Motivating and Impeding Factors Affecting Faculty Contribution to Institutional Repositories

1. Introduction

In this paper, I discuss the problems surrounding faculty contribution to Institutional Repositories (IRs) and propose a theoretical model for studying the diverse factors surrounding this issue.  A growing body of literature regarding IRs has emerged since 2002 when major research universities in the U.S., such as MIT and the University of California launched their own IR systems. Over the past 4 years, an increasing number of research universities has implemented or plans to implement an IR (Markey et al., 2007). Lynch and Lippincott (2005) found that out of 97 universities categorized as Carnegie "doctoral universities", 40% already operated IRs. Among non-implementers, 88% were found to be in the planning stage of IR implementation. This finding indicates that IRs are becoming a component of the technical infrastructure in doctoral research institutions. Whether they become a part of the intellectual infrastructure depends on the extent of faculty contribution.

While the rise in IR deployment looks promising, Shearer (2003) suggests that the success of IRs will be determined eventually by "their uptake and use by researchers" (p.106). She points to the critical mass of content that led to the significant usage of disciplinary e-print repositories. Translating this to IRs, she argues that the success of an IR should be determined by its use, and one of the measures of usefulness is contribution of content. Although potential contributors include faculty, students and staff in universities, faculty members are considered the crucial contributors of scholarly content. However, several studies note that it has been difficult to get faculty members to contribute (Chan, 2004; Foster and Gibbons, 2005; Pelizzari, 2005; Young, 2002).

Foster and Gibbons (2005) interviewed 25 professors at the University of Rochester in order to investigate the factors affecting contribution.  They suggested that the primary impetus for faculty contribution is to enable other scholars to find, use, and cite their work. Foster and Gibbons also identified reasons why faculty did not submit their content, such as copyright infringement worries and disciplinary work practices (e.g., co-authoring or versioning). Faculty members developed their own routines to create and organize documents. Finally faculty members perceived that IR contribution involved additional work, such as metadata creation for contributed objects.

Other more quantitative studies using survey methodology also exist (Rowlands et al., 2004; Swan and Brown, 2005). These studies deal with a broad range of Open Access (OA) practices including OA journals, disciplinary e-print repositories and IRs. Yet, these empirical studies also fail to outline a theoretical basis to analyze the benefits and barriers that faculty face when confronting IRs. In the present article, I investigate the factors that affect the faculty contribution to IRs with a theoretical framework which combines on Socio-Technical Networks and Social Exchange Theory. Research questions are presented as follows:

• What are existing ways that faculty members make research/teaching materials publicly accessible on the Internet?
• Why do they want to make their research/teaching materials publicly accessible on the Internet? Why don't they do so?
• Why do faculty contributors submit their research/teaching materials to IRs?
• What makes faculty members reluctant to contribute to IRs?

2. Conceptual Framework

2.1. Conceptualization of IRs

IRs can be conceptualized as the following three entities: (1) electronic Scholarly Communication Forums (e-SCF); (2) digital libraries; (3) knowledge management systems. Several studies explained that an IR would be a new strategy for facilitating changes in electronic scholarly communication (Chan, 2004; Crow, 2002; Lynch, 2003; Shearer, 2003; Van House, 2003), and therefore, it can be embraced into a broad family of e-SCFs termed by Kling et al. (2003).

The first and second conceptualizations of IRs are particularly related to the socio-technical network model, which provides a framework to investigate the interactions of social and technical elements in e-SCFs (Kling et al., 2003). The socio-technical network combines participants with different roles, rights, responsibilities, resource flows, legitimacies and taboo behaviors. In this network, social and technical elements were neither separable nor reducible to one another, but mutually constituted through the interactions between participants, technologies and artifacts. The socio-technical network model provides a general framework that helps understand the interactions between social and technical elements in IRs, although it does not mention specific variables to be examined, especially for the incentive structure. The incentive issues pertain to factors that motivate and impede the faculty contribution to IRs - on which the present study focuses. In this respect, another theory is necessary to frame those factors in a concrete manner.

