Citations Gone #Social: Examining the Effect of Altmetrics on Citations and Readership in Communication Research

Abstract

Altmetrics are a relatively new phenomenon in research. These metrics measure the attention that research articles receive from nontraditional venues such as social media and the Internet. This study examined how these metrics affect both the readership and citation of articles in communication research. The study examined citation data alongside altmetrics data from academic social networking sites ResearchGate and Mendeley, as well as mentions on Facebook, Twitter, and Google+. Results indicated that all altmetrics positively correlated with citation. Posting articles on sites such as ResearchGate and Mendeley not only impacted readership, it increased the likelihood of citation. Other variables that improved readership and citation were social media mentions, downloadable articles, coauthorship, and an active online presence among scholars.

Keywords

academic social networking altmetrics bibliometrics citation Mendeley ResearchGate

“Research is conducted, published, read and cited,” state Casson and Al-Qureshi (2010, p. 655) in a discussion of the role of citation in journal editorial policy. The statement aptly surmises the traditional mode of scholarly production and dissemination. It is the last part of the statement, read and cited, that is the focus of the current study. Here, I used both traditional citation data and the newer, internet-driven altmetrics data to examine the effect of social media attention on peer citation in communication research. Altmetrics measure the attention that research articles garner from nontraditional venues such as social media and the Internet. With altmetrics, Casson and Al-Qureshi’s quote could as well read as follows: Research is conducted, published, read and cited—but also tweeted, blogged, posted, shared, commented on, uploaded, downloaded, bookmarked, and followed. This study seeks to examine the role and effect of altmetrics in this process. Specifically, the study examined high-impact articles in the top journals in communication and how the attention they got from Mendeley, ResearchGate (RG), Twitter, Facebook, and blogs affected citation and readership.

Several reasons merit this study. First, there is the need to understand the impact of academic social networking via platforms such as RG and Mendeley. These recent ventures have the potential to transform scholar-to-scholar communication just as more established platforms like Twitter and Facebook did to communication in general. Some studies have already reported such developments (Han et al., 2019; Liu & Fang’s, 2017; Park & Park, 2018; Yu, Xu, Xiao, Hemminger, & Yang, 2017). Since launching in 2008, RG and Mendeley boast 15 and 6 million users, respectively (Mendeley, 2019; RG, 2019). Most important, however, is the uniqueness these sites add to scholar-to-scholar communication. As the author discuses in later sections, scholars can now connect with others worldwide, not only to share articles but also to get updates, seek collaboration, comment on articles, and even participate in research-related crowdsourcing sessions. The end result is that academic work gets more attention, which is then quantified and measured by altmetrics data. These developments have not gone unnoticed because major publishers like Sage, Taylor & Francis, Wiley, and Elsevier now post and track altmetrics data for the articles they carry.

This study goes beyond observing trends. It adds to the burgeoning body of work on altmetrics. The study also provides a lens through which we can examine factors that affect attention to scholarly work in communication and the role that social media plays in the process. This way, we get a better perspective of communication research and also get to build a framework that guides future scholarly endeavors. Research on bibliometrics and citation supports this. Such endeavors help identify research trends, both emerging and past (White, 2011). Bibliometric and citation data also reveal areas that generate attention and those that do not, and this information can usefully guide editorial policy (Antonakis, Bastardoz, Liu, & Schriesheim, 2014). Additionally, such data can help emerging scholars create more productive and influential research agendas (Cohn, Farrington, & Iratzoqui, 2017). It is also useful to know the reach and influence of scholars both within and outside the discipline (Koo, 2017). The advent of social media compels us to expand our horizons. We not only have to examine how scholarship is disseminated within traditional channels but beyond into nontraditional spheres such as social media outlets. I further discuss of the role of citation data and altmetrics in the next sections.

Altmetrics

It is no surprise that the term altmetrics was coined via a tweet, by then doctoral student Jason Priem in 2010: “I like the term #articlelevelmetrics, but it fails to imply diversity of measures. Lately, I’m liking #altmetrics” (Roemer & Borchardt, 2015, p. 5). Priem proposed altmetrics, not to supplant traditional metrics but as a complement. This way, altmetrics can be defined as quantitative and qualitative “citations on Wikipedia and in public policy documents, discussions on research blogs, mainstream media coverage, bookmarks on reference managers like Mendeley, and mentions on social networks such as Twitter” (Altmetric.com, n.d., para. 1). Even though altmetrics are still at a nascent stage, research indicates that they are gaining ground and prominence. First, major publishers like Sage, Wiley, Taylor & Francis now provide altmetrics scores alongside journal articles. Second, most journal articles in most academic databases now come with an altmetrics score. While some emerging research indicates that altmetrics are more likely to accompany newly published articles (Roemer & Borchardt, 2015), the author came across altmetrics accompanying articles published from as far back as 1951. This means that at the very least, altmetrics are revitalizing online attention to older publications.

Emerging research on altmetrics shows moderate to strong significant relationships of these metrics with citation (Brandon, Midway, Sackett, Lynch, & Cooney, 2016; Ebrahimy, Mehrad, Setareh, & Hosseinchari, 2016; Priem, Piwowar, & Hemminger, 2012; Thelwall, Haustein, Larivière, & Sugimoto, 2013). However, little has been done in the way of examining pertinent trends in communication research and specifically regarding the relationship between citation and altmetrics in the field. Studies on these metrics keep finding bigger effects as alternative methods of reading and sharing articles go mainstream. The caveat here is that altmetrics are still new and their impact is tied to the diffusion and adoption of related media such as Mendeley and RG. One early study found a strong correlation between mentions on Twitter and blogs, and citation (Zahedi, Costas, & Wouters, 2014). These findings align with Priem, Piwowar, and Hemminger’s (2012) study that found significant but moderate correlation between Mendeley, a major altmetrics index, and citation. Similarly, Thelwall, Haustein, Larivière, and Sugimoto (2013) found a correlation between mentions on social media outlets such as Twitter and Facebook with citation. In another early study that examined altmetrics and citation within communication research, Torres-Salinas, Cabezas-Clavijo, and Jiménez-Contreras (2013) also found a relationship between two popular altmetrics indices (Mendeley and CiteUlike) and citation (r = .52 and r = .30, respectively).

More recent communication-focused research has also found a relationship between altmetrics and citation. In a study of communication journal articles listed in the Social Science Citation Index, Repiso, Castillo-Esparcia, and Torres-Salinas (2019) found a strong correlation (r = .97) between Mendeley reads and citation and between Twitter mentions and citation (r = .88). Additionally, Stuart, Faucette, and Thomas (2017) found correlations among Google Scholar, RG, and Scopus citation metrics in a study of communication science authors. Correlations ranged from r = .44 to r = .99 with an average correlation of r = .77 among all metrics examined. In another communication-related study on altmetrics, Blackman (2016) discusses the role of altmetrics in academia. For instance, placing articles on sites such as RG allows for the easy sharing of work, which also improves dissemination.

Academic Social Networking

There are several social media venues through which scholars can share, discuss, download, and comment on their work and that of others. Mendeley and RG are the two most prominent. RG, which currently has over 15 million members, launched in 2008 to allow for easy collaboration among scholars (RG, 2019). The site allows researchers to upload copies of their work and give and/or receive comments from others. RG also allows users to freely download articles if available, and if not, users can directly message authors for copies. Unlike Google Scholar, which only provides data on citation, RG also provides citation data and how many times an article was read, who read it, how often it was commented on, who commented on it, and so forth. Also, unlike Google Scholar and traditional indexing services like Scopus or Web of Science, RG’s citation data are uniquely derived from articles housed within the site.

