Students perception of videos in introductory physics courses of engineering in face-to-face and online environments

Digital videos have an important (and increasing) presence in learning processes, especially within online universities and schools. However, creating videos is a time-consuming activity for teachers, who are usually not expert in video creation. Therefore, it is important to know which kind of video is perceived as more satisfactory and useful by students, among the videos that docents usually create. In this paper we show a structural model with the relation between satisfaction, the way in which a video has been created, the kind of video (with or without the hands of the teacher and with or without the body/head of the teacher), perceived usefulness, contents of the video (theory or problems) and the potential impact of videos on passing rates. The experiment has been performed in an introductory Physics of Engineering course with over 200 first year students in both: at 100% online university, Universitat Oberta de Catalunya (UOC); and at a face-to-face university, Escola Universitària Salesiana de Sarrià (EUSS). Tests have been performed with around 100 videos of two types: videos created with a digitizing tablet and screen capture, and videos created by recording the hands of the teacher. Results have been quantitatively analysed. The research shows that results are independent of the environment and that students prefer videos with hands. On the other hand, little effect has been found regarding the content of the video in the perceived usefulness or satisfaction. The performance results show that videos can improve the chances of passing the subject. Thus, the paper shows that videos with hands are a useful complement to challenging subjects, like introductory physics in Engineering, to effectively assimilate scientific knowledge. The main contributions of this paper are: to analyse the perception that students have of video in a specific context, introductory course of Physics in Engineering, in different environments; and to analyse the perception of the video regarding the way in which it has been created, and the kind of content.


Introduction
video can facilitate the use, entry and access to information [2]. In addition, the use of videos in schools has been very positive in some aspects, such as getting the desired effect on students [3,48,53] and facilitating the educator's job [13].
From the student's point of view, it can be a crucial teaching tool of science subjects in general, and physics in particular, for several reasons: • It helps acquire the abstraction capacity [19,58].
• It facilitates the acquisition of scientific language [21] .
A key element is the non-verbal communication associated with the lesson [52]. Information entry to memory can be divided into three channels [60]: verbal, visual (non-verbal), and the input-output information of the working memory. The dual coding theory [44] points out that reading and writing involve activities in two independent cognitive subsystems of codification: verbal and non-verbal coding.
Among non-verbal information, several studies have shown the importance of iconic gestures [29,57] that also have semantic information [16,35,61] and are semantically processed [43]. Hand gestures in particular, can alter the interpretation of discourse, eliminate ambiguities, increase understanding and memory, and transmit information not explicitly integrated into the discourse [10,17,25]. In the case of STEM, Stull [52] showed the importance of gesture in the case of chemistry. Some studies have pointed out the importance of showing the teacher's head and his/her gestures in distance learning [27]. However, there are extra elements that also play an important role in the non-verbal communication, like physical appearance [59] and body language [5], but they can be decontextualized from the information to transmit [54]. On the contrary, Videos with hands can contribute to reduce the cognitive load by providing non-verbal communication [9,55].
Recently, Nagy [41] evaluated an application of the Technology Acceptance Model (TAM) to the usage of video by students, where it is shown that perceived usefulness, attitude, and internet self-efficacy has a direct effect on the video usage.
From the point of view of physics courses, problem solving is used as instructional model [14,24,26,56], which is common in many faculties [11].

Hypotheses
Taking into account the aforementioned state of the art, we propose the following hypotheses: H0: within a physics subject, satisfaction with videos is independent of the environment.
This hypothesis can allow us to detect if the results are environment-independent (face-to-face, or virtual).
H1: Satisfaction of students increases when non-verbal information appears in the video. H2: Perceived usefulness has a positive effect on learning satisfaction.
H2 corresponds to Nagy's proposed TAM to online video usage and learning satisfaction.
H3: The presence of videos of problems increases the perceived usefulness of videos. H4: Perceived usefulness has a positive effect on video usage.
This hypothesis corresponds one of the Nagy's.
H5: Students prefer watching problem solving videos to theory videos. H6: Students see videos as complementary material instead of the main resource.
The first proposed hypothesis (H0) analyzes the impact of the environment to student perception. The second hypothesis (H1) deals with the satisfaction of the students with the videos. The third hypothesis (H3) focuses on the effect of the content of the videos. Hypotheses H2 and H4 assess the impact of perceived usefulness and will be tested by checking whether Nagy's results are hold in the proposed scenario. Finally, hypotheses H5 and H6 are related to the preferences of students and the relevance of the videos in the user learning experience, in a physics course.

