Use Of Video Materials on a Flipped-Classroom Physics Course

Practice Paper Recommended citation: Kontro, I., & Kuuluvainen, H. (2025). Use Of Video Materials on a Flipped-Classroom Physics Course. In Kangaslampi, R., Langie, G., Järvinen, H.-M., & Nagy, B. (Eds.), SEFI 53rd Annual Conference. European Society for Engineering Education (SEFI), Tampere, Finland. DOI: 10.5281/zenodo.17631381. This Conference Paper is brought to you for open access by the 53rd Annual Conference of the European Society for Engineering Education (SEFI) at Tampere University in Tampere, Finland. This work is licensed under a Creative Commons Attribution-NonCommercial-Share Alike 4.0 International License. USE OF VIDEO MATERIALS ON A FLIPPED-CLASSROOM PHYSICS COURSE I. Kontro a, 1 , H. Kuuluvainen b a Physics Unit, Tampere University, Tampere, Finland, 0000-0002-1874-3756 b Physics Unit, Tampere University, Tampere, Finland, 0000-0002-6003-3547 Conference Key Areas: Teaching mathematics and physics in engineering education, Digital tools and AI in engineering education Keywords: physics, flipped classroom, video ABSTRACT We studied the use of video materials on a flipped classroom engineering physics course. The course’s video materials consisted of theory videos, worked examples and motivational videos. We found that most students watched the theory videos completely or almost completely (median percentage watched 96%), but only partially the videos with worked examples (median 40%). Most students did not watch the motivational videos (median 0%). There were no statistically significant differences in the viewing metrics of students of different majors or genders. The different video types were used differently: the viewing data of theory videos had more searching and short views whereas the motivational video, when watched, was viewed for longer consecutive amounts of time. Viewing theory videos correlated moderately with exam points (Spearman’s 𝜌 = 0.43). Viewing the other types of video materials had weak or no correlations with exam points. 1 Corresponding Author I. Kontro [email protected] 1 INTRODUCTION The flipped classroom is an instructional model, in which traditional classroom activities such as lecturing are transferred to individual study at home, and the practice of the new material, such as completing assignments, are done with teacher guidance (Akçayır & Akçayır, 2018). Flipped-classroom teaching often utilizes lecture videos. Short videos are recommended (Brame, 2017). For example, based on data from a massive online open course (MOOC), Guo et al. (2014) recommend that video length does not exceed 6 minutes. However, results from MOOCs may not be generalizable to degree students, whose motivation and engagement are likely higher than for nondegree students studying online. Recently, the demand for very short videos has indeed been criticized. Seidel (2024) reported that using segments in longer videos promotes learning compared to non-segmented videos, and that video length is of lesser importance. Ahn and Chan (2025) showed that playback speed did not affect learning, but students have individual preferences for the style of the videos, which affects their engagement of videos. In introductory university physics, many topics are not easily presented in six minutes. Calculations and derivations take time, and adding context and discussing the results lengthens the video but is important for linking the results to prior knowledge. Focusing on conceptual learning is important in university-level physics. Traditional, lecture-based instruction produces lower conceptual understanding that interactive methods (Hake, 1998, Formic et al., 2010). Many active learning strategies such as flipped classroom or Just-in-Time-Teaching, which introduces pre-assignments, improve conceptual knowledge (Formic et al., 2010). However, interactive methods are not necessarily better at teaching problem-solving skills (McDaniel et al., 2016). A variation of the flipped classroom model, prime time learning, also introduces active group work and the group’s reflection with a teacher (Koskinen et al, 2018). Flipped classroom in general and prime time learning in particular has advantages for learning. Prime time learning promotes conceptual learning and improves passrates in physics (Koskinen et al, 2018). In this work, we study how engineering students use video materials on a flippedclassroom course, which utilizes videos on theory, worked examples and motivational videos. Specifically, our research questions (RQ) are: • RQ1: How do the students make use of the video material? • RQ2: Do students with different background (gender, major subject) differ in how they use the videos? • RQ3: How does watching the video material correlate with learning outcomes, as measured by exam results? 2 CONTEXT AND PRACTICAL WORK 2.1 Physics 3 Physics 3 is a calculus-based physics course for engineering students who major in technical physics, chemistry or mathematics at Tampere University. Physics 3 covers introductory electromagnetism and is based on the book Understanding Physics (Mansfield and O’Sullivan, 2020). The instruction on Physics 3 loosely follows the prime time model of Koskinen et al. (2018). The study week starts with video lectures, which include a longer video on theory, shorter videos on worked examples, and a motivational video called ”Physicists’ table”. Based on the materials, students solve pre-assignments. Students also turn in two concepts, which they are willing to explain, and two questions, which they have not found the answer to, in preparation for the prime time sessions. This process is similar to Just-in-Time-Teaching (Formic et al., 2010). During prime time, the students discuss previous weeks’ exercises and model solutions with their group of 4-6 students. They discuss the concepts and questions of all participants, and finally the solutions to the pre-exercises. The teachers participate in the discussion as needed. After the prime time sessions, students solve calculation exercises relating to the week’s material either in tutorial sessions or individually. The exercises are handed in at the end of the tutorial session, after which the students start studying for the next week (Fig. 1, top). The course’s grading includes the final exam, exercises and pre-assignments, participation in the prime time sessions and laboratory work (Fig. 1, bottom). The final exam mostly contained calculation problems, but also one essay question. Fig. 1. Top: the weekly schedule of the course. Bottom: the weights of the different grade components. 