1 Introduction

As the importance of information and communication technology (ICT) transformation processes continues to increase in educational systems and school digitalization contexts (e.g., Medienberatung NRW 2019; Mußmann et al. 2021; Scheiter 2021; Tondeur et al. 2021), the integration of technology into teaching and learning at school has become an increasingly essential aspect of teachers’ professional knowledge base (Mishra and Koehler 2006; Scherer and Teo 2019). It is widely assumed that (pre-service) teachers acquire the knowledge required to contribute to their success in teaching during the theoretical and practical phases of their initial teacher education programs and that such knowledge is further developed by means of professional experience and reflection (Bromme 2008; Mußmann et al. 2021; Stigler and Miller 2018). The opportunities to learn (OTL) concept is thus considered relevant to examinations of teacher education programs’ key characteristics (Floden 2015; König et al. 2017; Kunina-Habenicht et al. 2013; McDonnell 1995; Schmidt et al. 2011).

Empirical research on approaches to testing teacher knowledge has proliferated in recent years (Ulferts 2021). However, few test instruments that encompass technology-related knowledge domains and associated competence facets in a standardized manner have been established (Baier and Kunter 2020; Gerhard et al. 2020; Lachner et al. 2019). Student teachers’ acquisition of general pedagogical knowledge (GPK) and the impact of pedagogical OTL on their knowledge acquisition have been investigated extensively by several empirical studies in Germany (Depping et al. 2021; König et al. 2017; Tachtsoglou and König 2018; Watson et al. 2018); meanwhile, however, knowledge regarding the impact of ICT-related OTL in teacher education on student teachers’ academic learning outcomes remains limited (e.g., Wilson et al. 2020; Zlatkin-Troitschanskaia et al. 2020). Similarly, the contribution of personal factors, such as student teachers’ motivations for using ICT in the classroom, to academic learning outcomes remains significantly underexplored in empirical higher education research (e.g., Zlatkin-Troitschanskaia et al. 2020).

The present study addresses the research gap outlined above. From a theoretical perspective, our work aligns with the technological pedagogical content knowledge (TPACK) framework, which describes technological knowledge (TK), content knowledge (CK), and GPK as three dimensions of knowledge that—together with their intersections (e.g., technological pedagogical knowledge [TPK]; pedagogical content knowledge [PCK])—are decisive for ensuring teaching quality (Mishra and Koehler 2006). Using TPACK as a reference, we focus on the TPK that bachelor student teachers have accumulated following three years of study and investigate how their teacher education OTL and personal factors relate to TPK as a learning outcome.

Current research suggests that the degree of OTL with respect to ICT that is incorporated into pre-service teacher programs at most universities in Germany is sparse (Bertelsmann Stiftung 2021; Jäger-Biela et al. 2020; Mußmann et al. 2021). However, a comprehensive range of initiatives have recently been launched with the aim of fundamentally changing this situation (BMBF 2018). Given that TPK represents an intersection of TK and GPK (Mishra and Koehler 2006), we hypothesize that GPK—the more conventional pre-service teacher knowledge category in existing teacher education curricula (KMK 2004)—has the potential to mediate between pedagogical OTL and TPK. OTL measures are typically used in analyses of the teacher education curriculum, even when based on student teachers’ self-reports (e.g., König et al. 2017; Schmidt et al. 2011). Given that several studies have revealed the importance of personal factors in addition to OTL variables (e.g., Blömeke et al. 2010, 2012), a selection of those factors will also be incorporated into our analyses, including student teachers’ motivations for integrating ICT into the classroom (Bürger et al. 2021). Data from 338 student teachers approaching the culmination of their bachelor studies (6th semester) collected during summer 2021 will be analyzed.

The present study’s originality lies in its application of a novel standardized test measuring student teachers’ TPK as well as its examination of TPK’s association with the technological pedagogical OTL to which these students had been exposed during the first three years of their initial teacher education. In generating new insights into the link between OTL and teachers’ knowledge, the present study aims to contribute to a better understanding of how teacher education program characteristics relate to key academic outcomes. The findings are discussed in the context of expectations regarding the effectiveness of teacher education (Blömeke et al. 2012; Hascher 2014).

2 Literature survey

2.1 TPK as part of teachers’ professional knowledge

In accordance with the expertise research paradigm that adopts a cognitive perspective on the teaching profession, an extensive knowledge base is considered crucial for mastering professional tasks and achieving key teaching objectives (Bromme 2008; Stigler and Miller 2018). Current research indicates that teaching is regarded as teachers’ core activity, for which the facets of professional knowledge CK, PCK, and GPK are important (Baumert and Kunter 2006; Bromme 2008; König 2014; Shulman 1987; Ulferts 2021). This multidimensional structure of professional knowledge has been validated for teachers of varying types from multiple countries (e.g., Blömeke et al. 2010; Ulferts 2021).

