The number of publications on nudge research is presented in
Figure 2, which depicts the development from year 2012 to 2021. The overall growth trend is supported by an increasing number of published articles. In terms of the average annual number of publications, the observed trend in research can be divided into three phases. (1) Before 2015, it was a slow growth period for nudge research, with the number of publications remaining below 100 per year. The average number of publications was about 76 per year. (2) In 2016–2017, the number of publications experienced a slight decline. To explore the reasons for the decrease in publication, we compared the publication categories between the two years. The comparison found that research in economics, law, and ethics were the categories that attracted the most attention, with 20 more publications in these three categories in 2016 than in 2017, while publications in 2017 were more dispersed across disciplines. (3) Research in nudge has grown rapidly since 2018, with a slight negative growth in publications in 2020 likely due to COVID-19, with the number of publications peaking in 2021 (n = 315, 18.46%), and a further increase in research trends likely to be seen in 2022.
Data for the research categories were generated from the search results of the WoS database.
Table 2 shows the top 20 research categories of nudge publications. Nudge originated from behavioral economics, and Thaler, the founder of behavioral economics, pioneered the introduction of psychology into economic research, focusing on human behavior, especially human economic behavior. It was later applied to various categories as a tool to intervene in human decision making. As such, the statistics show that nudge research consists of a wide variety of disciplines, with “Economics” (n = 226, 13.25%) remaining the most studied discipline in nudge. “Public Environmental Occupational Health” (n = 116, 6.8%) and “Ethics” (n = 102, 5.98%) were also important research categories for nudge. The categories “Psychology Multidisciplinary”, “Political Science”, “Public Administration”, and “Law”, as the first areas of applied research in nudge, continue to receive attention. In addition, nudge-related research covers such categories as “Social Science Biomedicine”, “Business”, “Environmental Science”, “Nutritional Dietetics”, and other categories, aiming to focus on better serving people in their daily lives and guiding them to choose healthier and more sustainable decisions. Notably, an emerging trend in applying nudge in computer-related fields such as “Computer Science Information Systems” and communication is witnessed, especially after the introduction of digital nudging by Weinmann et al. [
47].
3.1.1. Journal Distribution
In general, Nudge Theory can be applied in research pertaining to decision making. We found that articles on nudge-related research are published in a wide range of journals, indicating a significant development in the field. The 1706 articles screened and selected by the researchers were published in a total of 1006 different journals. The top 20 most productive journals are summarized in
Table 3 and their number of citations and the impact factors of the journals are reported. It should be noted that we found two journals (i.e.,
Journal of Chemical Physics and
Journal of Chemical Theory and Computation) from the data obtained from the WoS database, whose keywords are “nudged elastic band”, which is a technique for finding transition paths between a given initial state and final state in chemical research. It does not match the original meaning of “nudge” and is hence excluded. Two books and conference proceedings were also excluded, and the final top 20 journals were ranked according to their number of publications. In the case of a tie, the impact factor of the journal was taken into account.
When journals are ranked based on the number of published articles, their citation counts do not correspond to their rankings. Despite publishing a small number of papers, several journals have relatively high citation counts. To explore the reasons for this, we created a treemap of the average citations (AC) of the top 20 journals (as shown in
Figure 3), which is an efficient way to visualize the data in a space-saving manner [
48]. The data in
Table 3 show that six of the most productive journals are related to medicine and health, while three of the top five journals with the highest AC in
Figure 3 are also related to health behavior and health consumption. Although nudge originated from behavioral economics, economics journals are not the most productive. This is presumably because in medicine and health, which have more to do with individual behavior than economics and management, the design and interventions of nudge are more operational. Researchers hope to improve personal health or promote public health by slightly intervening in people’s behavior, for example, by encouraging people to eat more healthily or to have regular health checkups.
Specifically, the most published journal,
American Journal of Bioethics, with a total of 33 nudge-related studies and 483 citations, is with a modest citation count per paper (n = 13.7). The main reason is that the paper by Blumenthal-Barby and Burroughs [
49] has been cited 194 times in WoS. The journal with the highest total and average number of citations was
BMC Public Health, with studies such as Hollands et al. [
50] and Arno and Thomas [
51] contributing the major citations.
