From Screen to Place: Testing Media-Driven Interest and Local Information Interventions

Participants are randomly assigned to one of three conditions: (1) a control group that receives neither intervention, (2) a video treatment in which participants watch a short video about Deadwood, (3) an information treatment in which participants are given the opportunity to access information about living, working, visiting, and finding housing in Deadwood, and (4) a combined treatment that includes both the video and information interventions. The video is intended to increase participants’ interest in Deadwood, thereby potentially enhancing the effect of the subsequent information treatment. We are interested in the effects of the video alone, the information intervention alone, and the combination of the two interventions.

2026-10-20

2026-10-22

The primary outcomes are three binary measures indicating whether participants are very or extremely likely to (1) visit Deadwood, (2) move to Deadwood, or (3) visit Deadwood conditional on visiting Mt. Rushmore. We estimate the effects of the video treatment, information treatment, and combined treatment separately.

The three primary outcomes capture participants’ interest in engaging with Deadwood in three distinct ways: as a place to live, work, and visit. We use binary measures that identify participants who report being very or extremely likely to engage in each activity, focusing on relatively strong expressions of interest rather than small differences in stated preferences. These outcomes allow us to assess whether the video and information interventions increase interest in Deadwood across different dimensions of economic and tourism engagement. We hypothesize that exposure to the video will increase people's stated likelihood of i) visiting Deadwood, ii) moving to Deadwood, iii) visiting Deadwood conditional on visiting Mt. Rushmore.

We will estimate heterogeneous treatment effects based on a persons i) interest in western cinema, ii) geographic proximity to Deadwood, iii) experience with the HBO Series/movie Deadwood, iv) experience with Deadwood and the Black Hills of South Dakota, and v) degree of surprise about the information they learned about Deadwood, South Dakota. We will additionally estimate the effect of exposure to the video on both the frequency that the informational buttons are used and the length of time a subject spends on the informational page. Finally, we will also estimate the effect of previous exposure to the HBO series Deadwood on i) the likelihood a person clicks on the various information buttons, and ii) the three core outcomes (the stated likelihood of visiting Deadwood, moving to Deadwood, and visiting Deadwood conditional on visiting Mt. Rushmore.

Participants are randomly assigned to one of four conditions: (1) a control group that receives neither intervention; (2) a video treatment in which participants view a short video promoting the show "Deadwood"; (3) an information treatment in which participants are given access to information about living, working, visiting, and finding housing in Deadwood; or (4) a combined treatment that includes both the video and information interventions. The video is designed to increase participants’ interest in Deadwood and attention to the subsequent information, allowing us to test both the independent effects of each intervention and whether the video enhances the effect of the information treatment. Participants’ responses are then used to measure their stated likelihood of living, working, and visiting Deadwood.

Not available

We will randomly (and evenly) assign subjects to treatments using Qualtrics logic.

Individual level

No

NA

2,500. We will sample 1,500 subjects from the states of az, co, id,ia, ks, mn mt, ne,nv, nm, nd, ok, sd, ut, and wy. We will also sample 1,000 subjects from states outside of this list.

625 per treatment arm.

With a sample of 1,250 respondents (625 in each group), we calculated the minimum detectable effect (MDE) for binary outcomes assuming 80% statistical power and a 5% significance level using a two-sided test. The MDE varies with the expected control-group outcome rate. If 10% of control-group respondents are expected to answer yes, we can detect an increase of 5.26 percentage points, from 10% to 15.26%. If the control-group rate is 25%, we can detect an increase of 7.15 percentage points, from 25% to 32.15%. Finally, if the control-group rate is 50%, we can detect an increase of 7.89 percentage points, from 50% to 57.89%. These estimates indicate that the experiment is powered to detect effects of approximately 5–8 percentage points, depending on the underlying control-group outcome rate.