2019/10/31 by Victoria Y. Martin, Victoria Y Martin · 1 voice · 80 citations
Environmental Science · Social Sciences · #Climate Change Communication and Perception #Environmental Education and Sustainability #Environmental ethics #Environmental planning #Environmental resource management #Environmental science #Geography #Natural (archaeology)
paper · doi:10.1093/biosci/biz128
published in BioScience 70(1), 13-16 (Oxford University Press)
openalex publication_date 2019/10/31 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/29
Why do natural scientists continue to conduct and review environmental social science research without training and experience in the social sciences? Perhaps they have come to realize that many of the environmental challenges we face are, fundamentally, human problems. Perhaps they assume that asking people questions is easy, or their well-intentioned efforts are attempts to address the long-standing calls for better integration of the social and natural sciences (Heberlein 1988, Mascia et al. 2003, Metzger and Zare 1999). Whatever the reason, rather than working with social scientists, many natural scientists continue to “[step] over disciplinary boundaries to conduct attitude studies. And, this is a problem.” (p. 583; Heberlein 2012). The problem is that when researchers do not have adequate training, knowledge, and experience, their social scientific studies are often poorly designed, neglect vast bodies of social scientific knowledge, and are full of methodological flaws. Ultimately, these problems lead to misinterpretation of the results and unsubstantiated conclusions. The term social science encompasses a large number of disciplines and subdisciplines, ranging from social psychology to economics and law (Bennett et al. 2017). Natural scientists have been involved—without any training—in topics that extend well beyond the “attitude studies” Heberlein (2012) discussed, into areas such as behavior change, education, and communication. Not all natural scientists do this, of course. Many acknowledge that social research is not their area of expertise, which makes the mindset of those who take a nescient approach to doing social science confounding. If the roles were reversed, natural scientists would be horrified at the thought of someone without adequate skills, knowledge, and supervision coming to their lab or doing whatever field experiments they believed to be a good idea, however they please to conduct them. Unqualified people also create problems in reviewing, approving, funding, and publishing social research, which slows (and in some cases sets back) the advancement of socioecological knowledge. For example, interdisciplinary journals that do not have the capacity to review social research adequately (Teel et al. 2018) continue to publish questionable environmental social studies or reject valuable research when reviewers do not understand social scientific theories and methods well. Publication of substandard research encourages others to conduct and publish research in a similar way, which not only does a disservice to environmental social science but may also be hindering our ability to understand and respond effectively to some of the most serious environmental issues we are now facing (Bonebrake et al. 2018, Stenseke and Larigauderie 2018). It is important to point out that social scientists can also produce weak research, especially when their work suffers from excessive constraints or as newly trained but inexperienced researchers launch into their careers or when researchers from other fields move into social science in postgraduate studies. Although this article focuses on natural scientists (specifically, those without any social science training), clearly we all need to ensure the social research that informs our collective environmental knowledge, policies and decisions is of the highest possible standard. Strong criticisms will sit uncomfortably with some readers without further explanations and examples, so in the present article, I will briefly discuss four common problems found in environmental social research undertaken by natural scientists. The issues are often matters that are fundamental to quality social research, which makes it disturbing to see them persist. My purpose is to open discussion about problems that result when researchers naively delve into fields they are unfamiliar with and to emphasize the necessity of bringing different types of expertise together to deepen our understanding of environmental issues and develop effective solutions. To illustrate specific areas of concern, I draw from many discussions with researchers in social science networks around the world, reviews of natural scientists’ social research, and the lessons I have learned in my work with natural scientists since the mid-1990 s. The examples considered here are topics I am most familiar with (environmental social psychology, human behavior and communication), however similar issues present themselves in other areas beyond my expertise (e.g., environmental economics). One factor driving numerous problems is unfamiliarity with the social scientific literature. Surprisingly, many proposals and manuscripts written by natural scientists provide a scant review of (or completely overlook) the relevant literature. This is particularly concerning when large bodies of work (some dating back more than 100 years—e.g., theories of human behavior) have been ignored. Oversight of the existing knowledge eventually leads to serious issues with the methodology and subsequent problems with analysis and interpretation of results. For example, several manuscripts by natural scientists described their attempt to measure behavior change following an intervention—for example, participation in activities such as litter removal, environmental restoration work, or an educational event. Because the researchers did not build their study on behavior change literature, their research questions were based on mistaken assumptions about which factors were important. In doing so, these studies failed to address any of the well-known drivers of human behavior or ask the participants appropriate questions. In the end, their conclusions were unsubstantiated. Similar examples exist for studies in which researchers assume that improving environmental attitudes or increasing people's knowledge will automatically lead to behavior change. Social scientists have known for many decades that attitudes and knowledge, on their own, do not change behaviors (Nilsson et al. 2019). Common problems also proliferate when researchers are unfamiliar with the application and documentation of social scientific methods. Social science, like any other science, uses well-established methods for the development of study designs, data collection, and analysis (Bryman 2012, Moon and Blackman 2014). Also like other sciences, careful consideration of methods is essential for the validity, reliability, replicability, and generalizability of the study (Walliman 2006). The standards social scientists adhere to depend on the type of research (e.g., quantitative, qualitative, mixed methods) and