Intervention to Promote Ethical Authorship Practices in Graduate Education

Abstract

Authorship credit is essential for researchers’ success in academia. For academics collaborating with others, differing perceptions of how to value different contributions, disciplinary differences in authorship conventions, and power differences among collaborators can make authorship decisions more difficult to navigate in ways that feel fair and transparent to all involved. Graduate students may feel particularly disadvantaged in authorship decisions due to their relative lack of publishing experience. Here we tested the effectiveness of an educational training intervention designed to promote ethical authorship practices by supporting graduate students’ knowledge of authorship and authorship ethics and their ability to effectively navigate authorship conversations with collaborators. Students (n = 185) underwent an online training program and used an authorship agreement form to discuss authorship on a research project with their faculty mentor. We randomly assigned half of the students to undergo an additional small group workshop to test the level of institutional investments needed to see benefits for students. We found the online training and authorship agreement forms boosted students’ perceptions of their authorship knowledge and confidence effectively navigating authorship conversations with collaborators. The additional workshop did not yield further benefits for students’ outcomes, suggesting that institutions can help promote ethical authorship through low-cost, scalable educational resources.

Introduction

Academics across many fields of study are facing increased pressure to publish their research (Cutas & Shaw, 2015; Papatheodorou et al., 2008). Moreover, popular press articles suggest that the publication pressure is beginning earlier than ever, with graduate students and junior assistant professors feeling significant stress to publish early and often in their career (Evans, 2019; Flaherty, 2017). Accompanying this pressure is a rise in the number of co-authors listed on the average academic journal article (Bennet & Taylor, 2003; Borenstein & Shamoo, 2015) and uncertainty about different authorship standards (Bain et al., 2022; Satalkar et al., 2020; Puljak & Sambunjak, 2020). Surveys of academics also suggest authorship conflicts may be on the rise, and that these conflicts are also occurring between graduate students and their faculty advisors (Savchenko & Rosenfeld, 2024).

Although researchers have reasonable disagreements about what qualifies someone to be an author and how much to weigh an intellectual contribution (Bozeman & Youtie, 2016; Moffatt, 2018), there is broad agreement that certain practices are unethical, like ghost authorship (i.e., wrongful exclusion of a contributor), guest authorship (i.e., wrongful inclusion of a contributor), bullying behaviors, or automatically listing senior department members as authors (Bennet & Taylor, 2003; Meursinge Reynders et al., 2024; Smith et al., 2020; Satalkar et al., 2020). Unethical authorship practices are clearly harmful to the individuals directly impacted, but perhaps more concerning is their potential to create negative downstream impacts on a university’s research climate (e.g., Norman et al., 2021; Smith et al., 2020). For instance, Smith et al. (2020, p. 1968) note when “researchers believe they are treated unfairly, they are more likely to misbehave, which in turn, may compromise the integrity of research.” Norman et al. (2021) make a similar argument, suggesting that unethical authorship practices can create tensions between colleagues, hurt collaborations, and damage careers. These tensions can also result from misunderstandings or different assumptions about authorship. For example, multi- or interdisciplinary research is increasing, but collaborators often bring different expectations about authorship to these collaborations. When these differences are not recognized and addressed, these tensions can also lead to increased likelihood of research misbehavior (Smith et al., 2020, p. 1968). Internally, we have observed that many of the more serious research misconduct allegations on our campus include a dispute over authorship in some form (T. Reynolds, Dean of the Graduate School and Research Integrity Officer; personal communication, 2018; see also UNC Charlotte Graduate Council statement, 2017). Thus, when it comes to ethical issues in the conduct of research, authorship plays a central role.

Higher education often assumes faculty mentors will provide training to their students on what constitutes ethical authorship practices (Anderson et al., 2013; Ripley et al., 2012). This assumption is problematic for at least three reasons. First, faculty may not provide the necessary training, leaving students unaware of the importance of the problem (Ripley et al., 2012). Second, faculty may not receive training themselves, meaning they will be expected to provide mentorship on ethical authorship decisions without knowing how to provide it (Alexander et al., 2023). And third, faculty may not be the best resource for authorship training given that they face pressures to publish as well (e.g., some faculty and students may feel the need to compete to be first author on publications). As Cutas and Shaw (2015, p. 1322) argued, there may be a “conflict of interest” built into the mentor-mentee relationship when it comes to authorship naming and ordering.

Together, these concerns necessitate a more intentional approach to authorship training. Researchers have begun to focus more systematically on how ethical authorship practices might be promoted and normalized within academia. For example, many journals and publishers now require standardized authorship reporting according to the CRediT taxonomy system (https://credit.niso.org/adopters/). We suggest that in addition, intentional training on healthy authorship practices can increase open and transparent communication in teams even when harmful authorship practices are not in play.

In this paper, we present the results of a training intervention that we developed for graduate students and their faculty mentors at a large, public, research-intensive university. Our research goals were twofold: to test the effectiveness of the authorship training intervention and to determine if a synchronous component (i.e., the 60-minute workshop) yielded additional benefits. In doing so, we hoped to determine the kinds of institutional resources and investments required to provide effective training. To this end, we randomly assigned graduate student participants to one of two intervention groups. One group received authorship training in the form of an online authorship ethics course and a discussion of an authorship agreement form with their faculty mentor. The other group received the online authorship ethics course and then participated in a 60-minute virtual workshop with two project leads before completing the authorship agreement form with their faculty mentor. We assessed graduate students’ knowledge, confidence, and stress levels regarding authorship practices using online surveys distributed before and after the training intervention. We hypothesized the intervention would result in significant improvements in terms of participants’ self-reported knowledge, confidence, and anticipated stress. Further, we hypothesized that the addition of the synchronous workshop would yield additional benefits beyond those generated by the intervention.

To preview our results, we found the authorship training course and authorship agreement positively improved students’ knowledge and confidence and reduced anticipated stress about authorship conversations, but there was no added benefit of the workshop. Our results suggest that, rather than relying on informal mentorship alone, universities can successfully support ethical authorship training for graduate students with cost-effective and scalable resources.

Materials and Methods

Description of Institution

Our institution is a large, public university located in the southeastern United States. It is classified as an R2 (high-research activity) institution by the Carnegie Classification System. The University enrolls around 5,900 graduate students, including ~ 1,100 doctoral students, with the majority in STEM fields.

Participants

All study procedures were approved by our University’s Institutional Review Board (IRB). Participants consisted of graduate students and their faculty mentors, both from our institution. Student participants were recruited by emails forwarded by graduate program directors on the study team’s behalf. In total, 209 graduate student participants completed informed consent for this study. All students were currently engaged in research with a faculty member. The research may or may not include additional collaborators, and there were no exclusions placed on the research areas. The student participants each identified their primary faculty mentor. The study team then obtained the faculty mentor’s consent to participate. This resulted in 118 unique faculty participants, with some faculty listed by multiple student participants. Of the students enrolled, 18 withdrew before undergoing the study intervention and 6 failed to complete the final post-training survey. This left us with 185 student participants who completed the study (see Table 1 for student demographics; note demographic information for faculty participants was not collected).

Table 1.

Graduate student participant demographics and academic characteristics. This sample represents roughly 14% of our doctoral students. Data on how many master’s students are engaged in research with faculty at any time is not readily available at our institution, making it difficult to estimate a response rate for our master’s students. However, most of our master’s programs do not carry research or publication expectations. Our participants were generally representative of our institution’s overall graduate population in terms of their gender, race/ethnicity, undergraduate degree from U.S institution or institution outside the U.S., and their area of study

Demographics & other characteristics	Percent of student participants (n = 185)
Degree Sought
Master’s Degree-Seeking	19.4%
Doctoral Degree-Seeking	80.5%
Gender
Men	41%
Women	58%
Non-binary	1%
Race & Ethnicity (multiple selections allowed)
Asian	41.5%
Black or African American	13.9%
Hispanic or Latino	4.2%
Native American	1.5%
White	45.9%
Undergraduate Degree
Institution outside the U.S.	45.9%
U.S. institution	54.1%
Area of Study
Arts or Architecture	2.2%
Computer & Information Systems	18.4%
Education	7.6%
Engineering	22.7%
Health	6.5%
Humanities	1.1%
Natural Sciences and Math	23.7%
Social Sciences	17.3%

Authorship Intervention Description

We randomly assigned half of our graduate student participants to complete an online training course (Holladay-Sandidge et al., 2023) and then to complete an authorship agreement form with their faculty mentor (Rasmussen et al., 2023a, b).

The other half of our participants were assigned the same online training but underwent a 60-minute virtual workshop before completing the authorship agreement form with their faculty mentor (see Fig. 1). Random assignment was stratified by students’ area of study so that disciplinary representation was balanced across the two conditions.

Fig. 1.

Study design

The online training course consisted of learning modules on authorship and authorship ethics delivered within our institutional learning management software, Canvas (see Table 2). The course incorporated intentional active learning activities designed to engage students in a self-paced, online setting (Holladay-Sandidge et al., 2023). The modules focused on enhancing students’ knowledge about ethical authorship decisions for collaborative academic works, including the considerations that go into authorship naming and ordering decisions, how authorship conventions vary across disciplines, and ethical and unethical authorship practices. The modules sought to empower students to feel confident and comfortable in their ability to navigate authorship conversations with collaborators by engaging them in activities where they were required to make and justify authorship decisions in realistic but complex scenarios. The training also sought to lower the stress students might anticipate regarding discussions about potential authorship-related conflicts with collaborators by offering supportive resources, practical advice, and language to use in such discussions.

Table 2.

Educational content of authorship training intervention. Case studies “You Make the Call” and “Romero and Reed” are adapted from the Survival Skills & Ethics program at the University of Pittsburgh

Authorship training course
Estimated time	Learning objectives	Content	Assessments
Module 1. What is authorship?
20 min	1. Define authorship & common terms related to authorship. 2. Describe common criteria for authorship in a scholarly work. 3. Describe differences in authorship conventions across disciplines. 4. Apply authorship considerations to a case study to determine who merits authorship.	1. Video introduction to academic authorship, “What is authorship?” 2. Slides explaining authorship naming and ordering concepts and common conventions.	1. Self-check quiz. 2. Written assignment asking participants to assign authorship to characters in a provided scenario, “You Make the Call”
Module 2. Ethical Authorship
30 min	1. Summarize key ethical concerns that can arise in authorship considerations. 2. Explain how authorship considerations can differ for graduate students & faculty. 3. Begin to conceptualize & justify author naming & order decisions.	1. Video with voice actors reading real quotes from campus constituents about their authorship experiences, “Murky Waters: Real Stories of Academic Authorship.”	1. Written responses to a case study about a student & faculty research collaboration where the student was not given authorship: “Romero & Reed” 2. Written responses to 7 short authorship scenarios.
Module 3. Handling Authorship Challenges
25 min	1. Explain the complexities sometimes involved in authorship decisions. 2. Feel more confident in your ability to justify certain authorship decisions. 3. Identify potentially unethical authorship practices.	1. Choice-based interactive simulation of a communication breakdown between faculty and student.	1. Written responses to 7 discussion questions based on the interactive simulation.
Module 4. Authorship—Outlook & Resources for Graduate School and Career
60 min	1. Describe what an authorship agreement is. 2. Identify resources (on campus and external) for supporting good authorship practices. 3. Complete an authorship agreement form with your faculty mentor for an existing research project.	1. Authorship agreement instructional video & form. 2. Course wrap-up, post-training survey, & additional resources.	1. Self-check quiz. 2. Authorship agreement conversation with faculty mentor; sign & submit provided authorship agreement form.
Synchronous Workshop (half of participants received)
60 min	1. Discuss & debrief on Canvas course materials. 2. Apply knowledge of authorship norms & ethical considerations to a series of short case studies.	1. 10 short authorship cases designed for rapid reactions. 2. Institutional authorship agreement form. 3. Institutional authorship policy.	1. Reaction & discussion to short authorship cases. 2. Participants role-play authorship conversation between graduate student and faculty member.

The workshop consisted of a 60-minute group session held on Zoom with eight to ten students and two study team member facilitators per session. This workshop enabled a deeper discussion of the training course topics, a discussion of case studies designed to promote good ethical reasoning in authorship decisions, and an opportunity for students to engage in role-playing activities designed to simulate authorship conversations between graduate students and faculty mentors (for materials see: https://osf.io/vrezh/?view_only=d92fecb462a74be1a912a8f48f84c256). All student participants then engaged in an authorship conversation with their faculty mentor to provide students an opportunity to practice making authorship decisions. Participants were given an authorship agreement form to use as a guide for the conversation and were asked to focus on a collaborative research project currently planned or underway (Rasmussen et al., 2023a, b). The authorship agreement form was included as part of the training because, as other researchers have suggested, a prospective discussion of authorship criteria can reduce the chances of problems arising later (Bennett & Gadlin, 2012; Roberts, 2017). Importantly, however, the form allowed for flexibility by clarifying that changes can and should be expected (Youtie & Bozeman, 2016). The completed authorship agreement form was submitted to the study team. Student participants received a total of $250 USD for completing all study components.

Faculty mentors watched a ten-minute instructional video introducing good authorship practices, promoting authorship agreement forms as a tool for facilitating transparent authorship conversations for collaborative research projects, and explaining the overall goal of the study. The video also explained the faculty’s role as a participant in the study was to have an authorship decision-making conversation with their graduate student about a research project and to use the provided authorship agreement form to guide and document that conversation.

Evaluation of Intervention

We evaluated the effectiveness of our intervention primarily through a survey delivered to student participants before and after the intervention (see Table 3). The author team, which included subject matter experts from multiple disciplines, developed the survey questions. We refined the questions in an iterative process and asked two graduate student authors to assess the questions for relevance, clarity of language, and appropriateness to the graduate student population. We designed the survey questions to assess the intervention’s effectiveness in achieving positive outcomes for students’ knowledge (about authorship practices and ethics in different disciplines, awareness of institutional authorship resources), confidence (perceived ability to ethically weigh and balance different authorship-related concerns, confidence effectively self-advocating, comfortability navigating authorship conversations), and reduced stress (regarding anticipating discussing potential authorship conflicts with collaborators). Collaborators were defined in our survey instructions as including “people who work with you on a research project, potential coauthors, and/or your faculty mentor(s).” The survey items had good internal consistency (Cronbach’s alpha reliability scores all > 0.8). We conceive of these knowledge, confidence, and stress categories as useful communication tools rather than as higher-level constructs. Our survey was designed to specifically align with our intervention’s learning goals and we do not claim they reflect more generalizable constructs.

Table 3.

Descriptive statistics for evaluation survey items. Graduate student participants received a survey pre- and post-intervention to evaluate students’ self-reported knowledge and their self-perceptions of their confidence and attitudes regarding collaborative authorship decisions. Additional evaluation questions regarding the training quality and effectiveness were included in the post-intervention survey. Data represent the survey item means and standard errors around the means (SEM) for participants in the study group with the added workshop (n = 92) and for the group without the added workshop (n = 93)

Survey item & scale	Pre- or post-intervention	With workshop Mean (SEM)	Without workshop Mean (SEM)
Knowledge-related items. Scale: 7-point Likert scale, very low (1) to very high (7)
‘Your awareness of common authorship conventions in your discipline.’	Pre-Intervention	4.30 (0.16)	4.23 (0.15)
	Post-Intervention	6.10 (0.08)	5.95 (0.10)
‘Your awareness of common authorship conventions in disciplines other than your own.’	Pre-Intervention	2.88 (0.15)	2.95 (0.17)
	Post-Intervention	5.24 (0.10)	5.22 (0.13)
‘Your general knowledge of ethically questionable authorship practices.’	Pre-Intervention	4.42 (0.14)	4.44 (0.16)
	Post-Intervention	6.23 (0.06)	6.15 (0.09)
‘Your awareness of institutional resources available to help resolve authorship-related disputes.’	Pre-Intervention	3.02 (0.20)	2.97 (0.19)
	Post-Intervention	6.25 (0.10)	6.23 (0.11)
‘Your overall knowledge of authorship ethics.’	Pre-Intervention	4.29 (0.16)	4.15 (0.17)
	Post-Intervention	6.14 (0.08)	6.23 (0.07)
Confidence-related items. Scale: 7-point Likert scale, very low (1) to very high (7)
‘Your confidence in your ability to proactively identify potential authorship-related concerns that other people you work with may have.’	Pre-Intervention	4.14 (0.15)	4.24 (0.16)
	Post-Intervention	6.24 (0.07)	6.23 (0.08)
‘Your ability to weigh and balance different authorship-related concerns in order to make ethically sound authorship decisions.’	Pre-Intervention	4.20 (0.15)	4.19 (0.18)
	Post-Intervention	6.07 (0.08)	6.13 (0.07)
‘Your confidence in your ability to effectively discuss authorship plans with collaborators.’	Pre-Intervention	4.50 (0.17)	4.31 (0.19.)
	Post-Intervention	6.17 (0.10)	6.41 (0.08)
‘Your comfort level in discussing authorship plans with collaborators.’	Pre-Intervention	4.29 (0.18)	4.30 (0.19)
	Post-Intervention	6.15 (0.11)	6.28 (0.08)
‘Your confidence in your ability to effectively advocate for yourself in authorship conversations with collaborators.’	Pre-Intervention	4.09 (0.18)	4.17 (0.18)
	Post-Intervention	5.91 (0.10)	6.12 (0.09)
Stress-related items. Scale: 7-point Likert scale, very low (1) to very high (7)
‘Your stress level at the thought of discussing an authorship-related conflict with collaborators.’	Pre-Intervention	4.55 (0.17)	4.45 (0.19)
	Post-Intervention	3.66 (0.18)	3.69 (0.19)
Additional post-intervention survey items
Scale: strongly disagree (1), disagree (2), neither agree nor disagree (3), agree (4), strongly agree (5)
‘Participation in this course changed my understanding of authorship practices.’	Post-Intervention	4.55 (0.05)	4.42 (0.06)
‘Participation in this training deepened my appreciation for the ethical complexity of authorship.’	Post-Intervention	4.66 (0.05)	4.57 (0.06)
Scale: very unlikely (1), unlikely (2), undecided (3), likely (4), very likely (5)
‘Moving forward, how likely are you to initiate authorship plan discussions with collaborators?’	Post-Intervention	4.49 (0.08)	4.52 (0.06)
Scale: very poor (1), poor (2), fair (3), good (4), very good (5)
‘How would you rate the quality of this training?’	Post-Intervention	4.68 (0.05)	4.61 (0.06)

Implementation fidelity was ensured by tracking student engagement with the training intervention course content and tracking completion of the authorship agreement forms. We monitored students’ assignment responses in the training course and asked them to engage with an assignment more deeply if their response was overly short or insufficient to address the question. Finally, we ensured that all authorship agreement forms were completed and signed by the students and faculty mentors (100% completion rate).

Statistical Analyses

For the 185 student participants, we had no missing data on the survey items repeated pre- and post-intervention. We investigated our research questions using regression models appropriate for ordinal data (cumulative link mixed regression models; clmm2 function of the ordinal package (v2023.12.4) in R (v4.3.2)). For each survey item this model evaluated if students’ survey responses changed significantly post-intervention compared to pre-intervention (treatment factor intervention with two levels, pre-intervention and post-intervention), if students’ responses differed significantly between the two study groups (treatment factor study group with two levels, workshop and no workshop), or if there was any significant interaction between the intervention and group factors. Deidentified data and analysis code are available (https://osf.io/vrezh/?view_only=d92fecb462a74be1a912a8f48f84c256).

Results

The descriptive statistics of our study are reported in Table 3. The results of our regression models analyzing the pre- and post-survey data from student participants are reported in Table 4.

Table 4.

Inferential statistics for evaluation survey items. Survey item responses were entered into cumulative link mixed model regression analyses appropriate for ordinal scale data with the treatment factors of intervention (pre-intervention dummy coded as 0, post-intervention as 1), study group (no workshop dummy coded as 0, with workshop as 1), and an intervention x study group interaction term. The table shows the results for each survey item, with the data representing the fitted model’s parameter estimates, standard errors and Wald-based p values. These results show the effect of the treatment factors on participants’ survey item ratings. Parameter estimates (or coefficients) show the effect of different treatment factors on the likelihood of item ratings being in one category versus another (e.g., “agree” versus “strongly agree”). The parameter estimates show the strength and direction of the relationship between the treatment factor and the survey item ratings. Positive estimates mean as the predictor increases, the likelihood of being in a higher category (e.g., “very likely” instead of “likely”) also increases. A negative estimate means the opposite. The standard error is a measure of the confidence in the accuracy of the parameter estimate, with smaller standard errors meaning increased confidence estimate is close to the true effect. The p values indicate whether or not the treatment factors had a significant effect on the item ratings, with p values < 0.05 deemed significant

Survey Item	Treatment Factor	Estimate	Std. Error	p value
Knowledge-related items
‘Your awareness of common authorship conventions in your discipline.’	Intervention	3.50	0.38	< 0.00001**
	Group	0.12	0.33	0.72
	Intervention x Group	0.14	0.40	0.72
‘Your awareness of common authorship conventions in disciplines other than your own.’	Intervention	3.42	0.35	< 0.00001**
	Group	0.03	0.31	0.91
	Intervention x Group	0.08	0.38	0.84
‘Your general knowledge of ethically questionable authorship practices.’	Intervention	3.51	0.37	< 0.00001**
	Group	0.09	0.30	0.77
	Intervention x Group	0.15	0.39	0.71
‘Your awareness of institutional resources available to help resolve authorship-related disputes.’	Intervention	3.90	0.39	< 0.00001**
	Group	0.06	0.29	0.82
	Intervention x Group	0.14	0.40	0.73
‘Your overall knowledge of authorship ethics.’	Intervention	3.51	0.38	< 0.00001**
	Group	0.16	0.31	0.60
	Intervention x Group	0.35	0.40	0.38
Confidence-related items
‘Your confidence in your ability to proactively identify potential authorship-related concerns that other people you work with may have.’	Intervention	3.85	0.38	< 0.00001**
	Group	0.12	0.28	0.68
	Intervention x Group	0.10	0.39	0.80
‘Your ability to weigh and balance different authorship-related concerns in order to make ethically sound authorship decisions.’	Intervention	3.28	0.36	< 0.00001**
	Group	0.03	0.30	0.93
	Intervention x Group	0.12	0.39	0.76
‘Your confidence in your ability to effectively discuss authorship plans with collaborators.’	Intervention	2.66	0.34	< 0.00001**
	Group	0.24	0.31	0.43
	Intervention x Group	0.79	0.40	0.049*
‘Your comfort level in discussing authorship plans with collaborators.’	Intervention	2.78	0.34	< 0.00001**
	Group	0.04	0.32	0.89
	Intervention x Group	0.25	0.40	0.52
‘Your confidence in your ability to effectively advocate for yourself in authorship conversations with collaborators.’	Intervention	2.84	0.34	< 0.00001**
	Group	0.13	0.33	0.69
	Intervention x Group	0.35	0.39	0.37
Stress-related item
‘Your stress level at the thought of discussing an authorship-related conflict with collaborators.’	Intervention	-1.03	0.27	0.0002**
	Group	-0.11	0.31	0.73
	Intervention x Group	0.12	0.38	0.75

* Denotes significance at p < .05, ** at p < .0001

Intervention Improves Graduate Students’ Self-perceptions of their Knowledge About Ethical Authorship Decision Practices and Norms

Our results show that students rated their knowledge of authorship issues higher after completing the intervention than before (Tables 3 and 4; Fig. 2). Students reported increased awareness of authorship conventions in their discipline as well as conventions in other disciplines. They reported higher levels of knowledge of ethically questionable authorship practices (e.g., gift or ghost authorship), and more awareness of relevant institutional resources (e.g., our institutional authorship policy and dispute resolution procedures; University Policy (2021); Rasmussen et al., 2020). Students also rated their overall knowledge of authorship ethics higher after completing the intervention.

Fig. 2.

Pre-intervention and post-intervention survey responses. Data represent the means and standard error around the means (SEM) for student ratings on evaluation survey items before and after completing the intervention, collapsed across the with and without workshop groups.

The addition of the synchronous workshop did not have a significant effect according to our results. Students in the synchronous workshop group did not rate their knowledge as higher than students in the no synchronous workshop group. The pattern of results was equivalent across the two study groups.

Intervention Improves Graduate Students’ Self-perceptions of their Confidence in Navigating Authorship Decisions Effectively

In terms of confidence, our results show that students reported higher confidence in their ability to proactively identify potential authorship-related concerns collaborators may have after the training as compared to before the training (Tables 3 and 4; Fig. 2). Relative to pre-training ratings, students’ ratings of their general comfort discussing authorship plans, as well as their ability to appropriately weigh and balance others’ concerns to make ethically-sound authorship decisions, increased. The training also successfully increased students’ confidence in their abilities to effectively advocate for themselves in authorship conversations, achieving an important goal of the training intervention.

We hypothesized that students receiving the added synchronous workshop would show greater gains pre- to post-intervention in their self-ratings about their confidence in their ability to discuss authorship plans with others. However, students participating in the added workshop did not generally show different outcomes than students who did not receive the added workshop. There was, however, one item (‘Your confidence in your ability to effectively discuss authorship plans with others’) on which students who received the workshop reported slightly smaller gains post-intervention compared to students who did not undergo the workshop (Intervention x Study Group interaction, p = .049, Table 4).

Students Anticipate Less Stress in Discussing Potential Authorship Conflicts

Participating in the authorship training intervention successfully decreased students’ stress levels at the thought of discussing an authorship-related conflict with collaborators. There were no differences observed for students receiving the added workshop compared to students who did not receive the workshop.

Students Shared Positive Feedback About the Training Intervention

The post-intervention survey included four additional questions designed to allow students to provide additional feedback regarding the effectiveness of the authorship training (Table 3). Students rated the quality of the training very positively and indicated that participating in this training deepened their appreciation of the ethical complexity of authorship. Encouragingly, students indicated they were likely or very likely to initiate authorship conversations with collaborators in the future.

Discussion

We investigated the effects of an intervention on graduate student participants’ self-reported knowledge, confidence, and anticipated stress regarding authorship decisions and conversations. We also investigated if participation in an additional workshop yielded any additional positive outcomes for our student participants. We found that the authorship training course and authorship agreement form had a significant positive impact on students’ self-reported knowledge of authorship conventions and confidence in making ethical authorship decisions. Further, it reduced their anticipated stress at having authorship conversations with faculty and other collaborators. We found no significant differences between the participants who attended the additional virtual workshop and those who did not, suggesting the asynchronous training and authorship agreement form process may be sufficient for supporting positive student outcomes. The training and resources were well-received by both faculty and graduate students. We received informal feedback from participants that the training was a light amount of work and valuable to them.

Again, one finding of note is that the synchronous, 60-minute workshop did not appear to influence study outcomes over and above the asynchronous elements of the training. In some sense this is good news as this “low cost” approach is less resource intensive for universities who wish to offer authorship training to faculty and students. Our sense is that asynchronous learning has become more familiar to participants since the COVID-19 pandemic and the quality of such materials is perhaps of higher quality than years before. The asynchronous training was exhaustive in terms of covering policies, professional standards around authorship, as well as presenting cases for consideration. These cases covered scenarios where there was ambiguity in the authorship determinations, but solutions were offered in terms of how to prevent ambiguity in the future and how to handle such scenarios when they occurred. Therefore, there appears to be diminishing returns of 60 more minutes of training over and above the original 135 min that all participants engaged in.

Many higher education institutions offer credit-bearing courses or formal training on responsible conduct of research, often including authorship discussions. Our authorship training course and authorship agreement form can enhance these existing efforts. The training course described here is specifically designed for graduate students. Responsible conduct of research instructors may find value in sharing our resources or in assigning our online training course as homework. Institutions can also offer the authorship training course on demand for their trainees, as we do. We publicize the course at graduate student orientations and on the Graduate School’s website. Additionally, we have added a step to doctoral students’ proposal defense paperwork to confirm the dissertation committee discussed authorship plans related to scholarly products generated by the student’s work. For faculty, we developed a short authorship training course emphasizing their mentoring role in training junior scholars on authorship practices and ethics. Institutions can similarly offer this faculty resource on demand on their websites, or tie it into other faculty mentor training. While the student and faculty training courses are now voluntary at our institution, other institutions may see merit in making authorship training mandatory.

Both students and faculty can benefit from using the authorship agreement form to document authorship conversations. Faculty have shared the form serves as a helpful checklist to ensure comprehensive discussions with students and is useful for navigating authorship decisions in complex research collaborations. We link to the authorship agreement form on our Graduate School’s website and promote both the courses and the form as valuable resources for faculty and students to engage with early and continuously during collaborative research projects. These resources can become part of researchers’ practice, promoting ethical authorship behaviors sustainably. The authorship training courses, agreement form, and other resources are publicly available on our project website (2025).

Our work complements recent efforts by Norman and colleagues (2021, 2023). In a biomedical research context, Norman et al. (2023) tested how using an authorship agreement form could impact the process of collaborative research projects. They found that the use of an authorship agreement form by early-career biomedical researchers could reduce tensions in an author team and perhaps even speed up the publication process. Taken together, Norman et al.’s (2023) work and our own shows that authorship agreement forms can be one important component of healthy authorship practices, and that these forms may be useful in terms of research team productivity (increasing speed), cohesion (e.g., reduce tensions), and ethics (e.g., making fair and justifiable decisions).

In a context of increasing publication pressures (Cutas & Shaw, 2015) and multi-author collaborations (Borenstein & Shamoo, 2015), authorship naming and ordering has become critically important for career outcomes. With the potential for high reward (an impactful publication) comes the chance of unethical practices (Smith et al., 2020). For instance, junior researchers may feel pressure to add a senior author to a paper to increase its likelihood of being published, or a senior researcher may choose an author order primarily based on who “needs the publication” more. In this context, the need for training regarding the importance of ethical authorship practices has never been greater. We suggest it is especially important to focus on training graduate students because they may be particularly vulnerable to unethical authorship practices, given they have relatively less publishing experience compared to faculty and because of the power difference between them and faculty members. Although our training cannot fully protect students from the decisions of faculty members, it can provide them with greater knowledge and confidence to address these issues with their advisor or with others. In other words, a student with more knowledge and greater confidence may be better equipped to know when they have seen an unethical authorship practice and to speak up about it.

Future research can contribute to the broader goal of normalizing transparent and ethical authorship practices. One key question concerns the downstream impacts of authorship training on the research climate. In the same way negative experiences can harm research climate (Smith et al., 2020; Martinson et al., 2010), positive experiences – such as institutionally encouraged training – could foster a more positive research climate.

Author Contributions

All authors contributed to the conceptualization of this article and revision of the manuscript. LMR, GB, KHH and ED secured funding for this study. ED, AM, HHS, LMR and GB supervised the study. ED and AM analyzed the study data and prepared the first draft of the manuscript. All authors approved the final version of the manuscript.

Funding

This material is based upon work supported by the National Science Foundation under grant number 2024200. Any opinions, findings and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the National Science Foundation. The funder had no role in the decision to publish or the preparation of this manuscript.

Data Availability

Data, Materials and Code Availability: Deidentified data, materials and code are available in an OSF repository: https://osf.io/vrezh/?view_only=d92fecb462a74be1a912a8f48f84c256. Authorship courses and authorship agreement form are also available: https://www.authorshipproject.org/.

Declarations

Ethical approval

This study was approved by the UNC Charlotte Institutional Review Board (IRB). The procedures used in this study adhere to the tenets of the Declaration of Helsinki.

Consent to participate and publish

Informed consent was obtained from all individual participants included in the study.

Competing interests

The authors declare no competing interests.