Please note that the software referred to in this paper bears no relation to Michio Arai's software of the same name. At the time this paper was written I was unaware of the existence of his software ("TypeIt"/"Type It"), and unwittingly used the same name for my own. My apologies for this unfortunate coincidence, of which I became aware only recently, thanks to a communication from Mr. Arai. (Jan. 25, 2000)
0. Outline of paper
1. Background to TypeIt
2. Integration into the course
3. Description of software and hardware
4. Looking back on TypeIt as homework
5. Analysis of TypeIt data: An example
6. Conclusion
*. E-mail to author
In this paper I will present a brief summary of my TypeIt software and how I have used it as a supplementary homework during the past year and a half in the Oral English classes I teach at Tokiwakai Junior College. After giving some background information on the motivation for assigning this kind of outside work to students, I will describe: how TypeIt was introduced to students and implemented as homework in the courses; the software functionality of TypeIt and the hardware available to students; and my own subjective impressions of its merits and shortcomings based on actual use. Although TypeIt was assigned for purposes other than the gathering of statistics, this paper will conclude with an example of how data collected with it might be used to provide a measure of students' improvement in basic typing skills.
I firmly believe that dictation of spoken English can be a useful tool at all levels of study in developing an ear for phonetic detail as well as improving general listening comprehension ability, particularly when it is done as homework from a cassette tape to allow for repeated listening at a student's leisure outside of class. For this reason, I frequently distribute a homework tape containing dialogs or readings to students at the beginning of the term and devote a portion of the class to questions and answers or discussion based on the content of an assigned dictation. This dictation may either be of the "cloze" type, with part or all of the words deleted and represented as underlined blanks on a printed handout, or "complete," with no prior clues whatsoever as to the form or content of the taped passage. Depending upon the focus of the class and the philosophy of the instructor, it may or may not be deemed appropriate to check all or even part of the dictation per se in class. After all, if the purpose of the dictation is to provide the basis for eliciting students' opinions on its content or, at a more basic level, to check whether they have grasped the idea of what it is about, it can be a significant drain on already-limited classroom time resources to actually go through, word by word, what may be a significant volume of English. When this is done, the traditional methods for doing it, together with some common problems associated with them, are:
- Writing out the words or sentences on the blackboard or a monitor and having students correct their mistakes and copy down missed words.
- [Some students can't keep up or make mistakes in transferring the answers to their own papers, so in the end they don't have correct English written down anyway.]
- Going through the answers orally, perhaps spelling out difficult words.
- [While this can be valuable additional listening practice, if for any reason a student misses a word, beyond a certain point there is "no going back" and slower students can become lost and frustrated.]
- Passing out a printed answer sheet.
- [In addition to the obvious use of yet more paper, less motivated students, now having the answers, will not feel any need to check their own answers against the correct ones, thereby failing to receive the benefit of reviewing and consolidating the content in their minds by a careful "second pass" through the material.]
Yet, if the answers are not checked in some way, many if not most students feel dissatisfied not knowing whether what they have written is actually correct, and from the teacher's standpoint it is difficult to expect the students to fully understand the meaning of the material and to test them on it later. If this basically mechanical process of checking fixed answers could be accomplished in such a way that it could be done individually on students' own time outside of class and, preferably, would require some sort of active participation on the part of the student at the same time, some of the above-mentioned shortcomings of in-class checking could be alleviated, while most of the benefits would be retained. TypeIt, which requires that students type in the complete text of the assigned dictations and keeps a record of their individual progress, was created with this in mind.
In the beginning students receive a personalized bootable program disk (FD) containing the TypeIt program (executable), an encrypted student ID file, and encrypted files representing at least a portion of the dictation materials to be assigned during the course. (All key files, including those logging the progress of the students as they work their way through the material, are encrypted to avoid the possibility of someone tampering with the disk contents.) The students are told that, rather than having dictation-checking done in class, they are expected to, first, do the dictations by listening to the homework tape at home and, then, once the dictation has been done, to check their work on their own time, but by a specified date, at an available console during the "free practice" hours in the school computer room, after which time it will be assumed that the student has checked and understood the content of the dictations.
