Section II: Chapter 2: Review of Literature: Development of Keyboarding Skills: Research: Theory & Research: Universal Design for Learning: CAST

Before we can discuss the probability that individuals with learning disabilities can or cannot learn keyboarding, it is necessary to understand how patterned fine motor skills are developed. A great deal of study has been devoted to the ways in which individuals acquire and automatize fine motor skills, and a significant subset of this literature specifically addresses typing or keyboarding. It is impossible to summarize all findings of this literature in the current thesis, but a brief review of representative works is appropriate.

There is general agreement that the learning of complex fine motor skills is achieved through the progressive combination of smaller subunits of motor control (Adams, 1984; Fischer, 1987; Hay, 1984; Heier, 1984; Kihlstrom, 1987; Kornhuber, 1984; Luria, 1973; Neuman & Morocco, 1985; Pew, 1984; Roberton, 1982; Rosenbaum & Saltzman, 1984; Seefeldt & Haubenstricker, 1982; Shiffrin & Schneider, 1977; Sternberg, Monsell, Knoll, & Wright, 1978). At the most basic level, learning of the smallest of motor control subunits cannot be demonstrated and these subunits are generally assumed to be innate.

How are Keyboarding Skills Learned?

Researchers examining the acquisition of keyboarding skills have accepted this overall view, and have looked closely at movements and feedback loops as their subjects developed proficiency. Cooper, in his introduction (Cooper, 1983) makes a distinction between transcription and generative typing but with few exceptions addresses the former. The recognition of the differences between these two forms of typing is seen as important and progressive, and the paucity of research on generative typing both in Cooper's otherwise excellent book and in the literature as a whole is seen as a serious deficiency.

Cooper has provided a model of the psychology of the transcription typing process that is reproduced here. The stages of this model remain the same no matter the degree to which typing skills are developed in the individual.

Overall control is maintained by the executive pacer, which oversees recognition of the individual character to be typed and selects an appropriate motor pattern for execution. The character is held in a short term memory buffer until the motor pattern is initiated, and this initiation results in a keystroke. During and after the completion of the keystroke sensory feedback in the form of kinesthetic, visual, and auditory cues feed back to the executive pacer to evaluate the success of the motor program's execution. However, the type of feedback that is most heavily relied upon, and some internal details of stages, vary as the individual's skill improves from highly conscious control towards automaticity.

In the initial stages of learning to type the process is mediated almost exclusively visually. The novice typist scans the keyboard to locate the individual key to be struck, locates the key, strikes it, and goes on to locate the next key. Rosegrant has noted that children as young as 4 years of age develop first a quadrant strategy for subdividing the keyboard, and then a row scanning strategy to locate their target letters. Rosegrant's subjects used their dominant hand exclusively (Rosegrant, 1984). This process is slow because the individual must scan visually to locate the correct key and then must assemble a series of small units of motor performance in a relatively novel pattern in order to execute the keystroke. Further, because kinesthetic feedback loops have not been developed, the novice typist relies exclusively on a visual examination of the typed character to determine if the correct key has been struck. These individuals type demonstrably faster in meaningful text than in either nonsense syllables or in syntactically nonsensical strings of actual words. Cooper suggests that this is a result of the relative ease with which meaningful text can be memorized (Cooper, 1983). This memorization results in fewer visual scans of the model passage which in turn results in time savings when contrasted with typing tasks where memorization is more difficult and presumably can be carried out only on smaller blocks of text.

As expertise is developed through practice and perhaps with the aid of instruction the process changes in several interesting ways. Most importantly, even though skilled typists have been shown to rely to some extent on visual feedback for positioning, the primary motor control mechanism is no longer visual, it is proprioceptive (Cooper, 1983). This distinction is seen as parallel to the stages of general motor skills development described by (Hay, 1984) who termed initial efforts "ballistic" and referred to practiced and refined motor sequences as "feedback oriented".

Finger movements, which were strictly serial in the initial stages, begin to overlap as skill develops, with as many as three distinct keystrokes in the process of execution by individual fingers at any given instant. The difference between the rates at which connected text and syntactically nonsensical strings of words can be transcribed is much less pronounced than it is in novice typists (Cooper, 1983).

In a study of speed and accuracy where comparisons were made between performance of skilled typists when theywere and when theywere not able to see their keyboards and typed products, it appears that skilled typists still look at the keyboard to an extent, but are much less reliant upon visual cueing than are novices (West, 1967). Speed was unaffected by masking the keyboard and typed product, but errors increased by 60%. Cooper concludes from analysis of errors that the preponderance of these result from typists' inability to visually orient themselves and correctly position their fingers on the home row keys (Cooper, 1983). Perhaps the addition of tactile feedback in the form of the small bumps or ridges found commonly on computer keyboards could be used to decrease still further the need to rely upon visual information for initial finger positioning.

Further evidence for kinesthetic, rather than visual, mediation of skilled typewriting comes from the study of errors made by skilled typists. Cooper, citing Rabbitt, (1978), notes that skilled typists, who are executing several finger movements simultaneously as they work, frequently depress incorrect keys with less pressure than they depress correct keys. This suggests that the subjects are detecting errors while in the process of making them, but succeed in only partially stopping the motor patterns' inaccurate executions. The speed of this response is said by Cooper to rule out visual mediation and to point to kinesthetic feedback generated after motor commands have been initiated.

Perhaps the best summary of the process of learning transcription typing is the description of handwriting generated by Luria, who observed that:

The same could be said of keyboarding, where in the early stages of learning each letter must be located and struck individually, but with practice automatized finger movements are executed smoothly, simultaneously, and accurately.

During the childhood years motor skills acquisition generally occurs more rapidly in older individuals than it does in younger. Some skills may have a developmental threshold, before which they may prove difficult or impossible to learn. For motor skills in general as well as in the acquisition of typing skills (Book, 1925) the acquisition process is not reflected in a smooth, upwardly sloping curve. Rather, development is characterized by a series of periods of rapid development punctuated by plateaus during which little or no observable development takes place. It has been suggested that during these fallow periods the individual is performing some form of accommodation or consolidation of the newly acquired skills, in preparation for using these skills and perhaps incorporating them as subskills within a more complex yet-to-be-acquired skill.

Five Critical Questions About Keyboarding Instruction

As with any complex motor task, it seems that instruction may improve subjects' rate of acquisition of keyboarding skills. Fry has advocated formal keyboarding instruction as desirable and necessary (Fry, 1988). Support for this view is found in the writings of several authors who have investigated the potential of the computer to aid the writing efforts of students with learning disabilities (Follansbee & Meo, 1987; Hoffman & Welk, 1986; Katz & Hoffman, 1987; Meyer & Pisha, 1987; Meyer & Rose, 1987a; Pisha, 1989; Pisha & Follansbee, 1988; Saka, 1986). Interestingly, this question is infrequently addressed in the literature, and fundamental questions pertaining to keyboarding are either incompletely answered or simply ignored. Among these are: 1.) What rate of keyboarding is necessary or desirable for emerging writers? 2.) Should keyboarding instruction precede access to the computer for writing tasks? 3.) Is keyboarding instruction necessary? 4.) If keyboarding instruction is to be provided, at what age should it be initiated? 5.) Can all students acquire this skill? Current research and practice relating to these questions will now be reviewed.

1) What Keyboarding Rate is Adequate for Writing?

There appears to be universal agreement that more keyboarding speed is better, in that it is likely to translate into automaticity of routine motor output tasks. This automatization of relatively mechanical task components can be expected to free up cognitive capacity for higher level writing activities (Cooper, 1983; LaBerge & Samuels, 1974; Neumann, 1984; Shiffrin & Schneider, 1977). This point is well expressed by McAndrew who noted, "As long as the mechanical processes involved in writing are themselves highly conscious, slow, or even labored, writers are not likely to have easy access to their thoughts." We may anticipate that individuals who have thoroughly over learned and automatized the basic motor patterns necessary for writing with either a pencil or a keyboard will need to think very little about the physical act of writing. They will be able to devote a greater proportion of their finite cognitive resources to more interesting aspects of writing, such as content, organization of ideas, choice of words, and the development of a personal "voice".

Only a few researchers have stated what keyboarding speed is adequate for generative typing. Several authors suggest that relatively modest speeds, from as few as ten words per minute (Wetzel, 1985), to fifteen words per minute (Hoffman & Welk, 1986), or twenty words per minute (Kissner, 1984) are sufficient. Kissner suggested that some formal instruction by teachers with specific expertise in typing instruction would be required to attain that rate. Arguably, an individual who can enter text at the keyboard as rapidly as she or he can handwrite knows the keyboard (Fidanque, 1986; Hoot, 1988; Kissner, 1984; Pisha, 1989).

