Introduction

Managing the challenges of the first year of study is a crucial task for students who are laying the foundation for an academic career (Heublein et al., 2017; Trautwein & Bosse, 2017; van der Meer et al., 2018). Upon starting their studies, students enter a new learning environment that places new challenges on students’ time management and academic self-efficacy (McKenzie & Schweitzer, 2018; van der Meer et al., 2010).

Accordingly, students’ time management and academic self-efficacy are already known as crucial in students’ first year in higher education: Research emphasised the important role of time management for students to cope with the challenges in the first year of study (Haarala-Muhonen et al., 2017; van der Meer et al., 2010; Wolters & Brady, 2020) and, in accordance with the social cognitive theory (Bandura, 1986, 1997), of academic self-efficacy as a vital predictor of commitment and retention in studies (Brahm et al., 2017; Chesnut, 2017; Lent et al., 1994).

We also know that students’ commitment to their studies and the teaching profession is a notable determinant of their retention (Chesnut & Cullen, 2014; Hagenauer et al., 2018; Hong et al., 2018; Klassen & Chiu, 2011). Although, commitment shows up comparatively broadly in current research (Bowman & Holmes, 2018; Chesnut, 2017; Hong et al., 2018), to our knowledge, no previous study has investigated whether and how pre-service teachers’ commitment to their studies changes in the first and second semester at higher education.

Often research considered the first year of study as a whole (e.g. Bowman & Holmes, 2018; Fryer, 2017) or focused only on the first semester, when students encounter a broad range of new requirements at university (Girelli et al., 2018; Gorges, 2019; Perander, et al., 2020). To look at the whole year does not consider that students, who have no previous experience at the university in the first semester, can build on the experience from the first semester in the second semester. For example, prior researchers point out that students’ self-efficacy beliefs and commitment are supposed to change because of their personal experience of success and failure at university (Bandura, 1986; Chesnut & Cullen, 2014). Additionally, in the first semester, students may improve their time management or realise that it is worse than they expected (van der Meer et al., 2010; Wolters et al., 2017). Therefore, it seems appropriate to distinguish these semesters in the research on the first year in higher education.

Summarising, prior research indicates that students’ commitment is positive related to their academic self-efficacy and time management. However, longitudinal studies that take a more differentiated view on first- and second-semester development of academic self-efficacy and time management and their influence on commitment are lacking. Prior longitudinal research mostly based on between-person designs which describes whether differences between individuals in one construct are associated with differences in individuals in the other construct (e.g. self-efficacy and interest; Nauta, et al., 2002). But these designs cannot account for intra-individual effects. Based on this, we derive the following research questions:

  • RQ 1: How do time management and academic self-efficacy interact on the intra-individual level in the first year of study?

  • RQ 2: How do time management and academic self-efficacy influence commitment on the intra-individual level in the first year of study?

  • RQ 3: Does the relationship between time management, academic self-efficacy and commitment differ in the first and second semester?

As we expect changes in all three constructs over time within one person, we used state-of-the-art statistical analyses which allows us to differentiate trait-like between-person and state-like within-person fluctuation over time: We applied the random intercept cross-lagged panel model (RI-CLPM, Hamaker et al., 2015) for the first time to longitudinal data of pre-service teachers. Moreover, by using three measurement occasions, we examine not only the interplay over time between time management and academic self-efficacy and their impact on commitment but also the extent to which these relationships differ in the second semester from the relationships in the first semester. This seems like a promising deeper dive into first-year processes, as second-semester students can build on experiences they did not have in the first semester. Thus, our research expands existing literature on pre-service teachers’ first year.

The effect of time management and academic self-efficacy on commitment to studies

First-year students enter a new learning environment when deciding to study. Along with increasing opportunities comes the challenge to manage all the new demands occurring with the transition to university (Jansen & van der Meer, 2012; Trautwein & Bosse, 2017), such as gaining an overall orientation within the university system, scheduling learning activities, and encountering different challenges regarding the content of their study programme and subject matter of their courses (Trautwein & Bosse, 2017). In that process, students need the ability to regulate the amount of time available to complete their academic work, to make plans and set priorities (Haarala-Muhonen et al., 2017; Mah & Ifenthaler, 2018; van der Meer et al., 2010; Wolters & Brady, 2020). Claessens et al., (2007, p.262) define time management as ‘behaviours that aim at achieving an effective use of time while performing certain goal-directed activities’ that includes planning behaviour such as setting goals, planning tasks, and prioritising and monitoring behaviours.

