Abstract

The global pandemic of 2020 caused a paradigm shift in engineering education. In a matter of weeks, and sometimes days, faculty members across the world had to move their hands-on engineering courses to an online environment. During this shift, educators relied on technology more so than ever to improve student design learning without an empirically understanding of the impact of this shift on students' cognition and understanding. The current study was developed to determine the cognitive underpinnings of such shifts by exploring the impact of Augmented Reality (AR) and animation impact engineering student learning, cognitive load, and recall during a virtual product dissection educational activity. This was achieved through a full factorial experiment with 117 first-year engineering students where students were divided into one of four conditions: baseline of virtual dissection; virtual dissection + animation, AR dissection, and AR dissection + animation. The results of the study show that students in the virtual dissection + animation showed an increased understanding of the product over the three other conditions. In addition, participant cognitive load and recall in the AR condition were not significantly different than in a non-AR virtual environment. The results are used to provide recommendations on how technology can be utilized in a virtual classroom environment, providing crucial insight into the steps needed to virtualize engineering education during the pandemic as well as future steps toward possible education reform.

1 Introduction

“We can learn some things through this crisis about online delivery of not only instruction, but an array of opportunities for learning and support. In this way, we can make the most of the crisis to help redesign better systems of education and child development.”—Reville [1]

The global pandemic of 2020 caused a massive paradigm shift in engineering education resulting in uncertainty; education was forced to move online on an untested and unprecedented scale requiring much trial and error [2]. In a matter of weeks, and sometimes days, faculty members across the world were forced to move their hands-on engineering courses into a remote environment to stop the spread of COVID-19 [3]. As one engineering student describes it, “engineering-specific classes are much harder because there aren't as many resources available to help with engineering-specific skills” [4]. Students and educators need more resources as well as an empirical understanding of the impact of technology on students learning, particularly when it replaces hands-on in-person instruction. As shifts are made to accommodate learning during this pandemic, it is important that the advancements to virtual education are recorded and retained, to continue to aid future virtual engineering education efforts.

If there is one potential silver lining in the 2020 COVID-19 pandemic, it is its role as a catalyst for educational reform. Recently, technology has become an item of necessity in engineering education rather than a luxury. During the pandemic of 2020, nearly 300 universities in the United States were forced to close campuses and switch to remote/virtual education [5], relying heavily on technology to drive educational paths. In cases where the proper technology was not available, universities decided to temporarily close rather than offer online courses [6,7]. While lecture-based courses can be moved to virtual environments with little effort, hands-on learning, like engineering education, is more difficult to virtualize [7]. As such, testing new technology is critical to improve virtual engineering education. Exploring technology for virtual education is not only important to education during this global pandemic, but an important step in a much-needed educational reform [1].

Engineering education should be at the forefront of technology integration due to the exponential rate at which engineering technology grows [8]. Learning in engineering education has typically been measured in relation to real-world tasks, usually an evaluation of knowledge, skills, attitudes, and values [9]. Technology has been shown to improve these metrics in other areas of education. For example, the use of virtual laboratories is on the rise due to the ease and affordability of implementing the newest available technology [10]. In fact, a study utilizing virtual reality (VR) showed that virtual experimentation allows the practice to be repeated easily and provides increased freedom [11] and can aid students in understanding difficult concepts [1214]. However, adopting technology without an understanding of how it impacts learning is problematic. Selway [15] summarizes their problematic experiences in technological implementation by saying “technology is not meant to supplant, it's meant to enhance” (2019). One such area of engineering education that has experimented with integrating technology is product dissection.

“Product dissection is the systematic disassembly of products” and has long been a staple of the engineering classroom [16,17]. The first formal introduction of product dissection in engineering education was in 1991 at Stanford University [16]. It was initially implemented in order to develop a basic aptitude for engineering and engineering design and to develop mental visualization skills [18]. While there are many educational benefits to product dissection, it also has its downfalls; for example, high costs are linked with physical product dissection; as well as significant resource usage and facility requirements which can make implementation difficult [19]. Physical product dissection also has the disadvantage of requiring maintenance in order to maintain consistency [20]. Due to these disadvantages, and with the advancement of technology, virtual product dissection has been investigated as an alternative to physical product dissection. Virtual product dissection is the use of computer software to disassemble and gain an understanding of virtual objects [19]. Several studies have been conducted to investigate the differences between virtual and physical product dissection in engineering education [11,2124]. This research has shown that there were no significant differences in student learning between virtual and physical product dissection environments [21]. The next step in the advancement of engineering education is to enhance virtual education so that we truly leverage the benefits of the technology and extract its true value.

Several technologies exist that are potentially beneficial to engineering education, including both animation and augmented reality (AR). Animation, a collection of images shown in rapid succession to highlight the changes from image to image [25], can be used to show motion and increase understanding of a topic [26,27]. Animation has proven to be useful and has already become a part of education in several areas of science [2830]. One of the biggest beneficial aspects of animation is that it is a complementary technology [31], meaning it is a technology that has the potential to enhance other learning tools when used together. One such technology that can be paired with animation is AR. Augmented reality can be defined as the utilization of computer-generated objects to coexist with and enhance real-world objects [32]. Due to recent advancements in AR technology, along with increased accessibility, AR technology has the potential to be an effective educational tool by adding previously missing information to real-world scenarios [33]. Research has shown the effectiveness of AR may stem from the reduction in cognitive load users experience, allowing them to learn the material more efficiently [34]. While in theory, these technologies seem greatly beneficial, no studies have been conducted to test this in the classroom. Therefore, the purpose of this research was to evaluate the impact of AR and animation on engineering student learning, cognitive load, and recall during a product dissection educational activity in an effort to help educators identify how these technologies can be leveraged in engineering education.

2 Related Work

While AR and animation have both been studied in an educational context, there is little research on their impact on components of engineering education such as cognitive load and recall. The following sections summarize the existing research in these areas and also to highlight the use of product dissection in engineering education, the educational topic applied in this study.

2.1 Product Dissection in Engineering Education.

In order to understand the impact of AR and animation on learning in an engineering classroom, these technologies were applied to an activity that has been a staple of engineering classrooms since the early 1990's: product dissection. The benefits of product dissection have been harnessed in engineering education through both physical and virtual dissection over the years. Virtual dissection has made a name for itself due to the efficiency of dissecting virtually [21] through the reduced time and effort required [24]. One major boost to efficiency is the ability to quickly and easily reset the dissected product back to the original state [35], allowing multiple virtual product dissections to be performed in the same time it takes to perform a single physical product dissection [22]. While research has shown the benefits of integrating virtual dissection technology, it is not without its drawbacks. For instance, research has found that participants in both conditions still had room for better understanding. In fact, participants scored below a 75% on their post-dissection assessments, indicating they did not have a perfect understanding of how their product worked [36]. This indicates that there is still room for improving educational technology.

This potentially better technology could stem from old technology, new technology, or a combination. Previous instances of virtual product dissection have utilized three-dimensional (3D) portable document formats (pdfs) and computer-aided design (CAD) model viewing software to allow for manipulation of products in virtual space, but have missed the opportunity to include additional details such as animations of how the product functions. With the addition of animation, the functionality of parts and mechanisms can be seen within products, which is typically unseen due to the outer casing obscuring the inner workings in physical products [26]. This can be seen in Fig. 1, where the only moving parts that are visible in the physical condition are the slide (to load the spring) and the trigger to release the dart. Unlike the physical condition, in the virtual dissection condition with animations, we can see that pulling back the slide loads the large spring into the firing position, the mechanism that locks the spring in the firing position, and the spring that brings the slide back to the home position. Due to the core difference between how products work in physical environments as compared to virtual environments, there is an opportunity for AR and animation to provide an understanding of product functionality that is not possible in the physical world. Without a full understanding of the functionality, learners fill in the blanks by guessing about the purpose of different parts of a product. This could cause unnecessary mental stress (cognitive load) on the user. Although these are possible benefits of virtual environments, they have yet to be explored.