This socio-technical approach was also used by Van House (2003) who studies digital libraries as socio-technical networks - networks of technology, information, people, and practices. Since the concept of socio-technical network was applied to both e-SCFs and digital libraries, which implied the nature of IRs, the model is appropriate to examine IRs. When IRs are considered in the context of digital libraries, they are viewed not just information retrieval systems or digital resources collected on behalf of user communities, but also a component embedded in information-related activities of those communities and several information institutions, such as libraries and archives (Borgman, 2003). In this sense, IRs can be regarded as a type of digital library constructed by a university community through contributions of scholars and other members of the community. IRs also serve a wide range of user communities with the collaboration of multiple information institutions.

The third perspective of viewing IRs is as knowledge management systems (Branin, 2003; 2005). Knowledge management systems are defined as "IT- based systems developed to support and enhance the organizational processes of knowledge creation, storage/retrieval, transfer and application" [1]. Branin argues that more and more faculty and students in a university utilize information technology not only to access information but also to create new intellectual output in digital form. He suggests that the approach to knowledge management is relevant to the implementation of IRs that manage a wide range of digital information created in a university. Moreover, knowledge management requires investigating knowledge workers themselves, as well as social and cultural issues regarding knowledge creation and sharing. Several studies investigated factors that contributed to knowledge repositories or intranets in corporate environments based on social exchange theory (Hall, 2001; Kankanhalli et al., 2005). Social exchange theory posits that there are many social interactions outside the economic marketplace involving exchange of different resources, such as favors between neighbors (Molm, 1997) or information. Unlike most economic exchanges which are one-time events, social exchanges of information can take place recurrently based on the history of relations and the mutual contingency of behavior. This recurring exchange results in patterns of interactions and interdependence between people over a period of time, thus strengthening the system.

2.2. A Model of Factors on IR Contribution

Applying Social Exchange Theory to IRs, faculty members may consider costs and benefits implicitly in terms of IR contribution. Based on this assumption, my study suggests extrinsic and intrinsic benefits relating to IR contribution. Extrinsic benefits include accessibility, publicity and trustworthiness of documents in IRs (Kling and McKim, 1999), professional recognition (Swan and Brown, 2005), institutional recognition (Kankanhalli et al., 2005), and academic reward (Kling and Spector, 2003). Intrinsic benefits concern altruistic intention of and self-interest in the IR contribution (Cronin, 2005). Cost factors relate to copyright concerns (Gadd et al., 2003) and additional time and effort required to make the IR contribution (Foster and Gibbons, 2005).

In addition to cost and benefit factors, Nahapiet and Ghoshal (1998) suggested that relational social capital - trust, identification, and pro-sharing norms - influence the motivation to exchange knowledge. Kankanhalli et al. (2005) used those as contextual factors affecting the contribution to knowledge repositories. In the IR context, trust and identification are considered important factors. Trust indicates belief in good intent and competence of other actors, such as a university and users. Identification indicates faculty members' concerns with collective outcomes, membership and loyalty toward universities. Instead of pro-sharing norms, the IR literature mentions pre-print culture, in which researchers distribute drafts of research articles before they have been peer reviewed to colleagues around the world, as a factor. Figure 1 presents a model that depicts the relationship between various factors and the contribution to an IR.

Furthermore, individual traits might affect IR contribution. Kling et al. (2003) suggests that one of the central socio-technical features of e-SCFs is that the use of one or more existing communication channels may either encourage or discourage the use of a newly introduced e-SCF. Centered on this argument, the present study examines the relationship between faculty members' self-archiving experience and their contribution to IRs. Self-archiving refers to depositing scholarly content in publicly accessible web sites. If faculty members already have self-archived their research/teaching materials in web spaces, such as a personal home page, research group web sites or disciplinary repositories, they are more likely to contribute to IRs. However, since they already utilize one or more Open Access venues, they might not have the impetus to contribute their content to IRs. The present study, therefore, investigates whether or how much faculty members' exposure to self-archiving practices (exclusive of IRs) affects their contribution to IRs. Other individual characteristics that might relate to IR contribution include faculty rank, journal-related editorial service, and administrative roles. These positional variations imply individuals' control over resources in their universities or disciplines. Perceptions and experiences of self-archiving might differ across people in these positions. Based on these theories, I have developed the following conceptual framework to model incentives for faculty contribution to IRs (Figure 1).