Research has shown that students’ time management practices is positive associated with satisfaction, academic performance, and negative associated with procrastination (Chang & Nguyen, 2011; Claessens et al., 2007; Gibney et al., 2011; Haarala-Muhonen et al., 2017; Limone et al., 2020; Wolters et al., 2017). As a part of non-cognitive attributions, time management shows a positive relationship to commitment both directly and indirectly via social adjustment and grade point average (Bowman et al., 2019; Collier et al., 2020). Thus, in terms of robustness checks, we firstly analysed the relation between time management and commitment. We expected that students who report good time management skills would be more committed to their studies than students who struggle with their time management in the first year of study.

Thus, we expected to find positive relationships over time between time management and pre-service teachers’ commitment (Fig. 1).

Fig. 1
figure 1

Relationships over time between academic self-efficacy, time management and pre-service teachers’ commitment

Full size image

Moreover, social cognitive theory highlights the importance of self-efficacy as a cognitive component regarding commitment and retention (Bandura, 1986; Bowman et al., 2019; Chesnut, 2017; Lent et al., 1994). Self-efficacy refers to ‘people’s judgments of their capabilities to organise and execute courses of action required to attain designated types of performances’ (Bandura, 1986, p.391). Students’ academic self-efficacy refers to their beliefs in their abilities that they can succeed academically (Bowman et al., 2019).

Research has shown that self-efficacy is a major predictor of in-service and pre-service teachers’ commitment (Chesnut, 2017; Chesnut & Burley, 2015; Chesnut & Cullen, 2014). Based on multiple prior research, we analysed the relation between academic self-efficacy and commitment as part of our robustness checks. Students who report high self-efficacy might be more committed to their studies than students who report low self-efficacy. Thus, we expected to find positive relationships over time from academic self-efficacy to pre-service teachers’ commitment to their studies.

The interplay of time management and academic self-efficacy over time

Several previous studies showed that prior self-efficacy predicts subsequent self-efficacy; along the same line, prior competencies predict the competencies that follow (e.g. Alisic & Wiese, 2020; Fryer & Ainley, 2019). Furthermore, social cognitive theory assumes triadic reciprocity (Bandura, 1986), meaning that individuals’ attributes (e.g. academic self-efficacy), behaviour (e.g. time management strategies), and external environmental factors operate as an interlocking mechanism, affecting one another bidirectionally. Research shows that self-efficacy and time management skills are positively associated (Bowman et al., 2019; Galindo-Domínguez and Bezanilla, 2021; Wolters & Brady, 2020). Based on these findings, in terms of robustness checks, we analysed the relation of academic self-efficacy and time management in our study. We expected academic self-efficacy and time management to be reciprocally interrelated over time.

Comparison of first and second semesters

Additional to the reciprocity between individuals’ attributes and behaviour the social cognitive theory assumes that both are influenced by the environment (Bandura, 1986; Lent et al., 1994). By entering universities, students encounter a very different learning environment in contrast to their experience at school (Christie et al., 2008). This transition requires students to make major adjustments to their learning strategies (Coertjens et al., 2017).