Method
This section describes the methodology followed in the research that comprised the following stages: the creation of the videos, the definition of the population under study and the collection and analysis of data.

The creation of the videos
Two main kinds of video were created: screen capture (see Fig. 1), created with a Wacom tablet; and videos with hands (see Fig. 2) created with a camera that filmed the hands of the Fig. 1 Example of video created with a digitalizing table teacher while he was explaining. Both kinds of video correspond to the same topics and they follow the same structure and notation used within the text material provided to students. In some videos, the head of the teacher was included in a small box inside the video to allow students to see his face. Table 1 shows the number of videos created per topic and kind: 94 videos with hands (nearly 10 h); and 46 screen capture videos with tablet (over 7 h).
Videos are mp4 with codec H.264. Aspect ratio is 16:9 with 1280 × 720 pixels at 25 fps. Sound is in AAC LC, stereo with maximum bit rate of 128 kb/s and sampling rate of 48.0 kHz.

The population under examination
The experiment was performed at two different universities: EUSS, which is a face-to-face university; and UOC, which is a 100% online university. The courses chosen were: Physics of the degree of Industrial Engineering at EUSS ("EUSS"); Physics I of the degree of Telecommunication at UOC ("Tl") and Fundamentals of Physics of the degree of Computer Sciences at UOC ("Inf"). Table 2 shows the topics covered in every course. Within each topic, only the videos that correspond to the corpus of the course were given to the students.
The teacher who recorded the videos is the same for all videos (but for 7 videos, that were created by another teacher and we think are too few to affect the results) and is the teacher of  all courses, which is beneficial since students' engagement increases when the videos have been created by their own teacher [20]. Videos are given to students at the beginning of every lesson as a complementary material. Students have also access to all of the usual material: text material at UOC and summary and faceto-face classes at EUSS. UOC has a proprietary virtual campus, where students have all the materials corresponding to the subject, and the videos in a tool named Present@ [46,47] and communication tools. At EUSS, students have Moodle. Videos are available in Vimeo®, protected by a password to be accessible only to the students of the course.
Students received, every week, a guideline message from the teacher with links to the relevant videos for the week topic and the expected order to visualize them. Watching the videos was not mandatory in any case.

Data collection
Data collection for the goal of the present paper was performed in two ways: through a questionnaire and through a semi-structured interview. Semi-structured interviews will be analyzed in a future work.
The questionnaire was compiled by using Google Forms and sent during the semesters between September 2016 and February 2018. That corresponds to one semester of EUSS, two semesters of Tl and three semesters of Inf. The number of students to whom the questionnaire has been sent is 62 of EUSS, 129 of Tl and 423 of Inf; and the number of answered questionnaires is 15 (24%) in EUSS, 25 (19,4%) in Tl and 85 (20%) in Inf (see Table 3).

