Using DynEd and ALC with low-level university freshmen

In this paper we share our experience implementing a CALL project with first year, lower level, non-English major, university students, using two CALL packages popular in Japan: DynEd and ALC. In addition to comparing the pros and cons of both, we will also discuss how the project was launched and how it evolved over time. Implementing commercial CALL software in a classroom setting involved a number of challenges, both technical and pedagogical, necessitating a variety of adaptations to suit the needs of both learners and the institution. Along with describing how changes were made, both in our methods of implementation and how the software was used, we also share our process of dealing with two key issues that arose: student motivation and learner assessment. We hope that by recounting our entire experience with this project, as well as sharing results from a series of student surveys, we will provide ideas for other teachers and faculty departments interested in introducing CALL projects with lower-level students.


Introduction
CALL is increasingly being used in Japanese universities in a variety of ways, from individual teacher efforts to larger, more campus-wide projects similar to the one discussed in this paper.Bingham and Larson (2005, pp. 39-40) attribute the increase in many of the larger projects to reasons ranging from staff cuts and shrinking budgets in Japanese universities to marketing strategies.They go on to say that many of these expensive CALL labs, built as "selling points for school brochures, " are then underused or misused with some simply becoming untended "independent study centers" rather than the active self-access centers their planners may have envisioned.The educational use of many of these expensive labs has not always been fully explored or exploited.
In their study of e-Learning in Japan, Ozkul and Aoki (2006) support this claim and examine its causes citing reasons from lack of teacher training and support, to cultural and social issues, such as Japan being a "high-context" culture, and to managerial issues.Another cause they cite is the nature of education in Japan with its Confucian value systems and hierarchical position of the teacher that lead to "little consideration to the pedagogy of e-learning, especially the learner centered approach and use of interactive tools such as discussion boards." Pagel and Reedy (2007) also echo some of these points, as well as those of Bingham and Larson, in their own study of a project similar to the one in this paper.They mention that whilst many universities invest in CALL labs and software, there are very few studies in the implementation and success of these projects.The implementation is often left to the individual teachers charged with teaching the classes even though those teachers took little or no part in the choice of the software.Johnson and Brine (2000, p. 266) in their study of CALL in Japan concluded by saying … in the design and development of effective CALL applications more than just the curriculum is important.The entire context of teaching, technology, learning and culture, and social interactions must be considered explicitly and the complex interactions of these various components have to be incorporated into instructional design.
With this in mind, we feel it would be beneficial to share the details of one such institutional CALL program in Japan from the perspective of some the contract teachers charged with implementing and facilitating it.The project involved the use of two popular CALL software packages, DynEd and ALC, in newly built, highly resourced computer labs with lower-level, non-English majors as part of a compulsory first-year course in English oral communication.In addition to describing how the project came about and was structured, we will also provide a comparison of DynEd and ALC, both of which are commonly used CALL software packages in Japanese universities.Finally, we will discuss the main issues and challenges we faced when implementing the program for the first time, sharing some of our insights along the way.We hope that by doing so, more light will be shed upon the manner in which CALL programs arise in Japanese educational institutions and how they are implemented and sustained, particularly with regard to lower level students.