subdisciplinary norms. In the reviews of natural scientists’ social research, the methods section often receives the most criticism. Consider the example of a quantitative questionnaire, which is one of the most widespread tools used by natural scientists in social research. Rather than developing their questions from previous research, some natural scientists said they asked whatever questions they felt were important. Many of the questions they used were poorly formed, unclear, and untested, and the response options and analysis were given little thought. Consequently, most of these studies have substantial amounts of unpublished data that will never see the light of day. The adage “garbage in, garbage out” rings true. In addition, methodological details are frequently missing in manuscripts, such as descriptions of how the survey questions were developed and tested, details about how, when, and where the fieldwork was conducted, what type of sampling was employed, and the response rate (along with many other reporting requirements, discussed below). When it comes to analyzing social data, some natural scientists are unaware of appropriate methods. Similar to the natural sciences, there are myriad ways to measure and analyze social data, which means there are also numerous, context-specific problems. In the interests of brevity, the examples here will focus on two common problem areas in the analysis of quantitative questionnaire data. Both examples use questions that ask people to respond on a rating scale (e.g., a 5 point scale; 1, strongly agree; 5, strongly disagree), which natural scientists often incorrectly call “Likert scales” (that is a topic for another discussion, along with how many points to use in a rating scale). In the first example, several recent studies used response scales but the authors did not know how to analyze the data. Rather than analyzing the data using statistical methods such as regression, they collapsed the scales into a binary response (agree versus disagree). This makes the use of a response scale redundant, and loses a lot of information in the process. In the second example, many researchers overly relied on previously validated groups of questions (called scales) that were designed to measure particular constructs such as people's level of “environmental concern” (the New Ecological Paradigm is a good example). In a surprising number of cases, no scale reliability tests were provided (a relatively straightforward procedure), because the researchers were unaware of the necessity for it. To make matters worse, most researchers did not pretest the scale prior to the study, which is vital because different populations and settings can produce different results (in some situations, the scale fails altogether). The results are questionable at best, but some of these studies were published nevertheless. In reporting social research results, natural scientists frequently leave out important information. Although this discussion could include many examples, reviewers often comment on two problems. The first is the omission of basic demographic and other background characteristics of the respondents. This essential information tells us who the study includes, which speaks to the reliability, replicability, and generalizability of the findings. Second, the means and standard deviations (or similar) for questions using response scales are often not reported. These statistics tell the reader how people responded to each question and how much variation there was in the responses. For readers interested in learning more, the APA Style Journal Article Reporting Standards (JARS; www.apastyle.org/jars) is a helpful guide. The JARS website and documents provides step-by-step instructions for different research methods and how to report them and are also useful references for the design phase of research projects. Misunderstandings and mistakes made early in the research process, such as those described above, usually amplify problems in the discussion and the conclusions. When the problems are so extensive, social scientists wonder why researchers who are unaware and untrained in the social sciences continue to insist on conducting social research and why this dilemma has persisted for decades. In the end, these researchers waste a great deal of everyone's precious time, effort, and resources. With fundamental problems being so common, what are the solutions? Collaborations between natural and social scientists who have the training, experience, and skills to conduct robust and reliable research are essential (Alexander et al. 2018, Moon et al. 2019, Schultz 2011). There are many excellent examples of genuine interdisciplinary collaborations that have produced positive outcomes for the natural environment. For example, social and natural scientists have worked together to protect critically endangered species (Struebig et al. 2018), better understand the extent of non-compliance in recreational fishing (Thomas et al. 2015), facilitate indigenous knowledge into environmental management and monitoring (Eckert et al. 2018, Thompson et al. 2019), and change farming practices to improve environmental health (Pickering et al. 2018). Numerous networks of social scientists welcome the opportunity to work on environmental issues. These include the Network of Environmental Social Scientists (www.nessaustralia.org), the Conservation Social Science network (#ConsSocSci on Twitter), the Social Science Working Group of the Society for Conservation Biology (https://conbio.org/groups/working-groups/social-science), the Marine Social Science network (www.marsocsci.net), many “human dimensions” or social science subcommittees in other societies, and more. I encourage natural scientists to stop going it alone in social research and reach out to these networks. Involving an experienced social scientist from the outset of your research will help ensure the work is valuable and publishable. Over the last few decades, an increasing number of environmental research centers and organizations have integrated social scientists. Interdisciplinary teams who have the training, skills, and knowledge to conduct quality research are on the rise. Although we still have a long way to go, we owe it to the planet, ourselves, and future generations to continue this progress, recognizing and valuing the contribution each discipline can make to reliable and robust research of the highest possible standard. I am deeply grateful to Rick Bonney and Scott Collins for their feedback, encouragement and support in the development of this article. Many other environmental social scientists shaped my thinking about these issues, especially Edd Hind-Ozan and Angela Dean. I would like to thank the many wonderful natural scientists I have collaborated with. They have taught me about the environmental challenges we face, and the value of bringing the natural and social sciences together. I also appreciate support from the Cornell Lab of Ornithology and the Rose Postdoctoral Program. Victoria Y. Martin is affiliated with the Cornell Lab of Ornithology, at Cornell University, in Ithaca, New York.