....

/

 

.v...
1..

 

1

INITIIIIIII\\\\\\\\;§\\\}§\T\\§T\\\\i r

3 1293 0078

  
   
   

  

LIBRARY
Michigan State ‘_
University

 

This is to certify that the

dissertation entitled

THE EFFECT OF PRIOR KNOWLEDGE AND LANGUAGE
PROFICIENCY ON THE NONNATIVE SPEAKERS'
WRITTEN DISCOURSE

presented by

AH SaIeh Khabti AI-Ghamdi

has been accepted towards fulfillment
of the requirements for

Ph.D. degreein EnQITSh -

gm

Major professor

Date May 5, 1989

MSUl'rnnAffirm/n' A - r, m" 042771

 

Ad —_v4_7...4._,_.~*4-.‘A ,

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.

DATE DUE DATE DUE DATE DUE

m I!
moi

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU Is An Affirmative Action/Equal Opportunity Institution
c:\circ\datedm.pm3—p.1

 

‘ THE EFFECT OF PRIOR KNOWLEDGE AND LANGUAGE
PROFICIENCY ON THE NONNATIVE SPEAKERS'

WRITTEN DISCOURSE
By

Ali Saleh Khabti Al—Ghamdi

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY
Department of English

1989

f§Q,rb:::uD

ABSTRACT

THE EFFECT OF PRIOR KNOWLEDGE AND LANGUAGE
PROFICIENCY ON THE NONNATIVE SPEAKERS'
WRITTEN DISCOURSE

BY
Ali Saleh Khabti Al—Ghamdi

This study was designed to further investigate
the clahms of Langer (1984), Chesky and Hiebert (1987),
Stroethoff (1988), and others that. prior‘ knowledge is
related to success in writing. Using 49 nonnative
speakers studying English at the advanced levels in an
intensive ESL program at a large, Midwestern university,
the study investigated the relationships among various
measurements of" writing' skill (writing jproficiency
scores), background knowledge, language proficiency,
reading habits, and demographic information.

Each subject was asked to write on the topic of
eclipses, a topic on which students showed a large spread
0f knowledge. Before writing the essay, the subjects
took free-association and prior knowledge tests on the
toPic. Then they filled out a questionnaire which
reflected the degree of their involvement with the topic,

their background characteristics, and the amount of

Ali Saleh Khabti Al-Ghamdi

reading they did. Trained raters assigned holistic
writing proficiency scores and rated the writing samples
for amount of content, text sophistication, global
coherence, linguistic complexity, and revision
strategies. The raters also scored the prior knowledge
test for the degree of fluency and organization.

Data were analyzed using frequency analysis,
correlation coefficient, regression, and analysis of
variance. IPrior knowledge significantly correlated with
writing proficiency scores, content length, text
sophistication, global coherence, and involvement. Prior
knowledge was not significantly related with T-unit
length, amount of subordination and revision strategies.

Overall language proficiency was significantly
correlated with writing proficiency scores, text
sophistication, and linguistic complexity, but was not
significantly related with content length, global
coherence, revision strategies, and involvement.

Among the other variables measured, age appears
to have a significant relationship with content length
and involvement with the text. Level of education and
the amount of reading appear to have a significant
relationship only with involvement. The results also

showed that there was a difference among the subjects'

Ali Saleh Khabti Al—Ghamdi

writing related 1x3 their native languages. Implication
of these findings are drawn and suggestions for further

research are provided.

 

Copyright by
Ail s. K. Al—Ghamdi
1989

 

 

vi

Dedicated to my father, Saleh,

and my mother, Khadra.

May God rest their souls in Heaven.
My brother, Mohammad, and

my wife, Azizah, with my love and thanks.

ACKNOWLEDGMENTS

At the successful completion of this project, I
would like first to thank God for the opportunity,
patience, and strength to enable me to finish this simple
work. Thanks also goes to my government for its
sponsorship and support, which made this work possible.
I also wish to extend my sincere thanks and appreciation
to all those who have helped me along the way. In
particular, I am indebted to the following people:

Dr. Paul.‘E. Munsell, friend, academic advisor,
and committee chairman, for encouragement valuable
comments, and helpful suggestions.

Dr. James C. Stalker, Dr. Marcellette G.
Villiams, and Dr. Marilyn Wilson for their willingness to
:ake the time and responsibility to serve on my guidance
:ommittee and for their thoughtful comments on my
ussertation work. Dr. E. Dean Detrich for serving as a
acuity representative for the College of Arts and
etters at my dissertation's defense and for the comments
e provided.

Dr. Betsy Becker, for answering my statistical

lestions during the analysis phase of this manuscript.

viii

 

 

Mrs. Ila Baker, Mrs. Alice Dungey, and Mr. Joel
Boyd, Jr., for their time and effort to serve as raters
for the study. Also my appreciation goes to Margot Hynes
for her assistance as a rater.

Director of the English Language Center, his
staff, and D and E levels writing teachers and students
of Spring, 1988, for their cooperation and help in
collecting the data for this study.

Nancy Heath for her cooperation and meeting of
deadlines in the typing of this manuscript.

Last, but not least, special thanks go to all my
friends and relatives for their help, encouragement, and
financial support. My appreciation goes to my wife, my

aughters Abeer, Najla, and Ro'aa and my sons Saleh and

ohammed for their love, support, and patience.

ix

 

 

 

Table of Contents

Page

:IISTOFTABLES..eenooeooeooeoo'boao-oeeoooosq ...... XiV

LIST OFFIGURES................... ........ . ....... xvii

LIST OF APPENDICES..................... ..... .....Xviii
Ihapter

I. INTRODUCTION...................... ..... .......l

The Problem.................................1

Reading and Writing.........................2

The Goal of the Study.................. ..... 5

Hypotheses..................................6

Limitation of the Study... ..... . ..... .......7

Generalizability of the Study...............7

SumarYOooeoeeouoooeooo-ocooooooooo ....... .08

II. REVIEWOF LITERATURE.CIOOIOCOOOOIODOOOC.00... 9

 

Introduction ..... ...... ..... ... ........ .... 9
Prior Knowledge. .......... ......... ..... 9
Prior Knowledge and Reading. . ... ....... 9
Top— —Down Processing in Reading. ..........10
Reader and Text Interaction.... .......... 11
In Explaining Schema Theory. ..... ........12
Schema Theory and L1 Reading........ ....... 14
Schema Theory and L2 Reading. .. ........ 16
Content Schema... ..... ...... ...... ....16
Content Schema vs. Formal Schema ........ .18
Cultural Schema........ ........... ..19

Schema Theory and Old Views of Reading. .21
Prior Knowledge and Writing Among
Native Speakers............ ..............22

Writing and Topic Choice......... ........ 28
Language Proficiency and Writing Among
Nonnative Speakers. ..... ...... ......31
Measuring Prior Knowledge and Writing
Performance................... ...... .....34
Prior Knowledge.... ....... .... ...... .....34

 

Chapter

Writing Measurement............. ...........
T—Unit Length................ ..... .......
Amount of Subordination..................
Revision strategies....... ......... . .....
Global Coherence........ ,
Involvement..............................

Summary..................

III: METHODOLOGY.oe-tooctet-toe...ooeooo...ooooo-e

The Study's Variables......................
Subjects. ................................
Materials .... ..... . ...............
Selecting the Prompts for

the Prior Knowledge Test. ...........
Prior Knoweldge Test and Background
Questionnaire........... ...... ......

Scoring the Subjects' Associations

in the Prior Knowledge Test............
Writing Performance and

Involvement Questionnaire..............
Writing Measurement................. .....
T—Unit Length and Amount of

Subordination..........................
Writing Proficiency Scores. ..............
Revision Strategies......................
Content..... ....... ..... ..... . ........ ...
Global Coherence............. ........... .
Text Sophistication.............. ..... ...
Involvement. ..... ............ .....
Language Proficiency and Writing. .....

Data Analysis....................... .......
Summary...... ...... ........................
IV. RESULTS.............. ..... ... ..... .. ..... ....
Hypotheses........ ........ ...... .......
Characteristics of the Participants. ...... .

 

Characteristics of the Sample Performance.
The Sample's Prior Knowledge Task........
The Characteristics of the Sample's

Essays.......... .......

Relationships of Demographic Information

and Amount of Reading with Writing...

Comparisons by Native Language... ..... .....

Writing Proficiency Scores........ .........
Writing Proficiency Scores and

Prior Knowledge.......... ..... .........

xi

 

 

 

Chapter

VI

 

Writing Proficiency Scores and

Overall Language Proficiency ........... 76
The Additive Effect ..................... 77
Content: Number of Words and Number of
Ideas... . ............................... 79
Content and Prior Knowledge .............. 79
Number of Words .......................... 79
The Number of Ideas ...................... 81
Content and Overall Language Proficiency...83
Number of Words .......................... 83
The NUmber of Ideas ................... ...84
The Additive Effect ........ . ............. 84
Text Sophistication ........................ 86
Text Sophistication and Overall
Language Proficiency ................... 89
The Additive Effect ...................... 90
Global Coherence ........................... 91

Global Coherence and Prior Knowledge.....91
Global Coherence and Overall

Language Proficiency ................... 92
The Additive Effect ..................... 94
Linguistic Complexity: T- Unit Length
and Subordination ........................ 96
Linguistic Complexity and Prior
Knowledge .............................. 96
Linguistic Complexity and Overall
Language Proficiency ................... 98
The Additive Effect ............... . ..... 99
Revision Strategies ...................... 101
Revision Strategies and Prior
Knowledge ............................. 101
Revision Strategies and Overall
Language Proficiency .................... 103
The Additive Effect ..................... 104
Involvement ............................... 105
Involvement and Prior Knowledge ......... 105
Involvement and Overall Language
Proficiency ........................... 107
The Additive Effect ................ .....107
Summary ................................ ...108
CONCLUSIONS ................................. 111
Introduction .............................. 111
Personal Characteristics and Writing ...... 113
Participants' Naive Language ............ 114
Prior Knowledge and Writing Performance...115

 

 

 

Chapter

Suggestions for Further Research.
Summary...

APPENDICES

REFERENCES

 

Writing Tests................

Content Length....

eoooooooooee

Text Sophistication..........

Global Coherence..
Involvement...

to.

00.00.0000...

coo-0000....

o o o o o t
o eeeee o o
o o o e 9 0

clove...

Linguistic Complexitiy................
Revision Strategies..........
Prior Knowledge and Writing

Performance of the Subgroups......
Overall Language Proficiency and

Writing Performance........
Additive Effect.....
Implication........

coo-oo-

tote-o.

....DIOIOOOOOOOOI

eoeeooovoul

eaeoeooooooucooeoooeo

0.300.000.-

xiii

coon-one.

ego-coco

.0000...-

ooooleaoun

115
116

.117

118
119
120
121

121

122

.123
.124

127
130

.132

191

 

 

 

LIST OF TABLES

Page
Gender, Age, Level of Education, and Native
Language.............. ........ ............ ....... 61
ELC Scores of Participants.................. ..... 62

Characteristics of the Sample's Prior Knowledge
Task 64

Characteristics of the Sample's Essays...........65

The Relationship of Age, Level of Education,
and Amount of Reading with Written Discourse
......68

Measures............... .........

oeoeooeeceo

MANOVA and ANOVA Results for the Difference
Between the Population Means Writing Measures

by Native Language...............................7O

Means of Organization, Combination, and

Subordination by Native Language.................72

Means of Writing Proficiency Scores, Number
of Words, Number of Ideas, Text
Sophistication, Global Coherence, T—Unit
Length, Revision Strategies, and Involvement
of Writing Components by Native Language.........73

Correlation Between Prior Knowledge and Writing
Proficiency Scores............ .........74
Correlation Between Overall Language
Proficiency (OLP) and Writing Proficiency
Scores.................. ........
Regression Results of Writing Proficiency
ScoreSoooyeo ..... OIUOODIOOODIIO IIIIIIIIIII 0.9.0.078

. ....... 76

coo-cues:

Correlation Between Prior Knowledge and
Content: Number of words and Number of

Ideas for all Groups and Each Group..............80

xiv

 

Tab:

 

Table

13.

14.

15.

16.
l7.

l8.

19.

20.

24.
25.

26.

28.

 

Page

Correlation Between Overall Language
Number of

Proficiency (OLP) and Content:
Words and NUmber of Ideas............. ..........82
Regression Results of the Content: Number of
Words......... ............ . . .. ............. . ..85
Regression Results of the Content: Number of
Ideas............................................86
Correlation Between Prior Knowledge and Text
Sophistication for All Groups and Each Group.....87
Correlation Between Overall Language
Proficiency and Text Sophistication... ...... .....89
...91

Regression Results of Text Sophistication.....