CALL for Lower Level Students
Literature on teaching CALL with lower level students is difficult to find, in contrast to the wealth of studies with intermediate and much higher-level students.This lack of balance is often reflected at CALL conferences, where a majority of presentations and papers pertain to projects involving English majors, and/or students of a higher linguistic or motivational level.This is particularly surprising in Japan, where the great bulk of students studying English at the university are non-English majors in their first and second years of compulsory English study, many of whom, despite 6 years or more of previous study, still tend to be at the false-beginner/ pre-intermediate stage of English and who often show little motivation for learning more.
A common theme in some of the papers found on teaching English to low-level learners from outside of Japan was an increase in student motivation with the use of CALL.Al Jarf (2005), in her study from Saudi Arabia comparing two classes studying the same English grammar course -one using CALL and one not using CALL at all -reported "heightened motivation" with the CALL class.She reports students found features offered by CALL, such as online exercises with instant feedback, were useful for their study and that the students felt that CALL helped create "a warm climate between the students and instructor and among the students themselves." Berzosa and Rokowski (2000) in their CALL study from Spain similarly found motivational benefits and that the students appreciated the features that CALL, through computers and the Internet, offered them for their study.On motivation they say that it was "the most outstanding characteristic of the new CALL practitioners.Both integrative and instrumental motivation have increased creating a favorable atmosphere for language learning acquisition." Alias and Hussin (2002) in their e-Learning study from Malaysia dealing with writing and the Internet found that motivation of their students was much higher in the end of the program despite the students' generally having a "negative attitude" to writing in English at the beginning of the course.Students' confidence, low at the beginning of the course, built up though the course with accompanying increase in motivation.
However, studies in Japan with lower-level students similar to those in this project made little-to-no mention of motivational benefits.Ito (2006) in his study using CALL with non-English major first year students (similar to those in this study) at Iwate prefectural University found that students felt that the CALL course required them to work harder than in traditional face-to-face classes and made no mention of any motivational benefits.Redfield and Campbell (2005), in their study from Osaka comparing hybrid and self-access use of CALL, found better results with the self-access students but surmised this could be due to these students being forced to spend more actual time on the computer than the hybrid students.Pagel and Reedy (2007) in their study of a university wide project with ALC report on the use of bonus points as a way to encourage their students to participate in the use of CALL with at least 10% bonus points being necessary.Our observation of students and their motivational issues in this study found results more similar to the above-mentioned Japanese studies than to the overseas studies.One reason for this could rest with the general language learning motivation of the non-English major students, as Pagel and Reedy found with their first-year students, citing a lack of motivation toward learning English as a possible cause of failure for their initial project.
Concerning the use of DynEd in a CALL project at a Japanese university, O' Connor and Gatton (2004) found considerable success in the use of the speech recognition capabilities of the the DynEd software, suggesting an increase in confidence and ability of the language learners involved."These are students speaking, listening and responding.This is language being acquired through a wealth of busy, noisy, productive interactions." However, neither Redfield and Campbell, using Side by Side Interactive, nor Pagel and Reedy, using ALC, made any mention of specific findings on the benefit of the particular software involved, instead concentrating on issues related to the CALL project as a whole.It is our intention, therefore, to include some discussion of our observations on the actual software used, along with our observations on the entire CALL project, when describing our experiences below.

The KSU Course
In 2006 Kyoto Sangyo University (KSU) implemented a CALL program as part of their firstyear, non-English-major, compulsory English Oral Communication course.KSU places these students into five levels on the basis of their own in-house placement test.The lowest two levels were chosen for the CALL program because it was felt that use of CALL could help improve their most basic English skills.Many of these students are at the false beginner or elementary level and need basic reinforcement of grammar, vocabulary, reading, and listening.It was thought that use of tutorial-style CALL programs once a week could supplement lessons from the classroom teacher in traditional face-to-face sessions, also held once a week.Each year the course consists of about 500-550 students from a total population of around 3,000.
The English Oral Communication course at KSU for the lower two levels of students requires students to participate in two 90-minute classes per week: one class held in a brand new, specially designed, state-of-the-art computer room; the other a face-to-face (f2f) class in a traditional classroom setting.In most instances, students have the same teacher for both classes.However, due to economic and scheduling issues, the weekly CALL class combined two classes of up to 35 students each simultaneously, while the same teacher taught the f2f classes separately, essentially cutting what would have been a total of four classes in a week to three, thus requiring fewer teachers and CALL classrooms.The integration of content between the CALL and f2f classes was not required or specifically fostered at the beginning of the project; however, as the course developed, individual teachers saw the need to do so and began to integrate content and materials from the two classes.
The CALL component of the course involved the use of two software packages: Dyned and ALC, both self-paced, tutorial-style CALL programs.All classes also had their own Moodle setup for extra support, which is standard for all classes at Kyoto Sangyo University (Section 5 contains more information on what and how the software was used).In the f2f component teachers conducted standard university oral communication classes, most of which were carried out using modern communicative language teaching practices.Teachers were free to choose from a list of texts put together by a committee of teachers that compiles the list yearly.Assessment for the course was an amalgamation of 3 grades: one for the CALL component (35%), one for the f2f component (35%), and one for a standard proficiency test (30%), given to all first year students at the end of each semester.Teachers were free to calculate the f2f component grade in any manner they liked, usually using a combination of attendance, homework, class tests, and assignments.
The teachers of the course were native English speakers on contract with KSU and generally with some interest, experience and expertise in teaching CALL.Whilst not involved in the creation of the CALL project, a CALL committee was established for improving and fine tuning the project as it progressed including teacher training and support.The three authors of this paper were on the first CALL committee at the start of the program and two have continued with this committee in subsequent years.

