Abstract
Purpose
From a public health perspective, vaccination programmes significantly add to long-term, safe co-existence. However, because there is no social consensus about their benefits and risks, the promotion of vaccinations is difficult. Based on Kim and Grunig’s situational theory of problem solving (STOPS), including communicative action in problem solving (CAPS), this paper both proposes a model for identifying the involvement of mothers of young children in communication regarding vaccination and advocates for a novel approach to STOPS and CAPS data analyses.
Design/methodology/approach
The methodological design develops empirical analyses of the data yielded by the STOPS model. Two approaches to determining associations between situational-motivational variables and communicative-action variables in random-sample survey data obtained in Slovenia in 2016 (N = 1704) are implemented – i.e. visual methods and multivariate agglomerative clustering algorithm.
Findings
The STOPS model has been confirmed and both data-analyses approaches have shown potential by clearly demonstrating associations and patterns in the data. Based on these findings, we conclude that they have the potential to be the norm in analysing STOPS models.
Research limitations/implications
Limitations of the study, which are still to be overcome, involve drawing on one sample in one country and testing only one set of indicators.
Practical implications
From an academic point of view, confirmation of both the model and the analytical power of the pragmatic data-analyses methods significantly add to communication studies. From practical and social points of view, relationships among attitudes and communication behaviour, as outlined in the exposed segments of the public, enable the improvement of every step in strategic-communication planning and implementation.
Originality/value
This paper fulfils an identified need to establish a theoretical framework and methodology of segmentation in vaccination-communication studies.
Keywords
Citation
Kropivnik, S. and Vrdelja, M. (2024), "STOPS and multivariate hierarchical aglomerative clustering: segmentation of the public regarding children’s vaccinaton communication in Slovenia", Corporate Communications: An International Journal, Vol. 29 No. 7, pp. 109-129. https://doi.org/10.1108/CCIJ-01-2024-0007
Publisher
:Emerald Publishing Limited
Copyright © 2023, Samo Kropivnik and Mitja Vrdelja
License
Published by Emerald Publishing Limited. This article is published under the Creative Commons Attribution (CC BY 4.0) license. Anyone may reproduce, distribute, translate and create derivative works of this article (for both commercial and non-commercial purposes), subject to full attribution to the original publication and authors. The full terms of this license may be seen at http://creativecommons.org/licences/by/4.0/legalcode
1. Introduction
1.1 Significance and challenges of vaccination communication
Vaccination is one of the greatest achievements in public health (Baumgaertner et al., 2018; Gostin et al., 2020). Vaccination programmes have contributed to the reduction of mortality (Ratzan et al., 2019) and morbidity caused by various infectious diseases (Guidry et al., 2015). They have been credited with the eradication of polio in Nord America and Europe, as well as the elimination of smallpox worldwide (Dubé et al., 2013). To prevent illness and mortality associated with diseases preventable by vaccination, vaccination programmes must achieve and maintain high vaccination coverage rates (Eskola et al., 2015). Otherwise, the benefits of “herd immunity” will be lost (Blume, 2006). Herd immunity can protect the lives of the most vulnerable individuals and reduce the societal and economic burdens of crises, such as the COVID-19 pandemic (Neumann-Böhme et al., 2020).
However, it is challenging to maintain high vaccination coverage when the prevalence of certain vaccine-preventable infectious diseases is decreasing and the likelihood that individuals will doubt vaccination recommendations is increasing (Fine et al., 2011). In situations in which some infectious diseases have become rare, vaccine hesitancy and vaccine rejection have already increased (Oraby et al., 2014). The fear of potentially deadly diseases has been replaced by the fear of the fictional, in addition to real side effects of vaccination (Cintulová, 2019). Despite clear and unequivocal scientific and medical consensus on the benefits of vaccination, an increasing number of people perceive vaccines as dangerous and unnecessary (Dubé et al., 2021). Consequently, in recent decades, the coverage rates of vaccination against preventable diseases have decreased in many parts of the world, leaving many children unvaccinated (de Figueiredo et al., 2016; Ricciardi et al., 2018). The consequences of the decline in high vaccination coverage are extensive and significant, preventing progress towards general sustainability. As vaccination coverage rates decline, the spread of illness and death from vaccine-preventable diseases will increase, affecting not only those who refuse vaccination but also the general population (Ropeik, 2013).
These changes have led to shifts in vaccination communication. Effective communication is crucial for maintaining high vaccination coverage, but the content must be appropriate for the audience (Altay and Mercier, 2020). Poor or inappropriate communication may lead to greater risk exposure (Burgener, 2020), significantly influencing vaccine acceptance and contributing to vaccine hesitancy and refusal (Leask et al., 2012; MacDonald and the SAGE Working Group on Vaccine Hesitancy, 2015; Olowo et al., 2020).
1.2 Situational theory of problem solving as the segmentation framework
Approaches to vaccination communication vary, and their success depends on numerous factors. Various perspectives may be applied to segment the public – e.g. demographic characteristics, actual behaviour, lifestyles and values or attitudes. However, in the field of communication and public relations, Grunig’s situational theory of publics (STP) (Grunig, 1997) is widely recognised, since it includes factors that drive individuals to communicate in favour of or against a point at issue, and correspondingly explains when communications aimed at people are most likely to be effective (Grunig, 2005, 2013; Kim and Grunig, 2011; Kim et al., 2014). Its latest and extended form, developed in cooperation with Kim (Kim and Grunig, 2011), is the situational theory of problem solving (STOPS), which includes communicative action in problem solving (CAPS) and most precisely defines situational-motivational and communicative elements of the model and their theory-based relations (Kim, 2013; Kim et al., 2010). Key constructs, such as involvement and information selection, are proposed in the theoretical model. However, in the application to a specific situation, the relevance of key constructs, as well as their conceptualisation and operationalisation, are strictly contextual, as they all depend on the problem under investigation.
2. Research purpose and the structure of this paper
Based on Kim and Grunig’s situational theory of problem solving (STOPS), including communicative action in problem solving (CAPS), this paper proposes a model for identifying the involvement of mothers of young children in communication regarding vaccination, as well as a novel approach to the STOPS and CAPS data analyses. The purpose of this paper is therefore twofold.
The generic purpose addresses the application of the STOPS model to vaccination communication. On the one hand, there is a general lack of vaccination-communication studies, and existing studies are built on different foundations. On the other hand, STOPS has so far been hardly ever and almost always partially used in vaccination studies, despite its firm theoretical foundation and meticulously elaborated empirical model (see “Literature Review” below). Thus, the corresponding research question is: “Can our conceptualisation and operationalisation of the indicators be confirmed against the theoretically anticipated relations in the STOPS model?”
The specific yet equally or even more important purpose of the article addresses data analyses. STOPS applications typically include ordinal communicative-activity variables and numerical motivational variables. That makes analyses of relations between the two groups rather complicated, since many complex statistics must be interpreted simultaneously to effectively draw conclusions (e.g. in ANOVA tables). Thus, the relevant research question is: “How can researchers parsimoniously, yet productively and vividly, disclose relations in a STOPS model?”
Consequently, the next section contains a literature review of the intersection of segmentation studies with STOPS applications. Special attention is paid to health communication studies, particularly to vaccination-communication studies. The following section, “Methodology”, describes how STOPS has been applied as well as how the data were obtained and analysed. The outcomes are presented in the “Results” section, followed by the final section, “Conclusion and Discussion”.
3. Literature review
In health communication, the existing public-segmentation studies are built on different foundations. Members of the public can be divided into several different subgroups (sub-publics) based on different criteria – age, sex, ethnicity, self-efficiency, values, personal characteristics, social status, etc. – and into different levels. By combining different criteria, thousands of subgroups suitable for targeted actions can be formed. Yet the fundamental assumption is the same: the audiences differ because of different lifestyles, motivations, relationships, etc., and they follow different behaviour patterns (Moss et al., 2009). As a result, audience segmentation in relatively homogeneous subgroups enables targeted planning and implementation of communication activities (Noar, 2006), which is more efficient than communication with a large, heterogeneous population (Kim et al., 2008). Shaping relatively homogeneous subgroups enables strategists to prepare various messages adapted to specific characteristics and predispositions of each subgroup, and to choose the right communication channels in order to reach a targeted audience (Atkin and Freimuth, 2012). In general, segments should be plausible, mutually exclusive, measurable, relevant to the organisation’s mission, achievable by communicating in an affordable way and large enough to serve a communication campaign economically (Grunig, 1989).
STOPS stipulates factors that drive individuals to communicate in favour of or against a point at issue, as well as respects personalities of individual communicative action. Numerous distinctive applications of STOPS have contributed to defining and identifying diverse publics as well as explaining motivational and perceptual factors that increase communication activities among members of specific publics. Applications include various topics and fields – for example, environmental studies (Ismail et al., 2017; Jiang et al., 2017; Li et al., 2019), internal communication (Kim and Rhee, 2011; Lee, 2019), corporate social responsibility (Kim et al., 2019), crisis communication (Chon and Harrell, 2024; Kim, 2016; Liu et al., 2019; Poroli and Huang, 2018) and socio-political issues (Kim, 2018; Lee et al., 2014; Lovari et al., 2011; Lovari et al., 2012).
On the other hand, STOPS is less frequently used in the health field. Three comprehensive studies are specified in this paragraph and a further three, dealing specifically with vaccination, are specified in the next paragraph. Chronologically ordered, the first explored model usage in connection with organ donor shortage, and noted that factors or problems, being the subject of interest, were usually grouped and integrated into a problem net. One of the sub-publics that recognised the organ shortage problem simultaneously recognised a related problem: examining a cure for muscular atrophy with stem cell therapy (Kim et al., 2011). The second study is related to chronic diseases: two types of online communication were tested – information seeking and information forwarding. The effects of these two types of communicative action on perceived affective and physical coping outcomes were assessed (Kim and Lee, 2014). The most recent study applied the STOPS model to explore individuals’ communicative actions in relation to the COVID-19 pandemic. The authors found that perceptual backgrounds affect situational motivation, and situational motivation affects communicative actions. Communicative actions are a determining factor in individuals’ willingness to follow instructions (Akbulut, 2023). The study by Kim and Lee only partially used STOPS, since just the communication part is included. The remaining two studies fully used the STOPS model, and therefore confirmed the agreement of the theoretical model with empirical data.
According to our knowledge, only three studies in vaccination communication exist, all of them from the United States. Chronologically ordered, the first study (McKeever et al., 2016) focuses on how and why individuals communicate about vaccinations on digital platforms. The authors discovered that those who do not support childhood vaccinations are more likely to engage in communication about this issue, including information seeking, attending, forefending, permitting, forwarding and sharing. In addition, the importance of the issue, and the affective and cognitive involvement thereof, help drive communicative action regarding childhood vaccinations, which could affect public opinion or public perception about the issue. The second study (Krishna, 2018) explores situational and cross-situational factors that influence individuals’ attitudes toward vaccines. By using STOPS, this study identifies and tests a piece of knowledge: the attitude–motivation–behaviour framework of individuals’ negative cognitions and behaviours about vaccines. The third study with STOPS usage in the vaccination area (Xu et al., 2021) focuses on forecasting health behaviours in online health communities. These authors discovered that it was possible, from information selection and acquisition, to forecast HPV vaccination intention. Moreover, perceived seriousness and perceived susceptibility directly impact HPV vaccination intention and has an indirect impact due to information selection and acquisition. Perceived message credibility also indirectly affects HPV vaccination intention via information selection.
All the vaccination-communication studies above are founded on a part of the comprehensive STOPS model. In contrast, in our study STOPS is used more holistically, as an all-inclusive theoretical framework, providing solid fundaments for segmentation. Situational-motivation factors were identified in our study, as stipulated in STOPS. Communicative activity is completely covered in our study, in contrast to the first and last studies above, which both focused only on online communication.
4. Methodology
4.1 The STOPS model and indicators
In our STOPS application (Figure 1), situational motivation was measured according to (1) affective involvement, (2) cognitive involvement, (3) trust in the paediatrician, (4) trust in the health system and (5) trust in science. The values of these five situational-motivation variables were collected on a five-point Likert scale. Communicative action was measured by (1) information acquisition, (2) information selection and (3) information transmission on four or five channels of communication. The frequency of use of 14 forms of communicative action was measured on a five-point ordinal scale (Vrdelja, 2023: 14–96 and 146–202).
Survey data were collected as part of an interdisciplinary research project on controlling infectious diseases through vaccination (Kraigher, 2018) in which we participated as co-authors. Random sampling was applied to a population of approximately 40,000 mothers of newborns and babies up to two years old, registered in the national database of births. Responses from 1,704 mothers were obtained (44.5% sample realisation).
4.2 Data-analysis strategies
The first necessary step in data analyses is the empirical confirmation of the theory-based selection of indicators. Associations between the two groups of variables, situational motivation and communicative action, must be verified to authorise the application of the model. Once we demonstrate that variables are associated and – even more importantly – understand how situational motivation and communicative action relate in this situation, we can proceed with defining different segments of the population under investigation.
STOPS applications typically include ordinal measurement-level variables representing communicative activity and numerical motivational variables. From a statistical point of view, analyses of the association between numerical and categorical variables require comparisons of the mean values of numerical variables in subsamples, defined by categories. In other words, creation of aggregates of units (people) who share the same frequency of the same activity on the same communication channel, and calculation of mean values of situational-motivation variables measured on a Likert scale for these aggregates, is the first step of the analysis of the model. The results are presented in Table 1 (N refers to the number of units in an aggregate and is printed only for aggregates smaller than 100 units).
To expose associations between situational-motivation and communicative-action variables, we compared the mean values shown in the sectors of the Table 1. A sector is defined as the intersection of each motivational variable with each activity on each channel. If the mean values in a sector varied according to the frequency of an activity on a channel and a non-random pattern of the variation was recognised, then the motivational variable was associated with the communicative action on the channel.
However, as shown in Table 1, because five indicators measured motivation and 14 indicators measured communicative action, we derived 5 x 14 = 70 subdivisions (sectors). Additionally, in each division, there were five mean values, so we compared 10 pairs of mean values to define variations in each section of the table ((5 * (5–1))/2). Because there were 70 relevant subdivisions, we compared 700 pairs of mean values altogether. This considerable number of comparisons was difficult to summarise and present comprehensively.
Well-known statistical methods that can be used to compare mean values among aggregates are based on series of t-tests for independent samples or on multivariate analyses of variance (ANOVA). In addition to calculating all the mean values, these methods control for variations in an aggregate, calculate statistics for comparison between (t-test) or among (ANOVA) mean values, expose differences and provide tests for statistical inference (valid in the case of a random sample). Consequently, these methods demand an even larger number of comparisons. Since they produce additional statistics needed for in-depth comparison of the mean values under investigation, the number of required comparisons triples. That makes the problem of comprehensively summarising and presenting the STOPS models’ results even more difficult.
Therefore, in the following subsection we propose two pragmatic, efficient methods for analysing typical STOPS models, using a combination of numerical and categorical variables.
4.3 Methodological complements to STOPS
The first proposed method straightforwardly draws on a visual approach (Tufte, 1997, 2011, 2020) to determine associations between situational-motivational variables and communicative-action variables, which are depicted in line graphs. Each subdivision in Table 1 is represented by a line graph that connects the five mean values in that section (see Figure 2). Consequently, each line replaces five mean values and simultaneously vividly depicts the pattern of their variation. In so doing, a line exposes the association between a situational-motivation variable and a communicative action on a channel. Although altogether we are still dealing with 70 lines that must be examined, the distinctive advantage of the suggested visual approach is that line graphs are much easier to comprehend than sets of numbers. Consequently, it is also easier to combine pieces of information from individual lines in a comprehensive total outcome, summarising how situational-motivation variables are associated with communicative-action variables.
The second proposed method draws on formal mathematical procedures – specifically the multivariate agglomerative clustering algorithm. This hierarchical clustering method joins units that are similar enough to be regarded as equal in distinctive clusters. In our case, the main feature of the method is that it eventually reduces the number of necessary mean - value comparisons to a manageable number. A comprehensive explanation of the method is provided in Johnson and Wichern (2013, pp. 671–696), while the specific application to the STOPS model is outlined in the following paragraphs.
The rows of Table 1 are treated in the cluster analyses as units for classification (objects). Accordingly, each classification unit is an aggregate of mothers who shared a unique combination of an activity (e.g. information acquisition) on a channel (e.g. newspaper) at a certain frequency (e.g. never) and therefore represents a unique variant of communicative action. Concurrently, each classification unit is characterised by its average affective involvement, cognitive involvement, trust in a paediatrician, trust in the health system and trust in science, respectively. In other words, each classification unit comprises the typical levels of involvement and trust (i.e. the five situational-motivation indicators) of a communicative-action variant, uniquely defined by activity, channel and frequency.
At the core of agglomerative clustering methods are analyses of similarities among units, which result in merging all units that are similar enough not to consider them as separate units. Hence, all communicative-action variants characterised by similar levels of involvement and trust are organised in the same cluster and treated as joint in successive analyses. Usually, the number of extracted clusters is not high, but somewhere between two and eight. Based on such classification results, we can associate the variants of communicative action in a cluster with the level of involvement and the level of trust in the same cluster. The latter is indicated by the average values of the situational-motivation variables in a cluster, and the former is obtained by systematically summarising the characteristics of the communicative-action variants in a cluster.
In addition, the obtained clusters represent segments of the aggregates of the analysed public, created according to the trust and involvement indicators, then related to the communicative-action level. As such they are valuable for strategic communication planning. Finally, because agglomerative clustering methods are hierarchical, we may interpret results on multiple levels, from general to specific, as demonstrated in the following section.
5. Results
5.1 Graphs
As shown in Figure 2, each subdivision in Table 1 is represented by a line graph that connects the five mean values in that section. The order of the mean values from left to right in each line is defined by the increasing frequency of a communicative activity on a channel, which is specified below the line. Lines in the same row (of five rows) are based on the same situational-motivation variable, and all lines in the same column are based on the same communicative activity on a particular channel (14 columns). A line at the intersection of a situational-motivation variable and a communicative action on a channel depicts the association between both variables.
As shown in Figure 2, different lines indicate different types of association. The flat line indicates no association between affective involvement and the frequency of information acquisition using leaflets and brochures (the fourth line from left to right in the first row). In contrast, when the internet was used for information acquisition, a relationship with affective involvement is shown in the inclined straight line. An increase in usage frequency corresponds with a linear rise in affective involvement (the third line from left to right in the first row). However, to identify an association, the increase or decrease in the situational-motivation variable do not have to be equal in every change in the frequency of communicative action (i.e. shown in a straight line).
As shown in Figure 2, most lines are not straight, but they still indicate some kind of an association. For example, an increase in affective involvement is more significant in higher frequencies of newspaper information acquisition (the second line from left to right in the first row). Only a few lines are difficult to understand (e.g. the first line from left to right in the first row) because they significantly rise or fell only at the highest frequency, indicating that only the aggregate that utilised the channel the most frequently differs from all other aggregates with lower usage frequencies. Because such aggregates were sometimes composed of only a few individuals, N (Table 1) was consulted before making the final decision about the association (in large print or online graphs, N can be annotated on a graph). For example, the lines “Radio or TV, information transmission” in the 10th column of the graph decreases significantly regarding all three forms of trust at the highest frequency. However, only two individuals utilised radio and television channels for information transmission more frequently than 12 times per year.
Because lines are easy to comprehend, it is also easier to combine the content of individual lines into a comprehensive conclusion. Starting from the general, by observing all 70 lines simultaneously, we conclude that the situational-motivation variables are associated with the communicative-action variables. Lines that do not represent an association (i.e. flat lines) or represent random patterns (no general trend) are extremely rare – only three or four of 70. Some associations are obvious and therefore strong; others are moderate; and some are weak, yet still recognisable.
The findings above suggest that the model with our indicators is confirmed. Regarding the interpretation of associations – that is, how situational motivation and communicative action are related in our study – we drew the following conclusions:
- (1)
Comparatively higher levels of affective and cognitive involvement are typical of more frequent communicative action on all channels. Regarding information acquisition, information selection and information transmission, all lines but two in the first two rows of Figure 2 are increasing. The two exceptions are associated with information acquisition through leaflets. Predominantly, individuals are likelier to communicate a certain topic when they are more involved. This finding aligns with previous studies that have examined the relationships between involvement and communication behaviour and demonstrated a positive association between these two concepts (Kim et al., 2011; Kim and Lee, 2014; Chen et al., 2017; Kim et al., 2018; Won et al., 2018; Li et al., 2019; Shen et al., 2019).
- (2)
The opposite tendency is identified in all three varieties of trust: trust in a paediatrician, trust in the health system and trust in vaccination science. Comparatively lower levels of trust are characteristic of frequent communicative action on all channels regarding information acquisition, information selection and information transmission. Most lines in the bottom three rows of Figure 2 are decreasing; exceptions are associated with information transmission through radio or television, newspapers and events. This interpretation aligns with the findings of previous studies, confirming that lower levels of trust increased information demand and the variety of information sources (Hak et al., 2005; Ramanadhan and Viswanath, 2006; Yaqub et al., 2014). A problem arose when individuals accessed material from unreliable sources and false information, which they uncritically used to make decisions about health matters.
- (3)
All associations with the motivational variables are stronger when the internet was used for the acquisition, selection and transmission of information. As shown in Figure 2, the slope of the lines is evident in all columns regarding the internet. Simultaneously, the lines are straighter, indicating that each increase in the frequency of use was matched by a decrease in trust. This finding suggests that as a channel, the internet has the most significant effect on vaccination acceptance and consequently on vaccination prevalence.
- (4)
Equivalent but less far-reaching findings apply when verbal communication with a friend was used to select and transmit information.
- (5)
When radio and/or television and newspapers were involved as channels, the associations are weaker and less clear. The same applies to the organisation of public events for information transmission.
- (6)
The associations regarding information selection are stronger than those regarding information acquisition, and the weakest associations include information transmission.
- (7)
The findings described in the last four subparagraphs are cumulative. For example, when the internet was involved, the associations are always highly evident, and even more so if the internet was used for information selection. In contrast, when radio and/or television were involved as channels, the associations are less clear, particularly when the channel was used for information transmission.
5.2 Multivariate agglomerative clustering
Hierarchical-clustering methods, details and variants are comprehensively discussed in Johnson and Wichern (2013, pp. 671–696). In our case, clusters were obtained by selecting Euclidean distance as the dissimilarity measure regularly used with Likert scale variables. Non-standardised variables were used because they all were measured on the same scale. The Ward agglomeration method was applied because it is the most complex and productive method of hierarchical clustering. The optimal number of clusters was extracted from the dendrogram, presented below. Due to the hierarchical structure of the dendrogram, we were able to interpret the results on multiple levels.
