Introduction

Higher education (HE) institutes and educators are constantly seeking effective pedagogical approaches that enhance inclusive practices in HE. The increasing diversity reflected in the student population carries with it increasingly diverse educational requirements. Importantly, the Salamanca statement demands equal education for all while concurrently considering their individual needs (United Nations, 1994). Still, in HE, many students with learning support needs encounter different challenges and barriers during their educational journey (e.g., García-González et al., 2021). Particularly, students with learning disabilities invest more time and effort in their studies, and the challenges they face can negatively impact their quality of life (Lambert & Dryer, 2018) and increase the risk for different mental problems (Livingston et al., 2018). Nieminen (2023) noted how HE testing and even accommodation create disability and barriers for inclusion.

The National Joint Committee on Learning Disabilities (2016) defines learning disabilities as a heterogenous group of disorders that pose various challenges to learning and exerting listening, speaking, reading, writing, reasoning, or mathematical skills. In Finland, the Non-Discrimination Act (Finnish Law, 2014) obligates education providers to promote equality with reasonable adjustments, and in the Finnish HE context, efforts primarily seek to achieve this through accessibility (Korkeamäki & Vuorento, 2021), similar to many other countries worldwide (Barnes, 2007) with the means of removing barriers to, and limitations of, education for all people. This is in line with the social model of disability that views disability as a product of interaction with the environment (Nieminen & Pesonen, 2020; see also Corcoran et al., 2015). Furthermore, gaining support is not always easy (Lehto et al., 2019). Lecturers may lack appropriate knowledge to support students (Ryder & Norwich, 2019), and students themselves may not seek aid for learning adjustments in fear of stigma from other students and the staff (Pino & Mortari, 2014). Thus, more research on pedagogical approaches in HE from the perspective of student support is needed. One such approach may be the flipped classroom (FC; Bergmann & Sams, 2012), but pertinent study is scarce. Here, we investigated whether students with a history of receiving support in their previous educational context perceive FC differently than those who have not. Moreover, we sought to identify the factors that may influence students’ perceptions about the FC approach.

Learning disabilities in higher education

In 2021, 34.6% of students reported a health or functional limitation in the learning context (e.g.,Korkeamäki & Vuorento, 2021; Parikka et al., 2021). According to Lehto and colleagues (2019), accessibility in Finnish HE has improved but is still lagging in its ability to account for individual needs. Furthermore, Korkeamäki and Vuorento (2021) found that of those students in HE reporting health or functional limitations, 62.2% felt that they limit their studies, and 40.7% felt that they do not receive enough support. In fact, obtaining accommodations for disabilities itself contributes to many difficulties, including the considerable effort the process demands (Nieminen, 2023).

The most common learning challenge among HE students in various countries is dyslexia (Higher Education Statistics Agency, 2021; Parikka et al., 2021). While about 8% of HE students in Finland have been diagnosed with dyslexia (Parikka et al., 2021), the need for support may be higher, as the prevalence of dyslexia is estimated to be around 5–20% (Wagner et al., 2020). Additionally, since reading talent is normally distributed, diagnosed disabilities are a matter of cutoff points at the tail end of performance measures (Wagner et al., 2020). Not everyone who needs support in reading necessarily has a dyslexia diagnosis. In the Finnish comprehensive education context, pupils are not required to have an official dyslexia diagnosis to receive intensifying support for reading difficulties (Finnish National Agency for Education, 2015). Thus, from a historical perspective, students may have received assistance without an official medical or psychological diagnosis. However, students in HE often recall whether they have received support for reading or learning in general, especially as the subject matter becomes more difficult and the amount of requisite reading and elaboration increases. Some students have reported that difficulties in reading and writing led to a slowdown in their studies (Pirttimaa et al., 2015). To this point, Altemueller and Lindquist (2017) proposed FC as a beneficial strategy for students with learning disabilities, where their individual learning differences would not be as visible compared to a traditional teaching modality. However, more research is warranted.