Implementation
The implementation process at KSU began with the university hiring contract teachers who had some interest and expertise in using CALL from 2005.The two software applications were chosen by the full-time members of the committee supervising the English Program at the university (not by the contract teachers) after a search of the various software packages available.Two state-of-the-art computers rooms were set-up purposely for the program with the idea of being able to hold large classes of 80 students to accommodate the amalgamation of two f2f classes for each CALL class.Towards the end of that year a CALL Committee from the contract teachers was created to make recommendations on how to practically implement the new course.In March 2006 training sessions were held, both from the software manufacturers on the use of their products, and in-house on how to implement the new program based on the recommendations of the initial CALL committee.The course began at the start of the new academic year in April 2006 and, as one would expect, a number of issues arose concerning how best to conduct the course, how to deal with technical and pedagogical challenges, how to motivate students to stay focused, and how best to assess student work.Dealing with these issues and adapting the course to overcome the challenges is largely what occupied the CALL committee in the first two years of the program' s inception.
The KSU Software -DynEd and ALC

Overview of the software
The two primary software applications used in the Oral Communications CALL courses at KSU are DynEd' s New Dynamic English and ALC' s Basic English Listening Course.Japanese language support was an important factor in the selection process, as it was to be used with low-level students.Indeed, both DynEd and ALC provide Japanese support for students to differing degrees.
DynEd offers a suite of nearly a dozen different courses for schools and universities aimed at a range of ages and proficiency levels.At KSU, students mainly use New Dynamic English (NDE) from DynEd' s suite, although the other courses are also available for use.NDE was deemed the most appropriate for the level and interests of the students after a team of teachers evaluated the various courses available.
New Dynamic English is a listening/speaking-based interactive multimedia course that includes colorful graphics, animation, video, and speech-recognition activities.NDE consists of eight units, or Modules, that range in level from beginning to advanced.Students at KSU typically begin with Module 1 or 2 and progress through to Module 3 or 4 by the end of the year.NDE also includes native-language support in the form of audio translations and a Japanese version of the Intelligent Tutor, which tracks and evaluates all study activities and offers specific study recommendations to students.
In a typical NDE study session, students listen to a story or conversation, answer periodic questions about it, and then complete a variety of exercises to further develop their mastery of the grammar, vocabulary, and functions presented.Exercises include reordering words into sentences, fill-ins, dictations, and speech practice using the speech recognition system.Students are also encouraged through the Tutor to use the repeat button to listen to the sentences presented more than once and to record their own voices to compare them with the nativespeaker recordings.
One of the most useful features of NDE is the Records Manager, which allows teachers (and students) to track student progress, including time spent, percentage complete, and scores on the exercises and tests.The Records Manager is not only useful as an assessment tool, but it also helps to motivate students since they are able to follow their own progress as well.
ALC' s Basic English Listening Course is part of a suite of applications called NetAcademy produced by a Japanese publishing company.ALC was initially chosen mainly because of its Japanese language support, which is integral to the lessons and appreciated by many of the students.However, ALC offers little English support beyond the content of the lessons themselves.The listening course was chosen for the Oral Communications course because it seemed to fit best with the theme of the course (other applications, such as the PowerWords vocabulary course, are used in the Reading Skills courses at KSU).
The course consists of 30 listening units on various themes, from telephone calls and conversations to speeches.In each unit, students listen to a short passage, answer three comprehension questions about it, and then review the passage with the English transcript and a Japanese translation.In addition, the program offers a personal glossary and usage notes in Japanese to help with studying the vocabulary and grammar.The software also allows students to replay the recordings at various speeds for "speed listening" training.
ALC also provides a records management system, but it is not nearly as extensive or useful as DynEd' s.The main information provided is the time spent on each unit and the student' s score on the three-question comprehension quiz.

