Reference Staffing and Scheduling Models in Archives and Special Collections: A Survey Analysis of Prepandemic Practices

Reference services form the core function of any type of library. Even when faced with challenges like shrinking budgets and staff sizes, library and archives workers continue to provide reference services to meet the demands of researchers. The use of archives and special collections holdings by an increasingly diverse spectrum of people beyond traditional researchers guide internal decisions about processing priorities, digitization initiatives, the development of accessible tools and resources, and the scope of outreach strategies that introduce new or previously underrepresented groups to archives. While reference work is central to the life cycle of access and use and is often used to justify the value of archival institutions to stakeholders and local communities, research about current reference practices and their impact on employees is difficult to find.

Personal experience working reference shifts and managing reference scheduling and staffing provided the impetus for research on this topic and prompted the following questions: Are reference models in archives and special collections working? Why or why not? Have service and staffing models remained the same because “it's what we've always done”? Have other institutions discovered successful ways to approach reference services? Are staff happy with the approaches used at their institutions, and do they feel invested in the systems in place? How might the level of employee satisfaction with specific reference models positively or negatively impact both internal operations and the delivery of services to patrons? Using these topics of inquiry as a starting point, the author developed a survey to gather quantitative and qualitative data necessary for a detailed analysis of archival reference services representing practices as of January 2020.

Although many studies of archival reference services focus outward on the needs and satisfaction of patrons, this study examines employee perspectives on staffing and scheduling models used in archives and special collections. The survey intended to discover 1) which systems for reference staffing and scheduling are currently in use, 2) how satisfied participants are with their staffing and scheduling systems, and 3) what common challenges and successes institutions have experienced related to reference services. In addition to these questions, the information collected in this survey provides much-needed benchmark data from participating institutions about the number of special collections and archives staff who contribute to public services, the length and frequency of shifts, and typical service hours.

Why is this study important to conduct? Archives and special collections continually face budgetary reductions that lead to fewer staff, coupled with more collections and responsibilities to manage, yet are expected to maintain the previous level of service under these increased demands. More than a decade ago, RBM: A Journal of Rare Books, Manuscripts, and Cultural Heritage dedicated an issue to confronting “‘the new normal' of reduced resources and increased expectations.”¹ Librarians and archivists still experience this pressure and its effects. Recent examinations of time, labor, and workload among academic librarians and special collections librarians in public services roles revealed a reduction of staff that led to intensification of work demands, an increase in job responsibilities beyond position descriptions, and the need to balance competing institutional and departmental priorities.² Gaining a better understanding of the problems and successes found in reference work provides an opportunity for institutions to make more informed choices about staffing and scheduling approaches. Administrators and department heads reexamining reference services upon emerging from the COVID-19 pandemic should consider efficient reference models that save employees valuable time and energy. Finally, this article identifies a disconnect between the relatively high value of reference services within institutions compared to the lack of support given to the people who provide these essential services, paving the way for future study.

An extensive body of research has been conducted about reference service models in academic and public libraries over the past two decades as many institutions moved librarians away from service desks and into on-call or consultation roles. A combination of factors, including changes in the mode of contact (from in-person to chat messaging or virtual calls), a decline in reference questions, competing priorities, increasing teaching loads, and decreasing budgets, aligned to cause widespread changes in library service models.³ A 2014 survey of 600 academic librarians revealed that high-quality reference services can still be provided to users even as service desks transition from staff with MLS degrees to students. Institutions devoting additional attention to an expanded suite of services—such as chat, consultations, LibGuides, tutorials—and increased promotion of these services were more likely to be satisfied with their modified service model.⁴

Likewise, existing studies that focus on reference staffing and scheduling practices are most often written from an academic library perspective and rarely include archives or special collections. Results of these studies generally conclude that collaborative approaches to staffing and scheduling positively impact the organization. In one study, the authors posit there is no definitive approach to successful reference scheduling but suggest the scheduler must be “fair, flexible, good at resolving conflicts . . . and place the overall good of the library before the needs of the individual.”⁵ Another study investigated three different approaches to reference scheduling in an academic library before selecting a cooperative group scheduling model.⁶ As for examinations of reference staffing, Marlys Brunsting's survey of more than 250 midsize academic libraries compared the number of questions asked per week or per hour with the staffing model used at service points (e.g., librarian only versus student and staff run). Her study also identified specific factors that managers considered when deploying reference staff across service points.⁷ Carol Waggoner-Angleton's examination of student assistants in libraries and archives offered several possible approaches for effectively utilizing this segment of the workforce.⁸ The latter two studies include areas of inquiry worth exploring further in response to the staffing data gathered in this survey.

