By Edwin Agelink, Wim de Boer, Natascha Tebbe and Oscar Peters

Department of Education and Technology, University of Twente, The Netherlands

This article has been made as an assignment for the course 'Capita Selecta Instrumentation Technology' in association with Dr. Ir. F.B.M. Min

Enschede, November 1997


The interface of a computer provides information mostly in a similar way like with a television. According to Min, information for learning purposes should be presented in a parallel setting and not a linear way. Processing information in a parallel setting is called: Parallelism. Parallelism is the way the designer and the user organise the learning environment on the monitor as well as in the surrounding of the computer screen. On assumptions of parallelism Min developed the Parallel Instruction Theory (PI-Theory). The PI-theory attempts to explain why certain learning environments do not result in good learning behaviour whereas others do. It is a theory about shaping instruction with simulation and it possibly explains why linear learning and working environments are not really suited to the creation of good simulation environments and working environments with large amounts of data with electronic instructions. In this article the PI-theory is linked to a instruction learning theory, constructivism. In this article an effort has been made to use the insights derived with the PI theory and the experiences through the years making simulation software, to build a working environment based on this experiences. The example described in this article is the Randstad case, where a spreadsheet program was redesigned according to parallelism. Parallelism seems to be a good solution for learning as well as working environments.


Experiments with prototypes based on the MacTHESIS philosophy turned out that for a large number of pupils and trainees everything in a learning environments should be within reach and crystal clear. This seemed technically hard to realise by courseware writers or most of the educational software houses (Van Schaick Zillesen, 1990; Min, 1992). With tutorial multimedia it is known that certain learners are annoyed by the disappearance of the subject matter which has been read. Technically speaking calling back text which has been read is for almost all tutorials still problematic or not common sense. The problem with separate paper materials is that students and trainees think they can do without. Even teachers tend to think so. As a result beautiful computer simulations remain unused at schools, in spite of the perfect design, styling and user-interface. Our earlier learning environments are characterised by modern input and output techniques and a wide range of different kinds of visualisations, from abstract to concrete. In spite of all that it turned out that the instruction method was the decisive factor. Multimedia and hypertext techniques of in particular Macintosh computers, and the MacTHESIS system and philosophy, made it possible to build the parallel instruction theory.

First the MacTHESIS philosophy will be described. After that, the Parallel Instruction Theory itself will be described. The experiences with computer simulation will be discussed here. Constructivism is a knowledge acquisition theory and the PI-theory can be used as a instrument that support constructivism. In the last section the Randstad case is described. In this example a spreadsheet program was redesigned according to parallelism.

The MacTHESIS philosophy

The MacTHESIS philosophy behind the computer programs of Min et al. is described and showed in products on many conferences (Min et al. 1987 and 1992; Van Schaick Zillesen, 1990). The MacTHESIS philosophy is based on parallelism, which means that every part of information, feedback or instruction will be in view. The design environment, the teacher environment, and the learning environment each have their own characteristics design parameters and a simple human-computer interaction. The learning environment is very characteristic and will be described in this paper. Therefor first the experience with simulations will be described. A complete simulation learning environment consist in large series of elements. The most important requirements and characteristics are:

These points are based on concepts of parallelism and the PI-theory which stated that all these elements should be (parallel) available to the student. Some elements must be presented on-line (multimedia materials) others can be off-line (paper instructions). In the research of Min the balance between the simulation parts and the instruction parts is investigated.

The Parallel Instruction Theory

The Parallel Instruction theory is based on the concept that in a simulation environment everything should be in view, even the paper materials. Apart from the graphic output of the computer simulation program, also the instruction program, possibly a help system or a video window ('model-driven desktop video') and the paper instruction materials are in sight.

