References provide a means for acknowledging previous publications, integrating new studies with previous research works, and identifying primary sources supporting authors' statements (2). They facilitate continuity and novelty of research by avoiding redundant work. Previous publications may drive the scientific discourse and arguments. Acknowledging all pertinent sources may foster an exchange of ideas and data.

Proper citations scrutinize cited works and aid in elaborating new concepts (5). Authors should pick accessible and validated references to enable verification of the original sources by evaluators. It is therefore important to cite primarily peer-reviewed sources passed through multiple checks by reviewers and editors (1).

Authors should be skilled to search and cite references. Searches are largely dependent on ideas behind the objectives of research studies, and can be done through multidisciplinary and/or specialized databases by setting time limits, languages of the required sources, and article types (6). Authors themselves are responsible for correct searches, which can be expanded by legitimate requests of peer reviewers.

Inaccurate reference lists negatively affect the indexability and influence of a scholarly journal. Authors who cite references without retrieving and reading related full-texts may increase inaccuracies. To some extent, such practice is driven by skewed perceptions of the role of citations in boosting the notorious Impact Factor (IF) (5).

Accurate referencing is crucial for allowing readers to follow the flow of ideas and statements in scholarly works and for ensuring the integrity of science communication. Incorrect citation of author names, journal titles, volumes, or page numbers makes it difficult or even impossible to locate primary sources. Irrelevant in-text citations may result in the misrepresentation of the sources and dissemination of unchecked statements, ultimately raising concerns over 'vanity publishing'.

Online publishing has changed reference formatting and interlinking. As a prime example, tagging sources with Digital Object Identifiers (DOIs) has made them more functional and easily searchable. Open access movement and availability of various online communication and archiving platforms (e.g., blogs, Research Gate, LinkedIn) have increased the use and functionality of references.

In the digital era, new scholarly sources have emerged and add more complexity to reference formatting and citing. For instance, web-based materials require listing correct URL linking routes and access dates to facilitate re-access by readers at least within a certain period.

In line with the digitization, reference management software tools have emerged to generate text reference citations. However, these software tools may become cumbersome and duplicate references (1). They rarely correct chronological and alphabetical arranging details when several references are grouped together at first use (1). To overcome such problems, authors may use several software tools to manage their references and bibliographies in different styles (e.g., EndNote) (2).

Currently, individual and journal impact indicators influence academic promotion and research grant allocation globally. These indicators are heavily dependent on the accuracy of references. Electronic publishing now provides reference linking and swift navigation from primary and secondary sources (4). Importantly, reference linking not only cites a related work, but also provides a clickable link to additional information presented in the work (4). For instance, many publishers now provide e-links to gene sequences, chemical structures, multimedia, and Web data.

Impact analyses now consider the frequency, directions, and networks of citations in scholarly articles. In some cases, these may reveal 'citation cartels', abundant auto-citations and other inappropriate citation practices, which disqualify individuals and journals. Citation analyses are thus impossible with inaccurate references.

Current strategies of navigating through the references by CrossRef matching, Highwire, PubMed linking, or XML-encoding help in detecting and correcting inaccuracies. Many large publishers have already incorporated these navigation systems in the peer review systems.

Different styles of citing scholarly works have been practiced globally (5). The citing styles range from selective citation of studies to support the authors' statements, referring to the studies that support the authors' viewpoints, auto-citations to boost the authors' own reputation, or citations for convenience. Some authors refer to the sources, including grey literature without reading and understanding their entirety. Others give multiple examples of similar references to support a single statement or, conversely, use a single source to support multiple statements.

Basically, the relevance and accuracy of cited references reflect the authors' professionalism and writing skills. The authors' duty is to continuously upgrade their skills of processing scholarly information and referring to essential sources. They are prime contributors to the accuracy of formatting in accordance with the target journal's instructions. Rechecking the relevance and format of each reference by searching through evidence-based bibliographic databases is also the authors' responsibility toward their readers (7).

By filtering quality submissions, journal editors should pay attention to the overall quality and relevance of the references. They can instruct the reviewers and authors on proper validation of cited sources. Some editors may also take a step forward and organize online or on-site training courses for their authors on acceptable citation styles and formatting.

An important issue for editors is to set quantitative limits for reference lists in the instructions for authors and avoid referencing duplicate and retracted items, which may threaten the trustworthiness of science communication (8).

Authors' editors, who edit manuscripts at presubmission and resubmission stages, should also take an active part in selecting, verifying, and formatting references. This task requires from them deep knowledge of the role of references in the 'big science' era and the magnitude of any (un)intentional error.

Peer reviewers are the gatekeepers of current scholarly communications who are in the best position to pick incorrect or irrelevant references and suggest amendments. They should recheck the relevance and correctness of the reference lists, advise ethically sound additional references (preferably not auto-citations), or suggest shortening the lists by omitting redundant and retracted sources. Adding references is particularly required when the authors' statements are neither their own words of wisdom nor general knowledge.

A growing number of publishers use editorial management software to digitize the validation of references by the reviewers. Such an approach may increase the efficiency of publishing reliable and highly informative reference lists. Journals still relying on manual checks and uncertain policies over reference validation risk to be excluded from the global pool of science communication.

Some publishers employ stringent copyediting mechanisms for reference checking (4). Copyeditors either work on original Word format files with or without macros for reference editing or use XML-based editing tools or typesetting software. Copyeditors are now also equipped with access to multiple open and subscription platforms and databases, helping them in locating references and correcting the lists.

Other publishers have implemented effective practices to improve reference linking (4). Higher link-matching rates are now achieved thanks to the rigorous copyediting supported by automated reference extraction and checking processes through Manuscript Central™ and Editorial Manager™ programs (e.g., Aries Systems' eXtyles technology, Thomson Reuters' ScholarOne, Inera's eXtyles, SPi's SPiCE) (4).