DEPARTMENT OF TECHNOLOGY MANAGEMENT AND ECONOMICS 
DIVISION OF SCIENCE, TECHNOLOGY AND SOCIETY  

CHALMERS UNIVERSITY OF TECHNOLOGY 
Gothenburg, Sweden 2023 
www.chalmers.se 
Report No. E2023:108 

 

Libraries and values  

Building trust in Swedish public libraries through 
transparency 

Master’s thesis in Management and Economics of Innovation  

 
 

 
 
VIKTOR JOHANSSON 
 
 

 



   

 

   

 

 

  



 

 

 

 

 

 
 

 
REPORT NO. E2023:108 

 
 
 
 
 
 
 

 
 

Libraries and values 
 

Building trust in Swedish public libraries through 
transparency 

 
VIKTOR JOHANSSON 

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

 
 
 

 
 
 

Department of Technology Management and Economics 
Division of Science, Technology and Society 

CHALMERS UNIVERSITY OF TECHNOLOGY 
Gothenburg, Sweden 2023



   

 

   

 

 

 

 

 

 

 

 

 

 

 

 

Libraries and values 

Building trust in Swedish public libraries through transparency 

VIKTOR JOHANSSON 

 

 

© VIKTOR JOHANSSON, 2023. 

 

 

Report no. E2023:108 

Department of Technology Management and Economics 

Chalmers University of Technology 

SE-412 96 Gothenburg 

Sweden 

Telephone + 46 (0)31-772 1000 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Gothenburg, Sweden 2023



   

 

   

 

Libraries and values 

Building trust in Swedish public libraries through transparency 
 

VIKTOR JOHANSSON 

 

 

Department of Technology Management and Economics 

Chalmers University of Technology 

 

 

 

SUMMARY 
Democratic institutions around the world have been  
questioned recently, including libraries. Problems include 
polarization and populism. The purpose of this  
thesis is to increase the trust of libraries against this  
backdrop. This is done using critical  
algorithm theory on transparency, to improve the  
possibilities, for the public, of knowing how Swedish  
small to medium sized public libraries collections of books 
are created (via the libraries’ process of purchasing).  
13 interviews and six documents are used as empirical 
data. I conclude that: 1) The process of purchasing used  
at these libraries can be made transparent, to increase  
the trust in libraries, by showing empirically their neutrality 
based in the diverse group of citizens they serve. 2) A part 
of this is making the valuation of quality transparent, that is, 
how librarians balance the (literary, scientific etc.) quality of 
books against other criteria in the purchasing process,  
especially the criteria of what books library visitors demand.  
3) The debate between user-centric and collection-centric  
collections, that is the degree of power between users and  
librarians in the purchasing process, needs to account for  
another dimension, feedback from users – made possible by  
an increased transparency. By doing this, I contribute to critical  
algorithm theory through an empirical study recommending  
transparency. Second, to valuation studies, through 
investigating valuations at libraries. Third, to library studies,  
by adding a dimension to the debate about user-centric  
and collection-centric collections. 
 
 

 

 

 

 

 

 

 

Keywords: Critical algorithm theory, transparency, trust, valuations, library, library studies, 

Sweden, interviews, documents. 



   

 

   

 

 

 

 

  



 

 

 

 

 

 
 

Table of contents: 

1. Introduction ............................................................................................................. 1 

1.1 Swedish libraries .............................................................................................................. 1 

1.1.1 Libraries and trust ..................................................................................................... 2 

1.1.2 Librarians valuation of the quality of books ............................................................. 2 

1.2 Aim of the thesis .............................................................................................................. 3 

1.3 Contribution and research questions............................................................................... 3 

1.4 Delimitations .................................................................................................................... 4 

1.5 Structure of the thesis ...................................................................................................... 4 

2. Previous Research .................................................................................................... 5 

2.1 Trust and transparency .................................................................................................... 5 

2.2 Valuations ......................................................................................................................... 6 

3. Methodology ............................................................................................................ 8 

3.1 Research design................................................................................................................ 8 

3.1.1 Qualitative research .................................................................................................. 8 

3.1.2 Abductive design ....................................................................................................... 8 

3.1.3 Multi-methods .......................................................................................................... 8 

3.2 Data collection, Interviews and Documents .................................................................... 9 

3.2.2 Documents .............................................................................................................. 10 

3.3 Data Analysis .................................................................................................................. 11 

3.3.1 Translation............................................................................................................... 11 

3.4 Triangulation, validity, reliability, limitations ................................................................ 11 

3.4.1 Triangulation ........................................................................................................... 12 

3.4.2 Validity and reliability.............................................................................................. 12 

3.4.3 Generalization ......................................................................................................... 12 

3.4.4 Limitations ............................................................................................................... 13 

3.6 Ethical considerations .................................................................................................... 13 

3.7 Table of Interviews ......................................................................................................... 13 

3.8 Table of Documents ....................................................................................................... 14 

4. Results ................................................................................................................... 16 

4.1 Things that are opaque to the public concerning libraries ............................................ 16 

4.1.1 The typical process of purchasing ........................................................................... 16 

4.1.2 Avoiding BTJ ............................................................................................................ 17 

4.1.3 Purchase suggestions .............................................................................................. 17 

4.1.4 Removal of books .................................................................................................... 18 

4.1.5 The invisible non-users............................................................................................ 18 



 

 

 

 

 

 
 

4.1.6 Political interference in librarians’ work .................................................................. 18 

4.2 Librarian’s valuation of quality ....................................................................................... 19 

5. Discussion .............................................................................................................. 20 

5.1 Opaque to transparent – to increase trust .................................................................... 20 

5.1.1 Implementing transparency .................................................................................... 21 

5.2 The consequences of librarians' valuation of quality .................................................... 21 

6. Conclusion .............................................................................................................. 23 

7. References ............................................................................................................. 25 

Appendix A: ................................................................................................................... 27 

Interview guides: .................................................................................................................. 27 

Interview guide A: ............................................................................................................ 27 

Interview guide B: ............................................................................................................ 29 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 



 

 

 

 

 

 
 

 

 

  



 

 

 

 

 

 
 

Acknowledgements 

First and foremost, I want to say thank you to my supervisor Nicholas Surber. You gave me 

new insights into what it means to be a researcher in every meeting, often through 

metaphors which I will bear with me for a long time. Often, you made me see weaknesses, 

which in turn made me develop, refining my eyes into something new. Just like what Proust 

hoped to do with his masterpiece. 

Also, thanks to all the interviewees who made this thesis possible, I like you patrons of the 

libraries! 

  



 

 

 

 

 

 
 

 



   

 

 

  1 

1. Introduction 
 

Trust is an invaluable resource for societies (Holmberg & Weibull, 2013). It makes people go 

to the hospital when sick and order food without worries at a local restaurant when hungry 

(Holmberg & Weibull, 2013). In short, it makes people cooperate to reach a goal in a specific 

context (Holmberg & Weibull, 2013). Libraries are often seen as part of creating this trust 

between people as an important part of democratic societies (Rivano Eckerdal, 2017). Trust 

is also necessary for libraries to work themselves. 

Libraries have since the beginning of the Swedish democracy about 100 years ago been a 

part of the same (Lindsköld & Hedemark, 2023). These days, polarization is a common topic 

of debate and research, i.e., the reasons behind and the consequences of, among other 

things, questioning institutions from different sides of the political spectrum (Haider & 

Sundin, 2022). At best, libraries can act as an antithesis to the political rhetoric that goes 

over people’s heads, and instead make available useful knowledge, perspectives and ideas 

about the world. Among other things, this is how libraries have been argued to contribute to 

a more democratic society than would be the case without them (Rivano Eckerdal, 2017). 

The level of trust of libraries in Sweden is large relative to other institutions in Sweden 

(Eriksson & Michnik, 2019). In 2018, 58% of Swedish citizens had a very high or a fairly high 

trust in the libraries, while usually being over 65% between 2009-2015 (Eriksson & Michnik, 

2019), i.e., trust has decreased in recent years. The trust of institutions in Sweden in general 

has been relatively high compared to other countries (Segerberg, 2020). This trust in 

institutions has been stable over time, i.e., not declining much in recent years (Segerberg, 

2020). Whether or not the trust of libraries should be seen as large is not obvious (Michnik, 

2018). 

Regardless, only a minority of the Swedish population visits a library each year, and 30% of 

the population visits a library at least once each third month (Höglund, 2017). This could be 

seen as an indication that there is a gap between how trusted libraries are (as indicated 

from surveys) and the degree of usage, showing a potential increase in the value of their 

services for citizens. 

 

1.1 Swedish libraries 

The libraries of Sweden are guided by Bibliotekslagen (Bibliotekslag, 2013:801). It has been 

shown how this law is often reflected in library plans (Kungliga biblioteket, 2021), the plans 

which each municipality is obliged to formulate about their own strategy over a course of 

years, often 3-5 years.  

Public libraries are governed by their respective municipalities due to the Swedish municipal 

self-governance. The libraries are meant to be for all and are supposed to be neutral in 

terms of political, religious etc., ways according to Bibliotekslagen, which in practice creates 

a neutrality based on diversity (Haider & Sundin, 2022) due to a diverse set of citizens. The 

Swedish libraries' base of loan-takers is large but decreasing, around six books per person 

and year has been borrowed on average in recent years (Kungliga biblioteket, 2022). 



 

 

 

 

 

2 
 

“BTJ-häftet” is often used for purchasing at Swedish libraries. It is a service provided by the 

company BTJ. The service includes several reviews of upcoming books and helps librarians to 

make purchasing decisions. 