There are several challenges (e.g. self-directed learning activities, planning study-tasks, study independently, academic integration) in this new learning environment related to effective time management and academic self-efficacy (Perander et al., 2020; Schaeper, 2020). First-year students seem unable to accurately assess their own abilities to deal with these challenges (e.g. gaining an overall orientation, scheduling learning activities, confidence in long-term planning), as they have never been challenged to apply themselves in an environment resembling higher education (Lindblom-Ylänne et al., 2015; Mah & Ifenthaler, 2018; Perander et al., 2020; Trautwein & Bosse, 2017; van der Meer et al., 2018). Even students who have knowledge of time management struggled by applying their knowledge in practice (van der Meer et al., 2010). Thus, students’ self-perception of their time management and academic self-efficacy and their commitment to their studies might change due to their experiences at university (Chesnut & Cullen, 2014). However, one semester may provide students with enough opportunity to gain experience and develop a clearer picture of their time management, academic self-efficacy, and commitment (Wright et al., 2013). Thus, we expect the relationships between the constructs of interest to differ in the first and second semesters, with stronger relationships occurring in the second semester. In the first semester, the environment and feedback processes students experience in the higher education context might have a major impact on their perceived time management skills, academic self-efficacy, and commitment. This theoretical concept is supported by several studies. Van der Meer et al (2010) and Kantanis (2000) pointed out in their quality studies that the adjustment to the new learning environment is a process which takes time. Although students may have had good intentions and study plans, they are not immediately able to implement those plans (Lindblom-Ylänne et al., 2015). Students described that they firstly struggled but that they learned what worked for them as they proceeded in their studies (Perander et al., 2020). Thus, students might need the first semester to learn how to organise their study at university and apply their learnings from the first semester in the second semester. Quantitative research has been shown that students’ self-perception is highly unstable across the introductory study phase (Gorges, 2019). Longitudinal studies showed that the predictive value of self-efficacy for study performance depends on when self-efficacy was measured. Students’ self-efficacy at beginning of their studies was only an indirect predictor (Putwain et al., 2013) or even worser compared to self-efficacy measured during studies (Gore, 2006; Wright et al., 2013) that indicate that students’ academic self-efficacy changes substantial during their first experience at universities. This is in line with the social cognitive theory (Bandura, 1986), which supposed that self-efficacy beliefs develop because of experience. We therefore expected that students’ experience in the new learning environment in the first semester influences their perceived time management skills, academic self-efficacy, and commitment and as a result that the relationship between these constructs are stronger over the second than over the first semester.

  • Hypotheses 1 (H1): The positive relationship of academic self-efficacy to (a) pre-service teachers’ commitment, (b) time management, and (c) subsequent academic self-efficacy is stronger during the second semester compared to the first semester.

  • Hypotheses 2 (H2): The positive relationship of time management to (a) pre-service teachers’ commitment, (b) academic self-efficacy, and (c) subsequent time management is stronger during the second semester compared to the first semester.

Methods

Participants and procedure

The study was part of a comprehensive research project evaluating the development of preservice teachers. N = 579 participants of 2 cohorts enrolled in the Bachelor program at a university of technology in Germany took part in this complete survey of all pre-service teachers: N = 251 in the first cohort started their studies in October 2017; 328 in the second cohort started their studies in October 2018. Teacher training in Germany takes place in a three-stage process: six semester Bachelor program, four semester Master program, and 18 months teacher traineeship at schools. In the first year of the Bachelor program, students start with courses in pedagogy, didactic, and the two major they had to choose. Sixteen percent of the participants studied majors in science, technology, engineering, and mathematics (STEM), 11.6% studied majors in languages, 20.1% of the participants studied majors in social science and STEM, 32.2% studied majors in social science and languages, 18.4% studied their majors in languages and STEM, 5.7% of the participants studied majors in STEM and physical education, and 2.7% studied majors in languages and physical education. Table 1 displays the participant characteristics of the cohorts.

Table 1 Participant characteristics
Full size table

The longitudinal survey was integrated into the study programme for all pre-service teachers. The participants were recruited in the first week of their studies. The current study used the initial survey (T1) and two further surveys: The second survey (T2) took place 6 month after the first survey at the end of the first semester and beginning of the second semester. The third survey (T3) took place 12 months after the first survey at the end of the second semester and beginning of the third semester. Each semester is structured the same way including two phases: 3.5 months lecture time and 1.5 months examination time.