As the ICT transformation process in educational systems has increased in relevance, knowledge regarding best practices in technology integration at school has also become an essential component in teachers’ professional knowledge base (Mishra and Koehler 2006). Compared to previous considerations developed by Shulman (1987), according to which GPK and CK partially contribute to PCK, the TPACK framework regards TK as an additional factor (Mishra and Koehler 2006). In light of ongoing technological advancement and new developments of the current era of digitalization, the need to constantly adapt and expand TK has intensified: “TK is always in a state of flux—more so than the other two core knowledge domains in the TPACK framework (pedagogy and content)” (Koehler and Mishra 2009, p. 64). According to Koehler and Mishra (2009) and their definition of TK as it pertains to the TPACK framework, TK “enables a person to accomplish a variety of different tasks using information technology, and to develop different ways of accomplishing a given task. This conceptualization of TK does not posit an ‘end state’, but rather sees it developmentally, as evolving over a lifetime of generative, open-ended interaction with technology” (Koehler and Mishra 2009, p. 64).

Several additional intersections of “bodies of knowledge” that include technology-related knowledge components thus exist, frequently illustrated by using a Venn diagram (Koehler and Mishra 2009, p. 63). The technological content knowledge (TCK) domain includes knowledge regarding the ways in which the technologies and content relating to a given subject (e.g., mathematics) are interrelated, representing “an understanding of the manner in which technology and content influence and constrain one another” (Koehler et al. 2013, p. 16). TPK, meanwhile, comprises knowledge regarding which technologies are available and suitable for use in teaching-learning situations across subjects (Koehler and Mishra 2009). Teachers should be able to transfer technologies from their original application contexts to specific pedagogical and general didactic contexts and apply them in practice, irrespective of the specific subject at hand.

In our study, we focus on TPK for several reasons. Both theoretically and empirically, it appears to be an indicator for ICT integration in school classes (Koehler and Mishra 2009; Lachner et al. 2019; Mußmann et al. 2021; Scheiter 2021). TPK is regarded as the intersection of GPK and TK and considered highly relevant for all teachers and as having the potential to contribute significantly to the development of TPACK (Wilson et al. 2020). Previous studies on TPACK have relied primarily on self-report data (e.g., Hofer and Grandgenett 2012; Schmidt et al. 2009). However, self-report data can differ considerably from actual performance (e.g., Carrell and Willmington 1996; Dunning 2011). Relatively few standardized tests covering technology-related knowledge domains across subjects have been established to date (Baier and Kunter 2020; Gerhard et al. 2020; Lachner et al. 2019).

Lachner et al. (2019) developed a test to measure TPK that uses multiple-choice questions to elicit underlying conceptual knowledge and short text-based vignettes to measure situational TPK. The authors demonstrated the instrument’s ability to distinguish between different levels of teacher experience, reporting findings regarding the correlation between TPK and participants’ GPK but not their TK. In another study by König et al. (2020), who applied the test among in-service teachers during homeschooling in the context of the COVID-19 pandemic, the conceptual knowledge subscale of the TPK test successfully predicted whether teachers felt able to maintain contact with their students and to assign them differentiated tasks, thus confirming the test’s validity.

Baier and Kunter (2020) proposed a TPK test with open-ended items, several of which refer to specific technologies while the others are more general. Based on a student survey, the study identified 14 items that represent TPK as a latent construct. However, neither convergent nor discriminant validity was found when the test was related to self-report measures. The test was sensitive to pedagogical interventions, demonstrating a significant increase in TPK among student teachers who had attended a course on teaching with digital media.

Scheiter (2021) concludes that the predictive validity of the tests developed to date is likely related to the extent to which the items are related to teaching. Test items tend to be abstract, having been formulated independently of classroom practice so as to be as widely applicable as possible. Moreover, controversies have arisen with respect to how teachers develop their TPACK, since the integrative and transformative perspectives are opposed to one another: the former assumes that each dimension (TK, CK, and GPK) contributes individually to the TPACK framework, while, seen from the transformative perspective, TPACK develops independently of these dimensions. Scheiter (2021) emphasizes that the two perspectives have different implications for teacher training. From the integrative perspective, each component can be taught specifically, while from the transformative perspective, TPACK can be conveyed as a single unit. This highlights the need for further research on TPK as an outcome of initial teacher education.

2.2 TPK as an outcome of teacher education OTL

Scholars broadly agree that teachers’ acquisition of professional knowledge as part of their expertise and competence (Baumert and Kunter 2006; Stigler and Miller 2018) can be fostered through subject-related and pedagogical learning opportunities as well as in-school teaching practice opportunities, as provided by teacher education institutions (Cochran-Smith and Zeichner 2005; Flores 2020). According to this, the concept of OTL is relevant in describing the curricular characteristics of teacher education provision and analyzing pre-service teachers’ acquisition of professional knowledge (Blömeke et al. 2012; McDonnell 1995; Floden 2015; Schmidt et al. 2011). While research in pedagogical OTL has increased in recent years, consistently demonstrating links to GPK as an outcome of initial teacher education (Depping et al. 2021; König and Seifert 2012; König et al. 2017; Tachtsoglou and König 2018; Watson et al. 2018), few studies to date have addressed technological pedagogical OTL.