1.1.1 Libraries and trust 

As seen, libraries in Sweden are largely trusted relative to other institutions (Eriksson & 

Michnik, 2019). Recently, in a survey about politicians and trust, only about ten percent of 

the 206 library managers asked, answered that they have experienced political interference 

in their work (Ericsson & Clemens, 2023). Worries of increased political interference in the 

future were also experienced by about ten percent of the respondents, especially due to 

new political parties in charge (Ericsson & Clemens, 2023). Evidence in media (SVT, 2022) 

shows that libraries have become arenas of the so-called “culture war” in Sweden (Höglund, 

2023). 

Researchers have investigated what activities increase and decrease the level of trust of 

libraries (Michnik, 2018). Especially activities that fall out of the scope of what people expect 

of the libraries decrease trust (Michnik, 2018). I argue that actions which are seen as 

censorship can be considered one such activity. Therefore, transparency that can refuse 

such claims, is valuable. Both the empirical data to be able to make such claims trustworthy, 

as well as conditions to make this transparent exists according to my interviews, and my 

analysis. 

1.1.2 Librarians valuation of the quality of books 
The (literary, scientific etc.) quality of the purchased books at libraries has been observed to 

be a relatively non-influential variable in the purchasing process at Swedish libraries 

(Lindsköld & Hedemark, 2023; Hedemark, 2020). This could be due to many factors. One is 

libraries’ historically important step towards so-called user-centric collections in the early 

2000’s (Söderlind & Elf, 2014). This means the collection of books went from being mainly 

based on librarians' decisions, to being designed by the visitors (Söderlind & Elf, 2014). This 

means the most borrowed and demanded books are meant to be those mainly being in the 

library to increase the use of books and to satisfy citizen's needs (Söderlind & Elf, 2014). 

With that change, librarians started to think of many other factors than quality compared to 

before (Söderlind & Elf, 2014). 

A formulation on quality was first introduced in the latest version of Bibliotekslagen 

(2013:801) (Söderlind & Elf, 2014). The quality in this case concerns the quality of the books 

the libraries should buy, but it is not specified what quality means there. There are a lot of 

different views on quality in art and literature, but from looking at Swedish national reading 

policy through history (Lindsköld & Hedemark, 2023) it is obvious that the vision from 

decision-makers always has been to make people read more “high-quality literature”. 

KB’s summary of reading 78 library plans from 78 municipalities (Kungliga biblioteket, 2021) 

shows trends in library plans. The quality of books is seldom mentioned in the plans 

according to the authors, only in one of twenty plans, despite formulations of the value of 

literary quality in Bibliotekslagen which the same plans often cite from.  

The neutrality ideals libraries follow have probably contributed to the relative lack of focus 

on quality in libraries work with purchasing as well, due to the recurring controversies of 

implying any specific quality ideal in cultural settings (Söderlind & Elf, 2014). 

 



 

 

 

 

 

3 
 

1.2 Aim of the thesis 
When I write “libraries” or similar from now onwards, I mean Swedish, small- to medium-

sized, public libraries and foremost their collections of books. Saying that, I am not 

considering anything about the library as a social room in this thesis, except in direct relation 

to the collection of books. I specify this delimitation in more detail in the Delimitations 

section. 

With the thesis I aim to understand transparency in practice at libraries, i.e., how libraries 

can be more open about and inform their visitors about their way of working, in this case 

focusing on their purchasing of books. This is done by applying theory from the critical 

algorithm field. I especially apply theory on transparency from this field. In that way, 

potential solutions using transparency to increase trust in libraries are highlighted. 

I will also show how the quality of books is valued in different ways in today’s libraries. That 

is, in relation to the collection of books and how these books are purchased. I argue that 

these valuations could be opened up for feedback from citizens by making them 

transparent, and if the citizens want, lead to a different purchasing process and/or new 

valuations of quality. 

The conclusions in the thesis can be used to describe the connection between democracy, 

democratic institutions and libraries. Trust is tightly connected to transparency also for 

other institutions than libraries (Holmberg, Weibull, 2013). 

In their essay on “förtroende”, or trust (own translation), of institutions, Holmberg and 

Weibull (2013) mention five factors that decide the level of trust of an institution according 

to institution-based trust research, namely competence, integrity, empathy, transparency 

and degree of shared values. With this thesis I thus focus on two of the factors that build 

trust, competence (via the reasoning on librarians' valuation of quality), and transparency of 

the libraries (via transparency and trust). Researchers don’t know how much each of the five 

factors decide respectively (Holmberg, Weibull, 2013). The value of focusing on these two is 

thus at least not worse than focusing on any of the other. 

 

1.3 Contribution and research questions 

The problem which the thesis tries to address is how libraries can increase their already 

relatively large level of trust, and through this, address the issues of polarization that 

decreases the level of trust of democratic institutions overall in today’s society (Haider & 

Sundin, 2022). 

The thesis contributes to both the debate regarding the balance between user-centric and 

collection-centric collections within library studies (Söderlind & Elf, 2014), i.e., the power 

balance between librarians and the public over the collection of books, as well as to the 

critical algorithm field through an empirical study focusing on libraries. Thirdly, it contributes 

to valuation studies, again with libraries as a case. Taken together, the conclusions can give 

insights into what other democratic institutions could learn from libraries and the arguments 

of trust put forward in the thesis.  

The thesis is based on 13 interviews and a small document study of six documents as 

empirical data. Especially, the libraries collections of physical books were the focus of the 

interview questions. 



 

 

 

 

 

4 
 

I have three research questions for the thesis: 

1. Which aspects of Swedish small to medium sized public libraries are opaque to the public 

today? 

2. How do librarians value the quality of books in purchasing, as seen in Swedish small to 

medium sized public libraries? 

3. How can increased transparency of Swedish small to medium sized public libraries process 

of purchasing of books, and librarians value of quality in purchasing of books, increase the 

trust in and value of these libraries? 

The first two questions will be answered in the Results section, which both contribute to the 

third question which will be discussed in the Discussion section. 

 

1.4 Delimitations 
A few notes on wording. First, when I write “libraries” or similar, I mean Swedish, small- to 

medium-sized, public libraries and foremost their collections of books and how they are 

created by librarians and library users/visitors. Saying that, I am not considering anything 

about the library as a social room in this thesis, except in direct relation to the collection of 

books. By small libraries I mean libraries with 1-4 employees, medium libraries 4-10 

employees. In this thesis, mostly librarians working at small libraries have been interviewed. 

Often, libraries can be seen as medium-sized on a collected level in a municipality or 

"region" that the libraries cover, while individual libraries might be responsible for a town's 

library services, and then being defined as "small" in this thesis. I make the distinction 

between these libraries and what I would call large libraries since the large libraries was not 

interviewed to the same extent in this thesis, and since they often have so called floating 

collections, among other things, which the small to medium-sized libraries does not, and 

which makes the distinction relevant in this thesis. 

Secondly, by “non-visitors” and “non-users” I mean non-active loan-takers, i.e., persons that 

have not used the libraries’ services or borrowed a book in 1-3 years. 

 

1.5 Structure of the thesis 
The thesis is structured as follows: First, I describe some of the previous research which has 

been important for the thesis, mostly on transparency from a critical algorithm perspective 

as well as theory on valuations. Second, I will describe the methodology of the thesis and 

discuss some of the details and why I made the choices I made. Third, I answer the first two 

research questions, in the Results section. The empirical data which I chose to focus on 

based on the data collection is put into words there. Fourth, I relate the results to previous 

research to answer the third research question. Lastly, I conclude by summarizing the 

findings and answers to the research questions and propose next steps following the thesis's 

results and analysis. 

  



 

 

 

 

 

5 
 

 

2. Previous Research 
 

In this chapter I first discuss previous research on transparency within critical algorithm 

studies. Combined with the references about trust in the introduction, these sources will be 

used to discuss how transparency can influence trust in libraries, in the Discussion section. 

Secondly, I explain the idea of valuations and how they can be seen in library contexts.  

 

2.1 Trust and transparency 
Here, I delve into the concept transparency. I will show different ways it has been 

conceptualized and its usage in practice today as well as bring up some important nuances. 

Transparency can change both the direction of work and the way people interact with and 

use, e.g., the local library's services. In the case of libraries, it would be used to communicate 

with library visitors. Before that though, I contextualize the purpose of critical algorithms 

studies and relate it to transparency to showcase how transparency should be understood in 

this thesis. 

Critical algorithms studies use sociological perspectives on so called algorithmic 

assemblages, such as an organization doing pandemic surveillance, to understand, among 

other things, why different people know different things about the work an organization 

does (Lee, 2021). For example, some may know all aspects that create a certain risk of 

viruses spreading, and how this affects the map which a pandemic surveillance organization 

uses to inform decision-makers about the state of the day of the virus, while others think the 

map is simply colored in different ways but don’t know exactly why (Lee, 2021). In similar 

ways, only some people understand how purchasing is done in libraries today. Rather than 

being transparent according to critical algorithm theory (Lee, 2021), the purchasing process 

that creates library's collection of books is opaque. In this way, the collection can be seen as 

transparent only to some extent, anyone can go through the libraries’ catalogue of books if 

they know about the catalogue (like decision-makers can see the map but not how it is 

created in the pandemic surveillance example), while only a part of the group that knows 

this also knows how the collection is created by librarians. Since transparency is part of 

creating trust (Holmberg & Weibull, 2013), making this process known could be done via 

transparency to potentially increase trust. In this process, the following knowledge from 

critical algorithm theory can help. 

