In order to write and revise effectively with a word-processing program on a computer, one needs to know how to perform the physical procedures involved in altering text. Yet, skill at text editing has been virtually ignored in studies on word-processing programs and writing. Surprisingly, studies on word processing and revision indicate that although word processors facilitate surface-level revision, they are not associated with increases in meaning-level changes to texts (Collier, 1983; Daiute, 1986; Harris, 1985; Hawisher, 1989; Lutz, 1987; Womble, 1984). It is possible that limited skill at text editing might play a role in such findings. For instance, rather than making the physical operations of revision easier, as many authors assume, complicated text-editing commands that require two or three steps to complete (such as "cut and paste") may be difficult for students to perform. Lack of skill at such complex text-editing commands may restrict students to surface revision, or it may be the case that students have mastery over only those text editing commands that support their frequent practice of surface revision.

In order to examine the relationship between revision and word processing, it is necessary to know whether students can use text editors effectively. Researchers often assume that students who use word-processing software for extended periods of time become proficient at text-editing, even in the absence of direct instruction . However, recent research suggests that unless students are directly instructed in keyboarding and text editing, they are not likely to have mastered these skills (Britten, 1988; Larter, Branganca & Rukavina, 1987). Larter, Branganca, and Rukavina (1987) questioned students in grades one, three, and six about their text-editing and typing practices. 80 %, 50%, and 40% of these students, respectively, said that they used one finger and one hand to type with. About 37% of the grade three and grade six students said they "hunted and pecked with both hands and two fingers," and 22% of grade six students reported that they "hunted and pecked with both hands and more than two fingers" (p. 120). Most students in the study said that their deletions usually took the form of character by character deletion.

On the other hand, when students do receive instruction in typing and text editing, it is not surprising that they become skilled at these tasks. Gerlach (1987) found increases in the typing speed and accuracy of fourth grade students who had received typing instruction prior to using word processing systems.

Similarly, in a study by Britten (1988), second grade students were either instructed in typing or used a computer without specific keyboarding instruction, and only students who had been instructed in keyboarding improved in their typing skills. The results of these studies suggest that instruction in keyboarding would better enable students to make revisions to their texts. Further, instruction in text editing may facilitate both surface-level and meaning-level revisions.

Another possible consequence of weak text-editing skills is that the process of composing becomes cumbersome, leading students to view writing with a word-processing system in a negative light. Much of the research on students' attitudes has focused on their attitudes toward writing itself and not on writing in relation to word processing (Fitch, 1985; Kurth, 1987; Lansing, 1984). While Fitch (1985) and Lansing (1984) found that writers' attitudes toward writing were unaffected by instruction with word-processing systems, Kurth (1987), Teichman and Poris (1985), and Dalton (1987) reported that students felt more positively toward writing and word-processing systems after learning to use them.

Research also indicates that students' use of word-processing systems varies according to the phase of the writing process, and these writing practices may reflect different attitudes students have about writing with word-processing systems for drafts and revision. Bridwell, Sirc, and Brooke (1985) interviewed 48 college students about their writing practices and found that about half of the students used word-processing packages for composition as well as for revising drafts. The other half of the students used computers for preparing or text-editing drafts but did not compose on the computer. Larter, Branganca, & Rukavina (1987) also found that grade six students said they more often transcribed handwritten text on word-processing systems than composed new text, while students in grades one and three used word-processing systems primarily to compose text. Thus, while researchers have examined both students' attitudes toward writing and their writing practices when using computers, what remains unclear are students' attitudes toward writing with word-processing systems for different phases of the writing process, and the relationship between these attitudes and their skill at text editing.

The purpose of the present study was two-fold. The first goal was to investigate the text-editing skills of students who had considerable experience using word-processing systems. As pointed out above, unless directly instructed in text editing, students are not likely to be skilled at various editing commands. Although level of skill at text editing could have a powerful effect on the type of revision students engage in, this factor has been given little attention in studies on word processing and writing. Students in these studies have most often had little or no prior experience with word-processing systems (Collier, 1983; Hawisher, 1989), and their text-editing skills have not been evaluated. In the present study, we assessed the word-processing skills students had acquired in the ordinary course of instruction and practice over a two-year period. We collected descriptive information about average and above-average student writers' text-editing and keyboarding skills on word-processing systems. Our investigation is intended as a preliminary step in the study of word processing and revision.

The second goal of the study was to investigate students' attitudes toward word processing for different phases of the writing process. The findings on attitudes discussed above suggest that students' preferences for word processing or paper and pencil writing may differ for composing and revising. In order to examine this possibility, we interviewed grade eight students with respect to their attitudes toward writing with word-processing packages for both composing and revising. In addition, we interviewed samples of graduate students and professional writers to see whether the text-editing skills and attitudes of more experienced writers differ from those with less writing experience. Further, because we assessed both the text-editing skills and attitudes of the grade eight students, it was possible to relate these measures to determine whether students who were more skilled at text editing held more positive attitudes toward word-processing programs.

Subjects
The participants in our study were 14 male and 15 female grade eight students who were randomly selected from 3 classes in a middle-class junior high school. Fifteen student writers had been designated as "enriched" (above-average) writers prior to our study on the basis of teacher recommendations for their enrollment in an enriched writing program. The above-average writing group achieved significantly higher grades (M = 7953, S.D. = 6.95) than the average group on their English grades (M = 71.69, S.D. = 5.69), t(29) = 3.45, p < .01.

According to their teachers and to students' answers reported in a questionnaire, the students had from two to three years of experience working with various kinds of microcomputers in their English classrooms. The students' English teachers also had extensive experience with word processing and had jointly published a student manual on word processing with APPLEWRITER (a word-processing package for the Apple 2+ computer). All students were currently enrolled in a computer literacy course at their junior high school. Students' prior experience with different types of computers varied; and by using the ICON for this study, with which they were all roughly equally familiar, we hoped to minimize the possibility that prior experience would be a confounding variable.

The ICON computer is an Ontario Approved Educational microcomputer. The word-processing program constantly displays at the bottom of the screen small icons that stand for text-editing commands (see Figure 1). For example, if a student wants to cut text, she or he highlights some text, moves a command arrow to the bottom of the screen and positions it over a pair of scissors, and then presses an "action" key. Pressing "action" while the command arrow is on the "paste jar" icon restores the cut text, beginning at the point where the text cursor is currently located.

Procedures
Three experimenters conducted this study, and they were assigned to subjects at random. Students were interviewed and tested individually for approximately 45 minutes.

Students were first informed about the nature of the experiment and were then given a keyboard speed test, (the keyboard speed test was designed and programmed for this study by the second author) which provided a precise measure of their speed at pressing keys. Various letters appeared on the screen, one at a time (including capital letters); when the student pressed the correct key, the computer calculated the difference in time between the presentation of the letter and the key press. Students received 100 practice trials and were then tested on 100 items.

A questionnaire in which students were asked about their previous experience with computers and about their preferences for writing with word-processing systems and with paper and pencil was administered verbally to each student. We asked students questions regarding their preference for paper and pencil or word-processing systems for composing in general, for composing a first draft, and for revising a paper. The general question was always asked first, and the order of the second and third questions was counterbalanced across subjects.

The students were asked to state whether or not they knew how to perform various text-editing commands (such as deleting text, cutting and pasting) on the machine with which they were most familiar (see Appendix). If they responded yes, they were asked to describe how they would perform the text-editing command. The experimenter unobtrusively rated the students on their performance using a three-point scale: they received a 1 if they could not describe how to perform the task at all; a 2 if they could describe how to accomplish the task but not by making full use of the machine's capabilities (for example, moving text by rewriting and erasing rather than cutting and pasting); and a 3 if they described how to perform the task by exploiting the full capacity of the word-processing system.

Students were then asked to perform the same revisions they had just been questioned about on a text which was displayed on ICON, including insert, delete, cut and paste, search, and justify. All students were asked to make the same changes to a standard text (see Appendix). The experimenter again rated the students on their performance using the three-point scale: they received a 1 if they could not perform the task at all, a 2 if they could complete the task but not by making full use of the machine's capabilities, and a 3 if they accomplished the task by taking full advantage of the word-processing system.

Text Editing and Typing Assessments
Based on our analysis of the keyboard speed test, we estimated that students typed at a rate of approximately eight words per minute (this estimate is made by calculating 5 characters per word). We later administered a typing test to another group of students to see if typing speed would be greater if students were typing words rather than single letters. The second group of students had used word-processing systems for 40 minutes in their classroom every day over a one-year period and were, therefore, more experienced with them than the first group. Their typing speed was 22 words per minute which, although faster than that of the first set of subjects, would be considered a slow typing speed for adults. We can conclude that students in our studies typed from 8 to 22 words per minute, with the latter speed representing typing speed under optimal conditions of testing and prior experience. Although we did not examine the way in which students type, our results seem consistent with those of Larter, Branganca, and Rukavina (1987) and suggest that our students' typing abilities were limited.

According to the questionnaire, students reported that they had used the APPLEWRITER 2 program on the Apple 2+ computer for an average of 15 hours over 2.34 years. Their familiarity with the computer used in this experiment (the ICON) was more limited, with an average of 7 hours within 3 months. Students often used word-processing systems for short periods of time (e.g., 15 minutes per English class, several times a week) which is why the number of hours is small relative to the total length of time they used word-processing systems.

When queried about their text-editing skills on the machine that they were most familiar with, more than 75% of the students reported that they knew how to insert; and 93% said they knew how to delete. Thirty to 457O of the students said they knew how to search and replace, to cut and paste, and to justify. Students' responses to these questions reflect the pattern of their performance when they were asked to perform the revisions; although, in most cases, their actual performance was worse than their reported performance. This may be due to an overestimation of their own abilities, or to the fact that they are more competent at text editing with computers that are more familiar to them than the ICON.

When students attempted to edit on the ICON, we found that 66% could insert, 66% could delete, 3% could search, 21% could move text by cutting and pasting, and 7% could justify text. Although 6696 of students could insert and 66% could delete text, when asked to change text by both inserting and deleting (See Appendix, Task 2), only 49% could do so. This suggests that students' cognitive load is increased when they have to perform several text-editing commands to accomplish a task, resulting in decreased performance. Among the seven text-editing commands on which the students were tested, there was only one instance when a student made full use of the computer. The mean rating for skill at text editing was 1.89 (S.D.=.33) for the average students and 2.07 (S.D.=.34) for the above-average students. Our assessment of students' text-editing skills showed no significant differences between the two groups as a function of their skill level (i.e., above-average and average writers) t(27) =-1.45, p > .05. Because these students most likely had a greater amount of experience with word-processing systems than most students their age (due to the accessibility of computers in their school), their skills with word-processing systems ought to be at least typical of those of other students in their age group.

Moving text around by cutting and pasting and searching for text may be important commands for performing meaning-level revision. Students may not have learned, or may have forgotten, the text-editing commands necessary for performing meaning-level revision on word-processing systems because they do not frequently edit their text in this way. Deleting and inserting text may be adequate for carrying out the more commonly performed surface revisions. It is also possible that students will not engage in meaning-level revision on word-processing systems because they lack facility with the full range of text-editing commands. Our results suggest that one reason researchers may not find word-processing systems associated with increases in meaning changing revision is that students do not have the text-editing skills necessary to make such changes. However, it may be the case that students do not engage in much meaning-level revision regardless of the medium they used to write. Because none of the studies we re viewed on word processing and revision assessed their subjects' text-editing skills, it is still an open question as to whether or not the effects of word-processing on revision are mediated by students' text-editing skills.

One of the possible reasons we considered for students' limited skill on the ICON was lack of experience with the word processing. However, according to students' self-reports, even on computers they were more experienced with, less than half the students could describe how to cut and paste, search, or move text. We also hypothesized that the poor text-editing skills of the students were due to their immature writing skills. In order to examine the text-editing skills of more experienced writers, we assessed seven graduate students who had used word processing systems from two to five years. They used an editor called EDT on a VAX mainframe computer. EDT is a full-screen editor; students edit by pressing keys on a numeric keypad. One or two key presses generate a function associated with that key. The keys are not usually labelled with the functions. As with many text-editors, one key press is sufficient to delete character by character, while complex commands such as cut and paste require up to five key presses. In order to format text (such as justifying), commands must be inserted in the text. Non-credit courses on EDT are available to students, but many acquire their skills by means of informal instruction from other students.

We found that some of the graduate students had poor text-editing skills, while for others the process of text editing was completely automatic. One student who had used the word-processing system extensively for over two years said that when she wanted to move text, she recopied the entire section and then deleted the original passage rather than using "cut and paste." Many of these students did not touch type and had to look at the keyboard when typing. One student typed with only one hand, but nevertheless managed to type 32 words per minute. When asked if slow typing speed interfered with his composing process (he wrote and revised his drafts using the word processing systems), he said that he kept a piece of paper and pen on hand on which he would jot down notes if necessary, and that he would sometimes insert single key words in the text on the screen which would then serve as a reminder for a thought which would be developed later.

Clearly there was enormous variability in the text-editing and typing skills of these students. Assessment of a much larger sample of students would be necessary to determine how very experienced writers use text editors. However, we can speculate that the performance and responses of this small sample of students indicate that experienced writers devise ways to carry out "high level" writing procedures, regardless of technical obstacles. However, less experienced writers may be more influenced by limitations in their technological skill because their writing strategies are not as well established.

Interviews on Preferences for Medium of Production
The grade eight students were asked the general questions: "When you are composing, do you prefer to use paper and pencil or word processors? Why?" In answer to these questions, about half the students said they preferred paper and pencil, and half preferred word-processing systems. (See Table 1)

When asked which medium they preferred to write with when composing a first draft, most of the students said they preferred paper and pencil. Among the reasons they cited were: "It's easier to correct," "it feels right," and "it helps to keep ideas straight." One student said: "If I'm thinking about a story in my head, and there's a letter I have to delete, it stops me from thinking about what I'm writing." We categorized and tabulated the reasons students gave for their preferences, and the categories are presented in Figure 2.

For revision, many of the students said they preferred word-processing systems because these machines made it easier for them to correct their work using a computer. Students also mentioned their potential audience, saying either that the paper would be easier to read or that the reader would think the student knew a great deal about the computer.

T-tests were carried out to examine the relationship between the students' text-editing abilities and their preferences for medium of production. That is, we wanted to see if students who were better at text editing would prefer to use word-processing systems rather than paper and pencil. The results indicated that when asked both the general question and the question about writing a first draft, there was no difference between the text-editing abilities of students who reported that they preferred paper and pencil and those who preferred using word processing. However, when asked about their preferences for revision, those students who responded that they preferred using word-processing systems had a significantly higher text-editing skill index (M = 2.08, S.D. =.33) than those students who said they preferred paper and pencil (M = 1.82, S.D. =.32), t (127) = 2.03, one-tailed p = .028.

Chi-square analyses did not indicate significant differences between the average and above-average writers' preferences for medium of production.

To summarize the students' responses to the three questions, when asked which medium they preferred in general, half said they preferred word-processing systems, and half said they preferred paper and pencil. When we asked them more specifically about their medium of preference for writing a first draft and for revision (subsequent drafts), most of them preferred paper and pencil for the first draft; and most of them preferred word processing for revision. Those students who were more skilled at text editing preferred revising with a word-processing system to a greater extent than the less skilled students did. The reasons for students' preferences are presented in Table 2.

Because these students usually write their first drafts with paper and pencil in their classrooms and sometimes revise on computers, their answers may reflect their previous experience. It is also possible that their responses to the question regarding the first draft reflect preferences that are based on their experience of having the flow of their emerging text interfered with by the technical concerns of text editing and typing.

We asked our sample of graduate students about their writing practices on computers with word-processing programs. Most of the graduate students reported that they generally write their first draft by hand, type it into a computer, print out a copy, revise the hard copy by hand, and then make the corrections on the computer. They rarely typed a first draft directly with a computer and seldom made corrections on the screen without first correcting the hard copy. These results corroborate the findings of Tilly and Myers (1988), who found that 61% of their sample of post-secondary students reported usually writing with pen and paper, 29% said they wrote by hand first and then typed the text into a computer, and only 5% used a microcomputer throughout the writing process (the remaining students typed their papers). Teles (in press) found that graduate students who had more than three years of continuous use of computers with word-processing programs were "on-line composers" while those with less experience said they write by hand and then type this text into the computer . Teles attributes this to the fact that writers, in order to become on-line composers, have to change their ways of thinking and develop different cognitive approaches to writing; and this requires a considerable amount of experience.

We also interviewed three professional writers who had been using computers with word-processing programs for two to three years, and these writers reported that they made use of all the text-editing features of word-processing software. In addition, they compose their first drafts directly on the computer; two of the three writers said they revise on the screen without using a hard copy.

One difference between the graduate students and the professional writers is that graduate students use a mainframe computer, most often in a room filled with others writing on terminals. While the terminals are always available to the students, most do not have computers at home. These working conditions might encourage the students to print out a hard copy and edit in the privacy of their offices. The professional writers work at home on their own microcomputers and do not have to seek out private conditions in which to edit their work. Another possible reason for the differences in writing practices of the graduate students and professional writers is that because the graduate students in our study were less experienced with computers and word-processing programs than the professional writers, they may not yet have learned to approach writing differently in the way Teles (in press) suggests is necessary. It is also possible that differences in the writing practices and attitudes of mature writers are due to individual differences in their writing styles. This explanation is supported by findings in a study by Bridwell-Bowles, Johnson, and Brehe (1987), in which experienced writers who had published extensively were trained and used computer with word-processing programs for composing and revising over eight two-hour sessions. While some of these writers liked composing with computers, many of them did not; and the researchers attribute the "mixed reviews" to differences in the subjects' writing styles.

The results of our tests and interviews reveal several characteristics about students' skills and attitudes toward computers which are relevant to research and to instruction on writing with word-processing programs. First, the students in our study were poor at text editing and typing, and seemed to have settled into using an expensive piece of equipment as little more than a typewriter. It is possible that because students do not often engage in meaning-level revision, they do not require extensive text-editing skills to support their practice of surface revision. Thus, even when they are instructed in a range of text-editing skills, they may forget the ones which are necessary for high-level revision. However, some pilot data we collected suggests that even graduate students, who are sophisticated writers, fail to take full advantage of computers with word-processing programs after several years of experience with them. Clearly, we must design ways to help students develop the strategies and skills they require to work effectively with this new medium. That is, students should be taught text-editing and typing, with a focus on the functions that are more difficult to learn such as cut-and-paste and search-and-replace. Students' competence at these skills should be assessed periodically to ensure that they are not using inefficient procedures. If young students are not skilled at text editing, they may be distracted from high-level writing concerns by the amount of attention they must devote to text editing. This may explain why many students can only make use of the insert and delete functions, because these are the simplest operations and require the least amount of attention.

Our second finding was that our grade eight students generally prefer paper and pencil for composing first drafts and prefer word processing for revision only. At present, it is not clear what these preferences result from; they could be due to habit, to students' recognition that their poor text-editing skills may interfere with their composing process, to lack of access to computers, or to difficulty with the word-processing program used by the students.

Those conducting research on the effects of word-processing pro grams on revision and on the composing process in general would do well to note our findings. Researchers studying this topic should assess the word-processing competence of their subjects to ensure that all subjects are reasonably sophisticated in their use of text editors. Alternatively, level of text-editing skill could be included as a factor in the study, with several levels of skill represented. An experimental design such as this would help to determine what role a proficiency in word processing may play in the relationship between computers and writing. ~

Elana Joram, Earl Woodruff, Peter Lindsay, and Mary Bryson are researchers in the Centre for Applied Cognitive Science, Ontario Institute for Studies in Education.

Bridwell, L., Sirc, G., & Brooke, R. (1985). Revising and computing: Case studies of student writers. In S. W. Freedman (Ed.), The acquisition of written language: Response and revision (pp. 172-195). New Jersey: Ablex.

Bridwell-Bowles, L., Johnson, P., & Brehe, S. (1987). Composing and computers: Case studies of experienced writers. In A. Matsuhashi (Ed.), Writing in real time: Modelling production processes (pp. 81-107). New Jersey: Ablex.

Britten, R. M. (1988). The effects of instruction on keyboarding skills in Grade 2. Educational Technology, 28(4), 34-37.

Collier, R. M. (1983). The word processor and revision strategies. College Composition and Communication, 34, 149-155.

Daiute, C. (1986). Physical and cognitive factors in revising: Insights from studies with computers. Research in the Teaching of English, 20, 141-159.

Dalton, D. (1987). The effects of word processing on written composition. Journal of Educational Research, 80(6), 338-342.

Fitch, J. E. (1985). The effect of word processing on revision and attitude toward writing. Unpublished master's thesis. National College of Education, Illinois.

Gerlach, G. J. (1987). The effect of typing skill on using a word processor for composition (Report No. IR 012 797). Baltimore, MD: Johns Hopkins University, Center for Social Organization of Schools. (ERIC Document Reproduction Service No. ED 286 465)

Harris, J. (1985). Student writers and word processing: A preliminary evaluation. College Composition and Communication, 36, 323-330.

Hawisher, G. (1989). Research and recommendations for computers and composition. In G. E. Hawisher & C. L. Selfe (Eds.), Critical perspectives in computers and composition instruction. New York: Teachers College Press.

Kurth, R. J. (1987). Using word processors to enhance revision strategies during student writing activities. Educational Technology, 27(1), 13-19.

Lansing, M. L. (1984). Student writers and word processors (Report No. CS 208 577). Baltimore, MD: Johns Hopkins University, Center for Social Organization of Schools. (ERIC Document Reproduction Service No. ED 249 491)

Larter, S., Branganca, R., & Rukavina, I. (1987). Writing with microcomputers in the elementary grades: Process, roles, attitudes, and products. Toronto: Queen's Printer for Ontario.

Lutz, J. A. (1987). A study of professional and experienced writers revising and editing at the computer and with pen and paper. Research in the Teaching of English, 21, 398-421.

Teichman, M., & Poris, M. (1985). Word processing in the classroom: Its effects on freshman writers. (Report No. CS 210 196). Baltimore, MD: Johns Hopkins University, Center for Social Organization of Schools. (ERIC Document Reproduction Service No. ED 276 062.)

Teles, L. (in press). The adoption of word processing by graduate students in education. Education and Computing.

Tilly, A., & Myers, P. (1988). The influence of word processing on the attitudes and writing of postsecondary students. Toronto: Queen's Printer for Ontario.

Womble, G. G. (1984). Process and processor: Is there room for a machine in the English classroom? English Journal, 73(1), 34-37.

Assessment Procedures and Materials
All students are shown a computer screen that displays the Text A below. Each student is asked to change the text so that it looks like text B, which is given to the students on a separate piece of paper (this is why there are typographical errors in Text A).

Text A. The sponge crab carefully cuts a sponge it has found to the proper size and, with its hind claws, pushes the sponge onto its back and clamps it into place. The crab moves about easily, even though the sponge remains on its back. The sponge continues to live and grow there comffortably. As it grows, it forms a thicker and thicker protective cap for the crab. If the crab loses its sponge, it begins to search frantically for another one to replace it. Scientists have studied the sponge crab in laboratory aquariums. Once the crab loses its sponge, it appears to be very upset. Then, if the scientist puts a single piece of paper into the water tank, the crab will quickly trim the paper to the proper size and boost it onto its back. Early in life, the sponge crab searches for a sponge to carry around on its back.

Text B. Early in life, the sponge crab searches for a sponge to carry around on its back. The sponge crab carefully cuts a sponge it has found to the proper size and, with its hind claws, pushes the sponge onto its back and clamps it into place. The crab moves about easily, even though the sponge remains on its back. The sponge continues to live and grow there comfortably. As it grows, it forms a thicker and thicker protective cap for the crab. If the crab loses its sponge, it begins to search frantically for another one to replace it. Scientists have studied the sponge crab in laboratory aquariums. Once the crab loses its sponge, it appears to be very upset. Then, if the scientist puts a single piece of paper into the water tank, the crab will quickly trim the paper to the proper size and boost it onto its back.

Ask the student if they know how to:

• delete the words "carefully" and "frantically";
  
• change "studied the sponge crab in laboratory aquariums" to "studied the sponge crab in the lab";
  
• search for "about";
  
• change "about easily" to "about quite easily it seems";
  
• find and replace "comffortably" with "comfortably";
  
• move the last sentence up to the first sentence spot;
  
• justify the text.
  

Once the students have described what they would change and how they would change it, rate the students as follows:

1. don't know what to do
2. describes how to make the change, but by using strategies such as copy and delete for cut and paste
3. describes the appropriate text-editing commands
   

Then, ask the students to demonstrate the seven text-editing commands listed above, and record the success or failure of each change, and the strategy used (for example, note "copy and delete to move sentence," or "used cut and paste" to make the move) . Rate the students (1=don't know what to do; 2=make the change, but by using strategies such as copy and delete for cut and paste; 3=make use of the appropriate text-editing commands).