Contextual Analysis of Scientific Publications for Advancing Writing Skills (CASPArS): Self-Learning for Science Writing Using Top Scientists’ Literature

Abstract

Writing is an essential component of scientific activity. As such, it is necessary to develop strategies to provide equitable training opportunities for science writing. In order to provide learners with ways to improve their writing regardless of their language background and/or institutional and departmental environments, this article describes a self-learning strategy, Contextual Analysis of Scientific Publications for Advancing writing Skills (CASPArS), that employs contextual analysis of top scientists’ publications. By using full-text search software to analyze several hundred selected publications, the CASPArS method affords a visual representation of proper usage of words/phrases of interest, facilitating learners to understand writing patterns and rules in leading scientific papers. The utility of the method has been demonstrated with several examples by using 500 non-open-access papers of synthetic organic chemistry and chemical biology fields from various journals such as Nature and Science. In order to increase the accessibility of the method for diverse communities around the world, a free-search software sin3rou and 500 open-access chemistry papers (e.g., ACS Central Science and Chemical Science) by 150 U.S. researchers have been included in the Supporting Information (available online with the full paper). Using this CASPARS strategy, learners/students are able to develop appropriate writing through “professors’ eyes” by taking advantage of established scientific writers’ experience and knowledge of language usage.

Graphical Abstract

INTRODUCTION

Scientific writing is an integral part of dissemination of research and an essential educational component for various degrees as well as academic/industrial career development.^1,2 A common method of writing guidance in scientific graduate education is encouraging students to read and write frequently to improve their writing skills.³ This strategy captures the essence that good writing is learned by understanding and mastering common writing patterns in previously published work. However, without some clear instruction and writing tools, this strategy is analogous to expecting someone to become a great musician by being handed an instrument and recordings of great works. Plausibly, it would be unlikely for learners in chemistry, or any field, to become confident writing original research articles by themselves even at the end of their graduate study unless substantial training programs of science writing have been offered.⁴ The challenges of quality writing are exacerbated for non-native English speakers,⁵ with significant scientific writing training in English having become virtually imperative.⁶ A lack of effective writing instruction increases the likelihood of poor writing in international peer-reviewed journal publications. Multiple manuscripts have highlighted common problems in scientific writing of published literature such as paragraph structure,⁷ demonstrative pronouns,⁸ incomplete comparison,⁸ and superlatives,^9,10 summarized in Figure 1A. While numerous strategies to address the issues of quality writing are available (e.g., ACS virtual issues of writing art mastery,^11,12 increasing science impact/identifying scientific problems,^13,14 and figure presentation^15,16),^17–21 it is still challenging for learners to know whether writing patterns and expressions used in their own manuscripts are suitable for scientific publication.

Figure 1.

Issues and strategies for science writing. (A) Representative examples of frequent mistakes of learners in science writing. (B) Schematic depiction of strategies for science writing improvement. (C) Contextual Analysis of Scientific Publications for Advancing writing Skills (CASPArS) that allows learners to improve their writing skills at a personal level regardless of their institutional, departmental, or laboratory situations.

The availability of opportunities to learn scientific writing is often dependent on institutional, departmental, or even individual research group situations. Because writing instruction (like writing itself) is time and personnel intensive, the availability and effectiveness of writing assistance is highly variable (Figure 1B), making equitable access challenging. Organization level instruction, such as English training courses for international students, provide access to large numbers of individuals, but such courses tend to provide general writing techniques and lack subject-specific guidance. Department-level, subject-specific workshops or courses can provide better field-specific training that is still accessible for many students simultaneously;⁴ however, such opportunities are often dependent on a professor/instructor(s) willingness to provide writing instruction as above-load service. Similarly, availability of writing mentorship in specific research groups by a supervisor or senior student/scientist cannot be controlled by a learner, even though supervisor’s mentorship is known to have significant influence on positive learning experiences.²² While institutions and departments should continue to provide more English writing courses/workshops/resources, there is a need for resources that are readily accessible to individual learners. Although the increased availability of online editing services may be a suitable option for learner-writers, the cost may be an impediment for many, and the quality and subject-specificity of such tools is quite varied.

