Introduction

In recent years, science communication has received increasing attention from politics and society, and it has developed into a vital part of academic activity [1]. As science communication is of rising importance for scientists, and even for their academic careers [2], our work aims to bring a timely and important contribution to this area of research. We aim to contribute towards a broader understanding and deeper insights into science communication from the perspective of scientists in terms of individual viewpoints and the goals scientists focus on in their science communication practices as well as the goals scientists are trying to achieve when using social media for science communication.

As a growing area of practice and research, science communication over the last 50 years has developed into a field primarily concerned with answering questions related to science and society, science in the media, and the role of science journalists [3]. In the academic realm, specifically, these questions are addressed by researchers of sociology, journalism, and media studies. As it is the case with most multidisciplinary, rich research fields, science communication has produced different—and sometimes even contradictory—theorisations [4]. It comes as no surprise, then, that scholars have not agreed on a clear and shared definition of science communication [4, 5].

In very general terms, science communication refers to communication on scientific topics and the relationship between science and society. But as Burns, O’Connor and Stocklmayer state [6], the definition of science communication has been “plagued by an unfortunate lack of clarity” (p.183). The wide array of definitions offered by the science communication literature ranges from those referencing some set of signature methods and aims [7] to those focusing on “manifest or latent purposes of science communication practice” [4, p.4].

For example, Davies [8] defines science communication in a contemporary conception as “organized processes that seek to engage lay publics with scientific knowledge.” (p.163) While this definition seems simple and straightforward, it restricts science communication to those processes in which the public engages with scientific knowledge as a finished product. Such definitions are in clear contrast to those highlighting the importance of public engagement in the process of generating scientific knowledge [9, 10] and those advocating for a change in perspective—from a communication process to a ‘conversation’ [4]. Bryant [11] takes further steps to acknowledge the intangible cultural aspects of science communication and that science communication is a continual process rather than a one-off, linear activity and defines science communication as “… the processes by which the culture and knowledge of science are absorbed into the culture of the wider community.”

Scholars such as Burns [6] take a more pragmatic approach, formulating a definition aimed at providing the framework for evaluating the effectiveness of science communication from an outcomes oriented perspective:

“Science communication […] is defined as the use of appropriate skills, media, activities, and dialogue to produce one or more of the following personal responses to science (the AEIOU vowel analogy): Awareness, Enjoyment, Interest, Opinion-forming, and Understanding.”

The definitions exemplified above are by no means the only ones nor do they include all involved actors, beyond scientists and the lay public (such as institutions, journalists, etc.) and all societal contexts (e.g., policy-making). However, an exhaustive review of all the science communication definitions is beyond the scope of this article. Rather than attempt to further portray the breadth of science communication definitions, we would like to highlight the importance of exploring scientists’ personal science communication definitions. While some studies have attempted to understand scientists’ choices and objectives using the Theory of Planned Behavior [12] and the associated Integrated Behavioral Model [13], and others have focused on understanding scientists’ views about specific communication objectives and strategies [14], our contribution focuses on uncovering how scientists themselves define science communication.

A survey conducted in the UK by Illingworth and colleagues [15] further highlights the importance of understanding scientists’ personal definitions of science communication and how such definitions may be more nuanced and expansive than those presented by literature or various organisations. They show that, even though scientists believed their personal definitions were aligned with pre-existing ones, in reality the very wide range of personal definitions was more nuanced and expansive [15]. Thus, empirically assessing various aspects of science communication from the perspective of a pre-determined definition creates an unnecessarily rigid analysis framework.

Empirical assessment departing from the scientists’ personal science communication definitions can foster a broader understanding and deeper insights [16, 17] into individual viewpoints [18, 19]. By focusing on scientists’ “logic and the kinds of evidence they considered worthwhile and relevant, we suspend our own notions of what is right.” [20] In other words, the investigation of personal definitions can help us identify and understand what science communication is from the perspective of those who are closest to the scene. Here we argue that—while understudied—scientists’ personal definitions form the basis for their approaches to science communication and their communication activities. The importance of understanding scientists’ “individual conceptions of science communication” (or what the authors refer to as ‘mental models’) is further highlighted by a very recent study conducted by Kessler and colleagues across three European countries [21]. Their findings suggest that scientists’ individual conceptions largely align with the ways in which they practice science communication.

The scientific community has often been primarily focused on disseminating information and building public knowledge about science [22–24] and on developing theoretical models that describe the relationship between scientists and the public [25, 26]. However, very few studies have focused on understanding the specific goals scientists are trying to accomplish through their science communication efforts [21, 27], and scholars have generally not treated communication choices related to strategies, objectives, and goals as behaviours that can both be studied and reshaped [28].

Over the last decades, a rhetorical shift has taken place in the area of science communication in that calls for dialogue between scientists and non-scientists, as well as calls for more participatory approaches to science communication, have taken precedence over the scientific literacy deficits rhetoric. Together with this rhetorical shift from ‘deficit’ towards ‘dialogue’ and ‘participation’, scholars have developed a number of science communication models ranging from transmission of information to dialogue and public engagement. Such models are said to be frameworks for understanding what the “problem” is, how to measure the problem, and how to address the problem [29, p.13]. Models of science communication, emerging in the literature, represent how theorists believe that science has been, is being, or should be communicated [30]. However, as we will argue, the assumptions behind most of these models have not been comprehensively tested against real-life practices of science communication [31]. But let us first discuss three of the most prominent models, their underlining assumptions, and described goals.