Users are not limited to uploading journal articles. Conference papers, chapters, data sets, posters, theses, working papers, and more are also eligible. Account holders need not upload articles or data sets. They can also post information about upcoming projects and seek collaboration or send research-related questions to the community. The site also monitors user activity of which it uses to create a measure of a user’s presence. This is the RG Score, which measures a scholar’s impact and influence on the RG community (RG, 2017). Because RG uses peer evaluation to determine impact, it does not solely use citation as most other indices do with the h-index. The RG Score is based on an algorithm that considers a variety of factors based on contributions to the site and interaction with other users. Such contribution includes articles or data sets that one uploads online. Interaction includes comments and feedback to other people’s work and posting questions and answers to the online community. Additionally, users receive an RG Reach Score which denotes the number of users a researcher is connected to, including coauthors, followers, project collaborators, and so on. Just as it is with social media sites such as Facebook and Twitter, scholars can also follow others and get updates any time there is development in someone’s research activity.

Mendeley, which also launched in 2008 (Henning, 2008), is both a reference manager and an academic social networking site (Sud & Thelwall, 2014). The reference manager helps scholars cite articles as they write, using a downloadable plugin. This is similar to other services such as Endnote and RefWorks. On its social networking site, scholars can upload articles just like on RG as well as store and share data sets. Scholars can follow others and have followers. Users can also read articles and add other people’s work to their library. This is how Mendeley keeps track of how many times an article is read. Even though Mendeley’s readership metrics are drawn from member activity within the site, it uses third-party sources such Scopus to determine its citation metrics. RG on the other hand draws all its metrics from attention that articles garner within the site. However, unlike RG, Mendeley allows users to share article summaries via social media sites such as Facebook and Twitter via Bitly URL, a web address abbreviating service. Because Mendeley is now owned by publishing giant Elsevier (Dobbs, 2013, para. 1), not all articles housed here are available for download, and most articles published in major journals require subscription. Users can still share articles depending on the access and usage limits set by the publisher. Lastly, unlike RG, Mendeley only provides readership data and not citation data.

Researchers are not limited to RG and Mendeley for academic social networking. They can also tweet, blog, or post comments about certain articles on social media sites (Torres-Salinas, Cabezas-Clavijo, & Jiménez-Contreras, 2013). Publishers like Sage and Wiley use altmetrics tracking systems to gather such information, which is then posted alongside a published article. While these publishers give an overall altmetrics score incorporating mentions in nonsocial media sources such as news and policy documents, they also provide separate metrics for Facebook, Twitter, and blog mentions. This provides data that specifically reflect social media attention, the kind used in the current study. Additionally, mentions on academic social sites such as Mendeley are not counted, thereby preventing double counting within altmetrics data (Altmetric.com, 2016, para. 1).

The Role of Citation Data

Casson and Al-Qureshi (2010) continue their editorial by emphasizing the importance of the dissemination of scholarly work: “Meaningful scientific research demands effective communication: the dissemination of information to fellow scientists and to end-users of this information…” (p. 655). The first step in disseminating such information is publication. But for the intended impact of such works to occur, peers must promulgate by reading, replicating, and citing these works. Citation refers to the “acknowledgments of a scholar’s published work by another scholar in his or her publication (Dowling, 2014, p. 281). In academia, citation data play a key role in determining factors ranging from hiring, merit award and tenure, to determining the quality of an institution (Adler & Harzing, 2009; Merton, 1968; Strevens, 2006). Other uses of citation data include determining research funding, internal and external institutional evaluation, determining Nobel Prize winners, among others (Garfield & Malin, 1968; Gingras & Wallace, 2010; Thomson Reuters, 2008, p. 2).

The most common use of citation data is of course in determining journal rankings. Here, ranking indices use citation data to determine the journal impact factor. This refers to the average citation a publication elicits, and examples include the SCImago Journal Rank (SJR) run by Clarivate Analytics or the Google Scholar Metrics run by Google. SJR uses a series of citation-based markers to derive its impact factor and journal ranking. One such is the h-index, which by SJR’s standards represents the number of articles within a journal that have received a certain number of citations over a certain period of time. Citations per document (2 years) is another SJR marker, this one denoting the number of citations elicited in the year under consideration and citations received within the preceding 2 years. Similar metrics do the same for citations received 3 and 4 years prior to the year under consideration (SCImago, 2016). Google Scholar also uses the h-index to determine its journal ranking. Its calculation weights the frequency and newness of article citations, article length, location of publication, and who the author was (Google Scholar, 2017, para. 1).

Citation Analysis in Communication Research

Citation data have been studied and examined extensively in many fields including communication. Bibliometrics, or the process of using such data, is the “application of quantitative analysis and statistics to publications such as journal articles and their accompanying citation counts” (Thomson Reuters, 2008, p. 3). An early example of such an endeavor is Pasadeos’s (1985) examination of citation patterns in advertising research between 1981 and 1983. The study found, among other things, that advertising scholars cited heavily from business and psychology journals. Pasadeos also found an increase in scholarship as denoted by an uptick in citations from newer articles. Another example is the dedication of an entire 1989 issue of Communication Research to bibliometrics. Here, scholars examined issues such as co-citation in AIDS research (Small & Greenlee, 1989), citation patterns in Japanese studies (Miyamoto, Midorikawa, & Nakayama, 1989), co-citation mapping of scientific scholars (McCain, 1989), and how the invisible college manifested through citation (Lievrouw, 1989). The invisible college refers to a cluster of scholars with shared interests but geographically located at different institutions. The journal’s editor at that time also provided a model through which communication scholars could use bibliometrics to better the field. This included examining producers of information, information artifacts, and concepts. Further examination would focus on the evolution of scholarship, contributions, and the diffusion of communication studies in the wider scholarly universe (Borgman, 1989). Notable is that few of the bibliometric articles published in this issue were communication-related. This explains Paisley’s (1989) call for more bibliometrics studies in communication research, a notable gap at that time given the advanced development of bibliometrics studies in other fields.

Communication scholars seem to have answered Paisley’s call in later years. In the 1990s, for example, we see researchers start to publish more articles analyzing bibliometrics in communication. An example is the first bibliometrics study in public relations research. Here, Pasadeos and Renfro (1992) examined the citation patterns in advertising research over a 15-year period (1975–1989). They found a coalescing pattern where authors cited articles outside the field at first but cited more in-field articles later. The study also found that in that period, books were cited more than journals, with magazines coming in at third. Hoffman and Holbrook (1993) examined, among other factors, co-citation and research atypicality in articles published in the Journal of Consumer Research. Research atypically refers to citation elicited, not by an author’s similarity with others in an area of study but by his or her unique and distinct research. In an analysis of citation patterns in the Journal of Advertising from 1972 to 1991, Muncy (1991) found that authors mostly cited articles from consumer behavior, marketing, and advertising journals in that order. In turn, the journal was cited the most in advertising, consumer research, and marketing journals, respectively. Not all bibliometrics in this period focused on marketing or advertising. Rice and Chapin (1996) examined the Journal of Broadcast & Electronic Media’s reach and influence among 17 communication journals between 1977 and 1993. A notable finding was the considerable increase in influence and impact factor after the journal added & Electronic Media to its title. Influence and impact thereafter either matched or superseded that of the 17 communication journals under study.