Measures

Commitment

Pre-service teachers’ commitment to their studies was measured by the career decidedness scale (Nägele & Neuenschwander, 2015), which consists of three items. All items are presented in the Appendix A. The scale reflects students’ certainty about their chosen career path and their intention to continue in this area. Therefore, although the name of the scale suggests that it is only about career decidedness, in the study context, the items correspond to the definition of study commitment. Participants responded to items on a six-point scale from 1 (strongly disagree) to 6 (strongly agree). Cronbach’s Alpha was 0.855 at T1, 0.834 at T2, and 0.839 at T3.

Time management

Time management (related to students’ perception of their competences to plan study-related tasks and master everyday life at university) was measured by the planning competence scale (Janneck & Hoppe, 2018), which consists of five items (see Appendix A). Participants responded to the items on an eleven-point scale from 1 (0%) to 11 (100%). Cronbach’s Alpha was 0.889 at T1, 0.927 at T2, and 0.928 at T3.

Self-efficacy

ASE (related to students’ beliefs in their own capabilities to attain given goals) was measured by the self-efficacy beliefs scale (Jerusalem & Schwarzer, 1999). The scale consists of ten items whose phrasing was slightly adapted to the study context (see Appendix A). Participants responded to the items on a four-point scale from 1 (strongly disagree) to 4 (strongly agree). Cronbach’s alpha was 0.848 at T1, 0.923 at T2, and 0.929 at T3.

Missing data

The challenges in longitudinal research generally involve missing data and careless responding (Graham, 2009; Meade & Craig, 2012). For this reason, we conducted a careless response analysis for each of the three surveys, as recommended by Meade and Craig (2012; please see our missing data analysis in Appendix B), and used the full information maximum likelihood (FIML) estimation, making use of all available data points to estimate model parameters (Graham, 2009).

Analytic strategy

We used Mplus version 8 (Muthén & Muthén, 1998–2017) for exploratory structural equation modelling (ESEM) to analyse the factor structure and test for measurement invariance across measurement times. To examine our research hypotheses, we applied the random intercept cross-lagged panel model (RI-CLPM; Hamaker et al., 2015) to our data. This allows us to overcome the shortages of prior cross-lagged panel models by separating the within-person fluctuation from stable between-person differences. Please see Appendix C for a detailed description of the ESEM and the RI-CLPM.

We compared the RI-CLPM with the stability model (M0) and the CLPM (M2) to assess whether the hypothesised RI-CLPM fit the data better than those models. The RI-CLPM contained the paths between time management, academic self-efficacy, and students’ commitment as assumed in hypotheses 1, 2, 3, and 4. Afterwards, we stepwise placed equality constraints on stability and cross-lagged paths in RI-CLPM (M5–10) and compared the resulting models with the even less restrictive model to test hypotheses 5 and 6. Models that did not show a poorer model fit indicated that the effects were constant across the measurement occasion for the constrained paths.

Post hoc analyses

Finally, several control variables were included in the resulting model: cohort, gender, age, final school grade, majors of studies, and satisfaction with studies. All of them might have an impact on the relationships in the assumed model (Berens et al., 2018; Chesnut, 2017; Kim & Corcoran, 2018). Therefore, we dummy-coded the major groups. Since the two groups with physical education students were very small, they were assigned to the respective second major the students studied.

Results

Measurement model

The analysis of our measurement model allows the assumption of scalar measurement invariance over time (see Appendix D). Table 2 presents the model fit indices of the configural, metric, and scalar measurement invariance model across time.

Table 2 Measurement invariance models
Full size table

Our further analytics were based on the nine ESEM factors as manifest indicators of the three variables at the three measurement occasions (Table 3).

Table 3 Descriptive statistics, factor determinacies, and correlations among study variables
Full size table

Intra-class correlations

For commitment, the ICC was 0.637. This result indicates that 63.7% of the variance in commitment can be explained by differences between persons, and 36.3% of the variance can be explained by fluctuations within a person. For academic self-efficacy, the ICC was 0.673. In comparison, the ICC for time management was 0.587. All ESEM factors taken together consisted of within-person variance to a considerable extent, supporting the use of a RI-CLPM. However, it turns out that time management consisted of relatively more trait-like characteristics in contrast to commitment and ASE.