A simple "manual" consisting of screen dumps of menus, explanatory messages, user dialogs and error messages (all given only in English) is used in class at appropriate times during the term to prepare students for performing the various functions related to doing the homework. For example, which mode should be selected in order to get started on an assignment, and the explanations of how to correct a mistyped letter, move on to the next block of text and end a session need to be covered before the students attempt the first assignment on their own, lest there be total confusion in the computer room! On the other hand, where to read the version number and date of the program on their disk (in order to confirm that a new version has been properly loaded) and how to receive new files from a data disk can and should be deferred until the need for them has actually arisen and the students have become comfortable with the more ordinary, day-to-day use of the software.
Since my purpose is not to teach them typing per se, I do not consider how many mistakes students have made or how long it took them to finish the designated assignments when calculating grades. Instead, I make this homework a non-graded but required part of the course, which everyone can do regardless of their proficiency either in English or in typing, provided that they are willing to invest the necessary amount of time.
TypeIt was written in Microsoft QuickBASIC using an "incremental" style of development, which is to say that a very basic implementation was first produced and put into use, and increasing functionality (improved interface, more careful and detailed collection of data) was added later as shortcomings of the present version and possible ways of improving it became apparent "in the field." The following is an overview of the present incarnation of TypeIt.
At my institution, during the period in which this software was developed and used, two classrooms, each containing at least 60 NEC computers and doubling as a "free practice" room at times of the day when no classes were scheduled for it, were available to students on a walk-in basis. The hardware was, however, of the most basic sort: the equivalent of 8086 processors, a pair of 3.5" 1.2MB floppy drive units, no added RAM and no hard disk at all, which mandated the use of an early version of MS-DOS. (Hence the choice of non-Windows QuickBASIC, rather than, say, Visual BASIC, as the programming language.)
Upon bootup the student is presented with a personalized screen showing the student's name, ID number and class, which is then followed by a menu of modes/activities: begin or continue work on an assignment, review a previously done assignment, receive additional assignments or a new version (upgrade) of the program from the instructor, or submit a progress report to the instructor. The latter two activities involve a second data/report disk and were necessitated by the lack of a hard disk on student machines.
In "Do" mode, the student selects the desired material from a menu, which indicates the completion status (NONE/PART/DONE) of each material currently loaded on the disk. After a message from the instructor regarding the given assignment, the program displays, in succession, blocks of one or more lines from the passage, which the student must type out in full in the space below. The text can be complete (corresponding to a printed transcript) or contain blanks (requiring that the student understand the correct spelling of those words), but in either case the student must type each entire block of text correctly before moving on to the next. Here, "correctly" means "not leaving any errors uncorrected"; mistypes are pointed out as they occur by an auditory signal and remain color-highlighted until they are corrected. Along the way, the software keeps track of time elapsed and number and kinds of errors made, which, although peripheral to the main function of TypeIt as a dictation-checking program, may nevertheless be useful if it is used as a typing practice program and illustrates the possibilities for other types of data gathering which could be done. Upon exit the student's progress for that session is displayed and recorded on disk.
"Review" mode also begins with a materials menu, but provides for sequential display of partially- or wholly-completed assignments only, and is for the convenience of students wishing to separate the typing and correction (of their previously-done dictations) phases. Only blocks of text which have actually been typed in by the student can be re-displayed in this way, and any words originally blanked out in the dictation "skeleton" now appear color-highlighted.
New assignments and software updates are provided on non-personalized disks distributed by the instructor, inserted in the second drive bay and called from the main program. When called in "New" mode, the appropriate files are transferred to the student's personal disk. Similarly, whenever deemed appropriate, students can use the "Submit" mode to send an encrypted summary of their progress at that point to a second disk, which can then be turned in to the instructor for recording.