Interest in bibliometrics remains healthy in contemporary communication research. Here, scholars have not only expanded focus to a variety of topics but have used bibliometrics to examine communication research in different geographical regions. For instance, Masip and Fernández-Quijada (2011) found that authors in Catalonia, Spain, were more likely to collaborate when publishing in international journals than when publishing in local journals. The authors explained this to be the result of the low levels of participation among Spanish scholars in international journals. Another region-focused study is Zi-yu’s (2014) intercultural bibliometric profile of Library Materials about Macau. Among other findings, the study found this specific area of interest to be highly interdisciplinary with growing productivity. Likewise, Lei and Liao (2017) examined the influence of China-based authors within high-impact journals such as Applied Linguistics, Research on Language and Social Interaction and Journal of Pragmatics, among others. Their study found that articles originating from Mainland China elicited the most citations, followed by those from Taiwan, Hong Kong, and Macau, respectively. Other contemporary bibliometrics research has examined a variety of topics such as social capital (Chul-joo & Dongyoung, 2016), virtual media (Correia et al., 2014), video gaming (White, 2011), and gender and media (Navarro & Martín, 2013).

Even more recently, scholars have examined pertinent topics such as citation and open-access communication journals (Antell, Foote, & Foote, 2016) and citation patterns in translation research (Rovira-Esteva, Aixelá, & Olalla-Soler, 2019). Others such as Guo et al. (2019) have used citation analysis to examine the evolution of advertising literature, specific to product placement in mass media, in order to map out areas of emphasis and areas in need of further exploration. Similar studies have also used citation analysis to explore patterns in nonverbal communication research (Plusquellec & Denault, 2018), public relations research (Ki, Pasadeos, & Ertem-Eray, 2019), sport communication research (Harker & Saffer, 2018), and so forth. The literature discussed above indicates the existence of a healthy body of research on citation in communication. Therefore, the current study provides a timely extension of this scholarship by examining the effect of altmetrics on citation.

Method

Sampling

The unit of analysis was a journal article appearing among the 10 most cited articles in a journal ranked among the top 10 in communication research. First, this method aptly captures a large and representative sample of high-impact articles in the field that also reflects trends in communication research (Bajwa & König, 2019). This approach also reflects methods used in similar studies such as Liu and Fang’s (2017) examination of the top 100 papers on Altmetrics.com and Plusquellec and Denault’s (2018) study of the 1,000 most cited papers in nonverbal communication research. Therefore, a list of top-ranked journals was obtained from two leading journal ranking services: SCImago Journal Rank (SJR) and Google Scholar Metrics. Because the two ranking services use different methods to determine journals ranking, there are slight differences in their top-ranked journals. For instance, at the time of data collection, SJR’s No. 1 ranked journal in communication was Journal of Communication. Google Scholar’s top-ranked journal in communication was New Media & Society. Therefore, I selected all journals that appeared in the top 10 of either or both indices. This method resulted in a sampling frame of 15 journals and a sample of 150 articles.

Data

Both citation and altmetrics data were obtained secondarily. Citation data were derived from the Web of Science, a major citation indexing service. These data included other information such as a list of authors, the location of the corresponding author, year of publication, journal name, and so on. Because there is no central location to derive altmetrics data as is the case with citation, I visited several sources for these data. For Mendeley readership data, I searched the website for metrics for each of the 150 articles. I used the same procedure to derive readership, citation, and related data from RG. Social media data were derived from the respective publisher’s website an article was housed, and this process required searches on Sage, Wiley, and Taylor & Francis Online. Altmetrics data are more dynamic and change faster than citation, and therefore, the data reported here reflect attention the articles received by the data collection dates, February 5–19, 2017.

Dependent Variable

The lone dependent variable was citation. However, I analyzed citation data from both the Web of Science and RG. Web of Science citation data are similar to that of other services such as Scopus and even Google Scholar because they all derive data from the same sources. RG’s citation data are unique because these are derived in-house from articles available within the site’s database. I used both total citations and per-year citation to measure this variable. The per-year citation parameter was derived by diving total citations for each article by the number of years it had been in print. Antonakis, Bastardoz, Liu, and Schriesheim (2014) advocate for the use of the per-year citation to account for decay and/or increased citation over time. This is also advisable because the journal ranking services analyze impact over 2-, 3-, and 4-year periods. I report results from both total citations and per-year citations in later sections.

Independent Variables

Altmetrics

This variable includes readership from Mendeley and RG and social media mentions from Facebook, Twitter, blogs, and Google+. All altmetrics were considered as three separate variables, namely, (a) Mendeley, (b) RG, and (c) social media collectively.

Age

This variable denotes an article’s chronological age from the date of publication to 2016, the latest year for available citation metrics at the time of data collection.

Location

This variable denotes the general geographical location the corresponding author was situated at the time of publication, for instance, North America, Europe, the Middle East, Africa, South America, Asia, or Australia. This variable provides the spatial perspective of the reach of the articles.

Coauthors

This variable represents the number of authors listed for any article. The reasoning behind its inclusion is that if coauthors all have social media accounts and all post the same article or generate comments about it, chances of generating attention to their work increase than when it is a single author.

Download

This dichotomous variable tested whether articles that are available for download without subscription get read and cited more. The reasoning here is that such articles are easy to access and can be read in full by anyone even in the absence of a subscription. Because the download option is more common on RG, download data were collected and analyzed only for this source.

RG score

As mentioned, RG calculates a user’s impact to the community by the number of articles posted, citations, reads, comments, and so on. The RG Score represents this impact. The assumption for including this variable is that scholars with high visibility might attract more attention and this might, in turn, drive more traffic their way, thus impacting the reads and citations of their work.

Data Transformation

Citation data are susceptible to outliers because of the likelihood of seminal and groundbreaking articles that garner a lot of attention over time, and this might compromise the normal distribution of data (Priem et al., 2012; Sud & Thelwall, 2014). This was the case with both the citation and altmetrics data analyzed here, all which were positively skewed. Therefore, transformation of data was necessary to attain normality. This procedure is necessary when performing parametric tests such as the t test, analysis of variance (ANOVA), correlation analysis, and regression with skewed data. Because there still is a debate on the appropriateness of this method, I used advice and methods from extant literature to perform the necessary logarithmic transformations. For instance, it is advisable to perform an exploratory analysis of raw data as a first step (Packard, 2009, 2013). The raw data here returned significant relationships among variables, even with the skewness (see Appendix A for example of analysis using raw data). Second, research indicates that logarithmic transformations can and have been accurately used for correction even in data that are traditionally susceptible to skewness (Bland, Altman, & Rohlf, 2013; Mäkelä, Lagström, Kaljonen, Simell, & Niinikoski, 2013). Therefore, I used the appropriate log base 10 transformations for all citation and altmetrics data (Ferris, Miller, Glassman, Beck, & Diabetic Retinopathy Clinical Research Network, 2010; Liang, Li, Di, Zhang, & Zhu, 2015). Table 1 shows the transformation’s effect on the skewness.

Table 1.

Logarithmic Transformation Statistics.