Model comparison

Table 4 displays the fit indices and results of the model comparisons. The detailed model comparison is descripted in the appendix (see Appendix E).

Table 4 Fit indices and model comparisons for nested models
Full size table

The assumed model displayed in Fig. 2 showed good absolute model fit indices (M10: \(\upchi\)2(10) = 24.99, p < 0.01; RMSEA = 0.051; CFI = 0.991; TLI = 0.968; SRMR = 0.023). Additionally, we added the control variables to this model to gain our final model (M11) with mostly improved model fit (\(\upchi\)2(10) = 19.69, p = 0.03; RMSEA = 0.042; CFI = 0.996; TLI = 0.954; SRMR = 0.011). For clarity, the influence of the control variables on the constructs of interest is not shown in Fig. 2. But Table 6 displayed the influence of all control variables on all constructs of interest (see Appendix F). It turned out that only the cohort, the age of the students and their satisfaction with studies showed significant relationships to students’ commitment.

Fig. 2
figure 2

Random intercept cross-lagged panel model (RI-CLPM) of the relationships between academic self-efficacy, time management, and commitment across three waves

Full size image

Final model

At the between-person level, moderately strong correlations emerged between the random intercept factors of academic self-efficacy and time management (\(\upbeta\) = 0.614) and academic self-efficacy and commitment (\(\upbeta\) = 0.338), but not between time management and commitment (Table 5).

Table 5 Final model estimates with included control variables (cohort, gender, age, final school grade, major of studies, and satisfaction with studies)
Full size table

This finding indicates that individuals who reported higher stable trait-like levels of academic self-efficacy would likely report a higher stable trait-like level of time management and are more decided about their studies in general. But these results indicate that there is no relationship of time management and commitment on the between-person level. At the within-person level, all constructs of interest seemed to be reasonably stable over the first year (commitment (\(\upbeta\)T1-T2 = 0.364; p < 0.05;\(\upbeta\)T2-T3 = 0.260); time management \((\upbeta\)T1-T2 = 0.514; \(\upbeta\)T2-T3 = 0.605; p < 0.05); academic self-efficacy \((\upbeta\)T1-T2 = 0.302; \(\upbeta\)T2-T3 = 0.428; p < 0.05)). Against our expectation, cross-lagged paths from time management to commitment were only significant over the second semester (\(\upbeta\)T1-T2 =  − 0.051, p = 0.48; \(\upbeta\)T2-T3 = 0.468, p < 0.05), and both cross-lagged paths from academic self-efficacy to commitment were significant (\(\upbeta\)T1-T2 =  − 0.133; \(\upbeta\)T2-T3 =  − 0.205, p < 0.05). However, contrary to expectations, the effects were negative. That outcome surprisingly indicates that at the within-person level, high levels of academic self-efficacy led to lower commitment afterwards.

Considering the cross-lagged paths from time management on academic self-efficacy, we found significant effects over the second semester but not in the first (\(\upbeta\)T1-T2 = 0.104, p = 0.14; \(\upbeta\)T2-T3 = 0.292, p < 0.05). Furthermore, cross-lagged paths from academic self-efficacy to time management were not significant (\(\upbeta\)T1-T2 = 0.034, p = 0.38; \(\upbeta\)T2-T3 = 0.055, p = 0.38).

Notably, significant positive relationships emerged at the within-person level between commitment and time management at T1, T2, and T3, while there was only a significant positive relationship between commitment and academic self-efficacy at T2 and between time management and academic self-efficacy at T2 and T3.

Main analyses

After the robustness checks, we tested our hypotheses by adding constraints to our model. Interestingly, by constraining paths within the first semester and the second semester, we found that the paths from academic self-efficacy to subsequent academic self-efficacy, time management and commitment were the same for the first and second semesters without any decrease in model fit, thus rejecting H1a, H1b, and H1c. That said, it can be noted that the effects of academic self-efficacy on commitment and subsequent academic self-efficacy were somewhat greater in the second semester. However, further constraints led to a worse model fit. We found significant effects of time management on commitment and on academic self-efficacy over the second semester but not over the first semester. Additionally, the effect of time management on subsequent time management was slightly stronger in the second semester, supporting H2a, H2b, and H2c.

Discussion

By using the RI-CLPM our study contributes to prior research in two ways: First, we differentiated between trait-like between-person level and state-like within-person fluctuation over time and examined for the first time the dynamic interplay between time management and academic self-efficacy and their influence on commitment on the intra-individual level. Second, by using three measurement occasions, we examine the extent to which these relationships differ in the first and second semester.

In line with prior research on time management (Bowman et al., 2019; Haarala-Muhonen et al., 2017; Limone et al., 2020), our results highlight the relevance of time management for students’ development over time, especially in the second semester, as on the intraindividual level the students’ time management seemed to impact the level of their commitment at the end of the first year. In contrast, individual academic self-efficacy showed a smaller and unexpectedly negative relationship to subsequent study commitment on the intraindividual level, which is in contrast to the time-independent positive relationship of academic self-efficacy with study commitment on the interindividual level.

Theoretical implications

We have expanded prior research on students’ first year experience in higher education and development based on social cognitive theory by comparing the dynamic interplay of time management and self-efficacy in the first and second semester as well as distinguishing trait-like differences between individuals and fluctuation in the development of constructs over time within an individual.

For all constructs of interest, we found variance between persons and within persons, indicating that all constructs can be separated in a more trait-like component and a more fluctuating situational component. Prior research mostly considered the trait-like component (e.g. Chesnut, 2017; Girelli et al., 2018).

As expected, the trait-like components of academic self-efficacy and commitment were positively related (Chesnut, 2017; Chesnut & Burley, 2015). However, mention should be made of the small but negative relationship between academic self-efficacy and subsequent commitment that occurred in both semesters within an individual. This seems to be contrary to previous research on pre-service teachers’ self-efficacy and commitment (Chesnut, 2017; Chesnut & Burley, 2015; Chesnut & Cullen, 2014), but this research has dealt only with the trait-like differences between persons. A possible cause for our results could be due to the fact that pre-service teachers with higher measured academic abilities might be more likely to leave their chosen studies (Guarino et al., 2006). It could be that pre-service teachers who experienced an unexpected high degree of academic self-efficacy over the course of their first year at university doubt their initial career decision and consider another course of study that they previously labelled too difficult. Because of the strong positive relationship of the trait-like components, however, this should be interpreted with caution. Regardless of this, we have to note that the relationships of academic self-efficacy to subsequent commitment, time management, and academic self-efficacy are the same in the first and second semester. Thus, our results indicate no stabilisation process of students’ academic self-efficacy during the first year in higher education and its influence on commitment, as research suggested (Gore, 2006; Gorges, 2019; Kantanis, 2000; Putwain et al., 2013, Wright et al., 2013).

Contrary to our expectations, we found no relationship between time management and commitment on the between-person level, indicating that students with high time management skills are not in general more or less committed to their studies than students with lower time management skills. Interestingly, the constructs were related to each other on all occasions on the within-person level, indicating that a student who reported high time management skills at one time likely reported high commitment to studies at the same time. This fits into the picture painted by previous research (Bowman et al., 2019; Perander et al., 2020): Students are more committed to their studies and experienced less stress if they have positive experiences regarding their time management during their studies. This relationship appears to fluctuate over time and to be situational rather than general. Additionally, after the first and after the second semester, time management was also positive related to academic self-efficacy. This indicate that students’ situational perception of their time management skills is also directly linked to their situational academic self-efficacy after the first semester, supporting research which indicates that academic self-efficacy changed during the first semester related to students’ experience in this semester (Bandura, 1986; Gore, 2006; Gorges, 2019; Wright et al., 2013). First-year students are confronted with new demands on their time management again and again in the first year (Lindblom-Ylänne et al., 2015; Perander et al., 2020; Trautwein & Bosse, 2017), which they either master or not. Depending on their current experience, they report high or low time management skills and related high or low commitment to their studies and academic self-efficacy.

Furthermore, differences between the first and second semesters occurred regarding students’ time management and its relationship to subsequent commitment to studies, academic self-efficacy, and time management. Although pre-service students’ commitment appeared to be relatively stable over the first year, students’ time management seemed to play a prominent role in the second semester. These results are in accord with recent studies (Lindblom-Ylänne et al., 2015; Perander et al., 2020; Wright et al., 2013) indicating that students develop their time management skills and gain experience in managing their daily study routine in the first semester when they explore the new learning environment in higher education. This experience allows them to evaluate their commitment to their studies after the first semester. Students with poor time management might fail to achieve their first semester goals and as a result question their decision to pursue academic career. In contrast, students who experience their time management as successful might feel confirmed in their study choice. These findings suggest that the processes in the first and second semesters should be considered separately.

Practical implications

Our results provide guidance for universities, lecturers, and students by highlighting the relevance of time management within the first year at university and questioning the relationship of academic self-efficacy to following commitment.

Students often enter the higher education environment with a lack of understanding how to organise their study at university (van der Meer et al., 2010). This could be especially true for pre-service teachers who must study two majors at possible different faculties. Trainings and study skill guidance are recommended for the first year in higher education (e.g. Coertjens et al., 2017; Girelli et al., 2018). Our results suggest that universities could benefit from offering time management trainings to students especially in the second semester, when students have gained experience regarding how well they can manage their time at university. These trainings could be during the second semester providing students with the opportunity to apply time management methods to their daily studies, creating a sense of accomplishment and ultimately strengthening their commitment to their studies.

Lecturers could support their students by identifying students at risk of poor time management, integrating time management methods in seminars and providing support tailored to the students’ individual needs.

For students, it might be helpful to take up time management trainings, when they already know the learning environment, in which they will use time management methods (Perander et al., 2020).

Furthermore, it is of concern that higher academic self-efficacy is related to lower commitment over time, supporting that students with high academic self-efficacy tend to leave the teacher profession. Universities should be aware of this possible path. Maybe they could improve the study conditions in order to keep students satisfied with their studies, as this factor is positively related to both academic self-efficacy and commitment to studies (Hagenauer et al., 2018).

Limitations and implications for future research

Despite the sophisticated statistical analyses and longitudinal research approach, our study has several limitations. First, our sample was limited to pre-service students at one university of technology in Germany and mostly consisted of female students, which is typical for teacher training programmes. Nonetheless, generalisations of the findings to other universities and study programs should be made with caution (Hagenauer et al., 2018). At this point, transferring the investigated model to further samples and examining to what extent the negative effect of academic self-efficacy at the within-person level can be replicated would be desirable.

Second, our study examined the relationship of time management to subsequent commitment to studies. It also could be that very committed students are more engaged in their studies and thus create more sense of achievement regarding their time management, which results in higher time management skills. Thus, longitudinal research on the interplay of those, for example, with weekly measurements could provide deeper understanding of the relationship between these constructs.

Third, our study is based on self-reported measurements, raising common concerns like biased responses due to perceived coercion or inaccurate self-beliefs (Chesnut, 2017), however all constructs are measured by validated scales. Future research could investigate the extent to which these self-assessments can be verified using objective measures (e.g. grades).

Finally, although by including several control variables (cohort, gender, age, final school grade, major of studies, and satisfaction with studies) the problem of unobserved heterogeneity was addressed, the use of more objective, less distal control variables would have been desirable. Nevertheless, we were able to show that the addition of the control variables had a relevant effect on the obtained parameters and thus at least tended to serve its purpose.

Conclusion

The current findings overall substantially contribute to the research in higher education by highlighting two relevant findings on pre-service teachers’ first-year experiences, supporting the relevance of this study phase (Trautwein & Bosse, 2017; van der Meer et al., 2010). First, the results of our analysis show that different processes take place in the first and second semesters: In the second semester, students’ time management seems to have a significant impact on their following commitment. Second, it is worthwhile to distinguish between the trait-component between persons and the fluctuating component within a person within constructs of interest. For academic self-efficacy, we found contradictory results at these levels that require further research. From a practical point of view, these results are particularly relevant to universities and teaching staff to support their students to foster their commitment to their studies.