Fig. 1
Virtual nerf gun (top left), with spring engaged and casing removed (top middle), and casing removed (top right) compared to physical nerf gun (bottom left), with spring engaged (bottom middle), and with casing removed (bottom right) [37]
Fig. 1
Virtual nerf gun (top left), with spring engaged and casing removed (top middle), and casing removed (top right) compared to physical nerf gun (bottom left), with spring engaged (bottom middle), and with casing removed (bottom right) [37]
Close modal

2.2 Augmented Reality and Animation in Education.

Educational technology can come in many different forms, one of these forms could be animation. The definition of animation is a collection of images shown in rapid succession to highlight the changes from image to image [25]. Animation, in the simplest form, is used to show movement as a path or trajectory [27] and can be used to show the motion of physical objects, energy flow, or fluids [26]. Animation, as a form of multimedia, provides an effective understanding of a subject's content [38] when supported by up-to-date technology [39]. Research has shown that animation is beneficial as a complementary learning approach and helps motivate students and build interest in learning through interactivity [31]. Animation has been used in many fields of education including engineering [28], physics [30], and mathematics [29]. In engineering education, animation has been used to show the fluid motion of different viscosities of fluids interacting with different objects [40]. Animation has also been used to show robot kinematics without the high price associated with robotics education [41,42]. Additionally, animation has been used to teach thermodynamics and material science [43]. However, animation does not come without its drawbacks. A study conducted by Tversky et al. [27] showed that animation was only useful when it showed more information than a static image. This study also revealed that not having control over the animation (start, stop, pause, replay, speed up, and slow down) reduced the potential benefits [27].

Animation is a technology that can be described as a complementary technology [31]. Meaning it is a technology that has the potential to enhance other technologies. In this case, the technology we will be looking to enhance will be AR. AR is defined as the utilization of computer-generated objects to coexist with and enhance real-world objects [32]. AR was first introduced back in the 1960s, when mechanical and ultrasonic trackers were used to display simple wireframe objects through a head-mounted see-through display [44]. AR has advanced significantly since this time and can be broadly categorized into two main areas: marker-based and markerless-based [45]. Marker-based AR acquires a location through the use of a specific image, such as a barcode or quick response (QR) code, while markerless-based AR uses image recognition or location data to overlay information [45].

AR has many potential applications in education and training due to improvements in information technology as well as computers [46]. AR has the potential to increase motivation to learn as well as improve educational realism activities [47]. The ability to control objects and explore numerous perspectives through AR can attract and inspire new learners [48]. AR has particular usefulness in its ability to demonstrate information not easily experienced in the real world [48]. Johnson et al. [49] confirm, “AR has strong potential to provide both powerful contextual, on-site learning experiences and serendipitous exploration and discovery of the connected nature of information in the real world.” (p. 21) (2010). AR, however, is relatively new to education. An experiment performed by Shelton and Hedley [48] began its testing in formal education. The study found that AR was useful in teaching items that are more easily experienced in AR than in real life, in this case, the relationship between the sun and the earth [48]. It was found that the students who interacted with the virtual objects showed increased learning when compared to students who did not [48]. This experiment was taken further by Kerawalla et al. [50], which showed that animating the earth and the sun in AR resulted in even greater learning. AR has also been experimented with teaching molecular structure [51], storytelling for children [52], and learning in museums and exhibits [53]. In addition, an AR prototype was also developed to superimpose shape, functional groups, and component names in order to facilitate function diagram development and the understanding of the product's functionality during dissection [54]. While the prototype was not explored for its educational utility in this study, this work does highlight the potential utility of AR in a dissection environment. In addition, as AR is an up-and-coming technology it is likely to have issues. As the most common form of AR is through the use of a smartphone/tablet, AR is limited to the capabilities of the phone/tablet [45]. AR is also susceptible to technical difficulties [55]. Additionally, studies have found that AR can either lower or raise cognitive load [55,56].

AR has proven useful specifically in engineering education. An early example of AR in engineering education was an AR-enhanced textbook. This textbook was used for teaching civil and construction engineering, which contained simulated animations; it was reported that a majority of the students believed the AR tool was beneficial [57]. Another use of AR in engineering education was to improve orthographic and isometric drawing skills; the manipulation of a 3D object allowed students to more easily draw different views of an object when compared to 2D methods [5860]. A study performed by Gutierrez and Fernandez [61] showed that virtual object manipulation through AR led to increased student motivation and better academic performance. Overall, these studies showed that AR-enhanced engineering education methods were superior when compared to traditional teaching methods [5761].

2.3 Cognitive Load and Recall in Education.

Cognitive load is an important aspect to consider in education. The aspect of cognitive load that is important for this research is called working memory. Working memory is an area that cognitive load has an effect on, this area deals with the maximum level of elements an individual can process [62]. Extraneous cognitive load is useful when the goal is to strengthen working memory; however, extraneous cognitive load has been shown to interfere with learning as it requires the limits of working memory to be met [62]. For virtual learning, the ideal type of cognitive load is called Effective Cognitive Load. This type of cognitive load has been shown to enhance learning as the limits of working memory are not being met [62]. By decreasing cognitive load, working memory limitations are less likely to occur allowing for effective cognitive load and increased educational value. Since memory is limited, reducing cognitive load can make it easier for an individual to process information [63]. For complex tasks, these cognitive requirements are higher and require more resources [64]. Cognitive load can be broken down into different categories: perceptual/central, response, spatial, verbal, visual, auditory, manual, and speech.

When considering cognitive load with technology, an important aspect to consider is whether the participant is familiar with the technology. It has been found that students who are not familiar with the technology they are using will experience higher cognitive load and lower levels of learning [65]. This is most likely due to the split-attention effect, where the learner is required to split their attention between two different aspects [66]. The split-attention effect could be split attention between understanding the learning tool and learning the material or between different forms of learning materials [66]. Due to this, learners who have prior experience with a form of technology will likely have a lower cognitive load. With this in mind, the relatively new technology of augmented reality was looked at for this study.

Cognitive load has also been shown to have a relationship with recall. Recall can be defined as “the act of retrieving information or events from the past” [67]. While the cognitive load is used to determine how much of working memory is being used, understanding recall can be helpful in learning whether or not the working memory is being used efficiently [68]. Working memory refers to the process of both processing and storing short-term memory for tasks [69]. Recall refers to the ability to categorize short-term memories [70]. A study conducted by Xie et al. [71] concluded that a reduction in cognitive load led to increased recall ability. Research has shown long-term memory is obtained through short-term memory [72]. Due to this, technology that can increase recall, therefore increasing long-term memory, is vital in the efforts to enhance virtual education through technology.

3 Research Objectives

The purpose of the current study was to evaluate the impact of Augmented Reality (AR) and animation on engineering student learning and cognitive load. Our research questions are listed as follows:

RQ1: How does AR and animation in product dissection impact student cognitive load? This research question was developed to identify the impact of the use of AR and animation in a product dissection environment on student cognitive load. We hypothesize that AR and the use of animation would reduce students’ overall cognitive load. The motivation for this hypothesis stems from prior research showing that increasing the amount of information shared reduced the number of working memory elements that needed to be processed mentally [62]. As working memory is limited, reducing cognitive load can make information easier to process [63].

RQ2: How does AR and animation in product dissection impact knowledge gain? This research question was developed to identify the impact of the use of AR and animation in a product dissection environment as assessed by examining only new items learned through the Student Learning Assessment (SLA). We hypothesize that AR and animation will show increased knowledge gain in a product dissection environment. The motivation behind this hypothesis stems from a trend noticed in previous research between the pre- and during-SLAs indicating fatigue. Items that some participants had correct in the pre-SLA would not be restated the during-SLA.