	Before		After
	Skewness	SE	Skewness	SE
Total citations	3.343	.198	−0.051	.198
Per-year citations	3.656	.198	0.198	.198
Mendeley reads	8.381	.202	−1.292	.198
ResearchGate reads	7.533	.206	−0.703	.198
RG citation—total	3.946	.206	−1.675	.198
RG citation—per year	6.024	.206	−0.707	.198
Social media mentions	3.258	.285	1.266	.198
Age	1.48	.198	−0.032	.198

Note. RG = ResearchGate; SE = standard error.

Results

Descriptive Results

The descriptive results reported here reflect the raw data before the logarithmic transformation was performed. As Table 2 indicates, the oldest article has been in circulation for 65 years since 1951 and the newest one for 3 years since 2013. The most cited article elicited 2,009 total citations and the least cited elicited 22 citations. As mentioned, this study also used the per-year citations an article elicited on average for every year in print, computed by dividing the total number of citations by the age of the article. The reason was to account for the longevity, decay, and resurgent interest in later years. This way, all articles averaged 15.30 citations for every year in publication. The article with the highest per-year average was cited 127.27 times over the 9 years it had been in circulation. The lowest per-year average was an article that garnered 1.89 citations over 35 years in print. Journal of Computer-Mediated Communication was the most frequently cited, with a per-year average of 31.90 citations (see Table 3). Public Opinion Quarterly garnered the most total citations (7,950).

Table 2.

Descriptives.

	N	Minimum	Maximum	Mean	SD
Per-year citations^a	150	1.89	127.27	15.30	15.45
Total citations^a	150	22	2,009	256.07	283.70
RG per-year citations	138	0.86	405.30	29.94	46.12
RG total citations	138	0	4,053	410.17	590.61
Mendeley reads	144	1	8,100	314.65	736.34
RG reads	139	1	29,367	943.65	2,903.85
Social media mentions	71	1	79	9.23	15.43
Age	150	3	65	16.55	12.30
Length in pages	150	2	59	20.93	8.40

Note. RG = ResearchGate; SD = standard deviation.

^a World of Science citation data.

Table 3.

Raw Means and Standard Deviation for Citations and Altmetrics.

	Citations^a		Mendeley^a		ResearchGate^a		Social Media
	Mean^b	SD	Mean	SD	Mean	SD	Mean	SD
Journal of Computer-Mediated Communication	31.97	34.94	894.5	923.84	1,228.67	1,765.4	22.4	27.42
Jour. of Comm.	26.01	21.53	1,290.7	2,442.94	6,058.56	9,636.15	5.67	6.35
Public Opin. Quart.	25.12	17.35	220.3	148.77	861.1	1,501.92	4.13	4.94
New Med. & Soc.	20.89	12.56	487.6	311.86	203.4	215.94	15.38	21.41
Human Comm. Res.	18.81	8.49	195.3	188.31	1,296.4	2,223.57	3.29	3.3
Comm. Theory	17.1	3.86	342.6	123.8	742.89	699.8	3.75	3.59
Comm. Res.	16.6	11.06	107.3	91.89	475.1	642.02	3.29	3.09
Info. Comm. & Soc.	16.17	17.31	183.7	96.02	293.86	341.64	25	23.4
Res. on Lang. & Soc. Int.	12.34	8.92	87.33	43.88	180.11	110.56	—	—
Jour. of Prag.	11.55	8.08	169.1	153.57	385.33	320.46	—	—
P.R. Rev.	10	5	230.5	213.67	1,483.1	2,001.74	5.5	4.66
Media Cult. & Soc.	7.17	5.7	162.4	188	185	123.14	4.4	3.29
Int. Jour. of Comm.	7.15	4.78	237.2	161.65	261.8	581.14	—	—
Journalism	5.4	1.58	74.1	58.83	398.6	614.3	4.5	4.47
Comm. Educ.	3.24	1.25	50.89	34.86	66.1	51.49	—	—

Note. Comm. Educ. = Communication Education; Comm. Res. = Communication Research; Comm. Theory = Communication Theory; Human Comm. Res. = Human Communication Research; Info. Comm. & Soc. = Information, Communication & Society; Int. Jour. of Comm. = International Journal of Communication; JCMC = Journal of Computer-Mediated Communication; Jour. of Comm. = Journal of Communication; Jour. of Prag. = Journal of Pragmatics; Media Cult. & Soc. = Media, Culture & Society; New Med. & Soc. = New Media & Society; P.R. Rev. = Public Relations Review; Public Opin. Quart. = Public Opinion Quarterly; Res. on Lang. & Soc. Int. = Research on Language and Social Interaction.

^a Analysis of variance comparison of means: per-year citations: p < .001 (η² = .27); Mendeley: p < .001 (η² = .20); ResearchGate: p < .001 (η² = .22); social media: p < .05 (η² = .26).

^b Denotes per-year means, not means for total citations.

Regarding altmetrics (also shown in Table 3), articles appearing in the Journal of Communication were the most read on Mendeley, with 1,290.70 reads per article. This compares to 894.50 reads for the Journal of Computer-Mediated Communication in second place. The large difference in the readership means can be explained by the presence of Scheufele and Tewksbury’s (2007) article on agenda setting, which garnered 8,100 reads on Mendeley. Journal of Communication was also the most read publication on RG, with an average of 6,058.56 reads per article. Again, this can be explained by the presence of another highly read Scheufele (1999) article with 29,367 reads. The article discusses framing and media effects. Information Communication & Society was the most mentioned publication on social media with 25 mentions per article. Journal of Computer-Mediated Communication was a close second with 22.4 mentions per article—see Appendices B, C, and D for a list of the top 10 most read and mentioned articles. The most cited article on RG garnered a total 4,053 citations. This was Boyd and Ellison’s (2007) discussion of the definition, history, and scholarship of social networking sites. This was also the article with highest per-year citation (127.27 per-year citations from the Web of Science). ANOVA indicated significant differences among journals pertaining to all citation and altmetrics means (see Table 3). Geographically, only articles by first authors based in four regions were represented in the sample: Asia (2), Australia (1), Europe (31), Middle East (1), and North America (109). Articles from North America were the most cited. They were also the most read on both Mendeley and RG and the most mentioned on social media.

Citation and Altmetrics Results

The altmetrics data indicated that these metrics have diffused quite rapidly within communication research, a remarkable feat given their newness. Most articles (96%) had a Mendeley score for readership. Most had been read on RG (93%), and nearly half of all articles on RG (47%) were available for download. Also, nearly half of all articles had been mentioned on a social media platform (47%). This diffusion is important because data indicated that all three altmetrics significantly correlated with citation (see Table 4 for correlation coefficients and corresponding significance level values). Also, the negative association between age, Mendeley reads, and social media mentions shows that recently published articles are more likely to get attention than older articles. This is good for newer research articles because this way they get the needed attention faster.

Table 4.

Correlations Among Citations and Altmetrics.

	Mean^a	SD	1	2	3	4	5	6
Total citations	2.23	0.40	1
Per-year citations	1.04	0.34	0.74**	1
Mendeley reads	2.12	0.66	0.35**	0.57**	1
ResearchGate reads	2.19	0.92	0.30**	0.39**	0.27**	1
Social media mentions	0.36	0.48	0.22**	0.43**	0.40**	0.11	1
Age	1.10	0.32	0.56**	−0.14	−0.18*	−0.025	−0.21*	1

Note. SD = standard deviation.