Transparency has become a popular term to use in guidelines for many things, among them 

AI (Larsson & Heintz, 2020). The underlying logic is easy in principle, by informing of how and 

what is done, the organization is opened for internal and external criticism, increasing the 

ways performance may be checked. Though it quickly gets more complicated. As Kolkman 

(2022) has pointed out, “transparency to (non-)experts is at best problematic and at worst 

unattainable”. That is, in technically complex contexts (Kolkman, 2022). For libraries though, 

technical aspects are not as important, and therefore the transparency concept is, in this 

sense, easier to apply to non-technical rather than technical contexts. Still, a lot of empirical 

data and knowledge is needed to be able to understand the context well enough to be able 

to make realistic accounts of the reality that should be made transparent and communicate 



 

 

 

 

 

6 
 

this in a way that is effective toward the intended audience, also in less technical cases 

(Kolkman, 2022). 

de Fine Licht and de Fine Licht (2020) explain how transparency can describe a decision, 

describe the process to the decision or describe why the machine and humans work in the 

way they do – all potentially increasing trust, with the underlying logic that explanations 

yield trust. For libraries, I argue both are relevant as purchasing is a process that ends with a 

resulting collection. 

Transparency has been described in multiple ways, which can be summarized as a set of 

perspectives according to an extensive literature study (Albu & Flyverbom, 2019). The first 

view of transparency within the studied field indicates a tendency to think that more 

information to as many people as possible is good (Albu & Flyverbom, 2019). Transparency 

then works as a kind of light that does not change anything within the respective 

organization while making information open in new ways, which might be seen as unrealistic 

(Albu & Flyverbom, 2019). The assumed consequences of transparency with these 

perspectives include a belief that a rigid trust and information structure is created through 

transparency, creating information balance and increasing access (Albu & Flyverbom, 2019). 

The second perspective puts more emphasis on the processes, negotiations and 

complications involved in deciding the boundaries of transparency and then implementing it 

(Albu & Flyverbom, 2019). Transparency is consequently conceptualized as complex and 

largely social work (Albu & Flyverbom, 2019). Transparency might have unintended 

consequences (Albu & Flyverbom, 2019). 

In this thesis I believe libraries show characteristics in relation to transparency that reflect 

both strands of conceptualizations. Most things that I argue libraries should make 

transparent are already known, and the interviewees asked about this all agreed that these 

things are not secret and would be answered by them orally if they were asked. This reflects 

the light metaphor (Albu & Flyverbom, 2019). At the same time, making these things 

transparent would put things on the spot, and it’s not unlikely that one or more librarians 

disagree about details of what should be made transparent and how, highlighting the social 

perspective of transparency as well (Albu & Flyverbom, 2019).  

Ananny and Crawford (2018) lists several drawbacks of transparency, which are not likely to 

be problems in the case studied in this thesis, except for the risk of a decreased trust if what 

becomes transparent is disappointing to, in this case the library visitors. In relation to the 

idea of a critical audience, this eventual decreased trust can be dealt with by changing in 

response to the feedback of this critical audience (Kemper & Kolkman, 2019).   

 

2.2 Valuations 
In valuation studies, how valuations are constructed and can be seen in practice is studied 

(Lee et al., 2022). That includes, among other things, understanding which objects are 

valued and why, and how different valuations often yield different practices to reach the 

same end goal (Lee et al., 2022). In the context of libraries, the quality of books is a relevant 

value for their visitors. For some readers, the literary quality as defined by experts may not 

be valued, but by some it is part of their valuations and decides which books they choose. 

The books thus show a kind of multitude of value in relation to different visitors (Lee et al., 

2022). This can be likened to how fetuses in some situations are seen as having qualities of 



 

 

 

 

 

7 
 

life and are then handled with greater care, while in other situations they are seen primarily 

as useful for experiments in medicine, emphasizing the future value these experiments can 

bring (Lee et al., 2022). For librarians, the importance of the quality of books relative to 

other criteria, regarding purchasing books to the libraries, such as demand, genre etc., can 

vary, as will be seen in the Results section. 

Valuations also change over time. In medicine contexts, influential researchers changed the 

focus of most ethical research within the field from the doctors to focusing on the patients 

(Lee et al., 2022). In libraries, the introduction of the idea of user-centric collections in 

Sweden have been described as changing a lot of how collections of books are created 

(Söderlind & Elf, 2014). As such, the problem description, which is also part of valuations 

(Lee & Helgesson, 2020), had changed, and the practice to accomplish the new goals made 

new valuations necessary. The librarians were no longer spending as much time creating and 

withholding the collections, with less influence due to the increased influence of the library 

visitors (Söderlind & Elf, 2014). They worked on a different problem, creating new valuations 

(Lee et al., 2022). 

  



 

 

 

 

 

8 
 

 

3. Methodology 
 

In this chapter I start by describing the overall approach used to conduct the research. I then 

move on to describe the data collection, in the form of interviews and documents, as well as 

why these were used and in what way. Following that, I describe the data analysis procedure 

and reflect on the research quality of the thesis. I end by noting the ethical considerations of 

the research project and include two tables that give an overview of the interviews and 

documents used in the thesis. 

 

3.1 Research design 
Here, I give an overview of the research design. 

3.1.1 Qualitative research 

The whole study was guided by the general principles of qualitative research as described by 

(Bell, Bryman & Harley, 2019), e.g., focusing on the context that the research is concerned 

with, and being flexible to the best methods and research questions over time etc. I was 

using purposive sampling (Bell, Bryman & Harley, 2019), i.e., made a conscious choice of 

organizations rather than picking random participants, to find relevant organizations to 

answer the research questions.  

3.1.2 Abductive design 

In the interviews I listened to answers and asked questions in an inductive way (Bell, Bryman 

& Harley, 2019) in the sense that I did let the interviewees respond to what they consider 

the most important aspect of libraries collections and other aspects of libraries rather than 

define those aspects myself. This approach was used especially in the first four interviews. I 

changed the interview questions as I got input from earlier interviews. At the same time, I 

kept ideas of valuation, quality and transparency in the back of my mind, and thought from 

early on until the end that I should focus mostly on the collections. That was, since I both 

hypothesized this could be a good fit from the beginning, and since the interviews did not 

change my mind about this. In that sense, there are deductive (Bell, Bryman & Harley, 2019) 

aspects to the design as well. The analysis was based on the interviews, and it was not 

known what it would be about until after the data collection, but also previous theory was 

used and, in that sense, the analysis is both inductive and deductive. Overall, the research 

design can therefore be seen as abductive. 

3.1.3 Multi-methods 

The results of the thesis are as mentioned based on both interviews and documents. Both 

these types of sources have been seen as and analyzed as texts in this thesis (Bell, Bryman & 

Harley, 2019). Since the interviews were conducted before I got access to the documents, 

most of the interview analysis was made prior to analyzing the documents. Due to that, the 

focus of the analysis and results were chosen based on the interviews. The documents were 

mostly about similar topics that I had talked about in the interviews, and therefore a lot of 

the themes from the interviews and documents overlapped. I therefore mention them 

intertwiningly in the Results section. 



 

 

 

 

 

9 
 

The difference between analyzing the interviews and documents is mostly “the audience” of 

the “texts”. The interviews were conducted in a setting with a few people and meant only 

for this study, while the documents are either public, directed towards local politicians, or 

for library colleagues. Therefore, different readings must be made as well. For public 

documents, an outside view is depicted, while internal documents are less so. The 

interviews can be seen as part of the outside view due to the consciousness of interviewees 

that the thesis will be public to some extent, while it is not public to the same extent as 

public documents. 

 

3.2 Data collection, Interviews and Documents 
Interviews can be used to find out about the past, and different interpretations and views 

about it (Bell, Bryman & Harley, 2019). Semi-structured interviews can give insights into how 

respondents think, by not acting presumptuous based on the researcher's prejudices or 

knowledge going into the interviews and can showcase the respondent's knowledge on a 

topic (Bell, Bryman & Harley, 2019). As I was relatively new to the subject at the start of the 

thesis, semi-structured interviews were thus suitable (Bell, Bryman & Harley, 2019). 

The interviews were 15 in total, each 25-35 minutes long, with 15 different people, while 

one person was interviewed twice, and one interview was done with two people. All 

interviews except one were done over the internet, 13 in digital calls and one as an e-mail 

conversation. 12 of the interviews were recorded (sound). See Table 1 for the interviews 

used in the thesis. 

Ten of the interviewees are librarians at small- to medium sized libraries, two are library 

managers, and three are researchers within the subject of library- and information studies. 

Out of the ten librarians, two had worked with central purchasing and floating collections in 

large libraries. These two interviews were meant to complement the other interviews by 

giving another perspective on how things are done in the less technical and smaller libraries. 

The two interview guides for interviews with library personnel can be found in Appendix A. 

The interviews with researchers were concerned with questions each of them was 

interested in respectively and are therefore difficult to include in a general interview guide. 

Mostly, the library personnel are referenced in the Results section, more than the 

researchers. 

After sending out the report a few days prior to submission, two of the interviewees (two 

librarians) chose to not be part of the study anymore due to disagreements of the aim and 

understanding of libraries in the thesis. Therefore, the number of interviews used as 

empirical data in the thesis is 13. 