Pisha (1989) found that the late-elementary school aged children in his study could copy manually from a model at speeds from eight to seventeen words per minute. Output averaged eleven words per minute, and legibility was often compromised, even at this relatively modest rate of speed. Others (Pianko, 1979) found both remedial and normally developing writers composed at slightly more than nine words per minute and (Stallard, 1979) found 13.5 words per minute among skilled writers. These speeds seem realistic for short timed tests, but it is difficult to imagine sustained composition at these rates.

For the purposes of this study, the handwriting speed of each individual will be considered to be the minimum keyboarding speed necessary for proficiency. If speed of written output is accepted as a reasonable proxy for the cognitive load associated with the physical act of writing with a pencil or a keyboard, it follows that when individuals can keyboard as rapidly as they can handwrite the cognitive effort required for both tasks is approximately equal. Once the cognitive loading from output modality is equalized, the powerful features offered by word processing (spelling checking, ease of revision, typewritten final copies, potential for direct electronic transmission) can be expected to become accessible to the writer. Of course, as any individual continues to practice keyboarding by actually using the computer for writing he/she may reasonably be expected to increase speed and proficiency. The result of this increase will be a further freeing of cognitive resources for linguistic, rather then mechanical, aspects of writing. In contrast, it appears that after some relatively modest level of handwriting speed is attained, further increases in speed and freeing of resources can come only at the expense of legibility.

2) Is Keyboarding Instruction Necessary Prior to Writing With a Personal Computer?

In recent literature there is some agreement that it is not necessary to develop keyboarding skills prior to the inception of computer assisted writing. However, a dissenting opinion that students should be able to keyboard at 20 words per minute prior to the introduction of computer-assisted writing was found (Kissner, 1984), and (Cooper, 1983) offered an unsupported opinion that skilled transcription typing is a prerequisite to skilled generative typing. Other authors report that their students readily accepted the computer as a writing tool without any formal keyboarding instruction and were able to achieve reasonable output speeds in a short time using the "hunt and peck" technique (Freyd & Khan, 1989; Kahn, Avicolli, & Lodise, 1988).

Pisha (1989) found that students with little or no experience keyboarding in his sample of late elementary school aged boys with diagnosed learning disabilities were able to keyboard from a model at an average rate of four words per minute. These subjects had limited keyboarding experience, and no formal instruction, but still preferred word processing to more traditional forms of writing.

These students preferred to use the computer for writing tasks. Conversely, none of the students in Pisha's study showed any interest whatsoever in acquiring keyboarding skills for their own sake. Even though these students eagerly embraced word processing technology for their writing work, they would not use a computer based keyboarding instruction program to practice keyboarding independently. They would, however, use this software when directly supervised and encouraged by a teacher, and undue duress was not required. It appears that these students were capable of using the keyboarding instruction program, and perhaps even saw some degree of merit in its use. However, they characterized computer assisted keyboarding instruction as a boring activity when practiced in isolation (Pisha, 1989).

Another finding from Pisha's (1989) preliminary study is the relationship between the ability of subjects to use the computer to complete regularly assigned schoolwork and increase in measured typing speed. Without exception, the students who achieved significant increases in keyboarding skills in Pisha's 1989 study did not practice more than other students on the provided drill and practice keyboarding software. The students who learned to keyboard rapidly were those who were allowed to use their computers daily to complete meaningful assigned schoolwork. For them, it is suggested, keyboarding was learned because it was an element of the solution to a problem, satisfactorily completing demanding school writing assignments. The computer made writing easier for these learning disabled students, and the burden of learning keyboarding proved less onerous than the burden of doing this work without the aid of a computer. In contrast, for other students who were not allowed by their schools to submit computer generated or typewritten output, keyboarding did not present a solution to an existing problem. Rather, it presented a new problem: One more isolated and difficult to learn skill in an environment rife with difficult and isolated skills.

3) Is Keyboarding Instruction Necessary at All?

Some authors have suggested that it is inappropriate to provide any keyboarding instruction at all, maintaining that children will learn the keyboard on their own through "hunt and peck" strategies (Khan & Freyd, 1990a; Khan & Freyd, 1990b). These authors monitored the change in keyboarding speed of a group of elementary school students who, after some experience with Logo, were afforded one hour per week for writing with a computer word processor over a seven month period. They found a strong correlation between handwriting speed and keyboarding speed (P<.001) among individuals who typed at speeds from 3 to 21 words per minute, and noted that the average keyboarding speed for these students increased from 6.62 words per minute at the onset to 11.4 words per minute at the end of the study. Regrettably, no finer analysis of the rates of skills growth in subgroups of students, for instance those who could initially keyboard at only 3 words per minute, is provided by these authors. Further, Kahn and Freyd suggest that speed, accuracy, and "bad habits" are unimportant to students who use word processors. Although this may be accurate if one considers keyboarding different from typing as it has traditionally been conceptualized, these authors do not mount a detailed argument to support this contention.

It is certainly reasonable both from the viewpoint of motor learning theory and from the vantage of clinical experience to assume that students will acquire some knowledge of the keyboard without benefit of instruction. This view also is supported to an extent by Pisha's (1989) findings that students who used keyboarding for practical tasks seemed to learn to keyboard more rapidly than students who did not. Why, then, ought schools teach this skill?

Although at least some individuals can achieve some proficiency using hunt and peck techniques, it is certain that these individuals will never be able to achieve the speeds (up to 216 words per minute) reported by skilled touch typists. Informal measurements of several self-taught adult hunt and peck typists show that their keyboarding rates approximate 35 words per minute. This is substantially faster than these individuals can handwrite, but not nearly as fast as the keyboarding of other adults who have received instruction and use a more standard ten finger, home row oriented approach.

The hunt and peck typists relied more than touch typists on visual feedback, looking at the keyboard all of the time to guide their fingers. This is understandable when one considers that, using a limited number of fingers it is impossible to firmly establish any sort of a home position that would allow the development of proprioceptive feedback loops to govern the process. Hunt and peck typists generally make longer reaches because they do not use all of their fingers, and they tend to strike any given key with one of a number of fingers, depending upon the context. This, too, fosters reliance upon the less automatic visual feedback strategy to the detriment of more efficient and lower level proprioceptive feedback loops. As a result, untutored keyboarders may well be able to develop their skills to the point where they can type more rapidly than they can handwrite, but they are unlikely to be able to automatize the output process to the degree that a ten-finger typist may be expected to develop it. As a result the hunt and peck typist will always be forced to allocate more cognitive resources to writing output and will have less available for thinking about what is being written when contrasted to a comparable individual who has learned ten finger technique.

Even if the thesis that keyboarding skills can develop spontaneously in the school aged population with exposure to computers is accepted, recent research suggests that few students are fortunate enough to have enough time using computers in schools to develop proficiency. Further, much of the limited computer time today's students do have is devoted to drill and practice activities that are not particularly related to writing and probably do not require extensive use of more than a handful of isolated keys needed to respond to multiple choice queries (Becker, 1983; Becker, 1986a; Becker, 1986b; Becker, 1986c; Carey & Gall, 1986). It is unlikely that students will develop significant keyboarding skills under these circumstances. At least a modest amount of instruction and encouragement to adhere to a standard ten-finger approach to keyboarding appear to be highly desirable.

4) At What Age Should Keyboarding Instruction Begin?

Recently the introduction of personal computers and word processors in schools has lead some teachers and researchers to give thought to keyboarding instruction. There has been widespread agreement that keyboarding instruction is necessary, despite the objections of some writers (Khan & Freyd, 1990a; Khan & Freyd, 1990b). The most frequently asked questions have addressed the earliest age at which keyboarding can reasonably be introduced. One researcher (Britten, 1988) has suggested that with sufficient practice children as young as seven years of age can learn to use the keyboard. However, methodological weaknesses render these conclusions suspect. Rosegrant studied students in the four to five year old range, and found that they could at least develop logical search strategies for locating individual keys when needed (Rosegrant, 1984). Others (Cuffaro, 1984; Hoot, 1986; Hoot, 1988; Jackson & Berg, 1986; Wetzel, 1985; Wetzel, 1987; Wronkovich, 1988) have taken a more moderate stance, suggesting that instruction during the middle to late elementary school years is sensible.

A theoretical argument for beginning keyboarding instruction in the early school years can be mounted. First, keyboarding may be less cognitively and physically demanding than handwriting. At least two sources of evidence for this assertion can be found. On face, it is easier to locate one of approximately 60 keys and depress it than it is to copy a complex series of up to five distinct fine motor movements needed to write any individual letter. Further, the maximum rate of keyboarding found in a literature search was 216 words per minute on a manual typewriter (Long, Nimmo-Smith, & Whitefield, 1983), but it appears on-face impossible to produce legible handwriting a rate even approaching that. It is quite common for skilled typists to achieve 60 words per minute with excellent accuracy, but this level of speed is exceedingly uncommon in standard English handwriting. Using speed as a proxy for cognitive load, keyboarding is arguably less taxing than handwriting. Support for this view can also be found in the literature (Campbell, 1973).

A study of current practice of handwriting instruction in schools polled 400 teachers of kindergarten through grade three and found that a daily average of from eleven to twenty minutes were spent in handwriting instruction. This instruction took place largely in the form of whole-class drills and was most prevalent during the first grade year (Addy & Wylie, 1973). Using simple arithmetic, four 180 day school years of handwriting instruction at even eleven minutes per day amounts to 132 hours of instruction. This seems to be an immense expenditure of instructional time, particularly in light of the generally mediocre quality of handwriting displayed by most fourth grade students. We are asking very young students to struggle with a difficult psycho-motor task when we demand that they learn traditional handwriting in the early grades. Would it not be easier and more efficient to teach keyboardingfirst, prior to the introduction of pencil and paper writing tasks?

Reports in the literature suggest that children can learn keyboarding with significantly less instructional time than Addy found is being spent on handwriting instruction (Addy & Wylie, 1973). Jackson and Berg, (1986) have suggested that keyboarding instruction start in third grade, and that 30 hours of instruction, distributed across two or three school years is sufficient to achieve mastery. They also report a doubling of keyboarding speed after fifteen hours of instruction. Cameron (1986) shares Jackson & Berg's view that formal keyboarding instruction should start in third grade. (Khan & Freyd, 1990b) cite examples of students in grades four through six who improved their average keyboarding rates from 6.63 words per minute in October to 10.12 words per minute in May of the same year without any formal instruction. These students had access to word processors for their daily schoolwork. With practice, but no formal instruction, they approached their average handwriting speed of 11.4 words per minute. Other authors (Yuen, Carrillo, Bjonerud, & Chambers, 1961) focused on fourth grade students and found that these children learned to type faster than adults could be expected to learn. These subjects were able to type faster than they could handwrite after only ten hours of instruction. Pisha (1989) found that several students in his small sample were able to improve their keyboarding skills by a factor of two or even three with only two hours of formal instruction if they were able to use word processors to do meaningful homework between brief weekly instructional sessions.

Pisha (1989) provided instruction and opportunities for practice to a small sample of male students with identified learning disabilities in grades three through eight. He found a modest age effect, with older students achieving greater gains than younger. However, even the youngest students in this study were able to improve their keyboarding ability and some of these reported using their computers for meaningful work. On the basis of these studies, it appears that children of nine years of age and older can certainly benefit from keyboarding instruction. However, no experimental evidence has been found to suggest that younger students cannot learn keyboarding.

5) Can All Children Learn Keyboarding?

Surprisingly, given all of the research interest in the potential of word processing to improve writing instruction for individuals with and without identified specific learning disabilities, there is relatively little discussion of what might be considered a critical "gatekeeper" skill for word processing: keyboarding proficiency. A significant amount of time is spent in the primary grades teaching handwriting (Addy & Wylie, 1973). However, despite the clear difficulties many students experience in learning to handwrite and the promise of computer assisted writing, little recent research has addressed acquisition of keyboarding skills in the elementary years and few schools offer any systematic instruction.

A few authors have addressed keyboarding skills development among individuals with special needs (Hoffman & Welk, 1986; Katz & Hoffman, 1987; Pisha, 1989; Saka, 1986; Vacc, 1987; Campbell, 1973; MacArthur, 1986). Each of these cites some evidence to support a contention that children with special needs can acquire sufficient proficiency in keyboarding to allow them to effectively use the word processor or typewriter as a writing tool. However, with the exception of Pisha (1989), none of these has begun to study this question in a systematic way, and the conclusions of Pisha's study are rendered questionable by his small sample size, only twelve. A review of literature proved unable to locate any larger scale study directly measuring the rates at which children with special needs acquire keyboarding skills and contrasting these rates with a control group without identified special needs.

Chapter 2: Review of Literature, Section II

How are Fine Motor Skills Patterns Learned?