Revisiting variation affordance: applying variation theory in the design of educational software

ABSTRACT

Variation theory, which is a theory of learning developed by Marton and others, has quickly become popular in education research. Our purpose of this paper is to articulate the application of variation theory in the form of a number of concrete design principles that offer prescriptive and practical guidelines for improving the designs of educational software programs. To achieve this, we analyzed a wide range of educational software programs produced over the years in our previous projects for learning Chinese characters. From this analysis, we identified four design principles, namely, (i) not aiming to test but to bring about learning, (ii) focusing on a specific object of learning, (iii) allowing learners to explore variation to be learned, and (iv) keeping all other aspects invariant. These design principles are specialized for designing how learners interact with educational software programs, which is the major practical contribution of this paper.

KEYWORDS:

• Design of educational software
• design principles
• instructional design
• phenomenography
• variation theory

Introduction

Phenomenography and variation theory have recently gained much popularity in education research. Phenomenography was originally developed by Professor Ference Marton and his colleagues at the University of Gothenburg in their studies of learning. Originating from the discipline of education, it is now widely used as a research approach in Australia, Hong Kong, Sweden, and the United Kingdom (E.g. Åkerlind, 2018; Bowden & Walsh, 2000; Lam, 2018; Marton, 2015; Marton & Tsui, 2004; Rovio-Johansson & Ingerman, 2016).

Emerging from phenomenography is variation theory, which is a theory of learning that expounds what makes it possible for learning to occur. Variation theory has important implications to the practice of teaching, especially in terms of the design of teaching and instructional materials (E.g. Marton, 2015; Marton & Booth, 1997; Marton & Tsui, 2004). However, to the best of our knowledge, little has been written about putting variation theory to use in the design of educational software.

By way of background, we are teacher educators, doing research in phenomenography and variation theory in the area of learning and teaching Chinese characters. Besides this, we have Computer Science backgrounds and have substantial experience working in inter-disciplinary teams on various projects to design, develop, and produce educational software programs for children to learn Chinese characters in kindergartens and junior primary schools in Hong Kong (Dragonwise Project, 2003; Lam, Pun, Leung, Tse, & Ki, 1993; Lam et al., 2001; Lam et al., 2004).

The purpose of this paper is to concretely articulate in the form of a number of prescriptive design principles (Reigeluth, 1999) how variation theory is practically applied in improving the designs of educational software programs. To do this, we analyzed a wide range of educational software programs developed in our previous projects for learning Chinese characters. We hope that, through this paper, the identified design principles can be communicated to other researchers and instructional designers for critique in public professional discourse, which hopefully can foster more discussions for further advancing knowledge of educational software design.

This paper is organized as follows. It begins with the discussion of the ideas of phenomenography and variation theory. Next we introduce some of the features of Chinese characters. Then we explains the analysis of the educational software programs we previously produced for learning Chinese characters. Out of this analysis, we identified four design principles, which form the main thrust of this paper. The last section will highlight the major practical contribution of the design principles and their limitations as well.

Phenomenography, variation theory, and Chinese characters

Phenomenography as a research approach

In educational software design, objectivism and constructivism are predominately the perspectives of epistemology most commonly used. Objectivism assumes that reality exists independent of learners and meanings of the reality are knowable and transmissible to learners; while the assumption of constructivism is that learners construct their own realities and different learners may hold different meanings of the realities (Hannafin & Hill, 2007; Jonassen, 1991). Adopting a more encompassing stance between objectivism and constructivism, phenomenography assumes that there is only one reality but, as different learners focus on different aspects of the same reality, they see the reality as different. In other words, since their ways of seeing the reality differs, they get different meanings out of the reality (Marton, 2015; Marton & Booth, 1997).

In phenomenography research, reality often refers to a specific topic that learners have to learn (called the object of learning). The purpose of the research is to find out the qualitatively different ways learners see the object of learning, or put alternatively, the different ways the object of learning appears to learners as a collective sum. The result of the research is a number of categories of description that reflect different ways of seeing the object of learning. The unit of analysis is thus a way of seeing. Of particular importance is the qualitative difference in the ways of seeing, which stands in contrast to quantitative differences or the different extents to which learners have learned.