2.2 Video materials The video materials of the course are divided into three categories: 1) Theory videos, 2) Worked example videos, and 3) Motivational videos. Each week of the course includes one lecture video, 2–6 example videos, and one motivational video. In the theory videos, the theory is explained and visualized by using slides with bullet points and figures as well as visualizations and short derivations demonstrated with a touch screen. The theory videos cover all the course topics, and they follow the structure of the course handout. Each theory video is divided into parts according to the section titles and numbering of the course handout. When watching the videos, the students can see the content and section titles and they can easily jump to the beginning of each section. The length of the theory videos varies from 1h 22 min to 2h 37 min and the slides of the theory videos are available for the students as pdf files. Worked example videos demonstrate the key problem-solving techniques and applications of the core concepts of the course. One of the main purposes of these videos is to teach how to apply the Maxwell’s equations in problem-solving, which includes understanding the role of different symmetries, forming integrals, and vector field properties. The problems are solved in the videos mainly based on sketching figures and deriving equations by hand on a touch screen. The length of the example videos varies from 7 min to 29 min and the material produced in the videos are available for the students as pdf files. Motivational videos also called “Physicists’ table” videos are additional material in which some of the challenging concepts are discussed. In these videos, two physics teachers discuss 2–3 concepts per week in a classroom by using a blackboard and some demonstration equipment for visualization. The video format is meant to be somewhat entertaining and motivating. The purpose of these videos is to deepen the students’ understanding of the challenging concepts and encourage the students to discuss by themselves during the prime time. The length of the motivational videos varies from 24 min to 38 min and the content of the videos is visible for the students so that they can jump directly to the beginning of certain concepts or questions. 2.3 Participants and data The study comprised the three final weeks of the course Physics 3. The students were asked for informed consent to participate in the study. Of the approximately 80 students on the course, 56 students consented to participate, and only their data was collected. The exam points and the video view data (including time stamps, minutes watched, number of views and percentage completed by video) were collected. After combining the data, the data was pseudonymized and analyzed without identifying details. All data analysis was done in R. As the number of participants was small and the data generally not normally distributed, we used non-parametric statistical tests: the Mann-Whitney U-test for comparing two samples, the Kruskal-Wallis test for comparisons of more than two samples, and Spearman’s correlation for studying correlations between variables. 3 RESULTS AND INSIGHTS 3.1 Use of videos First we studied how students made use of the different videos (RQ1). Most students watched at least a part of the videos. As can be seen from Fig. 2, the median percentage watched was 68%. When the data is separated by video type, the theory videos have a median percentage watched of 96%. The least watched video type is the motivational Physicists’ table, which most students did not watch. However, many students also watched these videos in their entirety. Fig. 2. Boxplot of the percentage of videos watched by video type. We also looked at how many times students watched the videos. Data from two videos on week 4 show that not only did more students watch the theory video, but the viewing patterns show different qualities (Fig. 3). The heatmap of the theory video has a patchier appearance, as students have watched certain parts again. The Physicists’ table video has a more uniform watching pattern, but most students have skipped this video entirely. Fig. 3. Heatmaps of the different student’s watch count of video segments on week 4. A: theory lecture video, B: motivational physicists’ table video. Dashed lines indicate the position of search marks. It is apparent from the heatmaps that students click around particularly in the theory video. To study the duration of each consecutive view, we extracted the minutes delivered for each click of the theory video on week 4. The histogram of the duration of the delivery is shown in Fig. 4. The majority of clicks result in very short views of a few seconds (Fig. 4, left panel). However, when we filter out the shorter views below 6 minutes (Fig. 4, right), we see that the duration varies a lot. The cut-off of 6 minutes was chosen, as this was the mode of viewing time in Guo et al. (2014). A few students watch the video in one go, but the majority of students watch the video in pieces. However, as seen in both Fig. 2 and Fig. 3, most students cover a majority of the theory videos, even though they do not necessarily watch it in order or during the same session. Fig. 4. The duration of each interaction with the theory video on week 4. Left: all interactions following a click, right: interactions exceeding 6 minutes. The total length of the video is 149 min. Guo et al. (2014) observed in a MOOC that the mode of viewing time on videos of any length was 6 minutes and proceeded to recommend that lecture videos be very short. However, our data shows that degree students mostly do watch the full lecture videos, the longest of which are over two hours. 