Generally speaking, the response of students to the TypeIt homework has been positive. For most of my own students it has the novelty of being the first time for them to use the computer in connection with a class whose focus, English, is not itself related to computers. This, I feel, is a significant experience for Japanese students, considering that in other countries (in particular, my own--the United States) the use of a computer even in non-scientific fields is so widespread as to be taken for granted by virtually all college students; for example, many courses require that term papers be submitted as e-mail rather than as hard copy. The students who have finished the assignments on their disk--and not infrequently without a "marathon" session at the end of the term!--have a feeling of accomplishment when they see the menu of assignments flanked by a full column of "DONE" status indicators. These factors contributing to a positive attitude toward the use of computers would seem to make computerized homework a valuable addition, at least in moderate amounts, to an English course involving dictation.
There are, however, other considerations. Even after the software was written and debugged, the amount of time and energy required to manage the preparation, distribution, collection and recording of individual disks for several classes of students was considerable. Ideally, all students would do the assignments one at a time on schedule as planned, but in fact quite a few waited until near the end of the term to get started, and the resulting "crunch" is said to have had a negative impact on students of other computer-related courses with projects coming due at term-end. These problems should, however, be amenable to solution by a shift to hard-disk based program and data files on newer machines (now installed at my institution), which would eliminate the necessity for dealing with cumbersome stacks of floppies and the possibility of virus infection, and by stricter adherence to the assignment schedule.
In closing, I will present an example of the type of analysis that can be done on data gathered along the way with TypeIt. It should be borne in mind, however, that in order to achieve an acceptable level of confidence in the results, a number of variables should have been more carefully controlled and monitored, so this example should be viewed only as an illustration of techniques.
Let us look at the change in "average" typing speeds and numbers of mistypes per character during the course of a term. Three classes totalling seventy students were given a total of 15 assignments, all of which provided the complete texts of the dictations; i.e., the "printed answer sheet" model mentioned in Section 1 (Background) above. Although speedy typing was not a goal for them, perhaps some trend, either for the better or the worse, might be discernible. To chart this trend, we need to know the order in which assignments were done. Unfortunately, the version of TypeIt used for this group did not record the dates on which a student did individual assignments, and the situation is further complicated by the fact that students were allowed to complete assignments in multiple sittings. What is available, however, is the order in which each assignment was begun, and we shall use the order of these start dates as the basis for what follows.
What becomes apparent immediately upon computation and plotting of "normal" typing speed (expressed as its reciprocal: seconds per typed character) versus assignment number with this data set, is that many students show very wide fluctuations in their typing speeds. I hypothesize that one likely reason for this is that some students may have taken extended breaks in the middle of some assignments to talk with friends (or whatever), which would be consistent with the unusually long times taken to complete certain assignments, e.g., 1 1/2 hours for a few assignments, as opposed to their more typical 25 or 30 minutes for other assignments of comparable length. But how to "cull out" such anomalous data? Rather than rely on the arguably better subjective method of looking at each student's data individually and throwing out "obviously way off" data points, I decided to employ a purely mechanical, but systematic and reproducible, method which amounted to selectively deleting a limited (fixed) number of data until the remaining points expressed a relationship as close to linear as possible. More specifically, the individual data of each student in the data set was examined by a computer program written specifically to perform the following calculation: all possible subsets (= the candidates for anomalous data points) of a given size less than a predetermined cutoff value were deleted in turn from each 15-point data set and the remaining set of points checked for linear correlation as measured by the 2-variable correlation coeffient (see any statistics textbook), the set of best linear fit was determined, and lastly the students' average rates of improvement/decline over those sets were used to compute an "average of averages" profile for the 70-member group as a whole. The final results were somewhat disappointing, indicating as they did, not the hoped-for improvement, but instead a very slight decrease in typing speed and increase in mistype density over the course of the term. Here, too, however, it is possible that other factors were at work, for example, more errors being made as people rushed to complete remaining assignments under submission deadline pressure and the resulting increase in time needed to correct those errors. Obviously, a more carefully conducted study aimed specifically at elucidating this phenomenon is called for.
Computerized homework has great potential for motivating students and providing access to data trends, which in turn can lead to new avenues of research. At the same time, a significant commitment of time and energy resources will also almost certainly be required in the planning and implementation of a project of any appreciable size, as well as in the proper interpretation of quantitative data obtained therefrom.
Please direct any comments or suggestions regarding this article to the author
at the following e-mail address: