DOCUMENT RESUME

ED 313 002 IR 014 041

AUTHOR Joyce, Bonnie G.; Moxley, Roy A.

TITLE Comparing Children's Typing Skills Using the Dvorak
and QWERTY Keyboards on a Microcomputer.
PUB DATE 89
NOTE 21p.
PUB TYPE Reports - Research/Technical (143)

EDRS PRICE MF01/PC01 Plus Postage.

DESCRIPTORS Comparative Analysis; *Intermode Differences;
*Keyboarding (Data Entry); *Microcomputers;
Observation; Primary Education; *Young Children
IDENTIFIERS *Dvorak Keyboard; *QWERTY Keyboard

ABSTRACT
This study compared the typing efficiency of four
young children (5 to 6 years of age) who were novice typists on the
Dvork and QWERTY keyboards. A copying program on an Apple IIc
microcomputer functioned as the training instrument. Although the
children did not acquire proficient touch typing skills, they did
type accurate responses faster, keep their hands positioned on the
home row when they began typing, and show some conditioning to key
locations when using the Dvorak keyboard. While the children's
fingering was often inconsistent and they continued to glance at the
keyboard for confirmation of key locations, their acquisition of
touch typing skills would probably have been more extensive had they
stayed with the same keyboard throughout instead of changing from the
QWERTY to the Dvorak version. Although there is some evidence that
children as young as 6 years can acquire touch typing skills, it is
unclear what the recommended procedures should be for introducing
young children to keyboard fingering. Additional research in this
area acquires increasing importance as more young children use
microcomputer keyboards. Twelve references, 2 tables, and 1 figure
are included. (GL)

Reproductions supplied by EDRS are the best that can be made

from the original document.
 U.E. DEPARTMENT or EDUCATION
Office of Educational Rsearch and improvement
EDUCATIONAL RESOURCES INFORMATION
CENTER (ERIC)
)(This document has been reproduced as
received from the person or organtration
originating it
r Minor changes have been made to improve
reproduction Quality

Points of vie* or opinions statedin thsddcu

rnent do not necessarily represent otttcal
OER1 position or policy

Comparing Children's Typing Skills Using

the Dvorak and QWERTY Keyboards on a Microcomputer

Bonnie G. Joyce

New Medico Head Injury System

Community Re-Entry Services of Arkansas

Benton, AR

Roy A. Moxley

Division of Education

West Virginia University

Running head: Dvorak and QWERTY Keyboards

"PERMISSION TO REPRODUCE THIS

MATERIAL HAS BEEN GRANTED BY

Bonnie Joyce

2 TO THE EDUCATIONAL RESOURCES

INFORMATION CENTER (ERIC)."
dEST COPY AVAILABLE
 Dvorak and QWERTY Keyboards.

Abstract

This study compared the typing efficiency of four young

children, who were novice typists, on the Dvorak and QWERTY

keyboards. A copying program on an Apple IIc microcomputer

functioned as the training instrument. Although the children

did not acquire proficient touch typing, they did type accurate

responses faster when using the Dvorak keyboard. The results

are discussed in terms of the advantages of using the Dvorak

keyboard with young children.

3
 Dvorak and QWERTY Keyboards'

Comparing the Use of the Dvorak and QWERTY

Keyboards by Young Children

In the 1930s, August Dvorak and his colleagues designed a

keyboard for touch typing to replace the standard QWERTY

keyboard that was designed in the late 1870s before touch typing

came in use (Dvorak, Merrick, Dealey, & Ford, 1936). The Dvorak

keyboard increased the amount of home row typing, increased the

number of words that were typed exclusively on the home row,

reduced finger motions among rows, decreased the left hand

overload, and decreased specific finger overloads. In the

Dvorak arrangement of letter keys, only the letters A and M are

in the same position as they are in the QWERTY arrangement:

Dvorak PYFGCRL
AOEUIDHTNS
QJKXBMWVZ
QWERTY QWERTYUIOP
ASDFGHJKL
ZXCVBNM
Although the data vary on the degree of improvement offered

by the Dvorak keyboard, there seems to be a general consensus in

the research literature that it is a more efficient keyboard for

touch typing (Joyce & Moxley, 1988; Yamada, 1980). In his

review of the research, for example, Yamada (1980) estimates

that Dvorak users type 15% to 20% faster than QWERTY users for

timed copy-typing of limited duration and 25% to 50% faster in

4
 Dvorak and QWERTY Keyboards.

routine production typing. Despite its increased efficiency,

the Dvorak has not been adopted as a standard keyboard because

initial costs can be high for replacing the keyboards and

retraining typists.

With modern computers, however, keyboard conversion is much

easier than it was in Dvorak's day, and retraining costs can be

eliminated if touch typing is taught on the Dvorak keyboard from

the beginning. Although this might suggest that all children

should begin typing on the Dvorak keyboard, such may not be the

case. A keyboard that is superior for touch typing (e.g.

Dvorak) may or may not be superior for the hunt-and-peck typing

of young children who are typing novices. For example, when

fingers are positioned over the home row for ten-fingered

typing, visual searching by novice typists should be more

difficult on the Dvorak because the most frequently used keys

are covered. In contrast, the most frequently used keys on the

QWERTY keyboard are on the uncovered top row.

Because our review of the literature found no data that

favored or disfavored the initial use of the Dvorak keyboard by

young children, the following study sought to provide evidence

that would help to resolve this issue. To do so, we looked at
four young children's experience in copying printed words with

the Dvorak aria QWERTY keyboards in a naturalistic, after school

setting over an 8 month period.

5
 Dvorak and QWERTY Keyboards'

Method

Children and Setting

Four children were selected for instruction based on their

age (5 to 6 years of age), good attendance, and little or no

previous experience using typewriters or computers. Three of

the children attended kindergarten for a half day program (Beth,

6 years, one month; Kathy, 5 years, 5 months; and James, 6

years, 3 months) and one child attended first grade (Julian, 6

years, 4 months).

Pre Assessment of Keyboarding Skills

An informal assessment requiring the children to type all

the words they could from memory was used to initially assess

keyboarding skills. Results of the assessment indicated that no

child had knowledge of the keyboard arrangement as demonstrated

by their visual scanning of the letters and skimming the rows of

keys with a finger until locating a specific key. They

primarily struck the keys with their index or third finger and

frequently used the space bar inappropriate'y. Julian, for

instance, pressed the space bar after 13 words but failed to

press it after ' words. Beth, the only other child to use the
space bar, pressed it once in the middle of her name.

Experimental Design

In order to evaluate the efficacy of the two experimental

conditions, a single subject ABAB design, in which the emphasis

is on repeated objective measurement in a single individual over

an extended period of time, was counter balanced to more clearly

6
 Dvorak and QWERTY Keyboards

identify order effects that might result when learning on one

condition contributes to learning on another condition (Barlow &

Hersen, 1984; Best & Kahn, 1989; Kazdin, 1982). Both

experimental conditions, represented by the Qwerty and Dvorak

keyboards, were implemented twice for a total of four phases per

child (except for James who completed only three phases when his

mother removed him from the day care center at the end of the

academic year). Each phase consisted of approximately 20

sessions. After completing 20 sessions on one keyboard, the

computer was converted to the other experimental condition by

depressing a switch on the computer and rearranging the keys on

tne keyboard. Twenty sessions later, the computer was
re-converted. This process continued until all four phases were
completed.

Procedures

Each day, before instruction began, the teacher reminded the

children to tap lightly on the keys (as opposed to keeping the

ke. iepressed). They were also instructed to begin typing with

their hands on the home row and to strike each key with the

proper finger for touch typing.

Direct Copy Program. A Direct Copy program on an Apple IIc

microcomputer, which was a version of programs developed to

improve children's literacy skills, functioned as both a

training and testing instrument for increasing typing frequency

and accuracy (Moxley, 1986; Moxley & Barry, 1985; Moxley &

Barry, 1986). The program presented vocabulary words that were

7
 Dvorak and QWERTY Keyboards'

individualized for each child Approximately 15 words were

available for each daily session and the computer randomly

selected 10 cf these words. The program was stopped after the

child typed 10 words or when 4 minutes elapsed, whichever came

first. Shorter words were used in the beginning of each phase

and were replaced weekly by longer words as the children

expanded their vocabulary. The vocabulary words were presented

on the screen and were pronounced by a speech synthesis unit

connected ro the computer. Key presses were displayed on the

screen and were recorded by the computer. Correct letters

remained on the screen and were pronounced as a unit. The child

received posdtive feedback for correctly typed words from the

speech synthesis pronunciation of the word, praise from the

teacher, and an automatic display of a star on the screen.

Incorrect letters disappeared after a brief "flash" on the

screen.

After the Direct Copy program was administered, a computer

printout was made showing correct key presses for each word,

incorrect key presses, seconds between the presentation of the

stimulus and the first key press, seconds between each key

press, and cumulative time elapsed since the stimulus was

presented. Frequencies correct and incorrect were calculated by

the computer so the data could be plotted on an equal ratio

chart. The data were shared with the children and positive

comments given for improvements in rate and accuracy.

8
 Dvorak and QWERTY Keyboards'

Timed Samples. A timed sample of the student's typing of a

printed passage, was administered by the teacher after the

completion of each 20 session phase (with the exception of

Julian's first phase because the initial plan was to use only

the frequencies correct and incorrect from the Direct Copy

program as the dependent variable). The timed sample consisted

of one of five printed passages having the same words but in

different order.

Results

Post Assessment of Keyboarding Skills

Although no child fully acquired touch typing, increases in

keyboarding skills were observed. Such changes included keeping

hands elevated over the home row of the keyboard, keeping the

little fingers of each hand on the ends of the home row when

lifting the hands to search for a key, consistently pressing the

space bar between words, and quickly tapping a key rather than

leaving the key depressed. Some good ten-fingered typing was

acquired by Kathy. This, however, was still done mainly by

looking at the keys rather than by touch.

Direct Copy Changes

Daily frequencies of correct and incorrect typing were

compared across experimental phases. Frequency correct was

defined as the number of correct key presses divided by the

number of minutes. Frequency incorrect was defined as all

incorrect key presses (except those errors created by the same

 Dvorak and QWERTY Keyboards'

consecutive effect) divided by the number of minutes.

Frequencies correct and incorrect were calculated for each

vocabulary word and were totaled for each session. See Figure 1

and Table 1.

Insert Figure 1 and Table 1 about here

The Direct Copy data regarding celerations of frequencies

correct showed advantages for using the Dvorak keyboard. The

steepest accelerations occurred using the Dvorak keyboard for

all four of the children (Julian = X1.5, Beth = X1.5, Kathy =

X1.4, and James = X1.9). In addition, the accelerations on the

Dvorak keyboard were always higher than the accelerations in the

previous QWERTY phase. The QWERTY accelerations were also lower

than the previous Dvorak accelerations except for Julian's third

phase, which was slightly higher than the preceding Dvorak

phase. The Dvorak celerations for frequency correct were

steeper than the QWERTY, ranging from X1.2 to X1.9 with a median -

of X1.5, compared to X1.0 to X1,4 with a median of X1.2 on the

QWERTY. All celerPtion lines were calculated using the

Quarter-Intersect Method (PeLnypacker, Koenig, & Lindsley,

1972). Frequencies for each session are displayed in Figure 1.

Although celerations of frequencies correct were higher for

the Dvorak keyboard, celerations of frequencies incorrect were

highly individual for most children and generally did not favor

10
 Dvorak and QWERTY Keyboards'

10

either keyboard. James, however, was an exception. His data

clearly favored using the Dvorak keyboard. His highest

acceleration incorrect occurred in the QWERTY 1 phase, which was

the same phase as his lowest acceleration of frequency correct.

This was followed by his Dvorak 2 phase in which he had his

highest acceleration of correct responses and lowest

acceleration of incorrect responses.

The data also indicated the presence of order effects from

experience with the keyboards. Higher accelerating frequencies

across phases occurred in 7 of the 11 phase changes. The median

frequencies correct also tended to increase across phases, with

the exception that Beth and Kathy's fourth phases were lower

than their third. Except for James, the greatest gain in

frequencies correct occurred between the second and third

phases.

Order effects were also demonstrated by step values, the

measure of behavior change between the point where the

celeration line of one phase ends and the point where the

celeration line of the next phase begins (calculated by dividing

the larger point value by the smaller value). Julian, Beth, and
/

,
/ James exhibited greater step decreases in correct frequency
celerations with each phase (see Table 1). The incorrect step

values decreased across all phases for both Beth and James.

These values also decreased for Julian and Kathy until the last
 Dvorak and QWERTY keyboards'

11

phase when both children showed a slight increase in the step

values.

Further evidence of order effects were indicated by one

child who showed keyboard control from the previous phase by

striking keys that would have been correct on the keyboard in

the previous phase. In his 4th (Dvorak) phase, Julian hit J for

C in CAT (session 1), 0 for S in SAID (session 6), C for S in

WATCHES (session 8), and H for J in JUMPING (session 14).

Posttests of Typing Speed

There was always a gain in frequencies correct on the

posttests following the Dvorak phases but not always after the
QWERTY phases. Kathy and James, for instance, had an increase

in their frequencies of correct responses with each phase, while

Julian and Beth had a decrease in frequencies correct on their

second posttest (QWERTY keyboard). In addition, the gains in

frequencies correct were always greater for the Dvorak than the

QWERTY keyboard and errors were either nonexistent or less than

those obtained on the QWERTY posttests. Typing was also more

accurate in the Dvorak phases. Eleven of the errors occurred

using the QWERTY keyboard (three by Julian and eight by Beth)

while only one error (by Beth) occurred using the Dvorak

keyboard. See Table 2.

Insert Table 2 about here

12
 Dvorak and QWERTY Keyboards'

12

Discussion

None of the results supported the use o.:. the QWERTY

keyboards by novice typists even though home row finger

positions would seem to favor visual searching on that

keyboard. On the contrary, all the differences between the two

keyboards favored the Dvorak keyboard, and the extent of the

differences may have been suppressed by order effects.

Analysis of celerations of frequencies correct provided the

strongest support favoring the Dvorak keyboard. The range and

median of celerations correct on the Dvorak ws higher than the

range and median of celerations correct on the QWERTY. The

steepest acceleration of frequencies correct for all children

occurred on the Dvorak, and accelerations on the Dvorak were

always higher than those of the previous QWERTY phase. This was
not true of the QWERli keyboard where the accelerations were

lower than on the previous Dvorak accelerations with the

exception of Julian's third phase. In the timed posttests,

which are more conventional measures of t7ping efficiency, all

the children had increases in frequencies correct in their

Dvorak phases while two children had decreases in frequencies

correct during a QWERTY phase.
Analysis of the celerations of frequencies incorrect also

provides some support for using the Dvorak keybcard. For

example, the ratio of errors between the QWERTY and the Dvorak

keyboards on the posttests was 11 to 1 in favor of the Dvorak.

13
 Dvorak and QWERTY Keyboards

13

In addition, data from the Direct Copy program indicated that

James's lowest accelerations incorrect always occurred on the

Dvorak keyboard. Frequencies incorrect by the other children

provided little support for either keyboard. One reason for

this may be that key pressing errors are difficult to interpret

in the Direct Copy data. Errors may have occurred when children

inadvertantly pressed a key while the fingers were in the

"touch" position, leaned on the keys, held a key down too long

resulting in a repeated letter, or struck a key while waiting

for the next word presentation. In addition, children sometimes

struck a key after the word was presented but before the key

press was registered by the computer, causing the child's second

key press to be recorded as the child's first key press.

The tact that median frequencies correct and the step size

for frequencies correct tended to increase across phases

suggests an order effect that would tend to mask the differences

between the keyboards. A history of instruction on one keyboard

may have facilitated the acquisition of skills on the other

keyboard up to the point at which conditioning to one keyboard

interferred with performance on the other keyboard. In the

first two phases, when children used hunt-and-peck typing with

scanning, there was little interference from the previous

keyboard. At this point, the children basically may have been

acquiring greater fluency in scanning and identifying key

locations. By the third phase, the students benefitted from

14
 Dvorak and QWERTY Keyboards'

14

retraining on the first keyboard. In the fourth phase, the

retraining on the previous keyboard and the acquisition of

conditioned responses may have interferred with the next

retraining.

Because our study found no evidence favoring the initial use

of the QWERTY keyboard by young children, the argument for

introducing young children to the Dvorak keyboard from the

beginning receives further support. Indeed, all the evidence in

favor of one keyboard over the other, modest as it is, is in

favor of the Dvorak keyboard. Stronger evidence in favor of the

Dvorak keyboard would be expected when children acquired more

touch typing skills.

Even though the children in this study did not acquire touch

typing proficiency, they did keep their hands positioned on the

home row when they began typing and showed some conditioning to

key locations. Their fingering, however, was often inconsistent

and they continued to glance at the keyboard for confirmation of

key location. Their acquisition of touch typing skills would

probably have been more extensive if they had stayed with the
same keyboard. Although there is some evidence that children as
young as six years can acquire touch typing (Kaake, 1983), it is
unclear what the recommended procedures should be for

introducing young children to keyboard fingering (Hoot, 1986).

Additional research in this area acquires increasing importance

as more young children use microcomputer keyboards.

 Dvorak and QWERTY Keyboards

15

References

Barlow, D. & Hersen, M. (1984). Single case experimental

designs. New York: Pergamon.

Best, J. W., & Kahn, J. V. (1989). Research in education (6th

ed.). Englewood Cliffs, NJ: Prentice-Hall.

Dvorak, A., Merrick, N. L., Dealey, W. L., & Ford, G. C. 1936).

Typewriting behavior: Psychology applied to teaching and

learning typewriting. New York: American Book Co.

Hoot, J. L. (1986). Keyboarding instruction in the early grades.

Early Childhood Education, 63, 95-101.

Joyce, B. 3., & Moxley, R. A. (1988). August Dvorak (1894-1975):

Early expressions of applied behavior analysis and precision

teaching. The Behavior Analyst, 11, 33-40.

Kaake, D. (1983). Teaching elementary age children touch typing

as an aid to language arts instruction. The Reading Teacher,

36, 640-644.

Kazdin, A. E. (1982). Single-case research designs: Methods for

clinical and applied settings. New York: Oxford University

Press.

Moxley, R. A. (1986). Wordpictures: Personalized programs for

early literacy, The Ohio Reading Teacher, 20, 11-16.

Moxley, R. A., & Barry, P. J. (1985). Spelling with LEA on the

microcomputer. The Reading Teacher, 39, 267-273.

16
 Dvorak and QWERTY Keyboards

16

Moxley, R. A., & Barry, P. J. (1986). Developing microcomputer

programs for early literacy. (ERIC Document Reproduction

Service No. ED 182 465). Morgantown, WV: West Virginia

University, College of Human Resources and Education.

Pennypacker, H. S., Koenig, C. H., Lindsley, 0. R. (1972).

Handbook of the standard behavior chart (preliminary ed.).

Kansas City, KS: Precision Media.

Yamada, H. (1980). A historical study of typewriters and typing

methods: From the position of planning Japanese parallels.

Jot'nal of Information Processing, 2, 175-202.

17
TABLE 1

Phase Celerations, Median Phase Frequencies and Phase Steps

of Frequencies Correct

Child QWERTY Dvorak QWERTY Dvorak QWERTY Dvorak

Phase Celerations

Julian x1.0 x1.3 xl.' x1.5

Beth x1.4 41.0 x1.5 x1.2

Kathy x1.3 x1.4 x1.0 x1.2

James x1.5 x1.4 x1.9

Median Phase Frequencies

Julian 20.88 22.97 45.11 63.83

Beth 10.22 16.38 22.77 20.58

Kathy 20.58 28.26 41.82 30.27

James 21.46 31.76 33.66

Phase Steps Q/D Q/D Q/D Q/D Q/D

Julian +1.2 +1.5 +2.3

Beth +1.1 +1.2 +2.3

Kathy +1.8 11.1 +1.7

James +2.1 1.2.8

18
TABLE 2
Posttests glowing Frequencies Correct (FC), Frequencies Incorrect (FI) and Errors

Dvorak QWERTY Dvorak QWERTY Dvorak

FC 54.49 35.74 58.28

FI 1.28

Error/correct WAALKING/WALKING
GRIL/GIRL
SEEL/SEES

FC 11.16 11.9 13.17 18.54

FI 0.19 0.88 0.22

Error/correct WALKNG/WALKING IISSSSSSSSSSSS/IS TEE/THE

WSSLKING/WALKING
WATH/WITH
SEESD/SEES
BOY/BOY THH/THE

FC 7.07 24.26 31.94 35.77

FI

Error/correct

FC 14.80 16.42 25.41

FI

Error/correct

19
 Figure Captions
Figure I. Frequencies correct and incorrect for each

child during the experimental phases for the QWERTY and

Dvorak conditions. Celeration lines are drawn for each
phase.

2U
 MOVEMENTS PER MINUTE
Alb. _u Alb
..16 ul a _. at _u 01
V+ al a cl cs a co a
0 in -. cm a 0 00
I I I I I I I I I I I I I I I I

g
§