3.2 Video use of different student groups To see whether students of different gender or major used the videos differently (RQ2), we split the view data by the background variables. There was no significant difference in the distribution of percentage watched either by student gender or major. As seen in Table 1, men had a higher average percent watched than women, but according to the Mann-Whitney U test, the difference was not statistically significant (W = 1.48, p = 0.13). Similarly, while prospective physics majors had a higher average percentage watched than students of other majors, the difference was not statistically significant according to the Kruskal-Wallis test (H (4, N = 56) = 6.50, p = 0.16). Table 1. Mean and standard deviation of the watched percentage of theory videos. Gender N Mean (%) S.D. (%) Male 29 75.2 39.8 Female 26 64.1 42.2 Major N Mean (%) S.D. (%) Physics 17 85.4 32.7 Chemistry 13 51.4 47.4 Mathematics 9 72.8 30.9 Undecided 7 46.1 49.7 Other 10 75.0 42.4 As we found no statistically significant differences in the data, examining the details of viewing durations or heat maps by student major or gender is unlikely to yield statistically significant results with this number of students. However, as there is a difference in means (Table 1), it would be important to collect more data to study this further. 3.3 Correlation with exam results Finally, we wanted to see whether watching videos correlates with learning outcomes on the course (RQ3). To see how video use correlates with physics knowledge at the end of the course, we calculated the Spearman correlation coefficients for the percentage watched of videos and exam points. A visualization of the correlation matrix is in Fig. 5. From the figure, it is clear that while both the percentage watched of example and theory videos correlate with exam points, the correlation between theory videos and exam points is stronger. For this pair, 𝜌 = 0.43 (𝑝 = 0.0012), which indicates a moderate correlation (Dancey & Reidy, 2020, p. 183). For example videos and exam points, the correlation coefficient 𝜌 = 0.27 (𝑝 = 0.048), which indicates a weak correlation. Fig. 5. Spearman correlation table of the average percentage of video watched by video type (All, Theory, Examples, Table) and exam points. 4 CONCLUSIONS AND IMPLICATIONS Our study shows that students use video materials in different ways. Most students watched the majority of the theory videos. Students generally clicked around theory videos and skipped or re-watched segments. Fewer students used the motivational “At the physicists’ table” videos, and these videos were watched in longer segments. We did not observe statistically significant differences between students of different major subjects or between men or women, but this may be due to the small sample size. As students clicked around the videos, the duration of each consecutive view was very short. However, we did not observe that students would only watch the beginning of the video, much less that the mode of minutes watched would have been 6 minutes, as Guo et al. (2014) observed in a MOOC. The likely reasons for this difference are the better motivation of degree students and the inclusion of search marks in the video, as also recommended by Seidel (2024). We saw a moderate correlation between the percentage of theory videos watched and exam points. The correlation between the percentage of example videos watched and exam points was weak, and watching the “Physicists table” videos did not correlate with the learning outcome. However, a limitation of the study is that we did not measure initial knowledge of the students. It is possible that watching theory videos is linked to better initial knowledge or other variables in the learning process, which we were not able to measure. As we only included data from willing participants, it is possible that the data is skewed. We were also not able to measure engagement with the videos beyond clicks and viewing duration. However, the viewing data provides rich and interesting data, which can be used to understand which parts of the video students watch. The heat maps can be used to study video lectures and understand which parts are important for students in their learning processes. Studying students’ interaction with the materials has value for the pragmatic development of teaching materials. Further studies are needed to better understand the role of video materials for different students’ learning processes. ACKNOWLEDGEMENTS Microsoft Copilot (GPT-4o) was used to generate a part of the R code for the figures, in particular the data processing and the plotting of the heatmaps (Figure 3). We thank prof. Jorma Keskinen for discussions regarding the study. REFERENCES Ahn, D., & Chan, J. C. (2025). What Drives Student Engagement and Learning in Video Lectures? An Investigation of Instructor Visibility, Playback Speed, and Student Preferences. Applied Cognitive Psychology, 39(2), e70026.