Article on Research That Supports the Hunter Writing System: Sentence Sense Text*

For her M.A. thesis at Rowan College of New Jersey, Barbara Stubbs (1995a), a teacher at Clearview Regional Junior High School in Mullica Hill, N.J., carried out an experimental study in which my text The Hunter Writing System: Sentence Sense (Hunter, 1991) was the instructional instrument for the experimental group.  Her goal was to prove the text's efficacy for improving students' ability to revise and edit.

Since the results of the study bore out her hypothesis in a convincing way, I have taken it upon myself to do the reporting for her (she deferred to me because she is presently pursuing a doctoral degree).

My goals here will be to relate this text to the current discussion of grammars, to cite some misreported research that clearly supports instruction in grammar to help writing, to report the results of Stubbs' study, to report the nature of the experimental instruction, and finally to indicate some of the ways in which my Sentence Sense text material differs markedly from traditional grammar.

the need for a better grammar

In his article "Grammar Making a Comeback in Composition Teaching" in last October's issue of Composition Chronicle, Bill McCleary (1995) mentions a current "search for a grammar that works better for students than the other grammars available" (2).  He points out that "What we may need . . . is a `pedagogical grammar' that is both accurate and simple" (2).  He explains why:

Traditional grammar--the grammar taught in English textbooks--is inaccurate both in its description of real grammar and in the explanations given to students. . . . And the other grammars, scientific ones like descriptive, structural, generative, and the like, are too technical for use in teaching (2).

By "pedagogical grammars," he means "grammars intended to be used in teaching writing"--a kind of grammar which, he suggests, could be added to Patrick Hartwell's (1985) classification of basic approaches to grammar (4).

In particular, McCleary alludes to four pedagogical grammars that are now published or proposed (2). One of these four, one that he cited and made brief comments on as part of the article (2, 4), is the very textbook chosen by Stubbs as the strategic part of the treatment for her experimental group, Hunter's Sentence Sense (1991).

research that Supports the teaching
of grammar as an aid to composition

Two research studies unmistakably support the teaching of grammar as an aid to writing.

In the study of Donald Bateman and Frank Zidonis (1964), a grammar based on the generative approach of Noam Chomsky (1956), Robert Lees (1960, 1957), and Charles Fillmore (1964, 1963, 1962) was developed by the investigators in order to teach the process of sentence formation in relation to composition writing.  The experimental class was taught this special grammatical material during their ninth and tenth grades.  The control group was reported to have studied no formal grammar.  This study was exemplary not only in the duration of the instruction time (two years) but also in its controls (both students and teachers were randomly assigned) and in the quantity of writing analyzed (the first six and the last six compositions of every student).

Bateman and Zidonis (1964) reported that the improvement in writing of the experimental group was statistically significant at the .01 level of confidence in all three of the areas under investigation--in the quality of the sentences written, in the complexity of the sentences written, and in decrease in errors.  In addition, in the area of quality of sentences written, the difference in gain scores of the experimental group over the control group was also statistically significant at the .01 level of confidence.  The researchers also found a statistically significant greater improvement in the complexity of all sentences written by the experimental group over those written by the control group but at the .05 level of confidence and due, primarily, to the complexity of the writing of just 4 (of the 21) students.  However, had they contrasted just the complex sentences of the two groups and studied the degree to which the experimental group's sentences were better or less well formed than the control's--and not all the sentences to see which group's sentences were the more complex--they would have discovered a far more significant difference.  For the control group, 48% (3,141 out of 6,582) of their complex sentences were malformed,  whereas--for the group taught transformational grammar--only 9% (510 out of 5,534) of their complex sentences were malformed.  In other words, when the group taught transformational grammar wrote complex sentences, 91% of their sentences were well-formed, whereas when the group taught grammar informally wrote complex sentences only 52% of these students' sentences were well-formed (1966, 19­20).  [The latter part of this paragraph has been edited by the author.]

By the way, George Hillocks (1986) misreported this study by stating that "none of the studies reviewed for the present report provides any support for teaching grammar as a means of improving composition skills" (138), though earlier in the same chapter of this report he had stated, "[Bateman and Zidonis] did find a statistically significant difference between the number of well-formed sentences written by the experimental students and their control group counterparts (136)."  Moreover, Hartwell (1985) dismissed this study as an early experiment in sentence combining (127) though it is obvious from a review of the Bateman and Zidonis study (1964) that it was the instruction in transformational grammar--and not its being keyed to sentence combining practice (something not even mentioned in the study)--that caused such statistically significant improvement in the writing of the experimental students in all the areas under investigation.

In the Charles Thompson and Morris Middleton study (1973), a sequel to the Bateman and Zidonis study, teachers were randomly assigned to four sections of tenth graders (each with 23 students).  Half the students were taught transformational grammar and half were taught traditional grammar.  In all classes the students were required to apply the new concepts in their writing.  The instruction lasted for one semester. Thompson and Middleton report, "There was a significant difference at the .05 (and .01) [italics added] level between the means of the distributions of the Before and After structural complexity scores" (38­39).  Moreover, this was true for both sorts of treatment groups.  Though the groups taught transformational grammar had somewhat better structural complexity scores than those taught traditional grammar, that difference was not statistically significant.  What was statistically significant was the degree of improvement in structural complexity caused by both kinds of instruction in grammar.  As an aside, it might be noted that Hillocks (1986) dismissed this study by stating: "[Thompson and Middleton] found greater gains for students studying transformational grammar than for those studying traditional grammar.  However, the differences were not [statistically] significant" (136).

the results of Stubbs' study

In the Stubbs study (1995a), both the control and the experimental students had learning disabilities and were classified as "special education" students.  The mean I.Q. of the 10 students in the control group was 100; the mean I.Q. of the 10 students in the experimental group was 93--and the (slight) disability of these latter students was in the area of language. Instruction was carried out for six months.  Both groups were taught writing by the process approach, but the control group was also taught traditional grammar while the experimental group was given the added instruction of 10 chapters of the Hunter Sentence Sense writing program (1991).

The pre- and posttests given were a 15-minute writing sample based on the writing prompts and scoring criteria for the Test of Written Language­2 (TOWL­2) (Donald Hammill and S.C. Larsen, 1988).  The scores contrasted were the test's mean NPR percentile ranks. The investigator, who did the scoring of all writing samples, took the greatest care possible to score objectively and consistently (1995a, 40).

The reported gain/loss scores for the experimental group were as follows (1995a, 50):

A few things need to be noted here.  First, these students used up a few minutes of their writing time doing revising and editing; this adversely affected both the number-of-words-written score and, indirectly, the overall-writing-competence score--which is a composite score that is biased in favor of the number of words written.  Second, these students had not reached the chapter of the text where complexity (subordination) is first taught.  Third, if the scores of the one student who decided not to continue writing (an attention-deficit youth) were excluded, an overall writing competence gain for the remaining eight students of 67% would be statistically significant at the .05 level of confidence (Stubbs, 1995c, 3).  Fourth, there was another student (who had taken the pretest) who joined the class half way through the six months and was helped with the Sentence Sense material until he caught up the class: if his scores were included while the attention-deficit student's scores were excluded, an overall writing competence gain for these nine students of 59% was statistically significant at the .02 level of confidence (Stubbs, 1995c, 4).

It is apparent from the above discussion that the experimental group improved appreciably not only in their ability to revise and edit but also in their ability to write spontaneously and competently.

The reported gain/loss scores for the control group were as follows (1955a, 50):

A few comments relative to the instruction that the control group received are in order.  According to Stubbs, "Students in the [control group] learned grammar in a traditional manner taught in isolated skill exercises. Writing was taught as a separate skill from within a process approach (pre-writing, organizing, first draft, revising, editing, final draft)" (1995a, 42).  There were ten students in all, seven in one class in which most grammar instruction consisted of learning verb conjugations and three in another in which the source of the grammar instruction was Warriner's English Grammar and Composition (Stubbs, 1995c).

Since there were too few students in either the control group or the experimental group for the difference between their scores on any criterion to have statistical significance, the difference scores have been omitted here.

the content of the experimental instruction

In her study, Stubbs details the instruction for the experimental class [wording as printed]:

The . . . alternate approach to grammar instruction . . . which will be used in this study . . . evolved . . . from the doctoral dissertation of Hunter (1969) and was published as a text, Sentence Sense: The Hunter Writing System, in 1991.  In a preface to the student, Hunter explains, "though this text concerns itself with `grammar,' it has discarded the inexact definitions in current use.  It has replaced them with strategies that are easy, familiar, and fun" (p. vii).

The system teaches that the verb is the hub of the English sentence, and should be taught to 100% accuracy.  Then students learn to master the structure of the sentence through the use of instructional strategies and mnemonic devices.  Students perform the manipulation of word arrangement in order that they might experience grammatical boundaries and functions.

In this study, the Hunter system will be used in lessons taught immediately before writing opportunities.  The lesson just taught will be reinforced during writing.  The Hunter text is a two volume edition of text material and practice book which are used concurrently.  The newest edition of the practice book includes specific writing exercises with didactic models to help direct the student (Stubbs 1995a, 21).

In the study being conducted by the author of this thesis, many of these skills [sentence combining, sentence construction, paragraph construction, etc.] are taught using specific mnemonic devices and metacognitive self-questioning techniques as developed by Hunter, 1991 (Stubbs 1995a, 26­27).

Students in the experimental group were taught grammar using specific cognitive strategies.  They were also taught to identify problem areas within their own writing.  They learned mnemonic devices to identify different parts of speech and were taught the interrelationships of all parts of a well-constructed sentence.  The Hunter Writing System (Hunter, 1991) formed the basis of all grammar instruction (Stubbs 1995a, 42).

Writing was an activity designated for no less than two days per week with an ever increasing emphasis on successful revision and editing skills prior to publication.  Most of the writing strategies used were those based on the work of Englert, et al., (1991) and Atwell (1987).  Specific application of the grammar strategies were reinforced during the revision and editing stages.  As students perfected a skill, they were encouraged to record what they now understood in an ever-growing compendium of things they now knew (Stubbs 1995a, 42).

how the experimental instruction differs
from traditional grammar instruction

Stubbs discussed how the experimental instruction differed from traditional grammar (1995b); my commentary has been added in inset brackets:

My recollection of traditional grammar texts is less than positive.  Too often lessons are taught as drill and practice exercises in total isolation.  There is little relevance to anything substantive.  Each unit is unique in design and purpose.

Sentence Sense offers tools and strategies whereby the student can immediately see a relationship between each exercise.

[By way of example, the text uses this method to have students find the whole subject in statements:  as a first step, they find/label helping verbs; a sequel step has them reposition the helping verb to the start of the sentence to turn the statement into a question; the following step has them label all the words between the two positions of the helping verb as subject of the sentence.  Here is an illustration: for the sentence "Few of the children in the bus were just sitting still," they would find the helping verb "were," shift it to the start of the sentence to cause the question "were few of the children in the bus just sitting still," and then label as subject the words "few of the children in the bus."]

The lessons build on each other, so each unit actually reinforces those taught earlier.

[Here is an example of this: once students have found wording that does the work of a noun (for example, a whole subject), they use a test called the "The Test" to find nouns inside such wording; they test each word in this subject by seeing whether they can place "the" in front it outside of the sentence; whenever they can, they label those words as nouns. Here is an illustration: students would test the subject wording "rather old trees" by saying--apart from the sentence--"the rather," "the old," and "the trees"; only the word "trees" passes the "The Test."]

The parts of speech and of a sentence become meaningful because the strategies of the text instill in a student a sense of how they tie together. [This is my rewording--approved by Stubbs--of her original sentence.]

[Here is an example of this: Students use a helping verb as a springboard to find whether it has an accompanying main verb--by asking "what" after a helping verb and testing whether the word that answers "what" has an "-ing" way they can write it as part of English.  Here is an illustration: for the sentence "Birds will return to a former locality," they would say "Birds will what?"; they would label "return" as a verb because it answers "would what?" and has an "-ing" spelling (namely: "returning").]

The underlying philosophy of Sentence Sense is particularly useful.  Most children come to the middle grades with an ingrained, intuitive grammar that has served them adequately for more than a decade. They are rewarded for what they already know by virtue of the fact that exercises which make sense are probably correct.  Sentence Sense codifies what they know then offers them the tools to identify syntactical errors in their own writing.

[An example of this is the need for a question formed by the shifting of the helping verb (see above) to be an acceptable English question; when it is not, the sentence is a fragment.  You saw an example of the helping verb shifting successfully above; an example of one not shifting successfully would be the helping verb "could" in the expression "if I could get one more turn." The resultant wording is clearly unacceptable as English--namely: "could if I get one more turn"; consequently, the expression "if I could get one more turn" is a fragment.]

Revision and editing become positive experiences of improvement rather than negative experiences of remediation.

By way of further comment, Stubbs has indicated in conversation that a change in attitude toward writing developed in the experimental students that is not usually in evidence at her school.  She said:  "Students approached each [writing] task with confidence. . . . They took risks as writers because they knew how to correct their own work. . . .  They were not afraid of criticism."

In addition, Stubbs has made these comments: "Unlike the average student, they would stop to think and organize before starting to write. . . . I would never return to the old way of teaching."

To conclude this section, I feel that the most important way in which my text differs from traditional grammar is at the structural level.  It gives students command of structure due to its manipulative, chunking, and carefully sequenced strategies that cause students to experience, and therefore internalize, the system of the structure of the sentence as a backdrop to assist their composing.  As a result, they write with ever greater spontaneity, confidence, and competence.

additional notes regarding the experimental instruction

Here are Stubbs' responses to two other questions asked of her (1995b):

How did you make your teaching work?

I decided to devote the year to the first half of the text with the intent that next year's teacher would complete the book.  Two to three days per week were devoted to Sentence Sense with the remainder of the week being used for a writing workshop.  (Note, this year I have ten minute mini-lessons daily followed immediately by writing workshop activities.)

Writing workshop assignments were designed to reinforce the specific lessons being taught.  Two different models were incorporated into the writing workshop.  The first was the paradigm of Nancie Atwell as described in her book In the Middle: Writing, Reading, and Learning with Adolescents (1987).  The second was the Cognitive Strategy Instruction for Writing (CSIW) copyrighted by Carolyn Englert et al [1991].  CSIW recommends a highly structured approach to writing instruction that teaches transfer of abstract rules and principles by application within the student's writing.

The substance of Sentence Sense taught within the framework of Writer's Workshop using CSIW proved to be very successful.

What other materials and activities did you use in addition to Sentence Sense?

Since I knew that I was not going to be able to complete the text in one year, I taught some fundamental punctuation rules by means of discovery learning.  For example, the students were given samples of sentences demonstrating a particular comma rule.  Within cooperative learning groups they had to come up with a rule that seemed to fit each group of sentences.  When the class came together, we compared the various "rules" and developed a single composite which best described the use of commas.  The students seemed to retain these rules well since they had the ownership of having developed them.

Spelling was taught on two levels. First, each student was responsible for developing a list of words that they were currently using but were misspelled.  The goal was to have them learn to spell the words that they wanted to be able to use in their own communication.  Second was the class development of a list of homophones that seemed to be particularly troublesome (their, there, they're; to, too, two; its, it's; your, you're, etc.).  Every week they were tested on ten of these words by being able to identify which spelling was correct in a given usage.

summary and implications

It is apparent from the above discussions that, contrary to what is widely thought to be true, there have been research studies in the past that lend clear-cut support to the teaching of grammar as a means of improving writing.  It is evident, too, that the existence and importance of at least two of these studies have been misreported in at least one assessment of the subject.

In addition, McCleary (1995) has pointed out that there is a need for a "pedagogical grammar" that works better for students than the grammars tried in the past and that is accurate and simple.  In light of the findings of Stubbs (1995a), it would appear that one such grammar has been found--namely: the Hunter system (1991).  Its simplicity is testified to by the fact that slightly learning disabled seventh graders have learned it sufficiently well to be able to apply it to their writing.  Its accuracy is testified to by the fact that these students' writing improved so markedly--even as much as 67% in overall writing competence--in so short a span of time as six months.  Therefore, it would appear that this is a grammar that works better for students than many, if not most, of those tried in the past.

In conclusion, I feel (as does Stubbs) that the place of instruction in grammar in the curriculum must be reconsidered and that a textbook on sentence structure like the one used in this experiment can fill an important--if not indispensable--role in language arts instruction.  In addition, further research is needed with this kind of textbook material: the experimental group(s) must include a larger number of students; the instruction must be tried with a variety of age and ability levels; and an assessment must be made of the longitudinal benefits to writing from such instruction.

Indeed, there is evidence that the instruction embodied in this text is effective for the improvement of writing across the range of age and ability levels from sixth grade through high school (and the first year of college as well).

Atwell, Nancie. 1987.  In the Middle: Writing, Reading and Learning with Adolescents.  Portsmouth, NH: Boynton Cook.

Bateman, Donald R., and Frank J. Zidonis. 1966.  The Effect of a Study of Transformational Grammar on the Writing of Ninth and Tenth Graders.  ERIC ED 018 424.

________. 1964.  The Effect of Knowledge of Generative Grammar upon the Growth of Language Complexity.  ERIC ED 001 241.

Chomsky, Noam.  1956.  Syntactic Structures.  The Hague: Mouton.

Englert, Carolyn S., et al. 1991.  "Making Strategies and Self-Talk Visible:  Writing Instruction in Regular and Special Education Classrooms."  American Educational Research Journal 28.2: 337­72.

Fillmore, Charles J. 1964.  Desentential Complement Verbs in English.  The Ohio State University Research Foundation, GN 174, Report 1 (Apr.).

________.  1963.  "The Position of Embedding Transformations in a Grammar."  Word 19.2 (Aug.).

________.  1962.  Indirect Object Constructions in English and the Ordering of Transformations.  The Ohio State University Research Foundation, NSF-G18609, Report 1 (Feb.).

Hammill, Donald D. & S.C. Larsen.  1988.  Test of Written Language ­ 2.  Austin, TX: Pro-Ed.

Hartwell, Patrick.  1985.  "Grammar, Grammars, and the Teaching of Grammar."  College English 47.2 (Feb.): 105­127.

Hillocks, Jr., George.  1986.  Research on Written Composition:  New Directions for Teaching.  ERIC ED 265 552.

Hunter, Sr., Anthony Devereux.  1991.  The Hunter Writing System:  Sentence Sense.  Delhi, NY:  Hunter & Joyce.

________.  1969.  A Course to Aid Reading Based on Sector Analysis at the Sentence Level as Revised after Field Testing in the Eighth Grade.  Diss.  Columbia U.

Lees, Robert B. 1960.  "The Grammar of English Nominalizations."  Supplement to International Journal of American Linguistics 26.

________.  1957.  "Review of Chomsky, Syntactic Structures."  Language 33.3, Part I (July­Sept.): 376.

McCleary, Bill.  1995.  "Grammar Making a Comeback in Composition Teaching."  Composition Chronicle 8.6 (Oct.): 1­4.

Stubbs, Barbara J.  1995a.  "Specific Strategy Instruction to Enhance Revising and Editing Skills for the Learning Disabled."  M.A. Diss.  Rowan College of New Jersey.

________.  1995b.  "Answers to Questions about Instruction Given to the Experimental Group during the Time of Her M.A. Study."  Unpublished.

________.  1995c. " Answers to Questions Pertaining Primarily to the Control Group and to Statistical Significance."  Unpublished.

Thompson, Charles Lamar, and Morris Middleton.  1973.  "Transformational Grammar and Inductive Teaching as Determinants of Structurally Complex Writing."  California Journal of Educational Research 24.1 (Jan.): 28­41.

¹Copyright 1996 by the National Council of Teachers of English
  (www.NCTE.org).  Used with permission.

*This is the mostly unabridged edition of my English Journal article (85, No.7 [Nov., 1996], 102-107), which the editors had to shorten slightly. 

[Back to top]