To find participants to interview I emailed several library managers and asked them for 

people that would like to participate in the thesis. Many did not answer, and it took time to 

get 15 interviews. 1 interviewee responded to me in email after a post I made in a forum for 

librarians in Sweden, which I posted to at a late stage of the study.  

The librarians are the ones working hands-on with the things I became interested in and 

focused on in the thesis, hence the large number of interviews with them. The library 

managers instead can give an overview of how things are supposed to work in principle and 

are the ones uttermost responsible for communicating with decision-makers such as 



 

 

 

 

 

10 
 

politicians. Researchers were added since I thought they could give a better overview of the 

library system in Sweden, which actors are important not to ignore, as well as add more 

theoretical insight into sub-questions which I found interesting along the way of the thesis. 

The interviews were in most cases done digitally, sometimes in-place based on where the 

organization was located but mostly based on the organization and the informants’ 

preferences. The difference in effects between face-to-face and non-face-to-face interviews 

have shrunk, as people are much more used to doing similar activities with digital tools (Bell, 

Bryman & Harley, 2019). 

3.2.2 Documents 
The documents used in the thesis are partly public documents such as library plans, and 

purchasing policies, while a few are internal for the organization using them, i.e., the 

training document for substitutes and instructions for where to place books and how to 

approach library visitors. The documents were seven in total, while one document had to be 

removed since one of the interviewees’ that left the study had sent me that document. See 

Table 2 for the documents used in the thesis. 

Most of the documents were given to me after one of the last interviews, where I asked if 

they had any documents I could use in the thesis. I had also emailed some of the librarians I 

had interviewed earlier but did for different reasons not get documents from most of them.  

I also read two library plans which are publicly available. I read all documents that were 

given to me, which are all documents in the table below, but only three of them ended up 

being relevant to refer to in the Results section after all. 

Evaluating the relevance of, criteria for, and quality and possible interpretations of 

documents is critical in achieving good research quality using documents (Bowen, 2009). The 

criteria to consider for the quality of documents are “Authenticity”, “Credibility”, 

“Representativeness” and “Meaning” according to (Bell, Bryman & Harley, 2019). For all 

documents used in this study, as organizational documents, the authenticity and meaning of 

the documents are clear (Bell, Bryman & Harley, 2019). For public documents the credibility 

can be questioned since any organization rarely can or wants to show exactly how things are 

done in the organization (Bell, Bryman & Harley, 2019). Therefore, they must be read with 

the purpose of the document in mind. Internal documents that are used between colleagues 

are more credible since they are used in practice for a reason. The documents' 

representativeness is generally difficult to measure with such a small sample of documents 

as used in this thesis. Using (Kungliga Biblioteket, 2021) and their analysis of 78 library plans 

as a reference point, the public documents can be seen as representative due to the 

similarities between these library plans and the trends mentioned by Kungliga Biblioteket. 

The purchasing policy document’s representativeness can instead be evaluated by reference 

to (Söderlind & Elf, 2014) which analyzes media policy documents and shows similarities to 

the media policy document used in this thesis. The internal documents for colleagues I 

cannot evaluate due to no references and the small sample. 

The data quality of the documents is thus generally good. The exception being that the 

sample is relatively small, and especially for internal documents such as training documents 

more documents could give new insights and perspectives on things that have not been part 

of my analysis, but which may affect the answers to the research questions. 

 



 

 

 

 

 

11 
 

3.3 Data Analysis 
The data analysis has been done as a thematic analysis (Bell, Bryman & Harley, 2019). In this 

kind of analysis repetitive, engaging, contradicting and surprising aspects of the empirical 

data which is relevant to the chosen research question are developed into themes (Bell, 

Bryman & Harley, 2019). Not all themes are possible to develop most often, due to time 

constraints, so the researcher's skill, taste and background will influence the ultimate results 

of the process a lot (Bell, Bryman & Harley, 2019).  

In general, the analysis has been done continually throughout the research project, in an 

iterative way (Bell, Bryman & Harley, 2019). First, through changes in the interview 

questions based on previous interviews, and later through different stages of analysis where 

different themes seemed most relevant to analyze.  

Transcriptions of the interviews were done, but I mostly used my notes from the interviews 

in the initial stages of analysis to increase the speed. After I found interesting topics, I 

checked with the transcripts so that I had not misunderstood the context in which the things 

had been said. After that I could start structuring the arguments and complement some of 

the references. During writing, I started to realize new aspects of the data and found new 

themes to be the most valuable to continue working on, and so the focus of the thesis 

changed a few times before the final structure was chosen. 

As seen earlier, the documents were analyzed after the interviews, while both are brought 

up in tandem in the Results section. 

3.3.1 Translation 
Both the interviews and documents are in Swedish. This made it necessary for me to do 

translations. In general, translations can be seen as a technical, straight-forward work, from 

a positivistic perspective (Bell, Bryman & Harley, 2019). From constructivist perspectives, 

translations are instead seen as dependent on the translator's background and idea of the 

two languages and their cultures (Bell, Bryman & Harley, 2019). For this thesis the 

translation has mostly been straight-forward since the concepts used in Swedish have one 

appropriate English equivalent and since a linguistic analysis is not the type of analysis done. 

Applying a positivistic perspective on this, it works well to translate. From the constructivist 

perspective it is more complicated, but difficult to avoid. 

I mostly used my previous knowledge of English or used general translators online to 

translate. To translate concepts about libraries or research-related concepts I looked at 

other research papers to see what words they use, e.g., “public libraries” for “folkbibliotek”, 

i.e., the kind of library we most often call just “bibliotek” in Swedish. A few times, such as 

with the concepts of trust, quality and value, different interpretations exist both in different 

research fields and among readers, which could cause confusion or inconsistency when 

different persons get different ideas of what is meant. Hopefully the references about how 

trust have been discussed in previous research as well as how these concepts are discussed 

overall in this thesis can overcome these potential problems.  

 

3.4 Triangulation, validity, reliability, limitations 
Here, I discuss the research project’s quality in relation to different criteria. I note that some 

researchers have criticized the ideas of these criteria for qualitative research since this kind 



 

 

 

 

 

12 
 

of research always comes with the possibility of multiple accounts of reality (Bell, Bryman & 

Harley, 2019). At the same time, I think that for this research project these discussions can 

still be worthwhile for someone interested in its methodological choices. 

3.4.1 Generalization  
In qualitative research generalization is often considered impossible, since context and 

individual factors differ between settings. I still generalize between what I call small to 

medium-sized libraries in this thesis, arguing that the libraries I talked to are sufficiently 

similar to do so. Their process of purchasing, views of the value they provide to the visitors 

of libraries and how were so similar I argue the results of this research can be generalized to 

any small library using similar methods for purchasing. That is, since I interviewed five 

different small to medium sized libraries personnel and did not get contrary views on more 

than details. The overall process, the answers to my research questions, could be answered 

with any of the libraries as a representative candidate that the analysis would apply to. 

In short, this thesis can be considered as using a cross-sectional approach as defined in (Bell, 

Bryman & Harley, 2019). This means it compares multiple cases but focuses on 

generalization rather than each individual case's unique aspects (Bell, Bryman & Harley, 

2019). 

3.4.2 Triangulation 
Interviews with librarians from more technical libraries as well as documents were used for 

triangulation instead of comparing between the smaller libraries. That was due to the 

argued possibility to generalize. The technical libraries do things in a different way and could 

in this sense triangulate by showing alternative ways of creating the collection and having 

other processes for purchasing despite similar goals. The documents cover the same topics 

as the interviews and therefore add another type of source that could reinforce what I heard 

in the interviews. This gives some amount of triangulation to test the data from the 

interviews against other sources.  

3.4.3 Validity and reliability 
Regarding validity, whether the research is measuring what it is intended to (Bell, Bryman & 

Harley, 2019), the interview questions were a good fit to realize how the process of 

purchasing is done, and in that way find things that are opaque in today’s libraries. Also, 

valuations of quality can be seen through the same method since the criteria for purchasing 

and how librarians balance between different criteria for value is shown in their answers as 

well. Further information on what is opaque would be found through other methods, such 

as ethnography following librarians in their work with purchasing and the collection overall. 

That could show how they work with the collection hands-on as well as communicate about 

it in meetings etc. Using interviews will only highlight some of the opaque things. 

The internal validity (Bell, Bryman & Harley, 2019), whether there is a clear relationship 

between the results and the concepts developed in the research, can be discussed for this 

thesis. The focus of my research was one of many possible focuses, so in that sense there 

are many other possibilities of analysis based on the results. This is partly due to the large 

amount of data collected in interview studies (Bell, Bryman & Harley, 2019). With the 

chosen focus of trust, the large amount of data, and the repeated themes brought up in the 

interviews the connection between results and the discussed topics can be seen as close. 

This is typical for qualitative research, since the researcher spends a lot of time interacting 

with the participants, as in this study (Bell, Bryman & Harley, 2019). The internal validity is 



 

 

 

 

 

13 
 

though not as large as it could be, since ideas of transparency were rather seen as potential 

improvements based on the interviews, based on my interaction with these ideas and what 

was brought up in the interviews, while in the interview’s transparency was not discussed 

directly most often. 

External reliability (Bell, Bryman & Harley, 2019), the possibility of replicating the research 

process outlined here is relatively good. The interview guides in Appendix A shows what kind 

of questions I focused on in the interviews, while the type of documents I use exist in many 

libraries and are as mentioned in the Documents section probably looking like the ones I 

have read.  