In this Article, we introduce a self-learning strategy to enhance science writing through contextual analysis of published literature (Figure 1C and Supporting Information Figure S1). This method is accessible to diverse users regardless of a writer’s background or environment. Full-text search of selected literature enables facile identification and analysis of a particular word or idiom that learner-writers can then evaluate with respect to its usage in the given scientific field. This free tool allows for the visualization of frequency and usage examples and facilitates contextual analysis of the word or idiom in question. We present several examples of the contextual analysis in this manuscript using 500 non-open-access papers from top journals such as Nature, Science, and Journal of the American Chemical Society (JACS). However, the method can also utilize free open-access resources (the software sin3rou/its setup protocol and 500 open-access papers of 150 U.S. researchers such as those published in ACS Central Science and Chemical Science that are available as OA Files S1, S2, S3, and S4 in the Supporting Information), making it accessible to a wide variety of users. While the method was originally designed for writing manuscripts to be submitted for peer-reviewed journal publications, students and instructors in classroom settings can also take advantage of the search functions to enrich writing training as well.

METHOD

The Contextual Analysis of Scientific Publications for Advancing writing Skills (CASPArS) method utilizes full-text search software to analyze an extensive set of published literature and facilitate context-relevant word/phrase usage. This tool relies on analysis of a set of publications written by a selection of leading researchers in a given field. In the present embodiment, 500 non-open-access papers of 30 leading U.S. researchers in organic chemistry and biology fields were selected from various top journals (e.g., Nature journals, Science journals, JACS, Proceedings of the National Academy of Sciences U. S. A. (PNAS), Angewandte Chemie, etc.), with a focus on American English. The list of the 500 non-open-access papers and 30 U.S. researchers are shown in the Excel file of the Supporting Information (a different set of 500 open-access papers of 150 U.S. researchers are also available in Supporting Information OA Files S1, S2, S3, and S4 for learners to use the method; the open-access papers are from the American Chemical Society and Royal Society of Chemistry). Different groups of scientists can be chosen to emphasize different goals such as British English and/or other fields (e.g., material and physical science). The authors found the free software, sin3rou (also known as TanSaburou), to be most useful and convenient, since the total search results and preview of usage examples are displayed clearly in this software. The preparation and usage protocol of the software together with troubleshooting advice are described in the Supporting Information. Other search methods such as those built in Google Drive, Box, and Dropbox (as of February 2024) are also capable of full-text search of files, but the total results and/or previews are not as conveniently displayed (Supporting Information).

As an example of the use of the CASPArS method to provide contextual understanding of the usage of even basic wording (Figure 1A(6)), consider analysis of two words, “hypothetical” and “nevertheless” (Figure 2A). In the selected 500 papers, the word “hypothetical” was used at least once in 18 manuscripts (3.6%) and the word “nevertheless” in 84 of the manuscripts (16.8%). In addition to the number of usage results, the preview of results displays how the words are used, with representative examples shown in Figure 2A , where i−iii are for “hypothetical” and 1−3 for “nevertheless.” Examination of the results highlights that the word “hypothetical” is never used after a verb, but the word was exclusively used before a noun. This analysis demonstrates that standard usage of “hypothetical” in those 500 chemistry papers is placing the adjective before a noun, although placing the word after a verb like “is hypothetical” may not be grammatically wrong. The example of “nevertheless” shows its standard usage is at the beginning of a sentence (or after semicolons), unlike “although” that can be used at the beginning of a sentence, or as a contrasting transition between clauses.

Figure 2.