In the right half of Figure 3, the dendrogram represents the hierarchical agglomerative clustering process. The leftmost part of the dendrogram shows the individual variants of communicative action. The further to the right, the more clustered they are. In the rightmost part of the dendrogram, the final two clusters are joined. The appropriate solutions lie between the two extremes of all separate units and all units in one cluster. The most appropriate ones are marked by vertical lines, defining classifications into three and into five clusters.
The optimal two solutions were identified as a combination of statistical relevance (increase in the level of difference between the two uniting clusters at a certain step in the process) and contextual significance (what could be gained from a certain solution compared with other suggested solutions). Statistically, optimal solutions are on the interval from two to six clusters. Contextually, classifications in two, four and six groups include clusters that are not specific enough (i.e. too general and too close to the sample average), or not distinctive enough (i.e. too similar to another cluster).
Therefore, the classification in three clusters is presented in Figure 3 as full lines, enriched by a further split of one of the three clusters into three subclusters: the five-clusters solution (dotted lines). In full or dotted line boxes, the characteristics of clusters, shown as subdivisions in the dendrogram, are presented by average level of involvement (i.e. affective and cognitive) and average level of trust (i.e. in a paediatrician, in the health system and in science) in the cluster, as well as by a list of variants of communicative actions that comprise the cluster. These results are discussed in the text below the figure.
As shown in the full lines in Figure 3, the segments in the general three-clusters solution are significantly distinctive regarding involvement and trust. The first cluster (topmost, full line) is characterised by a low level of involvement and a high level of trust. Both involvement indicators are significantly below the sample average, and all three trust indicators are significantly above the sample average (see the statistics in brackets in the box in Figure 3 and the sample averages in Table 1). These represent the exact opposite of the third cluster (bottommost, full line), which was marked by a high involvement level and low trust level. Again, the involvement and trust indicators are internally consistent. Both involvement indicators are significantly above the sample average, and all three trust indicators are significantly below the sample average. In the second, middle cluster, the involvement indicators and the trust indicators do not significantly deviate from the sample means. Therefore, the cluster is identified by average levels of involvement and trust. Nevertheless, the indicators are internally consistent.
In contrast, the cluster characterised by low involvement and high trust, and the cluster characterised by high involvement and low trust, comprise greatly differing variants of communicative action. Although both include all three types of communicative action (i.e. information acquisition, selection and transmission) and almost all channels, an obvious difference is identified by the frequency of variants of communicative action. The cluster characterised by high involvement and low trust is composed of the highest frequencies of occurrence – i.e. 7–12 times per year and at least 13 times per year (Figure 3, bold font in boxes). The antagonistic cluster (low involvement and high trust) comprises the lowest frequencies of occurrence of all communicative actions – i.e. never or 1–2 times per year (Figure 3, italics in boxes). Some higher frequencies are found only in information acquisition. Lastly, the average involvement and trust cluster represents a mixture of low, middle and high frequencies, whereas high frequencies are absent from information transmission.
Based on these findings, the following main conclusions that confirm the model and specify theoretically anticipated associations were drawn:
- (1)
The level of involvement is associated with the level of trust. If one is low, the other is high –and vice versa (i.e. two opposite clusters). It is also possible that both are in the middle (i.e. the middle cluster).
- (2)
A high intensity of communicative action is associated with a high level of involvement accompanied by a low level of trust. Correspondingly, a low level of involvement and a high level of trust are associated with low intensity of communicative action. The middle level of communicative action is associated with a middle level of trust and involvement.
As shown by the dotted lines in Figure 3, the advance on the detailed level is in the split of the “high involvement, low trust and high intensity of communicative action” cluster, which was the most problematic and sensitive from the point of view of vaccination promotion. The other two clusters remained the same. This split resulted in three new clusters. The cluster shown in the bottom right of Figure 3 greatly resembles the mother cluster. Characterised by high involvement and low trust, this cluster is mainly comprised of the highest frequencies of occurrence of information acquisition, information selection and information transmission. However, the latter occurred frequently only when the channels were the internet and friends.
The other two new clusters are small, but they disclose particularities potentially important for strategic communication regarding the promotion of vaccinations in children. The first of these two is shown in the bottom right of Figure 3. It is composed of those who uniquely expressed high levels of both involvement and trust and who were also characterised by frequent information transmission. They engaged at least once a month in the transmission of information in newspapers, radio and/or television. They also frequently organised their own events to transmit information. The second cluster is shown in the bottom middle of Figure 3. This cluster includes those who used newspapers for information selection more than once a month and those who used radio or television for information transmission more than once a month. This cluster contains the highest average involvement level and the lowest trust level.
Based on these findings, the following additional conclusions regarding more specific aspects of the obtained clusters were drawn:
- (1)
On a general level, strategic vaccination communication should separately address at least three main segments of mothers who significantly differ in their level of trust, involvement and communicative action. The communication strategy should consider that trust and involvement are inversely interrelated, as well as that communicative-action level is inversely associated with the level of trust and directly with the level of involvement.
- (2)
What was established on a general level is observable on a specific level, with one small yet notable exception of mothers characterised by a high level of trust as well as by a high level of involvement. These mothers should be included in promotional activities as vaccination ambassadors to promote vaccination among equals, from one mother to another. This would be efficient in convincing mothers who are afraid of the side effects of vaccination and those who have doubts and questions, but regard health experts as not being personally involved and as disaffected. Promotion inside a group sharing the same parental responsibility would be the most effective (Attwell and Freeman, 2015).
- (3)
A cluster in line with general principles but representing an extreme emerged in the results: an engaged group of mothers who opposed vaccination and did not use the internet but consulted traditional media, such as newspapers, radio and television. Therefore, it would be important to address media publishing policies because some media have used anti-vaccine debates as a financial opportunity to increase readership and sell books, services and other products (Larson, 2018). It would be equally important to ensure that journalists have scientific knowledge and that they are motivated to explore the backgrounds of anti-vaccination activists.
- (4)
Specifically, the findings show that information transmission is a communicative activity that defines particularities. Therefore, information transmission should be the focus of vaccination promotion, as well as communication planning and monitoring. Information selection had only marginal effects, and information acquisition had no effects on the detailed segmentation level.
6. Conclusion and Discussion
A model for recognising involvement of mothers of young children in communication regarding vaccination was developed (Vrdelja, 2023: 14–96), employing Kim and Grunig’s situational theory of problem solving (STOPS), including communicative action in problem solving (CAPS). The model was empirically confirmed, utilising the random sample survey data obtained in Slovenia in 2016 (N = 1704) and employing a novel approach to the STOPS and CAPS data analyses. The proposed two data-analysis methods showed potential by straightforwardly revealing associations and patterns in the data; we therefore conclude that these methods have the potential to be the norm in analysing STOPS models.
The first method relies on traditional visual methods and focuses on associations between Likert scale situational-motivational variables and ordinal scale communicative-action variables, as shown in line graphs. The outstanding advantage of this method is that line graphs are much easier to comprehend than sets of numbers in conventional statistical tables. The second method, multivariate agglomerative clustering, draws on formal mathematical procedures and reduces the final number of mean value comparisons (in our case by approximately 90%) by combining units that are similar enough to be regarded as equal in further analyses. Both methods reveal associations and patterns in the data and provide insights based on differing yet compatible perspectives. The unparallelled additional benefit of the clustering method are provisional segments of the analysed public, created according to the trust and involvement indicators, and associated with the communicative-action level. As such, the obtained segments are particularly valuable for strategic vaccination-communication planning. In addition, due to a hierarchical nature of the method, segments may be recognised on multiple levels, from a rather general level (a few broad-spectrum clusters) to a very detailed level (a large number of specific clusters). Thus, we identify the methods’ potential to become the norm in analysing applications of the STOPS model, respectively or jointly.