Flipped classroom

FC is a pedagogical approach that sequences the traditional lecture method in a more meaningful way (Bergmann & Sams, 2012). Students learn basic elements on their own, before face-to-face meetings, and then deepen their understanding with the help of an expert instructor (Carbaugh et al., 2016). Often, pre-class material comprises videos or texts that provide rudimentary subject matter information, and class time is used to answer student questions and engage in relevant guided and independent activities (Bergmann & Sams, 2012). Äikäs (2021) suggest that the structure of FC gives students the time they need to comprehend the subject: if they need refreshing, they can go back to the pre-class material, and in the class, they can receive feedback and support from other students and instructors. In this study, the University of Eastern Finland trained their teachers in using FC for course instruction (Sointu et al., 2021), particularly because this approach may help teachers to know their students better, interact with them more, and differentiate learning content based on individual needs (Bergmann & Sams, 2012). Notably, students have indicated that their teachers are better experts in teaching with technology, and the majority seems to be satisfied with the FC approach (Sointu et al., 2019).

There are numerous studies on the effectiveness and outcomes of FC. Several reviews associate FC with better academic achievement (e.g., Akçayır & Akçayır, 2018; Al-Samarraie et al., 2020; Galindo-Dominguez, 2021; Zainuddin et al., 2019), as well as improved student engagement/motivation (Al-Samarraie et al., 2020; Galindo-Dominguez, 2021; Zainuddin et al., 2019), better understanding (Al-Samarraie et al., 2020), and better collaboration (Galindo-Dominguez, 2021). Some reviews also noted challenges for FC, and all were related to student preparation before class (Akçayır & Akçayır, 2018). While Zainuddin et al. (2019) highlighted the lack of motivation to watch videos or study outside class, Al-Samarraie et al. (2020) reported that the breadth of pre-class materials and the time required to study them were the issue. However, the following aspects support learning with the pre-class materials; adequate instruction for the use of materials before the class and how teachers can continuously assess the students’ knowledge also in this phase (Sointu et al., 2023a) as well as the use of learning analytics data to observe the use of materials (cf., Sointu et al., 2023b). Although these challenges might be greater for students with learning disabilities (e.g., Pirttimaa et al., 2015), there is some evidence that FC might in fact be beneficial for them (Altemueller & Lindquist, 2017). Attitude-wise, students have rated FC teachers and courses more positively compared to the traditional model, and overall, students hold positive views of FC (Sointu et al., 2019), describing a better learning experience in the FC context (Steen-Utheim & Foldnes, 2018). Still, the amount of research on students with learning disabilities in the FC setting is scarce, so we do not readily know if they obtain the same benefits or view FC to be more challenging.

Learning factors

Many factors have been linked to positive outcomes in academics in addition to FC. In this paper, we refer to them as learning factors—existing factors relating to an individual that, despite being independent of the FC course, can affect individuals’ learning and experience of the course. The first factor, reflective learning, has been associated with higher academic achievement (Köseoglu, 2016). Rogers (2001) defines reflection as (1) a cognitive process that requires the learners’ active engagement; (2) the consideration of learners’ responses, assumptions, and assertions in its context; and (3) the consequent integration or assimilation of new knowledge. In a study by Lonka et al. (2021), a reflective-collaborative group achieved the best academic grades and accumulated the most credits. The second factor, collaboration, has been connected to many positive learning outcomes such as better skill acquisition and an enhanced sense of community (Mitra, 2022) and social cohesion (Gratton, 2019). Cen et al. (2016) found that students outperform their individual performance expectancy when working in collaboration. The third factor, self-regulation, can be defined as the ability to alter dominant responses to promote desirable behavior (de Ridder et al., 2012). Differences in self-regulation have been found to impact differences in academic performance (Duckworth et al., 2019). Related to self-regulation, the fourth factor, self-efficacy, is one’s belief in his or her capabilities to handle specific situations or challenges (Bandura, 2006). In some studies, students’ self-efficacy for learning has been found to relate to academic achievement (Ayllón et al., 2019), mediating the relation of resilience to academic performance (Supervía et al., 2022), and to relate to motivation and homework procrastination (Katz et al., 2014). However, self-efficacy for learning does not necessarily predict future performance, but is an indicator of past performance (Sitzmann & Yeo, 2013). The fifth factor, the use of information and communication technology (ICT) in learning, is increasing in both classroom activities and home assignments (Pegler et al., 2010). A positive connection between academic achievement and ICT literacy has been found (Lei et al., 2021), specifically that interest and perceived competence in ICT is positively related to academic achievement (Park & Weng, 2020).