Phenomenography aims at uncovering the original nature of the object of learning as it appears to learners (i.e. not what the nature of the object of learning itself should be, as often laid down in the definitions in textbooks). Phenomenographers must “bracket out” their own presuppositions before inspecting the learners’ ways of seeing (Ashworth & Lucas, 1998). They must let the qualitatively different ways of seeing emerge from the data rather than fitting the data into some pre-determined categories of themselves. The result of phenomenography research is described from the learners’ ways of seeing (i.e. called the second-order perspective) rather than from those of the researchers.

In addition, phenomenography assumes that there are only a limited number of qualitatively different ways of seeing the object of learning. There is also no uniform method for researchers to arrive at an identical set of learners’ ways of seeing, which can only be justified in terms of results, but not of method. Marton and Säljö (1984) called this a “discovery procedure”.

Variation theory as a theory of learning

In the phenomenography tradition, learning is seen as a change in learners’ ways of seeing. Learners act in accordance with the object of learning as they see it. Powerful ways of acting originate from powerful ways of seeing. Thus, in a learning situation, some ways of seeing are more powerful and more desirable than the others.¹

To see the object of learning in the desired ways, learners must be able to see it in terms of certain critical aspects. The critical aspects are those aspects that learners do not yet attend to but must attend to in order to make the desired ways of seeing their own. Researchers identify the critical aspects through comparing and contrasting the qualitatively different learners’ ways of seeing. The critical aspects are those aspects that critically differentiate the desired ways of seeing from the other ways of seeing of the learners.

Learning outcomes are to become able to see the object of learning in terms of the critical aspects. Accordingly, the learners must have discerned each of the critical aspects, which entails the learners having experienced the variation in each of the critical aspects. To quote Bowden and Marton (1998), “Without variation there is no discernment” (pp. 7–8). This hypothesis that learners’ experience of variation is one of the necessary conditions for learning is called variation theory.

Teaching is viewed as a systematic effort to organize the work in classrooms to bring about the learning of learners. But “it is highly unlikely that there is any one particular way of arranging for learning that is conducive to all kinds of learning [original emphasis]” (Marton, Runesson, & Tsui, 2004, p. 3). As such, research into general teaching arrangements (e.g. Should less whole-class teaching, more project work, or more information technology be used?) is considered not focused enough. Rather than this, phenomenography adopts a content-based approach, assuming that “learning is always the learning of something: there cannot be any learning without something being learned. Focusing on what is learned implies focusing on the content of learning” (Marton, 2015, p. 22). Placing the emphasis on content, phenomenographers always focus on a specific topic as the object of learning. What matters in teaching is how the object of learning is enacted in classrooms.

Based on variation theory, teaching should enable learners’ discernment of the critical aspects of the object of learning (i.e. the learning outcomes). To make a critical aspect discernible, teachers must vary that critical aspect while keeping the other aspects unchanged. As Bowden and Marton (1998) put it, “when some aspect of a phenomenon or an event [i.e. the object of learning] varies while another aspect or other aspects remain invariant, the varying aspect will be discerned” (p. 35). After each of the critical aspects has been discerned, the learners must experience simultaneous variation in all critical aspects in the hope of fully making the desired ways of seeing their own.

Features of Chinese characters

Having mentioned phenomenography (a research approach) and variation theory (a theory of learning), we would now like to turn to some of the features of Chinese characters necessary for the readers to understand the subsequent sections of the paper. Chinese words are made up of one or more characters, which is the unit of written Chinese. For example, the word 週末 “weekend” /zau1mut6/² is formed by the two characters 週 “week” /zau1/ and 末 “end” /mut6/, adding up the meanings of “week” and “end” together into “weekend”.

Chinese has many homophones, meaning that different characters can sound exactly the same. For example, the characters 末 “end” /mut6/ and 沒 “do not” /mut6/ differ in meaning but share exactly the same sound. Thus, when there is only one sound, one often cannot tell which character is being referred to. But, when combined into a word, the character can be more precisely determined. For example, /zau1mut6/ can only refer to 週末 “weekend” as in Chinese there is no such word as 週沒. Mixing up homophonous characters (e.g. incorrectly producing 週末 as 週沒) is one of the problems children often encounter in learning characters (c.f. mixing up the homophonous English words “pear” /’per/ and “pair” /’per/).