Critical consideration of reference models in the archives and special collections field, however, is lacking. James Cross, writing in 1997, pointed to the scarcity of scholarly studies about reference work as a sign that reference may be “an ‘afterthought,' the end of a linear process that starts with appraisal and moves through accessioning, preservation, arrangement and description.”⁹ In surveying articles published in American Archivist, only six appear with key words for reference or public services, all published since 2013. Of those, only one is closely related to the provision of archival reference services—Wendy Duff, Elizabeth Yakel, and Helen Tibbo's study on developing a scope of competencies for reference archivists.¹⁰ Topics covered by other reference studies in the field include the shift from in-person to remote/online reference and changes in methods of patron contact,¹¹ the accessibility of reference spaces and collection materials,¹² the impact of disruptions like renovations or natural disasters on public services staff,¹³ and the role of the reference archivist in developing online tools and resources.¹⁴

Most studies in archival literature related to reference work naturally focus on users of archival collections and the services provided to them. A trend toward user studies began in the mid-1980s, encouraged by Paul Conway's seminal piece, “Facts and Frameworks: An Approach to Studying the Users of Archives,” published in American Archivist in 1986.¹⁵ As archives and special collections institutions gradually shifted their attention from preservation and maintenance of collections toward outreach and patron access to the collections, the need to learn more about researchers grew. User studies convey critical information about who is visiting archives and special collections, how users interact with collection resources and materials, and how successfully institutions deliver what researchers want.¹⁶ It is also crucial, however, to consider employees' perspectives in the methods used to deliver these services to patrons. The relationship between user needs and staff needs should be symbiotic, acknowledging the limitations of staff to provide everything that patrons desire as well as the invisible labor prevalent in library and archival work. This article is one attempt to fill the gap regarding data from an employee perspective.

In addition, differences between the nature of reference services in libraries and those in archives and special collections have allowed the field to neglect close examinations or reconsiderations of reference and scheduling models. Attitudes contributing to this dearth of research include the general perception that MLS-trained staff need to be available to provide archival reference services because the questions received are “substantive, obligatory, and continuing,” according to Mary Jo Pugh,¹⁷ and thus more likely to require a higher level of assistance than those received at a library public services desk. This perspective—while rooted in truth—minimizes the urgency to evaluate or reconsider reference service approaches as has happened at academic library information desks, described above. Practitioners continue to adhere to reference models that heavily rely on mediated help from trained professionals without closely analyzing the makeup of in-person reference questions received in archives and special collections.¹⁸ Existing research examining the content of archival reference inquiries has primarily been confined to email and chat reference services.¹⁹ The data analyzed in this paper suggest it is necessary to experiment with reference models that balance the need for mediated reference assistance from specialists with a triage system for basic informational requests. This type of approach would respect both the reference archivist's expertise and their time. This topic is deserving of further examination, though beyond the immediate scope of this article.

Perhaps the most comprehensive collection of data about reference and public services practices in archives and special collections to date is the Association of Research Libraries' (ARL) SPEC Kit 296, “Public Services in Special Collections,” released in 2006.²⁰ The report contains responses from seventy-nine ARL libraries on topics including reference staffing, offered services, patron registration and access, and major changes in reference and public services in the ten years prior to the survey. An obvious limitation of the ARL data set is its focus primarily on academic institutions. Qualitative information provided by SPEC Kit respondents suggested “the need for flexibility in staff workloads, training to provide skills that have been lacking, as well as greater cross-training between Special Collections staff and staff in other library units.”²¹ These challenges and suggestions mirror many of those reported by respondents in this survey. In addition, while the focus of this article is narrower than the SPEC Kit, there are a few overlapping points of comparison, like the number of trained staff providing reference services and each institution's service hours per week. Implementing (and sustaining) a holistic survey on public services in archives and special collections at a national level every five to ten years would provide institutions with valuable information that is currently missing.

This survey covered six areas of reference staffing and scheduling: service model, staffing model, reference frequency, scheduling tools, scheduling method, and challenges and successes with reference services. In four of the survey categories, participants reported their satisfaction with the system used at their institution. For the purposes of this survey, service model describes the portion of an institution's trained staff who participate in reference. Staffing model describes which types of workers by employment category participate in reference services. Scheduling method describes the system by which people select or are assigned reference shifts.

The survey instrument contained fourteen multiple choice questions, two open-ended questions, and one demographic question. To accommodate as wide a range of respondents as possible, the questions were necessarily broad. General terminology of reference shifts or reference services is used throughout rather than reference desk, in acknowledgment that the services provided are not always tied to a specific physical space. Special collections employees at the author's institution as well as two external librarians/archivists in the field tested the survey instrument for comprehension and flow.

The web-based survey was created in Qualtrics and deemed exempt by the Louisiana State University Institutional Review Board. The author distributed the survey primarily through listservs, with postings to multiple Society of American Archivists' sections, the Rare Books and Manuscripts Section of the Association of College and Research Libraries, regional archival organizations, and the Council of State Archivists, along with reposts to social media platforms. The survey instructions encouraged institutions to submit one response per library.²² Participation in this anonymous online survey was voluntary.

The survey was open from December 2, 2019, through January 31, 2020. It generated 234 total responses, 77 of which were discarded from the analysis as being less than 50% complete. The data set analyzed here reflects the 143 surveys of 100% completion, plus 14 surveys of 50% to 90% completion (N=157). The survey instrument and a summary of responses appears in the  appendix.

Although the survey was anonymous and none of the questions required a response, participants could choose to submit their institutional affiliation (see Table 1). Approximately 60% of respondents represented academic institutions, 30% represented nonacademic institutions, and the remaining 10% did not respond to the affiliation question. Participants in the “other” category listed the following affiliations: medical archives, historical society, nonprofit institution, and government contractor.

In addition, rather than reporting the total number of employees working in the library or a unit's full-time equivalency (FTE), participants provided the total number of people trained to serve reference shifts by employment category. A trained employee is a person from any employment category who is equipped with the knowledge to serve a reference shift, even if they were not actively scheduled for shifts at the time of the survey. The 7 employment options included librarians/archivists, staff, graduate assistants, hourly students, volunteers, interns, and other. According to survey respondents, the total number of trained employees participating in reference services ranged from 1 person to 57 people, with an average of 10.6 staff per institution (see Table 2).²³

Approximately 50% of the participants who reported their total number of trained employees use only librarians, archivists, or staff for reference work (N=75). The remaining half of participants also employ students, volunteers, interns, or people in the “other” category for reference work (N=76). While there is not a strong correlation between staff size and the use of students, volunteers, interns, or others, institutions with more trained employees were more likely to use workers in these 4 categories. The average trained staff size for institutions using students, volunteers, interns, or others for reference services was 11.5. The institutions using only librarians, archivists, or staff for public services reported an average trained staff size of 9.75, slightly less than the average for all institutions. Six survey respondents skipped this question.

Overall, respondents reported satisfaction with the systems in place at their institutions. Average rates of satisfaction in the 4 categories queried range from 64% to 77% (see Table 3). In analyzing the responses about satisfaction, however, it became clear that participating institutions unsatisfied with one aspect of their reference model were more likely to report dissatisfaction in other areas of the survey—and the same was true for institutions reporting high degrees of satisfaction. For example, among the 27 responses of “somewhat” or “very” dissatisfied with their service model, 20 also reported dissatisfaction with their staffing model. Among the 35 participants “very satisfied” with their service model, 34 reported the same level of satisfaction with their staffing model. This “very satisfied” group showed higher than average satisfaction rates for their scheduling tools (77%) and scheduling method (86%). Responses closer to neutral in satisfaction rate display more variability from question to question.

When analyzing satisfaction rates by institution type (see Table 4), academic institutions polled very close to the overall average, while respondents who self-identified as nonacademic institutions (the 5 demographic categories plus “other” that are listed in Table 1) reported above-average satisfaction rates. The lowest average satisfaction rates corresponded to respondents who did not select an institution type, though the response rate and small number of respondents in this group make forming any conclusions difficult. The “staffing model” category displayed the highest discrepancy in satisfaction when comparing academic institutions to nonacademic institutions, with 66% satisfaction for academic respondents versus 84% satisfaction for nonacademic respondents.

Staff size also impacted the average rates of satisfaction across all 4 categories (see Table 5). Institutions using 11 or more trained employees for reference services showed a higher degree of satisfaction with their tools and systems than those with 10 employees or less. The largest institutions—with 21 or more trained reference staff—reported the highest levels of satisfaction, while institutions with 6 to 10 trained staff were least satisfied. Conversely, staff size had very little bearing on the specific type of service model, staffing model, or scheduling model reported by institutions in this survey. No trends emerged that could point managers or employees toward the most effective type of approach based strictly on institution size. There is not a “one size fits all” solution to reference staffing and scheduling.

The open-ended questions about challenges and successes with reference services revealed many commonalities across institutions. Among the responses describing the biggest challenges in reference services, 95% of respondents reported problems with scheduling, staffing, training, or workload. More than a third of participants specifically cited staffing shortages. Successes in staffing and scheduling revealed a wider variety of responses including components of reference services that work well, such as the reference approach, scheduling methods, or service hours, as well as specific qualities that help create a successful system, like consistency in assigned shifts. Only 5% of participants said nothing worked well with their reference staffing or scheduling approach.

Despite these commonalities, the survey responses about challenges and successes again showed that there is not a single approach to reference staffing and scheduling that works well for every institution. Widespread buy-in for the chosen system, paired with employees' increased sense of agency to contribute to reference services and a willingness to collaborate with colleagues, increases the chances of maintaining a successful reference model. A detailed analysis of each survey category follows.

