The Use of Virtual Models for Training Procedural Tasks

Mona van der Vloed (Claessens), Rik Min & Jef Moonen


This paper reports on a research project in which it was investigated by what means Virtual models can be used to support the training of maintenance personnel at the company "Hollandse Signaal apparaten B.V.". On the basis of interviews and a literature study an experiment was planned to investigate suiting design criteria (leading to effective task performance) for virtual models, that are used to train procedural tasks. It is expected that for training simple procedural tasks a simultaneous presentation of information (coaching and virtual model) will be advantageous. However the involvement of complicated tasks is anticipated to benefit from a serial manner of presentation (instruction versus virtual model).

1 Introduction

Nowadays the use of virtual models in the industry is increasing. Grounds are because this method of designing and testing products is relatively quick, cheap and save. For example, to investigate the impact of a car crash virtual models can be used to save the expenses of a real car (VOLVO).

"Hollandse Signaalapparaten B.V.", a Dutch company that produces naval combat systems and trains their customers to operate or maintain these systems also uses virtual models. In addition they have a Virtual Reality room at their disposal in which VR-glasses can be used to see ‘real 3D’. Until now this company uses these models particularly as means of communication, both between personnel from different departments and between ‘Signaal personnel’ and their clients. With the help of these visualized systems personnel with different expert knowledge have the opportunity to discuss design issues in an early stage of product development. Also clients have the opportunity to see their ‘product’ and to take part in the design process in advance.

The department of ‘International Training’ at "Hollandse Signaal apparaten B.V." also wants to make use of these virtual 3D models but then for training purposes. In this paper we will discuss why it seems useful to do so and what the design should be like then.

2 Virtual models within the context of "Hollandse Signaal apparaten B. V."

One of the researchers involved was situated in the company of ‘Hollandse Signaalapparaten B.V.’ what enabled her to have informal conversations with experts from different divisions and gave her easier access to ‘Signaal-documentation’ [1, 2, 3, 4]. Interviews were taken from the instructors to explore the context within which these virtual models were to be used [5]. Questions were asked about the course in general, training methods used, the instructors and students.

This lead, among other interview conclusions, to the following list of reasons in favor of using virtual 3D models for training:

Educational reasons

Economical reasons

Thus, using virtual models as individual practice tools seems a safe and cheap manner to make students master the subject matter. To prove this hypothesis it was planned to investigate the difference between training with these models and training with traditional training methods (paper manuals and videos). However, first it had to be investigated how to design these three-dimensional models in order to improve task performance.

3 Virtual models for training procedural tasks

Here procedural tasks are defined as tasks in which one has to follow repeated similar actions step by step (procedure) to achieve a goal. This requires mainly skill- and rule-based knowledge [6]. Examples of training procedural tasks are programming a video, maintaining a machine or driving a car.

Certain aspects appeared to be important for efficient human information processing during training procedural tasks [7]:

Thus, context information, interaction and coaching (feedback and instruction) seemed to be important factors in order to gain effective task performance.

Virtual models do provide students with situational context information. It is possible to catch the whole system at once in a virtual model while in the manuals (the working of) systems are described with the help of text and technical drawings from the separate parts of the system. In addition virtual models can be dynamic and / or interactive. This makes them more compatible with real task performance than the manuals in which the subject matter is translated into words and separate pictures. This is supported by a study of Witmer et. Al [14] in which it appeared to be easier to find a route in a complex office building by rehearsing in a VR model, than by means of verbal rehearsal and studying maps of the building.

On the contrary coaching is more dependent on language than on moving images. After all, a great deal of human communication goes via symbols and signs. The question here is how instruction (which involves directions about theory and about practice) can be provided best when making use of virtual models. Should instruction be provided in advance of practicing with the 3D models or should instruction and virtual models be provided simultaneously?

There are several studies that contribute to the question whether text should be presented separately [8, 15, 16, 17]. We will discuss some of them in this paper. Diehl and Mills found that propositional representations lead to better recall, while (opposite to Paivio’s theory of ‘Dual coding’ [15] situational one's only seemed to increase task performance (inferencing, recognition tasks and problem solving) [8]. Propositional representations hold the text's meaning (text) and situational representations represent the situation described by the text (pictures, motoric enactment, etc.) [16, 17]. Their explanation is that when only text is provided students are able to concentrate purely on the meaning of the text without getting distracted (good propositional representation). This all implies that virtual models will only be useful with respect to task performance, while text is needed to enhance theoretical understanding.

Another study, from Benshoof and Hooper, investigated the effect of presenting information simultaneously on one screen against the effect of a serial presentation of information on multiple screens [18]. This study showed that low ability students performed equally for single-, and multiple window treatments, while high ability students performed better in the multiple window condition. These results were explained by the fact that the latter manner of presentation was more demanding with respect to student’s memory strategies (the formula that was needed to calculate sums was on another page than the sums were). It forced students to process information more deeply and accurately, resulting in better task performance. As a consequence high ability students that used short-term memory processing strategies more effectively [19] performed better than low ability students when they were forced to use their own learning strategies. Thus, when text and virtual models are separated a deeper information processing may be elicited. However different effects are expected when the text does not deal with complementary information .

There are also studies in which a simultaneous presentation of information is preferred [20, 21, 10]. A study that investigated the Parallel Instruction Theory of Min showed that better results were obtained with information presentation on multiple windows. The difference between these results and those from the previous study is caused by the fact that here they were dealing with strategy free tasks that were not as memory intensive [22]. Other evidence comes from literature about cognitive psychology [23].

Actively associating information as images in mind appeared to be a good mnemonic strategy. In an experiment people were either asked to repeat pairs of words (cow -cheese, cow - cheese, etc.) or to visualize combinations of these words (a cow that eats cheese) to remind them. People that had had the assignment to visualize combinations performed better on recall tasks. This is because active ‘participation’ was involved here and because images are more distinctive than pieces of text or words are. For example, when one is referring to a chair this can be related to a al kinds of variations of chairs, while a picture of a chair excludes al other chairs than the one represented.

In short, there are arguments for and against the merging of instruction and practice. These happened to depend on the kind of task and on the type of target group. On the basis of this and our previously gained knowledge about the situation at the company involved we designed an experiment to test our expectations. Expectations and experiment will be described in the next section.

4 Designing virtual models for training procedural tasks within the context of "Hollandse Signaal apparaten B.V."

From the large collection of tasks at "Hollandse Signaal apparaten B.V." a simple procedural task was selected for our experiments. The simple character of the task, like in strategy free tasks [22], was chosen here to decrease the change of having to deal with many different problem-solving strategies. The chosen task was selected on the basis of expert knowledge from teachers. However it was chosen to be difficult enough to avoid ceiling effects so changes between the different experimental conditions could be found when present. In addition, for "Holland Signaal apparaten B.V." it had to be a task that recurred in many courses and that dealt with an unclassified system [5]. Later experiments will be done with more complicated tasks, in which we assume serial presentation of information will be more useful.

Based on the literature and on the information gained from the interviews it is expected that it is best to present information together with the virtual models as much as possible. This is expected because we are dealing here with a simple task [22] in which task performance is more required than knowledge-based problem solving [8], and with a target group with a low entry-level [18]. Therefore it seemed better to support students and to maneuver them into an efficient learning situation. This will be done by providing them with appropriate ‘just-in-time’ information (hints and feedback) during the procedure. It is possible to do this either by pop up texts, by auditory instructions or by means of visualizations. For example, drawing maintainers’ attention on safety aspects by either giving them a written or spoken warning: "watch out high voltage!" or showing them a flickering red light inside of the system.

Research design We will maintain a between subjects design to prevent from learning effects that will distort the data. Groups of 15 subjects will each be assigned to one of the four conditions (in total 60 subjects will participate). The zero condition will be a serial presentation of the instructional text and the virtual model. Condition 1, 2, and 3 will respectively deal with instruction that is presented within the virtual models: auditory, by visual text and by visual imaging. In addition, the same experiment will be repeated with people that have a good control of the language used in the courses. This is because a bad control of language may distort the difference between the visual text and auditory condition on one hand and the visual imaging condition on the other. This experiment will enable us to generalize to other situations than specific circumstances given.

Subjects A sample of 120 people will participate. Subjects will be clients from the "Hollandse Signaal apparaten B.V." (60) and people that answer to the same requirements except that they will master the language used (60).

Procedure Subjects will be given a paper instruction in which it will be asked to study the materials in order to be able to perform the task afterwards. The time they will be exposed to the (instruction and) the model will be determined on the basis of some pilot experiments. After this they will be tested. This means they have to perform the task on the real system. Finally, they will be asked to fill in a questionnaire with questions about the perceived difficulty and satisfaction.

Data Analysis

The time it takes to perform the task together with the failures they make will be rated. A MANOVA will be performed on these data.

Answers of the questionnaires will be summarized, recorded and then classified on the basis of their contents. They will be arranged in tables according to their contents and the frequency with which they were mentioned.

Summarizing, our hypotheses are that for training simple procedural tasks a simultaneous presentation of information (coaching same time as the virtual model) will be advantageous, because it enables students to actively associate related information. When complicated tasks are involved we anticipate students to benefit from a serial manner of presentation (instruction before virtual model), because it induces people to process the information deeper.


[1] Signaal (2000). Versatile Interactive Computer Based Training for Operation and Repair. VICTOR Program Management Plan. Hengelo: The Netherlands.

[2] Signaal (1999). Manual for corrective maintenance. Hengelo: The Netherlands.

[3] Signaal (1999). Manual for adjustment procedures. Hengelo: The Netherlands.

[4] Signaal (1997). Van oorschelp tot radar; de geschiedenis van een bijzondere onderneming. [From auricle till radar; the history from a special company] Drukkerij Twente Hengelo, Hengelo: The Netherlands.

[5] Van der Vloed, F. M. M., Min, F. B. M. & Moonen, J. C. M. M. (in press). "The use of multimedia for training maintenance tasks" International Journal of Continuing Engineering Education and Life-Long Learning.

[6] Rasmussen, J. (1983)."Skills, rules and knowledge: Signals signs and symbols, and other distinctions in human performance models" IEEE transactions on systems, man and cybernetics. Vol SMC-3, 3, 257-268.

[7] Van der Vloed, F. M. M., Min, F. B. M. & Moonen, J. C. M. M. (unpublished manuscript)."The use of virtual objects in multimedia for training procedural skills; A theoretical foundation"

[8] Diehl & Mills, 1995) "The effects of interaction with the device described by procedural text on recall, true/false and task performance" Memory and cognition, 23, 6, 675-688.

[9] Shepard, R.N., & Cooper, L.A.. (1982). Mental images and their transformation. Cambridge, MA: MIT Press. In Winn, W. (1996). "Virtual environments in maintenance training." Human interface technology laboratory an college of education, university of Washington, Seattle, USA. December, 1996. klik hier

[10] Van Leyden, J. (1993). Psychologische Functieleer, theorie, techniek, toepassing; Psychonomy, theory, technology, implementatin Bohn Stafleu Van Loghum: Houten/Zaventem

[11] Winn, W.D. (1997, March). "The effect of student construction of virtual worlds on the performance of high and low ability students" Paper presented at the annual meeting of the American Educational Research Association, Chicago, Il.

[12] Osberg, K.M. (1997). "The effect of having grade seven students construct virtual environments on their comprehension of science" Paper presented at the annual meeting of the American Educational Research Association, Chicago, Il.

[13] Dix, A. J. (1998) Human Computer. Interaction London: Prentice Hall.

[14] Witmer, B.G., Bailey, J.H., Knerr, B.W., & Parsons, K.C. (1996). "Virtual spaces and real-world places: Transfer of route knowledge" International Journal of Human-Computer Studies, 45, 413-428.

[15] Paivio, A. (1991). "Dual coding theory: Restrospect and current status" Canadian Journal of Psychology, 45, 255-287.

[16] Van Dijk, T. A. & Kintsch, W. (1983). Strategies of discourse comprehension New York: Academic Press.

[17] Perrig, W. & Kintsch, W. (1985). Propositional and situational representations of text. Journal of memory & Language, 24, 503-518.

[ 18] Benshoof, L. A. & Hooper, S. (1993). "The effects of single- and multiple window presentation on achievement during computer-based instruction" Journal of computer-based instruction, 20 (4), 113-117.

[19] Case, R. (1985). Intellectual development: Birth to adulthood. New York: Academic Press Inc.

[20] Min, F. B. M. (1997). "Nieuwe ontwikkelingen op het gebied van computersimulaties en multimediale leermiddelen in het hoger onderwijs: Het World Wide Web en Java" [New developments in the discipline of computer simulations and multimedial learning tools in higher education.] Tijdschrift voor hoger onderwijs & management. Themanummer hoger onderwijs online, 4.

[21] Min, F. B. M. (1994). "Parallelism in open learning and working environments" British Journal of Educational Technology, 25, 2, 108-112.

[22] Sternberg, R. J. (1985). Human abilities. An information processing approach New York: W. H. Freeman and Company.

[23] Baddeley, A. D. (1990) Human memory: Theory and Practice Needham Heights, Mass.: Allyn and Bacon.