Academic Libraries as Partners in Data Literacy Education—An Explorative Case Study

Abstract

The concept of the ‘teaching library’, which originated in the Anglo-American world, describes all activities of libraries in the field of promoting information, media, and data literacy, as well as other skills in dealing with analog and digital media. Although data literacy is explicitly mentioned in this definition, many training courses in academic libraries seem to focus more on promoting library use, information, and media literacy. Given that the creation of data management plans, along with the indexing, storage, and reuse of research data, have become standard elements of the research process, this article discusses the growing importance of academic libraries in teaching data literacy. It presents a modular course framework, developed in exchange with the university library, as a reusable model for data literacy education. The primary objective is to introduce this framework and illustrate its application; preliminary, exploratory insights from a self-assessment survey are provided to support this presentation. The limited participant count in the pre- and post-evaluations restricts the statistical generalizability of the findings but provides a solid empirical impression of the effectiveness of the course format. Results indicate substantial learning progress in fields where academic libraries have proven expertise. The main conclusion is that such library-integrated interdisciplinary courses provide a valuable framework for data literacy education and highlight strategic areas for library involvement.

1. Introduction

Dealing and engaging with digital data is part of everyday research and course life for all students. Students from all faculties use and generate data and are repeatedly confronted with questions of data management, data evaluation, and interpretation, as well as ethical data handling and data protection. However, students often lack sufficient skills and basic knowledge, as well as higher-level practical data literacy (Bandtel et al., 2021; Oguguo et al., 2020). Data literacy enables students to analyze and interpret data critically and to make informed decisions. This process promotes critical thinking, problem-solving skills, and the ability to formulate evidence-based arguments (Heidrich et al., 2018; Schüller & Busch, 2019; Schüller et al., 2019).

To systematically anchor data literacy and thus the appropriate and ethical handling of data in the curricula of the degree programs at TH Köln, the Data Literacy Initiative (DaLI1) was launched as a research project in 2020. This paper presents a case study of the resulting modular course framework. The Data Literacy Basic Course, developed by DaLI, was designed to elevate a highly heterogeneous student body to a uniform baseline level of data literacy. Its modular structure, derived from a dedicated competence model, provides a scalable template for library-involved data literacy education. In the Data Literacy Basic Course, offered at TH Köln on an interdisciplinary, cross-semester basis since 2021, students from all faculties can acquire basic data skills. They are sensitized to the relevance of data, recognize a wide range of possible applications, and learn to deal with data in a well-founded and critical manner. To this end, they apply relevant process steps, including those related to the research data cycle, as they work independently on an open data project throughout the course.

In this paper an exploratory evaluation, based on pre- and post-course self-assessments, is presented to provide preliminary, illustrative insights into the framework’s application and its association with students’ perceived competence.

Franke et al. (Franke & Krähling-Pilarek, 2024, p. 192) describe how academic libraries offer numerous other training courses and workshops in addition to their focus on information literacy. They state that this offering can be particularly successful in cooperation with other institutions, enabling libraries to become partners in promoting future skills such as digital and data literacy or even AI (in the scientific work process). The Data Literacy Basic Course at TH Köln was also developed in exchange with the university library in several areas of expertise related to the data life cycle. Accordingly, this case study examines the library’s role as an important exchange partner within this modular framework, highlighting the strategic areas where library expertise was integrated into the curriculum.

2. Theoretical Background

Data literacy supports the responsible and thoughtful use of data and, due to the ever-increasing amount of data, is described as an essential and comprehensive skill for all situations, sectors, and disciplines (Heidrich et al., 2018; Schüller et al., 2021). Data literacy is defined as the ability to critically collect, manage, evaluate, and apply data (cf. Ridsdale et al., 2015, p. 3). Through the professional and critical use, analysis, and interpretation of data, data literacy aims to help identify issues and develop appropriate solutions for global economic and social phenomena. Schüller et al. (2019) see data literacy as a key competence of the 21st century and describe it as ’a cluster of all efficient behaviors and attitudes for the effective fulfillment of all process steps for value creation or decision-making from data’ (Schüller & Busch, 2019). Data skills are necessary to use data professionally and effectively in a specific subject and research area, both at university and in later professional life (cf. Schüller & Busch, 2019, p. 16).

2.1. Areas of Competence in Data Literacy

In order to integrate academic libraries into data literacy training, it is helpful to include models for structuring and describing the underlying competencies and tasks. Competence frameworks and models provide guidance by identifying the elements that are necessary to act effectively in a specific area of responsibility (Schüller et al., 2019). In the higher education context, they describe which areas of data literacy are necessary in a basic course of study to provide students entering an increasingly data-dependent working world with a sound basis for action (Ridsdale et al., 2015; Schüller et al., 2019).

Based on an extensive literature review, Ridsdale et al. developed a competence matrix that summarizes the core skills and competencies of data literacy. In total, this matrix describes 23 competencies and the associated skills, knowledge, and expected tasks (64 in total), which are divided into five competence areas: ‘Conceptual Framework’, ‘Data Collection’, ‘Data Management’, ‘Data Evaluation’, and ‘Data Application’ (Ridsdale et al., 2015, p. 3). Examples of such competencies include data discovery and collection, data organization, data manipulation, data visualization, and others. The ’Data Application’ competence area includes competencies such as critical thinking, data culture, and data ethics. The competence matrix developed by Ridsdale et al. forms the basis for other studies (i.e., Bandtel et al., 2021). Schüller et al. (2019) also use this as a basis for their data literacy framework, which encompasses the competencies necessary for decision-making and knowledge development in data-related contexts. The cyclical process model divides the respective process steps and the associated competencies into productive and receptive steps. The productive area encompasses the competencies necessary to derive data products from available data. The receptive area maps skills and tasks that are necessary for decoding data projects and uncovering the underlying data. The respective skills are divided into ‘basic level’, ‘advanced level’, and ‘expert level’ (Schüller et al., 2019, pp. 33–34).

2.2. The DaLI Competence Model of the TH Köln

The DaLI competence model (Echtenbruck et al., 2025) is based on the data literacy competence matrix of Ridsdale et al. and the data literacy framework of Schüller et al. and integrates the research data life cycle to emphasize the importance of data literacy in the higher education and research context. The research data life cycle describes the management of research data from collection and storage to use, archiving, or publication. The integration of the research data life cycle emphasizes scientific work with research data at universities and university libraries, rather than everyday data or data from a business context. Here, the data lifecycle often ends with data storage, whereas in the research data lifecycle, indexing, citation, and sharing of data (as open data) also play important roles. This enables an action- and process-oriented approach to the development and implementation of teaching and learning units on data literacy.

The DaLI competence model is divided into seven competence areas aligned with the research data life cycle: ‘Establish Data Culture’, ‘Provide Data’, ‘Manage Data’, ‘Analyze Data’, ‘Evaluate Data’, ‘Interpret Data’, and ‘Publish Data’, each of which summarizes a variety of detailed competencies (cf. Figure 1).

Special features of the DaLI competence model are the competence areas ‘Establish Data Culture’ and ‘Publish Data’. Both are directly linked to central topics of academic libraries. The overarching competence area ‘Establish Data Culture’ includes ‘Data Fundamentals’, ‘Data Ethics’, ‘Data Protection’, and ‘Copyright’ and should be seen as the basis for all other competence areas. The competence area ‘Establish Data Culture’ is the core value of the competence model and highlights the importance of data-conscious and responsible handling of data. The competence area ‘Publish Data’ describes university-specific features in the handling of research data. The focus is on the open reuse and recycling of data as well as its adequate archiving and publication.

2.3. Libraries as Cooperation Partners in Teaching Data Literacy

The role of libraries in teaching data literacy is viewed in different ways. The spectrum ranges from the use of premises as scientific infrastructure and training facilities to the actual integration of library-specific content, i.e., to concrete participation and input. The question is by no means new. As early as 2004, Schield (2004) called for an expanded role for libraries: they should go beyond traditional information transfer and become active in data and statistical literacy, because these skills are central to critical and evidence-based work. He argues that the three literacies—information literacy, statistical literacy, and data literacy—are closely linked and interdependent. He sees librarians as traditionally responsible for teaching information literacy and recommends that librarians and libraries integrate statistical literacy into their teaching programs in the future to enable students to evaluate data and its statistical analysis critically.

Almost 10 years later, Prado and Marzal (2013) identified five key areas for library involvement in data literacy: understanding data, finding and obtaining data, interpreting and analyzing data, data management, and communicating and presenting data. The latter encompassed visualization, citation, data sharing, and ethical issues such as data protection and copyright. Notably, their framework did not yet include the provision, indexing, and retrieval of research data, reflecting the field’s earlier stage of development. They proposed these five fields as a framework for libraries to structure their data literacy offerings.

The Higher Education Forum on Digitalization’s data literacy study also examines approaches to teaching data literacy and confirms the active involvement of academic libraries as key partners in higher education. On the one hand, it states that teaching data literacy skills should begin as early as possible in the course of study, and that various integration models are conceivable (Heidrich et al., 2018, p. 8f). All of the teaching options examined were interdisciplinary and collaborative with other institutions.

Since the beginning of the Data Literacy Initiative (DaLI) at TH Köln, libraries have been actively involved in curricular and extracurricular data literacy teaching. The initiative was scientifically monitored and documented, and it teaches data literacy skills with the direct involvement of library staff. The study by Fühles-Ubach et al. (2021b) suggests that libraries could take on specific content-related tasks in the areas of ‘Establish Data Culture’, ‘Provide Data’, and ‘Publish Data’, thereby assuming key roles within data literacy education. The extent to which these assumptions can be implemented will also be determined by the evaluation of the courses.

The current working paper by Demiröz and Lämmlein (2024) analyzes studies on integrating data literacy into university curricula and explicitly states that university libraries are key players in teaching this future-oriented skill. Various approaches and competency frameworks are presented, such as that of Calzada and Marzal (Prado & Marzal, 2013), which defines the five central areas in which libraries promote data literacy. The working paper summarizes the state of research and assesses libraries as ’key institutions’ that impart empirically verifiable competencies in practice.

2.4. Evaluation of Data Literacy and Comparable Research

Evaluating data literacy poses methodological challenges that justify our exploratory approach. In their study on the development of a competency framework, Schüller and Busch examine measurement instruments for data literacy in higher education and divide them into: instruments for measuring response, instruments for measuring learning success, instruments for measuring behavior, and procedures for evaluating results. The study highlights the importance of data literacy in higher education and emphasizes the need to develop appropriate instruments to evaluate data literacy teaching and learning events. The study makes it clear that there are currently no testing and measurement instruments that adequately assess the quality of data literacy courses. Most available evaluation instruments focus primarily on assessing general statistical, visual, or information-related skills and do not account for all the skills involved in the data life cycle (Schüller & Busch, 2019).

This methodological gap is reflected in current research practice. The study by Oguguo et al. (2020), which developed a separate questionnaire, provides insight into students’ self-assessment of data literacy competencies at five Nigerian universities. The results of the survey, in which 2550 students from various disciplines and degree programs participated, show that the participants rate their data collection skills as comparatively high. In other areas, such as data analysis, data interpretation, and visualization, they rated their own skills as average. The study shows that although students consider themselves capable of searching for data, they have difficulty effectively analyzing and interpreting this data. The authors of the study suggest that the limited curricular integration of data literacy in degree programs could be a possible reason for this discrepancy. Another reason for the moderate ratings of students’ data literacy skills could be teachers’ limited expertise in integrating all aspects of data literacy into their courses. In addition, according to the study’s authors, the rapid development of technology and data-handling applications may play a role, requiring continuous adjustments to data literacy education.

Similarly, Bandtel et al. (2021) employed pre- and post-questionnaires to assess the participants’ personality traits and perceived digital skills in their evaluation of an interdisciplinary data project, showing how such instruments can provide valuable insights into educational interventions.

The studies presented clearly show the need for suitable teaching and learning scenarios for the development of data literacy (Schüller & Busch, 2019; Schüller et al., 2019). Although students are able to collect data, they seem to have difficulties with analysis and interpretation (Carlson et al., 2013; Oguguo et al., 2020). The integration of practice-oriented projects and interdisciplinary approaches (Bandtel et al., 2021) can add value to the development of data literacy by creating an environment for targeted, action-oriented development.

Given these methodological challenges and the primary goal of introducing and evaluating our course framework, we employed a descriptive evaluation using self-assessments. This exploratory approach represents a pragmatic first step that aligns with current research practices while providing preliminary insights into our modular course design.

3. The Data Literacy Basic Course and Its Modules

Based on the goal of offering data literacy in an interdisciplinary manner at TH Köln, the DaLI Basic Course was developed as a modular, interdisciplinary course for all students at TH Köln.

This course enables students to view tasks with real data from an interdisciplinary perspective and to learn about and apply questions and methods from other disciplines in the respective areas of competence (Lerch, 2019). To adequately convey the areas of expertise addressed, expertise from different disciplines should be included in the design (Schüller et al., 2019). In this regard, ten lecturers from different faculties at the university were involved in developing the content of the DaLI Basic Course and took on responsibility for various data competence areas in accordance with their field of expertise.

The DaLI Basic Course is designed to last one semester, i.e., approximately 14 semester weeks, and consists of seven modules covering the topics in the various areas of competence. This modular design allows library staff and other domain experts to take responsibility for specific modules that align with their core competencies. During the course, students engage with the content in a self-directed manner and work together on a data project with open data2. The structure of the course follows the flipped classroom model (Ağırman & Ercoşkun, 2022) (see Figure 2), in which participants work through a module in the corresponding self-study program of the DaLI Basic Course on the KI-Campus3 between classroom sessions. The first module, ‘Establish Data Culture’, forms the basis of the entire model and also provides the first interface with the library, because good scientific practice, as taught in the library, also forms the basis for ethical, transparent, and high-quality research. This also includes the structured handling of research data and the creation of data management plans, which are also essential foundations of scientific work. In the second module, ‘Provide Data’, students at the AI Campus, for example, explore the concepts of secondary and primary research, learn what hypotheses are and how to form them, learn about the process of an empirical study, and what to look out for when collecting data. Based on the content developed there, they then work on a project task in an online exchange with other students from different faculties. Here, too, there are interfaces with library work, such as the introduction of re3data (www.re3data.org) as a global, web-based directory for research data repositories that helps researchers and students from all disciplines find repositories for research data in libraries.

The respective project tasks are interrelated and build on one another, guiding learners through a complete data project lifecycle—from data acquisition and cleaning to analysis and final documentation of results. This approach ensures the theoretical content is applied and implemented in practice right from the start. In the second module, for example, they formulate research questions and hypotheses based on open meteorological data, define the necessary variables, and actively download data from an open weather data pool for further processing in the project. Weather data was selected because, as a generic data type, it is similarly understandable to all degree programs. Students evaluate the quality and completeness of the available data in relation to their research question and prepare it for further processing.

After completing the task, students post their results and work steps on THspaces4, the social learning environment of TH Köln. The task also involves reading, evaluating, and providing constructive feedback on two posts by other students in a peer-learning format. Every two weeks, the topic currently being worked on is discussed and supplemented in terms of content in online meetings. These include group sessions in which students work together on the open data project and exchange their results. The DaLI Basic Course is offered each winter semester and has already been held four times.

While modules 3–6 deal extensively with the processing, analysis, and visualization of data from the student project, the seventh module, ‘Publish Data,’ once again covers key areas of library work, as the research data cycle is closed in the project.

The ‘Publish Data’ module, instructed by the university library’s dedicated Research Data Management (RDM) expert, provides foundational training for closing the research data lifecycle. The instructional content is structured around several core areas. It introduces the research data lifecycle and the German National Research Data Infrastructure (NFDI). Subsequently, it covers TH Köln’s specific ‘Guideline for Handling Research Data’, reiterating the FAIR Data principles (i.e., research data should be findable, accessible, interoperable, and reusable) (Wilkinson et al., 2016) and the importance of Open Access for scientific publications. A significant part is dedicated to long-term archiving, with a focus on the university’s own repository where students can publish their theses. Simultaneously, the global registry re3data is introduced to help students, scientists, funding organizations, and publishers find suitable data repositories. A film is used to demonstrate this tool. A third key topic is bibliometrics and altmetrics, in which the module discusses how scholarly success is measured through various indicators. This workshop-style session ensures students learn to navigate real-world archiving tools and standards, empowering them to share publicly and cite research outputs responsibly. This module exemplifies how the modular course structure facilitates the integration of library expertise.

4. Research Design and Methodology

The DaLI Basic Course is open to all students from all faculties and is advertised via a university-wide email distribution list. As a result, the participants come from a variety of disciplines and semesters. To evaluate students’ learning progress, this case study examines how the methods and content of the Basic Course build on and expand students’ existing skills.

To this end, a separate pre- and post-questionnaire was developed for the evaluation (Kaliva et al., 2023), which was based on the TH Köln’s competence framework for data literacy (cf. Section 2.2).The section dealing with the self-assessment of data literacy competencies, however, was identical in both questionnaires. In line with the learning outcomes (Wunderlich & Szczyrba, 2016) of the DaLI Basic Course, the focus was mainly on competencies in the lower levels of Bloom’s taxonomy, which relate to knowledge, understanding and basic application skills (Bloom et al., 1956). This evaluation summarises the research conducted between the winter semester 2022/23 and the winter semester 2024/25. The population consisted of all participants in the Basic Course during these three years. The data was collected using two online questionnaires that were made available to students at the beginning and end of the course in a single-group, pre-post design.

The evaluation addresses the following questions:

EQ1 (Participant Engagement and Challenges):

What challenges and barriers do students from heterogeneous academic backgrounds face when participating in the cross-disciplinary Data Literacy Basic Course?

EQ2 (Overall Impact on Data Literacy Confidence):

Does participation in the DaLI Basic Course lead to an increase in students’ self-assessed confidence in handling data?

EQ3 (Development of Competencies in Library-Related Modules):

For the library-related modules—‘Establish Data Culture’, ‘Provide Data’, and ‘Publish Data’—how does students’ self-assessment of their competencies develop after completing the course?

The sampling for the evaluations, which includes pre- and post-course surveys to address EQ2 and EQ3, as well as an interim qualitative survey to address EQ1, was conducted as follows:

Pre-evaluation: All students who enrolled and attended the first course session were invited to participate.

Interim evaluation: All enrolled students who were no longer actively participating by mid-course were surveyed. This survey also included an offer to catch up on missed work and re-join the course.

Post-evaluation: All students who completed the entire course (verified by attendance or submission of final project work) were invited.

This study used anonymous, voluntary surveys conducted for the internal evaluation and quality assurance of the DaLI Basic Course. In accordance with the institutional policies of TH Köln for such mandatory course evaluations TH Köln (2018) which are designed to protect participant anonymity and are integral to curriculum development, a separate ethical approval was not required. All participants were informed about the anonymous nature of the data collection and the purpose of the evaluation prior to their participation.

5. Results—Assessment of the Data Literacy Basic Course

A total of 130 students completed the pre-questionnaire, while 42 students answered the post-questionnaire. The dropout rate was therefore around 68% overall. The distribution of students by semester was similar in percentage terms in the pre- and post-surveys. There were greater fluctuations in terms of faculty affiliation.

5.1. Participant Engagement and Challenges (EQ1)

In the qualitative interim survey ‘Still with us or already done?’, the students cited time constraints and conflicting priorities as the most common reasons for leaving, followed by health or personal problems, different expectations of the course content, and difficulties getting started, especially among first-year students. The feedback also revealed differences in how students from various backgrounds perceived the course. An individual student with data science experience found the content ‘too basic,’ while a few students from non-technical disciplines reported challenges, particularly with programming and data preparation tasks. As one student noted: ‘Wer kein Data Scientist ist, für den ist es wirklich schwer hinterherzukommen’ (‘For those who are not data scientists, it is really difficult to keep up’). Additionally, one student from cultural and social sciences questioned the relevance of working with weather data for their majors, describing it as ‘ziemlich langweilig’ (‘quite boring’).

These findings highlight the challenge of designing a single course that adequately addresses the needs of a heterogeneous student population with varying prior knowledge and disciplinary interests.

5.2. Overall Impact on Data Literacy Confidence (EQ2)

Despite these challenges, the data from students who completed the course indicate a positive association between participation and self-assessed data literacy. This trend is evident from the noticeable increase in agreement with the statement ‘I feel confident in dealing with data’ at the end of the course compared to the beginning (Figure 3).

On a scale of one to five, they also rated their general data literacy higher after the course (Figure 4).

To complement these descriptive findings, we conducted independent t-tests to quantify the observed differences. The results align with the visual trends, showing substantial differences in scores for both the confidence statement and the general data literacy rating (p < 0.001).

The analysis reveals a consistent positive trend across all seven modules, with self-assessed competencies increasing from pre-course means (ranging from M = −0.38 to M = 0.49) to post-course means (ranging from M = 1.14 to M = 1.37) (cf. Figure 5). This uniform shift suggests that the training programme was associated with improvements in self-assessed data literacy skills. The post-course results also show a notable reduction in standard deviations compared to the pre-course assessment, suggesting that student competencies have converged towards a higher level (cf.

Table 1. Pre- and Post-Course Self-Assessment Results. The table displays the means (M) and standard deviations (sd) aggregated for all seven competency areas, alongside Cohen’s d, which quantifies the standardized mean difference for each area and the result of the Welch Two Sample t-test (with p < 0.001 in all cases).

Competence Area	Pre (M ± sd)	Post (M ± sd)	Cohen’s d 1	t(df)
Establish Data Culture	0.32 ± 0.72	1.33 ± 0.63	1.44	t(77.8) = −8.64
Provide Data	−0.39 ± 1.06	1.17 ± 0.85	1.52	t(41.0) = −7.73
Manage Data	−0.08 ± 1.03	1.12 ± 0.85	1.21	t(83.1) = −7.54
Analyze Data	0.01 ± 1.15	1.11 ± 0.85	1.01	t(44.0) = −5.36
Evaluate Data	0.30 ± 0.99	1.22 ± 0.83	0.95	t(39.4) = −4.72
Interpret Data	0.49 ± 0.84	1.37 ± 0.77	1.06	t(74.9) = −6.26
Publish Data	−0.35 ± 1.05	1.26 ± 0.88	1.57	t(39.3) = −7.78

¹ Cohen’s d represents the standardized mean difference between two groups. It is calculated as the difference between means divided by the pooled standard deviation. Common interpretations: d = 0.2 (small), 0.5 (medium), 0.8 (large) (Cohen, 2013).

).

The most pronounced gains were observed in the areas of ‘Provide Data’, ‘Manage Data’, and ‘Publish Data’ where median scores shifted from negative values, indicating a lack of confidence, to strongly positive values post-course (M > 1.1).

The pronounced improvements across these library-centric modules suggest that targeted training can effectively address critical competency gaps, underscoring the strategic value of integrating data literacy into library services and highlighting a clear pathway for future educational initiatives.

5.3. Development of Competencies in Library-Related Modules (EQ3)

The analysis of self-assessed competencies in the three library-related modules—‘Establish Data Culture’, ‘Provide Data’, and ‘Publish Data’—shows a substantial increase in self-reported confidence from pre- to post-course. Students rated their agreement with competency statements on a scale of −2 ‘not applicable’, −1 ‘rather not applicable’, 0 ‘partly applicable, partly not applicable’, +1 ‘rather applicable’, and +2 ‘applicable’.

In the competence area ‘Establish Data Culture’, the students stated already at the beginning of the Basic Course that they knew the difference between data, information and knowledge and that they handled personal data with care (cf.

Table 2. These are the results of the self-assessment questions for the area ‘Establish Data Culture’. For each question, the mean and standard deviation for the pre- and post-evaluations are given, along with Cohen’s d, which quantifies the effect size of the differences between pre- and post-evaluations and the result of the Welch Two Sample t-test (with p < 0.001 in all cases).

Question	Pre (M ± sd)	Post (M ± sd)	Cohen’s d	t(df)
I can distinguish between data, information, and knowledge.	0.46 ± 1.04	1.45 ± 0.74	1.02	t(97.06) = −6.8
I can recognize gross violations of data protection on the Internet.	0.33 ± 0.94	1.36 ± 0.85	1.12	t(75.62) = −6.63
I am familiar with guidelines, rules, and criteria that are relevant to data protection.	−0.04 ± 0.96	1.19 ± 0.86	1.31	t(76.5) = −7.81
I can recognize whether content on the internet is protected by copyright.	0.16 ± 1.02	1.24 ± 0.82	1.11	t(85.17) = −6.95
I handle personal data with care.	0.80 ± 1.07	1.38 ± 0.73	0.58	t(101.56) = −3.96
I am familiar with ethical issues/aspects relating to the use of data.	0.23 ± 1.09	1.36 ± 0.79	1.10	t(95.29) = −7.27

).

However, in their opinion, they were not good at assessing data protection violations, had difficulty recognising whether internet content is protected by copyright, and were not sufficiently familiar with guidelines, rules, and criteria relevant to data protection. These scores improved notably after completing the course. A pronounced positive shift was also observed in students’ self-efficacy regarding ethical questions. Table 2 provides a detailed breakdown of all the evaluation results for the module of ‘Establish Data Culture’.

In the ‘Provide Data’ area, students reported in the pre-evaluation that they found it challenging to plan and create a data set related to a research question/hypothesis. In addition, they stated that they were not familiar with quality criteria for research data and therefore rarely applied criteria to critically evaluate the quality of a data source.They also found it challenging to clean existing data.

After the course, self-assessed competency in these areas increased markedly.

Table 3. These are the results of the self-assessment questions for the area ‘Provide Data’. For each question, the mean and standard deviation for the pre- and post-evaluations are given, along with Cohen’s d, which quantifies the effect size of the differences between pre- and post-evaluations and the result of the Welch Two Sample t-test (with p < 0.001 in all cases).

Question	Pre (M ± sd)	Post (M ± sd)	Cohen’s d	t(df)
I can plan and create a data set related to a research question/hypothesis.	−0.37 ± 1.17	1.26 ± 0.86	1.48	t(94.27) = −9.75
I am familiar with quality criteria for research data.	−0.18 ± 1.25	1.48 ± 0.86	1.41	t(101.2) = −9.58
I apply criteria to critically evaluate the quality of a data source.	−0.25 ± 1.19	1.21 ± 0.87	1.30	t(94.34) = −8.59
I can clean up existing data.	−0.62 ± 1.17	1.04 ± 0.95	1.47	t(40.18) = −7.38

provides a detailed breakdown of all the evaluation results for the module of ‘Provide Data’.

The area of ’Publish data’ received the lowest rating, indicating the lowest level of competence. The students stated that they tended not to be familiar with the possibilities of sustainable data management. They also stated that they were unable to name the advantages and disadvantages of open access publications and open data or the opportunities and challenges of long-term archiving and reuse of data. They were also largely unfamiliar with methods for citing scientific data and publications in their own discipline.

Upon completing the course, students reported substantially higher levels of confidence in relation to these skills, with median scores shifting towards strongly positive numbers.

Table 4. These are the results of the self-assessment questions for the area ‘Publish Data’. For each question, the mean and standard deviation for the pre- and post-evaluations are given, along with Cohen’s d, which quantifies the effect size of the differences between pre- and post-evaluations and the result of the Welch Two Sample t-test (with p < 0.001 in all cases).

Question	Pre (M ± sd)	Post (M ± sd)	Cohen’s d	t(df)
I know ways to handle data sustainably and publish it.	−0.53 ± 1.15	1.26 ± 0.86	1.65	t(92.54) = −10.78
I am aware of the advantages of open access publications and open data.	−0.31 ± 1.32	1.48 ± 0.83	1.46	t(111.43) = −10.3
I am familiar with the challenges and opportunities involved in long−term archiving and reuse of data.	−0.6 ± 1.22	1.29 ± 0.86	1.65	t(97.76) = −11.04
I am familiar with methods for citing scientific data and publications in my field of expertise.	0.02 ± 1.32	1.21 ± 1.1	0.92	t(39.54) = −4.59

provides a detailed breakdown of all the evaluation results for the module of ‘Publish Data’. The need for library cooperation is by far the greatest in this area.

6. Discussion

This discussion evaluates the proposed modular data literacy course framework using the empirical findings. We analyze its effectiveness in building student competencies, the strategic role of libraries, and key challenges for implementation, as well as the limitations of the explorative study. The section closes with implications for future work.

6.1. Summary and Interpretation of Key Findings

In this paper, we propose a modular course framework that translates a theoretical data literacy competence model (Echtenbruck et al., 2025) into a practical and reusable curriculum. The structure, which, among others, maps specific library expertise onto dedicated learning modules, provides a scalable model for other institutions. The empirical findings from our exploratory survey serve to illustrate this framework’s implementation and provide preliminary insights into its reception.

Regarding EQ1 (Participant Engagement and Challenge), qualitative feedback reveals that the course faces challenges in addressing diverse student needs. The main difficulties reported by participants were not related to the course content itself but to time constraints and workload management. Many struggled to keep up due to high academic demands, overlapping study or work commitments, and personal circumstances. In particular, the lack of curricular integration or formal recognition as an additional qualification in some faculties appears to be a barrier to sustained participation.

Some feedback revealed a divide between data-savvy participants, who found the course too basic, and beginners, who felt overwhelmed. This highlights the need for a more differentiated approach to data literacy education, balancing foundational skills with discipline-specific applications.

Concerning EQ2 (Overall Impact on Data Literacy Confidence) the data from students who completed the course indicate a positive association between participation and self-assessed data literacy. After the course, students feel more confident in handling data and rate their data literacy higher. The students’ self-assessment shows improvements, primarily in the areas of ‘Provide Data’, ‘Manage Data’, and ‘Publish Data’. The increase was the highest in these areas compared to the pre-evaluation (Figure 5).

In response to EQ3 (Development of Competencies in Library-Related Module) the results demonstrate substantial improvements in the three library-related modules.

Competencies in the area of ‘Publish Data’ are considered particularly important for supporting transparency and dialogue in science and research (Borgman, 2015; Piwowar & Vision, 2013). A special feature of research data comes into play in the competence area of ‘Publish Data’. In the spirit of open science, this data should be publicly accessible (open access) and available for reuse and publication.

This is one of the key areas in which academic libraries can actively contribute and teach skills. It highlights the significant impact that academic libraries can have as partners in data literacy education, leveraging their expertise in scholarly communication and research data management.

6.2. Discussion in the Context of the Literature and Library’s Role

Our findings resonate with existing literature, which consistently shows students struggle with and have low self-efficacy regarding data analysis (Ghodoosi et al., 2024; Oguguo et al., 2020). Our pre-evaluation results, showing low confidence in data handling, further confirm this established baseline. The improvement in confidence around publishing data is particularly noteworthy. Academic libraries, with their established role in scholarly communication and research data management, are logically positioned to teach the principles and practices of data sharing, transparency, and reuse. The strong results in this area demonstrate the impact the academic library can have as a partner in data literacy education by focusing on its native competencies.

The mixed reception among students from different disciplines aligns with existing literature on interdisciplinary data literacy education. The feedback suggests that while the weather data provided a generic, accessible use case, it failed to engage some students from non-STEM fields. Future iterations might benefit from offering multiple data sets or project options with clearer relevance to different disciplinary contexts, potentially leveraging the library’s expertise in domain-specific data sources.

6.3. Limitations

Among other things, due to the exploratory nature of this case study, several limitations must be taken into account. Since the study is based on evaluations primarily intended to assess student satisfaction and self-perception regarding the course, the evaluations use self-assessments, and there is no independent comparison group. Therefore, other factors may influence improvements. Furthermore, the reliance on self-assessment data means the findings reflect changes in perceived competence and confidence rather than providing direct, objective evidence of actual skill development. These limitations are compounded by the limited sample size, the high dropout rate between pre- and post-evaluation, and the fact that the observations are based on a single-institution case. Additionally, the post-evaluation results primarily reflect the perceptions of the more motivated and persistent student subgroup who completed the course. Consequently, the results are not representative or generalizable. Nevertheless, they offer valuable preliminary insights into practice-oriented digital learning scenarios for promoting basic data literacy skills.

6.4. Future Work

Further research with larger and more diverse samples is needed to validate the results and draw more comprehensive conclusions. For future studies, we plan to increase sample sizes and include independent comparison groups in the evaluation. In addition, unique keys will be used in future evaluations to better correlate anonymous pre- and post-evaluations.

7. Implications for Cooperation with Libraries and Outlook

The results of this case study suggest that attending and completing the Basic Course is associated with increased student self-assessments of data literacy. These preliminary findings indicate potential for involving libraries more closely as strategic partners in embedding data literacy in the curriculum. Future collaborations should aim to systematically integrate library services with subject-specific teaching formats in order to cover both technical and ethical-legal aspects of data handling.

A promising approach is the expansion of interdisciplinary teaching and learning settings in which librarians work together with subject lecturers to support practice-oriented projects – for example, in the context of open data initiatives or research-related courses. In this context, libraries can act not only as knowledge mediators but also as infrastructure and service centers that support the entire research data cycle (Fühles-Ubach et al., 2021a).

In the long term, collaborations should also focus on building sustainable networks between university libraries, data centers, research groups, and external partners. International examples show that such networks increase the visibility of data literacy offerings, create synergies in the development of training materials, and promote knowledge transfer between institutions (Lavoie et al., 2023; Reichert, 2019).

Building on the exploratory findings presented here, future research should investigate these collaborative models through controlled studies to better understand their specific impact. As the field evolves, libraries that engage early with emerging trends like AI and automated data analysis can further establish their role as valuable partners in the higher education landscape. In the context of advancing digitalization and the growing importance of data-driven research, the role of academic libraries as competent partners in teaching data literacy will continue to gain relevance. Future developments—for example in the fields of artificial intelligence, automated data analysis, or research data infrastructures—will require continuous adaptation of training content. Libraries that get involved in these fields of innovation at an early stage can consolidate their position as indispensable partners in the higher education context and actively contribute to the training of tomorrow’s ‘data professionals’ (Clemons et al., 2024).