In the spring of 1982, the faculty in the Colorado State University Intensive English Program began a pilot project to determine the feasibility of using the WRITER'S WORKBENCH (WWB) text-analysis programs with English as a Second Language (ESL) students. Pre-and post-writing samples administered during the pilot project indicated that ESL writers were capable of using and benefiting from WWB (Reid et al, 1983). The following year, the Exxon Education Foundation funded extended research of international student writers and WWB. Four sections of a one-semester, noncredit "grant class" were offered, and student participants spent a minimum of one hour a week in the computer lab, typing their papers into the computers and receiving WWB output. Pre- and post-test holistic scores of writing samples (two topic tasks, one a description and interpretation of a graph, the other a comparison-contrast expository task) demonstrated that there was statistically significant improvement in the writing of students who participated in the class (see Table 1).

p. 54

	    Participants			 	     Contrasting Group

	pre-Graph	post-Graph 			pre-Graph 	post-Graph

	x = 3.78	x = 4.44 			x = 4.82	x = 4.36
	t-value = 3.11 at < .005 			t-value = -.9847 (n. s.) 


	pre-Other 	post-Other 			pre-Other 	post-Other 

	x = 4.64	x=5.14				x = 5.0		x = 4.68	
	t-value = 3 78 at < .005			t-value = -.9697 (n. s.)

        	 N=25              					N= 11

   *6-point holistic scale    				(Reid 1986)

Table 1: Matched t-Test of Holistic Scores for Pre- and Post-Writing Samples of ESL Students

Exxon project investigators also examined the teaching potential of the WWB programs by identifying the following:

1. General pedagogical advantages of using word-processing and text-analysis programs with ESL composition students;

2. ESL student writing problems that could be addressed directly by WWB programs;

3. WWB programs that needed to be adapted to make them more responsive to ESL writers;

4. Teaching methods and materials that could be used successfully in conjunction with WWB programs;

5. WWB programs that are most (and least) effective for ESL student use.

p. 55

Both teachers and ESL students benefit from the use of the word processors and WWB text-analysis programs. First, students are and highly motivated; even non-typists can master the keyboard and the function keys in an hour of on-line tutoring. Moreover, both the students' high interest and the motivation persist. In the Exxon study, pre- and post-affective questionnaires administered. At the end of each semester, more than half (59%) of the students believed that their writing skills had improved because they had used WWB; 87% thought they would be able to use their computer experience in their writing for other university classes, and 93% found that using WWB programs was an enjoyable experience. From the ESL teacher's perspective, the computer lab was valuable in another way as well; it gave the ESL student the opportunity to work side-by-side with native speakers of English, to interact with the lab monitors, and to become a part of the academic "society."

		   ***SPELLING***

	Possible spelling errors in your file are:

 	 	  simpel    negativ



		   ***VAGUENESS***

 	In this text. 7.6 percent of the words are vague

Below is a list of the vague words In your text Delete where possible
Otherwise.  replace or supplement with specific words, phrases, or
examples so that your index is as low as possible.  preferably below 3%. 

all		going		go		go		make
all		going		go		go		make
go		go		good		good		quality
go		some		very

p. 56

Opportunities for revision are perhaps the greatest advantage of the WWB programs. Generally speaking, ESL students are both inexperienced editors and novice revisers. Indeed, many of these students have never written a rough draft, never moved or significantly changed parts of their prose. Working with their writing is a new concept; teaching these students to revise, to `look again," to respond to their audiences classmates, computer text-analysis, teachers--is a primary focus of the course. During the Exxon project, students worked on their writing tasks in class (alone, in pairs, or in small groups), at home, and in conferences; gradually they began to learn, through practice, basic composing and revising techniques. Because of students' lack of experience, both with composing processes and with word processing, and because the computer lab was relatively small, they were required to bring a draft of their writing to the computer lab. (2) After they entered their writing, they ran WWB programs like SPELL (which identifies, but does not correct, possible spelling and typographical errors) and PROOF (which displays the text as it will appear in the draft) before they printed their drafts.(3) Then they printed a paper copy of their writing and took it home or to class. The next assignment was to make revisions according to the WWB suggestions and/or the suggestions of their peer editors, and then perhaps return to the computer lab to revise more completely before turning in the paper. Finally, after the teacher evaluated their prose, the students returned to the computer lab, called up their files, once again made corrections and revisions, and then ran a final draft.

Each time students worked with their writing, some learning occurred; initially the changes were primarily cosmetic, but as the semester passed, most students became aware of the opportunities for higher-level revisions. The ease of revision offered by the computer, and the clean, professional look of each draft, further encouraged revision. For teachers, the advantages were obvious: they read prose that had been drafted, revised, and then clearly typed.