^a Logarithmic transformed means shown.

*p < .05. **p < .001.

Regarding Mendeley reads, data showed a positive correlation with total citations (r = .35) and per-year citations (r = .57). RG reads also showed a correlation with total citations (r = .30) and per-year citation (r = .39). Similarly, social media mentions were positively correlated with both total citations (r = .22) and per-year citations (r = .43). Ordinary least squares (OLS) regression analysis also confirmed a significant relationship between per-year citation and all altmetrics: Mendeley reads (β = .40; t = 5.77; p < .001), RG reads (β = .25; t = 3.89; p < .001), and social media mentions (β = .25, t = 3.62, p < .001, R² = .43, $R_{adj}^{2}$ = .42). Additionally, both Mendeley reads (β = .25, t = 2.91, p < .001) and RG reads (β = .22, t = 2.88, p < .001, R ² = .18, $R_{adj}^{2}$ = .16) showed a significant relationship with total citations. Social media mentions did not.

Because the results indicate a positive relationship between altmetrics and citation, it is prudent to explore how thee altmetrics interact among themselves. Results show that mostly, they interact positively. For instance, Mendeley reads were correlated with both RG reads (r = .27, p < .001) and social media mentions (r = .40, p < .001). RG reads did not show a significant correlation with social media mentions (r = .11). When predicting Mendeley reads using RG and social media mentions, OLS regression confirms the existence of significant relationships: RG (β = .22, t = 3.04, p < .001) and social media mentions (β = .37, t = 5.05, p < .001, R ² = .21; $R_{adj}^{2}$ = .20). These particular findings are important because they suggest that these metrics complement each other rather than conflict with one another. It also means that the presence of an article in an online repository positively impacts readership and improves the chances of citation.

RG Results

Lastly, I report results for correlations among metrics within RG only. Examining RG as a stand-alone source is important because it is unique from Mendeley on several grounds. First, RG provides citation metrics from articles housed within the website, which gives an accurate measure of attention within this community of scholars. Mendeley does not provide citation data. Second, the availability of downloadable articles on the site, without subscription requirements, might improve access and attention to the articles because of easy availability. Even for those articles not available for download, users can request full texts from specific authors at no fee. Additionally, RG sends notifications to users whenever someone in their network updates their profile, for instance, when someone uploads a new article or if someone cites their work. On the other hand, most articles on Mendeley, especially those from major journals, are only available from the publisher’s website, mostly via subscription. Third, and as shown in Table 2, RG had nearly 3 times the reads articles on Mendeley received. Even without outliers, RG still generated almost double the average number of reads that articles on Mendeley received with its own outliers considered (609.47 vs. 314.65).

The results indicated that RG’s reads had significant correlations with all metrics from the site (Table 5). For instance, reads were highly correlated with total citations (r = .73) and per-year citations (r = .73). The availability of downloadable papers also improves reads (r = .55), as well as improving the chances of both types of citations. Also, coauthorship significantly improved reads (r = .48), downloads (r = .40), and both total and per-year citations (r = .36, r = .40). Coauthorship also improved a member’s RG Score (r = .35), which was positively correlated with both citation types. Articles posted by high scoring members were also more likely to be downloadable (r = .62) and be read (r = .47). T-test analysis further indicated that articles available for download were read and cited more, as were those that were coauthored (Table 6). OLS regression analysis also confirmed that RG reads had a strong relationship with both per-year citations (β = .74, t = 10.35, p < .001, R ² = .56, $R_{adj}^{2}$ = .54) and total citations (β = .83, t = 12.10, p < .001, R ² = .59, $R_{adj}^{2}$ = .58).

Table 5.

Correlations Among ResearchGate Metrics.

	Mean	SD	1	2	3	4	5	6
ResearchGate reads^a	2.19	0.92	1
RG total citations^a	2.24	0.79	0.73**	1
RG per-year citations^a	1.21	0.50	0.73**	0.89**	1
Downloads^b	—	—	0.55**	0.22**	0.33**	1
Number of authors	1.68	1.04	0.48**	0.36**	0.40**	0.40**	1
ResearchGate score	15.08	14.20	0.47**	0.34**	0.38**	0.62**	0.35**	1

Note. RG = ResearchGate; SD = standard deviation.

^a Logarithmic transformed means shown.

^b Downloads is a dichotomous variable—biserial correlations shown.

*p < .05. **p < .001.

Table 6.

Means Comparisons for ResearchGate Coauthorship and Download.

	Coauthorship		Download
	Yes	No	Yes	No
RG total citations	2.40 (0.47)	2.10 (0.98)**	2.42 (0.45)	2.10 (0.96)**
RG per-year citations	1.34 (0.37)	1.10 (0.57)**	1.38 (0.34)	1.05 (0.56)**
ResearchGate reads	2.52 (0.68)	1.88 (1.02)**	2.73 (0.63)	1.72 (0.88)**

Note. Logarithmic transformed means shown. Standard deviations are in parenthesis. RG = ResearchGate.

*p < .05. **p < .001.

Discussion

To add to the body of knowledge is a well recited adage in academia. The dissemination of that knowledge is equally important, and scholars have long pondered this process. Wilson, Petticrew, Calnan, and Nazareth (2010) aptly capture this sentiment in defining dissemination of research as “a planned process that involves consideration of target audiences and the settings in which research findings are to be received and, where appropriate, communicating and interacting with wider policy” (p. 2). Social media allows scholars to disseminate and share their work to a wider spectrum beyond the traditional channels of dissemination such as journal subscriptions, conferences, and conventions. We can now upload articles on academic social networking sites such as RG and Mendeley, read and comment on other scholars’ works, tweet, and even blog about the same.

The data reported here indicate that the new metrics of dissemination, that is, altmetrics, are useful toward this end. The positive relationships among altmetrics and traditional citation metrics indicate that it does not hurt for scholars to access these new media venues for research purposes. For instance, coauthorship improved attention to articles and citation on RG. This suggests that the more online presence that authors have, the more attention their work will get. I came across many instances where coauthors had inactive accounts on these sites. This means that there is room for improvement in disseminating work. It is also laudatory that most articles analyzed here were read on Mendeley and RG (96% and 93%, respectively). This is an improvement from Priem et al.,’s (2012) study of 24,331 articles in the Public Library of Science database, where 80% of articles were on Mendeley. However, the numbers within communication research stand to improve because less than half of the articles analyzed here had been mentioned on social media. Additionally, 4 of the 15 publications under study did not register a single mention on social media.

A simple solution is for scholars to join academic social networking sites. One does not need to upload articles because other research-related files such as data sets can be shared. Additionally, scholars can engage in question and answer sessions where members crowdsource for solutions. Both Mendeley and RG host questions on issues ranging from plugin issues to queries on multiple regression. Users can also post about upcoming and ongoing projects, and this way they can get suggestions and potential coauthors. Professors could also encourage students to join the many groups available or even form their own. Mendeley lists groups of all kinds of communication-related interests such as social media research, methodology, news, theses, and dissertations. The data also showed that there was no relationship between social media mentions and RG. Both Mendeley and RG provide links where users can connect with others on their Facebook, Twitter, and other social media accounts. This could be a conduit through which to disseminate scholarly work into the nonacademic sphere, especially in cases of research that generates interest among the wider public.

It is important to reiterate that the availability of downloadable articles on RG improved both readership and citation. However, only 47% of articles were available for download at the time of data collection. This number speaks only of journal articles and not conference papers or book chapters. This means that article downloads also stand to improve. Of course, authors should be wary of copyright and other legal issues when uploading articles online. Most journal publishers, however, encourage authors to do just that under fair access and usage rules. For instance, Elsevier allows authors to share preprint versions of accepted articles on most academic social networking sites. This publisher even provides a step-by-step tutorial on how to increase the attention given to an article, including how to use social media for this purpose (Elsevier, 2017a, para. 1; 2017b, para. 2). Elsevier’s statement on the article upload issue should be encouraging to wary scholars: “Now that your article is published, you can promote it to make a bigger impact with your research…Sharing your article is an important part of research and it’s important to share responsibly” (Elsevier, 2017c, para. 2).

I finish by addressing a common criticism that altmetrics have faced. Some view these metrics as an exercise in vanity. An example would be Hall’s (2014) criticism of these metrics by giving them the unfavorable moniker, the Kardashian Index (K-index). The index measures the disjunction between a scholar’s social media following and his or her citation count. The premise here is that some scientists have reached such high levels of fame on social media but produced little in the way of publications and/or citations. This is also a reference to reality television stars who become rich and famous, not because they are talented but due to celebrity and notoriety. There is much to dispute this comparison. For one, most users on academic social networking sites are researchers with scholarly intentions, not starstruck followers. Second, the burgeoning research on altmetrics shows that there’s a positive relationship between such metrics and article citation. Third, the results here show that the most cited journals in communication also happen to be the most likely to get attention via altmetrics. This denotes the complementary rather than the disruptive role of altmetrics. Besides, astrophysicist Neil deGrasse Tyson, who tops most K-index lists (You, 2014, para. 4), is renowned for his dissemination of science facts via the award-winning NOVA ScienceNow series, his social media presence notwithstanding. Coincidentally, as of March 2019, Hall’s article has garnered over 2,297 reads on Mendeley and 2,469 mentions on Twitter.

Footnotes

Appendix A

Correlations Among Citations and Altmetrics With Raw Data.

	Mean	SD	1	2	3	4	5
Total citations	256.07	283.70	1
Per-year citations	15.30	15.45	0.72**	1
Mendeley reads	314.65	736.34	0.24**	0.40**	1
ResearchGate reads	874.75	2,805.40	0.32**	0.30**	0.38**	1
Social media mentions	4.37	11.55	0.22**	0.51**	0.32**	0.10	1

Note. Raw means shown. SD = standard deviation.

*p < .05. **p < .001.

Appendix B

The Most Read Articles on Mendeley.

Author(s)	Year	Mendeley	Journal Title	Article DOI
Scheufele and Tewksbury	2007	8,100	Journal of Communication	10.1111/j.0021-9916.2007.00326.x
Ellison et al.	2007	2,812	Journal of Computer-Mediated Communication	10.1111/j.1083-6101.2007.00367.x
Boyd, Danah, and Ellison	2007	2,409	Journal of Computer-Mediated Communication	10.1111/j.1083-6101.2007.00393.x
Entman	1993	1,448	Journal of Communication	10.1111/j.1460-2466.1993.tb01304.x
Scheufele	1999	1,204	Journal of Communication	10.1111/j.1460-2466.1999.tb02784.x
Marwick and Boyd	2011	1,144	New Media & Society	10.1177/1461444810365313
Steuer	1992	975	Journal of Communication	10.1111/j.1460-2466.1992.tb00812.x
Livingstone	2008	897	New Media & Society	10.1177/1461444808089415
Debatin et al.	2009	765	Journal of Computer-Mediated Communication	10.1111/j.1083-6101.2009.01494.x

Appendix C

The Most Read Articles on ResearchGate.

Author(s)	Year	RG Reads	Journal Title	Article DOI
Scheufele	1999	2,9367.00	Journal of Communication	10.1111/j.1460-2466.1999.tb02784.x
Entman	1993	1,1643.00	Journal of Communication	10.1111/j.1460-2466.1993.tb01304.x
Scheufele and Tewksbury	2007	8,075.00	Journal of Communication	10.1111/j.0021-9916.2007.00326.x
Lombard et al.	2002	7,308.00	Human Comm. Research	10.1111/j.1468-2958.2002.tb00826.x
Ellison et al.	2007	5,500.00	Journal of Computer-Mediated Communication	10.1111/j.1083-6101.2007.00367.x
Waters	2009	5,314.00	Public Relations Review	10.1016/j.pubrev.2009.01.0
Kent and Taylor	2002	5,201.00	Public Relations Review	S0363-8111(02)00108-X
Kaplowitz et al.	2004	4,807.00	Public Opinion Quarterly	10.1093/poq/nfh006
Boyd and Ellison	2007	2,481.00	Journal of Computer-Mediated Communication	10.1111/j.1083-6101.2007.00393.x
Slater and Rouner	2002	2,313.00	Communication Theory	10.1111/j.1468-2885.2002.tb00265.x

Note. RG = ResearchGate.

Appendix D

Articles with the Most Mentions on Social Media.

Author(s)	Year	Mentions	Journal Title	Article DOI
Boyd and Ellison	2007	79	Journal of Computer-Mediated Communication	10.1111/j.1083-6101.2007.00393.x
Ellison et al.	2007	67	Journal of Computer-Mediated Communication	10.1111/j.1083-6101.2007.00367.x
Marwick and Boyd	2011	66	New Media & Society	10.1177/1461444810365313
Bennett and Segerberg	2012	56	Information Communication & Society	10.1086/678440
Lewis	2012	30	Information Communication & Society	10.1080/1369118X.2012.674150
Gil de Zúñiga	2012	23	Journal of Computer-Mediated Communication	10.1111/j.1083-6101.2012.01574.x
Livingstone	2008	20	New Media & Society	10.1177/1461444808089415
Noellene	1974	18	Journal of Communication	10.1111/j.1460-2466.1974.tb00367.x
Mccombs and Shaw	1972	13	Public Opinion Quarterly	https://www-jstor-org.web.bisu.edu.cn/stable/2747787
Fahy and Nisbet	2011	12	Journalism	10.1177/1464884911412697

Acknowledgement

I would like to thank Dr. Renita Coleman (University of Texas at Austin) for her counsel and encouragement towards this project.

Data Availability

The data are available from the author upon request. Email author at ben.wasike@utrgv.edu

Declaration of Conflicting Interests

The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author received no financial support for the research, authorship, and/or publication of this article.

Software Information

Data were analyzed using the Statistical Package for the Social Sciences (SPSS) software version 24.