Analysis examples of Contextual Analysis of Scientific Publications for Advancing writing Skills (CASPArS). The search was performed using the sin3rou software for a selected set of 500 published papers (non-open access). (A) Search results and examples of analysis for words “hypothetical” and “nevertheless”. Three example sentences for each searched word are from literature: i,²³ ii,²⁴ iii,²⁵ 1,²⁶ 2,²⁷ and 3.²⁸ (B) Pie chart showing the search and analysis results of a word “enables” (216 results total). If a single paper used the word multiple times, a sentence using the word appearing earliest in the paper is used for analysis. (C) Bar graph showing the search results for a word “important” and words that have similar meaning to “important”. The similar words were selected from the search results of Google search (Oxford Languages) of “important synonyms”. (D) Analysis of selected “casual” and “formal” words to highlight the importance of contextual analysis for identifying the suitability of a word for science writing. Quotation marks are shown for “making it”, as the marks are often necessary for idiom search.

Contextual evaluation is also helpful for analyzing the usage of different verb forms. Consider the verb “enables,” which was used at least once in 216 papers of the 500 non-open-access manuscript data set (or 43%). Within this set, 197 papers (91%) placed a grammatical object after the verb rather than having the word followed by a second verb such as “to be” (e.g., “enables substitution” vs “enables something to be substituted”). The minor usage (“enables object to be action”) is not grammatically wrong. Rather, this contextual analysis underscores the more standard word usage in the scientific field.

The CASPArS tool can be invaluable to minimize word overuse (Figure 1A(5)). For example, the word “important” is arguably one of the most overused words in scientific literature to emphasize the value of the work. As shown in Figure 2C roughly half of the 500 non-open-access papers in our reference set used “important” at least once. Here, it is useful for a writer to start with a “Ctrl+F” search of their manuscript to identify their own frequency of usage, raising the question of whether use of the adjective “important” is important. Or whether it would be more effective to provide an explanation of significance/value of the object rather than just declaring it to be “important.” If determined necessary, the writer might then consider synonyms as a means to reduce repeated usage. Microsoft Word’s thesaurus (or a Google search) listed the synonyms of “important” as shown below, and while those existing tools are useful for discovering proper words in general writing, these tools do not provide any insight into the common usage of the terms in a given context. On the other hand, the CASPArS method using full-text search software provides the frequency of use to help determine common usage, as some of which are shown in Figure 2C and Table S1, with the sin3rou-based percentage of our 500 manuscripts that use the synonym indicated in (): critical (45%), essential (28%), significant (26%), crucial (14%), prominent (9%), vital (6%), principal (6%), chief (3%), imperative (1%), weighty (0.4%), high-ranking (0.2%), and influential (0%). Clearly, some synonyms find more common usage than others in the selected scientific literature. Herein, care must be given to synonym replacement strategies. For example, “important” and “significant” could carry distinct meaning.²⁴ If something is important to a process, the process will not proceed without it, by contrast, if it is significant to the process, the process may proceed without the “it” but the “it” substantially influences the process. In addition to understanding context to address the more subtle meaning of synonyms, CASPArS analysis also can be invaluable to avoid grammatically correct but contextually nonstandard usage. Writers should be aware that even though avoiding casual and/or colloquial wording can be a valid strategy since scientific writing is generally more formal, there are numerous exceptions of commonly used casual/colloquial words as well as uncommon formal words in science writing (Figure 2D).

While there are many conceivable ways to take advantage of this contextual analysis method, the authors propose that the CASPArS tool may be most useful for learners-writers in three general ways: Recalibration, Replacement, and Redevelopment (Figures 3 and S2). Recalibration, as an example is described in Figure 2B for the verb-usage of “enables”, is for learner-writers to convert their writing into more standard usage by understanding contextual usage. Replacement is particularly important to minimize overuse of words, as described in Figure 2C for the word “important.” However, as discussed above, care must be taken since synonyms are not necessarily interchangeable with one another (e.g., the case of “importance” and “significance”).²⁹ Redevelopment is more broadly pointing toward the self-critique process of writing for which the tools of recalibration and replacement may be useful. Thesaurus-type replacement writing is not better writing than word over usage, for example. As such, the tools of recalibration and replacement should not be applied in a mechanical fashion. The redevelopment process of writing is a practice that even experienced writers continually work to master, critically analyzing each word and phrase in an effort to improve communication. Overall, while the CASPArS method is granular, the application should not be. The method becomes the flag/catalyst to evaluate that “grain” in the broader context of the writing. The replacement component of the method is significantly word-granular, but introduction of the ideas of recalibration and redevelopment take the tool beyond just a word-level tool.

