Aim: The main aim of this study was to examine whether citation counts of highly cited papers, both NON-OA and OA, published by authors affiliated in Croatian scientific institutions correlated with Twitter mentions and Mendeley readers. Based on the obtained data, the author will try to answer the following questions: RQ1. What are the numbers of citations and altmetric indicators, both in a group of NON-OA and OA articles? RQ2. Which sources provide the highest number of altmetric indicators, both in a group of NON-OA and OA articles? RQ3.What is the average number and the median of the most common altmetric indicators for most highly cited NON-OA and OA articles? RQ4. Do citation counts of most highly cited NON-OA and OA articles of Croatian researchers correlate with Twitter mentions and Mendeley readers?
Methods: The study was divided into two stages. The first stage was to collect and filter the data from the Scopus database. The metrics data collected from Scopus were related to the authors affiliated in Croatian scientific institutions (simple search in Scopus “affiliation country”: Croatia). The chronological scope of the study covered the years 2013-2017.Citation counts were collected for most highly cited OA papers (N=71) and Non-OA papers(N=71). The sample had DOI assigned. The second stage involved the use of the Altmetric Explorer (http://www.altmeric.com) to check whether the citations counts of most highly cited articles were influenced by Twitter mentions and Mendeley readers. Altmetric Explorerwas rated as a reliable and stable source of altmetric indicators (1). The fact that it collects data using digital identifiers of the documents, makes Altmetric Explorer transparent (2). The sample covered articles with a non-zero altmetric score. Altmetric Explorer collects data fromdifferent sources: news mentios, blog mentions, policy mentions, twitter mentions, patent mentions, peer review mentions, Weibo mentions, Facebook mentions, Wikipedia mentions,Google+ mentions, LinkedIn mentions, Reddit mentions, Pinterest mentions, F1000 mentions,Q&A mentions, video mentions, syllabi mentions, Mendeley readers. The data was collected on 23rd of June 2018.
Results and Discussion: RQ1. The NON-OA articles collected 16,751 citations and 28,236altmetric indicators whereas OA articles collected 9,725 citations counts and 16,330 altmetric indicators. RQ2. The highest number of indicators in this group was provided by Mendeley –17,369 and Twitter 8,830. Next, there were news mentions (926), Facebook mentions (512)and blog mentions (355). Other sources provided 244 mentions. In the grop up of The highest number of indicators in this group was provided by Mendeley – 8,454 and Twitter 6,823.Next, there were news mentions (511), Facebook mentions (232) and blog mentions (149).The remaining sources provided 161 mentions. The analyzed sample covered articles published in the highest quality journals. The previous studies showed that higher quality content has more Mendeley readers and tweets (3). Studies carried out by Croatian authors on papers published by the University of Zagreb School of Medicine showed that articles published in journals of international renown had much more intense altmetric activity than articles published in local journals (4). RQ3. The average number of Twitter mentions for most highly cited NON-OA articles was 126 and Mendeley indicators was 248. The median value of Mendeley readers was 148 and that of tweets 28. The average number of Twitter mentions for most highly cited OA articles was 97 and Mendeley readers was 120 whereas the median value was 85 for Mendeley and 25 for Twitter mentions. RQ4. The number of citations of NON-OA articles correlated significantly with both Mendeley readers and Twitter mentions(Spearman Correlation: Mendeley p=0.0005; Twitter p<0.0001). In the case of OA articles,citations counts correlated significantly only with Twitter mentions (Spearman Correlation:Twitter p=0.0001; Mendeley p=0.0508).
Conclusion: It is proved that citation counts of highly cited papers published by authors affiliated in Croatian scientific institutions correlate with both Twitter mentions as well as Mendeley readers in the group of NON-OA articles. In the case of OA papers citation counts correlate only with Twitter posts.