Questionnaire development, content validity
The questionnaire was adapted from the one proposed by Nagy. Table 4 shows the Nagy's questionnaire (column 2) compared with the one proposed in this paper (column 3). Since most of the students were in an online environment, with a very well-defined pedagogical model, 1 the questions needed to be adapted to fit the environment as well as the online model characteristics, while taking into account also the context of the other face-to-face students. On the other hand, it is important to take into account that in this paper we focus in more specific elements than Nagy's regarding the use of videos, since we would like also to see the effect of the kind of video and its content. The table also shows the name of the variable for every question and for every course. To identify whose course a variable belongs, we add "_Tl" for Telecommunication at UOC, "_Inf" for Computer Sciences (Informatics) at UOC and 1 https://www.uoc.edu/portal/en/universitat/model-educatiu/index.html "_EUSS" for Industrial Engineering at EUSS. In those variables where we need to distinguish between topics, the following identifiers appear: "Mec" is for "Mechanics", "Elc" is for "Electrostatics", "Mgn" is for "Magnetism" and "Cir" is for "Circuits". Finally, the capital letter "T" is for Theory videos and "P" is for Problems videos. Thus, for example, "SNV_MgnP_Tl" corresponds to the variable SNV regarding to Magnetism ("Mgn") in Problems videos ("P") in the course of Telecommunications at UOC ("Tl"). At the end of the questionnaire students had the opportunity to add any comment.

Data analysis
The construct reliability of the questionnaire was examined by using a Cronbach's alpha with 0.7 limit.
To validate hypothesis H0, Satisfaction with videos in a physics subject is independent of the environment, we analyzed differences regarding the environment. In particular, we compared the answers to every single question given by Tl course students and given by Inf and EUSS students. Quantitative questions were compared using t-test with α = 0.05, whose results are in Table 5. Columns t and p correspond to these values of t-test. Qualitative questions were compared using a χ 2 analysis. Its results can be seen in Table 6, where columns χ 2 and p show the corresponding values of χ 2 analysis. The usual limit to reject the null hypothesis are p values lower than 0.05, therefore, since values obtained are above this limit in nearly all cases, null hypothesis cannot be rejected [33] and this confirms that satisfaction is independent of the environment (H0 is validated). Therefore, results of one course can be taken as representative of all courses and results can be generalizable to any environment. In this paper we will take course Tl as representative.
To verify that there is no effect in the results, regarding the topic, due to different perception of the number of videos for every topic (mechanics, electrostatics, magnetism or circuits), two analyses were performed: an ANOVA and a principal component analysis. The results of the ANOVA (see Table 7) show (with an F = 0.546 and an α = 0.05) that we cannot reject the null hypothesis. Therefore, the satisfaction with the number of videos is the same, regardless the content or the topic of the video. Table 8 shows the results of the principal component analysis. As can be seen, one single factor explains 83.7% of the variance, which is compatible with the ANOVA conclusion that the satisfaction with the number of videos is the same for all the videos, regardless the topic. Therefore, students do not have different perceptions regarding the number of videos in different topics.
The previous steps helped us to verify that neither the number of videos nor the environment affect the results, regarding the kind of the video, the content or the topic. These results allow us to generalize our results to any topic.
The next steps of the methodology are to verify hypotheses H1 to H4: Hypothesis H1, Perceived usefulness has a significant positive effect on learning satisfaction, looks for a relation      Table 4; hypothesis H2, Perceived usefulness has a significant positive effect on video usage, looks for a relation between variables PU and U of Table 4; hypothesis H3, satisfaction of students increases when non-verbal information appears in the video, looks for a relation between variables SAT and TV of Table 4; and hypothesis H4, the presence of videos of problems increases the perceived usefulness of videos, which is an element specific of physics courses, looks for a relation between variables ToP and PU of Table 4. These four hypotheses are verified through the structural model proposed in Fig. 3. Figure 3a) shows Nagy's structured model [41] corresponding to the bubbles SAT, PU and U, and Fig. 3b) shows the model proposed in this paper. The model is built with the results of Tl course, variables SAT_Tl, TV_Tl, PU_Tl, ToP_Tl and U_Tl, because thanks to the validation of hypothesis H0 we can take one single course as representative. A multivariate regression and three linear regressions have been performed as shown in Table 9. The program R, with the extension R commander has been used to perform the calculations. Finally, results of students in the subject have been analyzed to see the true impact of videos in their performance.
The main limitations of the analysis are: 1) it was not possible to analyze the same number of semesters in all the analyzed courses because of uncontrolled external changes; 2) it was not possible to analyze whether different kinds of video had different effects since students had available all kind of videos and, therefore, we can only know which kinds of video (with hands, with tablets, etc.) were best perceived by students; and 3) we analyzed perception from students, but not the data about the videos they consumed and how they were consumed (what videos have been seen, when, for how long, etc.), which will be analyzed in a future work.

Results and discussion
The alpha of Cronbach for the questionnaire is 0.7 (see Table 4). From the t-test and Chisquare analysis we can see that the answer from students in the three subjects satisfy the null hypothesis that all the answers correspond to the same population and, therefore, results obtained can be considered independent of the environment or the degree. This satisfies the hypothesis H0: Satisfaction with videos in a physics subject is independent of the environment.
The ANOVA analysis confirms that the satisfaction with the number of videos is independent of the topic or the environment. The mean of the satisfaction with the number of videos is The satisfaction with videos (variable SAT) is of 4.16 over 5, with a standard deviation of 1.1, which means that students are satisfied or very satisfied with the videos. Regarding the perceived usefulness (variable PU), the qualification is 4.4 over 5 with a standard deviation of 1, which means that students perceive the videos as very useful. The mean of the frequency of visualization of videos is 2.28 with a standard deviation of 0.54. That means that students watched videos mainly weekly (75%).
When the results regarding the content of the video are analyzed (variable ToP), the mean of the variable that shows the content of video preferred by students is 2.5 with a standard deviation of 0.58, what means that 64% like videos of theory and problems equally, i.e. students prefer having both kinds of videos.
Finally, when asked about the way in which videos have been created (variable TV), the mean is 2.24 with a standard deviation of 0.66, what means that 52% of students prefer videos with hands, and 88% of online students prefer videos in which the teacher appear and, if they can choose, they prefer videos with hands.
If we analyze the satisfaction, in the structured model presented in Fig. 3, the making of the video (variable TV) explains only the 13% of the satisfaction of students. However, we cannot reject the null hypothesis (the kind of video does not affect the satisfaction of student) since as is shown in Table 9, we obtain p = 0.142 s. Therefore, we cannot prove hypothesis H3, satisfaction of students (SAT) increases when non-verbal information appears in the video (TV). However, the same multivariate regression shows that our results are compatible with H1, Perceived usefulness (PU) has a significant positive effect on learning satisfaction (SAT), which agrees with Nagy's paper.
Regarding the hypothesis H2, Perceived usefulness (PU) has a significant positive effect on video usage (U), with α = 0.05, we cannot reject the null hypothesis that there is no relation between PU and U. This is a different result than that obtained by Nagy's. However, in this research videos were not the only resource available and students have always different ways to access the same content: the teacher or text material with the same explanations than videos. Therefore, students do not need to go back to the video to review the concepts.
Finally, regarding the hypothesis H4, the presence of videos of problems (ToP) increases the perceived usefulness (PU) of videos, we see from Table 9 that there is a linear relation between both parameters. However, with a typical value of α = 0.05 we cannot reject the null hypothesis that the preference for one kind of video or other affects the perceived usefulness.
The performance results of two subjects of Physics where videos have been introduced are shown in Table 10 (for Inf) and Table 11 (for Tl). In bold are marked the results for the semesters in which the video, and only the video, was introduced. Physics at EUSS is not analysed here because in the period when videos were introduced, there were many other changes in the subject that affected the results.
Both tables show the percentage of students that fail or do not take the exam; and the rate that pass vs. those who take the course, or vs. the total number of students. The percentage of students that pass the subject vs. the total that takes the exam, changes when video is introduced. In the case of Inf, the semesters before introducing the video the mean of pass Fig. 3 Structural model a) Nagy's, b) of the hypotheses presented: SAT is Satisfaction, and corresponds to variable SAT_Tl; TV is the making of the video, and corresponds to variable TV_Tl: without any part of the teacher, with the hands of the teacher, with the body of head of the teacher; PU is perceived usefulness and corresponds to the variable PU_Tl; ToP is "theory or problems" and corresponds to the variable ToP_Tl; and U is for Usage and corresponds to the variable U_Tl Table 9 Linear regression between: SAT_Tl and PU_Tl, TV_Tl; RV_Tl and TV_Tl; PU_Tl and ToP_Tl; and U_Tl and PU_Tl. The first column represent the independent variable and the second the dependent variable; the column "Estimate" gives the values for the intercept and the coefficient for the dependent variable; "Std Error" gives the standard error of the estimate value; "t value" and "Pr(>t)" gives the t value and the probability that the result could be obtained randomly instead of because correlation In the case of Tl, the videos were introduced later. However, the mean of students that pass the subject over those who take the exam is 79% before introducing the videos, and raises until 92% after adding the videos. Nevertheless, data from previous semesters fluctuate around the mean value and in semester 20121 we find the same percentage. Maybe the reason for this fluctuation is that in the first semesters of Tl there was a lower number of students. As the number of students increased, the percentages are closer to Inf.
From these data we could conclude that for those students who follow the subject, videos can help them improve their chances to pass the exam, and therefore, the course.

Conclusions
In this paper we analyze the perceived satisfaction and usage of videos considering the environment (online or face-to-face), the kind of video (tablet or recorded with hands) and the content of the video (topic, theory or problems). The experiment has been performed in three equivalent Physics courses of three different degrees (Industrial Engineering, Telecommunication and Computer Sciences) of two different universities, one face-to-face (Escola Universitària Salesiana de Sarrià, EUSS) and one 100% online (Universitat Oberta de  Catalunya, UOC). All the courses had the same teacher, who was the one who created the videos. To avoid a possible effect of the number of videos with the satisfaction, it has been checked that students were equally satisfied with the number of videos, regardless the topic or if they were of theory or problems. The first conclusion found is that answers of students are independent of the course taken and the environment used (face-to-face or fully virtual). Students are very satisfied with videos and perceive them as very useful. Although they manifest to prefer problems videos, this preference neither affects the perceived usefulness, nor the number of times they watch every video.
We did not find a clear relation between the perceived usefulness and the number of videos, contrary to previous research [41]. This may be because in this experiment videos are not the main material and, therefore, students have also access to textual materials or face-to-face classes with the same content than in the videos.
Finally, students prefer videos in which they can see the teacher (hands, face or half body). This preference is a little more important in virtual students than in face-to-face students. However, there is not a conclusive effect with the satisfaction, what means that it is much more important for students having the video than the kind of video. On the other hand, results show that including videos in a physics subject increases the probability of passing the course.
Therefore, as a conclusion, students perceive the video in Physics as a very useful element and are very satisfied with it, although they perceive it as complementary material to textual material. This perception is confirmed by the results of the courses that improve when video is introduced.
As a future work, we plan to analyze the behavior of students when watching the video to study the effects of topic and the kind of video, from the true usage point of video, instead of the perception of students. We will deep also in how students perceive the fact that the head and the body of the teacher appear in the video.

Compliance with ethical standards
Ethical statement, conflict of interest and informed consent The authors certify that they have NO affiliations with or involvement in any organization or entity with any financial interest (such as honoraria; educational grants; participation in speakers' bureaus; membership, employment, consultancies, stock ownership, or other equity interest; and expert testimony or patent-licensing arrangements), or non-financial inter-est. (such as personal or professional relationships, affiliations, knowledge or beliefs) in the subject matter or materials discussed in this manuscript. All study participants, or their legal guardian, were informed and give consent prior to study enrollment; and all data were anonymized.
Contribution of this paper to the literature • This study aims to investigate the Physics students' perceptions on different kind of educational videos used in both virtual and face-to-face university contexts in Barcelona (Spain).
• The study shows that perception of videos by students is independent of the subject and of the environment (face-to-face or online).
• A relation is found between satisfaction, the way in which a video has been created, the perceived usefulness of videos, the contents of the video (theory or problems) and its usage.
• The effect of videos in students' performance is also shown.