[Proposal] Risk communication behavior about personal and impersonal risk

Risk Communication Behaviors about Personal and Impersonal Risk


Social scientists have been trying for decades to understand why people engage or disengage with risk issues (Hmielowski et al., 2019). Previous research indicates that distinguishing between personal and impersonal risks is crucial to understanding people’s risk attitude (Sjöberg, 2003) because the belief that one is personally at risk is the initial step toward risk reduction (Morton & Duck, 2001). In the case of environmental risks, people tend to view risk reduction as the government’s responsibility. Compared to other more urgent concerns such as health and financial risks, most people view environmental risks as distant and low-relevance issues. Kahlor et al. (2006) coined this type of risk as “impersonal risk” – that is, risks that are not viewed as posing a direct and immediate threat to individuals. As a result, scholars have emphasized the importance of establishing personal relevance to motivate people to act in environmentally friendly ways (e.g., Kahlor et al., 2006; Rogers et al., 2017). For example, even though climate change is an impersonal risk, people are more likely to act if they perceive high issue relevance in this topic. Therefore, it is important for communication scholars to identify socio-psychological factors that influence risk communication behaviors about risks with different personal relevance to the public.

Information use is often at the core of understanding individuals’ reactions to risks in their environment (Griffin et al., 1999). In this research, I plan to examine how individuals’ risk communication behaviors such as information seeking and information processing differ based on whether the risks are perceived as personal vs. impersonal . The key assumption is that personal relevance serves as a key factor that defines whether people perceive a risk as personal or impersonal. The research context is an emerging environmental hazard – Per- and polyfluoroalkyl substances (PFAS). PFAS are a family of man-made chemicals known as “forever chemicals” because they are almost impossible to break down once released into the environment and human body, which can cause negative health outcomes including cancer (Evans et al., 2020). Although a novel risk to most people, PFAS is ubiquitous in our everyday life – found in food packaging, nonstick cookware, water-repellent fabrics, and drinking water (Felton, 2020). PFAS contamination widely exists in the United States, ranging from both rural areas and metropolitan areas including New York City, Miami, and Philadelphia. It has been found that more than 200 million Americans could have a toxic level of PFAS in their drinking water (Amarelo, 2020).

I will apply the risk information seeking and processing (RISP) model (Griffin et al., 1999; 2013) to examine how personal relevance affects people’s information processing and information seeking behaviors. As one of the most integrative models in risk communication research (McComas, 2006), the RISP model has inspired over a hundred empirical studies. However, to date, little research has evaluated whether the RISP model functions differently when a risk is portrayed as bearing high or low relevance to the participants (see for exceptions, Yang, 2019; Yang & Zhuang, 2020). In terms of practical contribution, we believe the findings of this research can help risk communication practitioners design tailored messaging regarding PFAS contamination. Broadly speaking, beyond PFAS contamination, the findings of this research may also be extended to other environmental issues such as climate change and energy conservation.

A one-way between-subjects experiment will be conducted to collect data. Before the experimental manipulation, all participants will first answer questions assessing their general awareness of and past experience with PFAS. Then, they will be randomly assigned to either the high personal relevance condition or the low personal relevance condition. Participants in either condition will be asked for their zip code and then read a blurb. In the high relevance condition, participants will read a blurb detailing the presence of PFAS contamination in the drinking water of their residential area based on their zip code and the use of PFAS in consumer products. In the low relevance condition, participants will read a blurb presenting the fact that the drinking water in their residential area is not detected with PFAS contamination and the use of PFAS in industrial products. After the experimental manipulation, participants will first answer questions measuring information processing related to the blurb, and then answer questions measuring information insufficiency, perceived hazard characteristics, affective response, information subjective norms, information seeking, and information processing in general. Control variables such as individual characteristics and media attention will be assessed at the end of the survey. After answering all questions, participants will read a debriefing page.

As a pilot study, I plan to recruit participants from entry-level courses open to all majors to test the experimental stimuli and conduct manipulation check. Once the experimental stimuli are finalized, I will use Prolific for the main study.

Using a ratio of 1:10 (parameters to cases) recommended by Kline (2015), the minimum sample size needed to conduct structural equation modeling analysis is about 1020. To account for incompletion and failed attention check, we will over sample by approximately 25%. As a result, the sample size needed for this study is estimated at 1300.

The estimated median completion time is 15 minutes. According to Prolific’s recommended rate of pay, we will pay each participant $2.4 for their contribution. In addition, the service fee is $1040 for this study. Therefore, I am applying for $4160.

The study is preregistered at https://aspredicted.org/qh8qj.pdf.

All materials (including data, stimuli, and questionnaire) will be available from the leading author upon request. We aim to publish in open-access journals or a pre-print version of the manuscript will be available online.