It has become clear, in no small part due to the COVID-19 pandemic, that schools and teachers remain inadequately prepared for education and teaching in today’s digital world. While the infrastructure and technical equipment have evolved slightly, learning opportunities related to digitalization-related content have shown little development or remain only minimally used by students (Bertelsmann Stiftung 2021; Lorenz et al. 2021). For example, teacher education monitoring data (Bertelsmann Stiftung 2021) reveal that the curricular content designed to form the basis of media literacy and digitalization-related competence among pre-service teachers is still not a mandatory component in numerous teacher training programs. Between 2017 and 2020, little progress was recorded in the curricular anchoring of mandatory courses on the topic of teachers’ professional digital competence. Moreover, the cross-cutting topic of professional digital competence has not yet been anchored in all subdivisions of teacher training programs, including educational sciences, subject-specific educational sciences, subject-specific sciences, and practical phases, in a way that is binding and obligatory for all student teachers (Bertelsmann Stiftung 2021). Using the University of Cologne as an example, Jäger-Biela et al. (2020) examined the intended and implemented curriculum toward topical issues of professional digital competence. Results from a syllabus analysis based on 2018/19 academic courses and survey data from student teachers on the implemented curriculum were evaluated and compared. Evidence revealed that more than half of the students were wholly unaware of learning opportunities related to professional digital competence, clearly indicating a lacuna in initial teacher education programs.

In their meta-analysis, Wilson et al. (2020) offered evidence for the impact of teacher education courses in technology integration (TECTI) on pre-service teachers’ knowledge. Based on 38 studies and 46 corresponding independent effect sizes, an average positive effect on pre-service teacher knowledge at d = 1.057 demonstrated the courses’ effectiveness for technology integration. However, neither the course design features nor the quality characteristics satisfactorily explained the increase in pre-service teachers’ knowledge. One explanation is that a single course is not enough to identify relevant characteristics. This indicates the need for research that opens up the black box between learning opportunities and outcomes in technology-related teacher education.

2.3 Pre-service teachers’ personal factors affecting TPK

2.3.1 Affective–motivational variables

Research shows that in-service teachers’ successful technology integration can be related to their affective–motivational dispositions (e.g., Bürger et al. 2021; Eickelmann and Vennemann 2017; Scherer and Teo 2019). For example, applying the technology acceptance model (TAM; Davis et al. 1989; Scherer et al. 2019; Venkatesh and Davis 2000) in a meta-analysis, Scherer and Teo (2019) demonstrated corresponding variables’ relevance to the teachers’ behavioral intention to use technology. First, “perceived usefulness” (U) is defined as the prospective user’s subjective probability that using a specific application system will enhance their job performance within an organizational context. Second, the dimension “perceived ease of use” (EOU) indicates the probability that they will use a given application system (Davis et al. 1989). Eickelmann and Vennemann (2017) reanalyzed a sample from the International Computer and Information Literacy Study (ICILS) to investigate typologies of teachers’ attitudes toward and beliefs about the use of ICT in teaching and learning. Their findings demonstrated “that positive attitudes and beliefs (especially U and EOU) are regarded as crucial determinants and predictors for teachers’ use of ICT in instruction” (Eickelmann and Vennemann 2017, p. 754 f.). Against this background, the present study accounts for factors such as U and EOU as relevant affective–motivational variables of professional teacher competence.

Selecting student teacher motivation for possible ICT integration as a predictor of teachers’ TPK acquisition during training, we relate our research to general assumptions regarding learning and achievement motivation that may be considered the central motivational predictors for success in academic learning scenarios at university (Schiefele and Urhahne 2000). Initial teacher education programs in Germany focus heavily on academic knowledge acquisition (Zlatkin-Troitschanskaia et al. 2020), supported by formal learning opportunities—in particular, conventional lectures at university (König and Seifert 2012; Watson et al. 2018). It thus appears to be important that student teachers develop positive attitudes toward the use of ICT in instruction, since this, in turn, can support their acquisition of TPK during initial teacher education.

2.3.2 Demographic, social background, and teacher education entrance characteristics

Various frameworks have been developed to outline the potential impact of teacher education programs on pre-service teachers’ learning (e.g., Blömeke et al. 2012; Flores 2020; Hascher 2014) and have demonstrated that individual entry requirements influence (pre-service) teachers’ professional competence both directly and indirectly through their individual use of OTL (e.g., König and Seifert 2012; Kunina-Habenicht et al. 2013). Entry requirements include sociodemographic characteristics (such as gender and age) as well as cognitive prerequisites (such as grade point average [GPA]). Certain personality traits, such as cognitive prerequisites or sociodemographic characteristics, can also explain interindividual differences in student teachers’ professional knowledge acquisition (e.g., Blömeke et al. 2010, 2012). Regarding teachers’ intentions to use technology, gender differences revealed partial disadvantages for female study participants, but these findings are not consistent in the literature (e.g., Scherer and Teo 2019). Previous studies regarding knowledge among student teachers reveal at least minor correlations with GPA (e.g., Depaepe and König 2018; König and Seifert 2012).