Nature, one of the most prestigious scientific journals in the world, published nine papers with only 122 citations, five of which were related to vaccination against COVID-19. While
Review of Philosophy and Psychology published only 11 papers that focused on nudge in political and legal applications, the number of citations was 295, ranking third in average citations (n = 26.8). In addition,
Food Quality and Preference, a journal related to sensory science and food research that primarily focuses on the use of nudge to promote healthier food choices or healthier eating habits among consumers, also contributed the third highest number of citations (n = 324) with an average of 24.9 citations.
We used VOSviewer to analyze 1706 articles for co-citation and formed a journal co-citation network with four clusters containing 332 journals, each with a minimum number of citations of 30. As can be seen in
Figure 4, each node represents a journal, with its size indicating the number of papers published, and the lines between the nodes indicate the intensity of co-citation, with thicker lines representing higher intensity.
The most visible cluster in the co-citation network is red, with 94 nodes. Among them,
American Economic Review, with leading impact factors in its field, stands as the central position of the red cluster with the highest intensity of co-citations. The blue cluster has 77 nodes, and the most prominent one is
Science, which has co-citation links to the other three clusters, although it has published only five nudge-related papers.
Psychological Science and
Journal of Personality and Social Psychology, also in the blue cluster, are also important, as they both belong to the Association for Psychological Science (APS) and focus on the frontiers and applications of psychology, and their co-citation relationships are relatively strong. The yellow cluster contains 76 nodes and the journals in this cluster are mostly related to health and diet. The highest co-citation is Appetile, which focuses on the behavior of humans and nonhuman animals toward food and has made important contributions to the study of using nudge to guide consumers to make healthy dietary choices. The most distinguished one out of the 85 nodes within the green clusters is the book
Nudge, nudge, think, think, by John et al. [
52] based on
Nudge [
2]. While acknowledging the power of nudge, they argue that a particular democratic institutional framework is needed to provide an environment that evokes public thinking that promotes listening and reasoned argument among citizens, as well as the type of reflection that can lead to shifts in preferences. Other journals in the green cluster focus on ethics, law, and behavioral studies, all of which have played important supporting roles in nudge research.
3.1.2. Country and Institution Distribution
In bibliometric analysis, country and institution, as important analytical variables, can reflect the research intensity and contribution of different regions or institutions in the research field. By analyzing the citation and co-citation of publications from different countries or institutions, we can gauge their academic level and collaborative networks [
53].
From the data obtained from the WoS database, we found that the 1706 publications were distributed among 79 countries, and
Table 4 shows the top 10 countries with the highest number of publications, which accounted for 90.7% of the total number of nudge publications (n = 1548). The United States ranked first with 631 publications, accounting for 37.0% of all publications, far ahead of other countries. England stands as the runner-up (260/1706, 15.2%), followed by Germany (159/1706, 9.3%).
Next, we analyzed the most influential countries for nudge research through bibliographic coupling links. The logic behind bibliographic coupling is that two texts with a high number of shared literature references will be similar in content [
54]. This means that the coupling analysis shows the number of identical references cited by the documents as a measure of the collaboration of the country to which the publication belongs. We selected the analysis type “bibliographic coupling” in VOSviewer, and the unit of analysis was “countries”. Additionally, we selected the minimum number of documents for countries to be 1 in VOSviewer to obtain the maximum number of links generated between countries. The studies of Ukraine and Iraq among the 79 countries in the network were not interlinked with other countries, so the maximum number of linked items in the final generated coupling network was 77 countries.
The country analysis by bibliographic coupling is presented in a network visualization of the five main clusters, as shown in
Figure 5. The most striking one of the clusters is undoubtedly the blue cluster, represented by the United States. It is the first one to apply nudge in practice, not only because the authors of
Nudge, Thaler and Sunstein, were an American economist and an American legist, but also for the size of the country, the number of research scholars, and the investment in scientific research. During his term in office, former U.S. President Barack Obama signed an executive order establishing the Social and Behavioral Sciences Team, who translated Nudge Theory into improvements in federal policies and programs with success [
55]. Canada, South Africa, Thailand, and the Philippines are also in this cluster, indicating that these countries cite similar research articles in nudge research.