In addition, some characters look similar in form but have different meanings and different sounds, as they are different characters. For example, the similar-looking characters 末 “end” /mut6/ and 未 “not yet” /mei6/ have no relation with each other. In the former, the upper horizontal stroke is longer than the lower one, and vice versa in the latter. It is common that children do not realize the difference in the relative length of the two horizontal strokes, thus confusing the two similar-looking characters 末 and 未 (c.f. mixing up the letters “b” and “d” in the words “bay” and “day”) (Lam, 2015).

Designing educational software programs for learning Chinese characters

The tenets of phenomenography and variation theory discussed above were supported in general terms with empirical evidence in Marton and Booth (1997), Marton and Tsui (2004), and others. We incorporated these tenets implicitly into the learning and teaching of Chinese characters in our previous projects, the outcomes of which were the production of a wide range of educational software programs for children to learn Chinese characters. Although we characterized the distinctive feature of these programs as “variation affordance” (Lam et al., 2004), how we actually applied these tenets in improving the designs of the programs have not yet been explicitly deliberated on.

Thus, in the present paper, we would like to revisit and further elaborate this notion of variation affordance through analyzing those of our programs, the designs of which were improved through the application of variation theory. The main contribution of this paper lies in the elaboration of variation theory in the context of designing educational software rather than a report on the end products of our previous projects. Here we aim at identifying a number of design principles that offer prescriptive guidelines for designing educational software programs. This paper poses the question: What are precisely the design principles grounded in variation theory for improving the designs of educational software programs?

Data source

The data used in this paper were the different versions of our educational software programs, which were collected during the development phases of our previous projects. We chose to analyze these data because of the substantial amount of experience we had of over twenty years of designing educational software programs for the same purpose of learning Chinese characters based on theories of learning. ³

More specifically, the examples of educational software programs used in this paper came from two sources. First, some of the programs were the fully functioning products of the Dragonwise Projects we produced at the University of Hong Kong from 1998 to 2006. Funded by government funding, we set up an inter-disciplinary team (including a Chinese language teacher, software developers, a school liaison officer, graphic designers, etc.) to work in collaboration with the teachers of our partnership schools to produce the programs. The programs are now widely used by teachers in primary schools in Hong Kong.

A second source of the educational software programs was a course we taught at the Hong Kong Institute of Education from 2007 to 2009. The course was titled “Information, Communication, and Technology in Early Childhood Curriculum”. The participants were kindergarten teachers (7 classes with a total of 177 female and 2 male teachers). One of the tasks for the participants was to discuss how to design effective educational software programs for facilitating the learning of children. We developed a range of PowerPoint prototypes for the discussion of this topic. These prototypes, which were educational software programs, were continuously revised as the participants made suggestions about their designs during the course.

Identification of design principles

In this paper, we did not simply derive design principles deductively from variation theory at an abstract level. During the preparation of this paper, all of the educational software programs used in this paper had already been produced, mostly after a number of revisions on their designs. We had decided to let design principles concretely come out from the different designs of the programs during the analysis of these data. There were two steps in our data analysis.

First, we carefully revisited each of the programs repeatedly several times to re-familiarize ourselves with them and to understand their designs in their own contexts. Next, we compared and contrasted the designs of different programs as well as the different versions of the same programs. We contrasted the different designs with each other to identify the directions of improvement, which denoted in what way the design of the program was better than that of another, and in what way the revised version was better than the original version (i.e. the original versions were considered as counter examples). We believe the decisions we had made to improve the designs during the development phases were more instructive than the end products. We thus identified the design principles as directions of continuous improvement rather than as a binary checklist for verifying whether the programs had adopted the design principles or not. Put it alternatively, we were in this step analyzing the qualitative difference in how the object of learning was enacted differently in these different designs of the programs.

We went back and forth iteratively between the above two steps until coherence in the design principles was reached. The comparing and contrasting of the different designs to identify design principles is similar to the way critical aspects are identified in phenomenography research by comparing and contrasting the different learners’ ways of seeing. Out of this analysis, we eventually identified a total of four design principles. One may wonder why there were four design principles, not five or three. Regarding this, we would admit that the identification of design principles was a discovery procedure and that different researchers might arrive at a different set of design principles from the same set of data (similar to the stance of phenomenography research regarding the number of ways of seeing identified). The four design principles we identified will be explained in detail in the next section for readers’ scrutiny.

Results

Not aiming to test but to bring about learning

The first design principle we identified is that educational software programs should not only aim at testing learners, that is, merely to check if the learners have understood the object of learning or not. But, more constructively, the programs should aim at bringing about learners’ learning through their interactions with the programs. In other words, the programs should take into account their past interactions with the learners and, on this basis, provide specific feedback to facilitate the learners to grasp the object of learning. This design principle can best be illustrated with the program in Figure 1.

Figure 1. Multiple-choice exercise (Original). (a) Learners are asked to fill in a missing character to complete the sentence 週＿去哪裡玩？ “Where do you go on the week___?” Three options 沒 “do not”, 末 “end”, and 未 “not yet” are provided. (b) When the learners click on the correct answer 末 “end”, the sentence will be completed with the word 週末 “weekends”. The learners will be given a tick mark with applause. (c) Otherwise, when the learners click on the incorrect options 沒 “do not”, or 未 “not yet”, which cannot be combined with 週 “week” to form a word in Chinese, the learners will be given a cross mark with a “wrong answer” sound effect.

Figure 1 is a typical design of a multiple-choice educational software program. Learners will be given positive (or negative) feedback, depending on their choice of a correct (or incorrect) option. The feedback will be the same, no matter which one of the incorrect options (i.e. 沒 or 未) is chosen.

To further improve the design of this program, we revised it to the one in Figure 2, hoping to provide learners with more focused feedback that better facilitated their learning.

Figure 2. Multiple-choice exercise (Revised). When the learners choose different incorrect options 沒 “do not” or 未 “not yet”, different feedback will be provided. (a) When the learners click on the incorrect option 沒 “do not” /mut6/, which has the same sound as the correct answer 末 /mut6/ but cannot be combined with 週 “week” to form a word, the learners probably do not realize the difference between the situations in which the two characters 沒 and 末 are used. Thus, the program will offer the feedback that 沒 is used in words like 沒有 “have not”; while 末 is used in the word 週末 “weekends”. (b) When the learners click on the incorrect option 未 “not yet”, they probably do not realize the difference in the relative length of the two horizontal strokes of the two characters 未 and 末 “end” (the correct answer). Thus, the program will show slanting lines to highlight the relative length of the two horizontal strokes in the two characters to the learners. Moreover, it is also possible for the learners to explore the difference in sounds and situations in which the two characters are used by the provision of the word examples in red.

As can be seen, in Figure 2, the incorrect choices of learners are taken as an opportunity for us to understand the learners, and in turn to bring about their learning. Their incorrect choices are seen as indicating the specific kinds of difficulty they have and what kind of feedback the program should offer to help them to learn the correct answer. For example, when learners mix up the similar-looking characters 未 “not yet” and 末 “end”, a visual cue (i.e. the slanting lines) is used to highlight the difference in the forms of the two characters. Other kinds of feedback will be offered in other situations. The program takes into account the learners’ choices in determining how to help them to grasp the object of learning.

As mentioned earlier, we identified the design principles as directions for improvement (i.e. from Figure 1 to Figure 2), rather than as a binary checklist (i.e. Figure 2 has adopted the first design principle; while Figure 1 has not). The program in Figure 2 is considered to have implemented the first design principle to a larger extent than the one in Figure 1. In fact, the program in Figure 1 does not completely ignore the learners’ choices and to a very limited extent has taken them into account to give binary positive or negative feedback on the basis of the correctness of their choices. In theory, to extend this continuity, we can add that an educational software program purely for testing learners should give no feedback at all, without even providing positive or negative feedback to learners as in Figure 1. The first design principle essentially specifies a direction to move away from pure testing of learners to attuning feedback to past interactions of learners with educational software programs in order to bring about their learning.