3 Research questions and hypotheses

This study is guided by the following research questions and the hypotheses (abbreviated: H) that derive from them. The research model (Fig. 1) serves as a heuristic framework for these research questions and hypotheses.

Fig. 1
figure 1

Research model to explain TPK. OTL opportunities to learn, GPK general pedagogical knowledge, TAM technological acceptance model (U perceived usefulness, EOU perceived ease of use), TPK technological pedagogical knowledge, GPA grade point average (Abiturnote), ICT information and communication technology

Full size image

1

How do TPK and GPK relate to OTL among student teachers at the end of their bachelor studies?

Initially, we assume that OTL that is proximal to specific higher education outcomes serves best as a direct and significant predictor for both TPK and GPK test scores. As such, we hypothesize that GPK may be explained by pedagogical OTL (according to the state of research, e.g., Blömeke et al. 2010, 2012; König and Seifert 2012), while TPK can similarly be explained by technological pedagogical OTL (H1a). By contrast, being less proximal, pedagogical OTL is correlated with TPK, and technological pedagogical OTL is correlated with GPK (H1b) to a moderate extent only. One reason for this is that each of the OTL instruments (GPK: König et al. 2017; TPK: Gerhard et al. 2022) has been closely developed with reference to the concrete content that the specific knowledge test reflects (for further detail, see Sect. 4.2). Figure 1 illustrates H1b using dotted lines to indicate hypothesized weaker predictors. According to recent research, pre-service teachers—even graduate students—typically report that they have experienced little technological pedagogical OTL during the course of their studies (e.g., Jäger-Biela et al. 2020). As such, we hypothesize lower correlations between technological pedagogical OTL and TPK than between pedagogical OTL and GPK (H1c).

2

Does GPK mediate the usage of OTL and TPK of bachelor student teachers?

TPK is regarded as being at the intersection of TK and GPK (Mishra and Koehler 2006). Therefore, we assume a positive correlation between GPK and TPK (H2a). Moreover, we hypothesize that existing pedagogical OTL, which impacts the acquisition of GPK, may also be indirectly correlated with TPK (H2b)—not least because the link between pedagogical OTL (as part of conventional teacher education programs) and GPK may be stronger than that between technological pedagogical OTL (being part of recent implementations in teacher education programs) and TPK (see H1c).

3

Do pre-service teachers’ personal factors and entrance characteristics predict their TPK?

We assume that student teachers’ affective–motivational characteristics—namely, perceived U and EOU—are positively related to TPK (e.g., Bürger et al. 2021; Eickelmann and Vennemann 2017; Scherer and Teo 2019). Given that previous studies have largely analyzed self-perceptions, we hypothesize that the relationship between scores from standardized TPK tests and affective–motivational factors will likely be less tight than that between self-reported competence and self-reported motivational factors, as previous studies have demonstrated (H3a). Given that GPA, as an entrance characteristic of bachelor student teachers, constitutes a central predictor for success in higher education academic outcomes, we assume that it will affect both GPK and TPK test scores (H3b). Following Mishra and Koehler (2009), TPK represents the intersection of PK and TK and thus an extension of professional knowledge GPK by the technological dimension, and so we also expect that GPA will be a predictor for test performance in TPK.

4 Methods

4.1 Sample

Data were collected at the University of Cologne, one of the largest universities—both in Germany and in Europe as a whole—to provide teacher education programs. The analysis rests upon the survey data collected from student teachers in their 6th semester. Owing to the COVID-19 pandemic, the data collection took place online. All students in the population (N = 1204) were contacted between April and September 2021 via the university’s internal email platform and invited to complete the questionnaire, but only n = 338 students ultimately participated (28% response rate). A total of 279 completed at least three items of the TPK test and were thus included in our analysis (drop-out n = 59). Compared with the study population, the program type for primary schools is slightly overrepresented, whereas the program type for special needs education is slightly underrepresented in our sample (> 5% difference each) (Table 1). When the sample and the subsample (n = 279, with TPK data) are compared, drop-out analysis shows no significant differences with respect to gender or teacher education program type.

Table 1 Population of bachelor student teachers (sixth semester, summer term 2021), survey sample, and subsample with TPK data, each differentiated by teacher education program type
Full size table

4.2 Instruments

4.2.1 Teacher knowledge

The present study uses the GPK test instrument developed under the aegis of the Teacher Education and Development Study–Mathematics (TEDS‑M; Blömeke et al. 2010). The instrument conceptually draws on instructional models of effective teaching (e.g., Good and Brophy 2007) and didactics (e.g., Klafki 1985), integrating four content dimensions of the GPK (structure, classroom management, adaptivity, and assessment) that relate to core teaching tasks (KMK 2004). The test has already been applied in numerous studies, providing evidence for construct and curricular validity (e.g., Blömeke et al. 2010; König and Seifert 2012; König et al. 2017). In the present study, item response theory (IRT) scaling analysis using the software Acer ConQuest (Adams et al. 2015) demonstrated that the test with 42 items (both multiple-choice items and open responses) permits a reliable measurement (Table 2). The weighted mean square of each item is in an acceptable range and the discrimination index is, on average, 0.37, which is good (see Sect. 4.3).

Table 2 Empirical findings from the IRT scaling analyses
Full size table

To measure student teachers’ TPK, we used a standardized test instrument that had previously been developed by an interdisciplinary research team comprising educational researchers and media psychologists (Gerhard et al. 2020, 2022). The instrument was specifically designed with the aim of measuring pre-service teachers’ TPK as an outcome of initial teacher education. Following the conceptualization of GPK in empirical studies (Blömeke et al. 2010; Ulferts 2021), it builds on such content dimensions (e.g., structuring, classroom management etc.), which are valid across specific subjects and relate them to generic technology integration requirements for teachers (for item examples, see Table 3) to reach a conceptualization at the intersection of PK and TK, as outlined by TPACK (Koehler and Mishra 2009; Mishra and Koehler 2006). The TPK test items go beyond mere knowledge about different technologies as in the case of TK. The TPK items contain both technical as well as general pedagogical content aspects (Table 3). Moreover, as a generic instrument similar to the GPK test (Blömeke et al. 2010), the TPK test should be relevant to all teacher education program types, school levels, and school types.

Table 3 TPK test item examples
Full size table

The test development comprised two pilot studies and an expert review (for further details, see Gerhard et al. 2020, 2022). The final test comprises 34 multiple-choice items that are distributed across six content dimensions (classroom management, structuring, diagnosis, evaluation, motivating learners, and dealing with heterogeneity of learning groups) and two cognitive dimensions (recall and understand/analyze). Figure A1 in the online appendix provides an overview of the test items’ content. All items have a closed format, each offering four response options (i.e., one correct answer, three distractors). IRT scaling methods were used to estimate scores. EAP reliability (see, for further details on psychometric indicators, Sect. 4.3) was acceptable (> 0.60) for the total sample as well as for the subgroup of grammar schools and comprehensive schools or vocational colleges teacher education program types (Table 2). For the subgroup of other teacher education program types, EAP reliability was close to acceptable (0.59).

4.2.2 OTL

Two instruments are used to capture pedagogical OTL and technological pedagogical OTL (Table 4). All items require student teachers to indicate whether they have ever studied the relevant topic aspect (dichotomous answer with “yes” (coded as 1) or “no” (coded as 0)). The conceptualization of pedagogical OTL comprises four subareas corresponding to the design of the TEDS-M GPK test and is measured by 33 items. Both Cronbach’s alpha and IRT scaling reliability estimates of the four-dimensional model, representing the instrument’s structure, are good (Table 4).

Table 4 OTL item examples and scaling statistics
Full size table

The technological pedagogical OTL instrument comprises 31 items (see Fig. A1 in the online appendix) that were derived from the question stem of the 34 TPK test items. As some items were similar in content, they were combined into one item for OTL formation. This was done three times in total, so that 31 OTL items represent the content of the 34 TPK test items. For example, the OTL item “Cooperative learning in the classroom with etherpads” (Table 4) was derived from the TPK test item “You use etherpads in your classes. What are etherpads and what are they used for?” (Table 3). Analogous to the structure of the TPK test, the scale for the OTL instrument comprises several content dimensions. Each comprises several items (Table 4). Both the Cronbach’s alpha and IRT scaling EAP reliability estimates of the multidimensional model representing the instrument’s structure are good (approximately 0.80 on average, with the exception of TP-structuring, which has a Cronbach’s alpha of 0.69).

4.2.3 Personal factors

Both TAM dimensions U and EOU were measured using existing instruments (Eickelmann and Vennemann 2017; Goldhammer et al. 2017) with Likert-scale items (1 = strongly disagree; 2 = disagree; 3 = agree; 4 = strongly agree). U was differentiated into three dimensions (Table 5): accessibility and usefulness of digital resources (ACC), enhancement of learning processes with ICT (ENH), and improvement of student skills with ICT (IMP). EOU was measured using the perceived autonomy scale (AUT) and the introductory question: “Thinking about your experience with digital media and digital devices: to what extent do you disagree or agree with the following statements?” Each scale’s reliability is good (> 0.70, see Table 5).