The green cluster contains 23 countries, most notably England, which was also the earliest country to start research and application of nudge. The rest of the cluster is dominated by European countries (e.g., France, Switzerland, Belgium, Portugal, etc.), Asian countries (e.g., Vietnam, Korea), and Oceanic countries (e.g., New Zealand). These countries have cited similar articles in their nudge studies. Twenty-eight countries in the red cluster are spread out, ranging from European countries such as Germany, Italy, Sweden, and Norway, to Asian countries such as China, India, and Japan, and American countries such as Brazil and the Dominican Republic. In addition to being geographically distant, these countries are cited in similar articles. Regarding the relatively sparser clusters, Australia and Denmark are most visible in the purple and yellow clusters, respectively.
To gain insight into which countries have recently embarked on the nudge study actively, we created an overlay visualization of the country analysis. The score values are color mapped by taking the average year of country studies by default, as shown in
Figure 6. We found that the average year of active research on nudge across countries began in 2017, which indicates that within the decade when nudge was first presented, its effectiveness and academic potential for behavioral interventions was far underestimated. Since 2017, the United States, U.K., France, Denmark, and Canada have taken the lead in starting or expanding participation in nudge study, followed by countries such as Italy, Germany, The Netherlands, China, and Japan. In the past two years, more and more develo** countries have also invested in nudge research (e.g., Indonesia, Thailand, Philippines, Nigeria, etc.). We believe that this phenomenon is inseparable from the level of economic development and academic research in a country. Nudge was proposed to make better decisions for people’s health, wealth, and happiness, and countries with high levels of development focused on people’s well-being earlier and started research on nudge earlier. Therefore, based on the analysis and understanding of the status, we can expect and predict a vigorous development of research and applications in related fields.
3.1.3. Author Distribution
According to Thaler [
56], the emerging and interdisciplinary field of behavioral economics allows scholars to understand human behavior from a more humanitarian perspective. By the same token, nudge-oriented research has become attractive to a growing number of researchers. Consequently, to obtain a synopsis of nudge-related studies, one of the most pivotal tasks is to identify the most productive and influential authors in the field. We conducted an author citation analysis to identify the top 10 most productive authors and rank them by document and citation.
Table 5 shows the results of this analysis. Noticeably, that the number of publications is a metric that should be analyzed with discretion, taking into account factors including the length of each paper, the quality of the journal, and the number of authors per work [
57]. The table is sorted by the number of articles by author, and in case of ties, the citations per author were considered. In addition, the h-index, a composite indicator combining productivity and impact, was appended to the table.
Among the top 10 authors, Cass R. Sunstein is the most productive, with 20 publications, and he also ranks first in citations (n = 1057). Given his groundbreaking contribution in relevant fields (e.g., co-authored the book Nudge, founded the Behavioral Economics and Public Policy Program at Harvard Law School), his leading position in both publication and citation numbers is understandable and well-expected. He has worked closely with the U.S. Behavioral Insights Team since its inception. In 2020, the World Health Organization appointed him as chair of its technical advisory group on Behavioral Insights and Sciences for Health. His work and research laid the foundations of behavioral economics and provided the shoulders of giants for subsequent scholars. The runner-up in the author list is Peter John, a professor in the Department of Political Economy at King’s College London, with a total of 15 nudge research publications. He is adept at using randomized controlled experiments to explore how nudge can be applied in public policy, and how best to engage citizens interested in public policy and management, and in turn deploy behavioral interventions.
Notably, four of the top ten productive scholars have medical backgrounds. Mitesh S. Patel, Anne Thorndike, Douglas Levy, and Joline Beulens all have a medical and health perspective on interventions that use behavioral economics strategies to improve individuals’ dietary intake and health behaviors. Of those, Mitesh S. Patel has 14 publications, and his research focuses on integrating nudge with scalable technology platforms to improve health and healthcare. He has collaborated with health systems, insurers, employers, and community organizations to conduct clinical trials using nudge, such as digital health interventions using wearables and smartphones, and health system interventions using electronic health records, advancing the research and application of nudge. The most cited of this cohort of researchers is Anne Thorndike, with 10 publications and 465 citations. Her research concentrates on the use of nudge and choice architecture, such as traffic light labels [
58], to guide people to healthier food choices and maintain healthy lifestyles.
To further explore the authors’ collaborative research, we conducted a co-authorship analysis, which is a tool used to identify key organizations and scientists and examine their associations [
59]. VOSviewer identified 4557 authors based on the WoS data, and the calculation generated 600 items when the minimum co-authorship of articles was set to two. The largest set of connected items contains 32 items, and
Figure 7 shows the full co-authorship visualization. Interestingly, while Cass R. Sunstein was the most influential author, the highest scoring co-authorship was Denise de Ridder. She has 21 links, which implies that she has collaborated with 21 authors. Denise de Ridder is a professor of psychology at the Department of Social, Health and Organizational Psychology at Utrecht University and project leader of various research projects in the field of self-regulation and facilitation. She shows that she is proficient in collaborating with colleagues whose nudge research focuses on exploring self-awareness in nudge implementation and how nudge can help people to make healthier food choices. Besides her, the previously discussed authors are also relatively conspicuous in the co-authorship analysis visualization, with an average number of links of 10, which is inseparable from their overall number of publications. In conclusion, no scholars are in a dominant position in nudge research, as the field is still in its infancy and researchers can enhance their collaboration to explore more the potential of nudge in various fields of research. This might be an advantage for potential researchers and ongoing studies, as journal editors prefer a small group of highly productive researchers when deciding which articles to publish [
60].
3.1.4. Keywords Co-Occurrence Analysis
The last research questions for this study to explore concerned existing or future relationships between themes in the nudge studies by focusing on the content of the publication itself. For this purpose, keyword co-occurrence analysis was employed, which is a technique for examining the content of the publication by extracting keywords from the full text of the publications. Applied longitudinally, keyword co-occurrence analysis can be used to predict future research in the field with a view to enriching the study’s interpretation of co-citation analysis (in the past) or bibliographic coupling (in the present) and predicting the development of the field (in the future) [
35]. In addition, to obtain more accurate results, less relevant keywords were manually removed and a minimum number of occurrences of keywords of six was set as a threshold level. The network visualization was constructed based on the co-occurrence frequency of 294 keywords out of a total of 5444 retrieved keywords, as shown in
Figure 8.
In network visualization, each keyword is represented by a node, and the size of each node represents the number of publications in which that keyword appears. The clusters of the nodes are reflected by corresponding colors, with the distance between various clusters indicating the relatedness between them. Specifically, a close relatedness between two clusters could be identified if the distances between them is shorter, and vice versa. There are four main clusters in the network visualization shown in
Figure 8, which are red, green, blue, and yellow.
First, red clustering includes topics related to behavioral economic theory, public policy making, and ethical discussions. As an innovative approach to address policy issues, nudge is becoming increasingly popular in the field of public administration. In a recent study, John et al. [
61] designed randomized trials of support for nudge and deliberate nudge in response to top-down regulation and freedom of choice. The results of the experiment showed that public support for both nudge policy options is higher compared to top-down regulation. They also found that support for nudge and deliberate nudge is more correlated with perceived fairness than with perceived efficacy. Similarly, a study showed that nudge interventions positively moderated the impact of two-way risk communication on public value consensus [
62], which suggested that nudge can play a better role in public management than injunctive interventions. From another perspective, some studies have focused on whether there are ethical issues with nudge, such as doubting whether nudge may have the undesirable consequences of manipulating choice, reducing autonomy, and unintended behavior [
63,
64]. Conversely, others have assessed and argued that nudge does not usually interact with people’s rationality in a problematic way [
65], and that ethicists should remain open to its application [
66].
The green cluster focuses on online information, social media, digital nudging, and their impacts. Compared with traditional offline contexts, decision making in digital contexts is more dependent on human–computer interaction interfaces, so the interface design of human–computer interaction can have a significant impact on the decision-making process. This influence includes two main aspects: the interface provides the necessary elements for decision makers to access relevant information, and the way the interface provides this information affects the cognitive process, producing different decision outcomes than in a no-choice architecture context [
47]. Thus, social media and other online applications can exploit digital nudging to play a leading role, such as against the sharing of fake news [
67] and the protection of user privacy [
68]. Particularly in crisis events, local organizations can use digital nudging to disseminate topic-specific tweets (e.g., emergency notifications, evacuation information, etc.) to support emergency management objectives and to manage the crisis properly [
69]. In the long run, social media has become the most widespread channel for users to generate, access, and share all kinds of information. It is worthwhile to further explore how to use interface design and nudge to better assist users to search and share useful information more efficiently while privacy is effectively protected, and to guide other benign usage behaviors.