Comparison of ALC and DynEd
In the following section, we provide a detailed comparison of the DynEd and ALC programs used at KSU.To carry out this comparison, we chose to use Thorn' s (1995) 6-point criteria for evaluating software: Ease of use and navigation.1.
Cognitive load.

Ease of use and navigation
The first point is ease of use and navigation.According to Thorn, "A program needs to be very simple in its interface, so that learners don't have to compete between learning English and learning how the program works." In DynEd, a Control Bar appears consistently at the bottom of every window.Controls include the following: The basic operation of both programs is straightforward and easy to learn.DynEd' s controls are more consistent than ALC' s, which makes it marginally easier to learn at first.DynEd has more options, however, and the operation and significance of some of its more advanced features (e.g.voice record and playback) are difficult to grasp initially.The purpose of some of ALC' s features such as the checkboxes may also be unclear to a new user.

Cognitive load
The second criterion is cognitive load.Thorn states, "Users need to cope with the program' s content, its structure and the response options.The program needs to be intuitive, so that it works the way you might expect it to work." DynEd' s content and scope are more extensive than ALC' s, so it may not be completely fair to compare them on this point.ALC' s Listening Course focuses on listening and, to some extent, vocabulary development, while DynEd additionally includes speaking activities for developing pronunciation and fluency.DynEd starts from a more basic level of English than ALC, so the linguistic cognitive load of DynEd is probably more appropriate for low-level learners.
In terms of intuitiveness, both programs seem uncomplicated at first, but as mentioned before, both include features that are not immediately intuitive.DynEd' s Voice Record/Playback and Repeat buttons are integral to the program and its evaluation system, but a learner can move through the lessons without ever touching them.At no point does the program require the learner to use these features.Learners must be taught and constantly reminded of the importance of these features.This is especially true of low-level university students in a required English course, who risk ending up doing the bare minimum without consistent encouragement and support from the teacher.
Similarly, ALC' s checkboxes require initiative on the part of the learners.The checkboxes are there so that learners can mark portions of the listening passage that they couldn't hear.Most students, however, don't see the point of this option and it seems more of a distraction.
Overall, DynEd has a higher learning curve initially, and therefore training is necessary.DynEd provides video tutorials in Japanese on its website for this reason.ALC can mostly be figured out without special training, as it is fairly simple.

Knowledge space and information presentation
Thorn' s third point is knowledge space and information presentation.He says, "The concepts about learning English as a second language need to represent existing methodologies, where these methodologies have been proven successful.
The importance of practice and repetition in learning English as a second language is a key element in the design of both programs.In order for learners to see progress in their DynEd Study Records, they must review the lessons several times and use the Repeat button often while studying each lesson.As mentioned before, learners must be trained to do this.DynEd' s Intelligent Tutor reminds students while recommending the best ways to improve their study, while a review of the Study Records can at least provide motivation through ego involvement.
In ALC, repetition is built into the Steps of each unit.At each Step, learners are encouraged to listen to the passage again.Learners can also repeat individual sections of the listening passage by clicking on the numbered buttons.As in DynEd, this requires some initiative on the part of the learners, but ALC does not have a system like DynEd' s Study Records to motivate them.It is possible for learners to race through each unit and believe they have finished studying without repeating anything.
Both programs provide text support in English and Japanese, but over-reliance on text can hinder the development of listening and speaking skills.DynEd discourages the use of text support by deducting from the learners' study scores if they use the text buttons too much, but ALC makes it all too easy for the learners to depend on text and translation.This makes ALC very attractive to many students, but it can be detrimental to their English proficiency development.
We found the content of ALC' s listening course to be less than desirable.The strange choice of vocabulary and expressions (e.g."you go girl"), the inappropriate levels of some of it, the weirdness of the listening passages themselves, and the complete lack of continuity or any connection between them made for an unsatisfying experience.
Overall, DynEd seems to represent existing methodologies better than ALC.Neither application in our opinion, unfortunately, offers student centered content, as little of it is relevant to the average interests of young Japanese adults.