From an academic point of view, this paper fulfils an identified need to establish a theoretical framework and methodology of segmentation in vaccination-communication studies. Confirmation of the application of the model, and the demonstrated power of the advocated pragmatic data-analysis methods, significantly add to communication studies. In demonstrating that at first glance a homogeneous public of mothers, whose main drive is the well-being of their newborns, can in fact be incredibly heterogeneous, the necessity of segmentation for strategic vaccination promotion and communication planning is confirmed. Any communication strategy should regard that trust and involvement are inversely interrelated, as well as that a communicative-action level is inversely associated with the level of trust and directly with the level of involvement.
From practical and social points of view, the numerous outlined relationships among attitudes and communication behaviour will enable the improvement of every step in strategic vaccination-communication planning and implementation. The mothers in this study evidently used diverse sources to obtain information about vaccination, and they utilised them in varying frequencies. The findings show that some sources had stronger links with involvement and trust than others. Crucially, the internet was shown to be particularly influential. The problem with the internet is that when unofficial sources of health communication are consulted, the acquired pieces of information are far less reliable, often in fact deceiving. Hence, reliance on the internet, accompanied with increasing lack of trust in the health system and in medical science in general, has underlined the importance of promoting digital media literacy, particularly regarding health topics. Regarding the exposed low level of trust in some segments, health institutions must work to (re)gain public trust by practising expert, transparent and open communication. The competencies of health professionals in interpersonal communication must be improved to strengthen their direct influence on patients’ decision-making. Also, since the results demonstrate that conventional media are still influential, media policies and journalists’ approach to the topic of vaccination must be thoughtfully but exhaustively reflected on in view of ethics and objectivity. Finally, the limitations of this study, which are still to be overcome in further research, involve drawing on one sample in one country and testing only one set of indicators. The latter limitation is already being attended to in another ongoing project of ours.
Figures
Mean values of situational-motivation variables in communicative-action aggregates
Involvement | Trust | ||||
---|---|---|---|---|---|
Affective | Cognitive | Paediatrician | H.C. SYSTEM | Science | |
INF. ACQUISITION: radio or TV = never | 3.25 | 4.17 | 3.28 | 3.7 | 3.43 |
INF. ACQUISITION: radio or TV = 1–2 times per year | 3.36 | 4.17 | 3.36 | 3.37 | 3.34 |
INF. ACQUISITION: radio or TV = 3–6 times per year | 3.27 | 4.07 | 3.19 | 3.56 | 3.52 |
INF. ACQUISITION: radio or TV = 7–12 times per year; N64 | 3.27 | 4.03 | 3.32 | 3.53 | 3.33 |
INF. ACQUISITION: radio or TV = (at least 13 times per year; N15) | 2.6 | 3.87 | 2.73 | 4 | 4.33 |
INF. ACQUISITION: NEWSPAPER = never | 3.27 | 4.16 | 3.27 | 3.57 | 3.42 |
INF. ACQUISITION: NEWSPAPER = 1–2 times per year | 3.35 | 4.17 | 3.36 | 3.49 | 3.34 |
INF. ACQUISITION: NEWSPAPER = 3–6 times per year | 3.23 | 4.04 | 3.17 | 3.49 | 3.44 |
INF. ACQUISITION: NEWSPAPER = 7–12 times per year; N78 | 3.32 | 4.27 | 3.4 | 3.62 | 3.64 |
INF. ACQUISITION: NEWSPAPER = (at least 13 times per year; N24) | 2.79 | 3.87 | 2.91 | 3.75 | 3.96 |
INF. ACQUISITION: INTERNET = never | 3.45 | 4.27 | 3.48 | 3.45 | 3.08 |
INF. ACQUISITION: INTERNET = 1–2 times per year | 3.43 | 4.22 | 3.47 | 3.42 | 3.26 |
INF. ACQUISITION: INTERNET = 3–6 times per year | 3.32 | 4.14 | 3.3 | 3.49 | 3.42 |
INF. ACQUISITION: INTERNET = 7–12 times per year | 3.13 | 4.06 | 3.1 | 3.64 | 3.63 |
INF. ACQUISITION: INTERNET = at least 13 times per year | 2.91 | 3.93 | 2.85 | 3.82 | 3.86 |
INF. ACQUISITION: LEAFLET = never | 3.28 | 4.12 | 3.27 | 3.67 | 3.44 |
INF. ACQUISITION: LEAFLET = 1–2 times per year | 3.29 | 4.15 | 3.26 | 3.54 | 3.45 |
INF. ACQUISITION: LEAFLET = 3–6 times per year | 3.32 | 4.17 | 3.36 | 3.48 | 3.35 |
INF. ACQUISITION: LEAFLET = 7–12 times per year | 3.32 | 4.15 | 3.29 | 3.41 | 3.42 |
INF. ACQUISITION: LEAFLET = at least 13 times per year; N39 | 2.85 | 3.85 | 3.26 | 3.69 | 3.41 |
INF. SELECTION: radio or TV = never | 3.31 | 4.19 | 3.35 | 3.46 | 3.32 |
INF. SELECTION: radio or TV = 1–2 times per year | 3.31 | 4.12 | 3.29 | 3.55 | 3.45 |
INF. SELECTION: radio or TV = 3–6 times per year | 3.27 | 4.01 | 3.15 | 3.7 | 3.68 |
INF. SELECTION: radio or TV = 7–12 times per year; N46 | 3.17 | 3.91 | 3.11 | 3.67 | 3.71 |
INF. SELECTION: radio or TV = (at least 13 times per year; N18) | 2.44 | 4 | 2.65 | 3.83 | 3.61 |
INF. SELECTION: NEWSPAPER = never | 3.34 | 4.18 | 3.36 | 3.48 | 3.34 |
INF. SELECTION: NEWSPAPER = 1–2 times per year | 3.22 | 4.07 | 3.12 | 3.61 | 3.52 |
INF. SELECTION: NEWSPAPER = 3–6 times per year; N77 | 2.96 | 3.95 | 2.78 | 3.87 | 4 |
INF. SELECTION: NEWSPAPER = 7–12 times per year; N30 | 3.07 | 3.93 | 3.17 | 3.93 | 4.1 |
INF. SELECTION: NEWSPAPER = (at least 13 times per year; N10) | 1.7 | 3.2 | 2.3 | 3.9 | 4.33 |
INF. SELECTION: INTERNET = never | 3.6 | 4.31 | 3.68 | 3.16 | 2.84 |
INF. SELECTION: INTERNET = 1–2 times per year | 3.42 | 4.22 | 3.43 | 3.45 | 3.29 |
INF. SELECTION: INTERNET = 3–6 times per year | 3.21 | 4.06 | 3.22 | 3.6 | 3.57 |
INF. SELECTION: INTERNET = 7–12 times per year | 3.08 | 4.06 | 2.97 | 3.76 | 3.79 |
INF. SELECTION: INTERNET = at least 13 times per year | 2.75 | 3.86 | 2.67 | 4.1 | 4.16 |
INF. SELECTION: FRIENDS = never | 3.61 | 4.35 | 3.65 | 3.3 | 2.92 |
INF. SELECTION: FRIENDS = 1–2 times per year | 3.44 | 4.23 | 3.51 | 3.38 | 3.02 |
INF. SELECTION: FRIENDS = 3–6 times per year | 3.38 | 4.16 | 3.39 | 3.41 | 3.38 |
INF. SELECTION: FRIENDS = 7–12 times per year | 3.07 | 4.04 | 3.01 | 3.73 | 3.77 |
INF. SELECTION: FRIENDS = at least 13 times per year | 2.83 | 3.95 | 2.67 | 4.07 | 4.19 |
INF. SELECTION: HEALTH WORKERS = never | 3.47 | 4.18 | 3.46 | 3.27 | 3.04 |
INF. SELECTION: HEALTH WORKERS = 1–2 times per year | 3.31 | 4.16 | 3.36 | 3.49 | 3.3 |
INF. SELECTION: HEALTH WORKERS = 3–6 times per year | 3.24 | 4.15 | 3.24 | 3.57 | 3.64 |
INF. SELECTION: HEALTH WORKERS = 7–12 times per year | 3.1 | 4.05 | 2.93 | 3.9 | 4.01 |
INF SELECTION: HEALTH WORKERS = at least 13 times per year; N60 | 2.86 | 4.03 | 2.93 | 4.23 | 3.93 |
INF. TRANSMISSION: radio or TV = never | 3.3 | 4.15 | 3.3 | 3.51 | 3.4 |
INF. TRANSMISSION: radio or TV = 1–2 times per year; N48 | 3.17 | 4.04 | 3.24 | 3.76 | 3.36 |
INF. TRANSMISSION: radio or TV = 3–6 times per year; N28 | 3.25 | 3.97 | 3.43 | 3.59 | 3.66 |
INF. TRANSMISSION: radio or TV = (7–12 times per year; N8) | 3.75 | 4.75 | 3.38 | 3.88 | 4.13 |
INF. TRANSMISSION: radio or TV = (at least 13 times per year; N2) | 1.5 | 3 | 1.5 | 5 | 5 |
INF. TRANSMISSION: NEWSPAPER = never | 3.3 | 4.15 | 3.29 | 3.52 | 3.41 |
INF. TRANSMISSION: NEWSPAPER = 1–2 times per year; N28 | 2.96 | 4.03 | 3.14 | 3.69 | 3.07 |
INF. TRANSMISSION: NEWSPAPER = 3–6 times per year; N22 | 3.32 | 3.91 | 3.68 | 3.39 | 3.39 |
INF. TRANSMISSION: NEWSPAPER = (7–12 times per year; N5) | 4.4 | 5 | 4.2 | 4.2 | 4.2 |
INF. TRANSMISSION: INTERNET = never | 3.33 | 4.17 | 3.34 | 3.5 | 3.37 |
INF. TRANSMISSION: INTERNET = 1–2 times per year; N99 | 3.02 | 3.99 | 2.94 | 3.61 | 3.6 |
INF. TRANSMISSION: INTERNET = 3–6 times per year; N33 | 3 | 3.88 | 2.97 | 3.94 | 3.85 |
INF. TRANSMISSION: INTERNET = (7–12 times per year; N18) | 2.89 | 3.5 | 2.71 | 4.28 | 3.94 |
INF. TRANSMISSION: INTERNET = (at least 13 times per year; N6) | 2.5 | 4 | 2.5 | 4.17 | 4.33 |
INF. TRANSMISSION: FRIENDS = never | 3.48 | 4.3 | 3.48 | 3.41 | 3.18 |
INF. TRANSMISSION: FRIENDS = 1–2 times per year | 3.27 | 4.11 | 3.34 | 3.46 | 3.37 |
INF. TRANSMISSION: FRIENDS = 3–6 times per year | 3.1 | 4.03 | 3.03 | 3.66 | 3.76 |
INF. TRANSMISSION: FRIENDS = 7–12 times per year | 3.03 | 3.93 | 2.88 | 3.94 | 3.84 |
INF. TRANSMISSION: FRIENDS = at least 13 times per year; N35 | 2.69 | 3.61 | 2.56 | 4.42 | 4.11 |
INF. TRANSMISSION: ORGANISE EVENT = never | 3.29 | 4.14 | 3.29 | 3.52 | 3.41 |
INF. TRANSMISSION: ORGANISE EVENT = (1–2 times per year; N12) | 3.42 | 4.08 | 3.67 | 3.67 | 3.3 |
INF. TRANSMISSION: ORGANISE EVENT = (3–6 times per year; N12) | 3.42 | 4.15 | 3.38 | 3.38 | 3.38 |
INF. TRANSMISSION: ORGANISE EVENT = (7–12 times per year; N2) | 3.5 | 4 | 3 | 4 | 5 |
INF. TRANSMISSION: ORGANISE EVENT = (at least 13 times per year; N1) | 4 | 5 | 4 | 4 | 2 |
Source(s): Authors' own work
References
Akbulut, E. (2023), “The relationship between communicative actions, behavioural intentions, and corporate reputation in the framework of situational theory of problem solving in a public health crisis”, International Journal of Public Health, Vol. 68, December, pp. 1-9, doi: 10.3389/ijph.2023.1606301.
Altay, S. and Mercier, H. (2020), “Framing messages for vaccination supporters”, Journal of Experimental Psychology: Applied, Vol. 26 No. 4, pp. 567-578, doi: 10.1037/xap0000271.
Atkin, C.K. and Freimuth, V. (2012), “Guidelines for formative evaluation research in campaign design”, in Rice, R.E. and Atkin, C.K. (Eds), Public Communication Campaigns, 4th ed., SAGE Publications, pp. 53-68.
Attwell, K. and Freeman, M. (2015), “I Immunise: an evaluation of a values-based campaign to change attitudes and beliefs”, Vaccine, Vol. 33 No. 46, pp. 6235-6240, doi: 10.1016/j.vaccine.2015.09.092.
Baumgaertner, B., Carlisle, J.E. and Justwan, F. (2018), “The influence of political ideology and trust on willingness to vaccinate”, PLoS One, Vol. 13 No. 1, e0191728, doi: 10.1371/journal.pone.0191728.
Blume, S. (2006), “Anti-vaccination movements and their interpretations”, Social Science and Medicine, Vol. 62 No. 3, pp. 628-642, doi: 10.1016/j.socscimed.2005.06.020.
Burgener, A. (2020), “Enhancing communication to improve patient safety and to increase patient satisfaction”, The Health Care Manager, Vol. 39 No. 3, pp. 128-132, doi: 10.1097/hcm.0000000000000298.
Chen, Y.R.R., Hung-Baesecke, C.J.F. and Kim, J.-N. (2017), “Identifying active hot-issue communicators and subgroup identifiers: examining the situational theory of problem solving”, Journalism and Mass Communication Quarterly, Vol. 94 No. 1, pp. 124-147, doi: 10.1177/1077699016629371.
Chon, M.-G. and Harrell, J.R. (2024), “Building bridges for anti-racism activism: testing situational theory of problem solving and problem chain recognition effect”, Public Relations Review, Vol. 50 No. 1, 102425, doi: 10.1016/j.pubrev.2024.102425.
Cintulová, L.L. (2019), “The impact of the emotions that frame mothers' decision-making about the vaccination of toddlers”, Journal of Nursing and Social Sciences Related to Health and Illness original, Vol. 21 No. 2, pp. 189-196, doi: 10.32725/kont.2019.020.
De Figueiredo, A., Johnston, I.G., Smith, D.M.D., Agarwal, S., Larson, H.J. and Jones, N.S. (2016), “Forecasted trends in vaccination coverage and correlations with socioeconomic factors: a global time-series analysis over 30 years”, Lancet Global Health, Vol. 4 No. 10, pp. e726-e735, doi: 10.1016/S2214-109X(16)30167-X.
Dubé, E., Laberge, C., Guay, M., Bramadat, P., Roy, R. and Bettinger, J.A. (2013), “Vaccine hesitancy: an overview”, Human Vaccines and Immunotherapeutics, Vol. 9 No. 8, pp. 1763-1773, doi: 10.4161/hv.24657.
Dubé, E., Ward, J.K., Verger, P. and MacDonald, N.E. (2021), “Vaccine hesitancy, acceptance, and anti-vaccination: trends and future prospects for public health”, Annual Review of Public Health, Vol. 42 No. 1, pp. 175-191, doi: 10.1146/annurev-publhealth-090419-102240.
Eskola, J., Duclos, P., Schuster, M. and MacDonald, N.E. (2015), “How to deal with vaccine hesitancy?”, Vaccine, Vol. 33 No. 34, pp. 4215-4217, doi: 10.1016/j.vaccine.2015.04.043.
Fine, P., Eames, K. and Heymann, D.L. (2011), “Herd immunity: a rough guide”, Clinical Infectious Diseases, Vol. 52 No. 7, pp. 911-916, doi: 10.1093/cid/cir007.
Gostin, L.O., Hodge, J.G. Jr, Bloom, B.R., El-Mohandes, A., Fielding, J., Hotez, P., Kurth, A., Larson, H.J., Orenstein, W.A., Rabin, K., Ratzan, S.C. and Salmon, D. (2020), “The public health crisis of underimmunisation: a global plan of action”, The Lancet Infectious Diseases, Vol. 20 No. 1, pp. e11-e16, doi: 10.1016/S1473-3099(19)30558-4.
Grunig, J.E. (1989), “Publics, audiences, and market segments: segmentation principles for campaigns”, in Salmon, C.T. (Ed.), Information Campaigns: Balancing Social Values and Social Change, Sage, Newbury Park, pp. 199-228.
Grunig, J.E. (1997), “A situational theory of publics: conceptual history, recent challenges and new research”, in Moss, D., MacManus, T. and Verčič, D. (Eds), Public Relations Research: an International Perspective, International Thomson Business Press, London, pp. 3-46.
Grunig, J.E. (2005), “Situational theory of publics”, in Heath, R.L. (Ed.), Encyclopedia of Public Relations, SAGE Publications, London, pp. 778-780.
Grunig, J.E. (2013), “Situational theory of publics”, in Heath, R.L. (Ed.), Encyclopedia of Public Relations, SAGE Publications, London, pp. 834-836.
Guidry, J.P.D., Carlyle, K., Messner, M. and Jin, Y. (2015), “On pins and needles: how vaccines are portrayed on Pinterest”, Vaccine, Vol. 33 No. 39, pp. 5051-5056, doi: 10.1016/j.vaccine.2015.08.064.
Hak, E., Schönbeck, Y., Melker, H.D., Essen, G.V. and Sanders, E. (2005), “Negative attitude of highly educated parents and health care workers towards future vaccinations in the Dutch childhood vaccination program”, Vaccine, Vol. 23 No. 24, pp. 3103-3107, doi: 10.1016/j.vaccine.2005.01.074.
Ismail, I.B., Sabran, R. and Mohamed Ariffin, M.Y.B. (2017), “Study of situational theory of problem solving (STOPS) in conceptualizing farmer’S response towards insufficient information delivery in Malaysia”, Humanities and Social Sciences Reviews, Vol. 5 No. 2, pp. 124-133, doi: 10.18510/hssr.2017.528.
Jiang, H., Kim, J.-N., Liu, B. and Luo, Y. (2017), “The impact of perceptual and situational factors on environmental communication: a study of citizen engagement in China”, Environmental Communication, Vol. 13 No. 5, pp. 582-602, doi: 10.1080/17524032.2017.1346517.
Johnson, R.A. and Wichern, D.W. (2013), Applied Multivariate Statistical Analysis, 6th ed., Pearson Education, Pearson New International Edition.
Kim, J.-N. (2013), “Situational theory of problem solving”, in Heath, R.L. (Ed.), Encyclopedia of Public Relations, SAGE publications, London, pp. 831-834.
Kim, Y. (2016), “Understanding publics' perception and behaviors in crisis communication: effects of crisis news framing and publics ’acquisition, selection, and transmission of information in crisis situations situations”, Journal of Public Relations Research, Vol. 28 No. 1, pp. 35-50, doi: 10.1080/1062726x.2015.1131697.
Kim, J. (2018), “Why do people take to the streets? Understanding the multidimensional motivations of protesting publics”, Public Relations Review, Vol. 44 No. 4, pp. 501-513, doi: 10.1016/j.pubrev.2018.05.002.
Kim, J.-N. and Grunig, J.E. (2011), “Problem solving and communicative action: a situational theory of problem solving”, Journal of Communication, Vol. 61 No. 1, pp. 120-149, doi: 10.1111/j.1460-2466.2010.01529.x.
Kim, J.-N. and Lee, S. (2014), “Communication and cybercoping: coping with chronic illness through communicative action in online support networks”, Journal of Health Communication, Vol. 19 No. 7, pp. 775-794, doi: 10.1080/10810730.2013.864724.
Kim, J.-N. and Rhee, Y. (2011), “Strategic thinking about employee communication behavior (ECB) in public relations: testing the models of megaphoning scouting effects in Korea”, Journal of Public Relations Research, Vol. 23 No. 3, pp. 243-268, doi: 10.1080/1062726x.2011.582204.
Kim, S., Krishna, A. and Dhanesh, G. (2019), “Economics or ethics? Exploring the role of CSR expectations in explaining consumers' perceptions, motivations, and active communication behaviors about corporate misconduct”, Public Relations Review, Vol. 45 No. 1, pp. 76-87, doi: 10.1016/j.pubrev.2018.10.011.
Kim, J.-N., Ni, L. and Sha, B.L. (2008), “Breaking down the stakeholder environment: explicating approaches to the segmentation of publics for public relations research”, Journalism and Mass Communication Quarterly, Vol. 85 No. 4, pp. 751-768, doi: 10.1177/107769900808500403.
Kim, J.-N., Grunig, J.E. and Ni, L. (2010), “Reconceptualizing the communicative action of publics: acquisition, selection, and transmission of information in problematic situations”, International Journal of Strategic Communication, Vol. 4 No. 2, pp. 126-154, doi: 10.1080/15531181003701913.
Kim, J.-N., Shen, H. and Morgan, S.E. (2011), “Information behaviors and problem chain recognition effect: applying situational theory of problem solving in organ donation issues”, Health Communication, Vol. 26 No. 2, pp. 171-184, doi: 10.1080/10410236.2010.544282.
Kim, J.-N., Tam, L. and Kim, S. (2014), “Segmentation: health campaigns”, in Thompson, T.L. (Ed.), Encyclopedia of Health Communication, SAGE, Thousand Oaks, pp. 1223-1226.
Kim, S., Tam, L. and Bach, S. (2018), “Understanding the effects of perceived ethics failure, compassionate leadership, and communication strategy on anti-government sentiment”, Journal of Public Affairs, Vol. 18 No. 4, p. e1848, doi: 10.1002/pa.1848.
Kraigher, A. (Ed.) (2018), Cepljenje: Stališča in Odnos Ključnih Javnosti Do Cepljenja V Sloveniji, Nacionalni inštitut za javno zdravje, Ljubljana, available at: https://www.nijz.si/sites/www.nijz.si/files/publikacije-datoteke/monografija_8_07-03-18.pdf
Krishna, A. (2018), “Poison or prevention? Understanding the linkages between vaccine-negative individuals' knowledge deficiency, motivations, and active communication behaviors”, Health Communication, Vol. 33 No. 9, pp. 1088-1096, doi: 10.1080/10410236.2017.1331307.
Larson, H.J. (2018), “Politics and public trust shape vaccine risk perceptions”, Nature Human Behaviour, Vol. 2 No. 5, p. 316, doi: 10.1038/s41562-018-0331-6.
Leask, J., Kinnersley, P., Jackson, C., Cheater, F., Bedford, H. and Rowles, G. (2012), “Communicating with parents about vaccination: a framework for health professionals”, BMC Pediatrics, Vol. 12 No. 1, p. 154, doi: 10.1186/1471-2431-12-154.
Lee, Y. (2019), “Crisis perceptions, relationship, and communicative behaviors of employees: internal public segmentation approach”, Public Relations Review, Vol. 45 No. 4, 101832, doi: 10.1016/j.pubrev.2019.101832.
Lee, H., Oshita, T., Oh, H.J. and Hove, T. (2014), “When do people speak out? Integrating the spiral of silence and the situational theory of problem solving”, Journal of Public Relations Research, Vol. 26 No. 3, pp. 185-199, doi: 10.1080/1062726x.2013.864243.
Li, J.-Y., Kim, J.K., Overton, H., Bhalla, N., Zhang, N., Choi, M. and Moon, W.K. (2019), “What shapes environmental responsibility perceptions? Measuring value orientations as a predictor of situational motivations and communicative action”, International Journal of Strategic Communication, Vol. 13 No. 3, pp. 214-232, doi: 10.1080/1553118X.2019.1612407.
Liu, B.F., Xu, S., Lim, J. and Egnoto, M. (2019), “How publics' active and passive communicative behaviors affect their tornado responses: an integration of STOPS and SMCC”, Public Relations Review, Vol. 45 No. 4, 101831, doi: 10.1016/j.pubrev.2019.101831.
Lovari, A., Kim, S., Vibber, K. and Kim, J.-N. (2011), “Digitisation's impacts on publics: public knowledge and civic conversation”, PRism, Vol. 8 No. 2, pp. 1-20.
Lovari, A., Martino, V. and Kim, J.-N. (2012), “Citizens' relationships with a municipality and their communicative behaviors in negative civic issues”, International Journal of Strategic Communication, Vol. 6 No. 1, pp. 17-30, doi: 10.1080/1553118x.2011.634870.
MacDonald, N.E. and the SAGE Working Group on Vaccine Hesitancy (2015), “Vaccine hesitancy: definition, scope and determinants”, Vaccine, Vol. 33 No. 34, pp. 4161-4164, doi: 10.1016/j.vaccine.2015.04.036.
McKeever, B.W., McKeever, R., Holton, A.E. and Li, J.-Y. (2016), “Silent majority: childhood vaccinations and antecedents to communicative action”, Mass Communication and Society, Vol. 19 No. 4, pp. 476-498, doi: 10.1080/15205436.2016.1148172.
Moss, H.B., Kirby, S.D. and Donodeo, F. (2009), “Characterizing and reaching high-risk drinkers using audience segmentation”, Alcoholism: Clinical and Experimental Research, Vol. 33 No. 8, pp. 1336-1345, doi: 10.1111/j.1530-0277.2009.00963.x.
Neumann-Böhme, S., Varghese, N.E., Sabat, I., Barros, P.P., Brouwer, W., van Exel, J., Schreyögg, J. and Stargardt, T. (2020), “Once we have it, will we use it? A European survey on willingness to be vaccinated against COVID-19”, The European Journal of Health Economics, Vol. 21 No. 7, pp. 977-982, doi: 10.1007/s10198-020-01208-6.
Noar, S.M. (2006), “A 10-year retrospective of research in health mass media campaigns : where do we go from here?”, Journal of Communication, Vol. 11 No. 1, pp. 21-42, doi: 10.1080/10810730500461059.
Olowo, E.O., Hasan, H. and Wirza, J. (2020), “Information seeking and trust on vaccination practices”, Human Communication, Vol. 1 No. 2, pp. 126-138.
Oraby, T., Thampi, V. and Bauch, C.T. (2014), “The influence of social norms on the dynamics of vaccinating behaviour for paediatric infectious diseases”, Proceedings of the Royal Society B Biological Sciences, Vol. 281 No. 1780, 20133172, doi: 10.1098/rspb.2013.3172.
Poroli, A. and Huang, L.V. (2018), “Spillover effects of a university crisis: a qualitative investigation using situational theory of problem solving”, Journalism and Mass Communication Quarterly, Vol. 95 No. 4, pp. 1128-1149, doi: 10.1177/1077699018783955.
Ramanadhan, S. and Viswanath, K. (2006), “Health and the information nonseeker: a profile”, Health Communication, Vol. 20 No. 2, pp. 131-139, doi: 10.1207/s15327027hc2002.
Ratzan, S.C., Bloom, B.R., El-Mohandes, A., Fielding, J., Gostin, L.O., Hodge, J.G., Hotez, P., Kurth, A., Larson, H.J., Nurse, J., Omer, S.B., Orenstein, W.A., Salmon, D. and Rabin, K. (2019), “The Salzburg statement on vaccination acceptance”, Journal of Health Communication, Vol. 24 No. 5, pp. 581-583, doi: 10.1080/10810730.2019.1622611.
Ricciardi, W., Boccia, S. and Siliquini, R. (2018), “Moving towards compulsory vaccination: the Italian experience”, The European Journal of Public Health, Vol. 28 No. 1, pp. 2-3, doi: 10.1093/eurpub/ckx214.
Ropeik, D. (2013), “How society should respond to the risk of vaccine rejection”, Human Vaccines and Immunotherapeutics, Vol. 9 No. 8, pp. 1815-1818, doi: 10.4161/hv.25250.
Shen, H., Xu, J. and Wang, Y. (2019), “Applying situational theory of problem solving in cancer information seeking: a cross-sectional analysis of 2014 HINTS Survey”, Journal of Health Communication, Vol. 24 No. 2, pp. 165-173, doi: 10.1080/10810730.2019.1587111.
Tufte, E.R. (1997), Visual and Statistical Thinking: Displays of Evidence for Making Decisions, Graphics Press, Connecticut.
Tufte, E.R. (2011), The Visual Display of Quantitative Information, 2nd ed., Graphics Press, Connecticut.
Tufte, E.R. (2020), Seeing with Fresh Eyes: Meaning, Space, Data, Truth, Graphics Press, Connecticut.
Vrdelja, M. (2023), “Vpliv zaupanja ter afektivnih in kognitivnih dejavnikov na komuniciranje mater v povezavi s cepljenjem (The influence of trust and affective and cognitive factors on communication of mothers in relation to vaccination)”, Doctoral dissertation, Repository of the University of Ljubljana, available at: https://repozitorij.uni-lj.si/IzpisGradiva.php?lang=eng&id=146944
Won, J., Hon, L. and Lee, A.R. (2018), “Predicting public interest issue campaign participation on social media”, The Journal of Public Interest Communications, Vol. 2 No. 1, p. 129, doi: 10.32473/jpic.v2.i1.p129.
Xu, X., Li, H. and Shan, S. (2021), “Understanding the health behavior decision-making process with situational theory of problem solving in online health communities: the effects of health beliefs, message source credibility, and communication behaviors on health behavioral intention”, International Journal of Environmental Research and Public Health, Vol. 18 No. 9, p. 4488, doi: 10.3390/ijerph18094488.
Yaqub, O., Castle-Clarke, S., Sevdalis, N. and Chataway, J. (2014), “Attitudes to vaccination: a critical review”, Social Science and Medicine, Vol. 112, pp. 1-11, doi: 10.1016/j.socscimed.2014.04.018.
Acknowledgements
Survey questionnaire development and data collection were founded by The Slovenian Research Agency and Ministry of Health Slovenia, project number L7-6806 (see national database SICRIS https://cris.cobiss.net/ecris/si/en/project/9422).
The data were collected in accordance with Slovenian legislation and with the permission of the Commission for Medical Ethics of the Republic of Slovenia (decision number: 127/03/14).