There are other important learning factors that relate to academic performance (e.g., intelligence; Vazsonyi et al., 2022), but in this study, we focused on the aforementioned factors in their relation to FC. Chickering and Gamson (1999) indicated seven principles for good practices in HE, which are also noted by Bergmann and Smith (2017) as key points in why FC works. Two particular principles are encouraging active learning and giving prompt feedback. As a reflective learner is active and considers experiences in their context to form new understanding, both principles suit a reflective learner. Another principle is encouraging collaboration among students. Both Bergmann and Smith (2017) and Carbaugh et al., (2016) promote collaboration as being of paramount significance in FC, and this also seems to be the case in practice (Koh, 2019). Moreover, Sointu et al. (2023a) has indicated that successful FC requires guidance and the meaningful pedagogical skill of the teacher, which also includes the ability to combine theoretical and practical knowledge, and to create a safe learning environment, and an understanding of students’ technology skills. Zainuddin et al. (2019) found that the biggest challenges with FC concerned the students’ work on their off-classroom activities. This could be related to a lack of self-regulation (cf. Lai & Hwang, 2016). For example, lack of self-regulation and higher task avoidance may pose difficulties with the FC approach for students (Hyppönen et al., 2019), which may be particularly challenging for those students with a history of learning disabilities. However, FC has been shown to be effective in improving students’ self-regulation (Jdaitawi, 2019). FC has also been shown to have a positive effect on students’ self-efficacy for learning (Galindo-Dominguez, 2021), and is linked to students’ effective learning in online environments (Prior et al., 2016) and positive engagement in online learning (Deng et al., 2022). The use of ICT is common and ever increasing in HE in general and in FC in particular (e.g., Sointu et al., 2019). There is a long tradition of using ICT in education to support students’ collaborative and self-regulative learning practices (Cress et al., 2021). Especially within the FC model, the role of ICT as a platform for learning and as a way to deliver the pre-materials and assignments is vital. This again demands readiness and skills from the students in order to study with different ICT facilities (Sointu et al., 2019). In this framework, FC may influence those students in HE who need support. Still, studies on the effects of FC for students who have history of learning disabilities are severely lacking. Thus, the purpose of this study was to investigate how students with a history of received support related to reading on a pre-HE level view FC in HE compared to those students without a history of support. For this, we investigated (1) if student academic performance in FC is related to a history of received support, including between genders, (2) if there were any differences in learning factors or experiences related to FC among students with and without a history of support, and (3) the learning factors and FC experiences for developing a model to explain the academic performance of students with a history of support in an FC course.

Methods

Participants

The convenience sample included 1592 HE students (Nfemale = 1129, 70.9%; Nmale = 463, 29.1%) from the University of Eastern Finland. Data were collected from informed, voluntary students participating in FC courses during the years 2016 to 2018. Participation in the study did not affect the grading or completion of the related course. The participants’ age ranged between 18 and 69 years (Mage = 26.5, SDage = 8.18). Participants were taught in 46 independent courses in varying majors such as biology, chemistry, economic sciences, literacy, pharmacy, physics, and education. Data were collected via electronic questionnaires at two timepoints: background information including the received support status was self-reported at the beginning of the courses and other measures were collected at the end of the courses. To better understand the nature of our convenience sample, we tested for statistically significant differences with the available data. No statistically significant differences (p > 0.05) were found between grade and gender between responding and non-responding participants. Out of 1592 participants, 1367 (85.9%) had not received support and 225 (14.1%) had received support. All gathered information was protected according to the EU’s GDPR (2016) and National Data Protection Act (Finnish Law, 2018) institutional data protection guidelines, and Finnish National Board on Research Integrity (TENK) guidelines for the ethical principles of research with human participants (TENK, 2009) were strictly followed. Response rates between different courses varied from 22 to 90% of course participants. All courses’ teachers were new to the FC approach and were trained to apply the approach by a team of experts (e.g., planning, pre-materials, assessment and ICT) (Sointu et al., 2021). Instructors were specialists in the specific scientific content area (i.e., majority had a doctoral degree); however, not all had university pedagogical studies background. For this fact, the implementation of FC was seen suitable for instructors to develop their teaching and the course. The training involved curriculum work (e.g., how to sequence teaching to follow FC), technology in education (e.g., benefits and drawback of ICT, and how use software and hardware), students’ guidance and counselling (e.g., how to instruct FC and use of ICT to student), assessment practices (e.g., how to transform assessment from traditional to formative), and pedagogical practices (e.g., what approaches can be used in-class with students) (Sointu et al., 2023a). The instructor also developed their actual course based on training and the study was implemented in these courses.