Media integration
The fourth point is media integration, which means, "The multimedia needs to be combined to produce an effective whole." DynEd' s NDE consists of audio, video, flash animations, record and listen buttons, and voice recognition technology, all of it integrated to give the user a more engaging, high context experience.
ALC is just audio and text, nothing else.Audio can be manipulated in two ways: speed and listening to selected portions of the conversation.Also, the text is integrated with the listening, so that it appears in tandem with the conversation as it is spoken.
Overall, DynEd integrates media better than ALC.

Aesthetics
The fifth criterion is aesthetics.According to Thorn, "There needs to be a sense a beauty in the graphical interface.This adds to an effective learning environment.Neither program is unattractive, but DynEd is visually more interesting than ALC.DynEd has a variety of colorful illustrations, animations and video, while ALC offers a stark text-and-button interface.ALC could benefit from the addition of illustrations related to the listening passages, which would not only make the program more appealing, but would also provide contextual cues to aid the learners.DynEd is far from perfect, though.The illustrations, as functional as they may be, look somewhat dated at times, and the animations can be overly simple and unattractive.The videos, on the other hand, are well done.
Overall DynEd is more aesthetically pleasing than ALC.

Overall functionality
The final point is overall functionality.Thorn explains, "The program needs to provide learning in a way that users expect it to.Students need to go away from it having learned something.
No CALL software that we know of is capable of making students learn if they have no desire to and make minimal effort.If students take initiative and use the programs effectively, however, it is possible to learn something from both DynEd and ALC.
Of the two applications, DynEd gives students a better sense of having learned something, with its records manager (showing completion rates), study score (indicating good study habits), mastery test scores (implying a mastery of the material), and placement tests (which can show improvement in overall level if administered at the beginning and the end of the course).Furthermore, DynEd provides a wealth of assessment information for teachers via the Records Manager, including time spent, completion percentages, scores on individual tasks, Study Scores, and Mastery Test scores.The Records Manager also provides detailed data on information such as the number of times the student used each of the control buttons.We found that this detailed data was more than we needed for our assessment purposes, and decided not to refer to them when assigning grades.Indeed, Bingham and Larson (2006) found DynEd to be an effective tool for improving their students' overall English abilities.
Having said that, some of the functional difficulties noticed when using DynEd involved occasional crashes and subsequent loss of minor amounts of data for that particular student, difficulty with the students successfully using the voice recognition technology (a feature whose sensitivity can be adjusted by the administrator), and teachers never reaching a full understanding of how the application calculates completion percentages and study scores.
As for ALC, with only a crude records manager showing time of study and the three-question quiz result, there is no real indicator of progress.This could leave students feeling like they came away with little, and we feel it is one of ALC' s main weaknesses.Although ALC provides "Quiz Time" as one of the Steps in each unit, it is the second step (out of five), and serves only as an initial comprehension check.The only other assessment parameter available to teachers is the amount of time spent on each lesson, which is essentially worthless since students could run up the time without actually doing anything.
Since ALC does not provide any other means of evaluating the students' progress, we were forced to make our own listening tasks for assessment purposes.In these Moodle-based quizzes, students listened to entirely new conversations that we created based on the vocabulary in the ALC listening passages, followed by ten multiple-choice and true/false questions.These quizzes gave us a fair idea of which students had actually tried to learn the material.We did observe, not surprisingly, a correlation between diligent, focused study of ALC and success on the listening quizzes.

Motivation
Cultivating and maintaining student interest in the CALL classroom has been a constant challenge for KSU teachers each semester.It is difficult enough for anyone to concentrate on a given task for 90 minutes, particularly on a compulsory subject of little interest for these students, like English.To compound matters, both pieces of software were low in any sort of entertainment or gaming values -attributes which have been suggested to increase motivation (Meyer and Sørensen, 2007) and foster deep learning (Gee, 2005).Furthermore, the ubiquitous presence of cell phones, the Internet, email, Mixi, and YouTube provided constant distractions for some students, while others found the whirr of 70 computer fans conducive to a much needed nap.
For these reasons teachers implemented a number of strategies in the classroom to help students maintain focus on their study.First, teachers created and gave students a list of target learning goals for the semester, which included total number of units/modules and specific minimum completion percentages and mastery test scores for each.Grades for the semester pertaining to CALL study depended on successful completion of these learning goals -a strong extrinsic motivational factor.However, as one teacher pointed out, "The motivation is completion of the module to receive the grade.They are completing the software' s targets.They are not working on learning English." Second, teachers allowed students to study at their own pace while working toward the requirements for the semester.In most cases, students had freedom to choose how to best use their time in and out of the CALL room.Students enjoyed having this freedom, even though the challenge of meeting the semester learning goals limited the amount of "down time" or "slow study" in which a student could indulge.Indeed, student survey results suggest support for this point (see section 9).
Finally, teachers found that establishing and maintaining an active, physical presence by roaming around the room and interacting personally with the students helped to stem potential distractions while giving advice, help, and encouragement to students when needed.And as we will see in section 7 below, additional measures were eventually taken to restructure the CALL program in order to integrate study in the CALL classroom with learner-centered content from the weekly face-to-face meetings in the oral communication classroom.

Assessment
How best to assess student work in the CALL room was another major issue to contend with.As we saw above, DynEd provides teachers with a flood of data for each student, more than what is needed to arrive at a meaningful grade for the semester' s work in a reasonable amount of time.This prompted teachers to discuss which of the DynEd data was most useful for assessment purposes.It was initially decided that time of study, completion percentage for each module, and mastery test scores would be the main assessment factors.The CALL committee then generated a set of student learning goals for the semester based on these criteria, which were then recommended to CALL teachers, who were given the option of altering these for their respective classes.It soon became apparent that time of study was an inappropriate assessment factor, as students simply opened up both applications simultaneously and let them run freely whether the students were actually paying attention or not.In subsequent semesters, only completion percentage (for DynEd) and mastery tests were the official assessment factors, thus rendering 'time of study' valid again as a reflection of student effort.
While DynEd provided bountiful data, the ALC Listening Course generated almost nothing useful -time of study being the main item.As mentioned earlier, this led the CALL committee to develop a series of listening tests based on ALC content to compensate for these assessment deficiencies.Audio scripts for each test were designed to reflect the main grammar patterns and vocabulary words for sets of three to four ALC lessons.Recording and editing was done with Audacity on an iBook.Each mp3 recording was embedded in a unique Moodle quiz activity and delivered with a series of corresponding comprehension questions, resulting in a total of 6 tests covering 20 lessons.It was observed that students who performed well on these quizzes had used ALC diligently and consistently from week-to-week.
At the end of each semester an Excel-based grading spreadsheet was sent out to CALL teachers, allowing them to calculate grades for their students by entering the appropriate values from the DynEd records manager and ALC mastery test scores from the course Moodle, in addition to other potential grading factors such as other work, participation, attendance, etc.This grading worksheet was also designed to allow individual teachers to alter the weight percentage given to any particular value.The default percentages for the Spring 2008 semester were as follows: NDE Completion Percentage: 35% NDE Mastery Tests: 25% ALC Tests: 20% Other Work: 10% Attendance and Participation: 10% Use of the grading worksheet was optional, and only about half of the CALL teachers used it each semester.Again, the CALL grade accounted for 35% of the semester total.

Adaptation
Being an upstart CALL program, both the software and the teaching situation was new to all teachers involved.Thus the process of constant adaptation to arising challenges in the classroom was the third key issue in this project.As was previously mentioned, mastery tests for ALC had to be created, as well as time of study being dropped from assessment criteria.Another adaptation made was a move away from attempting to work with data exported from each application to Excel spreadsheets, massaging it, and using complicated formulas to render it meaningful for assessment.Teachers found it easier to simply eyeball the data in the records manager and input it all into the grading worksheet by hand.This unfortunate time consuming step is one of the current challenges still facing CALL teachers in the program.
In addition to assessment and software related adaptations, there were also technical difficulties to overcome, such as malfunctioning hardware and software, as well as a rapidly decreasing supply of headsets from the university-provided stock.Initially there was no specific system for dealing with technical difficulties; each teacher handled them differently, some reporting them to the CALL committee, while others contacted tech support directly.Headsets were eventually secured firmly to each computer with plastic straps, thus stemming the tide of their mysterious disappearance.This solution, however, led to another problem wherein a broken headset rendered the computer to which it was attached unusable until a new headset was eventually attached.By the second year of the program, a paper-based system of identifying problematic computers and reporting them to administration was developed, which made the problem solving process much more efficient, keeping the large majority of the computers fully operational from week to week.
Finally, once the adaptations described above were made, teachers began to address one of the obvious weaknesses of the CALL program: a total separation in both content and practice between what happens in the f2f classroom and what was being done in the CALL classroom.Indeed, student motivation is increased when attempts at CALL are integrated into the overall goals and structure of the course (Warschauer, 1996), so measures should be taken to link computer study with other coursework so that relevancy and educational benefit is achieved (Stepp-Greany, 2002).
Therefore, a move toward better integration between classes was called for, especially in light of the motivational issues discussed earlier.It was thought that since f2f classes contained much more learner-centered topics of discussion, that finding a way to integrate what goes on there into the CALL classroom would be more stimulating for students and lead to a better overall quality learning experience.The initial attempts at better integration included Moodlebased discussion and chat activities, the incorporation of digital materials from the textbook, web-based research and discussion activities, and the delivery of quizzes via Moodle related to f2f class material.

Student Surveys
Three separate intakes of students -from January 2007 to July 2008 -were surveyed to shed light on their perceptions of the software and their experiences with the CALL program at KSU.Although the survey was not anonymous, students were told that the completion of the survey would have nothing to do with their grades and that their answers would not be made available to their teachers until after the grading period was finished.The survey contained both quantitative and qualitative items, and was administered to students using the quiz activity feature on the Moodle course.When measuring student preferences quantitatively, a 6-point Likert scale was used in order to deter students from choosing the middle value -forcing a choice to one side or the other.
Overall, students felt that the use of the CALL software was significantly improving their listening ability, and to a lesser extent their pronunciation (thanks to voice recognition technology of DynEd), while they deemed their grammar and vocabulary to see only slight improvements, to no improvements at all (see Table 1).A look at the qualitative feedback showed that students overwhelmingly enjoyed being able to study at their own pace, something that other CALL teachers also noticed in their interactions with students.As one teacher said about the software, "It seems to encourage time on task for students with low motivation and those lacking in confidence." Negative responses were varied, but some included a preference for face-to-face classes, difficulty understanding the listening passages, feeling sleepy in class, and feeling bored with the content.Perhaps the most salient result noticed in the three surveys was the recent increase in overall satisfaction with the CALL program in terms of student perception of both how enjoyable and how educational the use of the software was (see Figures 1 and 2).We attribute this increase in satisfaction to the recent attempts at better integration between f2f and CALL classes.
Interestingly, the most negative comments received about the chosen software were not from students, but rather from CALL teachers themselves, describing the use of both applications as "tedious, " "tiresome, " "demoralizing, " and "clunky." As mentioned before, CALL teachers overwhelmingly agreed that a move toward better integration between f2f and CALL classes was necessary if the program was to improve.How educational is CALL?
of problems?How can it be ensured that those problems are attended to in a timely and efficient fashion?Think about classroom management.How will attendance be taken in large classes?How • can learner needs be best supported during class?How can students, who are prone to distraction (Internet surfing, cell phone messaging) and dozing off, be kept focused?We found it helpful to establish a physical presence in the classroom by walking around and interacting personally with students.This made our support and advice more accessible than it would have been otherwise to students shy about asking for help.Think about assessment when choosing software.How easy or difficult will it be to • assess?How will assessment data be gathered and massaged?How time consuming is this going to be? Are teachers going to do it individually, or will the approach be systematic?
We found the process of turning data into grades that accurately reflected the effort and success of each student to be a difficult and time consuming process.Think about the role the software might play in motivating students.Look at the "fun" • factor and at aspects that might stimulate ego involvement, like competitions, games, target goals, and progress measures.Also, what opportunities might there be for cooperation and outcome interdependence amongst learners?Think about integration potential with other English classes, other courses in the learner' s • major, or even integration with their personal daily lives.Integrating CALL activities with other areas of study and interest may increase relevancy and meaning, thereby tapping into intrinsic motivational factors.