Exxon project investigators identified three (overlapping) areas of ESL student writing that interfered with successful communication:

p. 57

form (organization), content (supporting detail), and mechanics (grammar, spelling, and sentence structure). Of course, the diversity of cultural and educational backgrounds among the ESL students complicated the investigation. The problems demonstrated by Arabic writers, for example, are often different from those experienced by Chinese or Spanish writers (Reid, 1984). However, most students, regardless of language background, needed experience in presenting written material that fulfilled the expectations of U.S. academic readers. Specifically, the Western organizational style of general statement (or heading), use of specific detail (supporting material), and the use of cohesion devices (transitions, repetition of key ideas) were writing techniques that needed explanation and practice. In addition, much of the ESL student writing was mechanically flawed by spelling, verb tense, and vocabulary errors, and a great percentage of the sentence structures were simplistic and/or flawed.

Some of the ESL language problems can be identified and solved by the WWB programs. For example, DICTION flags incorrect or overused vocabulary, and SUGGEST gives alternatives for solving those language problems. Additional assistance for ESL student writing has been incorporated into the DICTION/SUGGEST programs by adding common ESL errors to the computer dictionary (see Table 3). The FINDBE program, which flags excessive use of "to be" verbs and suggests more concrete verbs, was adapted for ESL writers by directing the students to monitor the capitalized "to be" verbs for subject-verb agreement. Other additions to be made in the GRAMMAR program include flagging incorrect infinitive verb endings and adjectives that end in "s" in order to assist ESL students in identifying their errors.

Other WWB programs are more useful as teaching tools in the ESL classroom. For example, the ORGANIZATION program (which prints

p. 58

			***FINDBE***

(Check All Capitalized to be Verbs for Subject-Verb Agreement)

	Example:

	Therefore, the saying refers to one thing that IS surprisingly
	exceeding the other, and the person who uses this saying IS
	usually holding a negative attitude. 



			***SUGGEST***

at which time:			when 
this: 				used with or in place of a singular noun
which is: 			omit or revise as necessary
couple of (informal): 		two, a few, several
date:				always a plural noun
listen:				the verb listen is usually followed by to



			***GRAMMAR***
		    (Instructions to WWB)

TO + verb that ends -ing or -ed (exceptions:  ring, sing, bring)
Adjective + s (exceptions:  various, curious, gracious)



			***GRAMMAR SUGGEST***
		      (Suggestions to student)

To + verb:       takes the simple form of the verb (to run, to study)
Adjective + s:     adjectives in English are always singular.

the first and the last sentence of each paragraph) and the TOPIC program (which prints a list of the most frequently used content words in the student text) are not regularly used by composition teachers of native speakers of English (N5); however, they are very helpful for teaching ESL students Western organization, style, and cohesion. By using several student ORGANIZATION and TOPIC programs as an exercise, the teacher can show students the necessity of clear thesis statements and topic sentences, the "reflecting" quality often evident in a concluding sentences, and the signals given to the reader about the focus of the text by the repetition of key words and phrases. In addition, such exercises are also experiences in audience expectation: "What do you expect this paragraph/essay to be about?

p. 59

What do you expect the writer to do?" Practice in audience expectation teaches reading as well as writing skills.

Still other WWB programs function both as revision tools and as catalysts for teaching. The PROSE program, for example, has subprograms that give the students quantitative information about their texts and suggestions that may improve their prose (see Table 4). The same is true for the STYLE program: students receive information about their texts and suggestions for improvement if the standards of such subprograms as Average Sentence Length, Sentences Types, and Sentence Beginnings are exceeded. The standards, of course, can be set by the programmer; at CSU, allowable percentages for the PROSE and STYLE programs are presently being re-evaluated and set for ESL students (as well as for technology majors, first-year composition students, business majors, etc.).

In light of the information they receive from WWB programs, students often provide the teacher with opportunities to present specific materials. For example, as the teacher explains the VAGUENESS program (a dictionary of vague words, e.g., some, thing, very, people), the question invariably arises: "How can I lower the percentage of vague words?" A lesson on concrete detail and on the substitution of specific for vague words is practical because the information is immediately applicable. In a similar situation, the DICTION/SUGGEST or the Passive Voice program may elicit questions about avoiding "you" or the appropriate use of passive voice. Finally, questions about STYLE or PROSE often spur successful learning experiences: "How can I get more complex sentences? Why is my Readability score only 4.2? What does it mean that 98% of my sentences begin with a subject?" The teacher thus becomes a dispenser, rather than an inflicter, of information while students, eager to "beat the machine," practice sentences combining and clause manipulation with a vengeance. Such practice increases their writing proficiency and gives students the necessary confidence to attempt more qualitative revisions of their prose. Of course, discussion of the computer programs is only a part-- sometimes a small part--of the ESL writing class. Depending on the needs of the students and the desires of the teacher, the computer lab can function as an integral part of the class, as an extension and an expansion of classwork, and/or as a simple word-processing convenience. However the lab is used, it is essential that

p. 60

			***PROSE***

VARIATION

  Variation in sentence length, type, and openings prevents monotony. 
More important, a lack of variation suggests that every topic and every
sentence has equal weight-a deficiency making it difficult for the
reader to pick out the important points. 
  In this text 73% of the sentences are simple, 27% are complex,
giving a difference of 46.  This difference should range from -40 to 10
for good papers of this type. 