The model considers the relationships between costs, perceived intrinsic and extrinsic benefits, individual characteristics, and contextual factors.  The goal is to isolate the factors influencing faculty contribution in order to better structure incentives and social mechanisms to foster contribution. This article reports the results of a pilot survey done to test the feasibility of the test of the model.

Figure 1. A model of factors that influence faculty contribution to IRs

3. Methodology

This article reports on a pilot survey of a sample of faculty members in ABC University, one of the 18 universities including members from each of the three strata of my study population. The university deployed a DSpace IR in 2004 and is currently in pilot testing phase. The sample included 67 assistant, associate and full professors whose materials are deposited in the IR.

The survey instrument consisted of four sections: (1) current IR contribution, including 1 yes/no question regarding awareness of IRs, 4 multi-choice questions, 30 Likert-scale questions and 1 open-ended question; (2) future IR contribution, containing 1 yes/no question about likelihood of future IR contribution, 8 Likert-scale questions about motivators, and 1 open-ended question; (3) self-archiving, consisting of 1 yes/no question about self-archiving experience other than IRs, 4 multiple-choice questions, 16 Likert-scale questions and 1 open-ended question; (4) demographic section including 9 questions. If surveyed faculty members indicate that they have awareness of the IR, plan to contribute to the IR in the future, and do other self-archiving practices, they are administered every section of the questionnaire. Otherwise, they will skip one or more sections depending on their awareness and experience of self-archiving. The questionnaire was developed in a web survey format.

Concerning the survey distribution, paper invitation letters were distributed first to 67 faculty members on April 5, 2006. The letter provided the URL of the web survey and an ID for each person. Faculty members were also given an incentive of a $5.00 gift certificate.  The invitation letter also indicated that a follow-up e-mail with a link to the survey would be sent immediately. The follow-up e-mail was sent twice after distributing the letters. As a result, 31 out of 67 (46.3%) professors responded. Out of 31 respondents, only 9 (29%) were aware of the IR. 13 out of 31 (41.9%) plan to contribute to the IR in the future. Also, 22 (71%) have made their research/teaching materials publicly accessible through venues other than the IR. The 30 Likert-scale questions for IRs, therefore, were answered by those 9 professors who were aware of the IR. However, the 16 Likert-scale questions for self-archiving practices were answered by all 31 respondents.

4. Preliminary Findings

4.1. Survey vs. Response Sample

The survey sample included 67 professors at ABC University; 31 out of the 67 responded. Table 1 presents the frequencies and percentages of both survey and response samples. It indicates that the response sample is proportional to the survey sample by disciplines, professional rank, and gender.

Table 1. Survey vs. response sample by discipline, professional rank and gender

4.2. IR Awareness and Contribution

Out of 31 respondents, only 9 professors were aware of the IR. This result indicates that although all of the respondents have materials in the IR, only a small number knew about the repository. This low level of awareness results from one current strategy used by libraries to populate IRs in which librarians collect and deposit materials on behalf of faculty members. The deposited items are generally pre-existing research papers in labs or departments, such as working paper series. Therefore, faculty members may not realize that their materials are in IRs. The other reason is that the IR of ABC University is still in a pilot testing phase and has not been widely publicized.

The 9 respondents who knew about the IR learned about it in various ways. Four came to know of the IR through publicity on a university/library web site, whereas each of the remaining five learned about it differently: contact from an IR staff member, presentation by an IR staff member at a faculty meeting, publicity through campus newspapers, results of a web search engine, and participation in an initial meeting of the IR.

When asked about his frequency of contribution to the IR, one respondent reported the frequency of contribution to the web site of his department saying, "I have been contributing through my department's web site even before the name 'IR' was coined." He had contributed to that web site more than five times, although not to the IR. Other than this case, no respondents reported contributing directly to the IR. This finding, again, suggests that there is a discrepancy between faculty whose materials are deposited in IRs and faculty who consider themselves as contributors. In addition, no respondents had searched the IR.

In spite of the low awareness of the IR, 13 (41.9%) out of 31 respondents planned to contribute to the IR in the future. Interestingly, 7 out of the 13 who were motivated to contribute to the IR had no awareness of the IR, but wanted to make IR contributions in the future. However, 3 respondents aware of the IR were uncertain about contributing to it in the future. Thus, among those 13 respondents who intended to contribute to the IR, 6 were already aware of the IR and 7 were not.  Of the remaining respondents, 5 (16.1%) had no plans to contribute in the future and 13 were undecided.

4.3. Self-archiving Experience

The present study is also concerned with respondents' self-archiving experience outside the IR. Twenty-two (71%) respondents had deposited their research/teaching materials on publicly accessible web sites other than the IR. Out of the 22 self-archiving respondents, 6 were aware of the IR and 9 planned to contribute to it. Therefore, most respondents had some IR awareness, and a majority of those who planned to contribute, already had experience with self-archiving. In addition, 3 respondents who were aware of the IR, planned to contribute in the future, and already had self-archiving experience in venues other than the IR.

.

Figure 2. Categorization of 31 respondents based on their awareness of the IR, likelihood of future IR contribution, and self-archiving experience

Figure 2 represents the number of respondents that fall into three groups based on their awareness of the IR, IR contribution in the future, and self-archiving experiences. It also shows the number of respondents in overlapping areas among these three groups.

In the introduction, the present study proposes four research questions; two of which relate to the self-archiving experience of university professors. In the following sections, therefore, I will focus on provisional answers to those two questions based on the pilot survey data.

4.3.1. Research Question 1: What are existing ways that faculty members make research/teaching materials publicly accessible on the Internet?
Out of the 22 respondents who had self-archiving experience, 9 (40.9%) had self-archived their work for more than 5 years and an additional 4 (18%) had done so for 4-5 years. Therefore, a majority had deposited their work in publicly accessible web sites for at least 4 years.

In the survey, I also collected data regarding what kinds of publicly accessible web sites the respondents used for self-archiving and what version of research articles they deposited there.  Figure 3 presents the frequency of deposit of refereed, published articles in the past 3 years in the following three types of web spaces: personal web pages, disciplinary repositories, and research group/lab/center web sites. Personal web pages were used most frequently by respondents (14), followed by research group/lab/center web sites (11), and disciplinary repositories (6). Interestingly, 4 respondents had never deposited any published articles in these types of web spaces. This result indicates that there may be other types of web sites that respondents employ for self-archiving, or they may not deposit published articles, but only other types of research work.

Figure 3. Types of web sites and frequency of self-archiving refereed, published articles

In addition, Figure 4 presents the frequency of deposit of pre-refereed articles in those three types of web spaces.  Similar to the self-archiving published articles, personal web pages were the most frequently used (10 respondents), followed by research group web sites (8), and disciplinary repositories (6). However, respondents were less likely to self-archive pre-refereed articles than they were to self-archive refereed, published articles.

Figure 4. Types of web sites and frequency of self-archiving pre-refereed articles

This finding is also supported by another result regarding the percentage of various research materials that respondents had self-archived in the past 5 years. These research materials included pre-refereed articles, refereed and published articles, unrefereed articles (technical reports or working papers), book chapters and data sets. Most respondents (15) self-archived published articles and unrefereed papers (Figure 5). Nine respondents self-archived more than 50% of their published articles, whereas another 5 self-archived more than 50% of their unrefereed papers. In addition, 10 respondents deposited pre-refereed articles on publicly accessible web sites, and 4 of those had self-archived more than 50% of their pre-refereed articles. Only two respondents self-archived data sets, and one of them had made 51-75% of his data sets publicly accessible on the Internet.

Figure 5. Percentages of self-archived research materials

Respondents also made other types of research/teaching materials publicly accessible. As can be seen in Figure 6, 15 (71.4%) and 14 (66.7%) respondents had self-archived lecture notes and course syllabi, respectively. Conference presentations were found to be the 3^rd most frequently self-archived materials. Yet, two respondents had self-archived none of the material types listed in the survey.

Figure 6. Types of self-archived research/teaching materials

In sum, the 22 respondents who had self-archiving experience used personal web pages more frequently than research group/lab/center web sites or disciplinary repositories. However, it is possible that some respondents might use other types of web sites that the survey did not cover for self-archiving. Refereed, published articles were self-archived by more respondents than pre-refereed articles. Respondents also tended to self-archive a greater percentage of refereed articles than other types. In addition to research articles, the majority of the respondents had self-archived lecture notes, course syllabi, and conference presentations.

4.3.2. Research Question 2: Why do they want to make their research/teaching materials publicly accessible on the Internet? Why don't they do so?
All 31 respondents responded to 16 Likert-scale items regarding benefit, cost, and contextual factors that influence their decision about whether to self-archive. These items were presented as statements offering a scale of options across five choices: strongly agree, somewhat agree, neutral, somewhat disagree, and strongly disagree. Just in case respondents did not sufficiently know what the items meant, the choice of "I don't know" was provided as well.

In order to answer this research question, I compared the 16 items' rating values between two groups: (1) The 22 respondents experienced in self-archiving; (2) The 9 respondents with no experience in self-archiving. For this comparison, I used a permutation test. Since the size of survey data was small, normality of data was not assumed. The permutation test is one of the non-parametric methods that can render reliable results even for a small set of data. The null hypothesis for this test is that there is no difference in the distribution of item ratings between the two groups and therefore, they have been drawn from the same population.

Out of the 16 items, I considered 7 items to be benefit factors regarding the increase in (1) accessibility and (2) publicity of their research work, (3) professional recognition, (4) positive impact of self-archiving on tenure and promotion, and (5) altruistic intention (See Appendix A). Two items concerned publicity, whereas two other items related to academic reward. As a result of the permutation test, I found no statistically significant difference in those items' rating values between the two groups. However, when I compared the mean values between the two groups, the results indicated that the group of respondents having self-archiving experience agreed with the benefit factors more strongly than the group with no self-archiving experience. In addition, two open-ended responses indicated that posting research papers on the web would save time compared to sending hardcopies or e-mailing files to those who requested the papers.

I considered 4 items as cost factors including (1) concerns about preservation of self-archived materials; (2) publishers' policies prohibiting self-archiving; and (3) additional time and effort required to perform self-archiving (See Appendix A). When I analyzed the 2 questions regarded as preservation issues, the permutation test did not show a statistically significant difference in item ratings or cost factors between the two groups. The groups also showed similar mean values for those items.  I did find a difference in mean value for the item regarding publishers' policies.

Considering the choice of 'strongly agree' as 5 and 'strongly disagree' as 1, the group having self-archiving experience rated the item 2.97 on average, whereas the other group's mean was 3.67.  Therefore, the group with no self-archiving experience perceived publishers as not allowing self-archiving more strongly than the group with self-archiving experience.

I created 5 items that were contextual factors, including 1) pre-print culture; 2) trust of readers; 3) influence of other actors - co-authors; 4) grant-awarding body; and 5) university or department actions - upon respondents' decisions to make, or not to make their materials publicly accessible (See Appendix A). The permutation test showed that there was a statistically significant difference in pre-print culture between the two groups. Table 2 presents the item regarding pre-print culture and its p-value.

Table 2. P-value of the item regarding pre-print culture from the permutation test

	p-value
In my field, it is common for researchers to post their work on publicly accessible web sites.	0.0318

Although there was no statistically significant difference in perception of trust in readers, two respondents expressed concerns about plagiarism of their research and teaching materials on the Internet. One of them mentioned that some researchers at other universities had taken the entire course outline that the respondent sent to them at their request, and simply replaced the headers with their information. He stated, "I do not believe we should let the bad acting of a few impair the access of the many." Another respondent mentioned, "Plagiarism is so common and widespread, that I would volunteer to put only dated (published) material or obvious material on the web."

Overall, respondents that had self-archived their materials tended to agree with benefit factors more strongly than those who had no self-archiving experience, although there was no statistically significant difference between the two groups. There was also no significant difference in cost factors, although the group with no self-archiving experience was more likely to perceive publishers as prohibiting self-archiving. The existence of pre-print culture was the only factor showing a statistically significant difference between the two groups. The group with no self-archiving experience might not have pre-print culture in their disciplines. This finding suggested that the existence of pre-print culture might be positively related to the decision to self-archive. Concerns about plagiarism might impede self-archiving as well.

4.4. Factors Affecting IR Contribution

The survey data showed that no respondents considered themselves to be contributors to the IR, and therefore, it was not possible to compare perceptions between IR contributors and non-contributors. Regarding future contribution to the IR, however, respondents were categorized into two groups: (1) 13 professors who planned to contribute to the IR in the future and (2) 18 professors who had no plan or were undecided about future contribution. It would be interesting to see the differences in perceptions of self-archiving between these two groups in order to conjecture about factors affecting IR contribution.

4.4.1. Research Question 3: Why do faculty contributors submit their research/teaching materials to IRs?
While the survey data did not directly answer this research question, the data does suggest some factors that would encourage professors to contribute to the IR. The 13 respondents motivated to contribute to the IR in the future rated the importance of reasons for the IR contribution using a 5-point scale. Table 3 presents those reasons based on ratings of respondents.

Table 3. Reasons for future IR contribution

I found that preservation was the most important reason for IR contribution, followed by the capability of the IR capability to show the frequency of viewing and downloading their materials. Institutional recognition was the 3^rd most important reason, although its rating was not as high as the first and the second reasons. Retaining copyright did not provide an incentive for future contribution. Respondents also did not connect functions provided by existing publishing systems with the IR. Thus, the peer review process and academic reward were considered least important motivators.

In addition, two respondents suggested that the IR would help researchers and students in the same field to communicate. Furthermore, one respondent mentioned that the IR might help to facilitate the coordination of interdisciplinary teaching and research efforts.

In order to examine motivating factors, I compared ratings of the 7 items representing benefit factors between the group that intended to contribute to the IR and the group that had either no intention or was uncertain about future IR contribution. Table 4 lists 4 items having statistically significant differences in ratings with their p-values from the permutation test.

Table 4. P-values (one-sided) of benefit factors from the permutation test

	p-value
Posting my research work on publicly accessible web sites will increase the chance to communicate my research findings to peers.	0.0459
Posting my research work on publicly accessible web sites will increase the potential impact of my work.	0.0018
Posting my materials on publicly accessible web sites will enlarge the readership of the materials.	0.0257
Posting my materials on publicly accessible web sites allows other scholars to access those that they could not otherwise use.	0.0202

The null hypotheses involved no differences of item ratings between the two groups. My alternative hypotheses were that the group motivated to contribute would have higher ratings on benefit factors. In order to test these hypotheses, I completed permutation tests to generate one-sided p-values. In these cases, all of these were statistically significant at the .05 level or less.

In Table 4, the first item represented an increase in accessibility, and the second and third referred to publicity. The fourth indicated altruism in contributing to the IR. This result suggested that professors who planned to contribute to the IR might acknowledge these benefit factors more than professors who either had no intention or were unsure of making an IR contribution in the future.

In sum, professors might be motivated to contribute to the IR by the prospect of an increase in the accessibility of their materials - both long-term preservation and an enhanced opportunity to make them accessible to peers. In addition, publicity factors - wider readership, increase in potential impact of their work, and knowing the usage statistics, would be positively related with IR contribution. Furthermore, respondents' altruism in making their materials publicly accessible was likely to be a motivator for contribution to the IR.

4.4.2. Research Question 4: What makes faculty members reluctant to contribute to IRs?
In order to investigate this research question, I compared ratings of 4 items regarding cost factors related to self-archiving between the group motivated to contribute to the IR and the other respondent group. The permutation test showed that there were no statistically significant differences in those items between the two groups. Yet, in responses to the item about the lack of secure maintenance of self-archived materials, the group that planned to contribute rated this question 2.90 on average, while the group with either no plan or uncertainty about future contribution to the IR rated it 3.71. The median rating of the former group was 3, whereas that of the latter group was 4. This result suggests that the latter group might be more concerned or skeptical about secure maintenance of open access materials than the former group. In order to attract the latter group, IRs might have to emphasize their function of long-term preservation and explain how it would be accomplished.

A permutation test was also conducted in order to compare ratings of 5 items regarding contextual factors between the two groups. I found that the single item concerning the influence of a grant-awarding body on respondents' decision showed a statistically significant difference based on a one-tail test (Table 5).

Table 5. P-value (one-sided) of the item about grant funders' influence on self-archiving decision

	p-value
My decision to make, (or not to make) my materials publicly accessible on the Internet was influenced by my grant-awarding body.	0.0176

This result indicated that professors who were motivated to contribute to the IR in the future perceived less influence of grant funders on their decision about whether or not to self-archive their work. Since the former group acknowledged benefit factors more than the latter group, it was suggested that grant funders' influence might be mitigated by respondents' already strong belief in positive outcomes from self-archiving. Conversely, this result suggested that the group with no intention or was uncertain about future IR contribution tended to perceive more influence of grant-awarding bodies on their decision to self-archive. If grant funders encourage self-archiving, the group of respondents would consider depositing their work into the IR. If not, they would have lack of motivation to contribute to the IR. Although this study did not determine the grant funders' attitude toward self-archiving, the lack of motivation for IR contribution might be led by grant funders that showed no interest in or discouraged self-archiving. 

5. Discussion

As depicted in Figure 1, I proposed that several benefit, cost, and contextual factors may influence professors' motivation for IR contribution, which will lead to the actual deposit into the IR. Since the survey data showed no knowing contributors to the IR, it was not possible to compare perceptions between IR contributors and non-contributors. Instead, the present study compared ratings of 16 items indicating those factors between the two groups of respondents: (1) 13 respondents who planned to contribute to the IR in the future; (2) 18 who had no plan to or were uncertain whether they would contribute or not. One or two of the 16 items corresponded to each factor except for identification. The items regarding identification were included in the scale questions answered by 9 professors who were aware of the IR. Due to the low responses to the items, those were not used for the analysis. In addition, those who planned to contribute to the IR in the future answered another set of scale questions regarding reasons for future IR contribution. Responses on the scale questions were analyzed to determine factors that encourage the IR contribution.

I found that faculty members who planned to contribute to the IR in the future agreed with some of the benefit factors resulting from self-archiving more strongly than professors who had no intention or uncertainty to make an IR contribution. Those benefit factors included an increase in (1) the chance to communicate research findings to peers, (2) potential impact of research work, (3) larger readership, and (4) an altruistic impetus for making research work accessible to other researchers. In addition, those faculty members motivated to contribute to the IR rated long-term preservation of their work as the most important reason for any future contribution to the IR. The next most important motivating factor would be the IR's capability to provide data on the number of viewing sessions and downloads of their work deposited in the IR. Currently, few IR installations provide such functions, although displaying the usage statistics would apparently encourage faculty contribution. Overall, faculty motivated to contribute to the IR appreciated the positive outcomes of self-archiving, especially growing accessibility and publicity of their research work, and displayed altruism.

I found no significant differences in the cost factors, such as copyright issues and additional time and effort, or in the contextual factors, especially trust and pre-print culture between the two groups. In particular, few respondents knew publishers' policies on self-archiving, 5 respondents who planned to contribute to the IR and 6 who did not plan or were uncertain to contribute provided "I don't know" answers. My conclusion here is that knowledge about copyright issues related to self-archiving is low. One respondent stated, "I have to sign copyright forms from most journals before they will publish research work. I don't know how this affects depositing [the] same in the IR."

Although my model did not include the influence of external actors - peers, university/department and grant-awarding bodies - on self-archiving decisions as a contextual factor, the item indicating grant funders' influence on such decisions showed a significant difference between the two groups; however, faculty who planned to contribute to the IR perceived less influence by grant funders on their self-archiving decisions than the other group. I speculate that the latter group's lower motivation for IR contribution may result from their grant funders' attitude to self-archiving.

My next step in this research is a survey of 1500 professors at 17 Carnegie-Doctorate Universities in the U.S., and 20-30 follow-up interviews. The analysis of the large-scale survey data will provide statistically more meaningful results than those of the present study based on the small set of data. As a result of this pilot study, I changed the focus for the larger study to examine the factors that effect self-archiving decisions in general, rather than solely on IR contribution. I will also investigate a wide range of self-archiving practices in greater detail. Furthermore, I think the 30-minute follow-up interviews will complement these survey data by providing interviewee's perspectives on self-archiving practices.

6. Conclusion

In the present study, I examined various factors influencing faculty contribution to IRs, based on a small set of survey data. Overall, permutation tests showed that benefit factors were more influential than cost or contextual factors.  Given the low number of respondents in the pilot, these findings are preliminary.  The pilot was also instrumental in developing the larger survey.  As a result of the pilot, the focus of the larger was changed to focus on self-archiving behaviors generally, with IR contribution as one instance.  My goal is that this latter survey and interview data collected from a larger sample of university professors will provide results with greater generalizability, and more insight into faculty motivations for and barriers to posting scholarly works on diverse open-access venues including IRs.