The PI-theory tries to explain why certain learning environments do not result in good learning behaviour whereas others do. It is a theory about shaping instruction with simulation and it possibly explains why computers which do not work with a multi window environment are not really suited to the creation of good simulation environments with electronic instructions. The theory deals with a number of issues including (Min, 1992, 1993):

Being collaborative, interactive and interdisciplinary, multimedia technology promises more than just an improvement in educational practices. It can boost learning up to five times compared to traditional one-way teaching methods (Erzurumluoglu, 1995). In 1992 Min searched and found that a multi-tasking operating system was extremely useful for educational computer simulations programs in two respects. Firstly the multi-tasking system allows the simulation part to be made with simulation tools and the instructional part to be made with cheap authoring tools. This situation is very natural for the design of this type of software and is of interest for the 'second' designer - the teacher in the school. Both parts are designed by specialists but in entirely different ways and with different tools. Secondly it allows a certain measure of a-synchronic working, as with paper-based instruction materials, because the two parts have stand-alone characteristics. Parallelism is thus the way the designer and the user organise the learning environment on the monitor as well as in the surrounding of the computer screen.

Min distinguish three types of parallelism, in the real world but also in the electronic learning environments. In the first type designers often use one half of the screen for instruction and the other for the actual learning environment. These solutions, called viewports, indicate the awareness of the designer of the user's need to put information side by side in order to be able to compare things. It proves that users need to have things parallel (Min, 1996). This is defined as parallelism of the first order. The problem with this is that the designer needs a larger screen in the end than that provided on the standard PC. Designers have come up with all kinds of solutions, e.g. by cramming a screen with all kinds of information. This is 1st order parallelism, but it has all kinds of ergonomic disadvantages: text is too compact, syntax is poor due to statements that are really too brief, an excess of information, letters on the screen are too small and so on.

In the second type of parallelism with windows, the effective screen surface became larger than 100% due to the arrival of multi-windowing applications. This is a big advantage. Pull down menu's also became rapidly popular all over the world. Not lastly because there is a quite distinct form of parallelism in pull down menu's of the second order. The last type is 1st order parallelism combined with something else: usually a relative linear coaching element. This is defined as 3rd order parallelism.

How does the PI-theory work in a practical setting? Most of the experiences of Min are in the field of simulation. Most simulation programs consists of a software program and a manual or other written instructional guidelines. In a simulation program given in a parallel setting the learner has a complete overview of the simulation and the instruction through the viewports. These viewports can consist of multiple monitors or one big monitor. The learner can make a decision in which way he wants to handle a given problem. Therefore a learner creates his own strategy. After the input, the output will immediately show in the output-window. In that way the learner keeps his survey and can directly adjust his strategy.

Parallelism and learning
Ever since cognitive science influences learning theory and research, learners are no longer seen as rather passive respondents to various environmental stimuli as was the case in the behavioural tradition. On the contrary, cognitive approaches to learning stress that learning is active, constructive, generative, goal oriented and depends upon the mental activities of the learner. Therefore, the currently accepted approach in learning and instruction, which is called constructivism, sees learners as actively constructing knowledge and strategies (Jonassen, 1991).

Computer simulations can offer a learning environment that is appropriate for active learning. A computer simulation is a program that incorporates a model of a process, phenomenon, system, etc. The learner is able to control input values of the simulation model and examine the resulting changes in the output. Computer simulations therefore invite the learner to activity, both in manipulating the domain, i.e. the variables and parameters of the model offered in the simulation (De Jong, 1991).

Learning with computers simulations is characterised as exploratory learning, which consists of active, constructive and goal oriented processes. The learner has to discover general rules, procedures, or higher order skills. The general idea is that this active attitude of the learner encourages meaningful incorporation of information into the learner's cognitive structure.

The main purpose of educational computer simulations is to help students to construct mental models of dynamic processes. We don't know the how such mental models in our head looks like, we do know how experts organise there knowledge of processes. We assume there is a relation between the knowledge and the external representations and that instruction that corresponds to the external format experts use, facilitate the construction of mental models by students. Elements in the knowledge representation of processes are:

The strength of traditional computer simulation is in the last element: the relations. Because of the active involvement of students and the required mental processing in building these relations computer simulations are superior to other forms of learning. A complete learning environment with computer simulation should provide parallel to the core simulation instructional possibilities on the other elements of the dynamic process. This instruction should include:

This instruction should always be accessible to users. Because of the nature of simulations we believe that compulsory instruction can cause cognitive dissonance and should be avoided. Open learning environments for simulation fail without instructions - or with poorly shaped instructions. Many experiments with over one thousand medical students using paper-based instruction materials have been carried out at Limburg University in co-operation with Struyker Boudier (1986). These simulations environments also used two parallel monitors and paper-based instruction materials. Miltenburg (1985) with his training course about dynamics of economics (1988) and Coleman with his simulation programs in university courses for medical students also showed that paper-based materials are absolutely essential to success (Min, 1993). The arrival of modern windowing computers such as Macintosh and SUN computers and MS-DOS computers today, meant a tremendous step forward. Together with Van Schaick Zillesen (1991) Min developed a designing system for simulation programs with many parallel windows, based on Macintosh desktop philosophy. A conceptual scheme or map of the mathematical physiological models was used in conjunction with the simulation output. In spite of all this researchers found that paper-based instruction materials remained a decisive factor. Recent research (Akkermans, Burg, Groenwoudt & Min,1996) found that students make less mistakes when they're using two screens with parallel windows, rather than one screen. They do not seem to work faster. The students preferred the two screen environment. The research confirms that students want to see all information (screens), in order to optimise their learning and that students want to have all information parts of the learning or working environments in view.

In the next paragraph the assumptions which the parallel instruction theory is based upon, is been measured with the well known theory of actively interpreting and constructing individual knowledge representations, called constructivism.

Introducing a technology, as simulations in education often goes together with a different view in the acquisition of knowledge and visa versa. Did the introduction of the computers in education and training changed the view in the acquisition in knowledge, as constructivism or situationalism did. Or was it the other way around? This question is not important for a educational scientist. Important for them to know is under which existing and not-existing conditions a new technology could be successfully introduced in education, or visa versa. This, to achieve the highest acquisition in knowledge.

At a lower level it is interesting to know what impact the PI-theory has on the nowadays view on education and training. As mentioned above constructivism changed the view on education. Constructivism holds that knowing is a process of actively interpreting and constructing individual knowledge representations (Jonassen, D. 1991). The PI-theory can, with the right tools, therefore be used as a instrument that support constructivism.

In the learning psychology as mentioned by Van Parreren (1984) a learner will get better learning results when he knows how he learns. Meaning that the learner is conscious of the learning task and in what way it should be handled. Therefore the PI-theory can be used in education and training because it applies to the new learning theories.

How can PI help the learner discover general rules, procedures, or higher skills? PI can be used in an efficient way to make the student aware of his cognitive skills. Not only creating his own reality but a student is conscious of those skills. He can choose from structures that are based on experiences and interpretations, which strategy is the best to solve a given problem. The strength of PI lies in the survey of information. This information lies all in the field of vision. In that way the student can make meaningful relations between the different information sources. So, if this instruction is proper introduced, the learner will discover general rules, procedures, or higher skills which can be optimised.

A teacher always will use an educational setting or environment where his matter of tuition is most effective. Min (1996) emphasises that parallelism is not a substitute but a additional tool which can help by simulations in education. The PI-theory is primarily meant as a theoretical framework to arrange learning-environments for simulations. In the following learning-environments the PI-theory can be used:

As the new educational paradigms, the PI-theory stimulates education to the individual. Therefore it is suited for teaching at different levels concerning time and pace. The PI-theory therefore can be used in education and training because it can be imbedded into a new learning theory, constructivism.

In the next paragraph the development of an electronic working environment, using the PI-theory, will be described.

Experiments and Practical Implementations: The "Randstad" case
In this next section a example of what can be done with the parallel instruction theory is been given. A design and a prototype for a working environment is described. The environment is thus not a learning environment, which doe not indicate that the assumptions of the PI theory can not be used. There are of course a lot similarities between learning and working environments, so that this example of experiences of bringing the PI theory in practice is also interesting for educational settings.

Randstad Holding NV supplies services to businesses and institutions through 35 operating companies in nine countries. The essence of all services is to provide personnel and skills to organisations, allowing them to operate more flexibly. The problem of this case can be stated as follows. The task of the division Strategic Planning of Randstad Holding NV in Diemen (the Netherlands) is to train all employees of Randstad in Europe and the US, helping them to plan the organisation. For this purpose a model is designed, developed and implemented in a spreadsheet program, that depicts a complex model, consisting of approximately 1500 variables. The spreadsheet covers an imaginary space that is 50 times the size of an actual computer-screen. The experiences which Min had with earlier developments of programs where valuable for the search for an answer for this problem. Computer simulations have in common with this spreadsheet program that it is a program that incorporates a model of a process, phenomenon, system, etc. The user is able to control input values of the model and examine the resulting changes in the output. Computer simulations and spreadsheet programs therefore invite the user to activity, both in manipulating the domain, i.e. the variables and parameters of the model offered in the input data.

Figure 1. Visualisation of a flat spread-sheet

The problem is that working in this environment, navigation and comparing the input with the output (resulting elsewhere in the program), demands a lot of the user. This is illustrated in the above figure (1). When the user changes a variable that is visual on the actual screen (b), the output (depicted in figure 1 as c and d) isn't shown to the user so mistakes can occur due this inaccurate procedure. The whole spreadsheet program (a) is too big to be placed on the actual screen (b).

The user, therefore, needs a environment which is based on parallelism, which means that every part of information, feedback or instruction will be in view. The following steps from the strategic planning process have been elaborated, such as the input of start up data, the input of data regarding the current year, starting points for a strategic planning and the summing of results. Within the steps of entering data for the model in most cases it concerns two windows. An input and a output window are displayed parallel on the screen. These solutions, called viewports, indicate the awareness of the designer of the user's need to put information side by side in order to be able to compare things. This is a example of parallelism of the first order. With these two windows the relations between input and output are depicted more clearly, so the user can make a better comparison (figure 2). The convenience of the user will increase even more due to the visual representation of the data in the form of line-diagrams and histograms.

Figure 2. Use of windows according to the Parallel Instruction Theory

The program has a different design than the original program (depicts the cards). The output in figure 2: c' and d' are visual on the screen by means of viewports. This approach leads to a better working environment for the user where everything that's important information for the user is in view, along with the paper materials which users need.

The elaboration of the ideas resulted in a pilot project that was finished at the beginning of 1995. This project has resulted in three prototypes that overlap each other. These prototypes have all been developed using the 'rapid prototyping' method (Tripp, D. & Bichelmeyer, B. 1990). Randstad went on with a new company (Kopal and Gritter Multimedia). They produced a second series of prototypes. At the end of august 1995 Kopal and Gritter had produced a working prototype in HyperCard. The final product is produced with Toolbook Multimedia 3.0 and fits in MS Windows. The final program has been completed January 1996.

Experiences with the prototype and the final product showed that the parallel approach was preferred by the users, rather then the linear approach which did take more attention, and led to more mistakes. The use of the program, which is now used throughout whole Europe, and which has been developed is a example which explains that PI is not only a (hypotical) theory, but also will work in practice. More tests, however, are needed to indicate if inexperienced users can swiftly learn handling a planning system.

Concluding Answers

In this paper an effort has been made to explicate parallelism and parallel instruction. People are best motivated to receive instructions when they can decide for themselves at what time and how much they need it. Linear instruction often arrives at the wrong time and in the wrong quantity . It has disappeared by time it is needed most. Most interactive media prove in practise to be a collection of often linear concepts, with consequent drawbacks. Using the Parallel Instruction theory and the MacTHESIS system will yield better results. This integrated instructional equipment will be easier to use for most users, and a deeper level of understanding will be achieved. The teacher has a wide range of tools to modify and change the instruction materials as well as certain aspects of the computer simulation program, e.g. the visualisation of the conceptual model, some texts and default positions of pages (windows).

Based on the experiences of Min can be stated that computer simulations have generally proved more successful when accompanied by paper workbooks. Computer based materials that did not take the parallel instruction theory in account, failed time and again until arrival of the multi-tasking operations systems. Total electronic learning needs a user environment in which things can be kept in view for an unlimited period of time.

Using the idea's of the parallel instruction in practice showed that learning and working environments improve when the information is parallel offered to the user. It seemed that the design of the interface is of great importance for a comfortable use of, in the case of Randstad, a database. In the future, more research should point out the possibilities of designing a learning (or working) environment according to the parallel instruction ideas, and weather this approach can be seen as a successful way of designing an environment.


Akkermans, W., Burg, J. ter, Groenewoud, U.A. & Min, R. (1996) Effect van mate van paralelisme op taakuitvoering en gepercipeerd gebruiksgemak University of Twente, Faculty of Educational science and technology.

Coleman, M. & Manns, T. (1988). Software quality assurance. Basingstoke: Macmillans cooperation.

Erzurumluoglu, B. (1995). New Horizons in Education. Multimedia and interactive computing. The Fountain, July - September, No. 11p. 4-7.

Jonassen, D. (1991). Obejectivism versus constructivism. Do we need a new philosophical paradigm. University of Calorado.

Jonassen, D. (1992). Constructivism and the technology of instruction: a conservation. Lawrence Erlbaum Associates - XI, 221 p.

Jong, T., de & Joolingen, van (1991). Instructional use of simulations: a learner's and authors perspective. University of Twente, Faculty of Educational science and technology.

Miltenburg, H. (1986). Software economische politiek, een simulatiespel. Bleiswijk: Economica

Min, F.B.M. (1992). Parallel Instruction, a theory for educational computer simulation. Interactive Learning International, Vol 6, no. 3, p. 177-183.

Min, F.B.M. (1993). CD.ROM titled: "Simulations"; "Experimental products with parallel instruction". University of Twente, Faculty of Educational science and technology.

Min, F.B.M. (1994). Parallelism bij coachien van gebruiker in open leeromgevingen. Simulatie: een state of Art. Proceedings van Conferentie 'Simulatie 1993' te Utrecht; organisatie: SISWO (TIM), Amsterdam.

Min, F.B.M. (1995). Simulation Technology & Parallelism in Learning Environments; Methods, Concepts, Models and Systems. De Lier: Academisch Boekencentrum.

Min, F.B.M. (1996). Parallellisme bij werk-, leer- en doeomgevingen: De parallelle Instructie Theory voor coaching en open leeromgevingen voor simulatie. Unpublished manuscript, Universiteit Twente, Faculteit Toegepaste Onderwijskunde, Twente.

Min, F.B.M. (1996). Shortcomings of the monitor. The problem of linear presentation media in learning situations; the importance of parallelism in open learning and working environments. Paper on WEB (

Parreren, M., van (1983). Leren door handelen. Onderwijsvernieuwingen in de klas. Apeldoorn: Van Walraven.

Raaij, van B., (1997). Leren via beeldscherm nog altijd onhandig. Parallelisme volgens Rik Min ideaal voor leeromgevingen. UT-nieuws 5 juni 1997, Universiteit Twente, Enschede.

Schaick Zillesen, P.G. van; Gmelich Meijling, M.; Reimerink, B. & Min, F.B.M. (1995). Computer support of operator training based on an instruction theory about parallelism. Kluwer Academic Publishers.

Schaick Zillesen, P.G. van (1990). Methods and techniques for the design of educational computer simulation programs and their validation by means of emperical research. Academisch proefschrift, Universiteit Twente.

Shuell, T. (1986). Cognitive conceptions of Learning. Review of Educational research. Educational Technology Research and Development: v56 n4 p 411-36

Struyker Bourdier, T.A.F., Min, F.B.M., Kan, H. van & Thijsen, H. (1983). Computer gestuurd onderwijs: onderwijskundige informatie van het hoger onderwijs: Computer simulatie in het medisch onderwijs aan de Rijksuniversiteit Limburg, pag 266-282. Antwerpen: Het Spectrum

Tripp, Steven D., Bichelmeyer (1990). Rapid prototyping: an alternative Instructional strategy. Educational Technology Research and Development: v38 n1 p 31-44.