Measures

For the first research question, achievement-related information was interpreted from course grading. Some courses gave out numerical scale assessments ranging from 0 to 5, but others were only pass or fail. Both instances were considered in the analysis. For the second and third research questions, data were collected with various measures previously used in the Finnish context and proven to be psychometrically sound. For learning factors, we used two measures from Vermunt’s (1994) Inventory of Learning Styles (ILS) relating to regulation strategies: (1) learning content self-regulation and (2) lack of self-regulation; two measures from Pintrich’s (1991) Motivated Strategies for Learning Questionnaire (MSLQ): (1) reflective learning and (2) self-efficacy for learning; one instrument for measuring collaborative learning from Wang et al. (2009); and one instrument for ICT performance expectancy (Chen, 2011); students’ experiences in FC (Sointu et al., 2023a) were measured with five subscales: (1) satisfaction with the flipped classroom, (2) experienced difficulty of the flipped classroom course, (3) studying the pre-materials together, (4) familiarization with the pre-materials, and (5) guidance to the flipped classroom method. Further information about measures is provided in Appendix 1.

Data analysis

The grading system used in Finland is comparable to the European Credit Transfer and Accumulation System (ECTS), with a grading of 0–5 compared to ECTS’s scale of F–A, but the University of Eastern Finland also uses pass or fail grading in some courses (University of Eastern Finland, 2023). We used the 0–5 grading system in all analyses, except for the second part of the first research question. (1) Grading 0–5 was analyzed with the independent samples t-test. The distribution of grades was not normal, so we used bootstrapping (Field, 2018) with 2000 samples and a bias-corrected accelerated confidence interval (Efron & Tibshirani, 1993). (2) The data for passing or failing an FC course was categorical in nature, and the relation of either category and membership in received support or not received support were analyzed with crosstabs and chi-square tests (Field, 2018). For that we used a new variable that combined the grade variable and the pass-fail mark into a dichotomic variable, where failing grades (0) and failing marks were given a value of 0, and passing grades (1–5) and passing marks were given a value of 1. After all of the data were examined as a whole, it was split based on gender and the tests were performed again. With this new categorical value, we examined if received support or gender were related to passing or failing the course.

For the second and third research questions, the following three steps were taken to investigate the measures’ psychometric properties. First, we ran exploratory factor analysis (EFA) for initial structural validity. We made two models, with the first comprising the learning factors (24 items) and the second comprising the experiences in FC (18 items). Oblique rotation was used to allow the factors to correlate with each other. Eigenvalue > 1 and scree plot were used in assessing the suitable number of factors (Field, 2018). The Kaiser–Meyer–Olkin (KMO) measure for sampling adequacy was expected to be over 0.8 and Bartlett’s test of sphericity to be significant (p < 0.01). Communalities needed to be over 0.2 and item loadings over 0.32 (Field, 2018). Second, after finding suitable EFA solutions, internal consistency (reliability) of factor solutions was tested with Cronbach’s alpha (α). Nunnally and Bernstein (1994) α > 0.70 criteria was used to interpret adequate internal consistency. Third, based on the information from the EFA and alpha, composite scores were calculated. We used mean composite score to keep the original metric of the measures.

For the second research question, we investigated the differences in learning factors or experiences related to flipped courses among students with received reading support and those without a history of support with bootstrapped t-tests, similarly as for research question 1. For investigating effect sizes, we used Cohen’s (1988) D effect size (ES) to investigate the magnitude of possible differences with the criteria of D ES < 0.2, no effect; 0.2–0.5, small effect; 0.5–0.8, intermediate effect; and > 0.8, large effect. For the third research question, we utilized linear regression (enter method) to make two models to predict the academic performance of those students who received support in their history. We made an initial calculation for the model to examine the residuals. Normality and homoscedasticy seemed acceptable. Multicollinearity was not observed, and the VIF statistic was at the highest under 2.4, and the tolerance was over 0.42 at the lowest. After casewise examination, we excluded one case with a standardized residual over 3. The next run made the model better, but again we excluded two more cases for optimal fit. The second linear model consisted of only two variables, with the worst variable and closely correlated variable being excluded. The same procedures were followed for the second model. Durbin-Watson (Durbin & Watson, 1950) was close to 2, plots were normal and homoscedastic, and VIF was just over 1. Initially, all cases of those who had received support were included, and after excluding two problematic cases, there were no standardized residuals over 3 and the model seemed fit. Linear regression was run with bootstrap 2000 samples and BCa. Data were analyzed with SPSS v27.