3.4.4 Limitations 
Optimally I would have conducted interviews until each role did not come up with anything 

new, having reached "saturation" (Bell, Bryman & Harley, 2019). For librarians, I reached 

saturation while the number of interviews with library managers could have been larger. I 

also would have liked to talk to library visitors to get their perspective on some of the 

questions. As mentioned above, only some of all the opaque things in libraries in relation to 

the public can be found through interviews. Especially ethnographic studies could find other 

aspects that are opaque in interviews as well. The questions on valuation could also be 

enriched by ethnographical studies. 

 

3.6 Ethical considerations 

For someone studying something via interviews and documents, ethical considerations 

include getting informed consent, avoiding harm of participants and deception and 

respecting the boundaries of privacy (Bell, Bryman & Harley, 2019). To solve this, 

recommendations about the anonymization of data and carefulness in reporting results, 

well-thought-out questions and being open about how the interview is conducted as well as 

how the material collected will be used, published and saved was followed (Bell, Bryman & 

Harley, 2019). I also chose to not mention library names for the interviewees’ sake. 

A form was signed before the interviews by the participants to agree on being interviewed 

as well as my decisions on storing data, etc., which followed Chalmers guidelines. I sent it 

out a second time in a slightly revised form to make sure that when the report was written – 

they still agreed to participate. This also gives an indication that they accept the thesis 

despite all the eventual problems relating to translation that exist in translated works of 

research (Schembri & Jahić Jašić, 2022). 

With my focus in this thesis there is no obvious connection to ecological aspects of ethics. In 

only one of my interviews ecological aspects were brought up. Taken together, I chose not 

to focus on these aspects in my analysis either. 

 

3.7 Table of Interviews 
Interviewee Role Size of library Library 

Interviewee 1 Library manager Small to medium 

 

1 



 

 

 

 

 

14 
 

Interviewee 2 Librarian Small to medium 

 

2 

Interviewee 3 Librarian Small to medium 

 

2 

Interviewee 4 
 

Librarian 
 

Small to medium 

 

2 

Interviewee 5 
 

Researcher - 

 

- 

Interviewee 6 
 

Librarian 
 

Small to medium 

 

2 

Interviewee 7 
 

Librarian 
 

Small to medium 

 

3 

Interviewee 8 
 

Librarian 
 

Small to medium 

 

4 

Interviewee 9 
 

- 
 

- - 

Interviewee 10 - 
 

- - 

Interviewee 11 
 

Researcher 
 

- 

 

- 

Interviewee 12 
 

Researcher 
 

- 

 

- 

Interviewee 13 
 

Library manager Small to medium 

 

6 

Interviewee 14 
 

Librarian 
 

Large 7 

Interviewee 15 Librarian 
 

Large 7 

 

Table 1. Interviewees of the study and their respective role and type of library. Non-

technical = Small to medium-sized library, Technical = Large library. 

 

3.8 Table of Documents 

Document Document type Content Library 

a Library plan The libraries 
strategy, an 
overview of goals 
and values. 

2 

b Library plan The libraries 
strategy, an 
overview of goals 
and values. 

4 

c - - - 

d Purchasing policy The criteria used 
when purchasing 
books. 

6 



 

 

 

 

 

15 
 

e Substitute 
introduction 

Varied info for new 
substitute 
personnel. 

6 

f Handling of physical 
books 

Where in the library 
to place books etc. 

6 

g How to approach 
and talk to visitors 

How to approach 
and talk to visitors. 

6 

 

Table 2. The documents used in the study. 

 

 

  



 

 

 

 

 

16 
 

 

4. Results 
 

In this chapter I answer the first two research questions of the thesis: 1. Which aspects of 

Swedish small to medium sized public libraries are opaque to the public today? 2. How do 

librarians value the quality of books in purchasing, as seen in Swedish small to medium sized 

public libraries?  

I do so by first listing aspects that are opaque in libraries today and by describing them in 

detail to be able to make the purchasing process possible to understand and potentially 

criticize by the public after it has been made transparent. I also highlight aspects that are 

opaque in the libraries today, but which are not necessarily relevant for the third research 

question, and which will therefore not be discussed in the next chapter. Secondly, I will 

show how librarians' valuations of quality manifest themselves in practice in their processes 

of purchasing and in relation to other criteria. 

In this chapter I use a non-conventional way of referencing. A number indicates that the 

interview with that number in Table 1 is referenced, a letter indicates that the document 

with that letter in Table 2 is referenced. 

 

4.1 Things that are opaque to the public concerning libraries 
Library visitors have been said to be influential in the creation of library collections, which is 

central to my trust argument. In this sub-chapter I describe in more detail the way in which 

this happens and why. This is relevant to be able to prove that the user-centricity described 

in library documents (a, b, d) today, is accomplished. That is, through the process of 

purchasing and removal of books described in this sub-chapter. 

4.1.1 The typical process of purchasing 
Libraries purchase books bi-weekly (4). They do so due to the bi-weekly BTJ publication 

which informs them about book releases (4). The books are purchased from Adlibris (an 

online and physical retailer of books) or other similar services which they have contracts 

with, in general, while some libraries use other services for specific kinds of books or book 

titles like books in some different languages (2, 8, 13). The number of books purchased each 

year varies among libraries but is in the range of thousands also in small to medium sized 

libraries. The same number of books therefore need to be removed each year, in general (6). 

To be able to work according to the purchase process each time, the time needed may vary 

but one interviewee said it takes about 2 hours per week to read reviews and choose which 

books to purchase (4).  

Libraries have increasingly over time started to use “demand” as a metric for management 

of their collections, and as a metric of success, which is seen in both the interviews (2, 3, 8, 

13) and documents (a, b, d). That’s what is called user-centric collection. Books are modeled 

as demanded before or after purchase via different criteria the librarians use to think about 

demand, e.g., media appearances of a book and length of queues waiting for a book at other 

libraries (2, 8). If the book is considered to be demanded in this way, it is purchased (a, b, d, 

2, 8). 



 

 

 

 

 

17 
 

Most of the book purchases are made by a librarian by reading reviews from BTJ of 

upcoming books and talking to people, and then the librarian makes the choice of what to 

buy and not (1, 2, 3, 4, 6, 7, 8, 13). The individual librarian's preferences can be influential in 

this purchasing, since in small to medium sized libraries, the interviewed informants have 

responsibility for their specific part of a library’s purchasing and does not discuss with peers 

frequently (3, 4, 13, 14). One person is often responsible for children and young people’s 

books, while two or more other people purchase fiction and non-fiction or similar (2, 14). 

The librarian becomes more knowledgeable in their respective fields of books over time. 

Time, resources and the number of books that are seen as a good amount to purchase limits 

the number of considerations a responsible librarian can reason about for each book (6). In 

general, new books are purchased rather than old, for multiple reasons, including popularity 

at the time, and the supply of books available (d, 2, 4, 6, 8, 12, 14). 

In general, the process of purchasing wasn’t problematized (1, 2, 3, 4, 6, 7, 8, 13). When 

asked about if there could be alternative ways to do purchasing, or if the purchasing process 

used had any downsides, the same interviewees only mentioned small details or nothing. 

4.1.2 Avoiding BTJ 

A quite different and alternative way to purchase books was explained by one informant 

(14). This informant was the only one who said that BTJ does not have enough experts in 

fields such as “mångspråk”, i.e., books in other languages than Swedish which many people 

living in Sweden have as their mother tongue, and science fiction. For these books, the 

informant thought services such as Goodreads and communication with other people the 

informant knew as well as the visitors to the library were better ways to choose books. That 

was, since the books on these subjects were few in BTJ, and those books that were included 

often were similar to each other, like historical or political novels for books in other 

languages (14). The informant probably waited a bit longer to make purchases, to be able to 

read reviews of the book done by others after it had been published. This interviewee’s (14) 

views contrasted a lot with what was said in the other interviews. 

As seen in these two sections, whether one uses BTJ or not influences the collection at the 

library a lot, since the supply of books is limited in BTJ compared to all books available. Only 

one of the interviewees (14) seemed to know of these limitations to this extent, which 

means many libraries may be showing similar limitations in their collections. 

Also, the BTJ services were expensive according to the same interviewee (14). 

4.1.3 Purchase suggestions 

In most cases purchase suggestions from library visitors are accepted as seen in (b, 2, 4), and 

if not, the books can be borrowed from other libraries (b, 2, 4, 8, 13). With this option, more 

books are available to any one person independent of which library they go to. Perhaps only 

some visitors know that they can borrow from other libraries, as indicated in a few 

interviews (2 & 13). The option to make purchase suggestions increases the level of user-

centricity. 

The libraries decide whether to accept purchase suggestions based on how applicable the 

book is to a larger audience. If a book is considered too niche it is not purchased (b, 4). This 

kind of consideration must be made according to the same interviewees.  



 

 

 

 

 

18 
 

4.1.4 Removal of books 
As described above, the process of purchasing is often accomplished by one dedicated 

person, in a partly subjective and specific way. The removal of books is instead often done 

informally in a larger group based on objective criteria (d, 2, 6). The removal process thus 

does not include as many details in the choice of which book to throw away compared to 

the criteria used for purchasing most often (6). A "removal list" is made based on a few 

objective criteria each year (6). These criteria are the age of the book, number of loans and 

how worn-out the book in question is (d, 6). This is in accordance with the CREW guidelines, 

which those libraries which I talked to this about were using (2, 6, 14). With that list, a book 

is thrown away if the book in a category of non-fiction is a certain number of years old, or if 

it has not been borrowed in a certain number of years or if it is considered too dirty to place 

on the shelves again (2, 6). Also, if the number of books going into the system is more than 

the number of books that can be categorized to be on the removal list then more books 

have to be included as well, based on the number of loans or something else (6). 

