Canadian professors’ views on establishing open source endowed professorships

Abstract To accelerate scientific progress by advancing the spread of open access and free and open source software and hardware in academia, this study surveyed university professors in Canada to determine their willingness accept open source (OS) endowed chair professorships. To obtain such an open source endowed chair, in addition to demonstrated excellence in their field, professor would need to agree to ensuring all of their writing is distributed via open access and releasing all of their intellectual contributions in the public domain or under OS licenses. Results of this study show 81.1% Canadian faculty respondents would be willing to accept the terms of an OS endowed professorship. Further, 34.4% of these faculty would require no additional compensation. Respondents that favor traditional rewards for endowed chairs were shown to greatly favor receiving funds that would help benefit research (28% for graduate assistants to reduce faculty load or 46.7% for a discretionary budget-the most common response). These results show that, in Canada, there is widespread shared sentiment in favor of knowledge sharing among academics and that open source endowed professorships would be an effective way to catalyze increased sharing for the benefit of research in general and Canadian academia in particular.


PUBLIC INTEREST STATEMENT
Most people believe if the public funds research they should be able to both read and use it. To help this process a new study surveyed university professors in Canada to determine their willingness accept open source (OS) endowed chair professorships. To obtain such an open source endowed chair, in addition to demonstrated excellence in their field, a professor would need to agree to ensuring all of their writing is distributed via open access and releasing all of their intellectual contributions in the public domain or under OS licenses. This would allow everyone to read and use their work for free. 81.1% of Canadian faculty respondents agreed and 34.4% of these faculty would require no additional compensation. These results show widespread shared goals of knowledge sharing among academics and that open source endowed professorships would be an effective way to catalyze increased sharing for the benefit of research.

Introduction
One of the foundational institutional norms of science is the idea that scientific knowledge is commonly owned (Sismondo, 2010: 24), yet contemporary scientific progress is being held back by lack of sharing knowledge due to: 1) restricting access to copyrighted scientific literature without payment (Gibbons, 1994;Heise & Pearce, 2020), 2) proprietary and expensive software (Damato, 2005;May, 2006) and 3) similarly expensive and restricted access hardware (Gibney, 2016;Pearce, 2014;Torrisi et al., 2016). All three mechanisms are based on profit seeking and divide scientists into the "haves and have nots" , which can be avoided by applying free and open source principles to the production and dissemination of academic scholarship.
First, the open access movement has grown to improve access to the peer-reviewed literature for everyone (Johnston, 2008;Joseph, 2013). Academics are increasingly calling for unlimited access to the entire scientific journal literature (Budapest Open Access Initiative, 2002) and increased transparency in the scientific knowledge process itself (European Commission, 2015). This is particularly true in Canada. The Canadian Journal of Higher Education (CJHE), which was founded in 1971, shifted to an open access journal in 2009 to promote the accessibility of this scholarship (Pidgeon et al., 2019). Federal funding agencies are increasingly requiring open access for publicly funded research (Suber, 2012)  Secondly, the use of free and open source software (FOSS) is growing rapidly (Fortunato & Galassi, 2021;Von Krogh & Spaeth, 2007). Anyone is free to use, copy, study, and change FOSS in any way, and the source code is openly shared so that voluntarily improvement in the design of the software is enabled and, in most cases, demanded by viral clauses. The fact that FOSS development leads to superior code is now established (Bonaccorsi & Rossi, 2003;Comino et al., 2007;Harhoff et al., 2003;Lakhani & Von Hippel, 2004;Lee et al., 2009;Osterloh & Rota, 2007;Raymond, 1999;Tozzi, 2017;Von Krogh et al., 2003;Weber, 2009;Zeitlyn, 2003). Even outside of academia, the majority of large corporations are now contributing to open source software projects (LeClair, 2016;Parloff, 2013). FOSS is the dominant form of software development with FOSS running on 100% of supercomputers (Vaughan-Nichols, 2018) and 90% of cloud servers (i.e. Facebook, Twitter, Wikipedia, YouTube and Amazon; Hiteshdawda, 2020), over 84% of the global smartphone market (IDC, 2020), and more than 80% of the internet of things (IoT) devices (Eclipse, 2019). Academic researchers rely on FOSS specifically for their research (i.e. Source Forge has >17,900 FOSS for science and engineering (Elsevier, 2021)). Major research funders understand the value of complete access to code for doing science (Chan Zuckerberg Initiative, 2019) as well as the fundamental right of citizens to use software they pay for in the case of tax-payer funded software development. Free and open source hardware (FOSH; Gibb, 2014;Powell, 2012) is growing rapidly but follows behind FOSS by about 15 years in the academic literature (J. Pearce, 2018). FOSH is defined as hardware whose design is shared so that anyone can study, modify, distribute, make, and sell the design or hardware (Open Source Hardware Association (OSHWA), 2021). FOSH provides the "source code" for physical hardware including the bill of materials, schematics, computer aided designs (CAD), and other information such as detailed instructions needed to recreate a physical item. FOSH development is established for improved product innovation (Dosemagen et al., 2017;Yip & Forsslund, 2017). In the form, most closely following FOSS, digital design files are shared, which enable users to replicate physical objects with 3-D printers, CNC mills or laser cutters (Niewiadomski & Anderson, 2014;Petersen & Pearce, 2017;Redlich et al., 2016). Thus, FOSH offers the opportunity to improve access to instrumentation, while enhancing the quality and diversity of tools by leveraging recent developments in these distributed digital manufacturing technologies (Baden et al., 2015;Gibney, 2016). FOSH saves scientists research funding while enabling high-performance customized equipment (Damase et al., 2015;Fisher & Gould, 2012;Murillo & Wenzel, 2017;Pearce, 2012Pearce, , 2014Pearce, , 2017 and supports a wide range of disciplines for open science Heikkinen et al., 2020;Hope, 2009;Robertson et al., 2014;Willinsky, 2005). In a 2020, review it was found that FOSH reduces costs compared to proprietary scientific equipment by 87% in general, up to 94% for those that utilize open source electronics and distributed manufacturing (Pearce, 2020). Calibrating FOSH scientific tools using open standards also has the potential to eliminate the replication crisis (Loken & Gelman, 2017). Already a few institutional programs like the 3D Print Exchange run by the U.S. National Institutes of Health (Coakley et al., 2014) and the Open Hardware Repository run by CERN (ohwr.org) and licenses (Ayass & Serrano, 2012) are adding to a cultural shift among scientists to full open sharing (American Association for the Advancement of Science, 2014). Research funders can benefit from the downloaded substitution value (Pearce, 2015) of FOSH, where scientific funding is primarily invested for the development of FOSH instead of simply purchasing proprietary equipment. Shockingly high returns on investment (ROI) (e.g., >1000%) are then realized by funders by direct digital replication of scientific devices for research and education at approximately the costs of materials (Pearce, 2016). FOSH international community networks are growing (Dosemagen et al., 2017).
Although open science is part of a larger social shift towards open production methodologies and decentralized, distributed models of collaboration (Von Hippel, 2016), there is still substantial progress needed because 1) most of the peer-reviewed literature is not accessible to the majority of researchers because it is held behind paywalls, 2) most researchers are dependent on proprietary software and 3) the vast majority of scientific hardware in each field is economically inaccessible to the majority of researchers. Increasing open source sharing in academia can provide solutions to all three of these issues, but some form of incentive may be necessary . To determine what such incentives may look like in the Canadian context, this study explores professors' perspectives towards the idea of an open source endowed chair. These developments are of particular importance for post-secondary institutions exist in Canada which include universities (offering bachelors, masters and PhD programs in all major academic disciplines) as well as community colleges and institutes (offering generally shorter and more applied programs in business, computer and mechanical technologies, health care, social services, agriculture and the trades; Government of Canada, 2021). Generally endowed chairs are supported at the universities either funded locally by university endowments (which are sometimes funded by industry) or more commonly through donors or the federal/provincial government. For example, The Canada Research Chairs Program invests over $300 million annually to attract and retain some of the world's most accomplished and promising minds to Canada (Government of Canada, 2022). Conventionally, endowed chairs (also called an endowed professorship) are a method of encouraging academic excellence while bringing prestige to professors, their institutions, and the donors that fund them (Schaeffer & Papalia, 1966). Top performing faculty in all disciplines can apply and are sometime nominated internally for endowed chairs. In this case, in addition to demonstrated excellence in their field, to hold an open source endowed chair, professors would need to agree

Methodology
The survey methodology, which effectively measures attitudes of individuals in a target population to inform generalizations about populations too large to observe directly (Babbie, 2010), was used to answer the following research questions: 1) Are Canadian academic faculty willing to accept an open source endowed professorship? and to 2) If so, what are their preferred terms?
To answer these questions, the study makes use of survey methodology. The survey methodology is considered appropriate because the study is trying to tap into a more extensive set of the target group (Gordon & McNew, 2008) of faculties in Canadian colleges and Universities. An online targeted survey using standardized questionnaires is used to systematically collect people's preferences, which is also the goal of this study. Even though the study is informative and exploratory in nature, the survey is considered a positivist approach with the inherent assumption that collected data can potentially support causal analysis (Creswell, 1999;King et al., 2021). The survey is designed as a means to explore participants' experienced and lived opinions of open source endowed professorship. The participants actively participate in the system for which the study intends to inspect the interest, need, and terms of the need, and hence was considered an appropriate way to reach the target audience.

Survey design and procedure
The survey design involved reviewing past studies using the survey method to study open source endowed professorship. The survey questionnaire was developed on the work done by Pascaris et al. (2021) in the U.S. context. Survey items were designed to measure the professional diversity among study participants, identify the preferred terms of an open source endowed professorship, and gauge attitudes towards the concept in Canada. The survey strives to measure this professional diversity, to get responses from various faculties and departments and faculty at different levels of their academic career. The primary variables analyzed in this study are willing to accept an open source endowed professorship and the preferred terms, which were captured in two survey items. Participants were first presented with information about open source endowed professorships to provide context and clarity for the subsequent multiple-choice question and open-ended question, which are part of Appendix 1.
This survey item provides direct insight into attitudes towards and preferences for an open source endowed professorship among Canadian academic faculty. Faculty members are the target audience that will be affected by the change or no change in the open source endowed professorship in their university. Hence their potential behavioral response is considered as a proxy indicator for the willingness to accept open source endowed professorships and the conditions of the same thereof. This is captured through the primary variable analyzed, which is the willingness to accept the terms of an open source endowed professorship. The other variables included were focused on measuring the demographic diversity among academic faculty who participated in the survey. Tables 1-4 show the options of the key demographic variables.
The survey questionnaire was submitted to the Institutional Review Board (IRB) as part of the package. After the IRB approval, the survey was launched online via email on 13 July 2021 and distributed to 9,835 participants. A reminder email was sent on 20th July to stimulate an increased response rate. The study ran for 1 month. Participants were prompted to follow an online survey link. Potential participants were found by scraping Canadian education institutional websites for faculty email addresses. The web scraper searched for any lines containing "@" and "email" on departmental faculty websites and then added the matching email addresses to a database. Their responses were recorded using Qualtrics software (Qualtrics, 2005).

Data collection and analysis
After the responses were recorded in Qualtrics, the data was taken offline, and the identifier information from the survey was removed. The data were analyzed using IBM SPSS Statistics (version 26; IBM Corp,). Each of the variables and responses was coded in SPSS, with each response having a unique identifier value. The analysis mainly used the descriptive statistics tab of SPSS to conduct the analysis.
The study aimed to achieve a sample representative of Canadian academic faculty, and the procedure resulted in a sampling frame of 9,835 individuals. Of the surveys distributed, 505 responses were recorded in total (5% response rate). After the elimination of incomplete responses, the effective sample size is 439 participants. The frequency and percentage of respondent key characteristics representing the target population of academic faculty in Canada are shown in Table 1 for professor type, Table 2 for tenure status, Table 3 for the type of college, and Table 4 for discipline.
There was one open-ended question to learn about if there were any points that were not captured and to learn what minimum additional compensation faculty would accept for the open source endowed professorship. Table 2 shows that 22.8% of faculty respondents were non-tenured, which is slightly lower than the national percentage of 30% (Government of Canada, 2019). Tables 1-4 show a reasonable diversity among survey respondents in terms of professor type, tenure status, college type as well as academic discipline given their relative popularity in the whole population. Instead of making statistical inferences about relationships about the entire population's key characteristics and their perspectives about and preferred terms of an OS professorship, the purpose of this study was to assess interest levels and therefore the effective sample (5% response rate) can inform logical conclusions about the sentiment of survey Canadian academic faculty towards the OS concept.

Results
The results of this study reveal that Canadian professors that participated in the study are overwhelmingly in support of an open source (OS) endowed professorship, with 81.1% denoting willingness to accept and only 18.9% being unwilling to go open source. The vast majority of Canadian faculty who participated in this study appear willing to divest of all IP for an endowed chair, based on their selection of compensation they seek in return (Table 5). As can be seen in Table 5, more than a third of respondents are willing to accept an OS endowed chair would need no additional benefits. This suggests that in Canada there is a growing shared sentiment in favor of knowledge sharing among academics. The results showing academics favor knowledge sharing supporting previous researches on scientists more generally (Ensign, 2008;Ensign & Hebert, 2004;Oliveira et al., 2019) as well the rapidly expanding open access movement (Johnston, 2008;Liesegang, 2013;Suber, 2009). It should be noted that percent (for Tables 5-8) was based on the responses, as a respondent could select more than one option the sum of percentages will be more than 100%. The tables show what are the most common terms of the OS endowed professorship among respondents. The percentage is based on the frequency divided by the total respondents.
As endowed chairs of any kind are generally supported by income from an endowment fund established by gifts and is made available to a distinguished faculty member in support of their teaching, research, and service activities. Traditionally, endowed chairs are highly sought after and held in high esteem in academia. They are associated with some form of additional compensation and a discretionary spending budget drawn from the interest of the endowment. In this study, the rewards for endowed chairs were most commonly selected as preferences among respondents included funds that would help benefit research with 46.7% for a discretionary budget and 28% for research assistant (RA) to help with research directly or a teaching assistant (TA) to reduce faculty load to help indirectly. A quarter were interested in increased salaries.
These results were further analyzed for three core disciplines 1) engineering/technology, 2) natural sciences, and 3) social sciences in Tables 6, 7 and Table 8, respectively. This was done to assess if there are differences in preferred compensation types among scholars of various disciplines. The first two disciplines were of specific interest because they produce physical products, software, and discover IP, and are the most likely to protect or patent their research findings (Owen-Smith & Powell, 2001) and it was possible there would be a major difference between and social science faculty, who typically do not. Approximately one-third of engineering professors indicating needing no additional compensation, while fewer (21%) natural scientists and more social scientists (39%) indicated desire for additional compensation. About 26% of engineers and social scientists wanted increased salaries, in comparison to only 15% of natural scientists. About a quarter of faculty in all three disciplines wanted increased graduate student support. About 60% of engineering and natural sciences faculty would want an increased discretionary budget, in comparison to 50% of social scientists. These small shifts in preferences could be based on the costs of research in the disciplines. Overall, only 21% of engineers, 13% of natural scientists, and 16% of social scientists in Canadian universities would be unwilling to accept the terms of an OS endowed professorship. Table 9 shows the overall relationship between a certain type of academic discipline and support for an open source endowed professorship. Table 9 shows that there are different sample sizes in   terms of responses from different disciplines. This is in part caused by the size of the disciplines but also the limitation of the sampling technique and anti-spam techniques used by the individual department websites. Of the responses, however, it is clear that all disciplines have a strong majority supporting OS opportunities. The strongest support was found in library science and law, with the former being a discipline dedicated to open knowledge transfer (Richard et al., 2009) and past work indicating that librarians feel a strong mandate to carry out open accessrelated activities (Greyson et al., 2009). The latter rarely use IP as a method to further their own work. In all the disciplines there is a strong majority for willingness to go for open source research. In all the major fields, based on the responses received, over 2/3 of respondents have shown willingness for open source work and the engineering field shows the same pattern Table 10 shows the relationship between rank of Canadian professors and support for an open-source endowed professorship. As can be seen in Table 10, most respondents are either assistant, associate or full professors and 81% of the assistant professors, 83% of associate professors and 77.6% of full professors reported willingness to accept an open source chair, well over 2/3rd of respondents.
The type of university and willingness to accept the terms of an OS endowed chair in Canadian professors was 81% of four-year colleges and 80% at research universities.
In the open-ended responses given by those willing to accept the terms of an OS professorship, the most common recurring themes were additional credit for tenure and paying for journal publications (e.g., article processing charges (APCs) waivers for open access journals). These fees are not strictly necessary to meet the requirement of an open source chair, as nearly all academic publishers have published policies that allow for preprint posting (e.g., Springer (2021), Taylor and Francis (2021), Elsevier (2021), Wiley (2021), and Sage (2018). Posting preprints has zero financial cost and there are numerous non-profit pre-print servers (e.g., arXiv, HAL, OSF, Zenodo, etc.), some hosted by publishers (e.g., SSRN owned by Elsevier and preprints.org by MDPI), and many universities house their own servers for their own faculty and students (e.g., Western University's Scholarship@Western). In addition, the "Directory of Open Access Journals" (2021) currently lists over 11,000 that have no APCs.

Discussion
Overall, most Canadian faculty member respondents that mentioned additional economic requirements were relatively modest as seen in Figure 1. The mean was $33,320 CAD per year with a standard deviation of $46,965 CAD/year. The minimum request was $1,000 CAD, presumably for APC charges, while the largest was for $250,000 CAD/year.
Open source sharing as a superior form of knowledge development is well supported in the literature (Von Krogh et al., 2003). As academics of every discipline work on the development of knowledge, it is not entirely surprising that the vast majority of Canadian academics would be in favor of OS endowed chairs for themselves even for modest investments. The results of this study indicate that the majority of   Canadian academics in the widest range of disciplines and school types view OS knowledge sharing as beneficial. Supporting OS endowed chairs thus appears to be a straight-forward way to increase scientific output and the social benefits associated with it, for modest investments.
As more than a third of faculty indicated that they would accept an endowed chair with no funding, it is possible to gain the positive benefits of an open source endowed professorship for nearly zero cost in Canada. University administrators and donors may want to consider other forms of compensation and rules (e.g., conventional measures of excellence) to maintain the prestige of endowed chairs in general. There are other nearly zero cost benefits that could be given to open source endowed chairs that may have real value and provide institutional prestige, like preferred/reserved parking spots. In addition, as can be seen in Figure 1, the majority of professors that did want additional funds were modest-primarily for APC costs. An endowment to cover these costs could be shared by many "assistant open source endowed chairs" or similar.
To ensure that open source endowed chairs carry the same prestige as all other named endowed chairs, institutions may want to maintain current funding requirements (e.g., $1 m or more to ensure the interest income is high enough to provide a real benefit to the researcher). Funding for open source endowed chairs can follow standard models for conventional donors and for government programs. The Canada Research Chairs program would only need to be adapted slightly to make these all open source as the first requirement of open access already exists for NSERC. A third opportunity exists in businesses that benefit from OS development in a sector that improves their bottom line (e.g., Facebook funding OS development of datacenters and Open Compute Project (Frachtenberg, 2012;Hazelwood et al., 2018; Open Compute Project (OCP), 2021)), as in the software-based business models that are well established in OS (Bonaccorsi et al., 2006;Chang et al., 2007;Helander & Rissanen, 2005;Munga et al., 2009;Shahrivar et al., 2018). Although OS business models for hardware-based companies are more challenging (Beldiman, 2018), there are now many profitable and proven options (Pearce, 2017) and several benefits, which include increasing customer perceived value, decrease costs for product development and sales, shorten product go-to-market time, and incubate startups with knowledge, experience, and resources (Li & Seering, 2019). As FOSH businesses thrive and expand they will be in positions to fund OS endowed chairs.

Limitations and future work
All universities were targeted in Canada, the construction of the sample frame was challenged by some school's anti-spamming features and website designs, which precluded full study participation. To overcome these limitations, future studies could have emails sent from interested funders or university officials internally. In addition, it is well known that response bias is issue of survey research (Mathiyazhagan & Nandan, 2010). Thus, Canadian faculty members with a strong opinion about open source (either for or against) were more likely to respond that those who had not previously formed an opinion. Fortunately, response bias has been found to be minimal in faculty surveys (Menachemi, 2011) and in this study did not appear to impact data quality. This exploratory study opens up the potential for future research in this area of open source endowed chair positions. Specifically, a broader survey could be conducted and personal interviews to obtain a high confidence interval and statistically significant sample size. Scientific funders could be surveyed to find those most willing to pilot such a funding mechanism. As such a high percentage of respondents indicating they would accept the terms of an open source endowed chair without compensation, university policies to mandate or opt in to such terms could be explored. Further work is also necessary to find the most efficient way of offering open source endowed chairs to top performing academics. This may vary by university, size, type or academic discipline.

Conclusions
This study found that an overwhelming majority (81.1%) of Canadian professors which responded to this exploratory survey would be willing to accept an open source endowed chair with the terms of open access for all of their writing and open source or public domain sharing of all of their intellectual efforts. These results indicate that offering open source endowed professorships may be an attractive way to increase accessibility to academic output and encourage more inclusivity by eliminating intellectual property paywalls. Interestingly, 34.4% of survey respondents would accept the open source endowed chair with no additional compensation of any kind. This again, points to a potential large percentage of the overall academia being willing to share all of their intellectual output. A quarter of the faculty respondents would want increased salaries as chairs, 28% increased graduate student support, and about half (46.7%) an annual discretionary budget. These latter two results, making up roughly 3/4s of those requesting some for of additional aid show that the faculty prefer to have funding going to increasing their output rather than increasing their own salaries. Of those that specified amounts of additional funding, the majority wanted only modest support to cover APCs. The support for open source was consistent across academic disciplines, faculty rank, and type of institution. It is clear Canadian academics share a support of knowledge sharing, which indicates that offering open source endowed chairs would be a viable way of accelerating scientific progress. This exploratory study opens up the potential for future research in this area to argue that Canadian professors would accept open source endowed chair positions. To test these theories scientific funders can investigate the most efficient way of offering open source endowed chairs to top performing academics in their disciplinary portfolios.