Results

The first research question compared the academic performance of students with and without a history of received support in FC courses. T-tests showed no statistically significant differences between the two groups (Table 1). The same investigation was made between female and male students, and no statistically significant results were found.

Table 1 Bootstrapped t-tests for FC course academic performance
Full size table

Second, we tested whether there were differences between the groups in passing or failing an FC course. With gendered examination, all cells were equal, without significant differences in females (χ2 = 0.003, p = 0.954) or males (χ2 = 0.012, p = 0.913). Thus, no statistically significant differences in students’ academic performance existed.

For investigating the learning factors and FC experiences, we ran exploratory factor analysis (EFA) and tested internal consistency with Cronbach alpha (α) to calculate mean composite scores. The results of EFA and α are presented in Table 2. For learning factors, EFA initially suggested a four-factor solution based on eigenvalue, but after examining the scree-plot, we ended with a similar five-factor solution as in previous research (Sointu et al., 2023a). KMO was adequate (0.85), BTS significant (p < 0.001), communalities ranged from 0.29 to 0.79, and the solution explained variability by 54.5%, all indicating appropriate structure. For FC experiences, KMO was 0.89 and BTS significant (p < 0.001), and communalities ranged between 0.32 and 0.87. The alphas were good for all factors (α > 0.70; Nunnally & Bernstein, 1994); thus, composite scores were calculated.

Table 2 Exploratory factor analysis for initial construct validity, and internal consistency (Cronbach’s α)
Full size table

Table 3 presents bootstapped t-tests for learning factors between students of no-support and received-support groups. Results indicate three significant differences between the groups. The students who have received support favored collaborative working more (p < 0.01; d =  − 0.288), had more lack of regulation (p < 0.05; d =  − 0.237), and lower self-efficacy for learning (p < 0.01; d = 0.323). The comparisons for FC experience variables yielded statistically significant differences in experienced difficulty of FC (p < 0.05; d =  − 0.192) and collaborative FC (p < 0.01; d =  − 0.260), indicating that students with a history of support experienced FC to be more difficult, and they favored more collaborative study methods during the FC courses. All effect sizes were small, according to Cohen (1988) criteria.

Table 3 Bootstrapped t-tests for learning factors and FC experiences
Full size table

The third research question examined if learning factors and/or FC experiences could form a model to predict FC course performance (i.e., dependent variable) of students who had received support. The variables were selected based on significant correlations (p < 0.05) to the dependent variable, so four variables were included: satisfaction with FC (rs = 0.342; p < 0.01), guidance for FC (rs = 0.327; p < 0.01), reflective learning (rs = 0.260; p < 0.05), and lack of self-regulation (rs =  − 0.312; p < 0.01). Model 1 was statistically significant (F (4,78) = 10.815; p < 0.001) and explained 36% of the variance (R2 = 0.357) (Table 4). Out of the coefficients, only lack of self-regulation had a significant independent effect (β =  − 0.302; p = 0.002) with guidance for FC falling just short of 5% significance. We then decided to make another model with lack of self-regulation and the next best predictor, guidance for FC. Model 2 (Table 4) was also statistically significant (F (2,86) = 19.687; p < 0.001), and it explained 31% (R2 = 0.314) of the variance. For the second model, both guidance for FC (β = 0.426; p < 0.001) and lack of self-regulation (β =  − 0.301; p = 0.001) had a significant independent effect for students with received support.