Some libraries still do removal as carefully and with similar criteria as their purchasing (13, 

14), while it may be difficult to do so when books are thrown in large numbers and info 

about books needs to be read to be remembered.  

The only library I asked about what happens with the books that are taken off the shelves 

said that they are thrown away and burned (6). This was something that they didn’t want to 

be open about to the public since it may seem cruel (6). Another person likened it to how 

people want to eat meat but not see how it is killed (14). The library that burned books said 

they considered it too time-consuming and messy to sell or donate books, since then many 

books will come back because others may think it is still the library’s books (6). 

4.1.5 The invisible non-users 
The libraries have little contact with, and therefore little knowledge of, at least on the local 

level, non-visitors (2, 3, 8, 13). Mostly, activities are mentioned as what the libraries do to 

get new visitors to the library (2, 8, 13). So, new visitors are welcomed to the library in 

different ways, but those who don’t come to these activities are not taken contact with by 

the library’s librarians I have talked to. These activities can be book talks, painting, language 

cafés, digital coaching or music events (2, 8, 13). Activities are more or less alike what the 

libraries usually do. 

The librarians I have talked to have not done any survey or other kind of action to contact 

non-visitors with the purpose of better understanding why they don’t visit (2, 3, 8, 13). This 

is true for all interviewees asked about this. A few librarians had plans to do such surveys in 

a relatively small format (2, 3). 

4.1.6 Political interference in librarians’ work 
Most of the interviews show that most librarians had not experienced any political influence 

from politicians in their daily work. As such, political influence is still relatively rare. 

One librarian (13) mentioned political interference as a problem when asked an open 

question on issues that they work with and/or debate a lot in the organization today. The 

informant said a new political majority in the municipality in question had a smaller group of 

people that questioned, among other things, parts of the library’s work and how the power 

is distributed between the library and politicians. 

 



 

 

 

 

 

19 
 

4.2 Librarian’s valuation of quality  
Here, I describe the interviewees valuation of quality of books and relate it to other 

considerations in the purchasing process to show how this idea of quality reveals itself in 

practice. 

As seen previously, the interviewees said that user-centricity is good, and their priority. 

Consequently, following this is prioritized above any quality considerations. The 

interviewees, when asked about the relevance of the quality of books and their preferences 

on what people should read, said that they would rather see that people read at all, than 

care about what they read or for what purpose (4, 7, 14).  

Books on different subjects are seen as important to have (a, b, d, 2), such as economy or 

philosophy books besides criminal fiction. That is one way in which demand is not 

necessarily prioritized the most (2, 9). Also, books for the prioritized groups in 

Bibliotekslagen, such as books in the minority languages are prioritized (a, b, d, 2, 7, 8, 13, 

14) despite low levels of demand in some cases.  

BTJ has been used a lot and has been used for a long time by some of the interviewees (2, 

13). Many of the informants seem to never have worked in a way which does not use BTJ as 

the main piece of advice for purchasing (2, 3, 4, 6, 13).  

The rating given to a book in BTJ is one of many variables used to determine if a book should 

be bought, which is one kind of quality criteria considered (2, 3, 4). Other criteria for which 

book to purchase include what is seen as demanded as said, which in turn is based on 

multiple variables such as demographics in the municipality where the library is located, 

experience from previous books the library has purchased, and books about the city or 

region that the library is located in (which are always bought) etc., (a, b, d, 2, 3, 4).  



 

 

 

 

 

20 
 

 

5. Discussion 
 

In this chapter I discuss what transparency could mean for libraries over time, in terms of 

trust. Secondly, I analyze the consequences of librarians' valuations of the quality of books in 

relation to other considerations in the purchasing process. All of this is done to answer the 

third research question, 3. How can increased transparency of Swedish small to medium 

sized public libraries process of purchasing of books, and librarians value of quality in 

purchasing of books, increase the trust in and value of these libraries? 

 

5.1 Opaque to transparent – to increase trust 

In this sub-chapter, I note some nuances of the transparency I propose in this thesis, as well 

as some hinders to the same. As was seen in the introduction, transparency is one of the 

aspects influencing trust of institutions (Holmberg & Weibull, 2013). 

Making the now opaque process (of purchasing) transparent (de Fine Licht & de Fine Licht, 

2020) could be a way to increase trust in the libraries' work. From the interviews it was clear 

that most books people ask for are purchased, and that any book that is not available at the 

local library is possible to borrow from other libraries. Purchasing is made so that most 

books are purchased, and the citizens themselves influence a lot and can consciously 

influence the collections if they know that the opportunity exists. Taken together, these 

aspects could reinforce trust that libraries are neutral. With that meaning a neutrality that 

takes in multiple perspectives (to mirror the diverse group of people they serve), while also 

not having to land in consensus on specific questions regarding the library or society (Haider 

& Sundin, 2022). 

The transparency I propose is meant to be a continuation of the documents’ libraries have 

today, which describe the libraries as user-centric and neutral but that don’t explain how 

these things are achieved in practice. Through the empirical data in the Results section, the 

purchasing process becomes clear, and whether it is neutral or worth trusting is up to each 

individual person. I propose it could add trust but realize that an empirical study testing this 

solution and if it changes trust in a positive direction is needed to make this idea rigorous. 

If the critical audience (Kemper & Kolkman, 2019), the public, does not agree with the 

purchasing process, they can argue for changes and perhaps the purchasing process change. 

That is true for all kinds of critique of the collections. This kind of feedback and response 

could be seen as increasing transparency (and potentially competence) and therefore trust 

(Holmberg & Weibull, 2013). 

In case of a practically non-technical case such as with the purchasing of books in libraries, 

transparency is not as problematic as in technical cases (Kolkman, 2022). The documents 

discussing media policy or purchasing policy that libraries already have on their website are 

common for libraries to have. The addition of another level of transparency for libraries 

activities is consequently close to what can be expected of libraries, increasing or at least not 

decreasing trust in that sense (Michnik, 2018). 



 

 

 

 

 

21 
 

5.1.1 Implementing transparency 

Of course, making something transparent takes time, and resources are needed. This must 

be considered, but since knowledge of most things I think could be made transparent is 

already known to librarians, an eventual implementation of this transparency could go 

relatively quickly.  

At least according to the interviews, the librarians often agree with one another about 

purchasing decisions at each library. With other words it does not necessarily have to create 

a lot of conflict to discuss what transparency they should go for regarding the purchasing 

process, which has been seen as a risk (Albu & Flyverbom, 2019). Instead, it could be 

possible to use the “light” metaphor for transparency (Albu & Flyverbom, 2019). At the same 

time, realizing that the trust of libraries has been questioned, it might be sensitive or 

controversial for a library to be fully transparent if some group sees an opportunity to 

criticize in destructive way. At least some details will likely involve some level of conflict or 

discussion (Albu & Flyverbom, 2019). 

There will be unintended consequences of transparency as well (Ananny & Crawford, 2018), 

therefore there is a need to work with the nuances of transparency and change over time. 

 

5.2 The consequences of librarians' valuation of quality 
The different degrees of power and knowledge in different parts of an organization yields 

different types and degrees of transparency for different persons (Lee, 2021), even within 

the same “group”, such as in a group of librarians or visitors. One such example in this study 

was the critique of BTJ’s book review-services. Only one of the interviewees stated that the 

output from this service is limited, especially regarding certain categories or genres of books 

such as science fiction and “mångspråk”. This knowledge made the informant go through 

other channels to find and buy new books. Others also mentioned that they use other 

information channels as well, but did not critique the BTJ services, implying that other 

channels were only used as a complement, and that the knowledge of BTJ’s limitations was 

not known to the same extent by them. So, even for the librarians themselves, the 

limitations of the process might not be known in the first place. Therefore, it would not be 

possible for them to make this transparent to the public on their own initiative. 

As seen, libraries can have all the specialized knowledge to purchase books and know the 

collection on all subjects only to a certain depth which has a limit. It is practically impossible 

to have a deep knowledge of all books for such a small group of people.  

Overall, the quality of individual books is not prioritized above what is considered demanded 

by the visitors in the process of purchasing, regardless of what is meant by quality here, 

apart from demand. Quality comes after demand in their valuations due to the goals of user-

centricity and the practices used to accomplish this (Lee et al., 2022). Only concerning some 

books, e.g., books for prioritized groups and classics, the books can be less in demand within 

their genre but still be kept on the shelves. This relative lack of focus on quality of books in 

libraries purchases has been shown to exist for a few years now (Lindsköld & Hedemark, 

2023; Hedemark, 2020; Kungliga Biblioteket, 2021). 

Quality is sometimes valued in different ways by librarians, but the purchasing process of 

the day does not allow much room for maneuvering (Söderlind & Elf, 2014). The 

problematization (Lee & Helgesson, 2020), the answer to the question of what should be 



 

 

 

 

 

22 
 

done, being user-centric, has made other valuations than the existing one difficult to 

implement in practice for the librarians. 

The authors note that the librarians did not seem to see a contradiction between user-

centric collections and quality (Söderlind & Elf, 2014), like how the purchasing process 

wasn’t problematized in most of my interviews. At the same time my interviews indicate 

that this contradiction exists in that the focus on demand limits how much quality can be 

considered and be a variable for purchasing decisions. Therefore, other ways to incorporate 

quality and increase the quality of books people read must be found (other than a changed 

purchasing process), if one wants to improve libraries in this sense, while still maintaining 

user-centric collections. 