The blue cluster focuses on individual behavior, preferences, and the specific nudge mechanisms used in the implementation of behavioral interventions. At the early stage in applying Nudge Theory, Thaler and Sunstein [
70] acronymized six mechanisms for optimizing choice systems to improve usage satisfaction into
nudge, i.e., iNcentives, Understand map**s, Default, Give feedback, Expect error, and Structure complex choices. Since then, researchers have extended their design insights into additional fields, focused on the measurement and examination of influences that optimize nudge effects. As the network visualization shows, the main common nudge mechanisms are default options [
71,
72], social norms [
73], incentives [
74], and feedback [
75]. Especially during the COVID-19 pandemic, nudge was used to promote vaccination, reduce social contact, disseminate trustworthy pandemic information, etc. [
76]. In summary, the designer of any choice environment must be aware of its effect on people’s choices. Choice architects should be aware of the goals of the intervention to design and test nudge to maximize the desired effect [
77]. Another study indicated that the vividness of image presentation increased gamification and improved the subjective usability of face-to-face counseling effects, which promoted counseling in real life for young people [
78].
Finally, the yellow cluster, which concentrates on the extension of choice architecture to consumer behaviors that intervene in people’s consumption, such as healthy food choices, sustainable consumption behaviors, chronic disease or cancer prevention and treatment, and pro-environmental behaviors. The ever-accelerating pace of life, irrational eating patterns, and obsession with digital media affect physical and mental health, especially among students and young people in the workplace. Because of health and pro-social factors, most people are more receptive to nudge [
79], although some studies have found that nudge to promote healthy eating is not effective [
80], or even that employees can accept nudge while students do not accept nudge, leading to more unhealthy food choices and making nudge ineffective [
81]. This may be because the nudge mechanism used is different and some nudges may be perceived as manipulative or uncomfortable, so the intervention is not as ideal [
82]. In the context of disease prevention and treatment, it was demonstrated that moderate interventions in individual rights and relatively unproblematic moral imperatives, nudge proved valid in various situations [
83], and in particular, that personally tailored and positively constructed messages were more persuasive than generic and/or negative messages [
84].
On top of the network visualization, VOSviewer offered different map** visualizations; the other two constructed are the overlay visualization (as in
Figure 9) and the density visualization (as in
Figure 10). Specifically, there are two groups of keywords in
Figure 9 with the earliest average years of nudge research. The first group is Nudge Theory-related terms and research, such as “neoliberalism”, “behavioral economic”, “libertarian paternalism”, and “autonomy”. The second group is related to public health, such as “healthy food”, “disease”, and “obesity”, which is in line with the number of journal articles and citations discussed earlier and is a topic that has been studied early and consistently in nudge research. The above keywords are directly linked to both “nudge” and “choice architecture”, while several keywords on the right of the figure, such as “stimulation”, “sensitivity”, and “precipitation”, are linked to each other, but only “sensitivity” is linked to “nudge”. As there is no other connection that indicates relevance to the nudge discussed in this study, none are further reviewed. After 2019, increasing attention is paid to nudge in the digital environment, such as the keywords “digital nudging”, “fake news”, and “social media”, especially since the outbreak of COIVD-19 in early 2020. The spread of the pandemic has led to the implementation of stay-away orders and social distancing in many countries, in which personal use of social media multiplies. Researchers have also turned to the effects of nudge in this special period, with the keywords “COVID-19”, “vaccination”, and “hand hygiene” surging after 2020. Remarkably, there are also emerging keywords such as “sustainable consumption”, “artificial intelligence”, and “gamification”, indicating a growing enthusiasm for nudge research that extends to a growing number of fields.
The depth of research in the field related to nudge can be observed in
Figure 10. In the present analysis, colors range from blue to green to yellow to red. The higher the number of items in the proximity of a point and the higher the weights of the related items, the closer the color of the point is to red. Contrarily, the sparser and less impactful the point, the closer its color to blue. Through density visualization, we can quickly observe that features around nudge research consumption, health, management, ethics, donation, sustainability, specific nudge mechanisms, and information systems are currently widely discussed topics.