References

Adler

N. J.

Harzing

(2009). When knowledge wins: Transcending the sense and nonsense of academic rankings. Academy of Management Learning & Education, 8, 72–95. doi:10.5465/AMLE.2009.37012181

Altmetric.com. (2016, November 24). How is the altmetric attention score calculated? Retrieved from https://help.altmetric.com/support/solutions/articles/6000060969-how-is-the-altmetric-score-calculated-

Altmetric.org. (n.d.). What are altmetrics? Capturing the online attention surrounding scholarly content. Retrieved from https://www.altmetric.com/about-altmetrics/what-are-altmetrics/

Antell

Foote

(2016). Scholarly publishing’s evolving landscape: Impact metrics, electronic-only journals, and open access in journalism and communication research. Journalism & Mass Communication Educator, 71, 309–328. doi:10.1177/1077695816668864

Antonakis

Bastardoz

Liu

Schriesheim

C. A.

(2014). What makes articles highly cited? The Leadership Quarterly, 25, 152–179. doi:10.1016/j.leaqua.2013.10.014

Bajwa

N. H.

König

C. J.

(2019). How much is research in the top journals of industrial/organizational psychology dominated by authors from the U.S.? Scientometrics. Advance online publication. doi:10.1007/s11192-019-03180-2

Blackman

(2016). Social media and the politics of small data: Post publication peer review and academic value. Theory, Culture & Society, 33, 3–26. doi:10.1177/0263276415590002

Bland

J. M.

Altman

D. G.

Rohlf

F. J.

(2013). In defence of logarithmic transformations. Statistics in Medicine, 32, 3766–3768. doi:10.1002/sim.5772

Borgman

C. L.

(1989). Bibliometrics and scholarly communication. Communication Research, 16, 583. doi:10.1002/aris.1440360102

10.

Boyd

D. M.

Ellison

N. B.

(2007). Social network sites: Definition, history, and scholarship. Journal of Computer-Mediated Communication, 13, 210–230. doi:10.1111/j.1083-6101.2007.00393.x

11.

Brandon

K. P.

Midway

S. R.

Sackett

Lynch

Cooney

P. B.

(2016). Twitter predicts citation rates of ecological research. PLoS One, 11, e0166570. doi:10.1371/journal.pone.0166570

12.

Casson

R. J.

Al-Qureshi

S. H.

(2010). Scientific information, journal impact factors and editorial policy: Editorial. Clinical & Experimental Ophthalmology, 38, 655–656. doi:10.1111/j.1442-9071.2010.02395.x

13.

Chul-joo

Dongyoung

(2016). Mapping the social capital research in communication: A bibliometric analysis. Journalism & Mass Communication Quarterly, 93, 728–749. doi:10.1177/1077699015610074

14.

Cohn

E. G.

Farrington

D. P.

Iratzoqui

(2017). Changes in the most-cited scholars and works over 25 years: The evolution of the field of criminology and criminal justice. Journal of Criminal Justice Education, 28, 25–51. doi:10.1080/10511253.2016.1153686

15.

Correia

Cassola

Azevedo

Pinheiro

Morgado

Martins

… Paredes

(2014). Meta-theoretic assumptions and bibliometric evidence assessment on 3-D virtual worlds as collaborative learning ecosystems. Journal of Virtual Worlds Research, 7, 1–18. doi:10.4101/jvwr.v7i3.7036

16.

Dobbs

(2013, April). When the rebel alliance sells out. The New Yorker. Retrieved from http://www.newyorker.com/tech/elements/when-the-rebel-alliance-sells-out

17.

Dowling

G. R.

(2014). Playing the citations game: From publish or perish to be cited or sidelined. Australasian Marketing Journal, 22, 280–287. doi:10.1016/j.ausmj.2014.09.005

18.

Ebrahimy

Mehrad

Setareh

Hosseinchari

(2016). Path analysis of the relationship between visibility and citation: The mediating roles of save, discussion, and recommendation metrics. Scientometrics, 109, 1497–1510. doi:10.1007/s11192-016-2130-z

19.

Elsevier. (2017a). Article sharing. Retrieved from https://www-elsevier-com-s.web.bisu.edu.cn/about/company-information/policies/sharing

20.

Elsevier. (2017b). Get noticed. Increase the impact of your research. Retrieved from https://www-elsevier-com-s.web.bisu.edu.cn/authors/journal-authors/submit-your-paper/sharing-and-promoting-your-article

21.

Elsevier. (2017c). Sharing and promoting your article. Retrieved from https://www-elsevier-com-s.web.bisu.edu.cn/authors/journal-authors/submit-your-paper/sharing-and-promoting-your-article

22.

Ferris

F. L.

Miller

K. M.

Glassman

A. R.

Beck

R. W.

, & Diabetic Retinopathy Clinical Research Network. (2010). A proposed method of logarithmic transformation of optical coherence tomography data for use in clinical research. Ophthalmology, 117, 1512–1516. doi:10.1016/j.ophtha.2009.12.014

23.

Garfield

Malin

M. V.

(1968). Can Nobel Prize winners be predicted? Paper presented at 135th Annual Meeting of the American Association for the Advancement of Science, Dallas, TX.

24.

Gingras

Wallace

M. L.

(2010). Why it has become more difficult to predict Nobel Prize winners: A bibliometric analysis of nominees and winners of the chemistry and physics prizes (1901–2007). Scientometrics, 82, 401–412. doi:10.1007/s11192-009-0035-9

25.

Google Scholar. (2017). Google Scholar metrics. Google. Retrieved from https://scholar.google.com/intl/en/scholar/metrics.html#metrics

26.

Guo

Hudders

Duffy

(2019). Product placement in mass media: a review and bibliometric analysis. Journal of Advertising, 48, 215–231. https://doi.org/10.1080/00913367.2019.1567409

27.

Hall

(2014). The Kardashian index: A measure of discrepant social media profile for scientists. Genome Biology, 15, 424–436. doi:10.1186/s13059-014-0424-0

28.

Han

Aung

H. H.

Erdt

Tai-Quan

Raamkumar

Yin-Leng

(2019). Social media presence of scholarly journals. Journal of the Association for Information Science and Technology, 70, 256–270. doi:10.1002/asi.24124

29.

Harker

Saffer

(2018). Mapping a subfield’s sociology of science: A 25-year network and bibliometric analysis of the knowledge construction of sports crisis communication. Journal of Sport & Social Issues, 42, 369–392. doi:10.1177/0193723518790011

30.

Henning

(2008). How our name evolved from B-movie monster to Mendeley. [Web log post]. Retrieved from https://blog.mendeley.com/2008/04/23/how-our-name-evolved-from-b-movie-monster-to-mendeley/

31.

Hoffman

D. L.

Holbrook

M. B.

(1993). The intellectual structure of consumer research: A bibliometric study of author cocitations in the first 15 years of the journal of consumer research. Journal of Consumer Research, 19, 505–517.

32.

E.-J.

Pasadeos

Ertem-Eray

(2019). Growth of public relations research networks: A bibliometric analysis. Journal of Public Relations Research, 31, 5–31. doi:1080/1062726X.2019.15777390

33.

Koo

(2017). A bibliometric analysis of two decades of aromatherapy research. BMC Research Notes, 10, 46. doi:10.1186/s13104-016-2371 -1

34.

Lei

Liao

(2017). Publications in linguistics journals from Mainland China, Hong Kong, Taiwan, and Macau (2003–2012): A bibliometric analysis. Journal of Quantitative Linguistics, 24, 54–64. doi:10.1080/09296174.2016.1260274

35.

Liang

Zhang

Zhu

(2015). Logarithmic transformation is essential for statistical analysis of fungicide EC50 values. Journal of Phytopathology, 163, 456–464. doi:10.1111/jph.12342

36.

Lievrouw

L. A.

(1989). The invisible college reconsidered. Communication Research, 16, 615. doi:10.1177/009365089016005004

37.

Liu

Fang

(2017). What we can learn from tweets linking to research papers. Scientometrics, 111, 349–369. doi:10.1007/s11192-017-2279-0

38.

Mäkelä

Lagström

Kaljonen

Simell

Niinikoski

(2013). Response to the letter to the editor “indices of insulin sensitivity and resistance: Adequate logarithmic transformation are needed to keep mathematical equivalence” by Tomoyuki Kawada. Early Human Development, 89, 517. doi:10.1016/j.earlhumdev.2013.03.009

39.

Masip

Fernández-Quijada

(2011). Mapping communication research in Catalonia: A comparative analysis of publication patterns in scholarly journals (2007-2009). Catalan Journal of Communication & Cultural Studies, 3, 95–108. doi:10.1386/cjcs.3.1.95_7

40.

McCain

K. W.

(1989). Mapping authors in intellectual space: Population genetics in the 1980s. Communication Research, 16, 667. doi:10.1177/009365089016005007

41.

Mendeley. (2019, March 8). Join a community of over 6 million users. Retrieved from https://www-mendeley-com-s.web.bisu.edu.cn/research-network/community

42.

Merton

R. K.

(1968). The Matthew effect in science. Science, 159, 56–63. doi:10.1126/science.159.3810.56

43.

Miyamoto

Midorikawa

Nakayama

(1989). A view of studies on bibliometrics and related subjects in Japan. Communication Research, 16, 629–641.

44.

Muncy

J. A.

(1991). The journal of advertising: A twenty-year appraisal. Journal of Advertising, 20, 1–12.

45.

Navarro

Martín

(2013). Bibliometric analysis of research on women and advertising: Differences in print and audiovisual media. Comunicar, 21, 105–114. doi:10.3916/C41-2013-10

46.

Packard

G. C.

(2009). On the use of logarithmic transformations in allometric analyses. Journal of Theoretical Biology, 257, 515–518. doi:10.1016/j.jtbi.2008.10.016

47.

Packard

G. C.

(2013). Is logarithmic transformation necessary in allometry? Biological Journal of the Linnean Society, 109, 476–486. doi:10.1111/bij.12038

48.

Paisley

(1989). Bibliometrics, scholarly communication, and communication research. Communication Research, 16, 701. doi:10.1002/aris.1440360102

49.

Park

(2018). Research evaluation of Asian countries using altmetrics: Comparing South Korea, Japan, Taiwan, Singapore, and China. Scientometrics, 117, 771–788. doi:10.1007/s11192-018-2884-6

50.

Pasadeos

(1985). A bibliometric study of advertising citations. Journal of Advertising, 14, 52–68. doi:10.1080/00913367.1985.10672971

51.

Pasadeos

Renfro

(1992). A bibliometric analysis of public relations research. Journal of Public Relations Research, 4, 167–187. doi:10.1207/s1532754xjprr0403_03

52.

Plusquellec

Denault

(2018). The 1000 most cited papers on visible nonverbal behavior: A bibliometric analysis. Journal of Nonverbal Behavior, 42, 347–377. doi:10.1007/s10919-018-0280-9

53.

Priem

Piwowar

Hemminger

(2012). Altmetrics in the wild: Using social media to explore scholarly impact. Retrieved from https://arxiv.org/html/1203.4745v1

54.

Repiso

Castillo-Esparcia

Torres-Salinas

(2019). Altmetrics, alternative indicators for web of science communication studies journals. Scientometrics, 119, 941–958. doi:10.1007/s11192-019-03070-7

55.

ResearchGate. (2017). About us. ResearchGate is built by scientists, for scientists. Retrieved from https://www.researchgate.net/about

56.

ResearchGate. (2019). RG score. Retrieved from https://explore.researchgate.net/display/support/RG+Score

57.

Rice

R. E.

Chapin

(1996). What’s in a name? Bibliometric analysis of 40 years of the Journal of Broadcasting (& Electronic Media). Journal of Broadcasting & Electronic Media, 40, 511. doi:10.1080/08838159609364373

58.

Roemer

R. C.

Borchardt

(2015). Introduction to altmetrics. Library Technology Reports, 51, 5–10.

59.

Rovira-Esteva

Aixelá

J. F.

Olalla-Soler

(2019). Citation patterns in translation studies: A format-dependent bibliometric analysis. Translation & Interpreting, 11, 147–171. doi:10.12807/ti.111201.2019.a09

60.

Scheufele

D. A.

(1999). Framing as a theory of media effects. Journal of Communication, 49, 103–122. doi:10.1111/j.1460-2466.1999.tb02784.x

61.

Scheufele

D. A.

Tewksbury

(2007). Framing, agenda setting, and priming: The evolution of three media effects models. Journal of Communication, 57, 9–20. doi:10.1111/j.0021-9916.2007.00326.x

62.

SCImago. (2016). Understanding indicators, tables and charts. Thomson Reuters. Retrieved from http://www.scimagojr.com/help.php#rank_journals

63.

Small

Greenlee

(1989). A co-citation study of AIDS research. Communication Research, 16, 642. doi:10.1177/009365089016005006

64.

Strevens

(2006). The role of the Matthew effect in science. Studies in History and Philosophy of Science, 37, 159–170. doi:10.1111/j.1460-2466.1999.tb02784.x

65.

Stuart

Faucette

Thomas

(2017). Author impact metrics in communication sciences and disorder research. Journal of Speech, Language, and Hearing Research, 60, 2704–2724. doi:10.1044/2017_JSLHR-H-16-0458

66.

Sud

Thelwall

(2014). Evaluating altmetrics. Scientometrics, 98, 1131–1143. doi:10.1007/s11192-013-1117-2

67.

Thelwall

Haustein

Larivière

Sugimoto

C. R.

(2013). Do altmetrics work? Twitter and ten other social web services. PLoS One, 8, 1–7. doi:10.1371/journal.pone.0064841

68.

Thomson Reuters. (2008). Using bibliometrics: A guide to evaluating research performance with citation data. Retrieved from http://ips.clarivate.com//m/pdfs/325133_thomson.pdf

69.

Torres-Salinas

Cabezas-Clavijo

Á.

Jiménez-Contreras

(2013). Altmetrics: New indicators for scientific communication in web 2.0. Comunicar, 21, 53–60.

70.

White

W. J.

(2011). The study of games in the ICA: A bibliometric analysis. Paper presented at the annual meeting of the International Communication Association, Boston, MA.

71.

Wilson

P. M.

Petticrew

Calnan

M. W.

Nazareth

(2010). Disseminating research findings: What should researchers do? A systematic scoping review of conceptual frameworks. Implementation Science: IS, 5, 91–91. doi:10.1186/1748-5908-5-91

72.

You

(2014, September 17). The top 50 science stars of Twitter. Science. Retrieved from http://www.sciencemag.org/news/2014/09/top-50-science-stars-twitter

73.

Xiao

Hemminger

Yang

(2017). Global science discussed in local altmetrics: Weibo and its comparison with Twitter. Journal of Informetrics, 11, 466–482. doi:10.1016/j.joi.2017.02.011

74.

Zahedi

Costas

Wouters

(2014). How well developed are altmetrics? A cross-disciplinary analysis of the presence of ‘alternative metrics’ in scientific publications. Scientometrics, 101, 1491–1513. doi:10.1007/s11192-014-1264-0

75.

Zi-yu

(2014). The contour of Library Materials About Macau: A bibliometric profile on the world stage. Intercultural Communication Studies, 23, 117–163.