Citizen Valuation of Police Services

See PAP

We use an online instrument to survey US residents who rent their homes. This instrument collects demographic information about each respondent, including age, race, gender, education, income, and number of children. Respondents who respond incorrectly to attention questions are not allowed to complete the survey.

The survey asks respondents to consider a hypothetical scenario where they must move to a new rental home. We ask respondents what number of bedrooms and monthly rent they would be interested in. We then show each respondent a choice set with 5 hypothetical homes with randomly generated characteristics and ask them to select their most preferred option. Respondents cannot move on to the next question until 10 seconds have elapsed. This question is then repeated 13 more times with new choice sets. Of the 14 decision rounds, the first two and the eight rounds serve as attention checks. Participants who don't pass the first two rounds (those that do not pick dominating options) will be exited out of the survey.

Half of respondents are randomly assigned to an information treatment about body cameras. For these respondents, the following paragraph is added to the description of police force and its reminder text:

"Assume that police officers wear body cameras as part of their uniforms to document what they see as they perform their duties. Department regulations require that use of force incidents are recorded, and citizens can request access to body camera footage involving themselves.

However, officers may fail to activate body cameras. A research study finds that police only activated these cameras around 24% of the time they were supposed to, and some officers disproportionally under-activated body cameras."

The control group only receives the first paragraph.

2022-08-20

2022-10-20

Outcomes of interest are willingness to pay for reductions in use of force and crime. We are also interested in whether information provision regarding body cameras affect these willingness to pay measures.

Willingness to pay for reductions in use of force and crime will be calculated by comparing the effect of these factors on home selections to the effect of price. The willingness to pay for a reduction of e.g. one use of force per 10,000 people is the difference in rent required to make respondents indifferent between a home with some baseline use of force and a home with one fewer uses of force.

We are interested in how the willingness to pay measures depend on various crime conditions (high and low crime) and demographic characteristics (e.g., minority status, education levels, and income levels)

See PAP

The hypothetical homes are given the following characteristics: monthly rent, violent crime per 10,000 people, police use of force per 10,000 people, square footage, and quartile of school quality. Respondents are asked to assume that all other characteristics are identical across the choices. Square footage and monthly rent are selected within a range that accords with the respondent's choice of bedrooms and rent.

In each of the 14 choice sets, only monthly rent and one other randomly selected characteristic vary. For example, the first choice set will present 5 homes with different rents and different square footage but other identical features identical. Then the second choice set presents 5 homes with different rents and different violent crime rates but other identical features.

In order to ensure data quality, we will use the first two choice sets to serve as data quality checks. In each of these two choice sets, there is a dominating option -- an apartment that is cheapest among the five options and also has the lowest crime rates. We will remove participants that do not pass either of these attention checks.

After filtering out participants, we will have participants consider police use of force as another feature of their decision-making. To treatment participants, we will say that police officers may fail to activate body cameras and give a low activation rate.

Prior to viewing the choice sets, respondents are shown detailed descriptions of our violent crime, police force, and school quality measures. They may also access reminder text while in the choice sets.

The 8th choice set will again serve as an attention check, where one option dominates the others.

We will conduct a follow-up survey with approximately 250 participants, randomly assigning half to receive information about quality of life measures. We will collect willingness to pay (WTP) estimates and the marginal rate of substitution between use of force and crime, analyzed separately based on whether participants received quality of life information. The analysis will compare statistical differences between the two groups. Additionally, we will ask participants how police use of force influenced their decision-making processes.

Randomization of the neighborhood features to vary and treatment into information provision will be done through Qualtrics' random number generator.

Randomization unit for neighborhood features are at the individual-by-choice level. Information provision will be given to half of study participants.

Yes

We will recruit 1000 individuals.

We will recruit 1000 individuals. Each participant will be asked to complete 11 rounds of choices (this excludes 3 rounds that will be used for data checks), which gives us 11,000 observations.

Half of 1000 individuals will be treated into information provision regarding body cameras.