RQ3: How do AR and animation in product dissection impact recall? This research question was developed to identify the impact of the use of AR and animation in a product dissection environment on student recall, as assessed through an SLA. We hypothesize that AR and animation will show improvements in recall in a product dissection environment. The motivation for this hypothesis comes from research showing that reduction in cognitive load has been associated with improvements in recall [71].

4 Methodology

To answer these research questions, a study was conducted with first-year engineering students. The remainder of this section highlights the methodological approach for this study.

4.1 Participants.

The participants of this study were first-year engineering students recruited through an introductory engineering design class. The class promotes hands-on learning as well as design-ability through design projects.

4.2 Procedure.

Before any procedures took place, an overview was provided and a signed consent form was obtained from each participant. Next, participants were given a pre-SLA, developed by Toh et al. [73] to gauge conceptual understanding of the product, before interacting with it. Participants were given 15 mins to complete the pre-SLA. After this assessment, the participants were assigned to one of the groups. This designation was depended on the layout of the room. The room consisted of eight, four-person tables. As each table only had two computers, two virtual conditions, and two AR conditions were held at each table. It was ensured that each table had either animation or no-animation conditions with no tables having both. The tables were then randomly assigned the conditions in the 2 × 2 factorial design (see experimental design) maintaining equal distribution. After the participants have been placed, they then began a 15-min product dissection activity. While they completed the activity, the participants were tasked with completing the during-SLA in order to gauge the effectiveness of the learning tool. Once the dissection activity was completed, participants were then asked to complete the cognitive load, novelty effect, and free-response surveys. Finally, the participants were given a post-SLA 48 h after the dissection activity in order to assess recall. The participants had 15 min to complete this post-SLA.

4.3 Experimental Design.

This study is designed to compare four different dissection conditions in a 2 × 2 factorial design: Virtual product dissection, virtual product dissection with animation, AR product dissection, and AR product dissection with animation. All participants in this study were assigned the same product, a Nerf Gun Sharp Shot.2 The animation was shown through two forms of media but was copied to ensure the same information was being expressed in each. Screenshots of these animations are shown in Fig. 2.

Fig. 2
(a) Screenshot from the animation in the virtual dissection and (b) screenshot from the animation in the AR condition
Fig. 2
(a) Screenshot from the animation in the virtual dissection and (b) screenshot from the animation in the AR condition
Close modal

Virtual: Participants assigned to this condition of the experiment (N = 28 participants) were tasked to dissect a product virtually. Participants were provided with the virtual files of the product as well as access to a computer with Solidworks eDrawings 2019. The participants were not provided with any instructions on how to use the program. The goal of the product dissection activity was to break down the product into its individual components and develop a deeper understanding of the products’ functions. The participants were not provided with instructions for dissecting the product and had the freedom to manipulate the object as they saw fit.

Virtual with Animation: Participants of this condition (N = 27 participants) were tasked similarly to that of the virtual product dissection group, as well as being provided the same materials. However, participants in this condition were also provided with an additional animation of the product. The animation was a 15-s video showing the internal functions of the product. The animation was provided to the participants through a YouTube link.

AR: Participants in this condition (N = 30 participants) were tasked with virtually dissecting the same product as in the previous conditions. However, participants in this condition were to use a different dissection tool. The dissection tool for this condition was the SimLab AR/VR Viewer version 3.1 application which was downloaded onto IOS smartphone devices. Simlab is a free, mobile solution that allows for viewing and sharing 3D models in 3D, virtual reality, and augmented reality modes (Fig. 2(b)) [74]. Participants were provided instructions on how to download the virtual product file but were not provided any instructions on how to use the application. Participants were also provided a file of the virtual product generated in SimLab Composer 9. Participants were not provided with instructions for dissecting the product. The application was set up so that when students tapped on a component of the product during the dissection, the part would be “hidden.” The students also had the ability to rotate the model, and they could zoom in by moving their phones closer to the artifact.

AR with Animation: Participants in this condition (N = 32 participants) were given the same task as the previous conditions. Participants were provided with the same material as the AR product dissection condition with the exception of the virtual product file. They were given a file of the virtual product that included an animation. The included animation had a duration of 15 s and showed the internal functions of the product. In addition to the features included in the AR condition, the application also allowed students to play, stop, and reset the AR environment.

4.4 Metrics.

In order to assess the differences in learning and cognitive load in the 2 × 2 factorial study design, the following metrics were used.

Student Learning Assessment (SLA): This metric was developed and validated in prior research by Toh et al. [73] to measure participant's understanding of the components in the assigned product. The SLA used in this study can be accessed.3 Each participant was to complete a two-sided worksheet composed of four categories. These categories consisted of power supply, mechanism that provides primary motion, energy flow of the device, and form and outer body. Under each of these categories, the participants had to provide visual representations as well as functional descriptions. Specifically, feature knowledge was computed by having three raters review the SLA and rated whether each of the 43 items was identified by the participant. The inter-rater reliability of the three raters was computed using Cohen's Kappa (κ = 0.80, κ = 0.74, and κ = 0.84). Examples of low and high scoring SLAs can be seen in Fig. 3.

Fig. 3
(a) An example of a section of an SLA that received a low score (6/43) and (b) an example of a section of an SLA that received a high score (33/43)
Fig. 3
(a) An example of a section of an SLA that received a low score (6/43) and (b) an example of a section of an SLA that received a high score (33/43)
Close modal

SLA Knowledge: The SLA knowledge metric was developed to indicate how many items on the SLA were learned during the activity. In order to consider this, the SLA knowledge score includes only the items that were incorrect on the pre-SLA and correct on the during-SLA. An example of this is shown in Fig. 4.

Fig. 4
(top) Visual representation of how SLA Knowledge Score was calculated; (bottom) visual representation of how Total Knowledge was calculated
Fig. 4
(top) Visual representation of how SLA Knowledge Score was calculated; (bottom) visual representation of how Total Knowledge was calculated
Close modal

Total Knowledge: Total Knowledge is a combination of the pre- and during-SLAs, where a point could be earned if an item was present on either or both of the SLAs. An explanation for this can be seen in Fig. 4.

Cognitive Load: Cognitive load theory suggests that working memory is limited and that we must reduce unneeded information, thus reducing cognition allowing more efficient processing of information [63]. In order to measure cognitive load, the Workload Profile Assessment developed by Tsang and Velazquez [75] was utilized. This measure breaks workload into eight dimensions with four different parts of processing:

Stages of Processing (Types of Attentional Resources) includes perceptual/central processing, or the attentional resources for detecting, recognizing, and identifying objects and remembering, problem-solving, and decision making, and response processing, or the attentional resources required to select and execute a needed response such as selecting the correct tool to put together Ikea furniture.

Processing Codes (How We Understand Information) includes spatial processing, or how we understand and remember the spatial relationship (e.g., distance) between objects, and verbal processing, how we understand and remember verbal or linguistic materials.

Input Modality (How We Take in Information) includes visual processing, or how we process and interpret the meaning of visual information gained through our sight, and auditory processing, how we process and interpret the meaning of auditory (sound) information gained through our hearing.

Output Modalities (How We Respond to Information) includes manual responses such as responding manually to a task such as physically maneuvering an object, and speech responses, including responding verbally (through speech) to an object such as answering the question “what is your phone number?”

It is important to note that spatial and verbal information can be processed through visual and/or auditory processing. For example, someone can either listen to the radio and process verbal information auditorily or read a book and process verbal information visually. The Workload Profile Assessment has been compared to the NASA task load index and subjective workload assessment technique, this metric is the least intrusive and all three methods were valid [76]. This metric was a self-assessment by the participants obtained through a survey after the dissection activity. Each participant rated the proportion of attention resources they utilized to complete the task on a scale from 0 to 100 for each dimension. These values were used individually to assess specific areas of cognitive process and collectively to assess overall cognitive resources required by the participant. In this study, there were no occurrences of auditory or speech activities so these forms of cognitive load were excluded from the results.

Dissection Technology Review: In order to gauge students’ experience with the technology for product dissection, a survey was provided at the end of the study. This survey asked students to rate their level of agreement on six items on a Likert scale from not at all agree (1) to strongly agree (7). For purposes of the current study, only the first item “the use of the dissection tool was a new experience for me” was assessed.

5 Data Analysis and Results

During this study, there were 117 products virtually dissected. Prior to the analysis, we compared the “newness” of the technology for each participant. Specifically, student responses to the Dissection Technology Review survey item “the use of the dissection tool was a new experience for me” were assessed using a Kruskal Wallace test because “Likert scales fall within the ordinal level of measurement” (pg. 2, [77] and thus should be assessed using nonparametric tests. As such, a Kruskal–Wallis test was conducted to determine if there were differences in the four conditions (VR, VR + Animation, AR, and AR + Animation) across this survey item. Distributions of this item were similar for all groups, as assessed by visual inspection of a boxplot. Median responses to this item were not statistically different across the four groups χ2(3) = 5.694, p = 0.127. This result indicates that participants did not differ in their perception of how new the technology was to them across the four conditions; the Median and mode survey response was 7 for all conditions indicating it was new to all. The remainder of this section focuses on the results according to our research questions. The results are mean ± standard error unless otherwise denoted. Effect sizes were classified according to Cohen [78]. The results are analyzed using SPSS v. 26 with a significance rating of 0.05. Data are mean ± standard deviations unless otherwise stated.

RQ1: How does the use of AR and animation during product dissection impact student cognitive load?

Our first research question was developed to identify the impact of the use of AR and animation in a virtual product dissection environment on student cognitive load. Cognitive load scores were used to assess whether the technologies in the 2 × 2 factorial reduced working memory, therefore showing potential for increased learning. Our hypothesis was that AR and the use of animation would reduce students’ overall cognitive load because prior work has shown that in today's world, people are more comfortable with technology [79]. It has been found more comfortable learning methods lead to reduced cognitive load [80].

Prior to the analysis, assumptions were checked. This analysis identified that there were no multivariate outliers in the data, as assessed by Mahalanobis distance (p > 0.001). There was homogeneity of variances, as assessed by Levene's test for equality of variances, p < 0.05. Outliers outside of three standard deviations were removed. However, the data violated the assumption of multicollinearity of the data (|r| < 0.9), and the data failed to meet the assumption of normally, as assessed by Shapiro–Wilk's test (p > 0.05). However, due to the fact that the two-way analysis of variance (ANOVA) is robust against violations of normality [81], and the fact that all of the data were within the range of normal univariate distributions for skewness (between −2 and +2) and kurtosis (between −7 and +7) [82,83], seven two-way ANOVA'S were computed in lieu of a two-way multivariate analysis of variance (MANOVA). The results of the seven two-way ANOVAs that had the six types of cognitive load (perceptual/central, response, spatial, verbal, visual, and manual), as well as the total cognitive load between the different product dissection conditions as the dependent variables and the method of dissection as the independent variable, failed to show statistically significant interactions between the technologies utilized for the cognitive load (Table 1). These results refute our hypothesis by showing that AR and animation would significantly reduce cognitive load over the other conditions.

Table 1

Main effects of dissection tool and animation on participants’ cognitive load values

Dissection ToolAnimation
Means (Standard error)Means (SE)
VariableFPartial η2PAR dissectionVirtual dissectionFPartial η2PAnimationNo-animation
Perceptual/Central0.000.000.9872.94 (2.35)73.02 (2.20)6.170.010.4374.24 (2.29)71.71 (2.26)
Response0.060.000.8162.06 (3.12)60.97 (3.18)0.910.010.3459.40 (3.07)63.63 (3.21)
Spatial0.010.000.9577.32 (2.96)77.03 (2.85)1.180.010.2879.40 (2.93)74.95 (2.87)
Verbal0.230.000.6322.59 (3.29)24.79 (3.13)0.050.000.8223.17 (3.19)24.20 (3.24)
Visual1.180.010.2892.19 (1.67)89.71 (1.56)0.900.010.3592.03 (1.59)89.87 (1.63)
Manual0.950.010.3374.63 (3.17)78.80 (2.86)0.670.010.6278.45 (2.98)74.97 (3.06)
Total0.250.000.62379.22 (9,73)385.97 (9.28)0.340.000.56386.53 (9.47)378.66 (9.55)
Dissection ToolAnimation
Means (Standard error)Means (SE)
VariableFPartial η2PAR dissectionVirtual dissectionFPartial η2PAnimationNo-animation
Perceptual/Central0.000.000.9872.94 (2.35)73.02 (2.20)6.170.010.4374.24 (2.29)71.71 (2.26)
Response0.060.000.8162.06 (3.12)60.97 (3.18)0.910.010.3459.40 (3.07)63.63 (3.21)
Spatial0.010.000.9577.32 (2.96)77.03 (2.85)1.180.010.2879.40 (2.93)74.95 (2.87)
Verbal0.230.000.6322.59 (3.29)24.79 (3.13)0.050.000.8223.17 (3.19)24.20 (3.24)
Visual1.180.010.2892.19 (1.67)89.71 (1.56)0.900.010.3592.03 (1.59)89.87 (1.63)
Manual0.950.010.3374.63 (3.17)78.80 (2.86)0.670.010.6278.45 (2.98)74.97 (3.06)
Total0.250.000.62379.22 (9,73)385.97 (9.28)0.340.000.56386.53 (9.47)378.66 (9.55)

RQ2: How does the use of AR and animation during product dissection impact student knowledge gain?

Our second research question was developed to identify the impact of the use of AR and animation in a product dissection environment as assessed on knowledge gained during the Student Learning Assessment (SLA). Our hypothesis was that AR and animation will show increased knowledge gain in a product dissection environment. We hypothesized this because prior research has shown that the students who interacted with the virtual objects showed increased learning when compared to students who did not [48]. Our prior research also indicated animation, as a complementary technology, can strengthen AR by increasing motivation and interest [28].

In this analysis, knowledge gain was defined by counting only the items that the participant got wrong in the pre-SLA and correct in the during-SLA. This is referred to as the SLA knowledge score. Prior to the analysis, assumptions were checked in preparation for running a two-way ANOVA comparing the SLA knowledge score to the conditions of our factorial design. Outliers were present in the data set but no changes were made to the data due to the robustness of the ANOVA [81]. Data were normally distributed, as assessed by Shapiro–Wilk's test (p > 0.05). There was homogeneity of variances, as assessed by Levene's test for equality of variances, p = 0.076.

The results of the two-way ANOVA revealed that there was no statistically significant interaction between dissection conditions for SLA knowledge score, F(1, 113) = 2.385, p = 0.125, and partial η2 = 0.021. Therefore, an analysis of the main effects was performed. The analysis failed to reveal a statistically significant difference in SLA knowledge score between the AR and virtual conditions, F(1, 113) = 0.662, p = 0.418, and partial η2 = 0.006. However, there was a statistically significant main effect between animation and no-animation, F(1, 113) = 5.280, p = 0.023, and partial η2 = 0.045. All pairwise comparisons were run where reported 95% confidence intervals and p-values are Bonferroni-adjusted. The unweighted marginal means of SLA knowledge scores for the animation and no-animation were 9.62 ± 0.526, and 7.91 ± 0.529, respectively. The animation was associated with a mean SLA knowledge score of 1.71, 95% CI [0.24, 3.19] higher than no-animation, a statistically significant difference, p = 0.023. The results can be seen in Fig. 5. These results partially support our hypothesis by indicating that animation was shown to have a statistically significant improvement in knowledge gain over the other conditions. However, AR showed no statistically significant improvement in knowledge gain over the other conditions.

Fig. 5
Means and standard error of SLA knowledge. The main effect of animation was significantly different (p < 0.05) with animations having a positive effect on SLA knowledge
Fig. 5
Means and standard error of SLA knowledge. The main effect of animation was significantly different (p < 0.05) with animations having a positive effect on SLA knowledge
Close modal

RQ3: How does the use of AR and animation during product dissection impact recall?

Our final research question was developed to identify if AR and animation impacted a student's ability to recall previously learned information gained from a product dissection environment. Our hypothesis was that AR and animation will show improvements in recall in a product dissection environment. Prior research has shown that reductions in cognitive load can result in an improvement to recall [71].

In order to assess recall, we separated our SLA scores by day. Day one score was computed as the Total SLA score (Fig. 4) whereas day two SLA score was post-SLA score. Prior to our analysis, assumptions were checked. There was homogeneity of variances for both day one (p = 0.743) and day two scores (p = 0.911), as assessed by Levene's test for equality of variances. The assumption of sphericity was met as there are only two levels within subjects’ factors. There was not a statistically significant three-way interaction between time, virtuality, and animation, F(1, 102) = 0.991, p = 0.342, and partial η2 = 0.009. There were insignificant two-way interactions (p > 0.05). As such, a three-way ANOVA was conducted to examine the day one to day two differences between the conditions. All pairwise comparisons were performed for simple main effects. Bonferroni corrections were made with comparisons within each simple main effect considered a family of comparisons. Adjusted p-values are reported. The results showed that the mean day two score was higher in the virtual dissection condition (17.40 ± 5.03) than those in the AR condition (17.36 ± 4.63), a mean difference of 0.18 (95% CI, −1.69 to 2.05). The mean day two score was higher in the animation condition (18.10 ± 4.71) than those in the no-animation condition (16.69 ± 4.81), a mean difference of 1.45 (95% CI, −0.42 to 3.32). The results are shown in Fig. 6. These results refute our hypothesis by showing that there were no statistical advantages to AR and animation when compared to the conditions of the 2 × 2 factorial.

Fig. 6
Means and standard error of recall for the four conditions
Fig. 6
Means and standard error of recall for the four conditions
Close modal

6 Discussion

The goal of this research was to identify the impact of AR and animation on engineering student learning and cognitive load. Our main findings are as follows:

  • There was no statistically significant difference between cognitive load in the four conditions tested (virtual, virtual + animation, AR, and AR + animation).

  • The use of animation was associated with higher knowledge gains regardless of the technology deployed (virtual or AR).

  • The recall was higher when participants performed the activity virtually (without AR) and with animation.

Our hypothesis for the first research question was that AR and animation would both reduce students' overall cognitive load. However, the results did not show significant improvements over the other conditions, refuting our hypothesis. The results may be attributed to students' lack of familiarity AR technology, and the AR interface itself (including its capabilities, workflow, and interactions provided) which may account for increases in cognitive load. On the other hand, a lack of cognitive load savings with an animation may be attributed to the split-attention effect [66] where people have to split their attention between a 3D model of the product and an animation showing the function of the product. While these two learning materials likely increased learning, it is likely this also caused the increase in cognitive load as well.

As digital natives, students are very comfortable with computers and their phones. The average person accumulates roughly five and a half hours of screen time between their phone and computer [79]. While both the virtual computer-based dissection and the AR dissection both required learning a new software/app, the computer-based software may be more intuitive than the AR app. For reference, the first version of eDrawings was released in 2003 and is very similar to other programs the students use daily, while the SimLab AR/VR viewer was initially released in 2017. Prior work has found that a higher cognitive load can be associated with unfamiliar technologies, which may be the culprit in this situation. In fact, previous research states that participants that are not familiar with the technology might have higher cognitive load and lower levels of learning [65]. Students may learn more through AR if they spend more time doing it, so future research should look at a more longitudinal type analysis. AR has many different forms; therefore, there are many different opportunities to be beneficial in the classroom.

The only condition in the 2 × 2 factorial to show a significant increase in knowledge over the other conditions was virtual with animation

Our hypothesis for the second research question was that educational technology, in the form of AR and animation, will show improvements in learning in a product dissection environment. The results showed that animation paired with virtual dissection had the only statistically significant increase when looking at knowledge gain. These results do not match our hypothesis, but these findings are consistent with previously run tests.

Previously ran tests utilized a three-way mixed ANOVA comparing the pre-SLA scores with the during-SLA scores to the conditions of the study. It was noticed that numerous participants would omit information on the during-SLA that the participants included on the pre-SLA. This was likely a symptom showing fatigue in the participants. In order to increase the accuracy of the study, SLA knowledge score was utilized. While two different methods were used to assess the effectiveness of the conditions, the consistency in the results provides increased assurance in the accuracy of the results of this study. The results that have been observed in both of these methods show that the virtual dissection paired with animation proved to be the best of the four conditions.

What we see is that animation may be beneficial as a complementary learning approach. This is in line with prior work that showed support for the animation to motivate students and build interest in learning through interactivity [31]. While only virtual with animation showed a significant increase over non-information, there was a numeric increase in the AR with animation condition. One reason we may not have seen as big of an increase in the AR with animation conditions may be due to the different versions of animation. AR had an animation included within the 3D model, while virtual dissection contained an animation in the form of a video clip. Both forms of animation showed the same information about the product. However, it is likely the participants found the video clip more intuitive, allowing the participant to more efficiently navigate the information.

We hypothesized that the highest SLA knowledge score would be seen in the AR condition. Our results showed that there was not a significant increase in knowledge gain in this condition. Previous research into AR has shown that it has the potential to be an effective learning tool [4850]. However, previous research in virtual learning, specifically in virtual product dissection has shown it to be an effective learning tool as well [21,23,24,29,84]. While our results show that the AR condition was not significantly different from virtual dissection, it did not disprove AR as a valid means of education. It was found to be a comparable tool to virtual learning in the context of product dissection. Along with being a comparable tool, AR has the added benefit of increasing motivation in the classroom [47]. Future work should be conducted to further explore AR as a tool for virtual learning in other aspects of education. Additionally, as many other forms of AR exist, further research should include investigating different forms of AR.

None of the conditions in the 2 × 2 factorial showed significantly improved recall when compared to the other conditions.

Our hypothesis for the last research question was that AR and animation would show improvements in recall in a product dissection environment. Our results did not show significant differences in recall between the conditions. These results do not match our hypothesis. While the results did not show a significant difference, they did show numeric differences. When comparing numeric differences, AR showed a slight advantage over virtual and animation showed an advantage over no-animation.

While virtual had a higher mean than AR for day one scores, AR had a higher day two score. This led AR to have the lowest difference in mean SLA score between day one and day two. While not significant this leads us to believe that further research in AR could yield significant results proving its value as an educational tool. Numerous studies have shown a correlation between the interest and curiosity of a student and its positive benefits on memory [8588]. It is possible AR's better recall score is a result of the technologies ability to increase interest and curiosity. This gives us further reason to believe that further testing of AR could result in increased recall ability.

When looking at animation, it had the highest day one mean as well as the highest day two mean. While not significant this provides evidence that animation is not only a valuable complementary learning tool, but a valuable learning tool itself. However, when looking at the difference between day one and day two scores, animation had a very slight advantage over the no-animation condition. While animation has not been shown statistically better than no-animation in the context of recall, we can say it is comparable to the current no-animation teaching methods. As animation is comparable in the context of recall, the highest day one and day two scores it has a net positive value on education.

7 Conclusions, Limitations, and Future Work

Due to the educational difficulties associated with the COVID-19 pandemic, research in virtual education has increased in importance. The purpose of this study was utilized to research new technology for virtual education. Specifically, the purpose of this study was to identify the impact of AR and animation on engineering student learning, cognitive load, and recall. This was accomplished through an experiment with 117 first-year engineering students. Overall, the results showed that AR was not statistically different from virtual product dissection. The results also showed that the virtual dissection condition and animations were able to add statistically significant value to education over the over conditions. Importantly, these results show that while AR is not a statistically different tool from virtual education, it is still a valuable learning tool with the potential to gain value with increased understanding. The results also show us that animation has value as a complementary learning tool with the potential to be a valuable learning tool by itself.

While the study has insights into the educational value of AR and animation, there are limitations to be considered. The study was conducted in the context of product dissection, so while product dissection education can benefit from the results, future work should focus on AR and animation outside of this context. In addition, AR and animation can take many forms but in this study they were limited. One of the biggest limitations of this study was the restriction to one product. Adding a second, or multiple products, could show the potential each product type has to influence the effect of AR and animation as learning tools. In addition, the results of this study may have been impacted by participants' previous background with the technology, their socio-economic setting, or other underlying individual factors. The effects of these differences may have a larger impact due to the relatively small sample size of each group or the between-subject nature of this work. As such, future work should look to replicate these findings with within-subject research design or by uncovering the impact of these, and other, individual differences. Finally, the study was conducted in a classroom setting of a first-year engineering design course. While this allowed for a more realistic data capturing experience, it also limited our ability to explore other more qualitative observational data such as video gathering and a video protocol analysis. In light of the findings here, future work should explore more qualitative investigations into why these differences do or do not occur in these different settings. Of particular interest are items such as usage patterns and interactive workflows.

Future work should explore the different forms of AR and animation in order to gain a longitudinal understanding of which methods are most effective for learning. Along with exploring different forms of AR and animation, different types of products should be tested as well. The product chosen in this study was selected due to its level of complexity and the amount of motion. Selecting products with different levels of complexity as well as different levels of motion could show which types of products AR and animations are better suited for. Finally, to improve these technologies for educational purposes, it would be beneficial to make the technology more accessible and user-friendly. More access to AR for students would allow them to become more comfortable with the technology, potentially increasing its value as an educational tool. The technology should be more accessible not only for the student but also for the teacher. Allowing the teacher control over the content enhanced by this technology would give this technology a greater chance to be beneficial.

Footnotes

Conflict of Interest

There are no conflicts of interest.

Data Availability Statement

The authors attest that all data for this study are included in the paper.

References

1.
Mineo
,
L.
,
2020
, “
Time to Fix American Education with Race-for-Space Resolve
,”
The Harvard Gazette
.
2.
Burgess
,
S.
, and
Sievertsen
,
H. H.
,
2020
, “
Schools, Skills, and Learning: The Impact of Covid19 on Education
,”
VOXEU CEPR
.
3.
Yilmaz
,
N.
,
2020
, “
Engineering Education in the Time of Covid
,”
National Society of Professional Engineers
.
4.
Alvim
,
L.
,
2020
, “
Constricted by Covid19: College of Engineering
,”
The Oracle
.
5.
Foresman
,
B.
,
2020
, “
Here Are The U.S. Universities That Have Closed Due to Coronavirus
,” edscoop.
6.
Lederman
,
D.
,
2020
, “
Preparing for a Fall Without In-Person Classes
,” Inside Higher Ed.
7.
Eisenberg
,
J.
, and
Escobar
,
A.
,
2020
, “
COVID-19: 10 Steps for Transferring Your Course Online
,”
World Economic Forum
.
8.
Brophy
,
S.
,
Klein
,
S.
,
Portsmore
,
M.
, and
Rogers
,
C.
,
2008
, “
Advancing Engineering Education in P-12 Classrooms
,”
J. Eng. Educ.
,
97
(
3
), pp.
369
387
.
9.
Hoskins
,
B.
, and
Fredriksson
,
U.
,
2009
,
Learning to Learn: What Is It and Can It Be Measured, Luxembourg: Office for Official Publications of the European Communities
.
10.
Potkonjak
,
V.
,
Gardner
,
M.
,
Callaghan
,
V.
,
Mattila
,
P.
,
Guetl
,
C.
,
Petrovic
,
V. M.
, and
Jovanovic
,
K.
,
2016
, “
Virtual Laboratories for Education in Science, Technology, and Engineering: A Review
,”
Comput. Educ.
,
95
, pp.
309
327
.
11.
Chittaro
,
L.
, and
Ranon
,
R.
,
2007
, “
Web3D Technologies in Learning, Education, and Training: Motivations, Issues, Opportunities
,”
Comput. Educ.
,
49
(
1
), pp.
3
18
.
12.
Lourdes
,
M.
, and
Cartas
,
M.
,
2012
, “
Using an Improved Virtual Learning Environment for Engineering Students
,”
Eur. J. Eng. Educ.
,
37
(
3
), pp.
229
241
.
13.
Messner
,
J.
,
Yerrapathruni
,
S.
,
Baratta
,
A.
, and
Whisker
,
V.
,
2003
, “
Using Virtual Reality to Improve Construction Engineering Education
,”
American Society for Engineering Education Annual Conference
,
Nashville, TN
,
June 22–23
.
14.
Trindade
,
J.
,
Fiolhais
,
C.
, and
Almeida
,
L.
,
2002
, “
Science Learning in Virtual Environments: A Descriptive Study
,”
Br. J. Educ. Technol.
,
33
(
4
), pp.
471
488
.
15.
Selway
,
M.
,
2019
, “
Tech Flop: When NOT to Use Technology in the Classroom
,”
KQED Education
.
16.
Sheppard
,
S. D.
,
1992
, “
Mechanical Dissection: An Experience in How Things Work
,”
Proceedings of the Engineering Education: Curriculum Innovation & Integration
,
Santa Barbara, CA
,
Jan. 5–10
.
17.
Lamancusa
,
J.
,
Jorgensen
,
J. E.
, and
Fridley
,
J. L.
,
1996
, “
Product Dissection – A Tool For Benchmarking in the Process of Teaching Design
,”
Frontiers in Education Conference
,
Salt Lake City, UT
,
Nov. 6–9
.
18.
Lamancusa
,
J.
,
Torres
,
M.
,
Kumar
,
V.
, and
Jorgensen
,
J.
,
1996
, “
Learning Engineering by Product Dissection
,”
ASEE Annual Conference
,
Washington. DC
.
19.
Devendorf
,
M.
,
Lewis
,
K.
,
Simpson
,
T. W.
,
Stone
,
R. B.
, and
Regli
,
W. C.
,
2009
, “
Evaluating the Use of Digital Product Repositories to Enhance Product Dissection Activities in the Classroom
,”
ASME J. Comput. Inf. Sci. Eng.
,
9
(
4
), p.
041008
.
20.
McKenna
,
A. F.
,
Chen
,
W.
, and
Simpson
,
T. W.
,
2008
, “
Exploring the Impact of Virtual and Physical Dissection Activities on Students’ Understanding of Engineering Design Principles
,”
ASME Conference Brooklyn
,
NY
.
21.
Starkey
,
E. M.
,
McKay
,
A. S.
,
Hunter
,
S. T.
, and
Miller
,
S. R.
,
2016
, “
Let’s Get Physical? The Impact of Dissection Modality on Engineering Student Design Learning
,”
Proceedings of the ASME Design Engineering Technical Conference
,
Charlotte, NC
.
22.
Olympiou
,
G.
,
Zacharia
,
Z. C.
,
Papaevripidou
,
M.
, and
Constantinou
,
C. P.
,
2008
, “
Effects of Physical and Virtual Experimentation on Students’ Conceptual Understanding in Heat and Temperature
,”
Res. Sci. Teach.
,
45
(
9
), pp.
1021
1035
.
23.
Starkey
,
E. M.
,
McKay
,
A. S.
,
Hunter
,
S. T.
, and
Miller
,
S. R.
,
2017
, “
Dissecting Creativity: How Dissection Virtuality, Analogical Distance and Product Complexity Impact Creativity and Self-efficacy
,”
Des. Comput. Cogn.
,
16
, pp.
59
77
.
24.
Toh
,
C. A.
, and
Miller
,
S. R.
,
2013
, “
Exploring the Utility of Product Dissection for Early-Phase Idea Generation
,”
ASME DETC Conference
,
Portland, OR
.
25.
Bétrancourt
,
M.
, and
Tversky
,
B.
,
2000
, “
Effect of Computer Animation on Users’ Performance: A Review
,”
Le Trav. Hum. A Biling. Multi-Disciplinary J. Hum. Factors
,
63
(
4
), pp.
311
329
.
26.
Ainsworth
,
S.
, and
VanLabeke
,
N.
,
2004
, “
Multiple Forms of Dynamic Representation
,”
Learn. Instr.
,
14
(
3
), pp.
241
255
.
27.
Tversky
,
B.
,
Morrison
,
J. B.
, and
Betrancourt
,
M.
,
2002
, “
Animation: Can It Facilitate
,”
Int. J. Hum-Comput. Stud.
,
57
(
4
), pp.
247
262
.
28.
Wozny
,
M. J.
,
1978
, “
Interactive Computer Graphics for Engineering Education
,”
Prof. Eng.
,
48
(
6
), pp.
14
18
.
29.
Hooper
,
P.
, and
Morton
,
J.
,
1982
, “
Fluctuations in the Dollar: A Model of Nominal and Real Exchange Rate Determination
,”
J. Int. Money Finance
,
1
(
1
), pp.
39
56
.
30.
DiSessa
,
A. A.
,
1982
, “
Unlearning Aristotelian Physics: A Study of Knowledge-Based Learning
,”
Cogn. Sci.
,
6
(
1
), pp.
37
75
.
31.
Xiao
,
L.
,
2013
, “
Animation Trends in Education
,”
Int. J. Inf. Educ. Technol.
,
3
(
3
), pp.
286
289
.
32.
Azuma
,
R.
,
Baillot
,
Y.
,
Behringer
,
R.
,
Feiner
,
S.
,
Julier
,
S.
, and
MacIntyre
,
B.
,
2001
, “
Recent Advances in Augmented Reality
,”
IEEE Comput. Graph. Appl.
,
21
(
6
), pp.
34
47
.
33.
el Sayed
,
N. A. M.
,
Zayed
,
H. H.
, and
Sharawy
,
M. I.
,
2011
, “
ARSC: Augmented Reality Student Card: An Augmented Reality Solution for the Education Field
,”
Comput. Educ.
,
56
(
4
), pp.
1045
1061
.
34.
Chen
,
C. J.
,
2006
, “
The Design, Development and Evaluation of a Virtual Reality Based Learning Environment
,”
Australas. J. Educ. Technol.
,
22
(
1
), pp.
39
63
.
35.
2019
, “
EDrawings Functionality Matrix
”.
36.
Starkey
,
E. M.
,
McKay
,
A. S.
,
Hunter
,
S. T.
, and
Miller
,
S. R.
,
2018
, “
Piecing Together Product Dissection: How Dissection Conditions Impact Student Conceptual Understanding and Cognitive Load
,”
ASME J. Mech. Des.
,
140
(
5
), p.
052001
.
37.
Kearney
,
K. G.
,
Starkey
,
E. M.
, and
Miller
,
S. R.
,
2019
, “
Leveraging Animations to Enhance Learning Through Virtual Product Dissection
,”
Proceedings of the ASME 2019 International Design Engineering Technical Conference
,
Anaheim, CA
,
Aug. 18–21
, pp.
1
6
.
38.
Badi
,
H. J.
,
Zeki
,
A. M.
, and
Othman
,
R.
,
2013
, “
Animation as a Problem Solving Technique in Mechanical Engineering Education
,”
Int. J. Sci. Eng. Res.
,
4
(
5
), pp.
96
99
.
39.
Badi
,
Hana
,
Zeki
,
Akram
,
Faris
,
Waleed
, and
Othman
,
Roslina
,
2013
, “
Animation: A Problem Solving Technique in Mechanical Engineering Education
,”
Int. J. Sci. Eng. Res.
,
4
(
5
), pp.
96
99
.
40.
Foster
,
N.
, and
Fedkiw
,
R.
,
2001
, “
Practical Animation of Liquids
,”
Proceedings of the 28th Annual Conference on Computer Graphics and Interactive Techniques
,
Los Angeles, CA
,
Aug. 12–17
, pp.
23
30
.
41.
Fok
,
S. C.
,
Tan
,
F. L.
, and
NG
,
P.
,
1994
, “
An Application of Computer Graphics for Teaching Robot Kinematics
,”
Proceedings of the 9th Annual International Conference on Frontiers of Technology
,
Singapore
,
Aug. 22–26
.
42.
Hejase
,
A. J.
, and
Hasbini
,
B. M.
,
1993
, “
Computer-Aided Instructions in Robotics
,”
Int. J. Eng. Educ.
,
9
, pp.
156
161
.
43.
Tan
,
F. L.
, and
Fok
,
S. C.
,
1995
, “
Development of Engineering Courseware for University Undergraduate Teaching Using Computer Animation
,”
Comput. Appl. Eng. Educ.
,
3
(
2
), pp.
121
126
.
44.
Sutherland
,
I. E.
,
1968
, “
A Head-Mounted Three-Dimensional Display
,”