Direction of improvement designated by the first design principle:

Table

No feedback

→

Positive or negative feedback (Figure 1)

→

Specific feedback for different choices (Figure 2)

Focusing on a specific object of learning

Consistent with the idea to bring about learning, the second design principle is that the designs of educational software programs should be based on a certain specific object of learning (i.e. what content learners have to learn from the programs). Emphasis is placed on a content-based approach rather than general teaching approaches such as a game-based approach, which is applicable no matter what content is taught. To illustrate the second design principle, we are going to use a few counter examples, in which the objects of learning are taught as a game. We will point out the limitations in the designs of these programs in order to illustrate the direction for improving these designs using a content-based approach (i.e. the second design principle). See the first counter example – Figure 3.

Figure 3. Jigsaw Puzzle. (a & b) Learners are asked to put the pieces of a jigsaw puzzle together to form the character 末 “end”. The learners can drag and drop the pieces anywhere. When a piece is placed at its correct position, it will be fixed with a “click” sound. When the piece is placed at any other position, no feedback will be given.

The program in Figure 3 is a fun-to-play game, which should especially appeal to young children. Learners are shown the character 末 “end”. But the question is: What is necessary for the learners to succeed in this task (i.e. to correctly put the pieces together)? To achieve the task actually does not require the learners to have any knowledge about the sound or meaning of the character 末 (/mut6/ “end”). Moreover, the learners need not even bother about the written form of the character. They can simply regard it as a picture and put the pieces together according to the “picture” with no need to have any knowledge about Chinese. In other words, it is possible for a westerner who has never studied Chinese before to succeed in this task.

Adopting the second design principle, we see that the interaction provided by the above program has little to do with the content to be taught. To slightly improve the design of the program, we would suggest that when all of the pieces are correctly put in place, the sound of the character /mut6/ can be pronounced and the meaning “end” can be shown. However, since the interaction has little to do with the content, only to a very limited extent has the program taken into account the object of learning in its design. What learners are expected to learn from the program, or what exactly the object of learning is, has even not yet been clearly figured out. For example, is it to recognize the character 末 “end”, or to pronounce it, or to realize its features? Thus we see the program in Figure 3 as a counter example of the second design principle.

Direction of improvement designated by the second design principle:

Table

Requiring an understanding of something other than the object of learning (Figure 3)

→

Requiring an understanding of the object of learning

Allowing learners to explore variation to be learned

As discussed earlier, variation theory espouses that, in order for learners to discern a critical aspect of the object of learning, they must have experienced variation in that critical aspect. Consistent with this, the third design principle is that the designs of educational software programs should allow learners to directly explore variation in those critical aspects that learners must discern in order to grasp the object of learning. Adopting this design principle, we produced the program in Figure 4 in the hope of helping learners to discern the critical aspect of the relative length of horizontal strokes in characters.

Figure 4. Direct exploration. Learners are asked to explore the difference between the characters 末 “end” and 未 “not yet”. (a) At the beginning, the character 末 is shown. The learners can drag either ends of the upper horizontal stroke left or right to change its length. (b) When the horizontal stroke is shortened, the character will become 未 and its sound /mei6/ will be pronounced. Explanations about the usage of the character such as 未來 “future” will also be provided. (a) When the horizontal stroke is lengthened, the character will change back to 末 and its sound /mut6/ and explanations will be offered.

As mentioned earlier, some learners do not realize the difference in the relative length of the two horizontal strokes in the characters 末 “end” and 未 “not yet”. The above program has been deliberately designed for helping these learners to discern the critical aspect of the relative length of horizontal strokes, and what variation in this critical aspect signifies. Hopefully through interactions with the program, the learners can understand that changing the length of horizontal strokes will result in producing different characters. In this way, the design of the program allows the learners to explore variation in the critical aspect of the relative length of horizontal strokes in characters and is thus an implementation of the third design principle to a very large extent.

The difference between the program in Figure 2 and the one in Figure 4 is worthy to note. In both cases, the critical aspect of the relative length of horizontal strokes in characters is at issue. But, in the former, the critical aspect is highlighted visually using slanting lines, which is shown to learners in a static fashion. In contrast, learners in the latter take a more active role. They can tinker with the length of horizontal strokes in characters. They can find out for themselves what happens if they change the length of horizontal strokes. Thus the exploration of the critical aspect is interactive.

Direction of improvement designated by the third design principle:

Table

No highlight of the critical aspects (Figure 1)

→

Learners are visually shown the critical aspects (Figure 2)

→

Learners can interactively explore variation in the critical aspects (Figure 4)

Keeping all other aspects invariant

As a complement to the third design principle, the fourth design principles is that in the designs of educational software programs, aspects not critical for understanding the object of learning should be kept invariant, whenever possible. Variation in these non-critical aspects draws the attention of learners away from the critical ones. As such, learners should not be led to focus their efforts on exploring these non-critical aspects, which eventually will not help them to grasp the object of learning. Figure 5 below describes the design of a program for teaching the recognition of Chinese words.

Figure 5. Drag and drop (Original). (a) Learners are asked to choose one of the four pictures of fruit that matches the word printed on the basket (i.e. 香蕉 “banana”). (b) The learners can drag and drop the four pictures. When the learners put a picture in a place other than the basket, the picture will bounce back to its original location. When the correct picture (i.e. banana) is dropped into the basket, the learners have successfully completed the task. (c) When the learners repeatedly drop an incorrect option (e.g. apple or pear) into the basket, the learners will be sent to explore again the Chinese words and the pronunciations of the four pictures of fruit.

In essence, the program requires learners to choose one among the four pictures of fruit that matches the meaning of the word 香蕉 “banana”. To achieve this task, learners indeed need to be able to recognize the word, which is the object of learning. But, when the fourth design principle is applied to analyze the design of the program, one question is posed: Is the drag-and-drop action necessary in light of helping the learners to grasp the object of learning? We would say that being able to drag a picture and drop it into the basket has nothing to do with being able to recognize the word. On the contrary, with the requirement of this drag-and-drop action, it will become possible that learners, especially very young children, do not achieve the task not because they cannot recognize the word but because of having problems with their fine motor skills in performing the drag-and-drop action. Thus the drag-and-drop action is a non-critical aspect, variation in which (i.e. dragging and dropping the pictures around) should be avoided, whenever possible.

To improve the design, we would suggest that there is no need for learners to drag and drop the pictures. Learners should be able to simply click on any one of the pictures and the picture will automatically drop into the basket (i.e. click to choose). See Figure 6. In this way, during the interactions with the program, learners can focus their attention on deciding which one of the four pictures matches the meaning of the word (i.e. the object of learning), rather than on performing the drag-and-drop action. Thus variation in the non-critical aspect can be kept invariant.

Figure 6. Click to choose (Revised).

Direction of improvement designated by the fourth design principle:

Table

Learners can explore the non-critical aspects. (Figure 5)

→

Variation in the non-critical aspects is avoided (Figure 6)

As a supplementary note, we actually adopted the fourth design principle to write this paper. We considered the use of the examples of Chinese characters as a non-critical aspect, which did not help readers to understand the four design principles. Thus we tried to minimize the number of characters used in this paper and to repeatedly use the same characters and words (e.g. 週末 “weekend”) as much as possible across the different designs of the educational software programs. In this way, we hoped to make this paper easily understandable for those readers with no knowledge of Chinese.

Discussion

Identification, not confirmation, of design principles

There is no attempt in this paper to measure or collect data to verify the design principles we identified. We see the purpose of this paper as to concretely articulate what the design principles are, rather than to confirm them (See the seminal book of Glaser & Strauss, 1967, for the idea to divide research studies into those that formulate and those that confirm a hypothesis). In our previous projects, case studies were conducted on how the teachers of our partnership schools put our educational software programs into practice in classrooms (Dragonwise Project, 2003; Lam et al., 2004). These case studies showed that, with the use of the programs, the teachers were empowered with more flexibility to deliver their teaching to meet the specific needs of their own classrooms. Besides this, empirical support for variation theory, in which the design principles are fundamentally grounded, is widely available (Marton & Booth, 1997; Marton & Tsui, 2004, etc.). Given this empirical evidence, we believe there is now a need for us to more clearly spell out the knowledge we used to improve the designs of our educational software programs with concrete examples. This paper is an attempt to formulate new knowledge in the form of not only a number of design principles per se but also a deliberation on how these design principles can be applied in the production of educational software (e.g. how to design multi-choice exercises, jigsaw puzzles, etc.) Such knowledge should form the basis for future research (e.g. future studies to test the design principles in other different settings).

Designing interactions between learners and educational software programs

The main practical contribution of this paper to the literature also merits discussion. The design principles identified in this paper are meant to be of practical value to instructional designers, that is, not design principles for general usability, but specific ones for the support of learning. Existing instructional design principles are mostly for teachers to use in classrooms, such as Merrill’s (2002) first principles of instruction, which however do not specialize in the design of educational software. Closest to the design principles of this paper is the work of Mayer (2002) in designing multimedia instructional messages (i.e. how to present material using words and pictures to better foster learning). In comparison, the design principles we identified in this paper do not place the focus on the presentation of messages from educational software programs to learners. Rather, we focus on how learners and educational software programs interact with each other. What can learners manipulate in the programs (e.g. drag and drop, or click to choose)? How do the programs react and give feedback to learners (e.g. take into account previous learners’ choices or not)? Such kind of interactions between learners and programs cannot be reduced to a one-way presentation of instructional messages. Put simply, the design principles in this paper are an attempt to contribute to the knowledge of designing the interactions between learners and programs. Future research around these interactions definitely deserves more attention of educational software design researchers.

Limitations of the design principles

Near the end of this paper, we would like to take a more critical stance and discuss some of the limitations of the design principles we have identified, or more broadly, the application of variation theory in the design of educational software.⁴ Undeniably, variation theory works best with certain kinds of learning but not with others. As discussed earlier, our approach is to take the object of learning as the point of departure for investigation, which has to be decided in advance of the actual delivery of teaching. Because of this, our approach can shed only limited light on those kinds of learning in which the learning outcomes of learners are expected to be original or unanticipated (i.e. cannot be pre-determined before the teaching). Moreover, as discussed earlier, the focus of the design principles in this paper is on the interactions between learners and educational software programs. Our approach falls short of capturing the learner-learner interactions (i.e. the social aspect of learning), which are however of crucial importance in language learning.

Disclosure statement

No potential conflict of interest was reported by the author.

Additional information

Funding

This work was supported by Incentive Scheme for Developing Instructional Software: [Grant Number 20061034]; Quality Education Fund: [Grant Number 1999/2422].

Notes on contributors

Ho Cheong Lam

Ho Cheong Lam is an Assistant Professor of the Department of Early Childhood Education, the Education University of Hong Kong. He is the Programme Leader of Chinese Language Enhancement Programme for Kindergarten Teachers and was former Associate Head (Programme Development) of the department. He began to work on the topic of computer-assisted-learning for learning Chinese characters as an undergraduate final year project some twenty years ago. His Ph.D. thesis was titled Orthographic Awareness for Learning Chinese Characters, for which he received Award for Outstanding Research Postgraduate Student. He was one of the founders of the Dragonwise Projects 現龍計劃, which are widely known by the primary school Chinese language teachers in Hong Kong. He has more than ten years of experience in early childhood teacher education. His research interest lies in the areas of teaching and learning of Chinese characters, phenomenography and variation theory, and designing technology for learning

Notes

1 Pang (2003) called this shift of the primary focus of the phenomenography tradition from methodological issues to theoretical questions about learning “new phenomenography”.

2 The sound of the character here is in Cantonese, which is a variety of Chinese widely spoken by about 90% of the people in Hong Kong. The sound is transcribed using the Romanization of the Linguistic Society of Hong Kong (2002), where the “z”, “au”, and “1” denote the onset, rhyme, and tone respectively.

3 We started in the early nineties of the last century at the University of Hong Kong to carry out a series of projects (called the Dragonwise Projects) to design and produce educational software programs for learning Chinese characters under the main leadership of Dr Ki, Wing Wah, Dr Chung, Ling Sung Albert, and Dr Lam, Ho Cheong. At the earliest stage (∼1991–1995), our projects mainly focused on exploring the emerging possibilities afforded by computers. During the second stage (1995–1998), the results of cognitive psychological research on processing and learning characters were taken into account in our projects. At the third stage (1998–2006), phenomenography and variation theory had predominately become the main theoretical and methodological framework that guided all of our projects from that time onwards.

4 Lam (2013) offers a more fundamental scrutiny of the epistemological assumptions of variation theory.