The first survey, in 2018, was designed to explore user perceptions of PSCC’s symbols with the addition of several new symbols for library collections and a new symbol for Education (L), after its original symbol was repurposed for the Curriculum Collection (Figure 2), and to test for recognizability and confusion. The survey included symbols and captions for the LCC classes and collections:

The 2018 survey first asked participants to match topic areas (e.g., subject heading or collection) with the best choice from the complete set of symbols. In order to collect data about all 29 of the initial symbols without overwhelming participants, five nearly identical surveys were constructed, each asking participants about only six of the symbols. Each survey also had two versions: one that included the LCC class letter for each symbol (e.g., N for Fine Arts, as shown in Figures 2 and 3), and one without class letters, to examine whether they were helpful or confusing, for a total of ten iterations of the survey. These questions were followed by three open-ended questions:

Is there anything that you like about any of these symbols, or the symbols as a set?

Focus groups during the spring of 2019, after the symbols were revised to address areas of confusion or concern raised by survey results, provided feedback on the second draft (see Figure 3). Questions for the focus groups were designed to illuminate preferences where multiple symbols were shown for a topic area. An additional goal was to gauge the success of revisions. In total, six questions were posed to each focus group:

While being shown the complete set of symbols (Figure 3), participants were asked:

What do you think these images are meant to be pictures of? This question was intended to identify those symbols that participants had trouble identifying, not merely connecting with a topic area.

Three rounds of in-person user testing of the signage system were conducted using the final draft of the symbol set (Figure 7). Round 1 gathered baseline data before the system was installed. Round 2 was conducted after the major components of the system were installed (Endcap Signs, Entrance Signs, and Floor Signs). Round 3 was conducted after the complete signage system was installed (Floor Maps, Elevator Signs, Elevator Maps, and Bump Point Signs).

The in-person testing employed methods similar to prior studies, including task scenarios, think-aloud protocols, and a post-task debriefing interview (Almeida & Tidal, 2022; Mandel, 2020; Mezick et al., 2015, 2016). Participants were timed as they completed two tasks:

Browsing a topic area: Participants were given a subject and asked to find its location in the library. Subjects were selected to test different physical areas in the library, including U.S. History (E), decorative arts (in the Fine Arts section, N), Political Science (J), botany (in the Science section, Q), and nursing (in the Medicine section, R).

Testers shadowed participants and took notes during each of the tasks. Participants all started in the second-floor lobby for the browsing task, and near the top of the stairs on the third floor for the locating task. They were also allowed to take advantage of any and all resources available, including signage, reference desk staff, and catalogue computers. A series of follow-up questions were asked after they performed the tasks:

Tell me about the path you took to locate each book and what you saw that helped you.

In early 2025, additional feedback was solicited from library users about the signage as part of a library-wide needs assessment. Information was collected via a short questionnaire and focused on the helpfulness of various design aspects of the different types of signage within the system. Samples of two main types of signage – floor signs and endcap signs – were shown separately and followed by several questions:

Do the pictures on these signs help you when you are trying to find something in the library?

Aspects unique to a specific type of signage, such as floor sign maps and lists of subject descriptions on endcap signs, were addressed in their respective sections. The questionnaire ended by asking participants to select aspects of the signage that they found most useful or would like to improve, such as titles and captions, pictures and symbols, colour coding, or maps, and finally by offering an opportunity to provide open-ended feedback.

There were 487 respondents to the ten iterations of the survey to gather feedback on the first set of symbols in 2018, with between 42 and 58 for each iteration (see Figure 2). Complete demographic data for survey responses and symbol matching are included in Appendices C and D. Overall, in terms of selecting the intended symbols to match topic areas, faculty and staff outperform students (Figure 7), with 39.7% of students selecting zero, one, or two incorrect matches, as opposed to 53% of faculty and 54.6% of staff. Older participants also outperformed younger ones (Figure 8), with 53.9% of those 26 years and older selecting zero to two incorrect matches, compared to 31.9% of 18- to 25-year-olds.

