A CONCEPTUAL HELP TO FACILITATE CHEMISTRY LEARNING Francois-Marie BLONDEL Institut National de Recherche Pedagogique Departement Technologies Nouvelles et Education (Tecne) 91, rue Gabriel Peri 92120 Montrouge - France tel : 33 1 46 12 87 15 - fax : 33 1 46 12 87 01 e-mail : [email protected] Abstract : When they are working with open computer environments, students often ask questions on terms and phrases, expressing a need for a specific help. In this paper, we present a method for the development of a conceptual help, which can be used in different learning activities. When we designed conceptual help, our purpose was to start from a minimal knowledge of concepts which are assumed to be mastered by the student and to develop extensions following his/her interests and the tutor intentions. Our authoring method is intended to enhance the reusability of elementary components, and to facilitate updating and adaptation. It uses standard tools to produce the different pieces of help that are defined as independent components. Using this method, a complete conceptual help system was prototyped in a project of learning environment (Schnaps) about the chemical reaction. Basically, conceptual help is made of short sequences of text and interactive graphics and is organised in a hierarchy of concepts. Interactive graphics are specially developed to demonstrate the dynamic part of concepts. The material displayed for one concept is generated from a selection of short sequences using an accumulation-restriction process which extends the minimal description of a concept with explanations and comments, and are supposed to meet the interest of the student. The adaptation to the curriculum is performed by the tutor who specifies the concepts and the explanations to be presented to the student. Keywords : development of CAL software, help design, application in chemistry learning, authoring techniques, distance education

1.Introduction When they are working with open computer environments using simulation, problem solving or data analysis software, students often ask questions on a large variety of subjects. These questions which range from "What can I do now ... ?" to "What is the meaning of ... ?" often express a critical need for help. Some of these questions can be directly answered by the teacher in traditional classes. In a context of distance education, it becomes necessary to consider different solutions : help can be provided by a tutor, by other students or by specific functions included in the computer environment [Vivet 92 ; Selker 94 ; Winkels 92]. We studied some of these questions and suggested solutions in developing a help system for a project of learning environment (Schnaps) in chemistry [Blondel & al 97]. In this environment, the student is given a set of tools to define and solve problems on the quantitative aspect of the chemical reaction. A diagnosis of the actions is performed and gives details about the student activity. The help system is composed of several modules that are assigned a unique role in the entire help system : conceptual help, procedural help, style hints, general advice. The purpose of this paper is to present the module which is in charge of conceptual help on chemical reaction. We first discuss the objectives and techniques of this particular kind of help. In the next sections, we describe the structure of concepts and the methods and tools used for authoring. Finally, we present the access facilities and answer adaptation features. Proceedings of 4th Int. Conf. On Computer Aided Learning and Instruction In Science and Engineering, Alvegard C. (Ed), Chalmers University, Göteborg, 15-17 juin 1998, pp. 295-304.

2.Objectives and methods 2.1Need for a conceptual help Among computer interactive learning activities, problem solving is considered as an opportunity for students to practice their existing knowledge in various situations. In a discussion of the role of coached problem solving in science learning, Van Lehn notices that human tutors often interrupt problem solving with discussions of basic principles and concepts of the task domain [Van Lehn 96]. On the other side, we observed that during problem solving in chemistry, students often request for clarification and explanation of terms. This specific need for what we named conceptual help, generally appears in questions about the phrases that are used in the problem statement and about the comments that are made by the tutor or displayed by the system. Students express their needs by questions about facts ("What is the physical state of matter of FeS2 ?"), about quantities (" ... the amount of substance, it's a number of moles ?"), and about features of the reaction ("Can I do the same thing with masses as I did with moles ?"). When they are working alone, students have some difficulties in finding the answers in textbooks or glossaries. When they are using computer learning tools, it seems appropriate to develop specific help functions capable to provide them with the required information. For these reasons, we think that conceptual help must include : 1- explanations of terms and phrases, sometimes called terminology help, 2- explanations about the fundamental notions of the domain, and 3- justification of the rules that can be applied in the problem. Our main purpose in designing a conceptual help is to rely on a minimal knowledge of concepts which are assumed to be understood and to develop extensions following the student interests and the tutor intentions. According to this idea, conceptual help is assigned two objectives : · to recall basic existing knowledge, which is generally known by students, · to expand this knowledge with detailed related information and new topics. Used in connection with problem solving tools or other learning tools, help about a concept must be available to the student by the time he/she uses it, and adaptation must be effective along working sessions.

2.2Help Techniques In most applications like those developed in Windows, on-line help, which often includes a glossary, is based on hypertext techniques [Kearsley 88]. Usage of static hypertext help has shown different limitations and especially the absence of adaptation to users who all see the same contents in the same way. But the most important shortcomings come from navigation problems : most users lack a representation of the structure of the hypertext and have difficulties in finding or recalling the pages they want to read. Current research on instructional hypertext in science [Dee-Lucas 96] has shown that "providing more structure in a hypertext facilitates the development of efficient strategies" and that "users with more segmented hypertext went directly to the most specific relevant units and were less likely to select irrelevant units". All this argues for an explicit structure (at least a detailed table of contents) and for small units to facilitate memorisation. [Wentland & Forte 95] have suggested sophisticated structures of concepts adapted for large documents such as textbooks. This interesting solution which puts a stress on the link types and their pedagogical usage, assumes that a reference book capable of receiving the agreement of future users is available. Otherwise, designing such a hypertext from scratch seems difficult to achieve.

Proceedings of 4th Int. Conf. On Computer Aided Learning and Instruction In Science and Engineering, Alvegard C. (Ed), Chalmers University, Göteborg, 15-17 juin 1998, pp. 295-304.

Conceptual help may also be seen as a reference manual like in ELM-ART [Brusilovsky & al. 96] in which a page of the manual contains a brief explanation of one of the course concepts and links to the related course units. Therefore, the manual (Introductory Lisp programming, in that case) performs the role of a conceptual help and a glossary of terms.

2.3Help main features From the above considerations, we choose : · to structure all concepts into a hierarchy, · to build help contents from easy-to-write and reusable small units, · to adapt the content to the user by generating the answer to his/her interrogation, enlarging the minimal explanation of a concept with its relations to others. We do not consider conceptual help as an isolated instructional product, but as a flexible resource which can be integrated with other resources into a more general environment. Even if we developed this first prototype in conjunction with other tools in a problem solving environment, the conceptual help module is independent and may be used with other applications.

2.4Collecting concepts Before specifying the contents of help, we started a pilot study of students' requests in problem solving. We recorded the interactions between a tutor and a student working with the computer environment, with no help function available, in order to determine the object of student questions and tutor answers. This first study was completed by a study of interviews of chemistry teachers and an investigation of the main current textbooks covering the domain of chemical reaction. From these data, we elaborated a list of concepts (about 180) which covers all the notions that can be encountered in solving problems. Besides the fundamental notions which can be identified in the chemical reaction, the list of concepts incorporates all the objects and entities that are manipulated in the problem statement, in the problem representation and also in the development of the solution. Quantities and units Basic quantities Units (International System of Units ) State of matter Homogeneous / heterogeneous states of matter Phases Different physical states of matter Change of state of matter Physical and chemical system Physical quantities that specify the physical and chemical system Relations between quantities for a compound Chemical reaction Chemical reaction in the laboratory Chemical equation Relations between quantities in the chemical reaction

Figure 1 : The two first levels of the hierarchy of concepts Concepts are organised in a hierarchy. This organisation, which uses only the generic-specific link, may introduce some recurrence but remains one the most simple structure to memorise for both students and authors. Furthermore, it provides a general overview of the domain and allows to exhibit some aspects often ignored by students.

3.Structure of the help contents 3.1Short sequences of text and graphics Basically, help is made from elements that can be used to produce a multimedia : text, pictures, animations and interactive graphics. They are aggregated into short sequences of a few elements Proceedings of 4th Int. Conf. On Computer Aided Learning and Instruction In Science and Engineering, Alvegard C. (Ed), Chalmers University, Göteborg, 15-17 juin 1998, pp. 295-304.

(between two and three on average) that are shown together in a predefined order (text + picture, text + animation + text, ...). To improve active reading by the student, we paid special attention to the design of interactive graphics. Some of them are specially suited to demonstrate the dynamic part of a concept. For instance, we created several graphics showing symbolic transformation during chemical reaction.

Figure 2 : Illustration of a chemical equation Another kind of interactive graphics are relational graphics like the ones introduced by [MorinetLambert 92], which reifies relations between concepts by spatial representation. For instance, a "star" graph shows relations between mass, volume, molar mass, molar volume and amount of substance and their algebraic relations for the same compound.

Figure 3 : In this "star" graph, when the user clicks on a quantity, the graph shows relations with adjacent ones. Most interactive graphics are generic objects in the way that some parts can be instanciated with specific data. This feature allows the author to adapt the displayed objects to a particular situation, for instance by choosing a compound, a reaction or a value.

3.2Organisation by concepts Short sequences are organised by concepts. The constituents of a concept are the following : · a phrase or a term, which identifies the concept and is used to give access to it, · a minimal description which represents the basic knowledge associated to the concept, generally a short sequence containing a definition-like text with minimal explanations, · extensions, which represent what a tutor would like to add to the minimal description when considering different viewpoints on the concept. The extensions may be of the following types : · a general explanation of the concept, · a specific explanation, which allows context-dependent comments, · extensions connected with other concepts, i.e. explanations about the relation of this particular concept with another one. For instance, the concept of mass has extensions connected with units, molar mass, and mass conservation. For each concept, there can be as many extensions as already defined ones. minimal description

Proceedings of 4th Int. Conf. On Computer Aided Learning and Instruction In Science and Engineering, Alvegard C. (Ed), Chalmers University, Göteborg, 15-17 juin 1998, pp. 295-304.

A gas has a definite mass but no definite volume and no definite shape ; its expands to fill the volume of its containing vessel. A gas consists of independent particles spread in the whole volume, in random motion.

extension : ideal-gas An ideal gas obeys to the universal gas equation : P . V = n .R.T where : V is the volume expressed in m3, P the pressure in Pa, n the amount of substance in mol, R the gas constant In most chemical problems, one assumes that gas are ideal.

T the absolute temperature in K

extension : molar-volume The molar volumes of all gases at the same pressure and temperature are equal, or, Equal volumes of all gases contain equal numbers of moles of substance at the same pressure and temperature. This is called Avogadro-Ampere principle.

extension : stp The molar volume of any gas at s.t.p. (T = 273 K or 0°C and P = 105 Pa) equals : V0 = 22,4 L . mol-1 or 22,4 . 10-3 m3 . mol-1 In usual conditions where temperature is 20°C or 25 °C, the molar volume of a gas is 24 L or 25 L, depending on temperature and pressure.

Figure 4 : The minimal description and some extensions to the concept of gas showing only text fragments

4.Authoring tools and architecture The method we took for authoring was intended to enhance the reusability of elementary components (we think that reusability is essential in instructional design), and to facilitate updating and adaptation by different authors. Therefore, we choose common and widespread tools to produce the different pieces of help, and we develop specific tools to combine them in a complete system. By separating the building and presentation processes, we tried to make help components as independent as possible. In brief, a conceptual help document is prepared as a set of specially structured Word files and Toolbook graphic objects. The task is then to convert these files into an appropriate format and to link the different elements for presentation requirements.

4.1Building graphics All graphic objects (still images, animations and interactive graphics) which are designed with a multimedia authoring tool (Toolbook) are independent components capable of answering to messages. These messages are used to adapt some parts of generic graphics, to set the initial views and to prepare the animations, before displaying the objects to the user. Authors may design graphic objects as they like ; some restrictions are placed on the size of the object (to be entirely visible in the viewer) and the syntax of the messages (object_type, object_name, command, list_of_parameters), like the following : group,equation,animate,1,CH4,2,O2,2,H2O,1,CO2

To facilitate duplication and reuse, all the elements of a graphic must be encapsulated in a single object. The methods corresponding to the messages must be attached to the object in a specific way (user-defined properties in Toolbook).

4.2Writing text and describing sequences Text fragments are written by the authors directly with a word processor (Microsoft Word). Fragments are organised in short sequences by concepts and extensions. The descriptions of sequences, which form the essential and complementary parts of a concept, are written along with text fragments in the same documents. Sequences are described as a succession of Proceedings of 4th Int. Conf. On Computer Aided Learning and Instruction In Science and Engineering, Alvegard C. (Ed), Chalmers University, Göteborg, 15-17 juin 1998, pp. 295-304.

text fragments and graphic objects. Graphics are referenced by their identification (page and name) in the Toolbook document, followed by the adequate messages with the desired values. The created files are then saved in an intermediate RTF format (Rich Text Format). *extension*molar-volume The molar volumes of all gases at the same pressure and temperature are equal, or, Equal volumes of all gases contain equal numbers of moles of substance at the same pressure and temperature. *tbk*avogadro,group,avogadroampere This is called Avogadro-Ampere principle.

Figure 5 : Excerpt of a sequence The resulting RTF text and description files are processed by a parser (written in Lisp) to store text fragments with only essential attributes and to create the complete original hierarchy of concepts. The resulting structure contains texts in minimal format and references to graphic objects. Word

RTF

Text fragments

minimal RTF

Lisp

Conceptual help structure and data

parser

references

Lisp Help provider

Toolbook

Toolbook

Graphic objects

Concept explorer / viewer

Author

User

Figure 6 : Author and user sides of the architecture

5.Access and adaptation The user interface is divided in two windows : a concept explorer and a viewer. The content of both windows displayed to the user is controlled by the help provider.

5.1Concept exploration The student can explore the set of concepts in three complementary ways : · a hierarchy, which can be developed branch by branch, · an alphabetic list of the concepts terms, · an index of the words used in the concepts terms. The set of concepts available for presentation is specified initially by the tutor who selects which concepts or branches of the graph to prune or to keep in the whole of the hierarchy. These settings initiate the concept student profile, which is saved along working sessions. Between sessions, the tutor has the capability to progressively add concepts to this student profile. When the user clicks on a particular concept term in the explorer window, a query is sent to the help provider which selects what kind of extensions are to be added to the minimal description of the highlighted concept and sends the required data to the viewer.

Proceedings of 4th Int. Conf. On Computer Aided Learning and Instruction In Science and Engineering, Alvegard C. (Ed), Chalmers University, Göteborg, 15-17 juin 1998, pp. 295-304.

5.2Content selection The selection of extensions is based on an accumulation-restriction process determined by a set of rules, which represent the following statements : 1. add extensions connected with concepts that have been already requested by the student. 2. add extensions connected with concepts whose content was read with enough attention by the student and remove extensions connected with concepts whose contents was not read by the student ; attention is estimated by comparison with a standard duration of reading ; below a threshold value, the contents is considered as not read by the student. 3. add extensions connected with concepts of special interest as defined by the tutor or determined by the analysis of the student's activity during learning ; this analysis is performed by the diagnosis module in our problem solving environment ; the set of concepts of special interest are stored in the student profile. The accumulation-restriction process that expands the contents of help during a session is applied along sessions. From one session to the other, the provider updates the student profile with the list of concepts he/she examined with a minimal attention. The consequence of this process is to increase the amount of relevant information presented to the student.

5.3Help viewer The viewer's role is to present in one or more pages, the text and graphics that compose the answer to a particular concept. This answer includes the added extensions prepared by the help provider and transmitted to the viewer using a simple communication facility (Dynamic Data Exchange between applications in Windows). The help viewer, also developed with Toolbook, contains all the procedures that are necessary to display the content. Text fragments are adapted in size and shape. Graphic objects are copied from the original book and then initiated with the values indicated in the related messages.

Proceedings of 4th Int. Conf. On Computer Aided Learning and Instruction In Science and Engineering, Alvegard C. (Ed), Chalmers University, Göteborg, 15-17 juin 1998, pp. 295-304.

Figure 7 : Concept explorer and help viewer

6.Conclusion and future work In this paper, we have presented an architecture and a method for the development of a conceptual help that can be used during several learning activities. The contents of help are build from short sequences of text and interactive graphics and are organised in a hierarchy. The material displayed for one concept is generated from a selection of short sequences using an accumulation-restriction process which extends the minimal description of a concept with explanations and comments about the relations between this concept and other concepts which are supposed to meet the student interest. The adaptation to the curriculum is performed by the tutor who selects the concepts and the kind of explanations to be presented to the student. The adaptation to the student activity is available through a list of concepts, which are to be given by an external diagnosis module. The authoring method uses standard tools to produce the different pieces of help that are defined as independent components and assembled in the generated answers. Using this method, a complete conceptual help system was prototyped in a project of learning environment about the chemical reaction. To develop a conceptual help for another domain, the author has to define the list of concepts in a hierarchy, to write down the text fragments using a RTF compliant word processor, and to build graphic objects using Toolbook. If the new domain is close to the previous one, some text fragments and graphic objects may be directly reused without any transformation. After this first prototype, we plan to port the contents and tools to a WWW platform. The authoring method we used will facilitate the port. The concept explorer and viewer need to be rewritten as software components capable to work under a Web client, the other tools and especially the help provider which is on the server side, remaining unchanged. All text fragments with usual attributes, all still image illustrations will be easily converted to HTML and graphic formats, respectively. Just interactive graphics have to be rewritten as Java components, to remain independent of the client platform.

7.Acknowledgements The development of conceptual help was made by a team that comprises Monique Schwob, Martial Tarizzo, Pierrette Max and Annie Ringard. This paper would not have been written without their helpful and significant contributions.

8.Bibliography [Blondel & al 97] Diagnostic et aide dans un environnement d'apprentissage ouvert : une exemple en chimie, Schnaps, Sciences et Techniques Educatives, vol 4, n° 4, 1997, pp. 377-412. [Brusilovsky & al. 96] Brusilovsky P., Schwarz E., Weber G., ELM-ART : An Intelligent Tutoring System on World Wide Web, in ITS'96, Intelligent Tutoring Systems, Third International Conference, Frasson C., Gauthier G., Lesgold A. (Eds), Springer, 1996, pp. 261-269. [Dee-Lucas 96] Dee-Lucas D., Effects of Overview Structure on Study Strategies and Text Representations for Instructional Hypertext, in Hypertext and Cognition, Rouet J-F., Levonen J., Dillon A., Spiro R. (Eds), Lawrence Erlbaum, Mahwah, 1996, pp. 73-107. [Didaskalia 94] Les cartes conceptuelles, Didaskalia, n° 5, INRP, 1994. [Kearsley 88] Kearsley G., Online Help Systems : Design and Implementation, Human Computer Interaction Series, Ablex Publishing Corporation, 1988.

Proceedings of 4th Int. Conf. On Computer Aided Learning and Instruction In Science and Engineering, Alvegard C. (Ed), Chalmers University, Göteborg, 15-17 juin 1998, pp. 295-304.

[Morinet-Lambert 92] Morinet-Lambert J. Interface Improvement using Graphics, An Example in Quantitative Chemistry, in Proceedings Ninth International Conference on Technology and Education, N. Estes and M. Thomas (eds), Paris, Morgan Printing, 1992. [Selker 94] Selker T., COACH : A Teaching Agent that Learns, Communications of the ACM, vol 37, n° 7, 1994, 92-99. [Van Lehn 96] Van Lehn K., Conceptual and Meta Learning During Coached Problem Solving, in ITS'96, Intelligent Tutoring Systems, Third International Conference, Frasson C., Gauthier G., Lesgold A. (Eds), Springer, 1996, pp. 29-47. [Vivet 92] Vivet M., Uses of ITS : Which Role for the Teacher ?, in New Directions for Intelligent tutoring Systems, Costa (ed), NATO ASI Series, vol 91, Springer-Verlag, 1992, pp. 171-180. [Wentland & Forte 95] Wentland M., Forte E., Modélisation pédagogique d'un domaine de connaissance : De l'arbre pédagogique d'un texte au réseau conceptuel d'un hyprtexte, in Environnements Interactifs d'Apprentissage avec Ordinateur, tome 2, Guin D., Nicaud J.F., Py D. (Eds), Eyrolles, 1995, pp. 125-136. [Winkels 92] Winkels R., Explorations in Intelligent Tutoring and Help, IOS Press, Amsterdam, 1992, 228p.

Proceedings of 4th Int. Conf. On Computer Aided Learning and Instruction In Science and Engineering, Alvegard C. (Ed), Chalmers University, Göteborg, 15-17 juin 1998, pp. 295-304.