My interviews also showed that the resources available for the librarians are not optimal, 

and as such developing the existing practice can be difficult. A time-efficient solution is 

therefore needed. 

The librarians’ valuation of quality, being prioritized below especially demand criteria in 

purchasing decisions, is also opaque (Lee, 2021) to library visitors today, being one of the 

aspects going into the purchasing process in a way that is not necessarily highlighted even 

following the previous sub-chapter’s ambitions. The valuation of quality as seen in libraries 

today, should, following that realization, be made transparent as well, increasing the degree 

of transparency and consequently trust (Holmberg & Weibull, 2013). If the existing valuation 

of quality is critiqued by library visitors, like a critical audience (Kemper & Kolkman, 2019), 

transparency can also lead to new practices, valuations and/or processes for purchasing, 

potentially increasing the usage of libraries as well. 

In sum, the quality of books available at a library is dependent on the company BTJ’s book 

services (the books they include in their bi-weekly book review services) and/or the local 

libraries librarians’ knowledge and taste in books, which is good but by necessity is limited. 

Quality criteria need to be balanced against ideas of demand when user-centric collection 

ideals are followed. This should be made transparent to library visitors. 

  



 

 

 

 

 

23 
 

 

6. Conclusion 
 

This thesis aimed at finding out new things about libraries via the use of concepts from the 

critical algorithm field, especially transparency. Following that, I argue that both the trust in 

and the value the public can get from libraries can increase through transparency and new 

valuations of quality. 

First, the process used for purchasing books at libraries should be made transparent. That is, 

since this process shows the kind of neutrality based on diversity which the library 

represents. It would describe the collection as constructed by the users, as a user-centric 

collection. The possibility of borrowing books from other places when the books are not 

available at the visitor’s closest library should also be made transparent since this also shows 

the ambition of libraries to include many types of books, for everyone. In this way I 

contribute to critical algorithm theory, showcasing how it can be applied to libraries. 

I suggest transparency can be applied in new ways in the context of libraries, but empirical 

studies that test to implement this transparency and do a survey to test the eventual 

increase in trust in this way would be needed to justify the suggestion. 

Second, I show that what is seen as (literary, scientific etc.) quality of books in each library is 

largely based on what books BTJ includes in their bi-weekly book reviews, as well as the 

librarians’ respective expertise and the user-centric collection ideal which libraries follow 

today. I argue that this creates limited collections, which have a lot of value, but could have 

more value for library visitors. With the existing purchasing process, librarians have to 

balance quality criteria (however quality is defined) against other criteria, especially what 

books are demanded, where demand is often prioritized. This should be made transparent 

to increase trust and perhaps change the valuation of quality in libraries as well if this is 

what the public wants in response to these things becoming transparent. Ways to 

incorporate (literary, scientific etc.) quality of books through other means than purchasing in 

libraries work is needed when the purchasing of books with user-centricity as an ideal is 

used. In this way I contribute to valuations studies, showing an example of valuations in 

libraries. 

Third, through making these two things transparent, and through the feedback of library 

visitors that can start to take part in libraries work in new ways when new things have 

become transparent, the valuation of quality of books by librarians as seen in the purchasing 

process for books, can align with the library visitors view of the ideal library collections. This 

puts a critical algorithm-perspective on the debate about user-centric and collection-centric 

collections. With that perspective, the users take part in the process of creating the library 

collections in another way than what is mainly thought of in the debate. In that sense it adds 

a dimension to consider in the debate between user-centric and collection-centric 

collections. In this way I contribute to library studies. 

Moving forward, I propose three possible future studies. First and foremost, testing whether 

the kind of transparency I propose can increase trust, could be tested empirically.  



 

 

 

 

 

24 
 

Whether or not library visitors of today understand that they "shape" library collections as 

much as they do today is unknown. Is there "informed consent", is the user-centric 

collection the way people want it to be? That can be a question of another empirical study. 

Lastly, how librarians think that visitors act in the library room in contrast to how the visitors 

use it, e.g., in relation to when visitors ask questions to librarians and not and why, could be 

understood in better detail. This affects what transparency would be useful on a more 

detailed level. 

For someone that wants to continue this work, of applying critical algorithm thinking to 

libraries and have other, more or better ideas than me of what could be done, my thesis can 

hopefully serve as inspiration. If so, that following work (if it would not have been done 

without this thesis) could be the greatest contribution of this thesis, even if indirect. 

  



 

 

 

 

 

25 
 

 

7. References  
 

Albu, O. B., & Flyverbom, M. (2019). Organizational Transparency: Conceptualizations, 

Conditions, and Consequences. Business & Society, 58(2), 268–297. 

https://doi.org/10.1177/0007650316659851 

Ananny, M., & Crawford, K. (2018). Seeing without knowing: Limitations of the transparency 

ideal and its application to algorithmic accountability. New Media & Society, 20(3), 973–989. 

https://doi.org/10.1177/1461444816676645 

Bell, E., Bryman, A., & Harley, A. (2019). Business Research Methods (4th ed.). Oxford 

University Press. 

(2013:801). Bibliotekslag. Kulturdepartementet. https://www.riksdagen.se/sv/dokument-

lagar/dokument/svensk-forfattningssamling/bibliotekslag-2013801_sfs-2013-801 

Bowen, G.A. (2009). Document Analysis as a Qualitative Research Method. Qualitative 

Research Journal, 9(2), 27-40. https://doi.org/10.3316/QRJ0902027 

de Fine Licht, K., & de Fine Licht, J. (2020). ‘Artificial intelligence, transparency, and public 

decision-making: Why explanations are key when trying to produce perceived legitimacy’. AI 

& Society, 35(1), 917–926. https://doi.org/10.1007/s00146-020-00960-w 

Eriksson, T., & Clemens, A. (March 23, 2023). Ovanligt med politisk inblandning – men oron 

finns. Biblioteksbladet. https://www.biblioteksbladet.se/nyheter/ovanligt-med-politisk-

inblandning-men-oron-finns/ 

Eriksson, C., & Michnik, K. (2019). Allas bibliotek?. I U. Andersson, B. Rönnerstrand, P. 

Öhberg & A. Bergström (Eds.), Storm och stiltje (pp. 301–320). Göteborgs universitet: SOM-

institutet. https://www.gu.se/sites/default/files/2020-04/301-

320%20Eriksson%20o%20Michnik.pdf  

Haider, J., & Sundin, O. (2022). Paradoxes of Media and Information Literacy: The Crisis of 

Information. Routledge. https://doi.org/10.4324/9781003163237. 

Hedemark, Å. (2020). Constructing the Literate Child: An Analysis of Swedish Literature 

Policy. Library & Information History, 36(2), 73–88. 

Holmberg, S., & Weibull, L. (2013). Det viktiga institutionsförtroendet. L. Weibull, H. 

Oscarsson & A. Bergström (Eds.), Vägskäl (pp. 225–245). Göteborgs universitet: SOM-

institutet. 

Höglund, L. (2017). Biblioteken fortsatt välanvända med höga betyg. I U. Andersson, J. 

Ohlsson, H. Oscarsson & M. Oskarsson (Eds.), Larmar och gör sig till (pp. 329–341). 

Göteborgs universitet: SOM-institutet. 

Höglund, P. (2023, May 11). 472: Jimmie och Fröken Skamlös [Podcast episode]. In Det 

Politiska Spelet. Sveriges Radio. https://sverigesradio.se/avsnitt/472-jimmie-och-froken-

skamlos-det-politiska-spelet 

https://doi.org/10.1177/0007650316659851
https://doi.org/10.1177/1461444816676645
https://www.riksdagen.se/sv/dokument-lagar/dokument/svensk-forfattningssamling/bibliotekslag-2013801_sfs-2013-801
https://www.riksdagen.se/sv/dokument-lagar/dokument/svensk-forfattningssamling/bibliotekslag-2013801_sfs-2013-801
https://doi.org/10.3316/QRJ0902027
https://doi.org/10.1007/s00146-020-00960-w
https://www.biblioteksbladet.se/nyheter/ovanligt-med-politisk-inblandning-men-oron-finns/
https://www.biblioteksbladet.se/nyheter/ovanligt-med-politisk-inblandning-men-oron-finns/
https://www.gu.se/sites/default/files/2020-04/301-320%20Eriksson%20o%20Michnik.pdf
https://www.gu.se/sites/default/files/2020-04/301-320%20Eriksson%20o%20Michnik.pdf
https://doi.org/10.4324/9781003163237
https://sverigesradio.se/avsnitt/472-jimmie-och-froken-skamlos-det-politiska-spelet
https://sverigesradio.se/avsnitt/472-jimmie-och-froken-skamlos-det-politiska-spelet


 

 

 

 

 

26 
 

Kemper, J., & Kolkman, D. (2019). Transparent to whom? No algorithmic accountability 

without a critical audience. Information, Communication & Society, 22(14), 2081–2096. 

https://doi.org/10.1080/1369118X.2018.1477967  

Kolkman, D. (2022). The (in)credibility of algorithmic models to non-experts. Information, 

Communication & Society, 25(1), 93–109. https://doi.org/10.1080/1369118X.2020.1761860  

Kungliga biblioteket. (2021). Trender i planerna 2020 - Kommunala biblioteksplaner antagna 

2020. https://urn.kb.se/resolve?urn=urn:nbn:se:kb:publ-99 

Larsson, S., & Heintz, F. (2020). Transparency in artificial intelligence. Internet Policy Review, 