  Nominalizations
  You have appropriately limited your nominalizations (nouns made from
  verbs, e.g., "description."


			***STYLE***

sentence information:
	av sent 5 words shorter than av:  23%
	% of sent 10 words shorter than av: 8% 
sentence typed:
	simple: 46% 		complex:  53%
verb choice:
	to be as percent of total:	24% (8) aux 24% (8)
					Inf 15% (5)
     	passives as % of non-inf verbs:  4% (1) 
	nominalizations:  2% (5) 
sentence beginnings:
	subject openers:  noun (1) pron (1) pos (0) 
			  adj (1)  art (1)  TOTAL 31%
	other openers:	prep 23% (3)	adv 23% (3)
			verb 0% (0)	sub-conj 23% (3)
			conj 0% (0)
other information:
	no. sent:  13			no. wds:  285 
	av word length:  4.53 
	no. questions: 0 		no. imperatives:  0
	no. content wds: 166  58.2%	av. length:  5.72
word types as % of total
	prep 9.1 % (26)	conj 2.5% (7)
	adv 9.8% (28)	noun 21.4% (61)
	adj 16.1% (46)	pron 7.4% (21)
readability
	(Kincaid) 12.1	(auto) 10.9  (Coleman-Liau) 9.5
	(Flesh) 13

p. 61

the teacher informs the students at the beginning of the course about the limitations of the text-analysis programs. Just as important, the teacher should assure the students that they are free to ignore the computer's suggestions without hurting its feelings. Indeed, the opportunity to examine the quantitative measures and the programmed suggestions critically, and then to decide what changes to make, is fundamental to the revision process.

There are, of course, some WWB programs that have only limited use for ESL students (see Table 5). CHECK, for example, lists "problem words" for native speakers of English (e.g., then/than) that rarely plague ESL writers; such a program can confuse ESL writers rather than illuminate ESL problems. The PUNCTUATION program only counts quotation marks and apostrophes, so it is useless in its present form. The GRAMMAR program, at present, only identifies split infinitives, an error that ESL students almost never make; the Nominalization subprogram flags nouns made from verbs, another "error" rarely encountered in ESL prose. The ABSTRACT program is a dictionary that flags many words considered abstract by teachers of native speakers; however, ESL writers have trouble replacing/eliminating those words successfully, and their energies could well be used elsewhere. At CSU, modifications to make the programs more viable for both ESL and native English speakers are underway and will continue: a limited comma splice program, a third-person singular program, and a sentence combining program.

For both ESL and native-English composition students, WWB can be a successful learning and teaching tool but even such a complex program is limited by its parser (which identifies parts of speech in text), by the people who program it, and by the enormous complexity of the English language. WWB, for example, cannot identify subject-verb agreement, and will probably never be 100% correct in any of its programs. More important, perhaps, is the inability of WWB to judge quality writing; it gives students quantitative, not qualitative, measures. As a result, the text-analysis programs serve as tools for teachers, not as teachers themselves.

p. 62

Very Useful		Somewhat Useful		Virtually Useless
_________________________________________________________________
ORGANIZATION		CHECK
TOPIC						PUNCTUATION
PROSE			PROSE			PROSE
  Variation		  Passives		  Nominalizations
STYLE			STYLE
  Sentence Length	  Verb Choice
  Sentence Types	  Word Type
  Sentence Beginners	  Word Length
  Readability		  Content Words

SPELL			ABSTRACT		GRAMMAR*
DICTION
SUGGEST
VAGUENESS
FINDBE

*Until new programs are developed.

Table 5: Usefulness of WWB Programs for ESL Writers

The limitations of WWB are in some ways its strengths. In order to learn about writing, students must be able to look carefully at their writing in various lights, from a variety of perspectives, and to make decisions about improving their writing. And students must have human experts, teachers, to teach them about quality writing. The benefits of WWB lie in its capability to provide another audience, to offer information to the students that would be too time-consuming for the teacher to collect, and to motivate the students to look again (and again) at their writing. The students, rather than the computer or the teacher, ultimately become responsible for their writing. The partnership between student and teacher, and the relationship between students and their writing, is strengthened by the addition of computer text analysis.

p. 63

1. CSU's online LEARN program, developed by Charles Smith, has been remarkably successful and efficient in teaching students to use the word-processing program.
2. At present, with an expanded Computer Laboratory and with ESL students who are becoming more experienced computer users; students are composing--not just revising--in the lab.
3. The titles of individual WWB programs are in small caps; subprograms have their initial letters capped.

Reid, Joy. (1984). English composition: The linear product of American thought. College Composition and Communication, 35(4), 449-452.

Reid, Joy. (1986). Using the Writer's Workbench in composition and teaching. In Technology and Language Teaching, H. Doug Brown (Ed.). Washington, D.C.: TESOL.

Reid, Joy, M. Lindstrom, M. McCaffrey, and D. Larson. (1983). Computer-assisted text-analysis for ESL students. Calico Journal, 1(3), 40-46.