Figure 3.

Schematic description of the three Rs of Contextual Analysis of Scientific Publications for Advancing writing Skills (CASPArS) for learners of science writing. Top, recalibration (red); middle, replacement (cyan); bottom, redevelopment (yellow). The writing examples are from published literature.^30–32

In addition to its use to improve original writing, the three Rs of CASPArS can be variously implemented for science writing. CASPArS analysis could assist instructors/teaching assistants in grading student writing or enhance referees review of manuscripts submitted for publication. The CASPArS tool also offers a unique framework with which senior students can mentor younger students in science writing by providing useful, and contextually reliable feedback even when the mentoring students may have limited writing experience. Detailed protocols for utilizing the three Rs of CASPArS are included in Supporting Information, together with a setup protocol for search software including sin3rou.

To evaluate the possibility of using the method beyond a research setting, the CASPArS tool was used in a teaching setting, and the student perceptions of the tool were assessed using an anonymous survey (Figure S3). The survey was performed with the Institutional Review Board (IRB) approval of North Carolina State University (the details are described in the Supporting Information) and used the Likert scale with response options ranging from 1 (strongly disagree) to 5 (strongly agree). The group of 13 undergraduate students who responded to the survey used sin3rou for writing analysis. Based on the survey, the majority of students’ responses were 4 (46.2%) or 5 (46.2%) about whether the benefits of the method to improve science writing were clearly explained in lecture and supporting materials. The majority of the students (69%) responded with a 4 (15.4%) or 5 (53.8%) to the question about the clarity of the protocol for setting up the method. A total of 77% of students responded with a 4 (23.1%) or 5 (53.8%) about recommending the method for students and researchers for science writing. Although significant issues were not reported by students using the tool itself, difficulty with identifying the words and phrases that should be replaced through the recalibration and replacement processes was expressed. We anticipated this issue, and in the first round of analysis, the instructors highlighted three words/phrases for each student to focus on. In the second round, students were asked to identify three words/phrases themselves and find appropriate replacements. We found that undergraduate students can use the tool successfully (Figure S4) as long as they receive guidance in selecting words/phrases to be replaced. It is likely that repeated use of the tool would be beneficial for learners not only to use the tool better, but also to improve general analysis skills by examining writing in published literature. We anticipate that the challenges of identifying words/phrases for CASPArS evaluation should not be a major issue for senior graduate students and postdoctoral researchers. However, exposure to scientific writing through guidance with the CASPArS tool by an in-person mentor would be beneficial for undergraduate students. Although the survey was conducted with a relatively small group of undergraduate students, the results indicate that the CASPArS method can be useful for undergraduate-level class or teaching laboratory settings.