Table 5 Item examples from U and EOU and scaling statistics
Full size table

4.3 Data analysis

IRT scaling analysis was performed using the software Acer ConQuest (Adams et al. 2015). Empirical reliability was examined using EAP estimation (de Ayala 1995), which allows an unbiased description of population parameters (Wu et al. 1997). We also use indices of infit and outfit mean square error (MNSQ), as reported by ConQuest. When an item’s weighted mean square is below 1.4 (Wright and Linacre 1994), that item fits the specific scaling model. Descriptive statistics and correlations were investigated using IBM SPSS Statistics for Windows, Version 27. Path modeling was specified with the program Mplus with latent and manifest variables (Muthén and Muthén 2017). In our path analysis, missing data are assumed to be at random (MAR), and full information maximum likelihood estimation was thus applied (Muthén and Muthén 2017, p. 443). In the following, we use the term “statistically significant” if the error probability is less than 5% (p < 0.05). To take the teacher education program’s provision of digital learning content at the studied university into account, we investigated the program based on content analysis of the course catalogues from the 2020/21 winter term using MAXQDA (VERBI Software, Consult. Sozialforschung GmbH, Berlin, Germany 2021) (see Figure A3 in the online appendix).

5 Results

5.1 Descriptive findings

On average, students solved around 53% of the GPK test items and around 46% of the TPK test items (Table 6). Both tests are generally suitable for measuring 6th-semester bachelor student teachers’ knowledge, since neither the floor nor ceiling effects of the assessments can be identified when examining test score distributions.

Table 6 Overview of study variables and statistics
Full size table

Regarding the student teachers’ self-reported OTL, mean differences were observed between pedagogical OTL and technological pedagogical OTL. Pedagogical OTL scales were rated between 0.43 (“P-Classroom Management/Motivation”) and 0.72 (“P-Assessment”) on average—meaning, for example, that 72% of the topics indicated in the scale “P-Assessment” were perceived as part of bachelor student teachers’ learning opportunities (see Table 4). By contrast, the technological pedagogical OTL scales range between 0.18 (“TP-Dealing with heterogeneity of learning groups”) and 0.33 (“TP-Classroom management”), indicating that these topics had been studied to a significantly lower degree. This weak agreement in our student survey data with respect to technological pedagogical OTL was somewhat confirmed by additional analysis of bachelor teacher education course descriptions that were provided during the winter term 2021/2022 at the University of Cologne (for further detail, see Figure A3 in the online appendix). These commentary descriptions were not particularly relevant to digital learning content for bachelor student teachers.

Table 7 contains inter-correlations of study variables. As assumed (H1a), pedagogical OTL scales are significantly correlated with GPK (except for the scale “Classroom Management/Motivation”); unexpectedly, however, technological pedagogical OTL scales are not significantly correlated with TPK. Against our hypothesis (H1b), pedagogical OTL is not significantly correlated with TPK, and nor is technological pedagogical OTL significantly correlated with GPK. Therefore, the theoretically postulated weak correlation with TPK (H1c) also fails to materialize. By contrast, the assumed (H2a) positive correlation between GPK and TPK, at 0.43, is striking.

Table 7 Inter-correlations of the study variables
Full size table

Another positive correlation is between TPK and one of the three analyzed motivational factors (U)—namely, improvement of student skills with ICT (IMP) (0.14). Contrary to our assumptions, the other factors are not significantly correlated with TPK. Moreover, the perceived ICT autonomy (AUT) is not correlated with TPK. Therefore, H3a is only partly confirmed. Finally, students’ higher GPA scores are correlated with higher scores in TPK and GPK, thus confirming H3b.

5.2 Path model explaining TPK

A path model was specified (Fig. 2) to analyze the correlations of the study variables according to our hypotheses simultaneously and to allow hypothesis H2b regarding direct and indirect OTL effects on knowledge to be tested. As Table 8 indicates, the model fit is good (χ2 / df ≤ 2.5; RMSEA ≤ 0.05 (good fit); CFI ≥ 0.9). Figure A2 and Table A1 in the online appendix additionally present the results without GPK as a mediator. The total explained variance for TPK is higher in the model with GPK (Fig. 2) than in that without GPK (Figure A2). Confidence intervals for all coefficients in Fig. 2 and Figure A2 are provided in Table A2 and A3 in the online appendix, respectively.

Fig. 2
figure 2

Path model explaining TPK. GPK general pedagogical knowledge (weighted likelihood estimates), TPK technological pedagogical knowledge (weighted likelihood estimates), P‑OTL pedagogical opportunities to learn, TP-OTL technological pedagogical opportunities to learn, P‑OTL ST P-OTL structuring, P‑OTL CL P-OTL classroom management/motivation, P‑OTL AS P-OTL assessment, P‑OTL AD P-OTL adaptivity, TP-OTL CM TP-OTL classroom management, TP-OTLST TP-OTL structuring, TP-OTL DI TP-OTL diagnosis, TP-OTL MO TP-OTL motivating learners, TP-OTL HE TP-OTL dealing with heterogeneity of learning groups, GPA grade point average (Abiturnote), U IMP Perceived Utility: Improvement of students’ skills and achievement, n = 277, *p < 0.05, **p < 0.01, ***p < 0.001

Full size image
Table 8 Fit indices for the path model explaining TPK
Full size table

Similar to the previous description of the correlative findings, only the path from pedagogical OTL to GPK is statistically significant, whereas technological pedagogical OTL do not significantly predict TPK. Therefore, H1a can only partly be confirmed. Pedagogical OTL do not directly predict TPK, and nor can technological pedagogical OTL directly predict GPK. This indicates the proximal meaning of OTL for student teachers’ learning outcomes, but goes against H1b, which anticipated moderate correlation between pedagogical OTL and TPK and between technological pedagogical OTL and GPK.