Table 4 Bootstrapped linear regression models for RS students’ FC achievement
Full size table

Discussion

The aim of this study was to investigate how students with a self-reported history of receiving support for learning view the flipped classroom (FC) approach. For this, our aim was to investigate academic performance, learning factors, and FC experiences between students with and without a history of support. Additionally, we tested predictive models for FC course performance for students who had received support. There were no differences in FC course performance in grading or in failing between students with a history of received support and those who had not received support, when examined together or as gender groups. For learning factors, students with a history of support needs preferred collaborative working and had lower assessments of self-regulation and self-efficacy for learning. Out of the FC experiences, they more often experienced FC as more difficult and studied pre-class materials in collaboration. In predicting the academic performance of students who had received support, the first model consisted of learning factors—lack of self-regulation and reflective learning—along with FC experience factors—guidance for FC and experienced satisfaction with FC. This model explained about a third of the variation in achievement, though only lack of self-regulation had a significant independent effect. Guidance for FC fell just short of significance, so we formed a second model around these two variables, which explained about 30% of the academic performance variance. According to these results, a significant amount of the performance variance of students who had received support can be accounted for by a lack of self-regulation and guidance for FC.

Research on FCs’ impact on students with learning disabilities is still limited, but Altemueller and Lindquist (2017) proposed that FC can be beneficial. The results of this study partially support this notion, as students with a support history performed as well as others which is not always the case (Showers & Kinsman, 2017). However, many different barriers can hinder students with learning disabilities (e.g., García-González et al., 2021), and studies have shown the importance of support in HE for students with learning disabilities (Sarid et al., 2020). Niazov et al. (2022) found that students with learning disabilities experience more difficulties in HE and have lower academic self-efficacy, which were also found in this study with the addition of lower self-regulation. FC may be considered as challenging because it requires self-regulation from the students (Hyppönen et al., 2019; Lai & Hwang, 2016), e.g., during off-classroom activities (e.g., Zainuddin et al., 2019). Moreover, although self-efficacy for learning is mainly an indicator of past performance and does not alone predict future performance (Sitzmann & Yeo, 2013), it is one’s belief in their capabilities to handle new challenges and is the main reason for procrastination with homework (Katz et al., 2014). In previous studies, most of the challenges in FC were related to students’ work before class (Akçayır & Akçayır, 2018; Zainuddin et al, 2019); low self-regulation and low self-efficacy for learning could therefore be a significant disadvantage for studying in FC, especially since FCs fundamentally rely on the student’s self-regulated effort prior to class. Despite the lack of self-regulation and low self-efficacy for learning that might most strongly show in engagement with the pre-class tasks where external regulation is at its lowest, there was no difference in pre-class effort. Mason et al. (2013) noted that most students took some time to internalize the need to be well prepared for class and credited this change to self-regulation. Thus, FC could be a bad fit for students with low self-efficacy for learning and low self-regulation, but past studies show that the opposite might actually be true. FC has been related to enhancing students’ self-efficacy for learning (Galindo-Dominguez, 2021) and to improving students’ self-regulation (Jdaitawi, 2019), and though lack of self-regulation can predict worse academic performance (Duckworth et al., 2019), in our study students who had received support performed as well as their peers in passing courses and achieving similar mean grades above 3 (scale: 0–5), which is described as good in the University of Eastern Finland. Nevertheless, based on the results, more emphasis should be placed on supporting self-regulation and self-efficacy of students, particularly students with history of support. If these aspects are taken into consideration in the training of instructors (cf., Sointu et al., 2023a) and use of learning analytics (cf., Sointu et al., 2023b), it may improve outcomes.

While the abovementioned results demonstrated challenges for students with a history of learning disabilities, it is also suggested that FC provides opportunities (e.g., needed time to comprehend the subject, ability to revisit the materials, support from peers and teacher) for all students (e.g., Äikäs, 2021). One major finding in this study was that students who had received support are more inclined toward collaboration in FC and overall. This may be considered as one pedagogical support approach in HE settings. Previously, FC has been found to increase collaboration (Galindo-Dominguez, 2021), which is one of its key components (e.g., Bergmann & Smith, 2017; Carbaugh et al., 2016). Collaboration has been associated with better performance in academics (e.g., Cen et al., 2016; Lonka et al., 2021), and it can be a source of social support that also predicts academic success (Musso et al., 2020). Previously, in a comparative study of FC, with and without collaboration and traditional lecture-based classroom, FC setting with collaboration yielded the best academic results (Foldnes, 2016). We posed that lack of self-regulation and lower self-efficacy for learning could cause challenges in FC course performance and pre-class tasks, but there were no such effects, and collaboration could be the interfering factor. The aforementioned self-regulation improvement was attributed by Jdaitawi (2019) to interactions in FC, and although Sun et al. (2017) did not find improvement in self-regulation, they found that there was more help-seeking. It may be, due to lack of self-regulation, lower self-efficacy and experienced difficulties in FC, that students with history of learning support sought out more collaborative learning methods as they identified their need for collaboration and to utilize collaborative support (i.e., co-regulation) more meaningfully. In Steen-Utheims and Foldnes’ (2018) study, students described how more talented peers aided them in their learning. Identifying one’s shortcomings and seeking help in collaboration can be a sound strategy for students with learning disabilities. Students who had received support might have adapted to their challenges in traditional pedagogical approaches by increasing collaboration in their studies. Working on their pre-class assignments with other students could counter the effect of low self-efficacy for learning and lack of self-regulation. The social support aspect of collaboration could also counter the effect that the experienced difficulties might have had.

The linear models explained about a third of the academic performance variance among students who had received support. In model 2, Lack of self-regulation and guidance for FC were both related to the achievement of students who had received support with opposite directional effects, with guidance’s degree of change (beta coefficient) being greater. This could be interpreted to mean that good instruction in FC can outweigh a lack of self-regulation’s negative effect on students’ academic performance. Mason et al. (2013) found that students needed some time to adjust to the FC setting, and for that they needed good instructions and guidelines. The importance of instruction to a new course format seems obvious, and the importance of guidance in the FC setting has already been recognized (e.g., Sointu et al., 2023a), but these findings underline the need especially for students with disabilities in learning. Many HE students already feel that they do not receive enough support during their studies (Korkeamäki & Vuorento, 2021), so a lack of instruction for an unfamiliar pedagogical approach could be severe. In addition, and particularly, the assessment culture in HE sets various barriers for disabled students (Nieminen, 2023). Thankfully in FC settings, more of the instructor’s time is freed up to guide and help students solve problems and deepen their understanding of the subject matter. This has led us to an approach that is associated with better learning results, where students having the possibility of reviewing pre-class material, help from collaboration and quality guidance from instructors led to no differences for achievement between students who have and those who have not received support. Still, even without performance differences, sufficient support is still required, for people with learning disabilities are a heterogeneous group with varying needs. Lecturers hold mostly positive attitudes for reasonable adjustments for students but may lack the knowledge to support them (Ryder & Norwich, 2019). Thus, we need professional educators adept with FC (Sointu et al., 2023a) and knowledgeable of varying students’ needs for the approach to be successful, as the proper implementation of FC with quality instruction and guidance may mean inclusive HE.

Limitations and future research

Limitations in the study exist. First, we had a convenience sample. In the future, randomized samples and experimental designs would benefit the field. Even though the differences between responding and non-responding participants were statistically non-significant with grade and gender, future research should use additional variables to investigate possible biases. Second, the received support status was informed by the respondents themselves. Thus, in the future, more detailed information about support status should be acquired, perhaps from the actual documentation. Moreover, actual support needs in higher education should be mapped and investigated in line with the FC approach. Third, the data were gathered from only one university. Thus, larger representativeness is needed in the future. Fourth, as more information about new insights from the support status and FC perspectives are now available, future research should consider more in-depth analyses such as measurement modeling, structural equation modeling and multilevel modeling. Additionally, understanding FC and support from qualitative and mixed-methods research would benefit the higher education field. Finally, there are many factors affecting learning, and we chose just some of them. In the future, various perspectives of FC and support should be investigated.

Conclusion

This study has provided some initial findings regarding how FC courses in HE impacts students with a history of learning disabilities. Students who had received support performed as well as their peers, but experienced the course to be more difficult, though there was no control condition to which FC could be compared. They also had a higher lack of self-regulation and lower self-efficacy for learning. Interestingly, the students who had received support favor collaboration in HE in general and in FC in particular. One finding is the predictive model for academic performance of students who had received support. It showed that guidance for FC and lack of self-regulation are significant predictors. Both components’ importance has been recognized individually in previous studies, but with students with learning disabilities in the FC setting, it can be even more so. A well-organized and well-taught FC course may be accessible for all students. However, research on the suitability of FC for all students requires more research with robust designs. Still, these initial results are encouraging in the task of ensuring inclusive HE for all students.