Limitations and Caveats

Although analysis of published literature may offer opportunities for enhanced self-learning experiences, users should take into account the following risks and pitfalls of the method. A risk of plagiarism should not be neglected because the learning method is predicated on applying other researchers’ writing to learners’ own, and users should be aware of the distinction between learning English and using others’ writing from the analyzed papers without sufficient changes.^33,34 The analytical method is also highly dependent on subjective judgements by users in the choice of papers for which substantial variation can exist between different fields and writing styles. Ideally, papers and journals for the full-text search should be selected by mentors or experts familiar with nuances of each field. Another risk of the method is a perpetuation of published misconceptions. For example, a reviewer of this paper during the peer-review process noted that the term “red-shift” is increasingly misused when describing the infrared (IR) region. The red/blue shift terminology is referenced to visible light. Both red and blue visible light are shifted in the same direction (i.e., to higher energy) from IR frequencies. The CASPArS method should not be considered a replacement of common science writing training that provides external feedback, as the method only addresses a selection of the common issues of science writing (Figure 1A5 and 1A6). Some words may not be easily analyzed properly by the method due to a common journal template; for instance, the search result of the word “introduction” tends to yield a high number of hits not only because of the common use of the word by writers but also because of journal’s labeling the first section of a paper as “Introduction”. In addition, inclusion of article titles in the cited literature is increasingly common in many journals, and it should be noted that words of the titles cited in papers would also appear as results even if the words were not used by an author of the paper that cited those papers. The users should also be aware that the method cannot differentiate words that contain the exact same words/letters (e.g., “hypothetically” and “domain” would appear when “hypothetical” and “main” are searched, respectively.). It is worth noting that if a word of interest is used multiple times in searched literature, the sin3rou software shows only the first example of the word usage in the preview, and this feature should be taken into account for analysis purposes like the example of “enables” in Figure 2B.

CONCLUSIONS

The presented method, Contextual Analysis of Scientific Publications for Advancing writing Skills (CASPArS), relies on contextual analysis of published literature by top scientists that can be done at an individual level, regardless of institutional or departmental resources provided to learners. Benefits of the method are providing learners opportunities to understand the standard use of certain words and idioms, which may eventually lead to future improvement of writing skills. It should be noted that the CASPArS approach is not meant to discourage creative writing, but the authors believe that visualization of typical usage of a certain word or idiom in top science publications can be of great help for learners to build a solid writing foundation. Even experts of science writing can make use of the approach for research purposes by comparing their own writing to others in order to obtain different perspectives. The tool can also be valuable for mentoring purposes by providing mentees with statistical analysis to show why certain phrasing may be more suitable than the others. Technological advances of search tools and artificial intelligence (AI) would potentially enhance the methods in the future, perhaps resulting in more facile and accessible search (e.g., online search of open-access papers through generative AI technologies) than the current version which requires several hundreds of PDF files on the users’ computer. Though using AI technologies for science writing is still controversial in many aspects,^35–38 the contextual analysis learning method could be combined with AI technologies for promoting writing training when clear rules and regulations of AI use become solidified in scientific communities.^39,40 Although ChatGPT can be helpful to new writers if the AI method is used in a specific framework,^41,42 the CASPArS method would provide a more field-specific tool and offer more independent opportunity of developing quality writing skills than ChatGPT-based approaches.

While this article focused on improving American English writing in chemistry- and biology-related fields, the strategy is applicable to other types of English/language and to scientific fields by selecting a different set of literature to be analyzed (sharing a list of literature in a research community can be assisted by a free tool such as Zotero where a web link of a list can be created and shared publicly. Examples of the links for the lists of this work can be found in the Excel file in Supporting Information). No universally perfect writing styles exist, with general manuscript preparation practices and language usage evolving over the time (e.g., use of personal pronouns,⁴³ multiple-panel figures,⁴⁴ and graph presentations⁴⁵). Moreover, when engaged in interdisciplinary research, it is important to recognize that language use and writing styles may vary between different fields.⁴⁶ The customizable capability of the CASPArS method can be helpful for writers to identify the most contextually relevant writing styles across disciplines by referencing a suitable collection of field-relevant literature. We recognize that good writing is much more complex than simply word choice. However, we find that this specific-word-level tool can serve as a catalyst for learners to think about the overall context of their writing.

Fundamentally, the published writings of established professors/researchers reflect their expertise and experience. As such, using the CASPArS method effectively serves as a tool whereby “professor’s eyes” can help learners evaluate their own writing patterns relative to the writing by field-specific experts.