their shopping decisions, they will be asked three quiz questions about the taxes charged in stores A, B, and C. This will be a test of whether subjects remembered and understood the instructions throughout the study. Subjects will also be asked some demographic questions about their household income, education, political party preferences, financial literacy, knowledge of tax rates, and ability to compute taxes.
Randomization Method
The experiment is within-subject. The order of stores-product pairs, as well the presentations of prices for each pair, will be randomly determined using the random number generator in Qualtrics.
Randomization Unit
The experiment is within-subject, and all randomization is conducted independently for each individual.
Was the treatment clustered?
No

Experiment Characteristics

Sample size: planned number of clusters
Approximately 1500 subjects who pass the inclusion criteria described in the pre-analysis plan. All standard errors will be clustered at the subject level.
Sample size: planned number of observations
The total number of subject-store-product-price observations will be 1500*3*(10+10+9)=130500.
Sample size (or number of clusters) by treatment arms
N/A. There is only one arm, as all variation is within-subject.
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
IRB

Institutional Review Boards (IRBs)

IRB Name
Dartmouth-Hitchcock Medical Center COMMITTEE FOR THE PROTECTION OF HUMAN SUBJECTS
IRB Approval Date
2016-08-19
IRB Approval Number
STUDY00029784
Analysis Plan

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

Request Information

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