In this literature review, we examined publications exploring the information behaviour of public service employees, also known as street-level bureaucrats. Street-level bureaucracy is an established concept in public administration research (Chang & Brewer, 2023). It refers to those public employees who interact directly with individual members of the public and have some degree of discretion in how they enforce the regulations, laws and policies in their daily tasks (Lipsky, 2010). While we do not discuss the concept itself in detail in this paper, we used it as a criterion for selecting the relevant literature. Specifically, we focused on studies that describe the information behaviour of public service employees who interact directly with individuals and have discretion over how they implement policies in response to individual requests. Asylum caseworkers, who, as mentioned in the introduction, require country of origin information, are an example of street-level bureaucrats (Severijns, 2019). Our objective was to gain insights into their information behaviour.

Although peoples’ professional lives are a common context for studying information behaviour (Case & Given, 2016c), relatively little has been published on the information behaviour of asylum caseworkers. Therefore, in our literature review, we focus on street-level bureaucrats in general.

We analysed studies on the information behaviour of government employees with similar executive duties, in other words, those working at a municipality who handle citizen applications, for example. Consequently, our findings are not only relevant to asylum caseworkers, but also to a broader group of street-level bureaucrats.

Information behaviour refers to the way people interact with information, particularly how they seek and use it (Greifeneder & Schlebbe, 2022). This includes both actively searching for information and accidentally encountering it, as well as behaviours such as avoiding information. The broader context in which information behaviour occurs is, amongst others, partly determined by factors such as time and emotions (Case & Given, 2016a).

Information behaviour research is an established discipline within library and information science. However, using information behaviour models to explain decision-making processes within administrative law, more specifically, within the asylum process, is an area which, to our knowledge, is under-researched. Information behaviour research is thus a promising, yet extensive, field of study, partly due to the large number of theories and models it encompasses.

In this literature review, we focus on person and context-oriented models and empirical studies of information behaviour, rather than system-oriented approaches (Case & Given, 2016b), and we do not address underlying theories (Fisher et al., 2005). Information behaviour models are simplified representations of the complex processes, practices and activities involved in information behaviour. These models are usually presented in graphical and textual formats (Savolainen, 2019). Preferably, they are based on empirical research findings and/or the analysis of previous empirical and theoretical work (Ford, 2015; Savolainen, 2021).

In recent decades, several information behaviour models have been developed in the field of library and information sciences, some of which have been assembled into meta-models (Shenton & Hay-Gibson, 2012). These meta-models are higher-order models based on or derived from existing models. We hypothesise that they contain elements that may potentially explain street-level bureaucrats' information behaviour. Alongside meta-models, studies examining information behaviour in the workplace help to explain how street-level bureaucrats like asylum caseworkers interact with information.

The main focus of our study was to examine how decision-makers working in executive government agencies collect and use information to inform their decisions.

We divided this main question into the following three sub-questions:

We used these questions to develop a theoretical framework which highlights the complexity of information behaviour and contributes to a more complete understanding of decision-making, focusing specifically on information gathering and use during this process. The framework is also intended to facilitate interdisciplinary research into information behaviour within legal decision-making processes, with the aim of improving the quality and efficiency of legal decision-making.

We addressed our research questions by conducting an exploratory, systematic and transparent literature review. By analysing theoretical as well as empirical studies on the topic, we charted the breadth of the research field (Munn et al., 2018, 2022) to gain a thorough understanding of the field and to identify its key characteristics (Arksey & O’Malley, 2005; Munn et al., 2022). The results have enabled us to reflect on the information behaviour of decision-makers in general and of those involved in asylum procedures in particular.

In the following sections, we first describe our methods in detail before presenting the results: a new conceptual framework. In the discussion, we use this framework to reflect on existing knowledge regarding the information behaviour of asylum caseworkers. Finally, we conclude with a summary and general considerations on how information behaviour can enrich our understanding of legal decision-making processes.

We took a two-phase approach, as outlined below. First, we searched for publications providing an overview of meta-models. Second, we looked for specific studies on the information behaviour of government employees. Figure 1 summarises this process and presents it as an adapted PRISMA flow diagram (Page et al., 2021).

In the first phase, we identified relevant overview publications on information behaviour. We searched for books via RUQuest (Radboud University Library database) and on publisher platforms, such as Oxford Handbooks, Routledge Handbooks, and Cambridge Core. We also searched for scholarly articles and conference papers in Web of Science (WoS) and Library, Information Science & Technology Abstracts (LISTA). Our search strategies included combinations of search terms, citation topics (WoS) and thesaurus terms (LISTA), supplemented by searches in leading journals. The full details of the WoS search strategy and the complete list of journals can be found in Appendix 1 (Tables 1A and 1B).

These searches yielded 590 references, of which 425 remained after deduplication. Based on criteria including publication type, accessibility, study design and relevance (see Appendix 2, table 1), we selected 34 publications. In the interest of maintaining a narrower focus, some well-known works were excluded from the study, including those by Kuhlthau (1991). However, these works are indirectly represented through the selected overview publications. For example, Wilson (1999) incorporates the results of Kuhlthau’s research in his overview publication. We assessed the quality of the 34 publications using a seven-question binary scoring system (see Appendix 2, Table 2). Only those that achieved at least five affirmative responses were retained. This resulted in 21 publications.

In the second phase, we extracted all referenced models from the 21 overview publications, identifying over 100 models in total. Approximately one third of these models originated in a work context, but only three were potentially relevant to street-level bureaucrats. To address this gap, we conducted additional searches focusing specifically on models related to information behaviour in executive government settings. These supplementary searches included structured searches queries in WoS, LISTA and the International Bibliography of the Social Sciences (IBSS), as well as in relevant legal journals not indexed in these databases. The search strategy and journal lists can be found in Appendix 1 (Tables 2A and 2B).

The second phase of our search process yielded 651 references. After deduplication, 564 remained. Using a broad definition of executive government, such as migration services and municipal departments, we selected 10 relevant publications. The selection criteria are outlined in Appendix 2 (Table 3).

To ensure reliability, the second author reselected a random sample of 10% of references from both phases. Any discrepancies were discussed and resolved collaboratively. A total of 34 publications were included in the final analysis (see Appendix 3).

For our qualitative analysis, we first examined the overview publications from phase 1, followed by the (empirical) studies from phase 2. We then checked and merged the codes, grouping them into three clusters. We explain this analysis process in more detail below, and Figure 2 provides a summary.

First, we extracted comparable meta-models from the selected publications. As the publications described different types of meta-models such as conceptual, process-based, or mathematical models, which are difficult to compare, we focused on process-based meta-models. We coded nine process-based models in order to identify their elements (see Appendix 4) and developed a summary of common elements. These were then validated by comparing them with the 12 publications without process-based meta-models (the control set). This consultation is shown as a dotted line in Figure 2.

In the second phase, we examined the ten specific studies on the information behaviour of government employees. We also incorporated the three models from phase 1 into our analysis. Moreover, we extracted elements not mentioned in phase 1.

In the final step, we combined all elements from phase 1 and 2 to form three content clusters. To avoid a possible one-sided or distorted view by the main author, this classification was discussed with the second author and adjusted where necessary. We describe the final version of the three clusters in the next section.

The first component activities and behaviours contains the practical actions aimed at seeking and using information. The idea behind this component is that it is manifested in observable behaviours which can include formal or informal searching (Whetsell et al., 2021) or passive or active searching (Wilson, 1999). Moreover, we identified the use of information to complete tasks (Byström & Järvelin, 1995) and communication about information (Robson & Robinson, 2013) as activities.

We compared our results with Zhong, Han and Hansen's (2023) systematic review for verification. They identified the following activities: information seeking; information sharing; information use; information creation; information management; and information avoidance. Their activities align with those identified in our analysis. Additionally, our findings indicate that these activities are carried out by both individuals and groups.

Regarding the term information use, we note that the authors of the analysed process-based models and specific studies interpret this term differently. To clarify, we consulted a publication without a process-based model, namely that of Greifeneder and Schlebbe (2022). In their conceptual model, they consider information use' to be an umbrella term for various forms of interaction with information, as well as an activity that occurs after the information has been found. This is also true when not using information. We adopted this umbrella term, considering information sharing to be a form of information use, for example.

The second component focuses on the various types of information users and their characteristics within a given environment. Information users can be individuals (Agarwal, 2022), intermediaries and groups of people (Prekop, 2002), organisations (Choo et al., 2008) or machines (Cantarelli et al., 2023). Users perform information behaviours. They have different characteristics and act within a certain environment that influences them. For our literature review, we selected publications that focused on users rather than systems, thus it is logical that users form one of the components.

When discussing the first component (activities and behaviours), we noted that information seeking and use can occur both individually and collaboratively. Therefore, users (the second component) may have shared information needs (Prekop, 2002) and collaborate with colleagues (Byström & Järvelin, 1995; Irvine-Smith, 2019). Additionally, users are influenced by organisational culture (Katopol, 2007).

By individual information behaviour, we describe how individuals seek and use information. According to Zhong, Han and Hansen (2023), studying information behaviour involves paying attention to individual cognition. In their textbook, Case and Given (2016a), explain that information seeking is usually studied as the work of one person at a time, but they also emphasise that people often work in groups and actively share information with each other. They see collaboration as a spectrum of solo and group activities.

Collaborative information behaviour is mentioned in several publications. For example, in his meta-synthesis, Ndumbaro (2023) analyses models of collaborative information behaviour and concludes that these offer an alternative to modelling human information behaviour from the group perspective.

The third component encompasses the characteristics of information, its sources, and the technologies used to find and access it. It is important to note that the terms information and information source are not synonymous, yet they are occasionally used interchangeably.

Shenton and Hay-Gibsen (2012), who categorised various meta-models, emphasise that there has been virtually no agreement on the definition of the term information for years. They point to the International Encyclopaedia of Library and Information Science, which suggests that it may be the least precise term used in the library and information sector. According to Bates, who describes seven categories of definitions, this lack of uniformity in the definition can be explained by the fact that the phenomenon has been studied by scholars from a variety of different disciplines (Bates, 2010).

However, it is generally accepted that data and information are not synonymous. In library and information science, information refers to data that has been presented in such a way that meaning can be assigned to it (Case & Given, 2016a). For instance, the number 22,400,000 is not informative in itself; it only becomes information when used in a report about people in Yemen who depend on humanitarian aid. This information is then represented as the number 22.4 million, in a document (physical or digital): for example, the Dutch Ministry of Foreign Affairs, General Official Report Yemen, September 2023. This general report is also the source of information. Incidentally, the report refers to UN aid agencies as the primary source.

How data and information are used depends on whether people can access them. The articles analysed in this literature review demonstrate that cognitive, practical, and structural forms of access influence the use of information in professional and public-sector settings. For example, Leckie et al. (1996) show that professionals only use information of which they are aware and which they can easily access. Their awareness of databases, colleagues, literature, and informal networks determines their choice of sources, making practical accessibility, knowing where information is and being able to retrieve it, central to information-seeking. Cantarelli et al. (2023) emphasise that the ease with which information can be recalled or retrieved influences decision-making. Finally, Whetsell et al. (2021) demonstrate that formal structures, such as hierarchies, departmental boundaries and permission pathways, determine who can approach whom for information. This structural access strongly influences the use of information, as employees tend to rely on sources embedded in the organisational channels available to them.

Together, the three components form a unit that encompasses the interaction between users, activities, and information in employee information behaviour. They are interrelated, but do not influence each other in a single direction. A change to an element in one component can thus affect elements in the other components. For instance, a change in the supply of information (component 3) can impact information behaviour (component 1) which can then affect both organisational culture (component 2) and the available sources of information, particularly if colleagues are also considered as sources of information. We have expressed the close relationship between the three components in Figure 3, the triad of information behaviour. We chose this term because a triad refers to a group of three things that belong together.

Research in the field of information behaviour typically involves a variety of methodologies, data collection methods, and analysis techniques (Case, p. 221). Depending on the research question, the chosen methodology and method, the framework can serve as a useful guide. It outlines the key areas of a study, particularly at the outset of a study, and serves as a map for the relatively broad topic of information behaviour among street-level bureaucrats. Frameworks like ours can also provide a template for the initial stages of data collection and analysis during and after fieldwork (Miles et al., 2020). For instance, it could help to define key topics for interviews and assist in the design of interview guides. During the data extraction and analysis phase, the framework can be used to develop a codebook by guiding the selection of relevant code categories or by grouping codes into broader patterns or themes.

We intended to test our new framework by conducting interviews and observations with Dutch asylum caseworkers, and we therefore sought permission from the Dutch Migration Service (IND). Although permission was not granted at this stage, the framework can be validated through future field research. For now, we are applying it to our existing knowledge of country of origin information in the asylum procedure, drawing on a comprehensive literature review published earlier. For this previous literature review, publications were searched for in five databases and additional searches were conducted in journal archives and on the websites of organisations involved in establishing guidelines for the use of country of origin information (Fötsch, 2022). This earlier comprehensive literature review of academic and grey literature revealed some elements of the asylum caseworkers’ information behaviour.

We consider these elements and additional recent information alongside the theoretical framework with its three core components: activities, users, and information. We map each of these components to the context of Dutch asylum caseworkers and provide an outline of what is known about their information behaviour related to country of origin information.

The first component of the framework comprises the activities and behaviours of individuals. In the context of asylum casework, one such activity is searching for information on countries of origin. Internal work instructions provide guidance for this task. One instruction, for example, explains how to search and use country of origin information, allowing caseworkers to either consult colleagues from the Team Research and Expertise Country and Language (TOELT) or carry out searches independently (IND, 2020).

Another key activity is using country of origin information at important stages of the asylum procedure, such as preparing for hearings or examining applicants’ statements (IND, 2020). The country of origin information is essential for assessing claims relating to identity, nationality, origin, and fear of persecution, harm or becoming a victim of random violence upon return (IND, 2020). However, final decisions may lack clear, traceable references to sources, which complicates judicial review. In one case, involving an asylum seeker from Sierra Leone, an outdated (11-year-old) source was used, and the judge ruled that more recent country of origin information could be expected to be used (Dutch District Court, 21 August 2024, NL23.5154). In another case involving an asylum seeker from Libya, the absence of any referenced country of origin information was criticized (Dutch District Court, 12 August 2024, ECLI:NL:RBDHA:2024:127642024). In a later interim ruling on an appeal by a Syrian asylum seeker, the District Court ordered the Minister for Asylum and Migration to include the General Official Report on Syria 2025 in the case file. Withholding this information would be contrary to EU law (Dutch District Court, 5 June 2025, ECLI:NL:RBDHA:2025:9840).

Beyond individual efforts, the search for and use of country of origin information often involves collaboration. Task division across roles, such as between hearing officers and decision-makers, means that multiple staff members may engage with the same information. Collaboration can also take place between asylum caseworkers and country of origin experts. In his study of the Austrian Migration Service, Van der Kist (2024) refers to such collaboration as tinkering: a process in which experts gather available data and adapt it to meet the specific needs of caseworkers.

The second component of the framework consists of the users and their environment. In the context of asylum caseworkers at the IND, the users are the caseworkers themselves. Little is known about their characteristics, apart from the significant high number of new staff recruited between 2018 and 2022. Many of the newly appointed staff begin their roles without sufficient experience (Maasland et al., 2023), or formal training, especially during periods of increased asylum applications (Severijns, 2019).

In addition to knowledge and experience, information needs form an important characteristic of asylum caseworkers: both formal and informal needs play a role. On the formal side, legal requirements mandate the use of country of origin information when examining (article 10 (3) Directive 2013/32EU and article 34(2) Regulation EU 2024/1348) and assessing (article 4(3) Directive 2011/95/EU and Regulation EU 2024/1347) asylum requests. Informal needs are less well documented, but Severijns (2019) observes that Dutch asylum caseworkers often experience uncertainties about the facts presented during hearings. This may, in turn, lead to additional or different information needs related to the search for and use of country of origin information.

Severijns' (2019) research also found that internal norms, referred to as an 'oral tradition of practice', shape the working environment within IND. These norms are considered part of the broader organisational or information culture and may influence the information behaviour of caseworkers. Beyond the recognition of these internal practices, however, little is known about the organisational or information culture within the IND. In contrast, a study of the Swiss migration system entitled Getting in line with the office revealed that institutional socialisation causes decision-makers to adopt a shared mindset that ultimately influences their decisions (Affolter, 2021).

The search for and use of country information is partly formalised in European legislation, such as Directive EU 2013/32, which is to be replaced by Asylum Standards Regulation EU 2024/1347, and Directive EU 2011/95, which is to be replaced by Procedure Regulation EU 2024/1348. The European Union Agency for Asylum (EUAA, 2023), is one of several sources that provide more guidelines on the use of country of origin information.

The extent to which legislation and the guidelines of the European Union Agency for Asylum affect the practice of searching for and using country information is not entirely clear. A study in Germany shows that caseworkers often adopt pragmatic approaches, balancing formal requirements with local constraints, and relying on both official and informal norms (Schittenhelm & Schneider, 2017). In the Dutch context, Severijns states that legal frameworks create both clarity and uncertainty among Dutch asylum caseworkers (Severijns, 2019), and this may influence their behaviour in seeking information.

The final component is the information and the ability to access it. For asylum caseworkers, country of origin information is particularly important to support their decision-making process. The characteristics of this information can influence how it is used which seems to be the case regarding the use of country of origin information. Caseworkers of other migration services consider government information to be more reliable than information coming from other sources (European Migration Network, 2015; Tsangarides, 2009). In the Netherlands, government information can include general, thematic or individual reports. Additionally, Dutch asylum caseworkers use reports from other organisations, news reports, websites, and social media (Adviescommissie voor Vreemdelingenzaken, 2020).

Within the IND, TOELT is a special department responsible for country of origin information. This department comprises four units: 'Country', 'Language', 'Regional Information Centre' (RIC) and 'Interpreter Coordinators' (TOELT, 2020). The 'Country' and 'RIC' units both use and supply information to caseworkers, who can request support (IND, 2020) or search for country of origin information independently.

The TOELT also prepares requests for individual reports, which are investigations by the Ministry of Foreign Affairs into asylum-related affairs (Ministerie van Buitenlandse Zaken, 2022). Asylum caseworkers cannot initiate or conduct these investigations themselves. The same applies to general reports: while TOELT staff can assist with drafting the Terms of Reference, the research and fact-finding are conducted externally (Ministerie van Buitenlandse Zaken, 2023).

Due to the limited publicly available information regarding access to country of origin information, we cannot comment on its impact on the information behaviour of asylum caseworkers. However, several reports have highlighted concerns regarding information channels and time constraints. For example, a 2022 Justice and Security report notes that information is fragmented and difficult to locate, which risks staff missing key updates. Time constraints can also discourage the staff from consulting experts such as TOELT (Inspectie Justitie en Veiligheid, 2022). A 2023 report notes that the growing volume of information requires more time and effort to process, leading to increased trade-offs in decision-making (Buimer & Maasland, 2024; Maasland et al., 2023). This is consistent with our earlier acknowledgement that access to information influences how it is used, as emphasised by Leckie and Cantarelli, for example.

Beyond information channels and time constraints, we expect emerging technologies to have a significant impact on access to country of origin information. As these technologies become more accessible, reliance on them is likely to grow (Meyer, 2016), conferences (Ministerie van Justitie en Veiligheid, 2023), and reports confirm this trend. For example, the IND 2024 annual report mentions exploring Artificial Intelligence (AI) (Inspectie Justitie en Veiligheid, 2022), while the 2025 plan states thatAI will be implemented in core processes (IND, 2024). However, their plans remain unclear on details such as the type of AI to be used and how it will be implemented. A briefing published by the European Parliamentary Research Service provides more transparency on this issue. It becomes clear in this briefing, for example, that the IND uses a text mining tool (Case Matcher) to identify applications made on similar grounds (Dumbrava, 2025).

We assume that AI technology is also widely used when it comes to country of origin information. We base our assumption on the fact the ACUTE research consortium is trying to get an overview of existing AI activities in the field of country of origin information and to provide guidelines on the use of AI (ACCORD, 2025). Furthermore, the international scientific literature highlights examples of the use of AI when researching country of origin information. For instance, Kinchin and Mougouie (2022) observe that machine learning can assist in identifying outdated, inaccurate or prejudiced sources. They also recommend incorporating humanitarian data analysis, such as satellite imagery and socio-political data, to develop consistent and current country databases. Similarly, Kanoulas (2025) explores the use of large language models, noting their ability to quickly translate and summarise texts containing country of origin information. However, Kanoulas also warns of the risks, as large language models may be trained using biased or outdated data.

While this literature review offers insights into the information behaviour of street-level bureaucrats such as asylum caseworkers, we note two limitations. Firstly, we did not include all publications on this topic, as we conducted a delimited search strategy and applied selection criteria. Secondly, we would like to note, as mentioned earlier in the introduction, that the components and their elements were identified via a comprehensive literature review and not by empirical research. Direct observation or interviews with asylum caseworkers could not be conducted, as permission from the IND was repeatedly requested but not granted. The refusal was attributed to time constraints and was not accompanied by substantive justification. This aligns with broader findings that migration authorities tend to operate with limited transparency (Rosset & Achermann, 2019) and that access for empirical research is often unpredictable and opaque (Lindberg & Borrelli, 2019).

Despite these constraints, the framework provides a basis for future empirical research on the information behaviour of street-level bureaucrats, including asylum caseworkers. As noted in Section 4.1, it can guide both data collection and analysis. A valuable direction for further study is the relationship between the level of freedom to act (discretion), a typical feature of street-level bureaucracy, and information behaviour. For instance, when faced with time constraints, Dutch asylum caseworkers may tend to rely on familiar or easily accessible sources rather than those that are more difficult to locate (Inspectie Justitie en Veiligheid, 2022). A key question is whether they actually have the discretion to make such choices.