Data from: Mitigating collision-caused bird mortality through message framing: Insights from residents' intentions for bird-safe windows
Data files
Oct 09, 2025 version files 5.89 MB
-
Codebook.csv
5.23 MB
-
Dataset.csv
658.03 KB
-
README.md
2.34 KB
Abstract
Efforts to change human behavior for the benefit of biodiversity often rely on the dissemination of scientific information about biodiversity loss to nonscientific publics. This approach to science communication, known as the knowledge deficit model, is often insufficient for changing behavior. Recent trends reveal a rise in the use of message framing as an alternative method of communication. To address biodiversity loss caused by bird-window collisions, we use experimental survey design to compare the effect of deficit model messaging and four message frames (efficacy, emotional, moral, normative) on residents' intentions to adopt bird-safe windows, while accounting for other potentially influential cognitive and contextual factors. Data from a sample of bird enthusiasts (n = 2854) and the general public (n = 2054) in the United States and Canada indicate efficacy and emotional message frames were the most effective for bird enthusiasts and the general public, respectively. Prior experience with collisions, perceived impact of collision prevention, centrality of birding, educational attainment, and mutualist wildlife value orientations were also positively associated with respondents' intention to adopt bird-safe windows. Normative message frames, age, identification as male, residency in the U.S., and domination wildlife value orientations were negatively associated with adoption intention. Beliefs about collision prevention, centrality of birding, age, education, and wildlife value orientations also had similar associations with respondents' intention to encourage others to make their windows bird-safe. Results provide important insights for the mitigation of collision-caused bird mortality through evidence-based message framing, and the actions people are willing to take on behalf of birds and biodiversity.
Dataset DOI: 10.5061/dryad.qbzkh18wd
Description of the data and file structure
We collected data through an online survey distributed to a sample of bird enthusiasts (n = 2,854) and the general public (n = 2,054) in the U.S. and Canada. Respondents from our bird enthusiast sample were recruited through Merlin Bird ID (hereafter, Merlin) – a freely available bird species identification app developed by the Cornell Lab of Ornithology. We sent email invitations containing a unique survey link to a random selection of Merlin users actively subscribed to receive email communications from the Cornell Lab of Ornithology with U.S. or Canadian-based Internet Protocol (IP) addresses. Respondents from our general public sample were recruited through a survey panel maintained by the sampling firm, Verasight. Verasight recruits panelists via random address-based sampling, random person-to-person text messaging, and dynamic online targeting. Panelists were then selected for participation in our study using quota-based methods to obtain adequate representation within our sample among age, sex, and race characteristics (Table S1 in related publication). The survey was administered via Qualtrics’ online software in April and May 2024. Data collection methods were approved by Cornell University’s Responsible Conduct of Research Office (#IRB0148296) and respondents gave their informed consent before participating.
Files and variables
Details regarding each variable in Dataset.csv (i.e., variable name, description, and coding) can be found in the corresponding Codebook.csv. Respondents were not required to provide an answer to every question in the survey, therefore, the dataset contains missing values. Missing values are represented by "n/a" or not available.
Code/software
Analyses were performed in SPSS (Version 30.0).
Human subjects data
<p>This research with human subjects was approved by Cornell University’s Responsible Conduct of Research Office (#IRB0148296) and respondents gave their informed consent before participating. Personally identifiable information was manually excluded from the present dataset.</p>