As H2a hypothesized, pedagogical OTL indirectly affects TPK (0.12 as the product of 0.25 and 0.46). Bachelor students’ motivation in relation to improving students’ skills with ICT (IMP) also emerged as a significant predictor as well, therefore our hypothesis (H3a) can partly be confirmed. Finally, as GPA constitutes a central predictor for success in higher education, hypothesis H3b is confirmed—at least for GPK—whereas its effect on TPK is mediated by GPK.

Other variables, such as gender, age, or teacher education program type, were analyzed in further path models. Given that they did not emerge as significant predictors for TPK, we decided to exclude them from the final path model depicted in Fig. 2.

6 Discussion

In consideration of the need to foster pre-service teacher competence with respect to ICT integration in school in the context of the current era of digitalization, we examined the link between the relevant characteristics of teacher education programs and student teachers’ learning outcomes at the University of Cologne, one of the largest universities in Germany to offer teacher education. Theoretically, we applied the TPACK framework and focused on pre-service teachers’ TPK as a relevant facet of their professional digital competence (Mishra and Koehler 2006). To test our hypotheses, we specified a path model with latent and manifest variables, which explains the 23% variance of TPK. Contrary to our expectations, however, TPK was not significantly predicted by specific—that is, technological pedagogical—OTL, and it did not matter whether students had more or less opportunity for exposure to technological pedagogical OTL during their bachelor teacher education program in terms of their performance in the TPK test. One possible explanation may be the presence of a slight floor effect regarding the distribution of technological pedagogical OTL usage among students, which is caused by the limited provision of such content in initial teacher education. This corresponds to the findings of an additional course catalogue analysis (Figure A3 in the online appendix), which shows that only a small number of courses provided by the University of Cologne included digital learning content for bachelor student teachers. Moreover, the limited provision in teacher education curricula has recently been highlighted as a relevant empirical finding in previous research (Bertelsmann Stiftung 2021; Jäger-Biela et al. 2020; Lorenz et al. 2021). Interestingly, no floor effect with respect to TPK could be observed in our data. As such, future research should investigate which other factors might explain student teachers’ TPK. Regarding OTL, informal OTL may contribute to accumulating knowledge alongside the formal OTL offered by the university (Greenhow and Lewin 2016), although findings regarding the effects of informal technological OTL on academic performance are mixed (e.g., Lau 2017). Our focus in the present study was on formal OTL exclusively, but future research that yields insights into the relationship between formal and informal learning opportunities with respect to pre-service teachers’ acquisition of TPACK is clearly warranted.

It transpires that student teachers’ GPK can explain around 9% of TPK variance and thus appears to be relevant in our analysis. We consider this to be consistent with the hypothesis that TPK is conceptually located at the intersection between pedagogical knowledge and technological knowledge. However, TK could not be considered as a separate measure in the present study. It would thus be interesting to investigate whether the previously unexplained variance percentage could be explained by the student teachers’ TK in line with the assumptions of the TPACK framework (Mishra and Koehler 2006).

In investigating the conceptualization of student teachers’ academic knowledge acquisition, we examined their motivation for using ICT as a predictor (following, e.g., Schiefele and Urhahne 2000). Using the TAM, we expected that perceived usefulness and perceived ease of use would explain variance of the TPK scores. Here, the improvement of students’ skills with ICT as just one affective-motivational factor could be identified as significant predictor. One explanation might be that our study analyzed the relationship between self-reported motivational factors and standardized test scores of TPK, whereas in other studies that may have detected higher correlations, self-reported competence and self-reported motivational factors are primarily correlated (e.g., Bürger et al. 2021; Eickelmann and Vennemann 2017). Analyzing the correlation of two different data sources (self-report and test scores) may alter the findings when only self-reported data of the same type is used. Nevertheless, given that TPK could not be explained directly by technological pedagogical OTL, the relevance of student teachers’ affective-motivational dispositions in explaining TPK test scores emerges as an important finding. With respect to student teachers’ GPA as a predictor of TPK, the findings were mixed. While GPA predicts GPK in the path model, similar to previous studies (e.g., Depaepe and König 2018; König and Seifert 2012), TPK could not additionally be predicted directly. In inter-correlational analyses, however, TPK was correlated with GPA almost identically to GPK’s correlation with GPA (−0.13 vs. −0.19, each p < 0.05, see Table 7).

