Design and Construction of a Virtual Environment
for Japanese Language Instruction
by Howard Rose

Chapter IV
Research Plan

The proposed research compares the relative efficacy of three instructional methods: virtual reality instruction with the Zengo Sayu system (VR), teaching with real-world objects (RW), and conventional text-based instruction (TB). This research attempts to answer the following research questions: 1) Can students using Zengo Sayu learn Japanese language? 2) How does Zengo Sayu compare to other teaching methods in terms of learning gains? 3) Does the Zengo Sayu approach have a positive affect of students' motivation and attitude toward learning Japanese?

Findings of this study will contribute to understanding the role of various learning factors in foreign language pedagogy such as: the merits of student control in the learning process, the importance of full body activity in learning language, the usefulness of a silent period and the effect of various instructional treatments on student attitudes and anxiety levels. Also, developing a greater understanding of VR's potential as a learning tool would be relevant to current research in human factors and human-computer interaction.

Research Hypotheses

The primary hypothesis for Phase I is that treatment with Zengo Sayu will yield significantly better achievement on listening comprehension and oral production tests than with TB, but will not differ significantly from the real-world RW instruction. This assumption is based on literature which supports the use of a silent period (Mangubhai, 1991; Atherton, 1993; Gary, 1975; Winitz & Reeds, 1973), the advantages of concrete, experience-based approaches for language learners (Baltra, 1992; Terrell, 1986) and the success of the Total Physical Response approach for early language learners ( Asher et al., 1974; Asher, 1966; Kunihira & Asher, 1965).

Hypotheses

1) Adult beginning language learners receiving instruction in virtual reality with Zengo Sayu will demonstrate significantly better listening comprehension achievement as measured by the Test of Listening Comprehension (TLC) compared to students receiving TB treatment. The VR and RW treatment groups will not show significant differences in listening comprehension achievement.

2) Adult beginning language learners receiving instruction in virtual reality with Zengo Sayu will demonstrate significantly better oral production achievement as measured by the Test of Oral Production (TOP) compared to students receiving TB treatment. The VR and RW treatment groups will not show significant differences in oral production achievement.

3) Adult beginning language learners receiving instruction in virtual reality with Zengo Sayu will demonstrate significantly more positive attitudes toward learning Japanese as measured by the attitude survey compared to students receiving both the TB and RW treatments.

Methodology and Procedures

This study compares learning gains between subjects in three instructional treatments: virtual reality (VR), computer-based (TB), and real-world (RW). Total N for this study is 42, with n =14 for each of the three treatment groups. Each treatment group will be divided into two sections of seven subjects and assigned to an instructor. As a precaution against confounding effects of instructors' attitudes, instructors will be informed only that this is a study to compare three different teaching methods.

All three treatment groups will receive two, 90 minute treatments. All treatments cover the same content and are designed to present instruction and offer rehearsal as consistently as possible. Time of instruction and time of rehearsal will be consistent for all three treatment groups. Students will be instructed to limit their practice to class time and not to seek outside instruction or help after leaving class.

All treatments will cover the same language content including: five colors, two nouns, two verbs, and five prepositions used in the context of full descriptive sentences and commands. A detailed lesson plan of the Zengo Sayu treatment is included in Appendix I.

Variables

Table 4: Independent and Dependent Variables

Independent Variable Comparison of Japanese with 3 levels of instruction: TB/VR/RW.

Dependent Variables

Listening comprehension achievement measured on the TLC.
Oral performance achievement measured on the TOP.
Change in attitude toward Japanese language as measured by gain scores on the attitude survey.

It is important to note that each instructional mode has unique strengths and limitations. For this study, each instructional treatment is designed to optimize the particular strengths of that medium in order to obtain the utmost positive instructional effect. In other words, in cases where one mode offered an advantage over the other in terms of presentation or practice, for example, the choice was always made in favor of optimal instructional effect rather than impose artificial consistency across treatments. Table 5 shows some of the strengths and weaknesses of each treatment. Table 5 is intended to facilitate comparison and is by no means a comprehensive description of any of the teaching modes.

Table 5: Comparison of Three Instructional Modes

Mode Strengths Weaknesses

Real-world
(RW) Strong sense of physical and tactile manipulation grounds learning in first-person experience. Not self-paced (i.e. instructor guides lesson)

Students get direct feedback from instructor High degree of instructor attention and peer observation may be anxiety producing

High level of student and teacher interaction High costs for classroom materials, props and other teaching aids.

No interface to learn

Flexible to any situation or student question

Low technology costs

Student pair work and collaboration easily achieved.

Text-based
(TB) Instructor determines pace and course of learning for students based on his/her pedagogical experience. Less physically interactive then other two treatments

Student pair work and collaboration is easily achieved. Students are not free to set their own pace and course of learning.

Low technology costs More abstract form of instruction adds an extra, intellectual layer between experience and use of the language

Students are most familiar with this teaching approach. High costs for classroom materials, props and other teaching aids

Virtual Environment
(VE) Immersive Students unfamiliar with the hardware and software interfaces

Interactive environment Inflexible: must be preprogrammed

Self-paced Text is difficult to read in the VE.

VE can contain video, animated objects, and spatial sound Moving and manipulating the VE can be difficult for novice users.

Voice recognition allows for constant error checking of students' speech Long exposure to VR may cause physical discomfort.

Software is portable, can be delivered over networks and customized to fill specific needs. High cost of technology and software development

Walking, grasping and moving objects in the environment is similar to real-world experience. Virtual environments are initially costly to develop.

Potential for technology to breakdown

Mode	Strengths	Weaknesses
Real-world (RW)	Strong sense of physical and tactile manipulation grounds learning in first-person experience.	Not self-paced (i.e. instructor guides lesson)
	Students get direct feedback from instructor	High degree of instructor attention and peer observation may be anxiety producing
	High level of student and teacher interaction	High costs for classroom materials, props and other teaching aids.
	No interface to learn
	Flexible to any situation or student question
	Low technology costs
	Student pair work and collaboration easily achieved.
Text-based (TB)	Instructor determines pace and course of learning for students based on his/her pedagogical experience.	Less physically interactive then other two treatments
	Student pair work and collaboration is easily achieved.	Students are not free to set their own pace and course of learning.
	Low technology costs	More abstract form of instruction adds an extra, intellectual layer between experience and use of the language
	Students are most familiar with this teaching approach.	High costs for classroom materials, props and other teaching aids
Virtual Environment (VE)	Immersive	Students unfamiliar with the hardware and software interfaces
	Interactive environment	Inflexible: must be preprogrammed
	Self-paced	Text is difficult to read in the VE.
	VE can contain video, animated objects, and spatial sound	Moving and manipulating the VE can be difficult for novice users.
	Voice recognition allows for constant error checking of students' speech	Long exposure to VR may cause physical discomfort.
	Software is portable, can be delivered over networks and customized to fill specific needs.	High cost of technology and software development
	Walking, grasping and moving objects in the environment is similar to real-world experience.	Virtual environments are initially costly to develop.
		Potential for technology to breakdown

Instrumentation and Testing

Testing will include a single pretest, and two post-tests. Because all test subjects will have no prior Japanese language experience, there will be no initial pretest to measure Japanese ability. An attitude survey will be administered prior to treatment to measure subjects' initial levels of motivation and anxiety toward foreign language learning.

Post test measures will include the Test of Listening Comprehension (TLC), Test of Oral Production (TOP) and an exit survey to measure attitude and anxiety changes due to the treatment. Post tests will be administered two times: first directly following the conclusion of the treatment, then again two weeks later to test retention. Both post-tests will use the same test instruments: TLC, TOP and the attitude survey.

Specific procedures will be followed in order to maintain consistency in test administration and rater performance. The TLC will be administered using an audio tape. Oral Production Tests will be video taped and rated by a skilled Japanese teacher with no prior knowledge or exposure to the test. The order of subjects on the tape shown to the rater will be randomized to eliminate potentially biased associations between groups of students.

Anticipated Results

The anticipated outcome is that both the whole language and the VR treatments will prove more effective then the text-based grammar-translation treatment, thereby supporting the superiority of interactive, holistic instructional methods. It is anticipated that the VR treatment will equal or exceed the RW treatment in learning gains, though the positive effects of the VR treatment may be mitigated by a number of factors such as hardware interface problems, users' adverse reactions to VR and unfamiliarity with virtual interfaces.

This study is susceptible to a number of potential confounds which could influence final results. Some of these confounds include: subjects' predisposed attitudes toward technology, gender and age differences among subjects, and excessive variance between treatments due to the nature of the respective teaching methods. Perhaps the greatest potential confound is that all three treatments may prove equally effective under these test conditions. Given that this study uses class sizes which are relatively small and all the instructors will be experienced teachers, it is quite plausible that the subjects' innate capacity to learn in different modes will enable them to succeed equally in all three treatments.

Human Interface Technology Laboratory

Independent Variable	Comparison of Japanese with 3 levels of instruction: TB/VR/RW.
Dependent Variables	Listening comprehension achievement measured on the TLC. Oral performance achievement measured on the TOP. Change in attitude toward Japanese language as measured by gain scores on the attitude survey.

Design and Construction of a Virtual Environment for Japanese Language Instruction by Howard Rose

Chapter IV Research Plan