Communication Matters:
A Study Of Opinion-Leader LinkedIn Post Characteristics and
Social Engagement
Ieva Innusa
Student number: 11007591
Graduate School of Communication
Submitted to: Dr. Pytrik Schafraad
MASTER THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR A DEGREE IN
COMMUNICATION SCIENCE: CORPORATE COMMUNICATIONS MASTER OF SCIENCE
UNIVERSITY OF AMSTERDAM AMSTERDAM
Abstract
In a world in which digital technology is rapidly advancing, individuals, as well as organizations, are progressively using social media as an important competitive tool. The benefits of ‘word-of-mouth’ communication have shown that interpersonal communication can be particularly beneficial for organizations’ maximizing their outcomes. The influence of opinion-leaders or those individuals, like chief executives, that are well informed within the field, and are positioned at the strategic location of the network are likely to change opinions, decisions, and choices of people around them. This research aimed to explore what are opinion-leader post characteristics and to what extent do they relate to social engagement on LinkedIn.
This paper reports on a study that investigated the effects of post length, mentioning the company, communication of personal information, and post sentiment. Using data from 101 chief executives, 1633 LinkedIn posts were analyzed. Results showed that the post length is a good and positive predictor for social engagement, and mentioning the company has no significant effect on social engagement. As for including personal information the findings did not reveal any significant relationship with social engagement; and lastly, after testing the relationship between four different types of sentiment – negative, neutral, positive, and mixes – only negative sentiment revealed a significant positive relationship with social engagement. In sum, the results suggest that the way opinion-leaders communicate on LinkedIn can predict social engagement, but to have a more understanding, more insight is needed to identify other communication characteristics.
1
Introduction
Influenced by changes in technology, over the past decade we have witnessed a monumental shift in business communication (Bonsón & Bednárová, 2013). In a world in which digital technology is rapidly advancing, individuals, as well as organizations, are increasingly using social media networks as an important tool for communication (Heath at al., 2013). Since stakeholder engagement is seen as a fundamental aspect of a business, the shift to a technology-based society has created new opportunities for companies to communicate more effectively and efficiently (Bonsón & Bednárová, 2013).
Not only can social media be particularly beneficial for enabling organizations to actively engage with their key stakeholders (including their potential employees, shareholders, customers, competitors), but based on the literature on strategic engagement, social media affords the possibility for organizations and their leaders to communicate their messages to a much wider audience (Heath at al., 2013; Gill, 2015; Prince & Rogers, 2012). As such, depending on the purpose of communication, organizations and/or their representatives may use different forms of strategic engagement to maximize the outcome.
Through “the process of developing and delivering an organization’s message by using narration about people, the organization, the past, visions for the future, social bonding and work itself to create a new point-of-view or reinforces an opinion or behavior,” corporate storytelling on social media networks can be a powerful public relations (PR) strategy, building stronger engagement within as well as outside the
organizations’ boarders (Gill 2015, p.665). While a message communicated via the official organization’s social media account can prove to be effective, it, however, may not be nearly as effective as communicated interpersonally (Araujo et al., 2017).
Literature shows that interpersonal communication or ‘word-of-moth’ can be particularly effective if communicated by opinion-leaders – those individuals that, due to their unique position in the network and their expertise in the field, are respected within the community (Araujo et al., 2017). As such, because these individuals are recognized as authorities in the field, they would likely receive more public attention on social media, causing opinion-leaders to seen as influential figures (Efimova & Grudin, 2007). Ultimately, if communicated well ‘word-of-mouth’ can be a successful tool promoting business development as well as market expansion (Brooks, 1957).
All in all, in a world in which the technology landscape is advancing rapidly, it has given opinion-leaders a great opportunity to act as messengers delivering information, while potentially reaching the largest masses of people. As such, considering all factors, it can be argued that opinion-leaders, such as company’ directors, CEOs and other similar senior position professionals, on social media networking platforms are creating, establishing and maintaining a function of corporate branding (Abratt & Kleyn, 2012; Hamzah, Alwi & Othman, 2014). By establishing a credible profile and image, opinion-leaders are thus partially taking over the role of what traditionally been considered as advertising. In turn, making the opinion-leader communication on social media platforms particularly important field of study.
However, despite the growth of technology as well as the growing influence of opinion-leaders, surprisingly little empirical research has aimed at investigating their use
3 of social media. Moreover, not only better understanding of opinion-leader communication on social media would contribute to the already existing literature on corporate communication, particularly, how corporate communication is changing over time (including PR, corporate branding, advertising); but also, companies, as well as opinion-leaders, could benefit from understanding what factors are associated with the highest social engagement (Grunig & Grunig, 1992; Abratt & Kleyn, 2012; Sundar & Limperos, 2013). Social engagement describing the overall social participation or interaction based on the number of likes and comments can be seen as an indicator of how large of an audience the post has reached. Assuming that the more people are reached, the better the expected outcome could be, social engagement can be considered a particularly vital aspect. Taking into account the algorithms, which social media platforms utilize, including LinkedIn, to promote both useful contents, and social interaction, results of this study would allow opinion-leader to use the findings as the guidelines for creation of more engaging posts, allowing them to be consequently more visible to wider audience for a longer period (Sehl, 2019; Foote, 2019). In other words, as it is safe it concludes that based on the logic of algorithms, those posts with a higher number of social engagement (likes and comments) would have a much higher lifespan, and thus, would appear on the follower timelines more frequently for a longer period. As such, understanding what factors are associated with higher social engagement is essential.
Since there have been no studies previously done specifically with regards to characteristics of company leader (chief executive and founder) online posts, this paper aims to shed light on their communication on online social networking platform
LinkedIn. Based on the fact that it is the world’s largest professional network, as well as affords the possibility to both observe and engage with each other, LinkedIn is considered the perfect medium for the study of opinion-leader computer-mediated communication. Additionally, since opinion-leaders are viewed as those individuals “positioned at the strategic location of the network into which useful information and resources flow, their remarks are likely to be worthy of gaining attention”, this exploratory study aims to shed light particularly on the communication of those opinion-leaders of high ranking positions, like company chief executive officers (CEOs), co-founders, and/or founders. (Choi, 2015, p.70). As such, aiming to shed light on opinion-leader communication within social (media) networks, the following research question is proposed:
What are the opinion-leader post characteristics and to what extent do they relate to social engagement on LinkedIn?
Theoretical framework
According to Grunig and Grunig (1992), “excellent public relations departments practice the two-way symmetrical model of public relations” (p.290). With the presence of new technologies, social media platforms, and attributes companies can practice the two-way symmetrical model much easier than in the past. Although, the model has also been criticized for its accuracy and limited possibilities for generalization, the two-way symmetrical communication model is nonetheless, considered not only an ethical
5 approach to public relations (PR) but also due to interactivity, the model allows for organizations to be more effective (Leichty & Springston, 1993; Grunig & Grunig, 1992, p.307). Grunig and Grunig (1992) indicate the core underlying assumptions of the two-way symmetrical model are - communication of truth; possibility of interpreting the client and public to one another; possibility for stakeholders to understand the viewpoints of each other; and understanding as the principal objective of public relations rather than persuasion (p.289). Moreover, as the model emphasizes the presence of feedback, the communication is considered a two-way process, as opposed to a one-way process (Ibid.).
The following chapter discusses if posts on the social media platform LinkedIn delivered by those opinion-leaders in the high ranking positions in organizations, fits the premises of the two-way symmetrical communication’s model. Three fundamental aspect, which allows the PR concept to take place are further discussed:
1. Social media and social engagement 2. The concept of opinion-leaders;
3. Corporate branding and impression management.
Social media and social engagement
Over the past decade, social media have proven to be a valuable asset not only because they satisfy the needs of users, but also because they can facilitate an interactive audience for advertising and market benefits of an organization (Di Gangi & Wasko, 2016).
Kaplan and Haenlein (2010) have defined social media as “a group of Internet-based applications that build on the ideological and technological foundations of Web 2.0, and that allow the creation and exchange of User Generated Content” (Ibid., p.2). While previously the notion of ‘media’ was associated with mass communication mediums such as radio, television, films, books, and newspapers, the contemporary academic conception of media also includes the generation of new communication technologies (from devices, like smartphones and robots, to channels, like Facebook and LinkedIn, which allows users to communicate with each other) (Sundar & Limperos, 2013). Additionally, nowadays, social media are considered to be more accessible, more interactive, and more community-like as compared to traditional media (Choi 2015, p.697). Due to the value of co-creation, which social media affords, it allows users to contribute, retrieve and explore the content with others (Ibid.).
Establishing interpersonal social connections, social engagement has been considered an important component of people's day-to-day lives (Park, 2009). Generally, social engagement can be defined as “a dynamic multi-dimensional relational concept featuring psychological and behavioral attributes of connection, interaction, participation, and involvement, designed to achieve or elicit an outcome on an individual, organization, or social levels” (Johnston, 2018, p.19). Johnston (2018) highlights that the value of social engagement for organizations “emerges as an outcome from engaged social relationships,” represented in a form of positive image and reputation, customer loyalty, or perceptions of being a socially responsible corporate citizen (p.20). For stakeholders (such as consumers, observers, users, and general audience), on the other hand, the advantage of being engaged suggests an outcome that is positively associated with their
7 experience, including, new insights or knowledge, stronger feelings, and/or a desire to act toward a product or service (Johnston, 2018).
The literature on social engagement separates three forms of engagement - cognitive, affective and behavioral engagement (Johnston, 2018). Cognitive engagement describes interest and involvement in a topic (Ibid.). For cognitive engagement attention, thinking and processing skills are considered as crucial aspects to understand and expand knowledge (Ibid.). Furthermore, effective engagement includes displaying emotional reactions (including fear, anger, joy) (Ibid.). However, as the behavioral engagement embodies concepts of action, collaboration, and participation (including “likes” and “comments” on LinkedIn), this study aims specifically exploring the concept of behavioral engagement (Ibid.). Overall, the literature shows that social engagement have beneficial effects on health and psychological well-being, including higher level of happiness (Graney, 1975; Thompson & Heller, 1990), increased quality of life (Thompson & Heller, 1990), as well as decreased rates of mortality (Rozzini, Bianchetti, Franzoni, Zanetti, & Trabucchi, 1991).
Studying consumer opinion platforms, Hennig-Thurau with colleagues (2004) found that because individuals desire social interaction, it creates an ideal opportunity for individuals, like opinion-leaders, to contribute to the community. While people desire social interaction it can be argued that social interaction (or social engagement) on social media platforms can only be achieved if a suitable space for social engagement is present. Considering that participation has been described as “the active involvement by community members to jointly develop meanings and negotiate solutions to an issue through dialogic processes in interaction with the focal organization,” for social
engagement to take place on social media, posts needs to be rich in its content, which is like to create a more suitable space for audience to interact (Johnston, 2018, p.24). Hence, I propose the following research question:
RQ1 = Does the length of opinion-leader posts, measured by the total amount of characteristics used in the post affect the social engagement on LinkedIn?
Opinion-leaders
Lazarsfeld, Berelson, and Gaudet (1948) first introduced ‘opinion-leaders’ as part of a two-step flow of communication model. The model suggests that information flows from the media to opinion-leaders and then to the general public (Lazarsfeld, Berelson, & Gaudet, 1948). Since opinion-leaders appear to be more exposed to media and their messages, their opinions, attitudes and personal influence mediates the way information is further perceived by the public (Katz & Lazarsfeld, 1955).
In essence, an opinion-leader can be defined as:
“Authorities on, or as simply well‐informed about, certain topics; they influence those
around them by passing on information and opinions that may result in learning, decisions, choices, and opinion changes among those they are in contact with.” (Case, Johnson, Andrews, Allard & Kelly, 2004, p.661)
9 Furthermore, Katz and Lazarsfeld (1995) highlight that opinion-leaders come to light from the give-and-take of information in daily interpersonal interactions. In this sense, opinion-leaders can be seen as a social construct that is based on interpersonal relations. Although opinion-leader’ influence is informal, they play a significant role in the stakeholder information acquisition process, potentially affecting people's decision making, which can thus, influence the outcomes of marketing strategies (Lyons & Henderson 2005).
Diving deeper, Katz (1957) specified that the classic concept of opinion leadership could be associated to ‘‘(1) to the personification of certain values (who one is); (2) to competence (what one knows); and (3) to strategic social location (whom one knows)’’ (Katz, 1957, p. 73). Essentially, to be positioned at the center of a specific social network, opinion-leaders would need to have expertise over the field, subject or matter, which one personifies or is associated with. Furthermore, Brooks (1957) indicates that the distinguishing characteristic of the opinion-leader is that “he is sought by others for information and advice regarding the field in which he is a leader” (Brooks, 1957, p.157).
Opinion-leaders and ‘word-of-mouth’ processes
Given the fact that opinion-leaders have expertise in a field as well as advantages strategic social location within a specific community, their effectiveness in ‘word-of-mouth’ (WOM) processes is larger (Araujo et al., 2017). Ladhari (2007) defines WOM as “informal person-to-person communication between a perceived non-advertising
communicator and a consumer about ownership, or characteristics of a brand, a product, a service, an organization or a seller” (p.1093). Not only WOM processes can be seen as particularly beneficial for organizational purposes (for example, selling the effort of the firm), but it can also be considered as part of a viral marketing strategy (Brooks, 1957; Kempe, Kleinberg & Tardos, 2003). Brooks (1957) highlights that because of increased influence on the community in which opinion-leaders are situated, it could ultimately make an impact on the customer (stakeholder) behavior.
Opinion-leaders in computer-mediated environment
Looking more specifically at the computer-mediated environment, digital technology allows easier communication between message senders and target audience (Choi 2015, p.697). In light of Web development e-commerce has become a strategic emphasis on business. Such development has forced WOM to be re-conceptualized as ‘electronic word-of-mouth’ (eWOM) (Kim, Kandampully & Bilgihan, 2018). While opinion-leaders are usually not directly involved in e-commerce, like selling the goods or services by using the Internet, opinion-leader communication can, nonetheless, contribute to influencing the public, which can ultimately shape the consumer behavior. Unlike the traditional concept of WOM, the influence of eWOM is considered to be much stronger, reaching a much larger audience, and lasting a longer period (since WOM influence diminishes quicker over time and distance) (Duan & Whinston, 2008; Xie, Miao, Kuo & Lee, 2011). Although Hennig-Thurau with colleagues (2004) defines eWOM communication as “any positive or negative statement made by potential, actual, or
11 former customers about a product or company, which is made available to a multitude of people and institutions via the Internet,” it can be argued that the concept of eWOM can also be applied to other interpersonal communication instances facilitated on the Internet, including those statements communicated by opinion-leaders (p.39).
Considering that the WOM consists of informal communication about either the “ownership, characteristics of brand, a product, a service, an organization or a seller”, it is likely that during the message the communicator will disclose the name of the company (Ladhari, 2007, p.1093). Moreover, the likelihood of a company leader mentioning the name of the company, which they are associated with, becomes stronger because this study aims exploring posts delivered on LinkedIn – “the world’s largest professional network” (LinkedIn, 2019). As such based on the underlying factors of WOM and the unique leader position within the network, I expect opinion-leaders to communicate (mention) the name of their company in their LinkedIn posts more frequently. Moreover, as company’s name alone mentioned in the opinion-leader posts is unlikely to drive social engagement, I propose the following two hypotheses:
H1a= Company’s names will frequently be mentioned in opinion-leader1 posts
on LinkedIn.
1
This paper focuses only on those opinion-leaders that are associated with higher-ranking positions (including, (co-)founders, directors, and chiefs), and/or are considered as leaders of the company.
H1b= Mentioning of company’s name in opinion-leader posts on LinkedIn
will have no correlation with social engagement.
Furthermore, considering that the “online WOM usually involves personal experiences and opinions transmitted through the written word,” it can be argued that opinion-leaders are likely to include their personal experiences, insights, or other personally related information in their posts (Sun, 2006, p.1106). Ryan and Deci (2000) highlighted that relatedness is a crucial innate psychological need of a human being, which when satisfied yields several positive outcomes, such as enhanced self-motivation and mental health. As such, communicating one’s personal experience can be particularly beneficial for the opinion-leaders, possibly, drawing more attention and social engagement to the post (Ryan & Deci, 2000). Hence, I propose the following hypotheses:
H2a = Personal information will frequently be mentioned in opinion-leader
posts on LinkedIn.
H2b = Posts that contain personal information will be positively associated
with higher social engagement as oppose to posts that do not include personal information.
Corporate branding and impression management
13 (Cottan-Nir & Lehman-Wilzig, 2018). Since corporate branding is considered one of the key drivers of an organization’s reputation management, it is vital to maintain communication and favorable relationship with those groups upon which the company is dependent (Abratt & Kleyn, 2012).
A well maintained corporate brand could ultimately create a supportive and advantageous long-term relationship with its stakeholders (Cottan-Nir & Lehman-Wilzig, 2018). Concerning corporate branding on digital media, “the emergence of the Internet has had a major impact on building corporate brand” (Hamzah, Alwi & Othman, 2014, p.2301). Through interaction, the Internet affords the possibility of greater involvement of brand experience and engagement in communities. However, due to the characteristics of opinion-leaders (e.g. their unique strategic location, their expertise over the field, and personal associations), companies are likely to use those opinion-leaders, which are associated with a higher social status (e.g. CEOs, CTOs, etc) as a representative figure of the company (Abratt & Kleyn, 2012). Since reputation of a company is built over time through the interactions between the company and its stakeholders, on social network such as LinkedIn, opinion-leaders may communicate company’s unique business model through their profiles (Abratt & Kleyn, 2012).
According to Bolino et al. (2016), impression management is a conscious and strategic behavior, which people (employees) use to shape the way they are seen by others, usually, targeted individuals and/or a group of people. Moreover, it is considered that the process of impression management is conscious and strategic, in that individuals may deliberately try to manage a particular image (Ibid.). While Bolino et al. (2016) focus on offline, person-to-person impression management, the literature indicates that
individuals are as likely to manage their impressions on digital context as they are in real life (Berkelaar, 2017). Berkelaar (2017) highlights that depending on how, where and when information is presented may affect how the audience perceives an individual. Moreover, the literature on the computer-mediated environment (Twitter and Facebook) shows that factors like personality (Rosenberg & Egbert, 2011), connections (Walther, 2011), as well as the design of the profile page (Gibbs et al., 2013) may contribute to impression management. Thus, to create the most desirable image, opinion-leaders are likely to be very considerate of what information to present, how and when, to reach the most desirable result.
As such, since opinion-leaders are likely to be seen as company representatives trying to sustain company’s brand and it’s (already created) reputation, as well as taking into account their likelihood to strategically manage their public appearance, it is reasonable to assume that not only the sentiment of their posts on LinkedIn will have a positive nature, but also those posts of positive nature will draw higher level of social engagement, as oppose to neutral, negative or mixed sentiment. Hence, I propose the following hypothesis:
H3a= Opinion-leaders will communicate positive sentiment posts on LinkedIn
more frequently.
H3b= Opinion-leader posts on LinkedIn whose sentiment is positive will have
higher social engagement, as oppose to posts that have either neutral, negative or mixed sentiment.
15
Methods
This section aims to introduce the data that were used for the study as well as how variables and the social media network LinkedIn, were selected.
Sample Selection
To answer the research question and hypothesis presented in an earlier chapter, a quantitative content analysis of opinion-leader (company’ leader) LinkedIn posts was conducted. After a careful selection of active 136 opinion leaders in the higher-ranking positions, 2000 LinkedIn posts were collected for the study (see Appendix B). Opinion-leaders were selected either based on Fortune 500 listed companies or companies that have partnered up with the Connecting Circles, a consultancy firm based in Amsterdam. Moreover, only those posts written in English were selected and studied. In order to increase the generalizability of the results, company leaders from the most diverse industries and fields were selected (e.g. sports, technology, health, finances, insurance, entertainment, pharmacy, banking, science, etc). In the end, due to a time limitation, by examining approximately 15 the most recent posts per person, a convenience sample of n=1633 LinkedIn posts of 101 leaders were divided among three coders (including myself).
Since LinkedIn is “the world's largest professional network with nearly 660+ million users in more than 200 countries and territories worldwide,” it was selected as the medium for the analysis (LinkedIn, 2019). Founded by Reid Hoffman, LinkedIn was
officially launched on May 5, 2003 (Ibid.). Ever since, it has been considered the fastest-growing professional social community (Bonsón & Bednárová, 2013). This platform allows not only their members to create, maintain and share their professional identity online, but also create and participate in the network (Ibid.). Moreover, since the mission of LinkedIn is to “connect the world’s professionals to make them more productive and successful”, the platform affords the possibility to maintain the communication, trade information as well as easily reach others (LinkedIn, 2019; Papacharissi, 2009).
Literature shows the use of digital communication technologies, such as social networking platforms like LinkedIn, has reshaped the working practices of multiple business industries, including public relations, human resources, technology (Quinton & Wilson, 2016). However, due to its benefits, LinkedIn creates a particularly beneficial environment for opinion-leaders. Allowing users to build and manage their professional network's opinion leaders can extend their reach by adding more connections. As such, based on its value, LinkedIn is considered the most fitting medium for this study.
Codebook and Inter-Coder Reliability
To shed light on opinion-leader communication characteristics and social engagement on social media platform LinkedIn, a codebook of 19 different variables was created. Although some variables were automatically obtained from the networking platform LinkedIn (such as length, likes, and comments), other variables had to be coded manually (such as opinionated, broader impact, personal information). Due to the high level of disagreement among the coders and the possibility of different interpretations of
17 variables - seven coder preparations took place in order to improve the agreement. Based on finding consensus and finding the appropriate definitions, the codebook was significantly improved in each of the preparation.
After the coder training was completed, the codebook was tested on a separate data set that consisted of N=500 of the LinkedIn posts (over 30% of the whole dataset). Based on the fact that the inter-coder reliability measure Lotus, “defines the proportion of agreement among all coders without using each individual comparison as a basis for calculation” and “is based on the agreement with the most commonly coded value (MCCV)”, the measure was selected to examine the agreement between all four coders (Fretwurst, 2015, p.2). As the coefficient for all manually coded variables was acceptable, including - Mentions the company – 0.91 (Lotus); Communication of personal information – 0.86 (Lotus); and Sentiment – 0.89 (Lotus), coders were allowed to proceed with the actual coding.
Variables
In order to answer the research questions proposed earlier, as well as to test the hypothesis, this paper focuses on six variables (4 independent variables and 2 dependent variables).
Independent variables
Measured by the total amount of characteristics of the post. As the indented variable ‘length’ was automatically scraped from LinkedIn it is considered 100% accurate indicator of a post.
2. Mentions the company
Code:
0 = No, the post does not include one’s company 1 = Yes, the post does include one’s company
Independent variable ‘mentions the company’ suggests that the text is centered around one’s company. This includes that the coder should be familiar the company that the author of the post is associated with. The post has to include company’s name or refers to the company using words such as “we”, “our”, or “us”, “my team” in order to be coded yes. This includes the full post text with tags and hashtags (#). This variable should be coded “Yes” if it is about any part of the company, including a sub-brand or a department. It can also talk about people joining “our” team, company. It can only be coded “yes”, if “we”, “our”, “us” is clearly related to the company.
3. Communication of personal information
Code:
0 = No, the post does not include one’s personal information 1 = Yes, the post does include one’s personal information.
19 The post is coded yes, if the text is about one’s personal experiences (different from opinion). It can be a personal reflection on their lives’ story, personal career, their individual learning, or other similar information. It can also describe one’s current activity (e.g. places visited, people met) or achievements. Communication of personal information refers to the idea that the story is about one’s own lived experiences, as such one may use 1st person language. However, not all posts written in 1st person language are personal stories/narratives.
4. Sentiment of the post
Code: 0 = Negative 1 = Neutral 2 = Positive 3 = Mixed
Sentiment refers to the overall feeling that is being conveyed in the text. These can be positive, negative, neutral or mixed.
A post that includes an adequate solution, mentions positive trends/events, conveying optimism about the future/topic/self can be defined as a positive post, and therefore should be coded 2.
Negative post has to be focused on problems and/or issues. It can also be critical towards solutions, pessimistic about the future/topic/self, or mentions negative trends/event. Post with such characteristics should be coded 0.
Posts that state facts without sentiment and/or do not match the positive/negative parameters above are neutral posts, and should be coded 1.
However, if a post talks about both positive and negative points on a topic/trend (i.e., tension/paradoxical), it should be coded as mixed (3).
Dependent variables
1. Number of likes by each post; 2. Number of comments by each post.
In this paper, I will use the term social engagement to describe social interaction or social participation based on likes and comments of each post. Since the two-way symmetrical model of public relations as described by Grunig and Grunig (1992) highlights the presence of feedback, likes and comments on social networking platform LinkedIn are elements that symbolize the presence of such occurrence. While, likes can be perceived as a more passive communication of feedback, comments, on the other hand, can be seen as an active form of feedback contribution.
Both, dependent variables - likes and comments - have automatically been collected (scraped) from LinkedIn. Since the data was gathered before the implementation of multiple reactions that LinkedIn now affords (like, love, celebrate, insightful, curious), the number of likes represents a cluster of all of the reactions by the audience.
21 Analytical Strategy
To study the research questions proposed and test the hypothesis, SPSS will be used for further data analysis. Answering the RQ1, or the possible correlation between the independent variable (length) and two dependent variables (likes and comments), linear regression will be applied. Considering that IV and DV are continuous variables, two independent linear regression analyses will be executed.
Furthermore, the first part of hypothesis 1 will be answered by exploring the frequencies and the overall presence of the independent variable (mentioning of the company). However, in order to examine the independent variable’s relationship with both of the dependent variables, the second part of hypothesis 1 will be answered by running the independent-samples T-test. The independent-sample T-test will be particularly beneficial based on the fact that the independent variable is categorical and thus, group statistics could be done.
To answer hypothesis 2a descriptive statistics will be used exploring the frequency of the independent variable. Furthermore, based on similar characteristics of independent variables of hypothesis 1 and hypothesis 2, hypothesis 2b will also be answered by using the independent-samples T-test. The method will allow getting insights into the group statistics (including the mean scores, significance).
Lastly, hypothesis 3 will be answered by firstly exploring the frequencies and the presence of the four different categories of the independent variable (sentiments – negative, neutral, positive and mixed). Secondly, Post hoc analysis will be used to shed
light on where the differences occur between the groups of sentiment. To determine which Post hoc comparison is the most appropriate for comparison analysis, one-way ANOVA will be tested. Lastly, due to data specifics the Post hoc tests – Tukey and LSD - will be carried out to examine the relationship between different types of sentiment and the social engagement (likes and comments).
Results
RQ1: RQ1 proposes a question if the length of the opinion-leader post(s), measured by
the total amount of characteristics used in the posts affects the social engagement on LinkedIn, in terms of likes and comments. Firstly, descriptive statistics found that the smallest post consisted of 7 characteristics, but the longest of 1419 characteristics (M=368.84, SD=271.8360). Furthermore, 91 posts (out of 1633) had no post text at all.
Furthermore, a simple linear regression was calculated to predict number of comments received based on the length of the post. A significant regression equation was found (F(1,1317)=4.015, p< .045), with an R2 of .003. The predicted number of comments is equal to 22.045 + .016 (length) when length is measured in the number of characters. As such, it can be concluded that the number of comments increased by .016 for each character of the post and, thus, the linear regression model is significant.
Secondly, similarly to previous analysis, another simple linear regression was calculated to predict the number of likes posts received based on the length of the post. Similarly to the relationship between comments and length, a significant regression
23 results indicate that the predicted amount of likes is equal to 360.778 + 1.331 (length) when length is measured in characters. Hence, the number of likes increased by 1.331 for each character of the post. As such, similarly to post length and comments received, results for post length and likes were too found significant.
Answering the research question, it can thus be concluded that post length, measured by the total number of characteristics used in the post, is associated with higher social engagement. In other words, the longer the post text is, the more likes and comments the post would receive.
Hypothesis 1a: Hypothesis 1a stated that the company’s name would be
mentioned frequently in opinion-leader posts on LinkedIn. The frequency table showed that from the total of 1633 posts that were coded, one’s company’s name was mentioned in 1049 posts. This means that hypothesis 1a is supported, as the company’s name was being present in the majority (64.2%) of the cases.
Hypothesis 1b: Furthermore, to see if mentioning one’s company’s name within
the post on LinkedIn has any relationship with social engagement (likes and comments), the independent-samples T-test was conducted. Testing the independent variable ‘mentioning the company’ (M=0.64, SD = 0.479), the results showed that for comments the mean comment scores did not differ, t(1631)= 0.519, p= .604. As for the likes, similar results were found, t(1613)= -0.986, p= .324. Considering that the p-value for both variables is > .05, no significance was found. As such, based on the results, it can be concluded that for social engagement, both for likes and for comments, the presence of one’s company’s name does not matter. Hence, H1b is supported.
Hypothesis 2a: Hypothesis 2a stated that personal information would be
mentioned frequently in opinion-leader posts on LinkedIn. Running descriptive statistics, frequency table showed that from the total of 1633 posts that were coded, personal information was mentioned in 461 posts. This means that hypothesis 2a is rejected, as personal information was being present only 28.2% of all posts.
Hypothesis 2b: Descriptive analysis showed that personal information was
communicated in less than 30% of all posts, more specifically, in 480 opinion-leader LinkedIn posts. Furthermore, testing the relationship between communication of personal information (M=0.28, SD=0.452) in opinion-leader LinkedIn posts and social engagement, the independent-samples T-test indicated no significant findings on either of dependent variables (likes: t(1612)= -1.922, p= 0.55; comments: t(1630)= -1.641, p= .101). Results show that in both of the instances p-value is higher > .06. Hence, it can be concluded that mentioning personal information does not lead to a higher level of social engagement, neither for likes nor comments and thus hypothesis 2 is rejected.
Hypothesis 3a: Hypothesis 3 stated that opinion-leaders would communicate
positive sentiment more frequently, as opposed to other types of sentiment (neutral, negative or mixed sentiment). In order to better understand the distribution of sentiments opinion-leaders use in their online communication, descriptive statistics were conducted (Table 1).
Firstly, results found that either of the four sentiments was used in all of the posts (M=1.64, SD=0.687). Furthermore, the positive sentiment was by far the most frequent sentiment posts are communicated in (60.6% of all posts or 990 posts in total). It is followed by neutral (445 posts and 27.3% of all posts), negative sentiment (112 posts,
25 and 6.9% of all posts), and finally, a mixed sentiment with only 86 posts, making it 5.3% of all posts. Thus, based on descriptive statistics, H3a is supported.
Table 1.
Frequency table – Sentiment
Type of Sentiment Frequency Percent (%)
Negative 112 6.9
Neutral 445 27.3
Positive 990 60.6
Mixed 86 5.3
Total 1633 100
Hypothesis 3b: Finally, following up hypothesis 3a, hypothesis 3b suggested that
those opinion-leader’ posts with positive would receive higher social engagement, as opposed to posts with neutral, negative or mixed sentiments. Results found that on average negative sentiment posts received more likes (M=1487.5, SD=2962.818), followed by positive sentiment posts (M=943.06, SD=7403.254), than neutral (M=668.39, SD=1871.057), and finally mixed sentiment posts (M=665.25, SD=793.870). As for the relationship between comments and different post sentiments, by far the most interaction received posts with negative sentiment (M=60.47, SD=93.663). However, the other three types of sentiment on average received a similar amount of comments (mixed: M=25.16, SD=36.077; neutral: M=24.04, SD=70.613; and lastly, positive with M=23.02, SD=73.233).
Comparing the difference between all sentiment groups and the social engagement, results indicated a significant difference between the comments received
and the type of post sentiment (F(3,1629)=9.070, p= .000). However, the relationship between likes and different groups of sentiment was not statistically significant (F(3,1611)=0.646, p= .586).
After normality checks and Levene’s test was carried out to test the homogeneity of variance, the dataset on likes (p= .625) met the assumptions to further carry out the post hoc comparisons using the Tukey test (since equal variance can be assumed if p > .05). Although, negative sentiment posts received by far more likes than other groups of sentiment, there was no statistical significance found (negative and neutral p= .563; negative and positive p= .795; and negative and mixed sentiments p= .772). As for comparison between other three sentiment groups (except negative) no significant results were found (see Table 2 in Appendix A).
Furthermore, comparing the means between the four types of the sentiment and comments, Fisher’s Least Significant Difference (LSD) was carried out. In this particular instance LSD test was chosen based on the fact that ANOVA test showed statistical significance in mean comments received between the groups (F(3,1629)=9.070, p= .000). Furthermore, the post hoc comparison using LSD found a significant difference between negative sentiment group and all other groups of sentiment (negative and neutral p= .000, negative and positive p= .000, and negative and mixed p= .001). More specifically, negative sentiment posts on average received by 36 comments more than neutral posts (Mdifference=36.435*), by 37 comments more than positive sentiment posts
(Mdifference=37.456*), and by 35 comments more than those posts with mixed sentiment
27 groups – neutral, positive and mixed – no statistical significance in received likes was found between any of those groups (see Table 3 in Appendix A).
In sum, with regards to likes received, based on the results it can be concluded that firstly, negative sentiment has more likes than neutral, positive or mixed sentiment posts; and secondly, no significant difference on likes between neutral, positive and mixed sentiment. However, with regards to comments and different types of sentiment, firstly, negative posts received more comments than positive, neutral or mixed (positive statistical significance). Secondly, there was no difference found for comments between other types groups of sentiment (neutral, positive or mixed). Hence, hypothesis 3 is rejected as positive sentiment posts fail to indicate higher social engagement.
Conclusion and Discussion
This study aimed to better understand which opinion-leader LinkedIn post characteristics is associated with higher social engagement. Due to technological advancements and the growing market competition, companies too need to adjust their marketing strategies in order to promote and sustain their corporate brand, image, and reputation. LinkedIn, “the world's largest professional network with nearly 660+ million users in more than 200 countries and territories worldwide,” has shaped the overall online interaction between companies and different groups of stakeholders (LinkedIn, 2019). Having the professional network at the center of attention, LinkedIn has stimulated a rise of the
specific group of opinion-leaders (directors, (co-)founders and chiefs) acting as the representatives of their associated company.
It is considered that social engagement is a fundamental factor that allows information to reach the largest audience (Sehl, 2019; Foote, 2019). Taking into account that LinkedIn uses an algorithm that is “working to amplify the most engaging posts from the most prominent creators,” individuals should be particularly interested in creating content that would attract the most attention (Hutchinson, 2019). Not only a higher social engagement, such as likes and comments, would account for the highest reach, but also depending on the content of the message, would allow successfully sell the effort of the company.
As such, based on the theory this paper studied four distinct post characteristics – (1) length; (2) mentioning one’s company; (3) communication of personal information; and (4) sentiment of the post (negative/neutral/positive/mixed). Returning back to the main research question set at the beginning of the paper - what are the opinion-leader post characteristics and to what extent do they relate to social engagement, the results of the analysis showed that different factors could increase the overall online social engagement (measured in likes and comments).
After analyzing 1633 opinion-leader LinkedIn posts, the results of the analysis indicated a significant positive relationship between the post text length and social engagement (for both comments and likes). Based on the results, it can be concluded that those posts with longer texts will receive more comments and likes than posts with the shorter post text. Such findings go hand in hand with the theory described earlier, suggesting that people are social beings who desire social interaction (Hennig-Thurau et
29 al., 2004). Hence, by including more information, elaboration and expression of a particular subject or topic can, therefore, help to create a space for the audience to interact. Although the results indicated a statistically significant positive relationship between the length of the post and social engagement, more research is needed to better understand which other post characteristics (such as topic of the post) is associated with higher social engagement.
Furthermore, shedding light on how often do opinion-leaders use the name of their company in their personal posts and to what extent does it influences social engagement, results showed opinion-leaders tend to include their company name in the majority of their posts. Since LinkedIn positions itself as “the world's largest professional network”, such finding is not surprising (LinkedIn, 2019). Moreover, taking into the framework of ‘word-of-mouth’ communication, which suggests the informal communication about either the “ownership, characteristics of brand, a product, a service, an organization or a seller”, these results are consistent with the literature (Ladhari, 2007, p.1093). As such, opinion-leader communication on LinkedIn can, therefore, be seen as 'word-of-mouth' communication. Furthermore, in alignment with Abratt and Kleyn (2012), such interpersonal communication indicates opinion-leaders acting as the company representatives. As such, having a higher social status, opinion-leaders can be particularly beneficial not only building the company’s reputation via their personal profiles, but also selling the effort of company (Ibid.).
As further analysis tested the relationship between mentioning of one’s company’s name and social engagement, no significant relationship was found. This
finding is in alignment with what has earlier been suspected implying that the company’s name alone is not an indicator for increased social engagement.
The analysis concerning the presence of personal information in opinion-leader LinkedIn posts showed rather surprising results. Firstly, contrary to Sun (2006), which suggested that online ‘word-of-mouth’ usually includes personal information, such as personal experience, results showed that only a marginal percentage of all posts included one’s personal information. Furthermore, unlike the initial assumption, the independent variable showed no relationship with either likes or comments. Such findings can be explained because LinkedIn is a professional network and thus, opinion-leaders are less willing to communicate their own personal information.
As for the results indicating no significant relationship between those posts including personal information and social engagement, a possible explanation can be that different types of personal information would have a different level of social engagement. For instance, according to Bolinio et al. (2016), people use conscious and strategic behavior to shape the way they are seen by others. Trying to manage their image opinion-leaders may accidentally frame their message in a way that could be perceived as ‘better than the rest’. As such, not only types of personal information might have an impact on social engagement, but also depending on how opinion-leaders present their new promotions, personal success stories and/or achievements might determine how much attention (likes and comments) they will draw to the post. Perhaps, the more arrogant or “show-off” they appear, the less social engagement they would receive; and, by contrast, the more vulnerable, emotional and humble they appear, the more social activity would follow. However, to makes such claims, more nuanced content analysis is needed to
31 better understand whether different types/frames of personal information communicated in their posts have a different level of social engagement.
Lastly, studying four different sentiments – negative, neutral, positive and mixed - the negative sentiment was found to be the most significant positive indicator for social engagement. Although positive sentiment was found by far the most used sentiment, it appeared not to have any significant relationship with either likes or comments. Considering that negative posts were found so few, it is likely that those are the posts that stand out from the rest of the posts. Based on the fact that people desire social interaction (Hennig-Thurau et al., 2004), as well as negative sentiment posts are about an issue or a problem, it creates a space beneficial for social interaction. As such, negative posts give an opportunity for people to react on them as well as discussing them, ultimately creating a higher social engagement.
Furthermore, linking the study back to the theoretical framework, it can be concluded that the opinion-leader interpersonal communication on social networking platform LinkedIn can only partly meets the criteria of the two-way symmetrical communication model as introduced by Grunig and Grunig (1992). Although, LinkedIn affords the possibility of two-way communication, and through social interaction (likes and comments) have the possibility for stakeholders to understand the viewpoints of one another, however, it is hard, if possible at all, to prove whether or not an opinion-leader is conducting a truly honest form of communication. As such, a truthful communication, considered as a fundamental aspect of two-way symmetrical communication, cannot be proven.
Limitations and future research
By analyzing more than 101 opinion-leader LinkedIn posts, this study has provided valuable social media communication insights into a better understanding of social engagement. Despite the contribution, this study is not without limitations.
The current analysis analyzed 1633 opinion-leader LinkedIn posts from which only 148 were female generated messages, which may impact the generalizability of its findings. However, since there were only 24 women chief executive officers found in the 2018 Fortune 500 list, this study can be considered as the most relevant (Mejia, 2018). Further research can be built upon this study investigating more nuanced online post characteristics of opinion-leader posts. Based on the fact that this study has not considered timing when the post is posted (day/time/frequency) it would be interesting to understand if the timing of the post can account for higher social engagement. For example, if a post is posted during the weekends when people are usually more with their families, the online social engagement might be lower, whereas, perhaps, during the mornings, when people are browsing the Internet – higher. In sum, this study has contributed a great deal to the already existing corporate communication literature (Abratt & Kleyn, 2012; Berkelaar, 2017; Ryan & Deci, 2000), opening a wide range of possibilities for future research.
33
References
About.linkedin.com. (n.d.). About LinkedIn. [online] Available at: https://about.linkedin.com/ [Accessed 10 Jan. 2020].
Abratt, R. & Kleyn, N. (2012). Corporate Identity, corporate branding and corporate reputations. European Journal of Marketing, 46(7/8), 1048‐1063
Berkelaar, B. L. (2017). How implicit theories help differentiate approaches to online impression management: A preliminary typology. new media & society, 19(12), 2039-2058
Berkman, H. W., & Gilson, C. (1986). Consumer Behavior: Concepts and Strategies, Boston: Kent Pub.
Bolino, M., Long, D., & Turnley, W. (2016). Impression management in organizations: Critical questions, answers, and areas for future research. Annual Review of Organizational Psychology and Organizational Behavior, 3, 377-406
Bonsón, E., & Bednárová, M. (2013). Corporate LinkedIn practices of Eurozone companies. Online Information Review, 37(6), 969-984.
Case, D. O., Johnson, J. D., Andrews, J. E., Allard, S. L., & Kelly, K. M. (2004). From two‐step flow to the Internet: The changing array of sources for genetics information seeking. Journal of the American Society for Information Science and Technology, 55(8), 660-669
Choi, S. (2015). The two-step flow of communication in Twitter-based public forums. Social Science Computer Review, 33(6), 696-711.
Cottan-Nir, O., & Lehman-Wilzig, S. (2018). CEO Branding: Between Theory and Practice—Case Studies of Israeli Corporate Founders. International Journal of Strategic Communication, 12(2), 87-106.
Di Gangi, P. M., & Wasko, M. M. (2016). Social media engagement theory: Exploring the influence of user engagement on social media usage. Journal of Organizational and End User Computing (JOEUC), 28(2), 53-73.
Duan, W., Gu, B., & Whinston, A. B. (2008). Do online reviews matter?—An empirical investigation of panel data. Decision support systems, 45(4), 1007-1016.
Foote, A. (2019). The LinkedIn Algorithm Explained In 25 Frequently Asked Questions. [online] LinkedInsights.com. Available at: https://www.linkedinsights.com/the-linkedin-algorithm-explained-in-25-frequently-asked-questions/ [Accessed 10 Jan. 2020].
Gibbs, J, Rozaidi, N, Eisenberg, J (2013). Overcoming the “ideology of openness”: probing the affordances of social media for organizational knowledge. Journal of Computer-Mediated Communication 19(1): 102–120
Graney, M. J. (1975). Happiness and social participation in aging. Journal of Gerontology, 30(6), 701-706.
Grunig, J. E., & Grunig, L. A. (1992). Models of public relations and communication. In J. E. Grunig (Ed.), Excellence in public relations and communication management (pp. 285‐325). Hillsdale, NJ: LEA
Hamzah, Z. L., Alwi, S. F. S., & Othman, M. N. (2014). Designing corporate brand experience in an online context: A qualitative insight. Journal of Business Research, 67(11), 2299-2310.
Hennig-Thurau, F.T., Gwinner, K.P., Walsh G. & Gremler, D.D. (2004) Electronic word-of-mouth via consumer-opinion platforms: what motivates consumers to articulate themselves on the internet? Journal of Interactive Marketing, 18(1), pp. 38–52 Hennig-Thurau, F.T., Gwinner, K.P., Walsh G. & Gremler, D.D. (2004) Electronic
word-of-mouth via consumer-opinion platforms: what motivates consumers to articulate themselves on the internet? Journal of Interactive Marketing, 18(1), pp. 38–52 Hutchinson, A. (2019). LinkedIn Has Outlined its Recent Algorithm Updates, Which
Focus on Broadening Engagement. [online] Social Media Today. Available at:
https://www.socialmediatoday.com/news/linkedin-has-outlined-its-recent-algorithm-updates-which-focus-on-broadeni/557609/ [Accessed 15 Jan. 2020]. Johnston, K. A. (2018). Toward a theory of social engagement. The handbook of
communication engagement, 1, 19-32.
Kempe, D., J. Kleinberg, and E. Tardos. 2003. Maximizing the spread of influence through a social network. In Proceedings of the Ninth ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, 137-46. New York: ACM.
Kim, S., Kandampully, J., & Bilgihan, A. (2018). The influence of eWOM communications: An application of online social network framework. Computers in Human Behavior, 80, 243-254.
35 Ladhari, R. (2007) The effect of consumption emotions on satisfaction and
word-of-mouth communications. Psychology and Marketing, 24(12).
Leichty, G., & Springston, J. (1993). Reconsidering public relations models. Public Relations Review, 19(4), 327-339.
Lyons, B., & Henderson, K. (2005). Opinion leadership in a computer‐mediated environment. Journal of Consumer Behaviour: An International Research Review, 4(5), 319-329.
Mejia, Z. (2018). Just 24 female CEOs lead the companies on the 2018 Fortune 500—
fewer than last year. [online] CNBC. Available at:
https://www.cnbc.com/2018/05/21/2018s-fortune-500-companies-have-just-24-female-ceos.html [Accessed 30 Jan. 2020].
Papacharissi, Z. (2009). The virtual geographies of social networks: a comparative analysis of Facebook, LinkedIn and A Small World. New media & society, 11(1-2), 199-220.
Quinton, S., & Wilson, D. (2016). Tensions and ties in social media networks: Towards a model of understanding business relationship development and business performance enhancement through the use of LinkedIn. Industrial Marketing Management, 54, 15-24.
Quinton, S., & Wilson, D. (2016). Tensions and ties in social media networks: Towards a model of understanding business relationship development and business performance enhancement through the use of LinkedIn. Industrial Marketing Management, 54, 15-24.
Park, N. S. (2009). The relationship of social engagement to psychological well-being of older adults in assisted living facilities. Journal of Applied Gerontology, 28(4), 461-481.
Rosenberg, J, Egbert, N (2011). Online impression management: personality traits and concern for secondary goals as predictors of self-presentation tactics on Facebook. Journal of Computer-Mediated Communication 17(1): 1–18.
Rozzini, R., Bianchetti, A., Franzoni, S., Zanetti, O., & Trabucchi, M. (1991). Social functional and health status influences on mortality: Consideration of a multidimensional inquiry in a large elderly population. Journal of Cross-Cultural Gerontology, 6, 83-90.
Ryan, R. M., & Deci, E. L. (2000). Self-determination theory and the facilitation of intrinsic motivation, social development, and well-being. American psychologist, 55(1), 68.
Sehl, K. (2019). How the LinkedIn Algorithm Works And How to Make it Work for You. [online] Hootsuite Social Media Management. Available at: https://blog.hootsuite.com/how-the-linkedin-algorithm-works-hacks/ [Accessed 10 Jan. 2020].
Sun, T., S. Youn, G. Wu, and M. Kuntaraporn. 2006. Online word-of-mouth (or mouse): An exploration of its antecedents and consequences. Journal of Computer-Mediated Communication 11, no. 4: 1104-27.
Sundar, S. & Limperos, A.M. (2013). Uses and grats 2.0: New gratifications for new media. Journal of Broadcasting & Electronic Media, 57, 504-525.
Thompson, M. G., & Heller, K. (1990). Facets of support related to well-being: Quantitative social isolation and perceived family support in a sample of elderly women. Psychology and Aging, 5(4), 535-544.
Walther, JB (2011). Theories of computer-mediated communication and interpersonal relations. In: Knapp, M, Daly, J (eds) The SAGE Handbook of Interpersonal Communication. Thousand Oaks, CA: SAGE, pp. 443–480.
Xie, H. J., Miao, L., Kuo, P. J., & Lee, B. Y. (2011). Consumers’ responses to ambivalent online hotel reviews: The role of perceived source credibility and pre-decisional disposition. International Journal of Hospitality Management, 30(1), 178-183.
37
APPENDIX A
Table 2.
Comparison between likes received and different groups of post sentiment (Tukey HSD test) (I) SENTIMENT (J) SENTIMENT Mean
Difference(I-J) Std. Error Sig.
95% Confidence Interval Lower Bound Upper Bound
Negative Neutral 819.104 630.006 .563 -801.09 2439.30 Positive 544.435 593.002 .795 -980.60 2069.47 Mixed 822.245 856.582 .772 -1380.64 3025.13 Neutral Negative -819.104 630.006 .563 -2439.30 801.09 Positive -274.669 341.204 .852 -1152.15 602.81 Mixed 3.142 706.046 1.000 -1812.61 1818.89 Positive Negative -544.435 593.002 .795 -2069.47 980.60 Neutral 274.669 341.204 .852 -602.81 1152.15 Mixed 277.811 673.235 .976 -1453.56 2009.18 Mixed Negative -822.245 856.582 .772 -3025.13 1380.64 Neutral -3.142 706.046 1.000 -1818.89 1812.61 Positive -277.811 673.235 .976 -2009.18 1453.56 Table 3.
Comparison between comments received and different groups of post sentiment (LSD)
(I)
SENTIMENT (J)
SENTIMENT
Mean
Difference (I-J) Std. Error Sig.
95% Confidence Interval
Lower Bound Upper Bound
Negative Neutral 36.435* 7.682 .000 21.37 51.50 Positive 37.456* 7.245 .000 23.25 51.67 Mixed 35.310* 10.419 .001 14.87 55.75 Neutral Negative -36.435* 7.682 .000 -51.50 -21.37 Positive 1.021 4.147 .806 -7.11 9.16 Mixed -1.125 8.560 .895 -17.91 15.66 Positive Negative -37.456* 7.245 .000 -51.67 -23.25 Neutral -1.021 4.147 .806 -9.16 7.11 Mixed -2.146 8.169 .793 -18.17 13.88
Mixed Negative -35.310* 10.419 .001 -55.75 -14.87
Neutral 1.125 8.560 .895 -15.66 17.91
Positive 2.146 8.169 .793 -13.88 18.17
APPENDIX B
Table 4.
List of CEOs for this research.
Opinion-leader Name Position Company
1 Karen Tamrazyan Co-founder and CEO Freeware Lovers
2 Hans Vestberg CEO Verizon
3 Kevin Johnson President and CEO Starbucks
4 Antonio Neri President and CEO Hewlett Packard
5 John Fallon CEO Pearson
6 Pierre-André de
Chalendar
Chairman and CEO Saint-Gobain 7 Stefan Oschmann Chairman and CEO Merck Group
8 Dominic D. Smith Founder WINETASTERY
9 Michiel Dijkman Head of Corporate Affairs
Samsung 10 Patrice Louvet President and CEO Ralph Lauren 11 Alex Gorsky Chairman and CEO Johnson & Johnson
12 Tom Kennedy CEO Raytheon
13 Thomas Sedran CEO Volkswagen
14 Charles Phillips Chairman Infor
15 Mohit Aron Founder and CEO Cohesity
16 Daniel Dines Founder and CEO UiPath
17 Henrique Dubugras Founder Brex
18 Ron Dutta Vice President of
Business Development
Support Services Group
19 Kevin Plank Founder Under Armour
20 Prasanth Manghat CEO NMC Healthcare
21 Girish Mathrubootham Founder and CEO Freshworks Inc.
22 Nicolas Sekkaki CEO IBM FRANCE
39
24 Bipul Sinha Co-founder and CEO Rubrik
25 Emanuel Chirico Chairman and CEO PVH Corp.
26 Heather Bresch CEO Mylan
27 Jack Salzwedel Chairman and CEO American Family Insurance
28 Carlos Eduardo Moyses CEO iFood
29 Paul Hermelin Chairman and CEO Capgemini Group
30 Robert Reffkin Founder and CEO Compass
31 Gillian Tans Chairwoman Booking.com
32 Tiago Paiva CEO Talkdesk
33 Margaret Keane CEO Synchrony
34 Mike Roman Chairman 3M
35 Brian Humphries CEO Cognizant
36 Dirk Van de Put CEO Mondelēz International
37 Frank Desvignes Global Head AXA
38 Bob Dudley CEO BP Group
39 Doug Baker Chairman and CEO Ecolab
40 Sebastian Mejia Co-founder and President Rappi
41 Thierry Bolloré CEO Groupe Renault
42 Yves Rannou CEO Senvion
43 Albert Bourla Chairman and CEO Pfizer
44 Robert M. Bakish President and CEO Viacom 45 Dominique Cerutti Chairman and CEO Altran
46 Méka Brunel CEO Gecina
47 Jan Zijderveld CEO Avon
48 Andrew Thompson CEO Proteus Digital Health
49 Lance Uggla Chairman and CEO IHS Markit
50 Dan Burton CEO Health Catalyst
51 Douglas Peterson President and CEO S&P Global
52 Ramin Sayar President and CEO Sumo Logic
53 Philippe Donnet Managing Director and CEO
Generali Group 54 Dave Ricks Chairman, President and
CEO
Eli Lilly and Company
55 H Fisk Johnson CEO SC Johnson
56 Jay Fulcher Chairman and CEO Zenefits
57 Gregory Hayes Chairman and CEO United Technologies 58 Ernest Spitzer, MD Board Member Cardialysis
