Keywords

1 Background

Along with spread of PCs in common, number of interface types for operating computer has increased. Among, Tangible User Interface (TUI) has been studied which is able to operate PCs by directly touching physical objects. TUI has made intuitive operation possible. Therefore, it is believed that users are possible to operate computer without learning its method of operation and focus on original learning purpose. However, it is believed that a learning system using TUI has not been spread commonly because it has not been recognized how affect people’s learning process.

2 Purpose

In this paper, When TUI and Graphical User Interface (GUI) were used, how an adopted interface may affect learning process was focused. It is believed learning and its method suitable for a specific interface is clarified by clarifying the effects making it possible to support achieving learning purpose. In the process, actions are classified based on Contextual Control Model and difference in the actions by interface is verified by transition of the classified actions.

3 Related Works

Learning using PCs includes GUI’s Light-bot [1]. It is operated with a mouse while watching PC screen. Light-bot is a product aiming at learning a concept of programming. There is also Osmo coding [2] based on TUI. Osmo coding is capable of moving characters on a screen by arranging panels on which illustrations of actions are depicted. It is also possible to learn a concept of programming similarly as Light-bot. Thus, GUI and TUI are different interface products but working on the same issue. However, they are not such products that have been developed by focusing on how a specific interface affects specific learning. It is believed that interface is able to support achievement of learning purpose by focusing on a relationship between interfaces and learning contents and figuring out advantages. Akahori [3] compared and verified tendency of each learning tendency when using PC, tablet terminal, and paper, in order to examine tendency of learning with PCs. It has been proved that paper is excellent in recognition of letters and text data and that PC and tablet terminal is excellent in input of letters and recognition of photos and the like, respectively. It has been also revealed that paper provides sense of learning completed and tablet terminal is suitable for motivation of further learning. However, it has not been focused on what kinds of difference in concept may appear at the time of learning.

4 Experiment

4.1 Objective of Experiment

We conducted the experiment to clarify the effect on learning process when a specific interface was used, difference in learning process is to be reviewed by using TUI and GUI.

4.2 Experimental Policy

As a learning task, logical circuit learning was selected in which GUI and TUI showed the same appearance. Subjects were asked to prepare a logic circuit based on a truth table. The three logical symbol used in this experiment was the “AND”, the “OR” and the “NOT”. In addition, the lead wire objects and the lamp objects, to connect among each gate and to confirm the output of the circuit respectively, were also prepared.

GUI is operated with a mouse using “click” and “drag” while watching on a PC screen. TUI is operated using cube- shaped blocks to create a logic circuit by applying them in 2-dimensional place. Asking subjects to voice what the subjects asked to “think aloud”, and, their behavior and verbal protocol data were recorded by VCR.

4.3 Experimental Methodology

Subjects were asked to solve problems using a system to learn logic circuit. Operation manual of the interface was written out on a paper for subjects to be able to read again during experiment and instruction of the logic circuit was also written on a paper likewise.

Experiment was performed for 6 weeks (6 times) by increasing the difficulty every week. Subjects were asked to solve problems taking for around an hour in each experiment. A personality assessment test was conducted in the first week in order to recognize their original strategies. They were asked to solve problems regarding a logic circuit in the second and later experiments. In the second week, such problem was set for them to fill in an output part in a truth table in reference to a logic circuit in order for them to get familiar with a way to read truth tables. From the third and later week, they were asked to solve problems to create a logic circuit from a truth table. The problems of 3th are make logic circuit from truth table. Logical symbol of 3th is 2 pieces, 4th is 3pieces, 5th and 6th are no specified pieces for increase the difficulty level.

In the 6th, number of available logic symbols was not designated for the problem with a setting of two lamps for output. Because of the increased two output lamp, subjects were required for applied skill different from the case of problems provided in the 5th or before.

5 Analytical Methods

We focus Contextual Control Model which makes analysis based on subjects’ behavior. It is one of cognitive models.

5.1 Contextual Control Model

In order to analyze what kinds of subject behavior will take, an analysis is made in reference to Contextual Control Model (COCOM) [4] by Hollnagel. There are five modes for COCOM.

Scrambled control mode: A control to select of random or panic.

Opportunistic control mode: A control to select next action just based on the current situation.

Explorative control mode: A control to seek for new ways at a venture without any other option available.

Tactical control mode: A control to select next action according to regulations provided in advance.

Strategic control mode: A control at high level in consideration of overall situation.

An analysis was conducted using an applied model which was adapted for problem solutions in reference to five modes advocated by COCOM.

Scramble control: The state that the user has no idea of what to do to solve the problem.

Explorative control: Their sub-goal is not clear and not to go toward the goal directly but to just try to find something tentatively.

Opportunistic control: Though subjects think looking at only one or a few sub-goals toward the goal, they do not know what kind of action should be done to achieve them.

Tactical control: Subjects think several sub-goals sweeping some part of path to the goal and knows what kind of action is needed to achieve sub-goals though they have not found the whole Tactical control: Subjects think several sub-goals sweeping some part of path to the goal and knows what kind of action is needed to achieve sub-goals though they have not found the whole path for the goal yet.

Strategic control: The path toward the goal is established more detail than tactical control, and subjects know what action are required to achieve it.

Then, we focus how the five modes of COCOM may transit in problem solving process. It was frequently observed that subjects tried ideas different from thoughts they had in the process of problem solving. With a concept to regard these thoughts as a cluster, how they bring about a transition in the cluster is reviewed.

Without any time limit for the experiments required, such a scrambled control mode to work on problems in panic did not occur. In addition, because the strategic control mode occurred when they were working on an easy problem and did not transit to other mode, the mode was excluded from the current analysis.

6 Results of the Analysis

6.1 Probability of Occurrence of Mode by COCOM

We analyzed the subjects’ mode transition based on COCOM by dividing the whole period of the experiment into 3 stages to focus on their learning process as follows:

  • Initial stage: The third week

  • The subjects learn the basis of the logic circuits.

  • Latter stage: The 4th and 5thweeks

  • The subject becomes accustomed to solve the problems of the logic circuit.

  • Practical stage: The 6th (using 2 lams)

  • The subjects are asked to solve some advanced problems.

We focused what kinds of transition occurred in each stage. It was also observed that many participants reset their thoughts. It was an action observed when they tried to switch their thought of the time to another. We regarded their thoughts between resets as a cluster of thoughts. We examine what kinds of transition were brought about in the cluster of thoughts. In the course of examination, such cases were observed that two to up to five modes were transited in a single cluster of thoughts. Probability of transition occurrence is shown in Tables 1 and 2. Transition is shown from the first line to lower lines in sequence of occurrence.

Table 1. Mode transition with GUI
Full size table
Table 2. Mode transition with TUI
Full size table

6.2 Initial Stage

In GUI, it was the most frequent cases that transited from Opportunistic to Tactical control mode followed by cases that transited from Tactical to Tactical control mode. From a perspective of mode transition from Opportunistic to Tactical control mode, it is understood that problems are solved by setting sub-goals until a correct answer is obtained while considering how a logic circuit is created in order to achieve the sub-goals. We are also understand that steady measures for new sub-goal were not figured out immediately after resetting in many cases. From the fact that it is the most common case that transits from Tactical to Tactical control mode, however, we are also understand that procedures until sub-goals are achieved have been figured out in many cases even immediately after resetting. In consideration of the fact that this type of transition is frequent, it is believed that they try to progress to a stage that a correct answer is obtained by accumulating achievement of sub-goals. However, procedures to achieve sub-goals in an accumulative manner have not been clarified in some cases. From this fact, it is also understood that GUI has got to a correct answer by accumulating achievement of sub-goals at the initial stage. It is characteristic that Explorative control mode did not occur and that transition of three or more modes was not observed in any case. For the reason why Explorative control mode did not occur, it is believed that any thought with a challenging spirit to try everything possible in an explorative manner was not induced because such thought of foreseeing was dominant due to accumulated achievement of sub-goals observed in many cases. In addition, there was no case that transited three or more modes because there was not motivation to transit through various modes because of the foreseeing thought.

In TUI, transition from Tactical to Tactical control mode was most frequently observed. Even though transition from Opportunistic to Tactical control mode as frequently occurred in GUI was also observed frequently, the percentage was at a level less than 20%. From the fact that transition from Tactical to Tactical control mode was frequently observed, we are also understand that they tried to get to a correct answer by repeated accumulation of sub-goals at the initial stage like the GUI. Different from GUI, however, transitions from Explorative to Explorative control mode as well as from Explorative to Tactical control mode were also observed frequently. Therefore, we are also understand that they have tried to find out clues by creating a logic circuit as a trial without setting a sub-goal. In case of a transition from Explorative to Tactical control mode, in particular, it is believed that they transited to Tactical control mode by a clarified sub-goal because clues were found in Explorative control mode. In addition, there were cases that transited through three or more modes from the initial stage in TUI, and transition of Opportunistic, Explorative, and Tactical control modes were often observed. While they often got to a correct answer by accumulating achievement of set sub-goals similarly to GUI, such actions were also observed that they got to a correct answer by taking different measures such as creation of a logic circuit as a trial.

From the above, we are understand that subjects tried to get to a correct answer by accumulating achievement of sub-goals in many cases of GUI and TUI, but such different approach to create a logic circuit as a trial was observed in TUI alone.

6.3 Later Stage

In GUI, transition from Opportunistic to Opportunistic as well as from Opportunistic to Tactical control mode became to be observed frequently. In this case, it is believed that Opportunistic control mode was repeated being unable to clarify measures for achieving sub-goals even though subjects tried to solve the problem by setting a sub-goal similarly to the procedures at the initial stage. It is believed as one of the causes that whether measures in hand is appropriate or not has become to be checked by GUI frequently before judging it in mind since they have got familiar with use of GUI recognizing that it has become easier to check it in that way than solving problems by foreseeing situations. Therefore, it is believed that sub-goals were achieved at the later stage by checking whether measures in hand was appropriate or not for the sub-goals by GUI rather than considering it in mind. Even though it has increased to transit to Opportunistic control mode due to increased use of GUI, it is believed that the fundamental concept to get to a correct answer by accumulated sub-goal achievements has not changed. In consideration the fact that transition to Explorative control mode has been seldom observed in spite of increased transition by three or more modes, it is believed that accumulation of sub-goal achievements is repeated with almost no challenge as a trial similarly to that at the initial stage.

Transition from Explorative to Explorative control mode was observed most frequently in TUI. We are understand that subjects create various circuit diagrams trying to find out clues without setting a goal. It is believed that subjects have become to use a method for seeking for clues such as consideration based on a circuit diagram created as a trial rather than adopting a thought to set up sub-goals for problems with increased difficulty. It is also understood that creation of a circuit diagram as a trial is not a simulation of a result but an action to create a circuit diagram dependent on TUI just like in GUI.

From the above, it is proved that results of the circuit diagram are checked promptly by using interface at later stage in cases of both GUI and TUI. Since different approaches were observed in TUI at later stage while similar actions were often observed at early stage, it has been proved that subjects became easy to be affected by use of each interface as they got used to the interface. Therefore, it has been revealed that GUI tries to get to a correct answer being dependent on interface while accumulating sub-goal achievements and that TUI tries to get to a correct answer being dependent on interface while seeking for clues to get to a correct answer.

6.4 Practical Stage

In GUI, transition from Opportunistic to Opportunistic and from Opportunistic to Tactical control mode, as well as transition through Opportunistic, Tactical to Opportunistic control mode has increased. In the transition through Opportunistic, Tactical and back to Opportunistic control mode, even though measures for sub-goals had been clarified, there were wrong and unable to be clarified eventually, and therefore it is believed that it became difficult to have explicit measures for problems required for application. Further, when mode transition ended up in Opportunistic control mode, the next mode transition started with Opportunistic control mode in many cases since later stage. It is believed that other procedures were performed by setting a new sub-goal because any clear procedure for the previous sub-goal was found but it is understood that measures for the new sub-goal are also not clear. In addition, as mode transition to occur Explorative control mode was seldom observed, it is understood that sub-goal had been set up continuously in mind without seeking for clues.

In TUI, we are understand that patterns to transit two modes appeared evenly except for a pattern of transition from Tactical to Explorative control mode. In particular, patterns of transition to start with Opportunistic and Explorative control mode were observed frequently. In other words, it is believed that various solutions such as a method to seek for clues and to accumulate sub-goals had been performed.

From the above, it has been proved that GUI tries to get to a correct answer also in application problems by accumulating sub-goal achievements and that TUI tries to get to a correct answer by combining various methods such as to seek for clues and to accumulate sub-goals.

7 Conclusion

In order to verify a learning process at the time when TUI and GUI were used, comparison experiments were performed for the interfaces. An analysis was performed on actions of the participants during the experiments in combination with an analysis based on COCOM.

In GUI, the participants got to a correct answer at initial stage by achieving a sub-goal with transition from Tactical to Tactical control mode which was frequently observed. At later stage for application, transition from Opportunistic to Tactical as well as from Opportunistic to Opportunistic control mode increased. Even though mode transition changed, such tendency did not change that tries to get to a correct answer by setting a sub-goal and accumulating the achievements. In other words, subjects tried to get to a correct answer by the same methods in case of GUI both before and after they got familiar with the interface.

In TUI at initial stage, the participants got to a correct answer by also achieving a sub-goal with transition from Tactical to Tactical control mode which was frequently observed like a case of GUI. At later stage when a pattern to start with explorative control mode increased, however, they became in frequently take a method to seek for clues in addition to a method to accumulate sub-goal achievements. At a stage of practical stage, it was proved that they tried to get to a correct answer by taking various measures with various transition occurred evenly. It was proved that they took different methods between before and after they got used to the interface.

8 Application

In consideration of the fact that an intention to get to a correct answer is induced by using GUI to set a subgoal until getting to a correct answer, it is believed to be suitable for such problems that the final completion is achieved by accumulating subgoals just like programming.

In consideration of the fact that an intention to get to a correct answer is induced by using TUI to take various methods, it is believed to be suitable for problems aiming at thinking from various directions like active learning.