Discrimination and Health Care: The Case of Dentistry Appointments

From the pool of dental providers that utilize a standard appointment request form, we draw a simple random sample across the United States to obtain approximately 4,000 providers across the country to request a first-time appointment for a fictitious patient. One potential concern is that providers, particularly ones in less populated locations, where there will be fewer new potential patients, may get suspicious if there is a significant increase in requests within a short span of time. To minimize this concern, we request first-time appointments once every two months over the period of a year, giving us 24,000 observations (4,000*6).

2025-03-01

2026-04-01

Response within 30 days (yes or no); timing of response

Our primary outcome will be whether or not a dental office responds (via e-mail and/or phone) to an applicant’s request for an initial appointment as a new patient. From a small pilot, we know almost all responses come within one week, so allowing for 30 days should be sufficient. We will also test to see if there are differences in whether the provider indicates that they are taking on new patients and how quickly a provider responds (measured in business days).

Heterogeneous effects by county characteristics (rural vs urban, racial diversity, political composition), dental practice characteristics (size of practice, racial and gender composition of practice), and difficulty of name pronunciation

We will test to see whether the degree of discrimination varies along a number of dimensions. Racial/ethnic and name-based discrimination may be more likely to occur in locations with fewer non-whites and in areas of the country and in dental practices that are less diverse.

We will be using both email addresses and phone numbers to collect data.

Instruments will be created for each persona that is developed so that we minimize the risk of repeating email addresses and phone numbers submitted to the dental practices. All emails will be forwarded to a central email account for ease of data collection. For phones, a separate Google phone number will be set up for each wave of applications and messages will be monitored and recorded by research assistants. Both instruments (phone and email) are standard methods that dental providers use to respond to appointment inquiries and they will allow us to track response dates and times along with the content of the response.

Undergraduate research assistants will be responsible for gathering information on each of the 4,000 dental practices (location, size, and dentist characteristics) and confirm that their online forms meet our criteria for the study. For the dental practice to meet our criteria for the study, the online form must:
● Have name box(es)
● Include a general comment box
● NOT have fields requesting patient address, social security number, or other information

During the experiment, the research assistants will be responsible for filling in the online forms with the appropriate information provided to them. If a response was received within 1 month of the initial inquiry then they will record the content of the response and the length of time that the response took (in business days).

To limit the number of hypotheses tested, we will restrict our sample to male-sounding names. For each request, we randomly select the fictitious person’s race, which is signaled by the name. For each fictitious person, we will randomly draw a name from a set of at least 6 race-signaling names. If the name is a repeat for the dental practice, then we will randomize until we draw a new name for that provider. Utilizing the comment box in the standard appointment request form, we randomize the type of insurance of the fictitious individual along with the grammatical correctness of the request. Together, this approach allows us to measure discrimination across the following dimensions:
1. Implied race (White, Black, Hispanic)
2. Insurance status (Private, Medicaid, No Insurance)
3. Implied socioeconomic status (High, Low)

Since we will be sending out 6 requests, one every two months, to each dental practice in our sample; data collection will take approximately one year. Initially, research assistants will collect background information on each dental practice that is selected for the sample.
Research assistants will fill in the online fields with the corresponding information from the project spreadsheet. Research assistants will then record the date and time in the data collection spreadsheet that the contact was made.

Randomization done in the office by a computer

Individual

Yes

4,000

24,000

1,333