Among 152 responses, most respondents used the “all trained employees” model (45%) for reference shifts, followed by “only a select group of trained employees” (32%), then “most trained employees” (23%). Participants also indicated whether their institution uses an on-call reference system, in which staff members are paged to a public service point only when they are needed to assist patrons. About 28% of participants (N=42) reported using an on-call system. Of this subset, about 40% utilize “all trained employees” as their service model, while another 40% use “a select group of trained employees.” There was no discernable correlation between use of an on-call system and institution size or type.

The service model responses revealed the highest average rate of satisfaction at 77% compared to the other categories queried. Yet when examining the rate of satisfaction for each service model, “all trained employees” and “most trained employees” earned satisfaction rates of 80% and 86% respectively, while institutions using a select group of trained employees reported only 68% satisfaction (see Figure 1). Almost a quarter of participants using a select group of employees for reference services said they were somewhat or very dissatisfied. Among nonacademic respondents, satisfaction rates showed a marked degree of success using the “all trained employees” or “most trained employees” model (90% satisfaction for both), compared to 67% satisfaction for a “select group of trained employees.” Institutions using an on-call system reported a higher-than-average satisfaction rate of 80%.

The staffing model questions divided employees into two categories: 1) archivists, librarians, and staff, and 2) students, interns, volunteers, and other. The first category included employees in professional, paraprofessional, and contract roles regardless of full-time equivalency (FTE). The second category generally represented nonsalaried workers. The largest portion of respondents (45%) indicated using “all archivists, librarians, or staff” for reference shifts, while 33% use “mostly archivists, librarians, or staff.” A smaller number of participants (14.5%) reported using an equal proportion of archivists, librarians, or staff and students, volunteers, or interns for reference services. Only 7% used mostly students, volunteers, or interns for reference: 8 academic institutions, 1 public library, and 2 from unknown affiliations.

This category displayed an average level of satisfaction of 70% (see Figure 2). Of the 4 staffing models provided, the most satisfied group (86%) were those using 50% archivists, librarians, or staff (N=22), followed by “all librarians, archivists, or staff” at 75% (N=68). The participants using mostly students, volunteers, or interns were the least satisfied, with 7 of 11 indicating dissatisfaction or neutral feelings. There was a slight correlation between institution size and staffing model. Smaller institutions were more likely to use only salaried employees. Institutions using all or mostly archivists, librarians, and staff had baseline staffing of 1 to 2 trained employees, while institutions using 50% or more students, volunteers, and interns for reference services reported at least 4 trained employees. When analyzing by institution type, nonacademic institutions were most likely to be satisfied with their staffing model. These respondents reported satisfaction of 80% or higher across all 4 staffing models. Academic institutions showed a higher degree of variability, reporting 88% satisfaction using an equal percentage of employees in all categories, but only 38% satisfaction when using mostly students, volunteers, or interns.

Participants were asked to provide the number of reference shifts per day and the average shift length, as well as scheduling frequency. The results show that most institutions (87%) offer reference services at least 5 days a week, as of January 2020 (see Table 6). Another 8% provide reference services 4 days per week. More participating institutions offered Saturday hours (31%) than Sunday hours (13.8%). Only 13 of 141 respondents offered service hours on Saturdays and Sundays: 1 museum, 1 nonprofit institution, 4 academic institutions, 6 public libraries/archives, and 1 of unknown affiliation. Among academic institutions, only 7 of 92 respondents were open fewer than 5 weekdays; however, of those, most were departments with only 1 full-time staff member or that were experiencing staffing shortages.

Across the data set, an almost even number of participants reported average reference shift lengths of 3 hours or less (54%) compared to those with shift lengths of 3 hours or more (46%). The highest percentage of respondents reported shifts in the range of 2 hours to 2 hours and 59 minutes (36%). Among the 93 academic institutions that responded to this question, a higher percentage than average reported shifts of 3 hours or less (62%). This could point to a greater number of trained employees serving reference shifts in academic institutions, allowing each shift to be shorter. Four participants did not respond to this question but indicated being solo archivists in subsequent questions about shifts or scheduling who may not serve set reference shifts.

Participants also reported how frequently their reference services scheduling period occurred, if applicable. The largest number of participants (38%) said their reference schedules change by semester or term, indicative of the high number of responses from academic institutions. This suggests that about 60% of the academic participants adjust their reference services schedule by semester. Twenty-four percent of total participants indicated that their institution does not have a scheduling period, which was more common among smaller institutions based on the average of 6.3 trained staff for responses in this category. Among the 13% of respondents who indicated “other,” the most common replies included quarterly scheduling (N=5), permanent schedule (N=5), and solo archivists (N=5).

One ambiguously worded question did not convey the information it was intended to gather (Q10: How often does your reference services schedule change?). The question meant to determine whether individuals serve the exact same reference shift (same day of the week and time of day) for a month, a semester, a year, and so on, or whether individuals' scheduled shifts (day/time) vary by each week or month. While some respondents understood the author's intended interpretation of the question, others clearly thought the survey question was asking how often their department's service hours change. The data collected for this question are reported in the  appendix, but not analyzed here due to this confusion.

The survey offered 7 options for tools used to schedule reference shifts, plus a write-in category. Participants selected as many tools as needed to accurately represent their approach: 110 of 145 respondents (77%) reported using 1 tool for scheduling, 31 respondents indicated using 2 tools, and only 1 participant indicated using 3 tools. The most common 2-tool pairings were 1) Google Calendar and Google Sheets, 2) Microsoft Outlook's calendar and Microsoft Excel, and 3) Microsoft Outlook's calendar and something else. Overall, approximately 9% of respondents do not use any tools for scheduling. All but 1 of the respondents not using scheduling tools were associated with institutions of 10 trained reference staff or less. The smaller the staff, the less necessary automated tools appear to be.

Satisfaction rate for scheduling tools was the lowest of the 4 categories queried at 64%. Among those unhappy with their library's tools, both “very dissatisfied” responses used Microsoft Outlook calendars. Four of six SpringShare users were “somewhat dissatisfied.” About 20% of dissatisfied participants use Microsoft products, yet 20% of Microsoft users reported being “very satisfied.” The largest percentage of “very satisfied” respondents use Google products (43%). Eight of the thirteen respondents not using scheduling tools reported being “very satisfied.” It is difficult to arrive at any general recommendations about scheduling tools due to the limitations of this section of the survey. The participants were not asked to explain why their scheduling tools worked well or not. However, additional details about problems or successes linked to specific scheduling tools appear below in the analysis of the open-ended questions.

Participants provided the scheduling method that best fit the system used at their institution. Almost 50% of participants reported that staff are assigned reference shifts by a person or team. Thirty percent of respondents ask staff to provide availability to a person or team before setting the reference schedule. About 10% of respondents do not require a scheduling method for reference. In the initial results, almost 30% of respondents replied with “other.” After reading through the open-text fields describing specific scheduling methods, some responses of “other” were coded to preset survey options. The remaining 25 “other” responses then corresponded to 5 new categories, listed in Table 7. Only 4 of the total responses indicated use of 2 different scheduling methods, usually divided by employee status such as “faculty send availability, staff are assigned.”

The average rate of scheduling method satisfaction for all respondents is 68%. Two methods achieved higher than average satisfaction rates: “staff sign up for shifts under their own initiative” (73%) and “staff members are assigned shifts by a person or team” (71%) (see Figure 3). When examining the responses according to institution size, the smallest institutions (1 to 5 trained reference staff) were most likely to choose “other” as their scheduling method. Participating institutions reporting the highest use of “staff are assigned shifts by a person or team” were those with 6 to 10 trained staff (56%) and 21+ trained staff (44%). “Staff send availability to a person or team” appeared as the most popular approach for institutions with 11 to 15 trained staff (45%) and 16 to 20 trained staff (56%). Among academic institutions, “staff send availability to a person or team” resulted in the highest degree of satisfaction (75%), while assigning staff to references shifts was the least successful method at 59% satisfaction.

The first of two open-ended questions asked respondents to consider their institution's biggest challenges in scheduling or staffing reference services. The 125 responses were hand-coded using Excel and grouped into 8 categories (see Table 8). Because of the complexity of some replies, 45% of the responses fit into more than 1 category. While the responses reflect the individual contexts of each participating institution, common challenges appear across the data set. Ninety-five percent of responses reported problems in at least 1 of 4 categories: scheduling issues, staffing issues, training issues, and workload issues.

Scheduling issues—the most frequently reported problem—revealed three recurring challenges: difficulty finding coverage for staff absences, difficulty planning around unpredictable student schedules, and difficulty creating an equitable reference schedule among trained employees. Most responses mentioned the need to quickly readjust schedules for typical staff absences like vacation and sick leave. Institutions of all sizes reported scheduling problems, yet reference coverage is especially problematic for libraries with 5 or fewer trained staff members because it can result in the temporary closure of reference services. Meetings and other work responsibilities take staff away from service points with no additional staff or students available for backup. As a survey respondent explained, “Sometimes the only day a researcher can come in doesn't work for us because we are all attending meetings that can't be rescheduled and the researcher misses out.” By institution type, a similar percentage of academic institutions (44%) and nonacademic institutions (47%) reported scheduling issues. Only 8 responses describing problems with scheduling explicitly mentioned staffing shortages. It is natural to assume that at least some of the scheduling issues could be resolved with additional staff.

More than a third of survey participants (36%) conveyed reference staffing shortages at their institutions. This category includes responses related to the need for additional staffing (either full-time or hourly) to adequately cover reference services or the need to reassign existing staff to take on reference duties. About 15% of respondents said that their institution does not have anyone permanently assigned to cover reference services, resulting in less-than-ideal reference triage systems. The average number of trained reference staff for institutions reporting staff constraints (8.3) is less than those reporting scheduling, training, or workload issues, signaling that smaller institutions have a greater likelihood of experiencing staffing problems impacting reference services. When analyzing responses by institution type, approximately 40% of academic institutions reported staffing issues compared to 27% of nonacademic institutions. These staffing shortages mirror the results reported in OCLC's 2010 research survey of special collections and archives. Participating institutions indicated to OCLC a 23% decrease in public services staffing in the ten years prior to the survey, more than any other library unit.²⁴ Though this survey did not ask respondents to report changes over time in FTE staffing for public services, the data suggest it remains a high-profile issue.

“If our mission is to serve the public, why aren't we investing in staffing to fulfill that mission?,” asked one survey participant. This question seemed to voice the underlying concerns of many other respondents. It also represents the correlation between staffing issues and lack of support from a supervisor or administrator. Six of the seven responses that reported lack of support from upper administration also reported staffing shortages. One respondent said their institution was forced to cut reference services hours after an archivist position was not refilled. Another conveyed their department's desire to reduce service hours based on visitor data, but their administration would not support additional hiring unless service hours expanded. A third participant described a lack of staffing to cover reference shifts and an administration that would not allow service hours to be reduced accordingly.

Challenges in the training category covered the various levels of subject knowledge or training among people providing reference services and difficulties in consistently applying institutional policies and procedures. Several respondents mentioned cross-training employees from outside units to assist with archival reference services or utilizing student workers to triage reference services with full-time staff on call. In both cases, the survey responses indicated that additional training is often required to ensure that staff understand the procedures and workflows unique to archives and special collections. For example, one respondent said, “Having students on the desk does not always work. They might not understand all [elements] of a good reference interview or [might not] think of other ways to help patrons.” Problems of this sort are most likely to have an external impact, resulting in uneven quality of service for patrons. Ten of the twenty-four responses in this category explicitly mentioned the negative impact on researchers resulting from inconsistent training.

The workload issues described by participants often included the struggle to find a balance between providing reference services and working on other projects or tasks. Some respondents communicated that reference shifts take time away from contributing to other work responsibilities. Responses about the unpredictable nature of reference work—for example, adequately staffing the desk to match demand—are included in this category as well. One participant's response succinctly reflected the perspective of many in this category in saying, “All staff who have reference shifts are also responsible for many other duties. There is always too much to do and too little time.” Several responses from larger institutions that use technical services staff for reference services conveyed a perspective that participating in reference duties equates to lost time for other work responsibilities.

Similarly, 4 participants reported struggling with colleagues who believe only certain employees should cover reference shifts; these replies are grouped together as “combating institutional perceptions.” One response indicated that few staff see reference services as being part of their job. Another said, “It would be helpful if employees who accession and process collections (back of the house) would be more willing to participate and share their knowledge of collections and materials with researchers through pulling a shift on the reference desk once per week. This a much more holistic model and would contribute to greater productivity and communication between colleagues working in the same department.”

The second open-ended question asked survey participants to consider what works well in staffing and scheduling reference services at their institution. Responses to this question were again hand-coded in Excel. The responses conveyed two different kinds of information: 1) specific components of reference services that work well, and 2) descriptive qualities of successful reference services (see Table 9). Many replies included 1 component and 1 quality; in fact, 50% of responses corresponded to 2 codes. The most common pairings of components and qualities were collaborative reference approaches (N=16) and consistent service hours or shift lengths (N=12).

Of the 116 responses to this question, however, 10 participants could not identify any aspect of their systems that worked well or did not provide enough information to be categorized differently. As an example, one institution linked its problems to a simultaneous reduction in staff and increase in work responsibilities resulting in a lack of quality service provided to patrons and difficulty maintaining collection security. Another pointed to a hierarchical issue in which senior staff members are permitted to forgo reference duties leaving the bulk of the demand to remaining staff. A third response linked their unit's troubles to a lack of training for student workers and a service model that hampered trained staff's ability with reference requests.

Among the 5 components of successful reference services provided by survey respondents, the highest percentage of participants (31%) reported that their reference approach worked well. The reference approach category included any mention of efficient staffing models or systems for triaging reference questions and interacting with patrons. The responses reflected the high degree of satisfaction in institutional service models discussed earlier (77%), whether preferring to use a core group of public services staff for reference or engaging a wider pool of employees to maintain reference services. Institutions using a small group of reference staff were able to mitigate the workload issue described in the previous section: the struggle to balance public services duties with other work. As a respondent replied, “Having two people whose primary responsibility is reference means they expect to do that work and it doesn't interfere with any other of their responsibilities.” Another said, “Having dedicated reference staff allows other staff to work uninterrupted on other tasks, knowing that it is a real necessity if reference staff ask for assistance.” Other participants emphasized the increased level of efficiency and clarity with ongoing reference requests when using only a few employees for reference services.

Respondents with service models utilizing all or most employees for reference work showed how a collaborative approach benefits the institution's work beyond public services. According to one participant, “All staff are required to do reference shifts. It helps keep everyone abreast of the types/topics of reference inquiries, which informs processing priorities and other management decisions.” Other responses echoed this one in emphasizing that participation in public services impacts descriptive and arrangement work, helps staff understand how their individual responsibilities translate to patron needs, and creates well-rounded employees. These responses also conveyed the positive impact of having a depth of knowledge among the employees providing reference services: “Due to how we work, someone is always available, with others who have specialized knowledge not far behind.”

About 25% of participants discussed scheduling methods or tools that function well. Twelve of the thirty responses tied their success, in part, to tools: Microsoft Outlook, Google Calendar, LibStaffer, Slack, and WhentoWork. Slack, which had not been mentioned in the scheduling tools section, was used by two different institutions to convey information about upcoming research visits, researchers, and collection retrievals among staff in real time. Numerous responses in this category indicated that employees at their institution feel empowered to proactively arrange their own shift swaps: a sign of an effective system.

Another 25% of respondents indicated that their institution's service hours or shift lengths contribute to success. The ability to maintain predictable shift assignments was the most common response in this category (9 of 27 responses). A set system allows staff to plan for other responsibilities further in advance knowing their reference shifts will remain the same. Several participants said this consistency also helps staff remember their shift assignments. Seven survey responses specifically cited shift length as a positive aspect of their reference system, with shifts ranging from 2 to 4 hours—similar to the average shift length (2 to 3 hours) reported in question 8. One respondent provided a detailed explanation in favor of longer reference shifts: “Each reference shift is four hours so that staff can build stronger rapport with researchers through orientation and the reference interview process. We have designed longer shifts because archival research requires more detailed conversation in order to develop an understanding of the user's information needs and how this can be met through resources in the archives.” Other participants preferred desk shifts of 3 to 4 hours to reduce the number of shift changes per day and total shifts per week. Many of the responses in this category overlapped with the 22% respondents who highly valued consistent reference scheduling or staffing practices.

Beyond specific components of reference services, collaboration appeared as the most common quality of successful systems across all institution types. Approximately one-third of survey replies (31.9%) reported that the collaborative nature of their institution's approach to reference positively impacts staffing or scheduling. As one participant reported,” The only reason it works is a shared commitment to providing good service.” Other responses echoed this perspective in sharing that their colleagues take reference responsibilities very seriously. Several noted that working as a team to develop policies and procedures contributed to success due to departmental buy-in. Collaboration also appeared in numerous responses related to employees helping each other with shift coverage. For example, a respondent said that their system works well “because most folks are willing to pick up the shifts of their colleagues when needed.” Supportive supervisors help promote collaborative attitudes, indicative in this survey response: “Our director is attentive to our scheduling conflicts and needs. She is open to suggestions and improvements. This attitude and leadership sets the tone and creates a work environment where we support each other and work as a team to staff reference services.” These responses illustrate how staff buy-in and a willingness to collaborate with colleagues increase the chances of maintaining a successful reference model.

With the onset of the coronavirus pandemic less than two months following the conclusion of this survey, reference practices reported in January 2020 were almost immediately disrupted as institutions closed or moved to virtual and appointment-only systems. At the time of this writing, very little published literature related to the pandemic has discussed impacts on public services operations in archives and special collections. One notable exception is Planton and Ferrara's survey about disruption in special collections public services units, which initially focused on renovations and natural disasters but also incorporated questions about the impact of COVID-19.²⁵ Another survey polled 145 public services librarians in academic libraries to determine the budgetary and staffing impacts of COVID-19, as well as its effects on decision-making, communication, and research productivity.²⁶ A study of academic libraries that are members of the Association of Research Libraries compiled data about hours of operation, access restrictions, and main library circulation services like interlibrary loan and reference consultations, as well as the prevalence of COVID-19 mitigation policies during the 2020–2021 academic year.²⁷ More informally, members of the community developed several public spreadsheets or LibGuides sharing data about institutional closures, materials quarantine, and other local policies since 2020.²⁸ Additional surveys are needed to determine long-term effects of COVID-19 in the field, especially whether archives and special collections have returned to previous practices or permanently modified aspects of their reference services.