Success was higher among participants with English as a first language than for those for whom English was not a first language (Figure 9), with 41% of students and 55% of faculty and staff making zero to two incorrect matches, compared to 11% and 25%, respectively. Participants who reported neither visual challenges, such as colour-blindness, nor trouble with reading, incorrectly matched zero to two symbols 50% of the time, but those who reported visual challenges did so 45% of the time, while those who reported trouble with reading did so 38% of the time (Figure 10).

Library experience did not appear to impact success, with 49% of lower-use participants making zero to two incorrect matches, and 40% of those reporting frequent use or greater experience selected zero to two incorrect matches (Figure 11). Non-traditional and first generation students outperformed minority students and all other students, with 48.7% and 48.6% selecting zero to two incorrect matches, respectively, as compared to 20% and 31% respectively (Figure 12).

Besides examining who made the most incorrect matches, the team wanted to know which symbols were most frequently incorrectly matched with their topic, and which symbols were most frequently selected as incorrect matches for other topics. Blank answers were removed before calculating the percentage of incorrect matches. Some symbols, such as those for Medicine (R) and Agriculture (S), had low instances of being incorrectly matched and incorrectly selected. Others, such as Curriculum (CRM) and Archaeology & Biography (C), had fairly high rates in both areas. Some, however, were high in only one area or the other, such as Social Sciences (H), which was rarely selected as the correct symbol for its topic area, and General Works (A), which was often incorrectly selected as the match for other topic areas (Table 1).

In response to open-ended questions, participants shared concerns about the symbols, including both design choices and imagery. Symbols mentioned and causes for concern included:

Some comments referred to the overall aesthetic or the cohesiveness of the symbols as a set. The symbols were called unprofessional, “dated,” and compared to clip art; one participant noted that “some look like they come out of a baby toy book.” As a set, the lack of cohesion was noted – participants called out the varying line weights, colouring, drawing style, and levels of detail.

After the first draft of symbols was revised, four focus groups were convened in 2019 to obtain further insight into user perceptions of the revised symbols (see Figure 3), with a total of 30 participants in groups of eight, nine, eight again, and five people. Students (37%), faculty (20%), and staff (43%) participated (Appendix A). Responses to prompts and the pursuant conversations revealed symbols which were difficult to identify, interpret, or connect to their topic area. Symbols which had been revised to good effect were also revealed, including several that were no longer difficult to identify or connect to their topic area.

Images that were difficult to identify included a line drawing of the Stewart Library (for General Works, A), a pair of stylised hands with laurel (for Philosophy, Psychology, & Religion, B), and a figure connected to a stylised network (for Social Science, H). The symbols for Philosophy, Psychology, & Religion (B) and Social Sciences (H) were also among those that were difficult for participants to connect to their topic areas when captions were not provided. Other symbols in this area included a ziggurat for Archaeology & Biography (C), a map with a pin for Geography & Anthropology (G), and a set of books on a shelf for Typography & Information Science (Z). The Great Seal, for Government Documents (GOV), was often mistaken as being the symbol for Political Science (J).

In response to questions about which of several symbols they preferred for specific topic areas, participants were largely in agreement, and several that were difficult to identify or connect to topic areas were nevertheless preferred over alternatives:

Other preferred options reflected new and revised symbols that participants no longer had difficulty connecting to a topic area:

Using dog-tags as the symbol for Military Science (U) was another choice that mitigated confusion and difficulty. When asked for specific ideas for improvement, participants suggested keeping the General Works (A) symbol generic, making the Geography & Anthropology (G) symbol more like a map, adding details or lines to the globe and map for the World History (D) and United States History (E) symbols, adding a second figure to the Social Sciences (H) symbol and making it less like a computer chip and more like a genealogy or more “sciencey” or using a family group with radiating connections, and adding a squiggle to the Language & Literature (P) symbol to show that the quill pen is writing. Suggestions for new or revised symbols included, for Archaeology & Biography (C), using a person with a stone tablet, putting a person on the ziggurat, or a bust of King Tut, and a typewriter or printing press for Typography & Information Science (Z). It was also suggested, again, to improve consistency between symbols, such as line weight, style, and use of shadows.

After the second draft of symbols was revised, a beginning draft of the signage system was developed (including only Endcap Signs, Entrance Signs, and Floor Signs) and tested in 2020. The final draft of the signage system, with all its components, was developed between 2021 and 2023, and tested in 2023. There were four participants in Round 1 testing (baseline/no signage), three in Round 2 (partial system), and seven in Round 3 (full system). All participants were students (Appendix A), and most had some library experience (Appendix B). Specifically, all in the Round 1 (baseline) testing and two out of three in the Round 2 (partial) testing reported taking a library class, and two in each round also reported having a library session with another class as well. In Round 3 (final) testing, library experience was more limited: only two out of seven reported taking a library class, and two others reported having had a library session with another class; none had both.

In order to control for the difference in destination areas between task iterations and compare search time accurately and consistently between testing rounds, tasks, and participants, the team measured the time it took to walk from the starting point to the destination point for each task iteration, and that time was deducted from participants’ total time to calculate actual search time (Table 2).

Overall, average actual search time to browse (Task 1) improved from Round 1 to Round 3 and, conversely, average actual search time to locate a known item (Task 2) roughly doubled. Notably, during Round 3, 71.4% of participants (five out of seven) began the tasks by using the signage system: 57% immediately utilised the Entrance Signs, and 14% used the Floor Sign on the destination floor. All participants who started by examining the Entrance Signs proceeded to also utilise the Floor Sign on the destination floor, and all participants used the Endcap Signs to facilitate their success.

During exit interviews, participants in Round 1 noted that the lack of signage made their searching more difficult: One got stuck in a dead end, and two others mentioned that signs with subjects or directions to subject areas would have helped. Round 2 and 3 participants, however, indicated that by and large most of what they needed was provided and their confusion came from other sources, such as expecting books to be organised alphabetically. Several wished for hanging signs in addition to those on walls and endcaps, and one – who got stuck in one half of the divided Medicine (R) section – wished for a sign indicating that the section continued. All said their overall impression of the signs was positive, and indicated that the words and numbers on the signs helped them. All but one also indicated that the pictures were helpful. Very few noticed the colours until prompted by the question, but two said they did help, and one noted that, upon reflection, they helped.

There were 31 respondents to the needs assessment questionnaire, and the response was largely positive. Components that were highly rated as helpful included pictures and maps on floor signs, at 80.6% and 74.2%, respectively. On endcap signs, pictures and subject description lists were each rated helpful by 74.2% of participants. The lowest rate of helpfulness across all areas was 48.4% for the colour coding on floor signs and maps, and it was still considered more helpful than not helpful (0%) or neither (12.9%), while all others reported either not noticing or not using them. Similarly, the highest level of unhelpfulness across all areas was 6.5% for the subject captions on floor signs, but these were also considered more helpful than not, with 64.5% rating them positively (Table 3).

Open-ended responses reflected this as well, with 79.3% (23 of 29 who answered the questions) responding that the Floor Signs and Endcap Signs help them find what they are looking for when they visit the library – though it should be noted that the two who did not answer the question marked all components of the signs as “helpful” in other parts of the questionnaire. Most respondents discussed the helpfulness of the signs, and some explained that they hadn’t noticed or used them. Several mentioned using the signs primarily to locate study areas or classrooms, and three stated that they are helpful in general but not always for locating books. One respondent was particularly positive:

Similarly, more respondents indicated helpful aspects of the signs than areas for improvement, with over half noting that they like the signage the way it was (Figure 13).

Ultimately, participant age was the most predictive of participant ability to select the intended symbol for a given topic area in the survey. Those in the 18–21 and 22–25 age ranges underperform older participants, and faculty and staff tend to be older than students, while among students, non-traditional, and first-generation students tend to be older than the general student population. Only 23% of non-traditional students, 49% of first-generation students, and less than 1% of faculty and staff reported being under the age of 26. The notion that age – or, perhaps, life experience – was the most important characteristic for interpreting the symbols indicated that they lacked the general recognisability necessary to be useful on signage that needed to appeal to and assist users of all demographics.

To decide which symbols to redevelop, the team considered the percentage of incorrect matches between a topic area and its symbol, the frequency with which a symbol was selected as the incorrect match for other topic areas, and responses to open-ended questions. A high rate of incorrect matches was considered indicative of a symbol that was difficult to recognise or an inappropriate representation of the topic area. High frequency of selection as a mis-match for another topic areas was considered indicative that the symbol was again a poor representation of its topic area, vague, or otherwise confusing, or that a representation resonated with users better for a different topic area – in short, that participants didn’t know what it was supposed to be (see Table 1). User comments were also used to guide design decisions, particularly where symbols were offensive, triggering, or confusing.

This data-driven redesign resulted in a second draft of the symbol set (see Figure 3) with new symbols for 13 topic areas, including A, C, D, E, G, H, J, L, P, T, U, and Z, and MDA (Table 4). In addition, minor revisions such as changes to line weight or shading were made to three symbols for clarity or to increase consistency within the set, including Q (Science), S (Agriculture), and MDK (Media Kit Collection). Focus groups examined these new symbols to test design choices and disambiguation, and to expose new or persistent issues.

In some cases, symbols with high rates of misidentification or mismatching were not redesigned, instead relying on disambiguation through the redesign of other symbols that were frequently confused with that symbol or its topic area. Among these were B, Philosophy, Psychology, & Religion, which had a 56.1% rate of incorrect matches, but was infrequently selected as the incorrect match for other topics, primarily H, Social Sciences, and C, Archaeology & Biography, both of which were redesigned, which the team hoped would disambiguate B.

Another was F, History of the Americas, which had a 41.8% rate of incorrect matches and was frequently selected as the incorrect match for other topics, especially G, Geography & Anthropology, and C, Archaeology & Biography. However, E, United States History, was frequently selected as the incorrect match for this symbol, so rather than redesign F, the other two history symbols were redesigned to be more similar to this one, with the hope that they would create a useful and recognizable “history subset” within the symbol set, and disambiguate them from other topic areas.

CRM, Curriculum Collection, and CRK, Curriculum Kits, were two additional symbols with high rates of incorrect matches (69.7% and 51.9%, respectively). In this case, however, frequently selected incorrect matches primarily consisted of the symbol for L, Education, and each other’s symbol. Similarly, these symbols were only frequent selections for these same areas, though CRM was occasionally selected for A, General Works, and CRK was occasionally selected for OVS, the Oversize collection. With these things in mind, the team decided to redesign the symbols for L and A, largely removing the book imagery, to disambiguate them.

The symbol for OVS, the Oversize Collection, was borderline, with 40% incorrect matches but only 39 selections as the incorrect symbol for another area – primarily for A, General Works, C, Archaeology & Biography, H, Social Sciences, and P, Language & Literature, which were already selected for redesign, which the team hoped would disambiguate these symbols as well.

Data from the focus groups supported the further development and revision of the symbols for the final draft of the symbol set, which also included new symbols for the Reference, Periodical, and Map Collections (Figure 14), and which was used to develop the signage. This final symbol set is freely available, in individual files and multiple formats, as a collection in the Harvard Dataverse¹ for others to use, test, adopt, or adapt in their own visual signage systems and other projects.

Minor changes to symbols in direct response to participant comments included refining lines and adding detail to show countries and states in D, World History, E, United States History, and F, History of the Americas, which was also changed to show more of Canada and South America. On G, Geography & Anthropology, the map was given more shading and made to appear slightly folded to differentiate it further from the new symbol for the Map collection. The symbol for P, Language & Literature, was given an additional line to emphasize the writing function of the quill.

Other changes to increase consistency among the set, which was also suggested by the focus groups, included simplifying the symbol for K, Law, to help it match the rest of the set more, and giving the barn in the symbol for S, Agriculture, more shading. The symbol for V, Naval Sciences, was changed from a submarine to an anchor that better matched the style of the set and would be easier to identify. The symbol for B, Philosophy, Psychology, & Religion, was changed to match the style of the set better and make the laurel wreath surround the hands on both sides. Finally, the apple in the symbols for CRM and CRK, Curriculum and Curriculum Kits, were made smaller and placed on top of the books instead of behind them, and the labels were removed from the spines of the books for Z, Typography & Information Science. The two remaining symbols from the original set were also changed to better match the final set as a whole as well as for the copyright considerations mentioned earlier: the symbol for M, Music, was changed from a stylised clarinet-treble clef combination to a treble clef on a staff, and the symbol for N, Fine Arts, was changed from a paint palette with two brushes through it to a palette with three brushes next to it.

More extensive changes were made to the symbols for A, C, H, Q, and R to respond to focus group results and participant comments (Table 5).

No changes were made to the other symbols (J, L, T, U, GOV, MCF, MDA, MDK, OVS, and YPL).

When the design process for the signage began with the final set of symbols to represent subject areas, the team was unsatisfied with the library’s existing signage, a shared a drive to make improvements, and a dedication to universal design. With the graphic designer on the team, the team created the new signage system with several specific goals in mind. First, the signs within the system should be cohesive and recognizable, using symbols, colours, fonts, language, and other design elements with consistency. Second, the system and each sign within it should rank high in usability and meet user needs clearly and functionally, following signage best practices as much as possible. Third, the system should be sustainable: easily iterated, modified, corrected, updated, or repaired to keep pace with growth and long-term changes as well as meet short-term needs. Finally, the signage must be accessible and follow seven Principles of Universal Design as much as possible, encompassing equitable use, flexibility in use, simple and intuitive use, perceptible information, tolerance for error, low physical effort, and size and space for appropriate use (Center for Universal Design, 1997).

Like the Principles of Universal Design, Universal Design for Learning (UDL) is organised around key principles: engagement, representation, and action and expression. The Center for Applied Special Technology (CAST) created the UDL framework “to improve and optimise teaching and learning for all people based scientific insights into how humans learn” (CAST, 2024). This includes optimizing access to information in learning spaces such as libraries for people of all abilities by providing effective signage for navigation and wayfinding. CAST explains, “UDL is based on principles that empower everyone to have agency over their own learning… like fingerprints, no two brains are alike” (2025). This variety supports the need for multiple modes of engagement with wayfinding within the library stacks. Providing points of access via symbols, verbal descriptions, and LCC call numbers empowers users to choose wayfinding methods they prefer, and creates a more accessible and equitable library experience.

A visual signage system not only supports inclusive wayfinding, but also minimises the margin for error while navigating the stacks. For example, browsing can be frustrating if one wishes to find a subject area but does not know what classification range it falls under, but understanding the symbol used to represent the desired subject area minimises the potential for error, yet requires very little physical effort to engage with or understand. Additionally, in developing signs that fit spaces appropriately and can support those experiencing colour vision deficiency, those for whom the primary local language is not their first, those with library anxiety, or other challenges, these elements also meet universal design and UDL guidelines. Together the symbols, colours, captions, and other visual elements create a visual language which reduces barriers and is more inclusive than any single method.

With these things in mind, the colour palette for the signage was selected with care and revised when testing showed problems with contrast. The final colour palette was selected to include and compliment the official Weber State University’s Wildcat Purple: blue, burgundy, lime green, aqua, and teal (Weber State University, n.d.a). Both the original and final colour palettes were tested in Matthew Wickline’s online colour vision deficiency simulator, Coblis (2001), to be as sure as possible that users who experience common vision deficiencies such as Protanopia or Deuteranopia (red-green colour-blindness) would be able to take advantage of the colour coding. The simulator ultimately prompted the change to the final colour palette (Figure 15).

Fonts, wayfinding symbols, and other design elements were also selected with UDL in mind. The designer selected sans-serif fonts for all the signage, and chose common, simple pointers, arrows, and logos to represent directional information, stairs, restrooms, elevators, exits, and information desks. The designs intentionally minimised text, and where text was used, it was provided in English and Spanish. Similarly, to maximise inclusivity, signs and signage elements were designed to be a reasonable size for viewing up close or from afar, and printed signs were placed at a height to enable any user to read them easily.

To support the maintenance of the signage system, including the ability to iterate and update as needed to improve accessibility or keep up with collection changes, the signs were printed in-house on the library’s large-format printer. Materials, sign-holders, and other materials for the production and display of the signage system were also selected with maintenance in mind so that, if an item became unavailable from a particular vendor or in a particular design, they would not be irreplaceable or upset the overall function or look of the signage system.

During the design of the signage system, three key types of signs were planned:

As part of a complete system, each sign can rely on the others for context, allowing for the minimalistic design while making it easier for users to parse new information and combine it with existing knowledge to more easily navigate the library, locate known items, and discover new resources. Using a complete signage system in lieu of individual signs also allows the signage to go beyond facilitating wayfinding and potentially become instructional. For example, including collection names on Entrance Signs, Floor Signs, and Elevator Signs helps users understand the range of content available in the library, with which many users are not familiar. Including subject captions in addition to LCC class letters (e.g., Medicine in addition to “R”) gives users insight into the classification system and call numbers, as well as a better idea of the academic subject areas available for their use.

Ideally, the system will guide a user from the door to the floor they need, and then to the specific stack, with help along the way regardless of the route they take. Figure 16 shows the placement of signs throughout the Stewart Library when the full system is implemented:

Navigation Signs placed at the entrances and main stairways (green stars) as well as in elevators (blue stars) give users an overview and floor-level destination.

The results from the User Testing are, on the surface, inconclusive about the efficacy of the signage. Overall, performance improved with signage for browsing and diminished when finding an item on the shelf, but the partial system (Round 2) times skew these results, being both much faster than both the Round 1 and Round 3 times for browsing, and much slower for finding an item on the shelf. This pattern of results remained consistent when controlling for library experience, such as a library class or one-shot session or frequency of library use, or removing outliers changes the pattern of results. Controlling for age or minority status also made virtually no change in the results, which is not surprising with so few participants, especially in Round 2.

When comparing Round 1 and Round 3, a few things stand out. Participants in Round 1, who took an average of 7:12 minutes to discover the area to browse, and 1:26 minutes to locate a book on the shelf, all had experience with a credit-bearing library class, and half also had other library sessions. Among Round 3 participants, only two of seven reported taking a library class and another two a library session; these participants took an average of 3:02 minutes to discover the area to browse, and 3:22 minutes to locate a book. In addition, only three out of four participants in Round 1 (75%) were successful in the browsing task, but all seven in Round 3 were successful (100%). Removing the unsuccessful participant’s time from the Round 1 calculation only slightly changes the average search time to browse, increasing it from 7:12 to 7:21. This is noteworthy because library experience apparently did not help users with browsing when there was not signage, but lack of library experience also apparently did not hinder users when browsing or locating a book on the shelf when there was signage.

In short, based on user testing results alone, access to the complete signage system may increase success rates for browsing, as well as times. The complete signage system may also have helped inexperienced users in Round 3 to locate a book on the shelf once they became familiar with the system during the browsing task, even if it didn’t improve their time. Regardless of their level of experience, all Round 3 participants were able to complete their tasks in five minutes or less, with the exception of one who nevertheless persisted and succeeded. In short, experienced users located books easily, and others succeeded without undue difficultly in spite of inexperience.

The Needs Assessment feedback reinforces this: after the system was in place for several years, students rate signage components as being helpful far more often than not, demonstrating that the design choices and commitment to UDL are supporting students in the library. Open-ended comments indicate that the system helps students navigate the library to find books, areas to browse, study areas, and classrooms. In addition to the extremely grateful comment noted above, other comments called the signs “very easy to read, understand and locate,” “very helpful” and “very informative.” One student said, “Easy to follow and I like the colors [sic],” and another said specifically that “they help you navigate the library.”

In response to the Endcap Signs, students had similarly positive comments. One stated, “I like being able to more easily narrow down my search and the signs help with that,” and another echoed them: “These give me some good examples of the books in a certain area so when I’m searching for something on a certain topic they give me a good idea of where to search.” One called out specific features, saying, “the words and pictures are especially helpful.” Overall, the comments demonstrate that the signage system is helpful to many students, but particularly to some who lack experience, face other barriers, or are hesitant to approach staff. Poblocki (2007, p. 18) says, “Soliciting input in the design process for the most high-profile signs can foster good will with the student body and a stronger sense of connection to the campus,” and one of the final comments shows the truth in this: “You guys do a pretty good job.”