In sum, our analyses correlate well with the literature, which suggests that TPK is partly determined through GPK and that motivational aspects are relevant for learning (e.g., Bürger et al. 2021; Schiefele and Urhahne 2000). Moreover, this study’s findings support the assumption that the implementation of learning opportunities with respect to ICT is proceeding rather gradually (Bertelsmann Stiftung 2021; Lorenz et al. 2021), which can also be observed in other countries comparable to Germany such as Norway (Gudmundsdottir and Hatlevik 2018). In line with expectations regarding the effectiveness of teacher education (Blömeke et al. 2012; Hascher 2014), the question arises as to whether the teacher education programs included in our study are sufficiently equipped to provide adequate support for student teachers’ competence development. If specific OTLs related to TPK are missing, student teachers are obviously left alone in their acquisition of TPK. However, they may profit from their GPK as well as from their affective-motivational dispositions to at least partly meet the requirements for developing TPK in the course of their teacher education program. However, the question of whether such a weakly defined approach to TPK acquisition among student teachers in initial teacher education already fulfills expectations, as outlined in the current discourse on teacher professionalism and technology integration in school (Caena and Redecker 2019; KMK 2017) remains open. We assume that in the current era of digitalization, a stronger emphasis is required to update initial teacher education curricula such that they are suitable for preparing future teachers for the schools of tomorrow (Starkey 2020).

One important implication of the present study’s findings may be that expectations of innovation are very high, potentially originating from the Quality Initiative for Initial Teacher Education (Qualitätsoffenisve Lehrerbildung), a joint initiative of the Federal Government and the Länder that aims to improve the quality of teacher training in the current era of digitalization (BMBF 2018). Accordingly, recognition of the need to reform initial teacher education programs appears to be very timely, not least brought forward by challenges that emerged during the COVID-19 pandemic with respect to digitalization in education (Carrillo and Flores 2020).

7 Limitations

Our study has several limitations that should be considered. First, our sample is small in comparison to the overall population of 6th-semester bachelor student teachers. During the COVID-19 pandemic, only distant learning was allowed, which partly accounts for the low response rate. This likely resulted in a positive selection of students participating. However, we do not assume that such a bias conceals the important correlations that we aimed to examine, such as the link between OTL and knowledge. Nonetheless, the sample derives from one university only, which limits the possibility of generalizing the findings beyond teacher education provided at that university.

The TPK test is considered relevant to all teacher education program types, school levels, and school types. Nevertheless, our analyses cannot wholly rule out the possibility that individual situation descriptions in the tasks are more likely to apply to a particular teacher education program type. Future research could apply differential item functioning analysis (e.g., Blömeke et al. 2013), for example, to yield further insights in this regard. Given that the test has been developed very recently, further steps should be taken to confirm its validity (e.g., predictive validity). A further limitation is that we used pre-service teachers’ self-reports to capture the learning opportunities used by the students. Although this facilitated data collection, we do not know whether their reports are biased, despite the comparison with the course descriptions (Figure A2, online appendix), which confirmed our findings to a certain degree. Future research should analyze the teacher education program in greater detail with respect to its qualitative aspects (Wilson et al. 2020). It should thus be borne in mind that we were only able to measure the quantity of learning opportunities and not the type and quality. From the perspective of teaching and learning in institutional settings, the type and quality of opportunities to learn may further affect outcomes (Praetorius et al. 2018; Whyte et al. 2018). Moreover, the novel technological pedagogical OTL instrument was validated on such a large sample for the first time, while earlier studies had provided indicators of the pedagogical OTL instrument’s validity (König et al. 2017).

Experts assessed the TPK items for their curricular relevance. The findings revealed that all items were evaluated as rather or even fully relevant on average, thus providing a content validity check (Gerhard et al. 2022). Given that the technological pedagogical OTL instruments were derived from the TPK items, we assume that they also represent what could be learned during the training courses. However, as the findings reveal, students reported that they had not been significantly exposed to these OTLs, likely because they had not been sufficiently implemented by their university. Further research is thus necessary to investigate the possibilities allowing a more comprehensive provision and usage of technological pedagogical OTL by students. Moreover, further validation of the technological pedagogical OTL instrument—for example, through direct observations that allow comparison of different sources of implementation across the curriculum—would facilitate better interpretation of our results.

Another limitation is that only cross-sectional data were available for our analysis. Future research should thus deploy longitudinal data to analyze the link between OTL and knowledge. Experimental study designs would also be preferable. As our data collection design continues, in the near future, we will be able to measure changes in TPK and how this is affected by OTL during initial teacher education. Such a research design could also yield insight into the relationship between GPK and TPK—for example, in asking to what extent GPK and TK together constitute TPK or whether it is possible to separate these two factors (integrative and transformative perspective; see Scheiter 2021).