Proceedings of Fall Joint Computer Conference
,
San Francisco, CA
,
Dec. 9–11
, pp.
757
764
.
45.
Pence
,
H. E.
,
2011
, “
Smartphones, Smart Objects, and Augmented Reality
,”
Ref. Libr.
,
52
(
1–2
), pp.
136
145
.
46.
Lee
,
K.
,
2012
, “
Augmented Reality in Education and Training
,”
TechTrends
,
56
(
2
), pp.
13
21
.
47.
Chang
,
G.
,
Morreale
,
P.
, and
Medicherla
,
P.
,
2010
, “
Applications of Augmented Reality Systems in Education
,”
Society for Information Technology & Teacher Education International Conference
,
San Diego, CA
,
Mar. 29
, pp.
1380
1385
.
48.
Shelton
,
B. E.
, and
Hedley
,
N. R.
,
2002
, “
Using Augmented Reality for Teaching Earth-Sun Relationships to Undergraduate Geography Students
,”
The First IEEE International Workshop Augmented Reality Toolkit
, p.
8
.
49.
Johnson
,
L.
,
Levine
,
A.
,
Smith
,
R.
, and
Stone
,
S.
,
2010
, “
Simple Augmented Reality
,” The 2010 Horizon Report, pp.
21
24
.
50.
Kerawalla
,
L.
,
Luckin
,
R.
,
Seljeflot
,
S.
, and
Woolard
,
A.
,
2006
, “
Making It Real’: Exploring the Potential of Augmented Reality for Teaching Primary School Science
,”
Virtual Real.
,
10
(
3–4
), pp.
163
174
.
51.
Feiner
,
S.
,
Macintyre
,
B.
, and
Selligmann
,
D.
,
1993
, “
Knowledge-Based Augmented Reality
,”
Commun. ACM
,
36
(
7
), pp.
53
62
.
52.
Billinghurst
,
M.
,
2003
, “
Augmented Reality in Education
,”
New Horizons for Learning
.
53.
de Freitas
,
S.
, and
Levene
,
M.
,
2005
, “
Wearable and Mobile Devices for Informal Learning
,”
Encyclopedia of Human Computer Interaction
.
54.
McPherson
,
C.
, and
Radkowski
,
R.
,
2016
, “
Tangible Augmented Reality and Engineering Product Dissection
,” Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 9740, pp.
633
642
.
55.
Akçayır
,
M.
, and
Akçayır
,
G.
,
2017
, “
Advantages and Challenges Associated With AR for Education: A Systematic Review of the Literature
,”
Educ. Res. Rev.
,
20
, pp.
1
11
.
56.
Wu
,
H.
,
Lee
,
S. W.
,
Chang
,
H.
, and
Liang
,
J.
,
2013
, “
Current Status,Opportunities, and Challenges of AR in Education
,”
Comput. Educ.
,
62
, pp.
41
49
.
57.
Shirazi
,
A.
, and
Behzadan
,
A. H.
,
2013
, “
Technology-Enhanced Learning in Construction Education Using Mobile Context-Aware Augmented Reality Visual Simulation
,”
Proceedings of the 2013 Winter Simulation Conference
,
Washington, DC
,
Dec. 8–11
, pp.
3074
3085
.
58.
Camba
,
J.
,
Contero
,
M.
, and
Salvador-Herranz
,
G.
,
2014
, “
Desktop vs. Mobile: A Comparative Study of Augmented Reality Systems for Engineering Visualizations in Education
,”
Proceedings of the IEEE Frontiers in Education Conference
,
Madrid, Spain
.
59.
Figueiredo
,
M.
,
Cardoso
,
P.
,
Gonclaves
,
C.
, and
Rodrigues
,
J.
,
2014
, “
Augmented Reality and Holograms for Tech Visualization of Mechanical Engineering Parts
,”
Proceedings of the 18th International Conference on Information Visualisation
,
Paris, France
.
60.
Gonzalez
,
C.
,
Dominguez
,
M.
,
Martin-Gutierrez
,
J.
,
Contero
,
M.
, and
Alcaniz
,
M.
,
2012
, “
Training With Augmented Reality on Engineering Degrees
,”
Proceedings of the ASME 2012 11th Biennial Conference on Engineering Systems Design and Analysis
,
Nantes, France
,
July 2–4
, pp.
1
8
.
61.
Guiterrez
,
J.
, and
Fernandez
,
M.
,
2014
, “
Applying Augmented Reality in Engineering Education to Improve Academic Performance Student Motivation
,”
Int. J. Eng. Educ.
,
30
(
3
), pp.
625
635
.
62.
Paas
,
F.
,
Renkl
,
A.
, and
Sweller
,
J.
,
2003
, “
Cognitive Load Theory and Instructional Design: Recent Developments
,”
Educ. Psychol.
,
38
(
1
), pp.
1
4
.
63.
Sweller
,
J.
,
1988
, “
Cognitive Load During Problem Solving: Effects on Learning
,”
Cogn. Sci.
,
12
(
2
), pp.
257
285
.
64.
Feinberg
,
S.
, and
Murphy
,
M.
,
2000
, “
Applying Cognitive Load Theory to the Design of Web-Based Instruction
,”
Proceedings of IEEE Professional Communication Society International Professional Communication Conference and Proceedings of the 18th Annual ACM International Conference on Computer Documentation
, pp.
353
360
.
65.
Moreno
,
R.
,
2002
, “
Who Learns Best With Multiple Representations? Cognitive Theory Predictions on Individual Differences in Multimedia Learning
,” EdMedia + Innovate Learning, Denver, CO.
66.
Chandler
,
P.
, and
Sweller
,
J.
,
1991
, “
Cognitive Load Theory and the Format of Instruction
,”
Cogn. Instr.
,
8
(
4
), pp.
293
332
.
67.
Britannica
,
2007
, “Recall,” https://www.britannica.com/topic/recall-memory
68.
Paas
,
F.
, and
Ayres
,
P.
,
2014
, “
Cognitive Load Theory: A Broader View on the Role of Memory in Learning and Education
,”
Educ. Psychol. Rev.
,
26
(
2
), pp.
191
105
.
69.
Baddeley
,
A. D.
,
2000
, “
The Episodie Buffer: A New Component of Working Memory?
,”
Trends Cogn. Sci.
,
11
(
4
), pp.
417
423
.
70.
Clair-Thompson
,
H. L.
,
2009
, “
Backwards Digit Recall: A Measure of Short-Term Memory or Working Memory?
,”
Eur. J. Cogn. Psychol.
,
22
(
2
), pp.
286
296
.
71.
Xie
,
H.
,
Wang
,
F.
,
Hao
,
Y.
,
Chen
,
J.
,
An
,
J.
,
Wang
,
Y.
, and
Liu
,
H.
,
2017
, “
The More Total Cognitive Load Is Reduced by Cues, the Better Retention and Transfer of Multimedia Learning: A Meta-Analysis and Two Meta-Regression Analyses
,”
PLoS One
,
12
(
8
), p.
e0183884
.
72.
Spence
,
K. W.
, and
Spence
,
J. T.
,
1968
,
Human Memory: A Proposed System and Its Control Processes
,
Academic Press
,
Cambridge, MA
, pp.
89
195
.
73.
Toh
,
C.
,
Miller
,
S.
, and
Simpson
,
T.
,
2015
, “
The Impact of Virtual Product Dissection Environments on Student Learning and Self-Efficacy
,”
J. Eng. Des.
,
26
(
1–3
), pp.
48
73
.
74.
75.
Tsang
,
P. S.
, and
Velazquez
,
V. L.
,
1996
, “
Diagnosticity and Multidimensional Subjective Workload Ratings
,”
Ergonomics
,
39
(
3
), pp.
358
381
.
76.
Rubio
,
S.
,
Díaz
,
E.
,
Martín
,
J.
, and
Puente
,
J. M.
,
2004
, “
Evaluation of Subjective Mental Workload: A Comparison of SWAT, NASA-TLX, and Workload Profile Methods
,”
Appl. Psychol.
,
53
(
1
), pp.
61
86
.
77.
Jamieson
,
S.
,
2004
, “
Likert Scales: How to (Ab) Use Them
,”
Med. Educ.
,
38
(
12
), pp.
1217
1218
.
78.
Cohen
,
J.
,
1988
,
Statistical Power Analysis for the Behavioral Sciences
,
Lawrence Erlbaum Associates
,
Mahwah, NJ
.
79.
MacKay
,
J.
,
2019
, “
Screen Time Stats 2019: Here’s How Much You Use Your Phone During the Workday
,” RescueTime:Blog.
80.
Chandler
,
P.
, and
Sweller
,
J.
,
1996
, “
Cognitive Load While Learning to Use a Computer
,”
Appl. Cogn. Psychol.
,
10
(
2
), pp.
151
170
.
81.
Norman
,
G.
,
2010
, “
Likert Scales, Levels of Measurement and the ‘Laws’ of Statistics
,”
Adv. Health Sci. Educ.
,
15
(
5
), pp.
625
632
.
82.
Byrne
,
B. M.
,
2013
,
Structural Equation Modeling with Mplus: Basic Concepts, Applications, and Programming
,
Rutledge
,
Abingdon, Oxfordshire
.
83.
Hair
,
J.
,
Black
,
W.
,
Babin
,
B.
, and
Anderson
,
R.
,
2010
,
Multivariate Data Analysis
, 7th ed.,
Pearson Education International
,
Upper Saddle River, NJ
.
84.
Hegarty
,
M.
,
2004
, “
Dynamic Visualizations and Learning: Getting to the Difficult Questions
,”
Learn. Instr.
,
14
(
3
), pp.
343
351
.
85.
Germain
,
C. M.
, and
Hess
,
T. M.
,
2007
, “
Motivational Influences on Controlled Processing: Moderating Distractibility in Older Adults
,”
Aging Neuropsychol. Cogn.
,
14
(
5
), pp.
462
486
.
86.
Hidi
,
S.
,
1990
, “
Interest and Its Contribution as a Mental Resource for Learning
,”
Rev. Educ. Res.
,
60
(
4
), pp.
549
571
.
87.
Isaacowitz
,
D. M.
,
Wadlinger
,
H. A.
,
Goren
,
D.
, and
Wilson
,
H. R.
,
2006
, “
Selective Preference in Visual Fixation Away From Negative Images in Old Age? An Eye-Tracking Study
,”
Psychol. Aging
,
21
(
1
), pp.
40
48
.
88.
Renninger
,
A.
, and
Hidi
,
S.
,
2011
, “
Revisiting the Conceptualization, Measurement, and Generation of Interest
,”
Educ. Psychol.
,
46
(
3
), pp.
168
184
.