Literature review

The emergence of social media took place simultaneously with Web 2.0. With the introduction of Web 2.0 into our lives, the Internet has become individualized, and use of social networks increased gradually. The Internet has become an interactive virtual world from a read-only state, and it has brought a different dimension to communication [12]. Today, social media is a wide network of interactions where people from many areas [15]. In particular, Internet has become a part of life due to the widespread use of smartphones. The majority of people actively use social media in daily life [16]. A recent study has stated that 70% of peoples in the USA have at least one social media account nowadays. The peoples over 65 years old of 62% have a social media account and they are regularly on social media. This observation is also similar for adolescents. The using of social networks has been reported as 77% for teenagers aged 13–16 years in 25 European countries [16, 17]. These social network users interact for an average of more than 2.4 hours a day on social media [18]. This increased instant interaction has further increased the use of social media. Beside, social media platforms allows independent sharing, regardless of age, venue, and gender. Thus, information spreads rapidly across a wide area [15, 18]. Additionally, social media platforms ensure simple interaction pathways between people, companies, and scientists without leaving the desk. Therefore, many scientists use social media in their personal or professional lives [19].

The most commonly used platform for the purpose of spreading science is Twitter [20]. Twitter is the most popular microblogging platform nowadays. This platform allows the publication of short messages by its users and enables them to communicate with each other. Evan Williams, Biz Stone, and Jack Dorsey created Twitter in March 2006 and brought into use in July 2006. Twitter has become one of the 10 most visited websites in 2013 and was defined as “the SMS of the Internet” [21, 22]. In 2019, it was reported that there were 330 million monthly and 145 million daily active Twitter users. Nowadays, it was reached 339 million users in 2020 [23]. The number of Twitter users has been increasing daily. Today, Twitter ranks as the world’s second most widely used social network. Twitter users can follow a conversation and discuss a topic using messages named “tweets.” Tweets are constrained to 140 characters of text. Later, this limit was increased to 280 characters. Twitter allows the sharing of photos or videos [24]. Twitter was initially used to share news about the lives of celebrities. Afterward, it reached a broader audience quickly, especially with the participation of famous names and the involvement of political campaigns. On average, approximately 98 thousand tweets are sent every minute on the platform, allowing an excellent interaction. Given the rise in popularity of Twitter, its use is increasing in all parts of society [25]. This widespread use has also caught the attention of the scientific community. The use of Twitter as a tool for the dissemination of academic articles has soon become the focus of attention in the scientific world.

Academic output has increased gradually worldwide. Therefore, eliminating the relevant from the irrelevant has become essential for the scientific world. Thus, an impact scale has been necessitated the published articles [26]. The primary impact of the published article has been measured by its citations [26]. However, article citations have recently become questionable due to negative factors, such as the slow process of identifying truly impactful articles. The long wait time required the emergence of the articles’ importance has led to need for an alternative metric scale [27].

Today, non-traditional metrics, altmetrics, are increasingly used to measure the real-time reach and influence of a scientific article [26, 28]. The term altmetrics was first proposed by Jason Priem in 2010 [29]. Thereafter, it has gained wide use in highlighting previously unknown and unrecognized scholarly impact metrics of studies [26, 30]. The alternative metric scores play a role in complementing traditional metrics or indicators [31]. Many publishers, such as SAGE, Taylor-Francis Group, Elsevier, Nature Publishing Group, and Public Library of Science provide much information to their readers by their altmetrics evaluating system [32]. These altmetrics are calculated by various methods, including “Altmetric” and “Plum Analytics” [33]. All altmetrics that are an alternative to the traditional citation system provide a score for the research output.

These scores have become attractive for researchers [34]. Many scientists believe that these alternative metric scores show the real impact of published articles [35–37]. Many web-based platforms play a significant role in obtaining an altmetric score [11, 31]. With the data obtained from these platforms, a digital score was acquired for academic output [11]. All altmetrics are based on the using social media and other online tools for disseminating scholarly information [38]. The use of social media platforms contributes significantly the spread of shared information in a wider environment. Thus, the sharing of academic output on social media accounts can reach more people faster by eliminating the waiting period in the traditional citation system [39, 40]. They have also allowed the impact of articles to be more immediately determined, contrary to traditional citation metrics [41].

However, some concerns remain that the altmetric score can be manipulated [42, 43]. In particular, the use of automatic bot can affect the altmetric scores of articles [44]. Further, Twitter accounts can affect the results of shared articles. These account holders may be social workers, companies, or politicians, and would have more followers than others [45, 46]. Thus, some researchers suggest the use of “alt‑index” to measure the social visibility of scientific research [6]. Similarly, Haustein et al. argued that social media metrics could not actually be regarded as alternatives to traditional citations; hence, they proposed these metrics as promoters of traditional citations [11]. The authors suggested the use of “Twitter Coupling” to deal with these concerns [38]. Although this may be a solution there is a consensus in many studies that social media usage will increase the impact of academic papers, thus ignoring these concerns [47, 48].

The potential of social media platforms to connect with other fields raises various scientific questions [49]. Therefore, the papers in the social sciences and humanities are more often found on social media platforms [11]. Today, social interaction is so intertwined with media that it is not possible to separate social media from the media sector. Thus, most studies related to social media have been published in the communication sciences. This increased usage of these social networks has led to the research question, “How are altmetrics and citation measures related in communication journals?” Tonia et al. (2016) stated that there were no statistical differences between tweeted and non-tweeted papers regarding their citation numbers [50]. Costas et al. (2015) found only weak correlations in citations suggested by altmetrics and traditional citation analysis [35]. Further, some articles can be more attractive than other on social media platforms. Hence, some researchers have argued that there is a difference between social impact and real impact [11]. However, Thelwall et al. (2013) found that altmetrics was associated with citation counts [51]. Another study stated that there is a positive relationship between social media posts and academic citations [52]. Similarly, Shuai et al. (2012) detected significant correlations between tweets and early citations on 4,606 pre-prints articles [8]. Nevertheless, there is no consensus in the studies that to evaluate between tweeted and non-tweeted papers regarding their citation numbers in communication science, which is the scientific field mostly associated with social media and Twitter. The studies related to social media are encouraged in the field of communication as people present themselves through digitally networked platforms today. Therefore, it was hypothesized that there might be a correlation between Twitter posts and traditional citations the articles in the top ten communication-based journals.

Study design

The present study was designed as a retrospective cross-sectional study. The aim was to examine the effect of Twitter and other social media platforms on academic citations. Therefore, the top ten communication-based journals [9] were evaluated based on the SCImago Journal & Country Rank (cited in 2019). It was used the SCImago Journal & Country Rank search field to select journals and filtered only “communication” journals. The selection criteria were the top ten communication-based journals according to their SCImago Journal Rank indicator. This indicator is a measure of journals’ academic impact and accounts for both the number of citations received by a journal and the importance or prestige of the journals in which the citations [53]. The impact factor of these journals was ≥ 2, and the quartile (Q) index was Q1. All ten journals had similar indexes and similar impact factors. It was assumed that this would reduce the potential for unwanted variation differences in social media activity.

Impact Factor: The impact factor has been defined as an indicator of academic journal. It reflects the year’s average number of citations per paper published during the preceding two years. It is often used as a relative indicator of a journal’s importance in its field. The journals with high impact factors are thought to be more important than those with low impact factors [54].

Q index: The journals’ rank in each specific category is separated into quartiles by the Journal Citation Report and SCImago Journal and Country Rank: Q1, Q2, Q3 and Q4. Q1 comprises the top 25% of journals in the list; Q2, Q3, and Q4 comprise 25% to 50%, 50% to 75%, and 75% to 100% of journals in the list, respectively (https://www.mondragon.edu/en/web/biblioteka/publications-impact-indexes).

These Journals “Political Communication (Q1, IF: 4.339)”, “Journal of Advertising (Q1, IF: 6.302)”, “Journal of Communication (Q1, IF: 4.846)”, Big Data and Society (Q1, IF: 4.577), Applied Linguistics (Q1, IF: 4.286), Communication Methods and Measures (Q1, IF: 5.281), New Media and Society (Q1, IF: 4.577), Human Communication Research (Q1, IF: 3.540), Public Opinion Quarterly (Q1, IF: 2.494), and Digital Journalism (Q1, IF: 4.476) were included in this present study (Table 1).

Download:

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

Table 1. The characteristics of journals.

https://doi.org/10.1371/journal.pone.0263725.t001

Data collection

All issues of these journals in 2018 were reviewed through the journals’ web pages. All published articles in this year were evaluated on the web pages’ archive. The date range was considered based on published issues in the journals. Only original articles (meta-analyses, systematic reviews, original research articles, and research notes) in 2018 were included in this study. Articles such as editorials, book review articles, case reports, letters to the editor, and other non-research correspondences were excluded.

The findings of the articles (title, doi number, article type) were recorded. In January 2021, the title of the article or doi number was searched one by one in Google Scholar (GS) and Web of Science (WoS) Clarivate. The traditional citation numbers (GS and WoS) of these articles were recorded. The tweet number and social media posts of these articles were searched by their metric evaluating system (https://www.altmetric.com/) and recorded. The data were appraised nearly three years after the articles’ publication date. It was thought that this period would allow the impact of published articles (the citations and Twitter mentions) to be fully observed.