9(2), 1–16. https://doi.org/10.14763/2020.2.1469 

Lee, F. (2021). Enacting the Pandemic. Science & Technology Studies, 34(1), 65–90. 

https://doi.org/10.23987/sts.75323 

Lee, F., Dussauge, I., & Jülich, S. (2022). Against Bioethicalization: Value Disjunctures in the 

History of Fetal Research. SocArXiv. https://doi.org/10.31235/osf.io/57ydj 

Lee, F., & Helgesson, C.-F. (2020). Styles of Valuation: Algorithms and Agency in High-

throughput Bioscience. Science, Technology, & Human Values, 45(4), 659–685. 

https://doi.org/10.1177/0162243919866898 

Lindsköld, L., & Hedemark, Å. (2023). Changes in the governance of the reading subject: 

Swedish reading policy, c.1949–1984. International Journal of Cultural Policy. 

https://doi.org/10.1080/10286632.2023.2187052  

Michnik, K. (2018). Samhällets allt-i-allo?: Om folkbibliotekens sociala legitimitet. [Phd 

dissertation, Bibliotekshögskolan, Högskolan i Borås]. https://hb.diva-

portal.org/smash/get/diva2:1191134/FULLTEXT01.pdf 

Rivano Eckerdal, J. (2017). Libraries, democracy, information literacy, and citizenship: An 

agonistic reading of central library and information studies’ concepts. Journal of 

Documentation, 73(5), 1010-1033. https://doi.org/10.1108/JD-12-2016-0152 

Schembri, N., & Jahić Jašić, A. (2022). Ethical issues in multilingual research situations: a 

focus on interview-based research. Research Ethics, 18(3), 210–225. 

https://doi.org/10.1177/17470161221085857 

Segerberg, T. (2020). Det motståndskraftiga förtroendet. Kriser och institutionellt förtroende 

i Sverige 1986-2017. (Report No. 83). Institutionen för journalistik, medier och 

kommunikation, Göteborgs Universitet. 

https://medarbetarportalen.gu.se/digitalAssets/1769/1769254_nr-83-hela-rapporten.pdf 

SVT. (2022). Dödshot efter sagostund med dragqueens på biblioteket i Malmö. 

https://www.svt.se/nyheter/lokalt/skane/dodshot-efter-sagostund-pa-biblioteket-i-malmo. 

Söderlind, Å., & Elf, G. (2014). Vi arbetar i medborgarnas tjänst: en kartläggning av 

mediestrategiskt arbete på bibliotek utifrån normerande dokument. Svensk 

biblioteksförening. http://hb.diva-portal.org/smash/get/diva2:884055/FULLTEXT01.pdf 

  

https://doi.org/10.1080/1369118X.2018.1477967
https://doi.org/10.1080/1369118X.2020.1761860
https://urn.kb.se/resolve?urn=urn:nbn:se:kb:publ-99
https://doi.org/10.14763/2020.2.1469
https://doi.org/10.23987/sts.75323
https://doi.org/10.31235/osf.io/57ydj
https://doi.org/10.1177/0162243919866898
https://doi.org/10.1080/10286632.2023.2187052
https://hb.diva-portal.org/smash/get/diva2:1191134/FULLTEXT01.pdf
https://hb.diva-portal.org/smash/get/diva2:1191134/FULLTEXT01.pdf
https://doi.org/10.1108/JD-12-2016-0152
https://doi.org/10.1177/17470161221085857
https://medarbetarportalen.gu.se/digitalAssets/1769/1769254_nr-83-hela-rapporten.pdf
https://www.svt.se/nyheter/lokalt/skane/dodshot-efter-sagostund-pa-biblioteket-i-malmo
http://hb.diva-portal.org/smash/get/diva2:884055/FULLTEXT01.pdf


 

 

 

 

 

27 
 

 

Appendix A: 
 

Interview guides: 
Overall, I asked questions about the topics mentioned in the interview guide but did 

reformulate questions, ask follow-up questions, and so on to go deeper into each issue that 

the interviewee responded to with engagement out of these topics. First, I have the guide 

used for the interviews with small to medium sized libraries librarians, in Swedish, and 

secondly an English translation of the same. Following that is the interview guide used for 

interviews with librarians from the large libraries. The headlines used were not used by me 

but are included for readability. 

Interview guide A: 
Presentation of myself and the purpose of the thesis and the interview. I ask if they have 

read the consent form including the Chalmers guidelines. If not, I ask them to do so as soon 

as possible after the interview and I explain what is in the form shortly. I then ask if I can 

record and/or take notes during the interview. 

Start of interview: 

Vad har du för roll på biblioteket? Vad jobbar du med? 

Purchasing 

Hur gör ni inköp? 

Hur vet ni vad folk vill ha (för böcker)? 

Hur väljs mängden böcker ni ska köpa in? 

Alternativa sätt göra inköp, finns det något ni vill utveckla/förändra? 

Hur gjordes inköp förr? 

Hur gör ni gallring? 

Evaluating their work 

Hur utvärderar ni er verksamhet?  

Vad är viktiga frågor för er på biblioteket idag? 

Non-users 

Hur lockar ni nya besökare och är i kontakt med icke-användare? 

Exceptions to the typical purchasing process 

Vilken litteratur har “allmänt intresse” och inte?  

Vilka böcker köper ni inte in? Exempel. Råder olika åsikter här? 

Documents 



 

 

 

 

 

28 
 

Var finns den så kallade inköpspolicyn, vilka läser den, besökare?  

Ending, rounding up 

Något mer du vill ha sagt om det vi pratat om, eller något annat? 

Någon mer jag kan intervjua? 

 

English translation:  

Start of interview: 

What role do you have in the library? What do you do in this role? 

Purchasing 

How do you purchase books? 

How do you know what books are demanded? 

How do you choose the number of books to purchase? 

Are there any alternative ways to do purchasing? Are there alternative ways you 

would want to do purchasing? Pros and cons of the purchasing process used now? 

How was purchasing done earlier? 

How do you do removal of books? 

Evaluating their work 

How do you evaluate your work at the library? 

What are some important questions for you at the library today? 

Non-users 

How do you work to get new library visitors and how are you in contact with non-

users? 

Exceptions to the typical purchasing process 

What literature is of general interest and not? 

What books do you not purchase? Example. Are there differing opinions about this? 

Documents 

Where is the so-called purchasing policy, and can I read it? Who reads it? Visitors or 

personnel? 

Ending, rounding up 

Anything else you want to bring up, concerning what we have already talked about 

or something else? 



 

 

 

 

 

29 
 

Do you know anyone else who I can interview? 

 

Interview guide B:  
This guide was made for the last two interviews, with what I call “large” and “technical” 

library’s personnel. Here, I only refer to the questions since the introduction was the same 

as in the other interviews. I also go between Swedish and English since the guide is shorter. 

 

Vad skiljer att jobba med flytande bestånd och inte, vad är viktigast för personal och 

besökare? 

What’s the difference between working with floating collections and not, for library 

personnel and visitors? 

 

Vad skiljer att jobba med centrala inköp och inte, vilka åsikter och fakta finns?  

What’s the difference between working with floating collections and not, what 

opinions and facts are there? 

 

Hur inkorporeras icke-besökare i era tankar om olika saker?  

How are non-users incorporated in your thoughts on different things? 

 

Hur görs inköp där? 

How is purchasing done there? 

 

Vilka övriga biblioteksfrågor är viktiga för er? 

What other questions are important for you at the library? 

 

Vilka resurser eller dokument finns som jag skulle kunna läsa om detta? 

What resources and documents are there where I can read about these topics? 

 

 



   

 

 

   

  



 

 

  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

DEPARTMENT OF TECHNOLOGY MANAGEMENT AND ECONOMICS 

DIVISION OF SCIENCE, TECHNOLOGY AND SOCIETY 
CHALMERS UNIVERSITY OF TECHNOLOGY  

Gothenburg, Sweden  

www.chalmers.se 


	1. Introduction
	1.1 Swedish libraries
	1.1.1 Libraries and trust
	1.1.2 Librarians valuation of the quality of books

	1.2 Aim of the thesis
	1.3 Contribution and research questions
	1.4 Delimitations
	1.5 Structure of the thesis

	2. Previous Research
	2.1 Trust and transparency
	2.2 Valuations

	3. Methodology
	3.1 Research design
	3.1.1 Qualitative research
	3.1.2 Abductive design
	3.1.3 Multi-methods

	3.2 Data collection, Interviews and Documents
	3.2.2 Documents

	3.3 Data Analysis
	3.3.1 Translation

	3.4 Triangulation, validity, reliability, limitations
	3.4.1 Generalization
	3.4.2 Triangulation
	3.4.3 Validity and reliability
	3.4.4 Limitations

	3.6 Ethical considerations
	3.7 Table of Interviews
	3.8 Table of Documents

	4. Results
	4.1 Things that are opaque to the public concerning libraries
	4.1.1 The typical process of purchasing
	4.1.2 Avoiding BTJ
	4.1.3 Purchase suggestions
	4.1.4 Removal of books
	4.1.5 The invisible non-users
	4.1.6 Political interference in librarians’ work

	4.2 Librarian’s valuation of quality

	5. Discussion
	5.1 Opaque to transparent – to increase trust
	5.1.1 Implementing transparency

	5.2 The consequences of librarians' valuation of quality

	6. Conclusion
	7. References
	Appendix A:
	Interview guides:
	Interview guide A:
	Interview guide B: