The role of space attributes in space-choice behaviour and satisfaction in an academic library

Abstract

Space attributes concern both users’ space-choice behaviour (i.e. where to sit) and satisfaction (i.e. level of satisfaction in the chosen seat). However, the effects of space attributes in academic libraries have most often been investigated with a separate focus on either space-choice behaviour or user satisfaction. Since library contexts vary greatly, it is unreasonable to compare findings from existing separate-focus studies. The present study thus aims to conjointly investigate the effects of space attributes on space-choice behaviour and user satisfaction from the same context of an academic library. For this purpose, a questionnaire survey was conducted and the data were explored in a series of statistical analyses. As a result, the study demonstrates that the influence of some space attributes differs between space choice and satisfaction. Once generalized in future studies of a wider variety of academic libraries, these findings should help architects to make better-informed decisions in library space-planning and design, leading to more efficient use of space and enhanced user satisfaction.

Keywords

Academic library satisfaction space attributes space-choice behaviour space-planning

Introduction

Space planning and design (SPD) seeks to optimize space utilization and to maximize satisfaction (i.e. the level of satisfaction with the space where building-users choose to sit down and work). Accordingly, during the SPD of academic libraries it is essential for architects to predict which library spaces will be selected by users and how satisfied, whether explicitly or implicitly, users will be as a result of their use of those spaces. At present, the architect’s professional judgment still plays an important role in SPD for small projects, where there is no guideline to which reference can be made (Cha et al., 2017). When guidelines have been generated on the basis of similar previous projects, post-occupancy evaluations, surveys, or architects’ cumulative expertise (Erhan 2003; Stanford University 2009), architects consult guidelines in order to determine the number and size of certain types of space under consideration for the user’s requirements. For example, the campus space-planning guidelines from Texas State University (Texas State University, 2018) suggest 25ft² as the recommended net assignable area per person for a reading room in an academic library. Although architects’ expertise and guidelines can thus assist in securing efficient space configurations for library services, this approach pays less attention to ways in which library-users choose spaces (i.e. where to sit) and ways in which library spaces influence user satisfaction.

Spaces have multiple attributes, such as furnishing, windows and lighting, so that users’ space choice and satisfaction with their use of space can be affected by a single change in any of these attributes. Without a clear understanding of the effects of space attributes, therefore, architects may be unable to come up with well-planned and well-designed library spaces for user satisfaction (Fulton, 1991; Hassanain, 2010). In this context, a number of studies have examined space choice or user satisfaction in academic libraries by asking about choice of seat location or satisfaction level by the use of the space or by a library building in its entirety. For example, Fox and Doshi (2013) measured satisfaction levels with common space in a library from four perspectives, i.e. power, lighting, aesthetics/furniture/ambience and “defined yet open” space. Holder and Lange (2014) showed different types of furnishing to users and asked them to rank the top three furnishing choices for quiet study, for group work, and for lounging and relaxation. Ojennus and Watts (2017) surveyed users’ space choice for individual study by asking what type of individual study space the library would need in greater numbers. May and Swabey (2015) conducted a multi-site study in an attempt to enhance the generalizability of findings, in which they asked how users choose a seat and about their satisfaction with the library for individual and group work. Despite the contributions made by those studies in understanding library-users’ space choice and satisfaction, there are limitations in their comparisons between the effects of space attributes in relation to space choice and satisfaction.

For this reason another line of investigation has focused more on the effect of each space attribute, in order to cast light on the relationship between space attributes and space choice or user satisfaction. As for the space attributes/space choice relationship, Cha and Kim (2015) conducted a survey in an academic library in which they identified the amount of space, noise level, crowdedness, comfort of furnishing and cleanliness as the top five space attributes to influence users’ space choice. In Brown-Sica’s (2012) action research, users of three libraries were asked about the best place to meet for group study and why. Webb et al. (2008) also sought to determine the relationship between space attributes and library-users’ space choice. They found that main space attributes for study-floor selection include furniture, a window view, computer or wireless access, the floor with the books required and friends who were studying on that floor. Weessies (2011) and Yoo-Lee et al. (2013) also studied space attributes affecting space choices, in this case for computer use or for study. Similarly, to understand the space attribute/user satisfaction relationship, Montgomery (2014) adopted the research question “What is it about this space that works well for you?” and conducted surveys in two consecutive years to track changes in the responses. Halling and Carrigan (2012) studied which academic library space attribute gave users least satisfaction and what space changes they requested in consequence. De Clercq and Cranz (2014) applied multiple data collection methods, such as covert observation, interviews and questionnaires, and found that adjustability, the comfort of furnishings, and spaces located near windows and along the periphery of the building were important space attributes for user satisfaction. These studies contribute to an understanding of the features of space that affect library-users’ space choice and satisfaction by their use of space. Quietness, furnishings and the availability of computers and space are generally-accepted space attributes which affect space choice and user satisfaction. Thus, in order to enhance space use and user satisfaction, architects can pay greater attention to those attributes when they design library buildings.

Nevertheless, the question of whether or not the space attributes that influence space choice in academic libraries are identical to those that influence user satisfaction has not been addressed by the existing literature. It is important to answer this question because, if some space attributes exert greater influence upon users’ space choices than upon their satisfaction and vice versa, architects could make use of this information during the SPD of academic libraries. For example, if there is a space with which readers are highly satisfied but which is infrequently used, then architects need to enhance the attributes known to be more important for space choice than for satisfaction. Unfortunately, findings to date on space attributes cannot simply be combined and used in comparisons because the importance of space attributes in space choice and user satisfaction varies according to university culture, library location, and user and activity profiles (Cha and Kim, 2015; May and Swabey 2015). Indeed, because the studies to date have been conducted in different types of university (e.g. research universities and small liberal arts colleges), in differently-balanced populations of undergraduate and graduate students, in different academic years and in different sizes of library building, there are some discrepancies between their findings.

For example, when Montgomery (2014) asked the same question when assessing library-users’ satisfaction levels in two consecutive years (i.e. 2012 and 2013), it was found that the responses were not identical. Moreover, while Webb et al. (2008) found that a window view is one of the main reasons for space choice in a library, in Weessies’ (2011) study this attribute was found to have a very low correlation with space choice. Therefore, to overcome the challenge, data on space attributes that affect users’ space choice and satisfaction levels must be collected on the same occasion if they are to permit such comparisons. Although May and Swabey (2015) studied both space choice and user satisfaction in a single multi-site study, they did not rank the importance of space attributes and compare their influence on space choice and user satisfaction. Instead, their study surveyed factors of space choice in seat level and factors of satisfaction at the level of the entire library building. It is thus difficult to compare the importance of space attributes upon space choices and user satisfaction on the same basis.

The present study aims, therefore, to answer to the following research question: Are space attributes affecting space choice different from space attributes affecting user satisfaction in academic libraries? For this purpose, a self-completion questionnaire was distributed to collect data on both space choice and satisfaction in an academic library. Respondents who were staying at and/or working in a space were asked to assess each attribute separately with respect to its importance for space choice and level of satisfaction. The data were then explored in a series of statistical analyses, to compare the importance of different space attributes on space choices and user satisfaction. It is hoped that the study will help architects to recognize different levels in the importance of space attributes before deciding whether a design needs to emphasize any particular attribute. By understanding the two space-attribute dimensions, architects can make better-informed decisions about effective SPD, and thereby optimize space utilization while enhancing user satisfaction.

Methods

Questionnaire design

The paper-based questionnaire had two parts. In the first part, participants responded to questions about 11 contextual factors, classified into three categories: demographic, activity type and additional considerations, to investigate effects of these library contexts on the importance of space attributes in terms of space-choice behaviour and satisfaction. The demographic information related to age, gender, ethnicity and occupation. Following Crumpton and Crowe (2009) and De Clercq and Cranz (2014), questions about the activity type addressed the four most frequent activities that users pursue in academic libraries: reading books or articles, working on assignments independently, meeting other people and studying online. The three additional questions specific to the survey asked whether respondents worked on weekdays or weekends, whether they were alone or accompanied by friends and how long they expected to remain in the building.

The second part had two subsections: (1) an enquiry into the importance of each space attribute when the respondent made space choices (i.e. its perceived importance), and (2) an enquiry into respondents’ satisfaction with each space attribute and their overall level of satisfaction in the chosen space. This part was designed based on the process of selecting and then using a space. When library-users make space choices, they first consider some space attributes to meet functional requirements of their activities. It is thus usual that those space attributes are likely to be satisfactory during the use of the selected space. The first section of this part is to identify those space attributes. After those space attributes have been considered, the second section of this part is to identify how influential each space attribute is on overall satisfaction. This part incorporated 18 space attributes, shown in Table 1, adopted from POE studies conducted at the Centre for the Built Environment (CBE) at the University of California Berkeley ( Frontczak et al., 2012; Lee and Guerin, 2009; Zagreus et al., 2004), together with studies of space usage in libraries (Applegate, 2009; Crumpton and Crowe, 2009; Webb et al., 2008;Weessies, 2011). The assessment of the 18 space attributes employed a seven-point Likert-type scale (i.e. for perceived importance, from 1: not considered at all to 7: strongly considered; and for satisfaction, from 1: very unsatisfied to 7: very satisfied).

Table 1.

A list of space attributes.

Space attributes	Access to facilities
	Adjustability to physical conditions
	Aesthetic appearance
	Air movement
	Air quality
	Amount of light
	Amount of space
	Cleanliness
	Comfort of furnishing
	Crowdedness
	Distance from an entrance
	Ease of interaction with co-workers or friends
	Noise level
	Sound privacy
	Temperature
	Visual comfort
	Visual privacy
	Window view

Location

The survey took place in an academic library occupying three storeys (basement, ground and first floors) of the MetaForum building located in the middle of the University campus at the Eindhoven University of Technology (TU/e), Netherlands. The basement is reserved for quiet study, while conversation is permitted on the above-ground floors. The ground and first floors have fixed curtain walls, giving respondents a view of the outside world. Heating, ventilation and air conditioning (HVAC) systems are centrally operated, with no user control over ambient conditions. There are about 950 seats, and all individual seats are adjustable. The building is furnished with different types of sofa, cabins for individual study, rooms for group study and approximately 50 seats at shared computers. Students and staff members can pursue diverse activities (e.g. meeting or reading books) with their own choice of seat in the building. The library’s diversity of seating makes it suitable for the purpose of study.

Data collection

We first undertook a pilot study (n=30) to test the consistency and clarity of the questionnaire. Then we conducted a questionnaire survey between 13:00 and 16:00 hours for two weeks in 2013. We approached individuals who had chosen a space in the building for working/studying and then distributed questionnaires to those who had agreed to take part in the survey. In all, 350 questionnaires were distributed, and 308 were collected. After excluding incomplete questionnaires, 268 valid questionnaires were used for analysis. The demographic information is summarized in Table 2.

Table 2.

Demographic information.

	Contextual factors	Description
User type	Age	<25 years		25–30 years		>30 years
		71%		25%		4%
	Gender	Male		Female
		78%		22%
	Ethnicity	European		Asian		Others
		81%		16%		3%
	Occupation	Undergraduate		Graduate		Others
		53%		42%		5%
Activity type	Reading books or articles	35%
	Working on assignments alone	66%
	Meeting	22%
	Online study	46%
Additional considerations	Weekday/Weekend	Weekday		Weekend
		63%		37%
	Accompanying friend(s)	With friend(s)		Alone
		52%		48%
	Expected length of stay	Less than 1 hour	1–3 hours	3–5 hours	5–7 hours	More than 7 hours
		2%	42%	35%	15%	6%

Because the survey was conducted in an academic library, nearly all respondents were either undergraduate (53%) or graduate (42%) students. Most were aged under 25 years. Males and Europeans accounted for around 80%, in terms of gender and ethnicity respectively. With regard to the activity type, respondents could select multiple activities if they wished. Two-thirds of respondents were visiting the building simply to work on assignments, while fewer than one in four were visiting for group meetings. The remaining two activities were reading books or articles (35%) and online study (46%), which were assessed separately. In response to the supplementary questions, approximately two-thirds of the respondents reported that they worked on weekdays, while the remaining third preferred to work over the weekends. Slightly more than half of the respondents visited the building unaccompanied, while the others were accompanied by friends. Lastly, two out of five respondents usually stayed for 1 to 3 hours per visit.

Data analysis

There were four steps in the analysis, as shown in Figure 1. At the first step, direct ratings (a direct method) were employed to assess the ranking of space attributes in terms of perceived importance. In addition, regression analyses (an indirect method) assessed the ranking of space attributes in terms of satisfaction. Overall satisfaction was the dependent variable, and the levels of satisfaction with each space attribute were the independent variables. As a rule, the methods for identifying important attributes are divided into two groups: direct or ‘subjective’ methods, and indirect or ‘objective’ methods (Alpert, 1971). In direct methods, the respondents are asked to give reasons for their choice, and highly-ranked or highly-rated attributes are classified as key attributes. The underlying assumption of direct methods is that the respondents know which attributes are more important than others. By contrast, indirect methods do not endorse that assumption. They employ statistical techniques such as regression analysis in order to make an objective evaluation of attribute importance on a dependent variable (Gustafsson, 2004). Thus, direct ratings are suitable for understanding the importance of an attribute in a fixed-choice situation (Van Ittersum et al., 2007). Moreover, in order to understand attributes that influence overall satisfaction as a dependent variable many POE studies have performed regression analyses. The present study thus employed both direct ratings and regression analyses for identifying general important space attributes in space choice and satisfaction respectively. At the second step, 18 space attributes, according to the two dimensions, were investigated in separate factor analyses (i.e. factor analysis P for perceived importance and factor analysis S for satisfaction) so as to divide the attributes into groups with commonalities. At the third and fourth steps, in order to identify the influence of contextual factors on perceived importance and satisfaction, t-tests were performed on the attribute groups of ‘perceived importance’, and dummy variable regressions were performed on the attribute groups of ‘satisfaction’.

Figure 1.

Survey data analysis framework.

Results and discussion

Rankings of space attributes: Perceived importance and satisfaction

All 18 space attributes for perceived importance and satisfaction were analysed using direct ratings and regression analyses respectively (Table 3). With regard to perceived importance, the five most important attributes (mean value >5.0) were: ‘Amount of space’, ‘Noise level’, ‘Crowdedness’, ‘Comfort of furnishing’ and ‘Cleanliness’. The overall satisfaction levels in the chosen space were regressed onto satisfaction with the 18 space attributes. Multiple regression showed the 18 space attributes to account for 42% of overall satisfaction (R² = .42, p<.0001). Space attributes with a statistically significant prediction of overall satisfaction (i.e. those with p-values <.1) were, in order of importance, ‘Visual comfort’, ‘Noise level’, ‘Window view’ and ‘Aesthetic appearance’.

Table 3.

Rankings of space attributes for perceived importance and satisfaction.

Space attributes	Perceived importance			Satisfaction
Space attributes	Mean	Std. Dev	Rank	Beta	Rank
Accessibility to facilities	3.83	1.79	10	0.099	6
Adjustability to physical conditions	3.58	1.75	14	−0.052	13
Aesthetic appearance	3.4	1.78	18	0.101*	4
Air movement	3.5	1.79	16	−0.013	17
Air quality	3.72	1.87	12	0.042	16
Amount of light	4.83	1.65	6	0.01	18
Amount of space	5.68	1.18	1	0.09	7
Cleanliness	5.17	1.37	5	0.044	14
Comfort of furnishing	5.32	1.45	4	0.074	9
Crowdedness	5.4	1.35	3	0.098	5
Distance from an entrance	3.51	1.79	15	−0.044	15
Ease of interaction	4.3	1.84	7	0.071	10
Noise level	5.6	1.65	2	0.165**	2
Sound privacy	3.67	1.87	13	−0.058	12
Temperature	3.76	1.87	11	0.064	11
Visual comfort	4.2	1.77	8	0.194**	1
Visual privacy	3.91	1.8	9	0.086	8
Window view	3.49	1.88	17	0.131**	3

*, **

statistically significant at the 0.1, 0.05 levels respectively.

Eight space attributes with approximately the same rankings between the two dimensions of perceived importance and satisfaction were identified: ‘Adjustability to physical conditions’, ‘Air movement’, ‘Crowdedness’, ‘Distance from an entrance’, ‘Noise level’, ‘Sound privacy’, ‘Temperature’ and ‘Visual privacy’. Of these attributes, ‘Crowdedness’ and ‘Noise level’ were ranked in the top five on both dimensions. The findings indicate that these space attributes are clearly salient for users when they make space choices. They are also highly influential on overall satisfaction. The results seem natural as many of students visited the library for academic purposes.

By contrast, ‘Adjustability to physical conditions’, ‘Air movement’ and ‘Distance from an entrance’ have low rankings on both dimensions, indicating that they have less influence on overall space satisfaction and are less important in guiding space-choice behaviour. The low rankings of ‘Air movement’ and ‘Adjustability to physical conditions’ for both dimensions may occur because the library has a centrally-operated HVAC system with no user control over ambient conditions. This provides constant heating and ventilation across rooms and/or spaces, so that users find little difference when making space choices and while staying there. ‘Distance from an entrance’ also ranked low, suggesting that the distance was not perceived as a strong determinant of space choice. This contrasts with a previous finding that distance is generally correlated with choice of computer workstation (Wessies, 2011). However, Wessies (2011) also acknowledged that many students came to use a computer near the entrance only for a quick print-out, which made the space near the entrance more attractive, while the two most remote spaces in the library were surprisingly popular. It thus appeared that students were willing to travel farther to find a better place. In this regard it seems fair that ‘Distance from an entrance’ was not perceived important for space choice, since the study found that only 2% of users stayed for less than an hour.

On the other hand, space attributes with wide differences in ranking between the two dimensions are ‘Aesthetic appearance’, ‘Amount of light’, ‘Amount of space’, ‘Cleanliness’, ‘Visual comfort’ and ‘Window views’. While ‘Amount of light’, ‘Amount of space’ and ‘Cleanliness’ were not influential on overall satisfaction, they had high rankings in perceived importance, indicating that they are important in guiding space-choice behaviour. While ‘Aesthetic appearance’, ‘Visual comfort’ and ‘Window views’ were ranked in the top five in satisfaction, they nevertheless had the lowest rankings in perceived importance, indicating that they are less important in guiding space-choice behaviour. The findings about space attributes with wide differences in ranking between the two dimensions suggest that when users make space choices they first consider whether or not the attribute conditions are suitable for their activities. Therefore, the attribute conditions considered for space choices are likely to be good enough for their activities, so that this may be a reason why they are not an influential factor for overall satisfaction. In other words, they first determine whether the attribute conditions can sufficiently support the requirements of their activities, and they do this before they consider satisfaction. Since ‘Amount of light’ was ranked sixth in importance but 18th in satisfaction, it suggests that users first consider whether the space has a sufficient lighting level for their activity when making their choice, but that their overall satisfaction is less influenced by lighting level. The level of satisfaction may be more relevant to ‘Visual comfort’, which had the greatest influence upon overall satisfaction. Indeed, many lighting effects can cause visual discomfort through glare, reflections and compromised contrast (Winterbottom and Wilkins, 2009), resulting in lower overall satisfaction. In short, users may first consider whether a space has sufficient lighting for their activities and then require visual comfort to enhance their performance (such as when working on assignments or pursuing online study).

The above-mentioned finding on space-choice behaviour, when users first determine whether attribute conditions can sufficiently support their activities before they turn their minds to satisfaction, can be also explain why ‘Cleanliness’ and ‘Amount of space’ have a low influence on overall satisfaction but yet have high rankings in terms of perceived importance. In addition, although the attributes of ‘Aesthetic appearance’ and ‘Window views’ had the lowest rankings for perceived importance, they were ranked high in overall satisfaction (at third and fourth), thus replicating findings from many POE studies ( DeClercq and Cranz, 2014; Organ and Jantti, 1997). This might be because users first consider important space attributes, such as noise level and lighting level, in order to ensure that the space meets the requirements for their activities. Their satisfaction level in the chosen space is then related more to ‘good views’ (i.e. ‘Aesthetic appearance’ and ‘Window views’), which enhance positive emotions (Ashkanasy et al., 2004).

Attribute groups for perceived importance and satisfaction

At this step, factor analysis was performed to examine the commonalities of space attributes in the individual dimension and then identify similarities and differences between attribute groups for perceived importance and for satisfaction. The space attributes with similar pattern for perceived importance and satisfaction were then sorted into attribute groups. Factor analysis P derived four attribute groups for perceived importance as shown in Table 4. Attribute group PA, comprising ‘Air movement’, ‘Air quality’, ‘Amount of light’, ‘Temperature’ and ‘Visual comfort’, represents dynamic attributes for air, temperature, lighting and visual comfort. This group was named ‘Indoor climate & visual environment’. Attribute group PB, comprising ‘Accessibility to facilities’, ‘Adjustability to physical conditions’, ‘Aesthetic appearance’, ‘Comfort of furnishing’, ‘Distance from an entrance’, and ‘Window view’, identified respondents’ perceptions from the perspectives of ‘Accessibility & furnishing’. Attribute group PC, comprising ‘Amount of space’, ‘Cleanliness’, ‘Crowdedness’ and ‘Noise level’, grouped attributes related to ‘Space congestion’. Attribute group PD, comprising ‘Ease of interaction’ and ‘Sound privacy’, was conceptualized as ‘Conversation, privacy & interaction’. ‘Visual privacy’ had no commonality with any of groups.

Table 4.

Attribute groups from factor analysis P.

Attribute groups	Attributes
Indoor climate &visual environment (PA)	Air movement
	Air quality
	Temperature
	Visual comfort
	Amount of light
Accessibility & furnishing(PB)	Accessibility to facilities
	Distance from an entrance
	Adjustability to physical conditions
	Aesthetic appearance
	Comfort of furnishing
	Window view
Space congestion (PC)	Crowdedness
	Amount of space
	Noise level
	Cleanliness
Conversation, privacy &Interaction (PD)	Ease of interaction
Conversation, privacy &Interaction (PD)	Sound privacy

Factor analysis S derived six attribute groups for satisfaction as shown in Table 5. Attribute group SA, comprising ‘Air movement’, ‘Air quality’ and ‘Temperature’, and attribute group SC, comprising ‘Amount of light’ and ‘Visual comfort’, resembled the attribute group PA (Indoor climate & visual environment) in terms of perceived importance. Attribute groups SA and SC were thus named ‘Indoor climate’ and ‘Visual environment’ respectively. Attribute group SB, comprising ‘Amount of space’, ‘Crowdedness’, ‘Noise level’ and ‘Sound privacy’, resembled the attribute group PC (Space congestion) shown in Table 4. Thus, the attribute group SB was named ‘Space congestion’. Attribute group SD, comprising ‘Accessibility to facilities’, ‘Aesthetic appearance’, ‘Cleanliness’, and ‘Distance from an entrance’, and attribute group SF, comprising ‘Adjustability to physical conditions’ and ‘Comfort of furnishing’, resembled the attribute group PB (Accessibility & furnishing) shown in Table 4. The attribute groups SD and SF were therefore named ‘Appearance and accessibility’ and ‘Furnishing’ respectively. The final attribute group SE, comprising ‘Ease of interaction’, ‘Visual privacy’ and ‘Window view’, comprised attributes related to ‘Visual privacy & interaction’. These attribute groups can be adopted for further analysis with contextual factors, in order to understand the contextual factors influencing space-choice behaviour.

Table 5.

Attribute groups from factor analysis S.

Attribute groups	Attributes
Indoor climate (SA)	Air movement
	Air quality
	Temperature
Space congestion (SB)	Noise level
	Sound privacy
	Amount of space
	Crowdedness
Visual environment (SC)	Amount of light
Visual environment (SC)	Visual comfort
Appearance and accessibility(SD)	Distance from an entrance
	Accessibility to facilities
	cleanliness
	Aesthetic appearance
Visual privacy & interaction(SE)	Ease of interaction
	Window view
	Visual privacy
Furnishing (SF)	Adjustability to physical conditions
Furnishing (SF)	Comfort of furnishing

To sum up, four and six attribute groups were identified for perceived importance and users’ satisfaction respectively. The commonality between the grouped space attributes indicates that they have similar patterns for each dimension. This implies that, once a grouped attribute has been promoted, the effect of the group can be leveraged for making space choices and delivering overall satisfaction. This is an important finding, in that in some cases certain space attributes cannot be modified on account of space constraints or policies, although architects know that those attributes need to be enhanced. For example, let us imagine a space at some distance from the entrances, with the consequence that it is unpopular with library-users. It might become more popular if we could modify other space attributes in the same group, such as ‘Adjustability to physical conditions’ or ‘Comfort of furnishing’, without changing the distance from the entrances, which might be impracticable. To take another example, a library building is likely to have many windowless spaces unless all of the seats are placed along the building’s perimeter. Since several studies (e.g. De Clercq and Cranz, 2014) have found that proximity to a window is an important component of satisfaction, windowless spaces may have a negative effect upon satisfaction levels. To complement such negative effects, architects can enhance the space attributes in the attribute group SE, such as ‘Ease of interaction’ or ‘Visual privacy’, either by installing partitions to enhance visual privacy or by allocating collaborative spaces with a high ease of interaction. In this way, attribute-groups such as these can help architects to understand the different facets of users’ space-choice behaviour and satisfaction.

Influential contextual factors in perceived importance

To identify influential contextual factors in perceived importance, the attribute groups obtained from factor analysis P were employed for t-tests. Contextual factors mostly comprised three categories (i.e. user type, activity type and additional considerations). Each contextual factor had varying effects upon perceived importance of attribute groups, as shown in Table 6.

Table 6.

Result of t-tests: Identification of influential contextual factors in perceived importance of attribute groups.

USER TYPE
Attribute group	Gender(M/F)Higher importance for		Age(18–24/25–30)Higher importance for		Nationality(European/Asian)Higher importance for		Undergraduate/GraduateHigher importance for
PA	Female	**	25–30	****	Asian	*	Graduate	**
PB	N/A		25–30	****	Asian	***	Graduate	**
PC	N/A		25–30	**	N/A		N/A
PD	Female	**	25–30	*	Asian	*	N/A
ACTIVITY TYPE
Attribute group	ReadingHigher importance for		MeetingHigher importance for		OnlineHigher importance for		Working aloneHigher importance for
PA	N/A		N/A		No	***	Yes	**
PB	N/A		N/A		No	**	N/A
PC	Yes	**	No	**	No	**	N/A	**
PD	N/A		Yes	***	N/A		No
ADDITIONAL CONSIDERATIONS
Attributegroup	Weekday/WeekendHigher importance for			With friends or notHigher importance for			Expected length of stayHigher importance for
PA	Weekend	*		N/A			longer time	**
PB	Weekend	****		N/A			longer time	***
PC	Weekend	*		N/A			N/A
PD	Weekend	*		with friends	****		longer time	*

**, ***, **** statistically significant at the 0.1, 0.05, 0.01, 0.001 levels, respectively.

As for demographics, gender was influential in perceived importance of the PA (Indoor climate & visual environment) and PD (Conversation privacy & interaction) attribute groups, which are more important for women. With respect to age, only two dominant groups, 18–24 and 25–30 years, were analysed since there were very few respondents aged more than 30 years. Age was an influential contextual factor for the perceived importance of all attribute groups, and it was much more influential for the PA (Indoor climate & visual environment) and PB (Accessibility & furnishing) attribute groups. In addition, Asians were affected more than Europeans by PB (Accessibility & furnishing), while graduate students were more sensitive than undergraduate students to PA (Indoor climate & visual environment) and PB (Accessibility & furnishing).

Each activity type has different purposes and characteristics, leading to different estimates of an attribute group’s importance for space choice. Participants who ticked the ‘reading’ activity box during their visit attributed greater importance to PC (Space congestion), which seems reasonable since quiet and uncrowded spaces are most suitable for reading. The ‘meeting’ activity places more importance on PD (Conversation privacy & interaction). Participants who engaged in online study tended to place less importance on most of the attribute groups, except for PD (Conversation privacy & interaction). Those who were working alone placed a high importance on PA (Indoor climate & visual environment), and a correspondingly low importance on PD (Conversation privacy & interaction).

In addition to the user and activity types, various additional considerations affected the perceived importance of attribute groups. The three main considerations are discussed here. First, participants who visited the building during weekends placed greater importance on PB (Accessibility & furnishing). Second was whether or not to visit with friends or co-workers. Those who were with friends or co-workers placed much greater importance on PD (Conversation privacy & interaction). Third was the ‘expected length of stay’, which had greater influence on the importance of the PA (Indoor climate & visual environment), PB (Accessibility & furnishing), and PD (Conversation privacy & interaction) attribute groups, implying that those who stayed longer were more sensitive to these attribute groups.

The findings contribute to a comprehensive understanding of how contextual factors affect space-choice behaviour. Since the contexts of academic libraries are variable along many dimensions, such as gender ratio, university culture, library location and activity profiles, architects can focus more on some attributes when considering the specific contexts of the library in question. For example, a library building might be more intensively used by people whose visits are protracted or take place at weekends, and the building’s space usage might have a poor balance between crowded and unused spaces. In this case, it would be effective to pay more attention to the PB (Accessibility & furnishing) attribute group to achieve a balanced use of library space. The provision of better furnishing or easier adjustment to physical conditions would be effective ways to attract users towards under-used spaces and away from crowded space. As another example, if some libraries are expected to accommodate more female students, then architects might pay more attention to the space attributes belonging to the PA (Indoor climate & visual environment) and PD (Conversation privacy & interaction) attribute groups for optimal space-usage in the library. However, it should be acknowledged that the influential contextual factors in academic libraries could be site-specific, whereas the present study is based on a single library.

Influential contextual factors in satisfaction

Multiple regression models were tested with the attribute groups of satisfaction (i.e. SA, SB, SC, SD, SE and SF), using contextual factors as dummy variables. The following five models were tested: Model 1, which was run with the attribute groups alone and without any contextual factors; Model 2, which was run with attribute groups and user type; Model 3, which was run with attribute groups and activity type; Model 4, which was run with attribute groups and additional considerations; and Model 5, which was run with attribute groups and all contextual factors. All models show p-values below 0.001. The coefficients of each attribute group remained almost unchanged across the different models, which implies that the influences of the attribute groups on overall satisfaction level remain unchanged regardless of contextual factors. Dummy variables in user type and activity type predicted very little change in overall satisfaction level, which indicates user type and activity type have little influence on overall satisfaction level. The additional considerations, ‘Weekday/Weekend’ and ‘Accompany with friends or co-workers/Alone’, only has an influence on overall satisfaction level with changing overall satisfaction in Model 4. In conclusion, the findings indicate that the importance of space attributes to overall satisfaction is less sensitive to contextual factors than to choice of space. This is reasonable, in that when users choose a space they seek the one that is most suitable for their activity and preferences, whereas a satisfaction level is determined in a chosen space after those factors have been considered.

Conclusions

Without an understanding of space-choice behaviour and users’ satisfaction, architects may not have enough information to design efficient and satisfactory spaces during space planning and design. The effects of space attributes in academic libraries have mostly been investigated with a separate focus on either space-choice behaviour or user satisfaction. Since the contexts of different libraries vary greatly, it might be misleading to assess the importance of particular space attributes for space choices and user satisfaction on the same basis as that employed in the present study. Addressing this gap in our knowledge necessitates a thoroughgoing investigation of both dimensions conjointly, from the same context of an academic library. The present study thus carried out a self-completion survey in an academic library in terms of the two dimension, namely perceived importance and satisfaction, in order to understand how library-users make space choices and what accounts for their satisfaction in the chosen space respectively.

As a result, the study demonstrated that the influences of some space attributes are non-identical between space choice and satisfaction. ‘Amount of light’, ‘Amount of space’ and ‘Cleanliness’ have only a slight influence upon overall satisfaction, although they are highly influential upon space choice; meanwhile, ‘Aesthetic appearance’, ‘Visual comfort’ and ‘Window views’ are highly influential upon satisfaction, but they have only a low influence upon space choice. The differences may result from the process of selecting and then using a space. Library-users first take some space attributes into account when selecting a space which meets the functional requirements of their activities: whether or not the lighting is bright enough, whether the space is big enough for their activities, and whether it is clean enough. The selected spaces are thus likely to have good conditions with respect to those space attributes, thereby less influential on overall satisfaction. In the present study, we believe that many library-users selected spaces meeting their functional requirements, so that as a follow-up to the space choice, users’ satisfaction levels were determined by attributes providing comfort and enjoyment during their use of the space.

However, if library-users cannot find spaces meeting their functional requirements and then have no alternative but to use less functional spaces, those space attributes could exert greater influence upon their overall satisfaction. Indeed, many academic libraries are likely to be crowded during the semester, with the result that library-users may be unable to occupy their favourite seats. Since the present findings were affected by the process of selecting spaces and then using them, our correlations between space attributes for space choice and satisfaction cannot be generalized. A supplementary method may thus be required, in order to determine the extent to which such a process affects satisfaction levels. One possible method might be to conduct a direct-ratings survey for satisfaction and then to compare the satisfaction results between direct ratings and regression analyses.

The study identified four and six attribute groups, for perceived importance and satisfaction respectively, with similar patterns in space choice and overall satisfaction. This implies that, when a grouped attribute has been emphasized, the effects of the group can be leveraged for space choice and satisfaction. In this connection, if a space attribute cannot be modified because of certain constraints, the same effects on space choice and overall satisfaction can be achieved by changing conditions of other space attributes in the same group. In addition, the study found that the attribute groups that influence overall satisfaction are stable across contexts, whereas the perceived importance of attribute groups related to space choice varies across contexts.

We should acknowledge that the findings cannot be generalized to all academic libraries, since the survey was conducted in only one building at a particular university. Library contexts vary along many dimensions, such as size, university culture, and rural or urban location. Moreover, the library in our study was located in a research university specializing in engineering and technology. The main library-users were male engineering students and there were more graduate students than undergraduate students. These specific contexts may affect the importance of space attributes in terms of space-choice behaviour and satisfaction. Future studies are thus required to conduct similar studies in diverse library contexts in order to test generalizability. For example, although age was shown to be an influential contextual factor in perceived importance in the present survey, the comparison was made only between groups aged 19–25 years and 26–30 years. In addition, although all activities were work-related, in future they could be differentiated from non-work-related activities. Moreover, some additional space attributes need to be identified and tested. For example, the building in which the survey was conducted has a good Wi-Fi signal in every micro-location, whereas in buildings elsewhere the strength of the Wi-Fi signal might vary from one micro-location to another. This variability can lead to different indoor space-choice behaviour and satisfaction level, since many space-users bring their own laptops to work or pursue online study in a space with a good Wi-Fi signal.

Once generalized in future studies, an enhanced understanding of how space attributes in libraries influence users’ space choice and satisfaction should help architects to plan and design effective library buildings in which users are satisfied. Moreover, the findings can contribute to the development of academic library space-plans that are both more efficient and more sustainable.

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Tae Wan Kim

Author biographies

Seung Hyun Cha is an assistant professor in the Department of Interior Architecture Design at Hanyang University. He received his Master’s degree in Sustainable Design and Construction from Stanford University and his PhD in Architecture from the University of Cambridge in 2011 and 2015 respectively. His research interests include new ways of working, activity-based working, space management, and post-occupancy evaluations.

Tae Wan Kim is an assistant professor at the Division of Architecture and Urban Design, Incheon National University. He received his PhD degree in the Department of Civil and Environmental Engineering from Stanford University. He is interested in modelling and forecasting human behaviour in the built environment.

References

Alpert

(1971) Identification of determinant attributes : A comparison of methods. Journal of Marketing Research 8(2): 184–91.

Applegate

(2009) The library is for studying: Student preferences for study space. Journal of Academic Librarianship 35(4): 341–346.

Ashkanasy

Ashton-James

Jordan

(2004) Performance impacts of appraisal and coping with stress in workplace settings: The role of affect and emotional intelligence. In: Perrewe

Ganster

(eds) Emotional and Physiological Processes and Positive Intervention Strategies. Bingley: Emerald, pp. 1–43.

Brown-Sica

(2012) Library Spaces for urban, diverse commuter students: A participatory action research project. College & Research Libraries 73(3): 217–231.

Cha

Kim

(2015) What matters for students’ use of physical library space? Journal of Academic Librarianship 41(3): 274–279.

Cha

Steemers

Kim

(2017).Modelling building users’ space preferences for group work: A discrete-choice experiment. Architectural Science Review 60(6): 460–471.

Crumpton

Crowe

(2009) Using evidence for library space planning. In: Proceedings of the 2008 library assessment conference: Building effective, sustainable, practical assessment (eds Hiller

, et al.), Seattle, WA, USA, pp. 51–64. Washington, DC: ARL.

DeClercq

Cranz

(2014) Moving beyond seating-centered learning environments: Opportunities and challenges identified in a POE of a campus library. Journal of Academic Librarianship 40(6): 574–584.

Erhan

(2003) Interactive computational support for modeling and generating building design requirements. Doctoral Dissertation, Carnegie Mellon University, Pittsburgh, PA, USA. Available at: http://www.cmu.edu/architecture/research/grad_work/2003_phdcd_erhan_halil.pdf (accessed 12 March 2016).

10.

Fox

Doshi

(2013) Longitudinal assessment of ʻuser-drivenʼ library commons spaces. Evidence Based Library and Information Practice 8(2): 85–95.

11.

Frontczak

Schiavon

Goins

, et al. (2012) Quantitative relationships between occupant satisfaction and satisfaction aspects of indoor environmental quality and building design. Indoor Air 22(2): 119–31.

12.

Fulton

(1991) A conceptual model for understanding the physical attributes of learning environments. New Directions for Adult and Continuing Education 50: 13–22.

13.

Gustafsson

(2004) Determining attribute importance in a service satisfaction model. Journal of Service Research 7(2): 124–141.

14.

Halling

Carrigan

(2012) Navigating user feedback channels to chart an evidence based course for library redesign. Evidence Based Library and Information Practice 7(1): 70–81.

15.

Hassanain

(2010) Analysis of factors influencing office workplace planning and design in corporate facilities. Journal of Building Appraisal 6(4): 183–197.

16.

Holder

Lange

(2014) Looking and listening: A mixed-methods study of space use and user satisfaction. Evidence Based Library and Information Practice 9(3): 4–27.

17.

Lee

Guerin

(2009) Indoor environmental quality related to occupant satisfaction and performance in LEED-certified buildings. Indoor and Built Environment 18(4): 293–300.

18.

May

Swabey

(2015) Using and experiencing the academic library: A multisite observational study of space and place. College & Research Libraries 76(6): 771–795.

19.

Montgomery

(2014) Library space assessment: User learning behaviors in the library. Journal of Academic Librarianship 40(1): 70–75.

20.

Ojennus

Watts

(2017) User preferences and library space at Whitworth University Library. Journal of Librarian-ship and Information Science 49(3): 320–334.

21.

Organ

Jantti

(1997) Academic library seating: A survey of usage, with implications for space utilisation. Australian Academic & Research Libraries 27(3): 205–216.

22.

Stanford University (2009) Stanford University Space and Furniture Planning Guidelines. Available at: https://vpasp.berkeley.edu/sites/default/files/stanford_space_guidelines.pdf (accessed 6 December 2017).

23.

Texas State University (2018) Texas State Construction Standards. Available at: http://www.facilities.txstate.edu/pdc/Projects_Documents/Construction-Standards.html (accessed 23 February 2018).

24.

Van Ittersum

Pennings

JME

Wansink

, et al. (2007) The validity of attribute-importance measurement: A review. Journal of Business Research 60(11): 1177–1190.

25.

Webb

Schaller

Hunley

(2008) Measuring library space use and preferences: Charting a path toward increased engagement. portal: Libraries and the Academy 8(4): 407–422.

26.

Weessies

(2011) A locational analysis of academic library computer use. Reference Services Review 39(3): 465–481.

27.

Winterbottom

Wilkins

(2009) Lighting and discomfort in the classroom. Journal of Environmental Psychology 29(1): 63–75.

28.

Yoo-Lee

Lee

Velez

(2013) Planning library spaces and services for millennials: An evidence-based approach. Library Management 34(6/7): 498–511.

29.

Zagreus

Huizenga

Arens

, et al. (2004) Listening to the occupants: A Web-based indoor environmental quality survey. Indoor Air 14(8): 65–74.