Historically, science communication had been described as a process of information transmission, which assumed “public deficiency, but scientific sufficiency” [32]. What has become known as the deficit model of science communication was first discussed in the literature during the 1980s and early 1990s [33, 34]. This model asserts that science communication is guided by a perceived need for science literacy, where the public is viewed as having a deficit of scientific knowledge until such knowledge is received from the communicator [35–37]. As Irwin [38] describes, the deficit model aims at ‘public understanding of science’ and is characterised as a one-way, top-down communication process. In other words, science communication is a one-way process from the scientists (‘experts’) to the public (‘non-experts’) in which scientists—with all the required information—filled the knowledge vacuum in the scientifically illiterate general public as they see fit [39].

The perspective of the deficit model is perfectly exemplified by the UK Chief Scientific Adviser to the House of Lords, in a report published in 2000 which purports “difficulties in the relationship between science and society are due entirely to ignorance on the part of the public” and “with enough public-understanding activity, the public can be brought to greater knowledge, whereupon all will be well.” [40, p.25]. This assumption that by fixing the deficit everything will be ‘well’ has led to the emergence of many large-scale projects focused on science literacy, often addressing revisions of national educational curricula [41]. However, research has shown not only that the deficit assumption is an incorrect one—as people with limited education can quickly come to understand highly complex technical information [42]—but that after more than 25 years of ‘filling the deficit,’ the public’s ‘knowledge’ seems unaffected and remarkably stable [43]. Under the deficit perspective, by making the knowledge deficits of the public central to the problem of science communication, many scientists ignore the possibility that their communication approaches might be part of the problem [44].

Scholars have theorised that the condescending claims of the public’s ignorance under the deficit model [45] and the increasing signs of public unease with science during the 1980s [30] have led to a new style of science communication as embodied in the dialogue model. This new model of science communication draws on the belief that “more active, open and democratic relations between science and citizens are both desirable and necessary.” [38, p.200]

The dialogue model is characterised by three main features: (1) engagement in a dialogue with the public to help explain the science [46], (2) listening to and consulting the public about their perceptions, concerns and needs with regard to science [37, 47], and (3) acknowledging that the public may have useful knowledge that can help scientific progress and policy-making [48]. Under this model, science communication is centred around dialogic approaches that aim for a two-way conversation, allowing an exchange of perspectives between scientists and the public.

The emergence of the dialogue model, together with the increasing demand for scientists’ involvement in public discussions and with policy documents shifting their language from ‘communication’ to ‘dialogue,’ have created a powerful narrative in which this type of science communication is portrayed as inherently superior [49]. However, many scholars have argued that this shift from deficit to dialogue has not been as complete as the narrative suggests [37]. More than ten years of research evidence has questioned the scale of this supposed shift and the extent to which the dialogue model goals and principles have been adopted in actual science communication practices [46, 50–54]. The apparent replacement of the deficit model, as Wynne [46] concludes, is more nominal than real. Thus, even though the science communication vocabulary has changed considerably over time, the underlying assumptions may still be inline with those that inform a deficit model perspective [37].

In recent years, the narrative shift has evolved even further to include public participation or public engagement. Public participation (or public engagement) models have emerged as a direct attempt to enhance social trust in science policy. These models focus on a series of activities—consensus conferences, citizen juries, deliberative technology assessments, science shops, deliberative polling, etc. [55, 56]—driven by a commitment to ‘democratize science’ through some form of empowerment and political engagement of the public. Although rarely linked in science communication literature, the public participation model shares many similarities to more established techniques, such as public meetings and public hearings [57, 58].

Participation models signal a more obvious shift in power than the dialogue model and emphasise the role of the public and other societal stakeholders in reflecting upon, sharing knowledge about, creating new knowledge, and making decisions about science that affects society [9, 10]. Participation models seek to set the sciences in a wider social context, addressing societal concerns and priorities that involve multiple stakeholder perspectives [38]. It is argued that such approaches to science communication are driven by the intent of ‘democratizing’ science by giving the public more control over science through some form of empowerment and political engagement [59].

Scholars have posed a number of criticisms related to participation models due to their focus on the process of science over any substantive content, their limitations in terms of the numbers of people they serve, and for sometimes exhibiting an ‘anti-science’ bias given their focus on lay/local knowledge over scientific knowledge [41]. The strongest criticism, however, relates to the fact that such models are said to address politics more than the public understanding of science due to their commitment to a particular stance about political relations [60, 61].

The three science communication models discussed above, driven by different assumptions, provide only schematic tools for theorising about science communication activities. While the deficit model centres around filling a knowledge vacuum and the dialogue model on an exchange of perspectives between the sciences and the public, the participatory model centres on the ‘democratisation of science’ through some form of empowerment and political engagement of the public. The emergence of these models (and many others) together with the strong rhetorical shift from deficit to dialogue and participation—both in scholarly and policy discourses—seem to indicate a unidirectional transition process. However, Bucchi and Trench [4] suggest that rather than looking at a ‘fixed triad of deficit, dialogue and participation’ (p.8), this range of models should be seen as a dynamic spectrum that is continually growing in both directions.

While the radical transition—from deficit to dialogue and participation—and the concurrent rhetorical shift have been deemed highly implausible over such a short period of time [37, 46], very few studies have endeavoured to empirically assess the specific goals scientists are trying to accomplish through their science communication efforts [27] and how their goals align with those of the deficit, dialogue, and participation science communication models. Although research has shown that many scientists accept and support dialog and participation as some kind of ‘gold standards,’ or are at least are aware of the push toward public communication [62], not many studies focused on assessing how this translates into actual science communication practices. One of the few studies investigating such practices found that—due to the continuing adoption of a simplistic contrast structure that opposes science and the public as two self-contained, antagonistic social entities—the shift towards more democratic engagement of the public has not been as profound and complete as expected [63].

In this article, we aim to address this research gap by investigating the goals scientists focus on in their science communication practices and the alignment of these goals with the three models discussed thus far. To do so, we build on the work of Metcalfe [30] that lists the primary goals inherent to each of the three science communication models, based on a thorough literature review. We further discuss these goals and their importance to our work in our Data and Methods section.

When discussing contemporary science communication practices, one must not omit to highlight the radical and important changes brought about by the emergence of social media channels. On a daily basis, millions of people all over the world are constantly consuming and creating content through social media platforms. Considering the popularity of such platforms (e.g., 192 million daily active users on Twitter), it is easy to see that information disseminated through these channels can reach millions in a matter of minutes and, as Van Eperen and Marincola [64] acknowledge, that successful communication can only be achieved by using the channels in which the public is currently engaged. Thus, social media channels offer a powerful venue through which scientists can act as a public voice for science in a quick and efficient manner by using a medium in which the public is already engaged [65].

Beyond reaching engaged and diverse publics—researchers, the general public, government, and all other stakeholders—social media platforms have the potential to enable multi-vocal, multi-way communication [39]. Such platforms are said to “facilitate interactive information sharing” [66] and to increase the depth and reach of engagement among stakeholders [67]. The potential of social media platforms to involve a wide variety of stakeholders in an interactional manner have long been emphasised by proponents of the dialogue and participation science communication models [39, 68]. However, in reality, many researchers are cautious in changing traditional scholarly communication patterns in response to social media [69] and the area of online science communication remains understudied [65].

When it comes to science communication, although social media seems to be the ideal environment for two-way and multi-vocal communication models, the few studies that have investigated online science communication have found that this is only one part of the story. science communication practitioners and science organisations use social media for one-way message dissemination more often than they truly engage with their publics [70, 71], and they generally underutilise social media’s potential to create true dialogue with their audiences [72]. For instance, when conducting a review of social media activities of 11 US based science agencies, Lee and VanDyke [73] found that their outreach activities did not facilitate two-way interactions very well, suggesting an adherence to a deficit-model way of thinking for these agencies. Similarly, analysing Twitter utilisation patterns of NanoDays science festivals organised between 2012 to 2015, Su and colleagues [74] found that most tweets exemplified a one-way, information-sharing model of communication.

Just as in the case of the proposed radical transition of science communication—from deficit to dialogue and participation—the belief that social media affordances foster unprecedented opportunities for two-way and multi-vocal science communication is becoming more or less an accepted fact. But, as the few studies that have investigated social media use for science communication have shown [73–75], one-way, information-sharing practices aligned with the deficit model still dominate. So if social media offers such opportunities, how come scientists still engage in information transmission behaviours? Here, we argue that in order to answer this question, we need to first have a better understanding of the goals scientists are trying to achieve through their social media science communication practices. To bring a contribution to this under-researched area of science communication, we investigate not only the goals scientists consider as most important in the context of general science communication, but we also assess which goals they find important in the context of social media science communication. To do this, we used a modified version of Metcalfe’s [30] list, specifically reworded for social media. We will further discuss this list in our Materials and methods section.

Considering the promises of social media as a participatory environment where already-engaged audiences can easily be reached, we also aim to investigate whether those actively using social media for their science communication efforts may approach science communication from a more participatory oriented perspective. To do so, we aim to compare between social media users and non-users and their general science communication goals.

Materials and methods

General science communication goals and models

To assess the goals respondents find most important to their science communication efforts and how these goals align with the conceptualisations proposed by the deficit, dialogue and participation models, we used a 28 item scale. Participants were asked to rate the importance of the 28 items to their science communication goals on a 5-point Likert-type scale, ranging from 1 = ‘Not important at all.” to 5 = “Extremely important.” This question appeared in the survey after respondents were asked to defined science communication in their own words.

The 28 items in our scale are based on the extensive literature review published by Metcalfe [30]. In her work, Metcalfe [30] used this 28 item scale with similar aims to ours: to assess theorised models against real-life science communication practice. However, her work, centred on the specific context of a 2012 Australian audit and 515 science engagement activities, found that most engagement activities had objectives that reflected a mix of deficit and dialogue activities and a lack of participatory activities in Australia [30]. Rather than focusing on activities, in our work, we will employ the 28-item scale to assess the goals our respondents find most important to their science communication efforts. Based on the literature review conducted by Metcalfe [30], each of the 28 items in the scale (listed in Table 1) align to one of the three science communication models as follows: goals 1 to 14 align with the deficit model, goals 15 to 20 align with the dialogue model (15-20), and goals 21 to 28 align with the the participation model. To ensure that the 28 items form a reliable scale, both as a whole and as three separate scales aligned with each of the three science communication models, Cronbach’s alphas were calculated for each. We find that the 28 items list as a whole (Cronbach’s α = .925) and also divided according to the three models form reliable scales (see Table 1).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. General science communication goals and models.

https://doi.org/10.1371/journal.pone.0277677.t001

To verify if the 28-items in our scale align to the three science communication models, as proposed by the literature review conducted by Metcalfe [30], we used factor analysis. More specifically, we employed confirmatory factor analysis (CFA), a statistical technique used to verify that the theory-postulated relationships between observed variables and their underlying latent constructs exist [80]. The CFA was conducted using AMOS (version 26.0.0). First, we confirmed the 28 items in our scale are suitable for factor analysis through a Bartlett test (χ²(378) = 7762.743, p = < .000) and the Kaiser–Meyer–Olkin measure of sampling adequacy (KMO = .913). Next, factor loadings were assessed for each of the 28 items and three items with factor loadings less than the acceptable threshold of 0.30 [81] were removed from the analysis. Thus, from the initial 28 items presented in Table 1, items 6, 11 and 21 were removed from further analysis, which resulted in a final scale of 25 items. Strong factor loadings were found across the remaining 25 items, ranging from.42 to.71. Next, we evaluated the three-factor model fit using the relative Chi-square (χ²/df) [82], the comparative fit index (CFI) [83], the goodness of fit index (GFI) [84], the standardized root mean square residual (SRMR) [85], and the root mean square error of approximation index (RMSEA) [86].

Our results confirm the unidimensionality of each construct in our model and indicate that the measurement structure of three factors (Deficit, Dialogue, and Participation) and 25 items produced good fit statistics (see Table 2). As the sample size of our study is relatively large (N = 584), it is within expectation that the model did not pass the χ² test (p < .05) [87–89]. Lastly, we found that even after the removal of the 3 items, the remaining 25 items form reliable scales, both as a whole (Cronbach’s α = .918) and as three separate scales aligned with each of the three science communication models: Deficit model Cronbach’s α = .830, Dialogue model Cronbach’s α = .817, and Participation model Cronbach’s α = .855.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Confirmatory factor analysis results for a three-factor model of 25 general science communication goals (N = 584).

https://doi.org/10.1371/journal.pone.0277677.t002

In our Results section, we explore these 25 general science communication goals individually (i.e., based on their item means across all respondents) as well as partitioned according to the three science communication models they represent, as confirmed by the factor analysis reported above. For the latter, we computed two aggregate variables for each of the three latent variables/models (Deficit, Dialogue, and Participation): 1) the unit weighted scale means and 2) factor scores using factor score weights produced by the CFA, normalized to allow for comparison across models [90–92]. While the bulk of the statistical analysis in this paper will focus on factor scores, the results from the scale means are also reported; these have been argued to hold complementary utility that is less dependent on sample characteristics (e.g., replicability of findings across studies [93]). After we computed the pairs of variables for each respondent and communication model, Shapiro-Wilk tests [94, 95] showed a significant departure from normality for all of our variables (see Table 3). Due to the variables’ non-normal distribution, and because the variables from each model are based on related samples, we used Wilcoxon signed-rank tests to determine whether the differences in how our respondents rate goals of the three science communication models are statistically significant [96, 97]. As a non-parametric statistical test that compares two paired groups, the Wilcoxon signed-rank can be used as an alternative to the t-test when the population data does not follow a normal distribution [97].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Shapiro-Wilk tests—General science communication models.

https://doi.org/10.1371/journal.pone.0277677.t003

Social media science communication goals and models

Additionally, using Metcalfe’s [30] scale, we selected nine of the 28 general science communication goals that were most suited to the particulars of social media science communication—three goals for each of the three models discussed—to create a scale for social media science communication goals (Cronbach’s α = .846). Although the items in this scale were specifically worded for the social media context, the question asked respondents to rate the importance of these items for their social media science communication activities, just as in the case of their general science communication goals. The nine items aligned to the deficit model (1-3), the dialogue model (4-6), and the participation model (7-9) and they form reliable scales (see Table 4).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 4. Social media science communication goals and models.

https://doi.org/10.1371/journal.pone.0277677.t004

Just as in the case of the general science communication goals, we ran confirmatory factor analysis to assess whether the nine items in our scale align with the three science communication models, as proposed by Metcalfe’s literature review [30]. We first confirmed that the 9 items in our scale are suitable for factor analysis through a Bartlett test (χ²(36) = 963.682, p = < .000) and the Kaiser–Meyer–Olkin measure of sampling adequacy (KMO = .872). Next, and unlike in the case of the general science communication goals, no items were removed from the three-factor model due to strong factor loadings, ranging from.60 to.76. Using the same model-fit evaluation measurements (χ²/df, CFI, GFI, SRMR, and RMSEA), we confirmed the unidimensionality of each construct in our model. Hence, a three factor model (SM Deficit, SM Dialogue, and SM Participation) with 9 items produced good fit statistics (see Table 5). Because our sample size is relatively large, the model did not pass the χ² test, as models based on large sample sizes have a higher tendency to be rejected [87–89].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 5. Confirmatory factor analysis results for a three-factor model of 9 social media science communication goals (N = 314).

https://doi.org/10.1371/journal.pone.0277677.t005

Just as we did for the general science communication models, in our results section, we explored the nine social media goals individually, but we also computed latent variables for each model, using the factor score weights as well as the unit weighted means of each scale. An exploration of these three new pairs of variables (for each SM Deficit, SM Dialogue and SM Participation) also shows a statistically significant departure from normality, as shown by the Shapiro-Wilk tests in Table 6.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 6. Shapiro-Wilk tests—Social media science communication models.

https://doi.org/10.1371/journal.pone.0277677.t006

Results

Before describing our findings, let us first describe the demographics and characteristics of our survey respondents. 54.6% of our respondents identified as female, 43.2% as male, and 0.9% as non-binary/third gender, while 1.4% preferred not to say. Most of the respondents in our data set held a PhD (58.6%) or a Masters’ (38.9%) degree, with 77.1% being employed under a full time contract. Interestingly, as per Table 7, the majority of respondents in our sample were early-career researchers (including 37.8% Ph.D. candidates and 11.8% Postdoctoral researchers). We further elaborate on the implications of this finding in our discussion.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 7. Demographics (N = 584).

https://doi.org/10.1371/journal.pone.0277677.t007

In the following subsections, we discuss our findings regarding our respondents’ personal science communication definitions, their general science communication goals, their social media science communication goals, and how these goals align with the three science communication models discussed in the introduction of this paper, namely the deficit, dialogue, and participation model.

Personal definitions of science communication

The thematic content analysis conducted on the 584 definitions provided by our survey respondents resulted in 30 codes and 6 themes. These codes and themes highlight the various science communication aspects, dimensions, and perspectives our respondents highlighted in their definitions. In Table 8, we show our codes, their occurrence frequency, and an example for each code as well as the themes emerging from collating our codes. Here, we should note that a code was applied to each definition provided by our respondents only once. In Table 8, the codes are organized by theme and ordered by frequency.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 8. Qualitative analysis results: Themes, codes and examples.

https://doi.org/10.1371/journal.pone.0277677.t008

Overall, our respondents highlighted six main dimensions of science communication, as shown by the themes that emerged from their definitions. These themes center on the type of communication (N = 620), the audience addressed (N = 511), the content of the communication (N = 389), the goals to be achieved (N = 148), the medium through which the communication takes place (N = 57), and the impact of science communication (N = 30). We will discuss each of these themes and their most prominent codes individually.

Type of communication.

Within the type of communication theme, the most prominent dimension we uncovered during our coding was that of information transmission/transfer. As shown in Table 8, most definitions (63.7%) made mention of science communication as an information transmission process, using terminology such as disseminating, transmitting, or transferring scientific knowledge or information to an audience. As we will discuss later, this process of transmission was highlighted regardless of the type of audience addressed. The second most prominent emerging dimension was that of translating scientific information (27.57%), while dimensions related to more dialogic, engagement, or participatory dimensions of science communication emerged in a very limited number of definitions. Thus, the focus on information transmission and translation seems to suggest that most of our respondents consider science communication to be a one-way process of information diffusion, from the knowledgeable scientist to the uninformed audience.

Audience.

Our qualitative analysis showed that the next most prominent theme our respondents focused on when defining science communication is the audience to be addressed. While most respondents mentioned a lay or non-academic audience (62.16%), a good proportion also included academic audiences in their definitions (23.29%), while very few made reference to interested (1.20%) or uninformed audiences (0.86%). In sum, when defining science communication, our respondents identify academic as well as non-academic stakeholders as main audiences for their science communication efforts. 111 (19.01%) definitions included both, academic and non-academic audiences. As we will discuss later, these definitions may be indicative of a dichotomy, a separation between scientists and other members of the public, which has further implications.

Content.

The third most common theme emerging from these personal definitions of science communication was the content of their communication efforts. Here, most respondents emphasized research findings or results as the most common dimension of their definition (42.81%). Fewer respondents made mention of involving their audience in the process of science (11.30%) or of communicating the relevance (7.02%) and methods (5.48%) of scientific research. Considering our earlier finding—that most personal definitions we investigated focused on a one-way, information transmission model of science communication—it then come as no surprise that the most predominant content mentioned was research results. This would suggest that our respondents look at science communication as a process of transferring a ready-made product of science (results) to their audiences, whether academic or not.

Goals.

Our analysis also uncovered a number of goals highlighted in the science communication definitions provided by our respondents. The most prominent of these goals was that of making science more accessible (5.48%), educating their audience (4.11%), raising awareness about science and scientific topics (3.94%), and enhancing interest in science (3.60%). The other six goals emerging in our analysis (see Table 8) were far less frequently mentioned by our respondents. As we will discuss later in this paper, most of these goals are very much in line with the deficit model of science communication, with the exception of the goal of making science more accessible, an inherently dialogue-oriented goal [30].

Media.

The fifth theme uncovered by our analysis was that of the medium through which science communication takes place from the perspective of our respondents. Here, 5.82% definitions mentioned traditional media (e.g., newspapers, radio, TV etc.) as the primary channel for science communication, while 3.94% made mention of digital and social media (e.g., blogs, Twitter, podcasts) as science communication platforms. This suggests that among those who include a means or channel of communication in their definitions, the majority made reference to more traditional forms of media, and fewer see digital and social media as a science communication venue.

Impact.

The last and least prominent theme that emerged from our analysis was that of impact. The definitions formulated by our respondents included communicating about the general (2.57%), societal (1.88%) and practical (0.68%) impact of their research.

To summarize, the qualitative thematic analysis of our respondents’ personal science communication definitions aimed at creating a broader understanding of what science communication is, from the perspective of those who are closest to the scene. While not trying to discount the very few personal definitions that have included aspects of dialogue-based and participatory science communication, our results suggest that the majority of these definitions are centered around a deficit way of thinking by emphasizing a one-way, transmission model of communication that aims to transfer and translate research outcomes and its impacts to a non-academic audience, with the purpose of making science more accessible, of educating audiences, and/or of raising awareness about science.

Having investigated how our participants define science communication, in the next sections of our paper we quantitatively analyze our respondent’s general science communication goals, as well as the goals they find most important when specifically asked about social media science communication. Furthermore, we discuss how these goals aligned with the deficit, dialogue, and participation models of science communication.

Science communication models

We start the description of our quantitative analysis results with the ways in which our respondents rated the importance of the 25 general science communication goals and the science communication models they represent. Our results indicate that, on average, our respondents rated deficit model goals as more important than dialogue model goals and participatory model goals. In Fig 1, we show the means of each of the 25 goals, in descending order and colored by the communication model they align with. The highest rated goals in our scale were deficit model goals. Based on the rating of our respondents, increasing understanding of science, educating, transferring information, raising awareness, and correcting misunderstandings or misconceptions were the most important goals to their science communication efforts. This particular finding resonates with our qualitative analysis findings, in that it seems to indicate that a deficit way of thinking was the predominant approach to science communication among our respondents.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. General science communication goals ordered by mean (N = 584).

https://doi.org/10.1371/journal.pone.0277677.g001

The means of each model—as already hinted at by our qualitative results and by looking at the highest rated goals—indicate that our respondents did indeed rate deficit model (M = 3.81) and dialogue model goals (M = 3.66) as more important than participation model goals (M = 3.38). A Wilcoxon signed-rank test of the factor scores further confirm that our participants rated Deficit model goals higher than Dialogue model goals and that this difference is statistically significant: T = 133157;z = −12.530;p < .001. We also find that on average, our participants rated Deficit model goals (T = 138057;z = −13.757;p < .001) and Dialogue model goals (T = 139912.50;z = −14.221;p < 001) significantly higher than Participation model goals. The same significance levels are also found between pairs of models when testing the scale means rather than the factor scores (p < .001 for each pair). Thus, we can infer that in terms of science communication goals, our respondents focused more on a Deficit and Dialogue perspective, rather than a participatory approach.

Analyzing the personal definitions and the goals our respondents find most important to their science communication activities and efforts has thus far provided a relatively straightforward portrayal of mostly deficit or dialogue oriented perspectives, with participatory approaches underrepresented or entirely lacking. However, considering the emergence of social media and the promises of such digital channels to be the ideal environment for two-way and multi-vocal science conversations, we move on to investigate whether those actively using social media for their science communication efforts may approach science communication from a more participatory oriented perspective.

First, we note that from our 584 our respondents, only 314 (53.77%) reported that they use social media for science communication. Our 314 participants who said they use social media for science communication reported that they used LinkedIn (38.9%), Twitter (33.4%), Facebook (15.8%), and YouTube (0.9%) for science communication in the past 12 months, and also other less frequently used platforms (e.g., Instagram). All of these most commonly used platforms among our respondents offer affordances that can be used to enter a dialogue with or involve already-engaged audiences into participatory science communication activities. Thus, in the next paragraphs, we assess what type of goals those participants who use social media for science communication rated as most important.

When looking at how our respondents rated their agreement to the 9 statements related to social media science communication goals (see Fig 2), we found that the highest rated goal is a dialogue model one, namely the goal of making science/scientists more accessible. Unlike in the case of general science communication, where respondents rated a deficit model goal as the most important, we see that when specifically focusing on social media, participants rated a dialogue model goal as most important. This would seem to indicate that perhaps in the context of social media, respondents moved from the deficit view to a more dialogue oriented one. However, the next two highest rated goals are related to raising awareness and transferring science information, two inherently deficit model oriented goals. Furthermore, we note that overall the highest rated goals (based on their means) were those of the Deficit model (M = 4.04) followed by Dialogue model (M = 3.90), while Participation model goals were rated lowest (M = 3.47). To further confirm the significance of this finding, we performed Wilcoxon signed-rank tests on the factor scores of our model variables. The results confirm that, on average, our participants rated Deficit model goals significantly higher than Dialogue model goals, T = 31957.00, z = −8.025, p < .001, and significantly higher than Participation model goals, T = 36779, z = −11.452, p < .001. The tests also show that, on average, Dialogue model goals were rated significantly higher than Participation model goals, T = 38018.50, z = −12.332, p < .001. Significance levels are found to be similar when using means instead of factor scores as well. Thus, when specifically asked about their science communication goals in the context of social media, and even though social media promises to be an ideal environment for dialogue and participation, our respondents’ focus remained mostly on the deficit-oriented perspective.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Social media science communication goals ordered by mean (N = 314).

https://doi.org/10.1371/journal.pone.0277677.g002

Additionally, our analysis found statistically significant differences in how respondents who reported using social media for science communication (N = 314) rate their general science communication goals, when compared to those who reported not using social media for science communication (N = 270). Overall, the factor score means listed in Table 9 indicate that social media users rated goals associated with all three models higher than non-users, and the Mann-Whitney U tests for each pair further confirm that these differences are statistically significant. Thus, in all cases, social media users rated goals significantly higher than non-users. When using scale means and Mann-Whitney U tests, we find similar significance, with the exception of the Participation model goals (p = .124).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 9. Mann-Whitney U tests.

https://doi.org/10.1371/journal.pone.0277677.t009

This particular part of our analysis showed that when comparing between social media users and non-users, we found important differences. Interestingly, social media users rated their general science communication goals higher than non-users, perhaps indicative of higher activity among social media users in communicating science.

Conclusion

Using a mixed-methods approach, this article set to uncover scientists’ personal science communication definitions; the goals they find important to their science communication efforts; the alignment of these goals to the deficit, dialogue, and participation models of science communication; and whether a deficit-model way of thinking remains dominant in social media science communication efforts. In the following paragraphs, we discuss our findings, starting with the qualitative thematic analysis results, followed by our quantitative findings and in the context of previous research. We discuss whether the highly popularised—and theorised—rhetorical shift from deficit to participation is reflected by our results, and we argue that we should be cautious of recent claims that we live in a new era of dialogue, transparency, and participation in the realm of science communication.

Discussion

Over the last two decades, scientific as well as funding and policy initiatives, in many countries, have shifted their rhetoric to move away from deficit ‘public awareness of science,’ and ‘science and society’ towards ‘citizen engagement,’ ‘dialogue,’ and ‘science in society.’ This change in terminology was suggested to be indicative of a transition from a deficit perspective on science communication, towards dialogue and participation. However, many have deemed this shift to be highly implausible over such a short period of time [37, 46] and that most science communication initiatives are still guided by the misconception that deficits in public knowledge are the driving force behind conflict over science [36]. Even those outreach projects and activities that claimed to follow other theoretical approaches have been shown to use the deficit model approach as a backbone [29].

With this study we add to research evidence that has questioned the degree to which science communicators have truly adopted the goals of this proposed shift towards dialogue and participatory science communication [50, 51, 103, 104]. Our results, though limited to scientists in the Netherlands, showed a predominant focus on the deficit model of science communication, with limited elements of a more dialogue centered approach and almost no emphasis on participatory approaches.

This narrow emphasis of the deficit approach does not allow for the recognition that scientific knowledge is not the only factor influencing how individuals reach judgments. Decades of research have shown not only that science literacy plays a very limited role in shaping public perceptions and decisions, but also that people with limited education can quickly come to understand highly complex technical information [42, 105]. Furthermore, simply sharing science information in a unidirectional fashion is not an adequate way of increasing science knowledge or changing attitudes about science [106, 107]. Taking a deficit approach to science communication presents a major obstacle for reflective exchange and mutual learning, and ultimately for the transition towards more democratic forms of public involvement [63].

As noted in our results and conclusion, our data also showed evidence of more dialogue-driven approaches among our respondents. This dialogue model perspective was evident in both the ways in which respondents defined science communication and in the general science communication goals they rated as most important. However, while the dialogue perspective is driven by aims of engaging in two-way conversations, this dimension was not prominent in our data. Rather, our participants highlighted the importance of making science and scientists more accessible. Because our respondents consider being and making science more accessible as important goals, perhaps a transition away from dissemination and towards an exchange of perspectives between scientists and the public is not out of the question. As Metcalfe [30] also argues, making science and scientists more accessible may make it easier for the public to engage with science, and it could be considered a perquisite for more sophisticated dialogue activities that have the potential to open up productive and surprising discussions [108].

Arguably, thus far, our discussion may read as yet another pejorative scholarly discussion about the deficit way of thinking that argues for a clear separation between deficit, dialogue, and participatory science communication approaches. This is not our intention and such a discussion would be rather unhelpful for the development of science communication. Rather, the few elements of dialogue-driven science communication approaches, together with the prominent deficit elements uncovered by our study, lead us to argue that scientists do not adhere to a single model of science communication and that a clear separation between these models is not realistic. Instead, as other studies have also found, a combination of goals and practices—inherent to different models—govern science communication efforts and perspectives [30, 52].

What is of great concern, however, is that participatory aspects of science communication were nonexistent in most of the definitions formulated by our participants and that participatory goals were invariably rated the lowest, whether in a general or a social media context. Here, we are not trying to argue that deficit or dialogue model based activities are inherently inferior, nor do we want to suggest that such activities should be entirely abandoned. Rather, we would like to emphasise the importance of true engagement with various publics, especially when dealing with the complexities of current contemporary issues. Even when deficit model approaches are at the backbone of such science communication activities, participatory elements can be invaluable tools for reflecting upon science and knowledge, creating new knowledge, and most importantly, placing science in a wider social context [10, 38].

This lack of focus on participatory goals among our respondents is especially problematic considering the demands of contemporary knowledge societies and in the face of fast-paced technological advancements. The rise of digital platforms and social media have created an entirely new ‘ecology’ of communication between science and the public [109]. In the past scientists and science organizations have relied on science journalists or professional communicators to interpret scientific information for the general public. While to a certain extent this still holds true today, digital media platforms have made it possible for scientists to communicate directly with publics [110]. This digital transformation of science communication comes with both positive and negative consequences. On the positive side, digital platforms can be used to facilitate broader involvement of citizens in science discussions and they can allow individuals to learn about science and to become involved in collective decision-making [111]. On the negative side, the abundance of information sources available to individuals and the prevalence of misinformation in digital media create an urgent need for science-society relationship and trust building [112].

While it may still be possible to communicate scientific findings in accordance with the traditional deficit model, it has become increasingly difficult to generate public trust in science on the basis of pure scientific testimony alone [100]. As Elam [100] states, in contemporary knowledge societies the challenges facing science communicators are no longer those of combating public ignorance, but rather those of creating “new forms of community around radical programmes of research still in the making.” (p. 251) Scientists need to re-frame their view of technology and digital platforms and consider their uses in more robust, relationship building activities [110].

The findings presented in this study are based on the results of a survey administered among scientists in the Netherlands. As it is with most surveys, and even though we made great efforts in disseminating and publicizing our study, only 584 complete responses were collected. Thus, our results and the subsequent discussion should be viewed in the context of a limited number of responses provided by scientists in the Netherlands.

Additionally, we saw that a large number of our respondents were early-career scientists (Ph.D. candidates and postdoctoral researchers). Previous studies discuss a shift in the culture surrounding more participatory approaches to science communication to be driven by early-career scientists [113]. However, our findings—although based on a sample in which early career scientists are over-represented—suggest an overall adherence to deficit model approaches, with some elements of dialogue. For a discussion on some of the factors (institutional and personal) that may impact the lack of more participatory oriented approaches to science communication, in the context of the United States, see [113, 114]. Moreover, while Howell and colleagues [115] found that both late and early career scientists held positive views on the role of social media in providing opportunities to engage with the public, our respondents’ focus nevertheless remained on deficit-oriented goals when asked about their social media science communication goals. While we can only speculate that perhaps the differences between the academic contexts of the Netherlands and of the United States might be the underlying cause of these differences, further research comparing academic contexts and science communication approaches among scientists (much like the one conducted by Kessler and colleagues [21]) could shed more light on the issue.

The demographics collected in our survey did not include information about the affiliations of our respondents. Although this particular choice in the survey design offered respondents more anonymity, our results cannot speak to differences between scientists affiliated with different universities in the Netherlands, nor can they make distinctions between scientists from different disciplines. Previous studies have found that scientists’ academic fields play a role in their approaches to science communication. For example, Burchell found that participatory approaches to science communication was considerably higher in fields like the arts, humanities, and social sciences compared to the STEM (Science, Technology, Engineering, and Mathematics) fields [116]. Similar results are also reported by Kessler and colleagues, with scientists from the humanities, social sciences, and life sciences being more likely to adhere to dialogue and participatory perspectives [21].

Our study builds further evidence that a deficit model approach to science communication is still very much present [73, 74, 99, 117], even in the face of increasing calls towards science communication approaches that foster dialogue and public participation [36]. Simis and colleagues [99] discuss four lines of reasoning for why the deficit way of thinking remains very much present in science communication. Their essay highlights the roles of scientists’ training in processing information in a rational manner, the lack of formal public communication training, viewing the public as “others”, and the fact that the deficit model works well for policy design as some of the most prominent reasons for the persistence of the deficit model.

Understanding that scientists are focused mostly on dissemination and translation of research results, with a willingness to make themselves and science more accessible, has important consequences for scientific organizations, the shaping of their long-term science communication strategies, the development of training programs for established scientists, and the inclusion of science communication curricula in graduate education programs. As Madden, Cacciatore, and Yeo [118] observe, without continued training in positive communication methods, grounded in social science research, “it is not surprising that scientists would follow the admittedly intuitive deficit model.” (p. 403) The importance of continued training in positive communication methods as well as the importance of engaging community members around scientific issues are also emphasized by Simis and colleagues [99] as ways of moving beyond knowledge deficit approaches.

Lastly, the abundance of science communication models proposed by many scholars over the years could not all be addressed in our work. For example, one established model we did not address is the contextual model [25]. This model is generally tied to particular audiences, focuses on particular context-dependant situations (e.g., locations), and highlights the ability of audiences to quickly become knowledgeable about relevant topics [29]. However, while this model highlights the importance of contextual factors in how individuals receive and process information, it nonetheless conceptualizes the same focus on the responses of individuals to scientific information, just as the deficit model does [119]. Another model that was not included in our investigation is the one described first by Scheufele [120] and later by Akin and Scheufele [121], namely the communication in context model. Also known as the science communication as political communication model, this expanded model of science communication takes into account the larger political contexts of science communication, acknowledges that scientists are just some of the many voices involved in scientific debates, and highlights the many science-related streams of information the public is exposed to. As Scheufele [120] argues, answering questions about emerging technologies will require “democratic decision making that draws on moral values, that weighs complex political options, and that includes debates about the ethical and legal aspects” (p.13590). Thus, it is reasonable to argue that such models can help us move beyond deficit perspectives and that they better fit with the realities of our contemporary societies and the complexities of current scientific advances. We encourage future studies to include such models in their investigations.

To conclude, science communication and public engagement can no longer be focused on persuading the public on the ‘facts’ of science, as we live in an era where most science debates are no longer small, localized events, controlled by scientists or science communicators. As such, and as Nisbet and Scheufele [36] also proposed, contemporary science communication and public engagement activities should move away from a deficit, top-down perspective, towards promoting conversations that “recognize, respect, and incorporate differences in knowledge, values, perspectives, and goals.” (p. 1777)