Despite the unexpected changes in reference services, however, the baseline data collected in this survey presents valuable information about staffing and scheduling which institutions can use to reevaluate their approach to reference when considering a return to prepandemic service levels. The data gathered in this survey lead to several conclusions. The most satisfied survey respondents used all types of service, staffing, and scheduling models, which suggests that methods do not automatically equate to success. Instead, high levels of satisfaction often relied on the willingness of staff to actively participate in the system used. Collaborative approaches to reference work, equitable and predicable schedules, and team-based creation of policies and procedures also increased satisfaction. This measure of success depends in part on the specific interpersonal relationships among people working in each unit and may be hard to replicate. Improving public services on a local level needs to rely upon feedback from all employees, with managers listening to understand common pain points or successes inherent in their current reference structure. Institutions wishing to break down silos and increase the cooperative nature of their unit may want to explore writing or reaffirming mission statements or developing yearly objectives as a department-wide exercise. Participants from larger organizations with many staff contributing to reference were more likely to be pleased with their staffing and scheduling systems. The limitations of this survey make it difficult to know if larger institutions are likely to obscure problems that are more easily identifiable in smaller special collections and archives units. Yet even participating institutions reporting high levels of satisfaction faced challenges reported by many other survey respondents, including trouble covering staff absences, the need for retraining or cross-training, and overwhelming workloads.

Adequate staffing for public services is one of the biggest issues participating institutions face, as explicitly stated and implied in the qualitative responses. Staffing concerns affect all types of institutions, including those generally happy with their unit's reference approaches. Among the most dissatisfied respondents, however, twelve of eighteen listed a lack of staffing as their primary challenge. Some report they do not have any staff dedicated to reference; others said the workload is too much for existing public services staff (sometimes only one person) to handle. In addition, staffing shortages are more likely to exacerbate problems with scheduling shifts or with employee workload. These responses mirror those reported in the literature.²⁹ Dedicating more resources toward staffing reference units, either through new full-time positions in public services or reallocation of existing staff responsibilities, is a logical first step in increasing employee satisfaction with the systems in place. However, this may not be a workable solution for organizations with strained resources or funding in the wake of the pandemic.

In the absence of better staffing for public services, institutions need to consider reference models that reduce the total time that archivists, librarians, and staff spend at public service points to allow overburdened employees the chance to attend to other demands while still assisting patrons when needed. Greater reliance on on-call systems (used by less than a third of survey respondents) offers one potential approach to triaging reference and freeing staff from service points when patrons are not present. Archives and special collections may be able to learn from academic and public libraries that have successfully transitioned to this type of service model. Implementing appointment booking systems could also help mitigate the amount of time spent at service points and allow staff to feel more prepared for research visits. Increased use of students and other types of paid, nonsalaried employees as front-line triage—with the appropriate training and librarians/staff available as backup—is another option not currently used by many survey respondents. Retooling or enlarging public services workstations to allow staff to complete other projects while working a reference shift might help minimize the feeling of “lost time,” especially when serving a shift with no researchers. Finally, closing service points until a researcher arrives may prove necessary for some institutions struggling to maintain reference staffing, but this approach could require additional justification to administrators worried about public perceptions.

Related future areas of inquiry on this subject include determining what kinds of resources departments need to improve their reference services (such as funding for new positions or implementation of staff training) and which skills are most valued by reference staff to provide high-quality services. A deeper look into reference service models outside archives is likely to reveal triage approaches that could reduce the workload burdens many survey participants reported. More data analysis of in-person reference transactions (such as the category of questions asked or the busiest times of the day or week) would allow departments to adjust staffing models and schedule employees accordingly. Additional surveys focusing on employee satisfaction—whether conducted informally on an institution-by-institution basis or conducted more broadly through national research studies—would bring light to specific areas of tension within organizations or possible solutions to increase employee morale. Further research about employee satisfaction with reference services might benefit from dialogue with the growing body of research on workplace burnout, morale, and engagement.³⁰

This project began with a desire to seek basic data about reference practices in archives and special collections that did not seem to exist elsewhere. Analyzing survey results and writing this article during a global pandemic with no knowledge of how long libraries and archives would be impacted added an extra layer of complication. It is clear, based on the author's own institution and their peers, that reference models in archives and special collections will continue to grow and change in response to both local and global concerns. Moving forward, evaluation of public services should strive to focus on ways to improve overall service to patrons, while also improving employee satisfaction and workload. As evidenced by the potential avenues for further research, there is much more to be studied about public services in the archives and special collections field. If reference work truly forms the core function of archives and special collections, then ongoing critical consideration of reference practices and their direct impact on employees deserves a greater degree of attention within our institutions and in the scholarly literature.