Personal Reflections on Conducting Societally Impactful Research

The article by Ozanne et al. (2024) is a must-read for anyone aspiring to conduct societally impactful research. The careful unpacking of the different dimensions of societal impact provides welcome conceptual clarification, and the Delphi study offers insightful and practical recommendations on how to advance toward this important goal. I found myself in broad agreement with their diagnosis and most of their recommendations.

In this commentary, I will offer my perspective on two key antecedents of societally impactful research highlighted in their article: the need to engage with stakeholders and the importance of a continuous process. In both cases, I will use as illustration two recent articles of mine that have been recognized for their impact (Chandon and Cornil 2022; Dubois et al. 2021), ¹ but only after generating unusually stressful media attention, going through an unusually difficult publication process, or both. These are my personal opinions and do not necessarily reflect those of my coauthors.

Engage with Stakeholders, but Beware of the Optics

As Ozanne et al. (2024) persuasively show, impactful societal research starts when the stakeholders are at the front and center of the research process. Even if researchers have identified an important societal issue and an effective way to study it, they will rarely succeed at getting companies, governments, or any other stakeholder to run a large, impactful study; the impetus must come from them. For example, I do not think any researcher would have dreamed of designing and conducting a randomized controlled trial (RCT) requiring manually adding 1.9 million nutrition labels on more than 1,200 products over ten weeks in 60 supermarkets. That study (Dubois et al. 2021) only materialized because the French health minister decided that such a study was necessary to select which of four competing nutrition labeling systems would get government endorsement.

Engaging with stakeholders must be done carefully because it can create conflicts of interest. Even just the appearance of a conflict of interest can jeopardize the societal impact of research, especially if it involves policy decisions like nutrition labeling mandates. Conflict-of-interest statements only partially alleviate this concern because they lack details and can be misinterpreted. For example, during the nutrition label RCT, a leading news outlet relied on the conflict-of-interest statements to highlight “links” between food companies and the researchers, sowing doubts about our integrity (Grandin et al. 2016). In one case, the suspicious “link” involved a researcher who had helped an alcohol company study ways to reduce binge drinking. In another case, the “link” was agreeing to give a pro bono talk at a conference, which was construed as a debt to the conference organizers, although independent researchers would see the debt, if any, as going in the opposite direction (Guy-Grand 2016).

Unfortunately, eroding levels of trust in science and scientists, particularly following the COVID-19 pandemic (Eichengreen, Aksoy, and Saka 2021), have increased the likelihood that researchers working on societal impact will face stressful accusations and innuendos. Care is therefore needed when choosing how to engage with stakeholders and with whom to engage.

Commit to a Topic and Context over the Long Term

The importance of a continuous process is another recommendation of Ozanne et al. (2024) that particularly resonates with me. I wholeheartedly agree that achieving societal impact cannot be measured one article at a time. In fact, I believe that impact requires committing to an issue (and hence industry or community) over the long term.

Among its many benefits, such a commitment provides opportunities for corrections and extensions and allows nonobvious results to emerge. For example, it was only through the follow-up work (Chandon and Cornil 2022) of an earlier article (Cornil and Chandon 2016) that we found that epicurean labels describing the aesthetic, multisensory properties of food better promote moderation in eating among French than among American consumers.

A long-term focused commitment is also essential to instill in researchers the confidence necessary to ask stakeholders to act. It was only after obtaining evidence from a field experiment involving paying restaurant customers (Chandon and Cornil 2022) that we felt confident enough to tell food marketers that pleasure-based interventions can lead consumers to pay more for less food.

Focusing on a specific issue and community is also essential to improve the translation of research into actions, the weakest stage identified by the researchers surveyed in Ozanne et al. (2024). Focus and time are necessary to acquire the institutional, legal, and tacit knowledge that builds the credibility and social capital needed before stakeholders act on research findings. For example, it was only after a decade of active participation in research, industry, and regulatory conferences that the French Health Ministry invited me to join the multidisciplinary team tasked with organizing the nutrition label RCT.

Expect Misalignments Between Academic and Societal Impact

Academic and societal impacts do not always coincide. If societally impactful research requires a long-term focus on a specific topic or industry, as I argue, its output may be seen by editors as incremental (Koole and Lakens 2012). With a few exceptions, such as the International Journal of Research in Marketing (Lynch et al. 2015) or Marketing Letters (Labroo, Mizik, and Winer 2022), which welcome replication studies, the most prestigious academic journals focus on novelty and rarely publish the kind of follow-up studies that societally impactful research often requires. This is not confined to our field. In medicine, for example, the current publication system rewards novel discoveries over the kind of translational research that directly improves patient outcomes (e.g., Kools et al. 2018).

The publication process of the two articles mentioned previously illustrates this misalignment. Dubois et al. (2021) was rejected by four journals, mostly for a perceived lack of contribution relative to studies of calorie labeling in restaurants. Yet, no government or food company would have mandated or adopted the Nutri-Score label in grocery stores based on studies examining the effects of displaying a different type of nutrition information (calorie content) in a different purchase context (restaurants). Chandon and Cornil (2022) was also rejected by four journals for insufficient contribution over our previous article (Cornil and Chandon 2016) despite examining a different intervention (menu labeling vs. multisensory imagery instructions) in a different context (a field experiment and cross-cultural analyses with representative French and American consumers vs. online or lab studies with specific samples).

A delayed publication in a less prestigious journal may affect promotion and tenure decisions. Ironically, it will also reduce the number of citations, which remains the traditional measure of impact, even though it is not necessarily correlated with actual societal impact (Davison and Bjørn-Andersen 2019). For example, Kapferer and Laurent’s (1993) follow-up article on consumer involvement profiles only received one-ninth of the citations of their first article (Laurent and Kapferer 1985) despite being based on stronger evidence (a diversified sample of 3,838 respondents vs. a convenience sample of 207 women) and updating the conclusions of the first article.

Forewarned Is Forearmed

No scholar targets low-impact research. Researchers aiming for societal impact, as opposed to other types of impact, will benefit greatly from reading Ozanne et al. (2024). Hopefully, being forewarned of the trade-offs and potential pitfalls of conducting societally impactful research should make researchers, editors, and stakeholders better prepared when engaging or evaluating this kind of research so we can all progress toward this important goal.