Conclusion
As with the implementation of any new CALL program in a university setting, a variety of challenges will arise, necessitating adaptations along the way, both in terms of how software is put to use and how assessment is carried out.With lower level students, cultivating and maintaining motivation is particularly important -a task highly dependent on the input of teachers and how they structure and implement the program.Studying at one' s own pace is vital toward fostering student satisfaction with process, as is a certain measure of relevancy and continuity, not only with students' other English courses, but also with their daily lives.We hope that this paper has shed some light on the challenges faced when starting a new CALL program with lower level students, as well as providing some feedback on the pros and cons of the software used, from DynEd and ALC.
Australia and Thailand.He has a long interest in CALL, especially in the areas of teacher training and making CALL practical for all teachers.
Aaron P. Campbell is a foreign lecturer in EFL at Kyoto Sangyo University, where he teaches oral communication, CALL, and Internet communication courses.He also serves on the CALL committee at Sangyo, and has been involved in the CALL program since its inception in 2006.
York Weatherford is an instructor at Kyoto Sangyo University.He has been teaching at universities in Japan for over thirteen years.His research interests range from CALL to childhood bilingualism.

Figure 1 .
Figure 1.Results from 3 separate intakes of students.

Figure 2 .
Figure 2. Results from three separate intakes of students.Practical Recommendations for Implementing a New CALL Program After reflecting on our experiences over the last three years with implementing and striving to improve the CALL program at KSU, we came up with the following recommendations, which we think might be of use to other practitioners thinking of starting their own CALL program: Designate a CALL coordinator or a CALL committee to oversee and develop the • program you set up.This will insure accountability measures and hopefully increase the quality of the program over time.It is best to do this before the program actually gets implemented, so that the committee can test the software, develop training materials, and devise the best strategies for integrating the CALL program into the department' s curriculum.Have a training program in place for new CALL teachers that involves as many face-to-• face sessions as possible, especially before and during the first few weeks of class.It is helpful to aggregate all training and support documents in a Course Management System, like Moodle or Blackboard, to provide convenient access and an easily searchable archive.Encourage new CALL teachers to post their questions, problems, and suggestions in a • discussion forum so as to make these issues available to all CALL teachers and to provide an archive of discussion, which can be searched and linked to in the future.An FAQ can be developed from these discussions over time if desired.Plan ahead for technical problems and develop a system of dealing with them, both • during and after the fact.For large classes, having a technical/teaching assistant in the first few weeks of class is helpful.At Kyoto Sangyo University we asked teachers who weren't teaching classes to support CALL teachers during the first week.This type of support is particularly important for novice CALL teachers.How can technical support be informed

Table 1 :
July 2008 survey (n=362)How well did CALL improve your English in these areas?