Emily's report.

A preliminary investigation into the use of Interactive Lecture Demonstrations in an online environment.

Abstract

1. Introduction This paper describes an investigation into the effectiveness of different formats of a multimedia online tutorial. Developments in the accessibility of video and computer technology have prompted the use of digital tutorials and other multimedia educational tools by many universities and high schools (Adhvaryu & Balbin, 1998), although little research into the effectiveness of the online tutorial format has been done. Online tutorials utilising videos of demonstrations represent a useful tool for distance-education students (Tsang, Hung & Ng, 1999), as well as providing teachers with additional resources which can be completed outside of class hours. The goal of this study was to create and test an online tutorial focused on the physics area of kinematics. The structure of the tutorial was based strongly on the work of Thornton, whose research into Interactive Lecture Demonstrations (ILDs) has shown ILDs to be extremely effective in a classroom environment (Sokoloff & Thornton, 1997). The translation of the ILD format into an online environment raises possibilities for effective Internet-based learning tools, and this study investigates its effectiveness in an online environment.

Thornton’s ILDs are based around the Predict-Observe-Explain model, in which a situation is described and demonstrated for students, who then predict the outcome of the demonstration, and finally have the correct result explained to them. This process is usually followed when investigating conceptual change (Tao & Gunstone, 1999), an important part of physics education, however in this investigation, the tutorial serves to directly teach students about kinematics, rather than focus on conceptual change. The structure of the Predict-Observe-Explain model is varied between three different conditions in the tutorial, each of which represent a different level of necessary prediction, from subjects making no predictions, to those being asked to make a prediction, and those having to make a prediction as part of the tutorial. It is expected that the students who do not make predictions throughout the tutorial will not learn as well as those who did make predictions.

Multimedia demonstrations involving some student prediction have been proven to be effective (Crouch, Fagen, Callan & Mazur, 2004) and popular amongst students in a classroom environment (Di Stefano, 1996), however when the demonstration is contained in a tutorial which is presented outside of class hours, as is done in this investigation, issues such as students’ motivation and interest become factors which may prove to be important in students’ learning. Research suggests that students’ interest and motivation must be maintained throughout a teaching intervention in order to optimise students’ learning (Wigfield & Eccles, 2000, Paris & Turner, 1994), and when students have no external authorities, such as when they are completing an online tutorial, the tutorial must rely on a design which optimises students’ situational interest in order to be an effective learning tool. Hence this study aims to investigate whether the ILD format which has been adopted in this tutorial incorporates a design which maintains students’ level of interest.

The use of web-based multimedia tutorials and teaching resources is thought to increase in the future (Goldberg & Salari, 1997), becoming an important tool for teaching in a modern environment. However, the fundamental issue remains of whether effectual methods of classroom-based learning, such as the Predict-Observe-Explain ILD model, successfully translate into a web-based environment, and to what extent students' predictions in web-based learning programs affect their results. The aims of the present investigation therefore are:

a. to compare learning involving mandatory prediction with voluntary prediction, or no prediction.

b. to assess the utility of the ILD Predict-Observe-Explain learning format in a web-based learning environment.

The first issue, that of the effect of students' levels of active participation was addressed through the split-stream tutorial model, where each stream had a different prediction condition. With the exception of the tutorial for the students who made no predictions, the structure of the online tutorials followed the Predict-Observe-Explain ILD format, and were closely based on the kinematics ILDs produced by Thornton (Ref) thus addressing the second concern of how successfully the Predict-Observe-Explain format performs in a web-based environment.

2. The domain

The domain of kinematics in physics, the branch of mechanics concerned with describing motions without considering the forces behind them, was used for the experiment, specifically the area of human motion. This topic is covered in university introductory physics courses, as well as high school physics courses. A learning tool in the form of a web-based tutorial was developed by the authors using a combination of video footage and data from a motion detector, which was edited in Macromedia Flash MX 2004™. The main focus of the tutorial was the exploration of position-time and velocity-time graphs corresponding to simple motions, and thus much of the time in the tutorial was spent focusing on graphical representations of motions, however, the tutorial endeavoured also to elaborate on the concepts behind the graphs, and strong emphasis was placed on students’ ability to appropriately and scientifically describe a motion in words. The tool allows students to view videos containing demonstrations of simple motions performed by students, and, in explanatory videos, watch the demonstrations occur whilst the output of the motion detector is shown real-time, alongside the video footage.

It was decided to study web-based learning in this domain for several reasons, the principal reason being its simplicity, and the lack of assumed scientific knowledge required, since the subjects participating in the experiment have little or no scientific training. The kinematics domain is also a well-structured area of physics, and the progression that students follow, from simple concepts to more complicated situations is clearly defined and easily re-created. The specific area of human motion was chosen due to the ease at which real-time demonstrations could be filmed in conjunction with a motion detector, and the increased accessibility of the information, providing students with familiar situations; it has been argued that, where it is possible to do so, a demonstration involving a person is more effective than one focused on a machine or other apparatus (ref).

3. Preparation of stimulus material

3.1 Designing the video treatments

Since the experiment aims specifically to assess the utility of the ILD format, developed by Thornton, the specific format of both the individual videos and the tutorial’s general structure was based on the kinematics ILDs produced by Thornton (ref), which follow the general structure:

¿ General Introduction → ¡ Introduction → ¬ Prediction Prompt → √ Explanation

Step 1 forms the beginning of the ILD, and steps 2 – 4 are repeated for each individual demonstration. Due to the nature of each specific intervention condition in this experiment, this format was not entirely followed in all cases in the tutorial, the specifics of the treatments for different conditions are detailed in Section 3.2, and a flow-chart representing the structure of the tutorial is shown in Figure 4.

All videos were narrated by a female student of a similar age to the subjects, and the physical demonstrations were done by a male student of the same age. Screenshots are given in Figures 1-3. The videos which were viewed by all three treatments are discussed below.

The general introductory video described to the subjects how the tutorial would operate, and some time was taken to define and briefly explain the terms ‘speed’ and ‘velocity’. An understanding of the vector nature of velocity is essential in achieving a full understanding of kinematics, however, since subjects were not assumed to be familiar with vectors, and since the tutorial did not require such an extensive understanding, the following simplified description was given: “Velocity is very similar (to speed), again it’s the speed of an object, but velocity also has a direction attached to it. Because the motions we're going to be looking at today are fairly simple, we only need to note down the direction in the velocity as whether it is either 'going away from some point', or 'coming towards some point', which we show with the sign of the velocity, whether it is positive or negative.”

Fig 1. Screen shot of general introduction

The introductory videos for each demonstration had the narrator describe a motion, and then showed that motion being performed, by walking to/from the motion detector. A full list of the motions is given in Appendix 3. To make it easier for the subjects to visualise the corresponding graphical representation, subjects were instructed:

“James is going to walk at a constant speed. Imagine he's moving perfectly smoothly.”

Fig 2. Screen shot of an introductory demonstration video

The explanatory videos repeated the video of the motion being performed, and the output of the motion detector (edited in Flash**) was superpositioned on the video, effectively showing the graph being drawn in real-time, as the motion was carried out. The video then zoomed in on the graph, and the narrator described the important features of the graph with respect to the motion. The graph was then ‘straightened’, that is, a straight line was superimposed which represented what the motion would look like if it were carried out at a perfectly constant speed, (See Appendix 3), and the narrator explained: “The line’s a bit wobbly, since James wasn’t really moving at a constant speed, but we can see that if he could move perfectly smoothly, it would look like *this*.”

Fig 3. Screen shot of explanatory demonstration video

3.2 Structure of stimulus material with respect to intervention conditions

Three learning intervention conditions were evaluated in the present study, and represented three forms of the Predict, Observe, Explain model, each with a different level of prediction required from subjects, the three prediction conditions being: (1) Mandatory prediction from a set of given options (MP condition); (2) Auto-prediction, in which students are asked to make a prediction but are not required to (AP condition); and (3) Non-prediction, in which students are not asked to make any predictions (NP condition). This section describes the structure of the tutorial used in the three intervention conditions (MP, AP & NP).

A flowchart depicting the structure of the tutorial for each condition is shown below.

Fig 4. Flowchart depiction of tutorial structure

3.3 Mandatory prediction (MP) condition stimuli

MP subjects were shown the general introductory video, followed by the five demonstrations, each of which began with an introductory video. The subject moved from each introductory video to a prediction screen that consisted of a variety of position-time and/or velocity-time graphs, from which the subject was asked to select the one they thought best described the previously shown motion.

Fig 5. The prediction screen for the first demonstration

The subject was asked to also justify their choice in words, and from there they viewed an explanatory video for each demonstration. After viewing all five demonstrations, the subject was directed to the post-test.

3.4 Auto-prediction (AP) condition stimuli

AP students were shown the same introductory and explanatory videos as the MP students, but instead of the prediction screen, they were shown a short video prompt asking them to draw, on paper, a position or velocity-time graph which corresponded to the shown motion. Screen shots of these prompts are shown in Fig 6. The script of these prompt followed the general form quoted below, but varied with each demonstration:

“On a piece of paper, please draw a set of axes like this, with position (velocity) on the Y axis, and time on the X axis. Draw a line on the graph indicating what you think the motion will look like.”

Fig 6. Screen shots from a prompt for subjects in the AP condition

The paper from these students was not collected, and their predictions were done of their own volition, as opposed to the MP students who were required to make a prediction from those shown onscreen. The subjects moved to the post-test after the five demonstrations.

3.5 Non-prediction (NP) condition stimuli

NP students were not asked to predict the results of the demonstrations, and simply viewed the introductory and explanatory videos, passively, before proceeding to the post-test. The NP condition moved much more smoothly from one demonstration to another than the MP or AP conditions, and took less time to complete, because it did not require the subject to make predictions.

4. Method

4.1 Subjects

Subjects were 71 undergraduate students from The University of Sydney, Australia. They were enrolled in a Science Foundations course, a mandatory unit for students enrolled in a Bachelor of Education (Primary), wishing to teach at a primary school level. The subjects had been taught some rudimentary kinematics in the Science Foundations course, but, through conference with the lecturer and course coordinator of three years, were understood to be largely unfamiliar with scientific graphical representations of motion. All subjects were randomly assigned to one of three groups: mandatory prediction, auto-prediction and non-prediction.

4.2 Experimental procedure

Subjects were tested individually via the World-Wide Web, at their own convenience. The experiment took the form of an online tutorial, which included a short quiz. All subjects were randomly sorted into a particular condition, and followed the treatments detailed in Section 3. Subjects were instructed to complete the program and post-test individually, without the aid of textbooks or the Internet. A small participation mark was awarded to subjects who completed the ILD, which went towards their mark in the Science Foundations course. All individual videos were under two minutes in length, and the estimated time required to complete the tutorial in each condition is tabulated below.



Condition Time spent watching Time spent predicting Time spent on post-test Total time taken Mandatory Prediction Auto Prediction Non Prediction 0 mins

Table 1. Comparison of time spent in the tutorial by subjects in different conditions

4.3 Post test instrument

The test consisted of fifteen questions, nine of which were multiple-choice; six were short answer questions. There was no time limit set for the test, but the times at which each question was answered were recorded automatically. Thirteen of the questions were focused on graphical representations of different motions, and two questions required subjects to discuss kinematics concepts without reference to graphs. The graph-centred questions consisted of two types of question. In the first type the subject was presented with either a position-time or velocity-time graph or both, and was then asked questions referring to that graph. The second type of question described a specific motion to subjects, after which subjects asked to choose the corresponding position-time and/or velocity-time graph from a variety of options. In several cases multiple-choice questions were grouped with a short-answer question asking the subject to elaborate on their answers. All questions related to material shown in the instructional program, no prior knowledge of physics or kinematics was assumed. The lecturer and coordinator of the Science Foundations course was consulted regarding the language used and difficulty of the questions, and asserted that the level of difficulty was appropriately challenging for the cohort. A complete list of the post-test questions is given in Appendix 1. The same post-test instrument was used for all subjects.

4.4 Instructions to subjects in experimental conditions.

The subjects in all conditions chose to view the recordings in either Windows Media Player or Quicktime format, depending on their preference, via the World-Wide Web, through standard Web browsers. Subjects could view the recordings as many times as they wished, and there was no time limit set for them to complete the tutorial. 5. Results 5.1 Quantitative results The results from the post-test were considered in three groups, the multiple-choice questions, the short answer questions, and the test as a whole. The results are shown, comparing the three conditions, in Fig 7. (a) (b) (c) Fig 7. Distribution of marks attained in post-test for each condition. The next step is to see whether or not there is a significant difference between the marks achieved by subjects in the different conditions. To do this we use Analysis Of Variance (ANOVA) ***Yet to find some info about this. A couple of lines or so. Will do. *** The means and standard deviations of the results for each different condition are shown in Table X. Condition Mean Standard Deviation N Mandatory Prediction 7.6667 1.23827 30 Auto Prediction 7.0556 1.43372 18 Non Prediction 7.2258 1.49910 21 (a) Condition Mean Standard Deviation N Mandatory Prediction 9.1429 4.09006 30 Auto Prediction 8.6111 3.91286 18 Non Prediction 8.7097 3.96816 21 (b) Condition Mean Standard Deviation N Mandatory Prediction 16.8095 4.66497 30 Auto Prediction 15.6667 4.83857 18 Non Prediction 16.3188 4.18110 21 (c) Table 2. Descriptive statistics for (a) multiple choice questions, (b) short answer questions, and (c) total mark, by condition Using one way ANOVA no significant difference in the results of the different conditions was found, F(2, 66) = 0.418, p = 0.660. 5.2 Results from Mandatory Prediction subjects From the subjects in the MP condition we gathered the predictions that were made throughout the tutorial. For each of the five demonstrations the subject was asked to make a prediction from a choice of several (See Fig X) and justify their prediction in words. The number of correct predictions made by each subject was then calculated, and the results are shown in Figure X. Fig 8. Distribution of marks from predictions in the tutorial. A breakdown of the number of correct responses for each demonstration is shown in Table 3. Demonstration 1 2 3 4 5 Correct predictions 17 16 6 13 14 Table 3. Number of correct predictions for each demonstration ( N = 21 ) The written explanations given by subjects for their choice of prediction tended to be brief and had little reference to physics terms, with some answers ranging from confusion “this is due to the fact that the man stopped at a point not beyond the point”, to a correct description lacking physics terminology, and reference to the graph, “James should travel twice the distance in the same amount of time.” 5.3 Qualitative results Throughout the responses in the short answer questions in the post-quiz, several main misconceptions emerged, which appeared to be equally spread across each condition. Briefly, these misconceptions were: • A point regarding the vector nature of velocity. Several subjects’ responses treated velocity simply as an indicator of direction, and speed as an entirely different concept. A typical answer (to Question 4, see Appendix 1.) reads: “…The speed will be faster however the velocity would be similar as they are moving in the same direction…” • A confusion regarding distance traveled on a position-time graph. In Question 3 (See Appendix 1) subjects are shown a position-time graph of two runners and asked who ran fastest, and who ran furthest. Subjects seemed comfortable in determining who was faster from the graph, but had difficulty in establishing who ran further. Many subjects took the length of the lines on the graph as representing the distance run, or noted that since the slower runner ran for a longer time, they must have run further (which they did not). A typical answer is: “They have run the same distance as their lines are the same length if they were to be placed side-by-side.” • Many students mistook the flat lines in a velocity-time graph, which represent motion at a constant speed, as pauses in motion, that is, times of zero velocity, even while the line is clearly in a non-zero position. This was most prevalent in Questions 9 and 10, in which many subjects mistook the initial constant velocity of the object for the object being stationary. • When subjects were asked if a particular position-time and velocity-time graph represented motion of an object which returned to its original position (Question 6 (d))many responded that since the line on the velocity-time graph did not return to its original position, the object did not return to its initial point (which it did). A typical response reads: " No, (the object did not return) because the velocity at the start being 1/2 m/s was different to the ending velocity which was negative (-2m/s)” There were also several areas in the test in which subjects performed beyond expectations, in particular: • The majority of subjects read values off the graphs accurately, and many were able to use this information when describing a motion in words. The level of detail and understanding shown in the following answer (to Question 6) was beyond what was expected, “Initially the person increased its distance whilst maintaining a constant velocity of 0.5 m/s. He/She then accelerated to 3m/s covering 6m in just 2 seconds. His velocity remained constant for 2 seconds before he slowed down to -2m/s as the person returned to his/her original position.”, and yet this answer is indicative of the level of many for Question X. • Most subjects successfully interpreted simple position-time and velocity-time graphs to a level beyond what was expected, describing the motion shown correctly with reference to the graph and physics terminology, giving correct examples of objects which could produce such graphs in real-life situations, “The object was initially moving at a constant velocity until it slowed down and stopped (hence the 0 velocity). A car may have produced this graph - driving and then slowing down and coming to a stop.” • The responses from subjects in the short answer questions in the post-test tended to make frequent reference to the material and situations presented in the tutorial, or from prior knowledge, for example “(a graph could have been produced by) a person moving steadily away from a motion detector and then quickly slowing down and stops.” 6. Discussion 6.1 Effectiveness of web-based ILD The results from the multiple-choice questions (Fig 7. (a)) show a clear skew towards full marks in all three conditions, with more than 50% of students achieving eight or nine out of nine. This trend can also be seen in the distribution of the total marks attained in the post-test, (Fig 7. (c)) with a total average score of 68.0%. This result is higher than was expected, and implies that either the level of the subjects’ understanding of kinematics was underestimated throughout the development of the post-test, or that the tutorial has, indeed, served as an effective tool in improving students’ understanding of kinematics. Without a pre-test, however, it cannot be determined which was the dominant factor affecting the subjects’ excellent results. The distribution of marks in the short answer questions appears to be quite flat for all conditions, without a general trend for any of the conditions, although, as in the multiple choice, there were no subjects who received one or zero marks in the multiple choice questions, yet there were several who achieved full or nearly full marks, indicating again that the subjects have performed at a level above what was expected. While the lack of a significant result in differentiating between the three conditions was not achieved, the subjects performed well in the post-test, indicating that they understood the material presented in the tutorial, and the references in subjects’ responses to the material shown in the tutorial indicate that the online format of the tutorial was effective in communicating the ideas it was intended to. 6.2 Non-significant difference It was expected that the subjects in the MP and AP conditions would perform at a higher level than those in the NP condition, as results in similar, classroom based kinematics ILDs show a significant increase in students’ understanding (Sokoloff & Thornton, 1997). Aside from the online format of the tutorial discussed here, the main difference between it and those presented by Sokoloff and Thornton is the fact that Sokoloff and Thornton presented their ILDs in a series, following from ‘Human Motion’ kinematic concepts on to ‘Motion with Carts’ demonstrations, and finished with an ILD discussing Newton’s first and second laws. Further research into whether such a combination of online ILDs leads to a higher level of understanding, rather than focusing on a particular topic could yield important results and implications for the design and format of online teaching resources. A discussion of other possible factors affecting the subjects in the three individual conditions, and their results, is given below. The split-attention effect, in which students presented with two or more different forms of information experience a high working memory load, adversely affecting their learning (Mousavi, Low & Sweller, 1995), comes from cognitive load theory (Chandler & Sweller, 1991), and can be extended to situations in which students have to split their attention between a computer and an external source, such as paper (Cerpa, Chandler & Sweller, 1996). The students in the AP condition were asked to make their predictions on a piece of paper, thus splitting their attention between the online tutorial and the paper on which they were working, and the subjects in the MP condition were shown prediction screens, the instructions for which were given in text (See Fig 5.), while the rest of the tutorial was narrated vocally, thus splitting the attention of the MP subjects between verbal and textual information. This may have adversely affected their ability to learn from the tutorial, and lowered their results in the post-test. Most of the research done on the split-attention effect has focused on paper sources which provide instructions, rather than paper on which the subject works, and thus further research into whether or not combining a paper worksheet or prediction sheet with an online tutorial adversely affects students’ learning outcomes would need to be done to determine the extent to which this effect is seen. Briefly considering the differences in time taken for subjects in different conditions to complete the tutorial, we see that subjects in the MP and AP conditions spent more time on the tutorial, and thus the issue of sustained interest arises. The subjects participated in the tutorial outside of a classroom environment, at their leisure, and did so in order to obtain a small mark towards their course. These factors suggest that the subjects’ interest and engagement in the tutorial will be lower than subjects in the NP condition. The tutorial represents extra work in the subjects’ own time, for little result, and if significant persistence and effort is necessary in the task, as is the case for the MP and AP groups, then subjects are likely to disengage from the task. It has been shown that levels of situational interest affect students’ ability to learn (Palmer, 2004), and while this is a factor which would affect subjects in all three conditions, it is also known that students’ level of interest drops quickly over time (ref), implying that the subjects in the MP and AP conditions could be expected to have a lower level of interest throughout the tutorial than the subjects in the NP condition, adversely affecting their learning. Hence the anticipated learning gain in the MP and NP subjects could be offset or contracted by subjects’ decreased interest, focus of attention and task engagement. 6.3 Results from Mandatory Prediction (MP) condition subjects Considering the results from the predictions done by the subjects in the MP condition, the results do show a skew towards full marks, showing that the subjects were not overly challenged by the level of difficulty in the situations presented in the tutorial, but there is a reasonable spread across the range of marks, indicating that many of the students had some difficulty with this area of physics. We see that the number of correct predictions for each demonstration was reasonably constant and high (Table 3.), with the exception of the third demonstration, in which only six of the twenty-one subjects correctly predicted the velocity-time graph. This graph corresponded to motion away from, and then towards the motion detector. There were few comments recorded by subjects regarding their predictions on this demonstration, and those that were left tended to correctly describe the situation with reference to the sign of the velocity, even though the graphs chosen were incorrect. This was the first point at which a velocity-time graph had been introduced in the tutorial, and while subjects did not do well in this particular prediction, there was no ongoing trend regarding understanding of velocity-time graphs following this, suggesting that the subjects learned from the explanatory video following the prediction. The predictions made about the fourth demonstration, in which James walks away from the motion detector, pauses, and then returns to the detector twice as fast as before, are of note because all but one of the subjects who predicted incorrectly chose the ‘distractor’ option, which had a velocity-time graph showing James’s velocity on both legs of the trip to be the same. This highlighted an area of difficulty, specifically the ability to differentiate between different velocities from a position-time or velocity-time graph, which was reflected in the answers to the post-test, as was mentioned above in 5.3. Although the level of successful predictions of the MP subjects cannot be compared with the results in the post-test, one can note that both the correct reference to physics terminology in describing motions, and the level of sophistication in subjects’ responses notably increased from the predictions to the post-test, which, combined with the subjects’ references to material in the demonstrations in their post-test answers, indicate that the tutorial has been effective in teaching subjects about the area of kinematics. 6.4 Misconceptions The misconceptions arising in the subjects’ responses in the post-test (See Section 5.3) form clear areas of difficulty, and the effectiveness of an online tutorial addressing these misconceptions directly, rather than simply presenting the subject area and simple demonstrations would be an interesting area of investigation, given that multimedia teaching tools have been shown to be an effective format for promoting conceptual change (Muller, Sharma, Eklund & Reimann, 2006**). These responses also highlight the possible use of online tutorials such as the one investigated here as methods for gauging areas of strength and weakness in students’ understanding of a particular subject, while acting to improve their understanding, in a non-assessment environment, and outside of class hours. 6.5 Graphicacy The nature of the subject matter within the domain demanded that the majority of the material within the tutorial related directly to graphs, a full understanding of which is underpinned by subjects’ level of graphicacy. The graphs were shown being produced in real-time, and presented as simply and plainly as possible, in an attempt to reduce the difficulty for students in interpreting them. It has been noted that high error rates can be encountered when subjects are asked to interpret data from graphs, and that their ability to do so corresponds to their level of scientific training (Schield, 2006), highlighting the area of graphicacy as one in which students, especially those in a non-science field, would benefit from additional learning resources. The high ability to interpret graphs shown by subjects in their post-test responses indicates that the online ILD format can be effective in improving students’ level of graphicacy, though further research would be required to investigate which factors in the design of the ILD (i.e. watching the production of the graphs in real-time, simplification of the appearance of graphs in the tutorial) are the most important in improving students’ understanding. 6.6 Implementation It is also important to note that subjects were instructed to complete the tutorial alone, and research suggests that some students benefit more in learning activities which are group-based, as some students act as explainers, others as ‘explainees’ (Greeno, 1998). In fact Thornton asserts that his ILDs must be combined with group discussions in order to facilitate learning. Research into situativity of learning suggests that the environment and conditions in which learning takes place are important factors in affecting students’ understanding, and our investigation had little to no control over the conditions in which subjects completed the tutorial. Running the tutorial as part of a computer laboratory session in class time would control the situation in which students completed the tutorial, and could be organised to have students working in small groups or alone, in order to investigate the effect of students’ environment on their learning from an online teaching resource. Alternatively, the web-based ILD could be run in a regular lecture with the use of data projectors, thus reducing the time taken for lecturers to prepare similar demonstrations. 7. Conclusion In an online environment, students who predicted, or were asked to predict the graph corresponding to a demonstrated motion did no better than students who made no predictions. Students did, however, perform to a standard higher than was expected, and their responses indicated that they had learned from the online tutorial, suggesting that the online form of the ILD was successful as a teaching resource. Further research into the effects of split attention in online tutorials, as well as the effect of students’ environment on their learning whilst completing such a tutorial would need to be done to determine the extent to which these factors affect the successful translation of a classroom-developed ILD into an online environment.