Correlation Between Prior Knowledge and Global
Coherence for All Groups and Each Group..........93

Correlation Between Overall Language
Proficiency and Global Coherence.......... ....... 94

Regression Results of Global Coherence .......... 95

Correlation Between Prior Knowledge and
T—unit Length and

Linguistic Complexity:
Subordination for All Groups and Each Group......97
Correlation Between Overall Language

Proficiency (OLP) and Linguistic Complexity......98
Regression Results of the T—Unit Length ...... ..100
Regression Results of Subordination ......... ....101
Correlation Between Prior Knowledge and

Revision Strategies for All Groups and

Each Group... ................................... 102
Correlation Between Overall Language

Proficiency (OLP) and Revision Strategies.......103
Regression Results of Revision Strategies.. ..... 105

XV

 

 

Fable Page

29. Correlation Between Prior Knowledge and

Involvement for All Groups and Each Group.......106
30. Correlation Between Overall Language

Proficiency and Involvement.. ................... 107
31. Regression Results of Involvement... ............ 108
1-1. The Study's Raw Data ...... ..... ................. 178

LIST OF FIGURE
Figure Page

1. Levels of organization and response
categories: Schizophrenia......... .............36

xvii

LIST OF APPENDICES

Page

Miscellaneous Correspondence.... ..... .........133
Background Information Questionnaire ........ ..140
Word Association/Prior Knowledge Test.. ....... 142
Scoring the Prompts............ ...... .........153
The Writing Task and the Involvement
Questionnaire.......................... ....... 157
Scoring the Participants' Written Discourse
Text Sophistication Questionnaire.... ..... ....161
Procedure for Selecting Prompts
Procedures of Scoring Knowledge Organization. ...
English Language Center Writing Scoring Scale
Bamberg Coherence Scale. ............... . ..165
The Study's Raw Data....... ........ ...........175

...179

Prior Knowledge Test and Writing Samples..

xviii

 

 

 

 

Sp;

CHAPTER I
INTRODUCTION

The Problem

This study attempted to determine the
relationship among prior knowledge, language proficiency,
and writing of nonnative speakers of English. This
investigation is important because no research has yet
>een done with nonnative speakers. Most of the studies
:ited below were conducted with native speakers. The
.nvestigation is also important because the measurement
if prior knowledge and its effect on writing have been
eglected relative to the emphasis placed on language
kills as summarized in the studies of Applebee (1984)
nd Newell and MacAdam (1987). The more traditional view
E writing tended to stress both language ability and
rammar skills and equate them with writing ability.
reffective writers were, therefore, often given a dose
language in the hope that it would remedy their
'iting ailments. To achieve these language ends, one

pic was often assigned to all students.
As Chesky (1984) stated, however, when native

eakers of English all wrote on the same topic, some

1

 

migh
litt
indi
writ
mech;
Howei
elems
cruc:
being
discr
and j
varic

When

 

indie

a giv

View
readii

might not know* much about the topic, and hence, have

little to write. Newell and MacAdam (1987) correctly

indicated that the role of language proficiency in
writing cannot be denied and that clarity of expression,
and form are affected by language proficiency.

also pointed out that these

mechanics,

However, these authors

elements alone do not constitute writing ability. Other

crucial factors, such as prior knowledge of the topic

being written about, must be taken into account in

discussing the writing ability of the students. Newell

and MacAdam said, "Even good writers who have mastered

various writing genres may seem to lack writing skills

when facing new topics" (p. 157).

Another writing theorist, Hoetker (1982)

indicated that little has been known about the effect of

a given topic on students' writing. He said:

Our understanding of topic effects is at such a
primitive stage that we cannot demonstrate
whether or in what ways students will write
differently in response to such thoughtless
topics than they will write in response to the
most carefully considered and closely edited

topics (p. 382).
Reading and Writing
It seems more likely that this language-based
iew of writing ignored a tremendous body of research on
aading which assumed that reading involves much. more

1an language. As demonstrated in the reading research

of

(197
(198
cons

the

ever
read
rese
stud:
DeGr<
(19M

Speai

becau
compr
Writt

°Vers

of' Johnson (1981), Rumelhart (1975), Anderson et a1.

(1983), Lee (1986), Hudson (1982), Cziko

(1977), Carrell

(1980), and others, prior knowledge plays a major role in

constructing meaning and organizing the information of

the text.

Based on the findings of these studies, there is

every reason to believe that what has been found about

reading probably applies to writing. Therefore, writing

investigate writing . Some

(1984),

researchers have begun to
studies of this type were conducted by Langer
DeGroff (1986), Chesky and Hiebert (1987), and Stroethoff

(1988) to clarify the role of prior knowledge in native

speakers' writing.

Langer (1984), for instance,
effects on reading

speculated that

because there were schematic

:omprehension, there were also schematic effects on

written expression and ignoring such effects would be an

>versight. She said:

. . writing as a skill is too intertwined with
knowledge of the subject matter itself to isolate
the two without considering how one affects the
other. Because topical knowledge so directly
helps shape a paper, the teacher's understanding
of what students know about a topic can be very
useful :u1 planning writing assignments, in
setting expectations for various students, and in
providing pertinent help while a writer is at

work (p. 28).

Chesky and Hiebert (1987) examined the essays of

:udents who wrote on two topics—-one in which the

 

 

 

 

the:

students had little knowledge and one in which they had

considerable knowledge. Chesky and Hiebert found that

when students wrote about a topic in which they had

background knowledge, they wrote quantitatively more and

qualitatively better; and they became more involved in

their writing and enjoyed it more than when they wrote on

a topic about which they had little knowledge.

DeGroff (1986) investigated the effect of prior

knowledge on conferencing' and revising texts and found
that students with high knowledge produced better content

by stating valued information within the specified

(nowledge domain. Furthermore, the students with high

:nowledge provided better comments during conferencing

:essions. In comparing the effect of prior knowledge and

:ohesive ties on the texts' coherence, Carrell (1982)

ndicated that if students lacked suitable background

nowledge, cohesive ties would not prevent their text

rom being fragmented and incoherent.

From these studies, we can conclude that writing

asearchers have not gone far enough to look for the role

5 prior knowledge in nonnative speakers' writing. None

these studies has looked at the effect of prior

owledge and language proficiency on writing in spite of

e opinion of Carrell (1986) and the other schema

eorists that both prior knowledge (represented on top-

wn processing) and language skills and vocabulary

 

dev

nee:

knor
non:

spee

The

prof
prof
SOph

revi:

Profi
inves
rOle

clear
ident:
effeci
tentat
effect
Study
relati

info rm

 

development (represented as bottom—up processing) are

needed for comprehension to take place.

Therefore, this study investigated whether prior

knowledge represents the same type of importance in

nonnative speakers' writing as it does in the native

speakers' writing. Likewise, it also tests the effects

of language proficiency on nonnative speakers' writing.

The students' levels of prior knowledge and language

proficiency will be correlated with their writing

proficiency scores, length of content, text

sophistication, global coherence, linguistic complexity,

revision strategies, and involvement with the texts.

The Goal of the Study
Even though prior knowledge and language
important and been

their

>roficiency would appear to be
.nvestigated to some extent among native speakers,
ole in the nonnative speakers' writing measures are not

lear. The goal of this study then is to simultaneously

dentify these roles. Without a study of the combined

ffects of these two variables on writing measures, only

antative predictions can be made about whether the

ffects are separate or interactive. Furthermore, this

:udy will undertake the investigation of the

iationship among writing measures and demographic

formation (such as age, gender, etc.), the amount of

readi
are i

Engli

 

reading students do and writing measures.

The following

are the major issues with regard to nonnative writers of

English:

The specific hypotheses

Is there any difference among the

participants' writing related to their

personal characteristics and the amount of

reading they do?
What is the relationship between prior
knowledge about specific topics and the

nonnative speakers' written discourse?

What is the relationship between overall

language proficiency and the nonnative
speakers' written discourse?

What is the additive effect of prior

knowledge and overall language proficiency on
written discourse?

Hypotheses
(stated as null hypotheses)

re the following:

There is no relationship between the

Hypothesis 1:
speakers' prior knowledge about

sophistication,
linguistic complexity,
and (g) involvement.

nonnative

specific topics and their scores in (a) writing

proficiency tests, (b) content, (c) text
(d) global coherence, (e)

(f) revision strategies,

 

0rd
get
the

Spa

Hypothesis 2: There is no relationship between the
nonnative speakers' overall language proficiency

and (a) writing scores, (b) content, (c) text
sophistication, (d) global coherence, (e)
linguistic complexity, (f) revision strategies

and involvement.

Limitation of the Study

A limitation is necessary related to applying the
prior knowledge testing instrument to nonnative speakers.
Swaffar (1988) said that students tested in their native
language had ani advantage over their second language in
recalling any'jpassage's information. Thus, because the
subjects used their second language in the prior
knowledge test, a decision of whether the students'
responses to the test reflected their actual knowledge of
eclipses or not can't be made. One potential solution to
this limitation would have been to have the students use
their first language in the prior knowledge test, then
either translating the students' responses to English in
arder to have them rated by English native raters or
letting raters who speak the students' language to rate

:heir responses.

Generalizability of the Study
The investigation was conducted with nonnative
Peakers who were studying English as a second language

t Michigan State University. There appears to be reason

 

prt

vie

Vie
alc
wri
eff

var

to believe that resulting generalizations can be applied

to subjects of similar backgrounds.

Summary
This chapter stressed that emphasizing language

proficiency and equating it with writing ability, as old
Views of writing appear to have done, did not accurately
represent the writing process. The recent views of
writing, however, do not completely rule out the effect
of language proficiency on writing. Instead, the recent
views of writing indicate that language proficiency,
along with students' prior knowledge, interact in the
writing process. In this study we will examine the
effects of both these variables on writing. Some other
variables, such as the students' demographic information
and the amount of reading the students do, are believed
to be related ix: ESL written discourse; these variables

will also be investigated.

 

ant

flO‘

lei

C0]

rei

C011

 

int

wri

knc

st

1an

ul
P

app

CHAPTER II
REVIEW OF LITERATURE

Introduction

This study undertook the investigation of the
effects of prior knowledge and language proficiency on
nonnative speakers' writing. Concerning prior knowledge
and its effects on writing, writing theorists generally
now assume similarity between reading and writing at some
level of abstraction since both. processes are used to
construct meaning. Thus, the first section of this
review of the literature will focus on the studies
related to prior knowledge and its effects on reading
comprehension. The second section of this review will
include a recapitulation of the studies related to
writing as a process affected by the writer's prior
cnowledge. The third section of this chapter reviews the
:tudies related to the measurement of grammar and

.anguage proficiency on writing.

Prior Knowledge
rior Knowledge and Reading
In outdated views of reading, readers passively

pproached the texts in order to absorb what the authors

 

sen‘
part
thai
resa
Semi
pro<
com!
rec:

reat
393:

the:

reac‘
0rde
util
"to

10

of those texts had produced. Those views were challenged
by recent language and reading theorists who indicated
that reading comprehension required an active
participation. of the comprehenders. This argument was
initiated by Chomsky (1959) who characterized the
behaviorists' view of language as a type of learned human
behavior with a predictable response to a known stimulus.
Chomsky insisted that understanding and producing a
sentence did not rely on our having learned that
particular sentence. We can say and comprehend materials
that we have not heard before. As a result of Chomsky's
research, the emphasis of teaching language shifted from
sentences themselves and their meaning to the learning
processes of the students themselves, and how they
comprehended sentences. This shift can be readily
recognized in the consequent theories in the field of

reading.

Top—Down Processing in Reading

Chomsky's research was followed by other reading

 

theorists, such as Goodman (1967) who stated that reading
LS a "psycholinguistic guessing game." Goodman said that
readers must interact with the text they were reading in
Irder to construct Ineaning. This interaction requires
tilizing higher level knowledge, to enable the reader

to get meaning from the print, to develop the sampling,

pred
para
Good
and
grap
was

made

the
know]
resee

EVOIV

Struc
text
delin

aids

(D
3’.
(D

thee;
text
With
c°mpre
two N(

Calle<

11

predicting, confirming, and correcting strategies
parallel to those they use in listening" (Goodman and
Goodman, 1982, p. 75). So, according to Goodman (1967)

and Goodman and Goodman (1982), it was not the text's
graphic representation that caused comprehension, but it
was the knowledge that readers brought to the text that
made reading possible.

A large amount of research has been conducted in
the past decade to assess the effects of the readers'
knowledge on reading comprehension. As a result of this
research, a new, interesting and sophisticated theory
evolved called Schema Theory. This theory demonstrated
the necessity of an active utilization of knowledge
structures within the comprehender in response to the
text to be comprehended. The following section will
delineate the way this interaction takes place and how it

aids reading comprehension.

Reader and Text Interaction

Carrell (1986) said the way in which schema
theory conceptualizes the interaction between reader and
text is that the input of the text must be compatible
with the reader's existing schema in order for
comprehension to occur. Schema theory provides us with
two modes by which information is processed. One mode is

:alled bottom-up or data-driven processing which

 

rep
inf
top
rep

rea

comi
faci
know
the

thos
proc
know
theo
mode
comp
r6co

Pr0c

SChen
infOI
the 5

readE

12

represents the reader‘s utilization of the graphic
information in the input data. The other mode is called
top—down or conceptually-driven processing which
represents the concepts from within the reader that the
reader draws on to search the data.

According to schema theory, top—down. processing
is when readers make guesses or predictions as to what is
coming in the text. Top—down processing is enhanced and
facilitated by the reader's past experience or prior
knowledge. Bottom-up processing is when readers decode
the text's graphic components to construct meaning out of
those linguistically decodable sentences. Bottom-up
processing is .enhanced by the reader's linguistic
knowledge (phonology, morphology, and syntax). Schema
theory clearly suggests that relying on one processing
mode at the expense of the other will impede
comprehension (Eskey, 1988). However, readers must
recognize that different texts may require more of one

processing mode than the other.

In Explaining Schema Theory

 

Anderson (1984) enumerates six functions of
schema. First, a schema is composed of slots. Learning
Lnformation easily takes place if that information fits
:he slots in the reader's schema. Second, schema enables

eaders to pay more attention to what aspects of the text

are
mak
hel
inf
rec
cas
aPP
wou
Fif

infw

Sixi
Spet
post
SOme
ment
the

What

body
exte
COmp
theo

revi

13

are more important. Third, a schema enables readers to
make inferences if the text is complex. Fourth, a schema
helps direct readers to search their memory for certain
information” As an example, we search our Inemory to
recall what kind of food a fine meal includes. In this
case and as a result of a inemory search, the kind of
appetizer, the kind of soup, and whether or not a salad
would be served are the types of information recalled.
Fifth, a schema enables the reader to summarize the
information read. In this summary, important information
is included and predictable portions might be deleted.
Sixth, and finally, a schema, along' with the recalled
specific text information, enables the reader to
postulate any missing information. For example, if
someone wants to recall the type of drink that was
mentioned in a text as served with a certain meal, and if
the person can remember the meal, that person can guess
what type of drink is usually served with that meal.

In light of these schema theory principles, a
body of research. has been. conducted to investigate the
extent to which schema theory can help readers readily
‘ comprehend texts. In the following sections, schema
theory and readings in L1 and schema theory and L2 are

reviewed. Following that is a review of the literature

 

 

abc
lit

wri

has
rea
hav:
(19E
voce
text
subj
asso
subj
The

they
info
knowj
know]

reCa]

P°Str
antic
Pest

14

about prior knowledge and writing. Finally, the
literature regarding how students' prior knowledge and

writing performance are measured is reviewed.

Schema Theory and L1 Reading

 

In the past decade, a large amount of research
has been conducted to investigate how native English
readers significantly comprehemd a text for which they
have high background knowledge. Langer and Nicolich
(1981) examined how the text's specific concept and
vocabulary affect the readers' processing and recalling a
text. Their subjects were 36 high school seniors. The
subjects' knowledge was tested by asking them to write
associations to the key words of two passages. Then the
subjects were given the two passages to read silently.
The researchers then asked the subjects to write, what
they could remember about the passages. When the
information recalled by the subjects was compared to the
knowledge they had, it was found that the level of prior
(nowledge was strongly related to the information
:ecalled from the passage.

In related studies, Pearson et al. (1979)
:ostulated that understanding, retelling, and
nticipating structures depend on elaborate schemata and
ast experience. These authors found that comprehending

text is a process of integrating text information with

 

 

preex
were
intng

possii

how fa
fifth-
differ
was r
Conver
reader
Taylor
succeg

the te:

who int
their 5
deficie
Poor re
between
memory

the ct}
related
comprehe
that t

15

preexisting schemata. They added that if the schemata
were not strong, comprehension would be low because the
integration of new information with schemata would not be
possible.

Taylor (1979) conducted a study to investigate
how familiar texts aided the comprehension of third— and
fifth-grade children. Taylor concluded that the
difference between poor and good readers' comprehension
was not significant when the texts were familiar.
Conversely, the difference between those two types of
readers was significant when the texts were unfamiliar.
Taylor suggested that both good and poor readers
successfully utilize the top—down processing Inode when
the text is familiar.

Contrasting results were provided by Eamon (1979)
who indicated that poor readers were not able to utilize
their schemata to comprehend a text. The reason for this
deficiency among poor readers as stated by Eamon was that
poor readers did not have the ability to differentiate
>etween related and unrelated information during the
remory search for relevant information. Good readers, on
:he other hand, were able to locate easily the text—
elated information in their minds and utilize it to
omprehend that text. The implication of this study was
hat the role of prior knowledge should not be

Jerestimated. Reading teachers must make sure that less

 

prc
act

rea

for

imp

tex
the
com
sin
corn
rea
res
fol
rol

the

km
461
38d
the}

EXaI

 

16

proficient readers get the assistance they need for
activating the schemata related to the text they are
reading.

This research illustrates how important schema is
for native speakers' reading comprehension. This
importance stems from the fact that schema theory gives
an important role to the comprehender as well as the
text. Unless successful interaction takes place between
the comprehender's knowledge and the text's knowledge,
comprehension will be incomplete. It was assumed that
since prior knowledge affects native speakers' reading
comprehension, it also affects nonactive speakers'
reading comprehension. Therefore, ESL reading
researchers started to investigate this assumption. The
following section reviews the studies investigating the
role of prior knowledge in L2 reading comprehension in

the light of schema theory.

Schema Theory and L2 Reading

Content Schema

 

Content schema was defined by Carrell (1987) as
nowledge relative to the content domain of the text (p.
61). A number of studies have been conducted with
econd-language learners to answer the question of how
hey use content schemata to comprehend a passage.

xamining the effects of background knowledge on native

 

and
stu
fir
or

tra
tex
fan
the
tha
tak

tex

sam
of

rep
pre
bot
lan
rec
Lee
tha
com:
fee.
was

red

 

17

and nonnative speakers' comprehension, Carrell (1983)
studied three components of background knowledge: the

first one was context; presence or absence of a picture

or a page title before the page. Second was
transparency: presence of lexical items which provide
textual clues. The third background component was

familiarity' with. the text or being’ knowledgeable about
the text to be read. The findings of this study showed
that unlike native speakers, nonnative speakers did not
take advantage of their prior knowledge to comprehend a
text.

Lee (1986) replicated Carrell's study using the
same components of background knowledge as Carrell. One
of the differences between Carrell's study and Lee's
replication of Carrell's study was that the instructions
preceding the passages to be read in Lee's studies were
both written and read aloud in the subjects' native
language. Also, subjects were asked to write their
recollections in their native language. The findings of
Lee's study contradicted those of Carrell. Lee found
that nonnative speakers were able to utilize the
combination of their background knowledge components to
recall the text. The reason Lee differed from Carrell
was that Lee asked his subjects to write their

recollections in their native language which gave them an

 

 

adv;

(19:

was
adva

text

Cont
addi
ESL
SChe

orga

form
DOSE
inte
Comp
text
Carr
and

fami

18

advantage in demonstrating their comprehension. Lee
(1986) said:
. . . the way in which comprehension is assessed
affects the evaluation of comprehension.
Assessing comprehension. with. a target language
task may limit learners' ability to demonstrate
what they comprehended. Assessing comprehension
with the native language allows learners to more
fully demonstrate their comprehension (1986, p.
353).
Commenting on this matter, Swaffar (1988) said it
was obvious that using a native language provided
advantages over using the second language in recalling a

text.

Content Schema vs. Formal Schema

 

Reading theorists have indicated that, in
‘addition tr) content schema, formal schema also affected
ESL reading comprehension. Carrell (1987) defined formal
schema as "knowledge relative to the formal rhetorical
organizational structures of different types" (p. 461).
In comparing the effects of content schema and
formal schema on reading comprehension, Carrell (1987)
posed the questions: what is the effect of the
interactions of content and formal schema on a text
comprehension? And what is the effect of one schema on a
text comprehension when the other is held constant?
Carrell had her intermediate ESL subjects read, recall,
and ranswer questions about two texts. One text was

Familiar and 111 a well—organizational rhetorical format.

 

The
uno
wit
comj
unf
Fur
witi
for]
tru.

COH‘

HOWT
fam;

imp<

SChE
rtad

abSe

 

19

The other text was unfamiliar and rhetorically
unorganized. Carrell's findings showed that the text
with familiar content and familiar form was easy to
comprehend; the text with unfamiliar content and
unfamiliar form was difficult to comprehend.
Furthermore, this study's findings showed that a text
with familiar content, but with unfamiliar rhetorical
form, was also easy to comprehend. The opposite was not
true. That is, if the text's form was familiar, but its
content was unfamiliar, comprehension was difficult.

So, content familiarity had a more important role
in the comprehension of unfamiliar content than did form.
However, the form familiarity did a job that content

familiarity could not do. That form familiarity was more

  
   
   
  
  
   
     

important in "the comprehension of the top—level episodic
structure of a 'text and in 'the comprehension of event
sequences and temporal relationships among events"
(Carrell, 1987, p. 476). Therefore, each type of schema
content and form appears to play an important, but unique

role in comprehension.

ultural Schema

As indicated by Carrell (1987), cultural content
chema was powerful to the extent that it facilitated the
eader's reading comprehension and compensated for the

bsence of formal schema. Carrell said: "Whether the

for
sig

bas

cul
The
Ind
cus
the
the
suh
mom
the
the
Wen
difl
imp:
Sche
fast
alth
demc
SUCc
fr0n

cUlt

effe

20

form was rhetorically familiar or unfamiliar had no
significant effect on subjects engaging in culturally
based elaborations or distortions" (p. 474).

Steffenson et al. (1979) studied how different
cultural background knowledge assisted comprehension.
Their subjects, native English speakers and native
Indians, were asked to read passages about wedding
customs in their cultures and in different cultures, and
then recall, in writing, information about the passages
they had read. The result of this study showed that
subjects were able to read faster and to remember details
more accurately, and make more accurate inferences when
they read about their own wedding customs. In addition,
the readers' recollections revealed that wrong inferences
were obtained as a result of reading passages from
different cultures. These findings indicated two
important points. First, readers with an elaborate
schema take advantage of that schema to read successfully
faster, remember better, and infer more details. Second,
although native English speakers and native Indians
demonstrated high. language proficiency, they could not
successfully remember details or make accurate inferences
from passages about cultures other than their own
:ulture.

In a related study, Johnson (1981) examined the

affect of interaction of linguistic complexity

 

 

(vor
cult
Ame]
Her
that
knor
groi

infr

M

on

emp]
pro:
matr
stur

tOp;

Shi:
rta<
(19:
indi
Com
This
beti

SChs

21

(vocabulary difficulty and length of sentences) and
cultural background knowledge. Johnson asked Iranian and
American subjects to read passages from both cultures.
Her findings were consistent with those of Steffenson:
that cultural background had more effect than linguistic
knowledge on comprehension. She also found that both
groups were easily and quantitatively able to elicit

information from culturally familiar passages.

Schema Theory and Old Views of Reading

The traditional way of reading put the emphasis
on processing language form and structure. It also
emphasized reading correctly, tied reading with language
proficiency, ignored the values of students (by assigning
materials that may have ignored the values of the
students), and focused on mechanics and cohesion. The
topic to be read was usually out of context.

With schema theory, however, the emphasis has
shifted from reading as a ineans to learn structure to
reading for meaning. The meaning of a text as Fish
(1980) and Iser (1978) indicated does not have an
independent existence. Rather, it developes in the
context of the interaction between readers and texts.
This interaction occurs as top-down processing and
Dottom—up processing take place, simultaneously. In

schema theory, one processing mode cannot be substituted

 

for
the
hav;
WOL‘K
prer
bot
ski

(Ca:

dif

comj

wri
mg,
The
Sin

aid

bel.

iHW

 

22

for the other. In top—down processing, readers overcome
the problem of insufficient background knowledge by
having slots to be filled during* reading. 111 other
words, background knowledge enabled readers to make
predictions that need to be confirmed during reading. In
bottom—up processing, readers employ "language decoding
skills, grammatical skills and vocabulary development"
(Carrell, 1986, p. 8).

These studies demonstrated the importance of
different types of prior knowledge in reading
comprehension. 1

Prior Knowledge and Writing Among
Native Speakers

The presumed similarity between reading and
writing rests on the fact that reading and writing are
cognitive processes aimed at constructing meaning.
Therefore, writing researchers have hypothesized that
since prior knowledge aids reading comprehension, it also
aids writing performance. Chesky (1984) says:

If reading, originally thought to be a simple
decoding process, is heavily dependent on prior
knowledge, then it is highly probably that
writing, possibly a cognitively more demanding
encoding process, is even. more dependent on
prior knowledge (p. 8).

Writing’ researchers with the guidance of’ this

elief started to conduct studies for the sake of

nvestigating the relationship between. prior 'knowledge

 

and
wit.‘
stur
ess;
semi
knor
Lani
bet'
knor
bet‘
the

mom

exa
Cohl
Cre.‘
Stlh
t0b:
Tea:
and
eSsa
the
685E

and

knot

23

and writing. All of the studies so far have been done
with native speakers. Langer (1984) asked 99 tenth—grade
students in four American history classes to write two
essays at two different periods of time during the
semester. Prior to that, Langer measured the students'
knowledge of the topics about which they would write.
Langer found papers requiring comparison or contrast were
better written by students who possessed organized
knowledge. Students who had fluent information wrote
better argumentative papers. On the whole, Langer found
that students with high, knowledge wrote quantitatively
more and qualitatively better.

In a similar study, Chesky and Hiebert (1987)
examined the influence of high and low knowledge on
coherence, syntactic complexity, essay length, content
creating context, and writing errors. They asked 80
students in the 11th grade to write about two topics:
tobacco price supports and school instruction. After
measuring the students' prior knowledge of those topics
and after scoring the coherence, syntactic complexity,
essay length, context creating statements, and errors in
the students' writing, correlation analysis was used to
assess the relationship between these writing' measures
and the students' prior knowledge.

The findings of this study showed that prior

:nowledge significantly correlated with the students'

 

holi
know
with
know
them
stud
leng
of t
thos
(309
know
show
make
disc

unde

24

holistic scores. In other words, students with high
knowledge had higher holistic scores than the students
with low knowledge. The students with high prior
knowledge were also able to organize their ideas and make
them into an integrated whole. Chesky and Hiebert's
study also showed that prior knowledge affected the
length of the students' essays. They said: "The letters
of the high-prior knowledge group were twice as long as
those of the students in the low—prior knowledge group"
(309). Concerning the relationship between prior
knowledge and context—creating statements, this study
showed that students with prior knowledge were able to
make an outline which served to organize the ideas to be
discussed and to create a context which facilitated the
understanding of those ideas.

In surveying the 80 students involved in the
study for investigating the effect of prior knowledge on
the students' attitudes toward the topic, Chesky and
Hiebert found that those with high knowledge liked the
topic and were more involved in it. This study showed no
relationship between prior knowledge and T—unit length
and writing errors (punctuation, capitalization,
spelling, grammar, and usage). To account for this
finding, Chesky and Hiebert suggested that the existence
3f, or lack of, grammatical errors was not affected by

:he students' prior knowledge because those errors may be

 

wri

dra
sec
Con
bas
abl
act
Dur
fils.
rel.

Wit}

 

25

a function of the writers, themselves, and not of what
they knew.

DeGroff (1986) extended Langer‘s (1984) and
Chesky and Hiebert's (1987) studies by examining the
influence of elementary school students' prior knowledge
of baseball on conferencing with peers about their first
drafts and revising those drafts of a narrative. Forty
students took tests originated by Graves (1983) and Calin
(1983) to determine the level of conferencing behavior.
Those 40 students also took a 49-item short answer test
of baseball knowledge used in Mosenthal (1984) and
Mosenthal et al. (1985) to determine the students'
background knowledge. The students were then asked to
write stories about a half inning of a baseball game.

Analyzing students' performance on the first
draft, conferencing with peers, and their revising of the
second drafts, DeGroff found the students' final draft
content showed the influence of prior knowledge of
baseball. That is, students with high knowledge were
able to provide more propositions about "an auxiliary
action," one of the grammatical baseball categories.
During conferences, students with high knowledge were
also able to make comments on specific content features
related directly to the goals of the game, while those
with low knowledge made only general comments not related

to baseball goals. Regarding the second draft, this

 

 

str
cha
con

inf

pri
McC

sut

con
knc
stu
coh
abl

mor

(19

USE

inc

ill

abi
The

26

study showed that students with high knowledge made more
changes than the students with low knowledge. The
conclusion of this study was that prior knowledge
influenced the students' writing processes.

Another interesting study about the effect of
prior knowledge on writing coherence ‘was conducted by
McCutchen (1986). McCutchen first tested her 30
subjects' knowledge of football by giving them 30
questions about a football game. Then she asked them to
write compositions about football. After the coherence
of those compositions was evaluated, McCutchen found that
composition coherence was correlated with the amount of
knowledge the students had about football. That is,
students with high knowledge were able to generate
coherent texts. Furthermore, those students were also
able to bring out and discuss important ideas of football
more coherently than the students with low knowledge.

To have successful written communication, Thomas
(1986) asserts three types of knowledge are crucial:

1. Knowledge of conventions which includes the
use of words and grammar.

2. Knowledge of language, or discourse which
includes the three speech acts: perlocutionary,
illocutionary, and locutionary acts. This means the
ability of the writer to do something to the audience.

The writer may want to move the audience, amuse them,

 

knc

int
nee
eas
Cal
man
the
(19
wel
she
the

mat.

 

27

inform them, or persuade them to perform the
perlocutionary speech act. When stating these goals, a
writer' may want to advise, assert, argue, promise, or
assess to perform the illocutionary speech act. Writers
will refer to entities, such as places, things, or ideas
to express their intentions while performing the
locutionary speech act.

3. Knowledge of the world which means that
writers should have a good sense of what their audience
knows. Thomas says:

a skilled writer will have a fairly
accurate idea of what she can expect her audience
to know about the words—-facts, common opinions,
and so forth. World knowledge is a good
understanding of what her audience already knows
and believes about the world (p. 587).

To do this, writers should provide only required
information without saying more or less than what is
needed. They also should provide true, relevant, and
easy—to-understand information. Those four rnaxims are
:alled the maxims of quantity, quality, relation, and
nanner, previous experience and writing. In suggesting
:he experiential approach to teaching writing, Judy

1980) says: "to write well, one must know something

ell" (p. 39). Judy indicates that composition teachers

 

hould rely on the students' experience and what they
ow in order for the students to produce better written

terials. Students, however, Judy says, should be given

 

 

 

 

 

an
WIS

res

(l!
as:
ati
tin
wit
ple
Kel
wit
gr:

in

and

tilt

 

28

an opportunity to talk about what they know before
writing. He calls this technique an experience
reexamination.

Chiesi et al. (1979), Hilgers (1982), and Kellogg
(1987) indicate that the reason prior topic knowledge is
associated with better writing is that students pay more
attention to their writing instead of spending effort and
time generating ideas. Kellogg indicated that students
with high knowledge demonstrated less cognitive effort to
plan, organize, and review their ideas. Agreeing with
Kellogg, Hilgers stated that if the writers are at ease
with the content, they will give arrangement, style,
grammar, and inechanics more attention which will result

‘in improvement of the quality of their writing.

Writing and Topic Choice

Kincaid (1953) hypothesized that the quality of
students' writing is stable regardless of the topic on
hich a student writes. Kincaid found out that the
ollege students' writing quality varied according to the
ssigned topic, and so his hypothesis was rejected.
ater, Wiseman and Wrigley (1958), in a similar study,
oncluded that the differences found in the children's
riting were caused by the titles of the topics. Wiseman
nd Wrigley suggested that writing teachers should employ

lternative title techniques when assigning an essay

paper

inten

knowl

:2 s2.-

p—I rt- mm rat—'0
...
moyp
..
0
am
(...

E?
a)
<:

might

Stude

there

 

infor
activ
the ;
McKay
10Urn

retri

iflStr

29

paper. In this way, a student could choose a subject of
interest.
McKay (1984) indicated that limited topic
knowledge may impede good writing,
. . we can allow students a great deal of
choice in the selection of writing topics. If
we want to select only one topic, we should
choose a topic about which the students are
likely to have knowledge or experience.
Therefore, topics which presuppose a great deal
of specific knowledge about American culture or
about a particular academic field with which the
students are unfamiliar should be avoided. For
example, an assignment which asks students to
compare and contrast the educational system of
their native country with that of the United
States assumes that the students know a great
deal about the American system of education. If
they do not have this knowledge, they will have
little or nothing to say (p. 189).
To avoid this problem, McKay said composition instructors
might need to spend some class time "familiarizing" the
students with the topic through reading and discussion.
Since, as Flower and Hayes (1981) indicated,
there can be a problem in retrieving a suitable amount of
information relevant to academic writing, McKay suggested
activating the students' schemata of any topic, even if
the students happened to know' a great deal about it.
McKay recommended heuristic devices, brainstorming, or
journal writing as techniques which might help students
retrieve amount of apt information.
Calkins (1983) indicated that successful writing

instruction allows children either to choose their own

 

top.
bet'
chi.

abor

(19:
Stat
sco:
ass:

rea:

Shor
tea:
Prir
EXM

meci

the}
the
ask

film:

30

topics or to employ brainstorming techniques as an aid to
better writing. Graves (1983) also contends that a
children's writing quality is better when they write
about something they know:
Writers who learn to choose topics well make the
most significant growth in both information and
skills at the point of best topic. With best
topic, the child exercises strongest control,
establishes ownership, and with ownership, pride
in the piece (p. 21).
These findings are consistent with Hoetker's
(1982) contention that when the scores of California
State University and College Equivalency Examinations'
scores in 1973 and 1974 declined drastically, the topics
assigned :hi the two examinations were reviewed and the
reason for the variations in scores was found to be due
to the topics assigned for the examination. The review
shows the 1974 topic "called for highly abstract
reasoning and was manifestly more difficult than the
prior year's topic which called upon a personal
experience" (p. 381). Hoetker concluded that the dull,
mechanical, and abstract pieces of text came as a result
of asking the students to write on topics about which
they knew little or nothing. The implication of
ioetker's report is that teachers and examiners should
ask students to write about topics with which they are

familiar.

 

 

pr:

thr

(l!

 

31

Language Proficiency and Writing Among
Nonnative Speakers

In reading, both prior knowledge and language
competence are important for the interaction of both top-
down and bottom—up processing modes. In discussing
reading comprehension, we have seen that without this
interaction, reading comprehension will be incomplete.
So far, all the studies which investigated the effects of
prior knowledge on writing revealed that prior knowledge
is also crucial for writing.

Concerning language proficiency in relation with
writing, Raimes (1987) suggests that decisions regarding
the placement of students in writing courses must not

depend solely <n1 the analysis of language proficiency.

   
   
   
   
   
  

Rather, such decisions must depend on further analysis of
the students' written products and information about
their experience with L1 and L2 writing.

Different views of the role of language
proficiency on writing were presented in current
research. For instance, Jones and Tetroe (1987) contend
that language proficiency is crucial to planning the
antity, but not the quality of writing. Jones (1985)
ays that mastering the linguistic code and being
roficient in it does not necessarily entail mastering
he written code. In contrast, Hanania and Shikhani

1986) argue that the correlation between multiple—choice

 

 

Cor
la<

Prt

 

 

32

tests and integrative tests (cloze tests) and performance
tests, such as writing tests, is significant.

Waltman (1983) conducted a study to assess the
relationship between competency in grammar and the scores
in a business communication course. His conclusion was
that the correlation between written report grades and
the usage test scores was marginal. He also found the
correlation between the final course grade and
grammatical competence to be moderate.

Grammatical competence which, according to Canale
(1983), includes knowledge of vocabulary, pronunciation,
spelling, and word and sentence formation, is crucial for
editing written material since it helps students handle
the structure of the language. Krashen (1984) contends
that teaching grammar does not provide the students with
writing ability. Zamel (1982) indicates that teaching
language in writing classes does not help to view writing
as discovering meaning. Pearl (1979) says that
emphasizing grammar will make the students obsessed with
editing. And if students are preoccupied with editing,
they will not worry significantly about formulating new
ideas (Sommers, 1980).

Reid (1987) gives two causes which lead to
ommunication failure of nonnative writers. One is the
ack of schemata. The other is the emphasis on language

roficiency. Reid says that unfortunately, many

 

 

non
inf

the

Rei
lan
the

inf

stu
con
ess
stu
pro
pro
Stu.
wri
hot

line

 

33

nonnative speakers do not have the necessary background
information and experience (schemata) that would allow
them to complete academic writing tasks successfully.
. . . This problem is not fundamentally a
question of language proficiency, although
frustrated university professors may lay the
blame for unsuccessful written communication not
on what is immediately obvious. "This student
does not understand what I want; I think he
needs work in grammar" (p. 34).
Reid proceeds to argue that despite the importance of
language proficiency in writing, it does not constitute
the core problem. Rather, it is the lack of background
information that causes the problem.

The message that has been obtained from these
studies is that grammar and language proficiency may
constitute an1 important factor, but not the only
essential one in learning how' to write ‘well. These
studies challenged the notion that emphasizing language
proficiency in writing classes may solve writing
problems. Doing so may aggravate them. Therefore, this
study will attempt to draw conclusions for teaching
writing based on the hypothesis that writing hinges on

oth students' prior knowledge of the topic and students'

nowledge of language.

 

 

am
Sh<
flr

34

Measuring Prior Knowledge
and Writing Performance

 

 

Prior Knowledge

A prior knowledge measurement was successfully
utilized with native speakers by Langer (1984), Newell
(1984), and Chesky and Hiebert (1987). In their recent
article, Newell and MacAdam (1987) explained the prior
knowledge measurement and how it was applied in two
previous studies. I

These studies provide the measurement with which
students' prior knowledge was assessed in the present
study. The first step is to choose a topic for the
students to write about. Next, the raters read a
selected article on the theme of that topic. From that
article, raters independently choose words or phrases
that represent the central ideas to that theme. Those
words/phrases constitute the prior knowledge measurement.
Students then are asked to write associations to those
words or phrases. For scoring the students associations,
the prior knowledge measurement indicated that students'
associations must be categorized by raters into three
levels: fluency, organization, and combination.

Fluency was defined by Langer (1984) as the
mount of knowledge the students had about the topics and
hould be scored on a three-point scale: (3) high

luency, (2) moderate fluency, and (1) low fluency.

 

 

COI

Sui

 

35

Organization is scored within the organizational
guidelines depicted in Figure 1 from Newell and MacAdam
(1987). Combination is the accounting at highly
organized and partially organized knowledge. These
procedures are explained in detail in the methodology

chapter.

Writinngeasurement

 

This section reviews the literature which was
utilized to elicit a measurement of T—unit length, amount
of subordination, revision strategies, global coherence,
and involvement. The rest of the writing measurement
(content and text sophistication) were originated by the

researcher.

T-Unit Length

T-unit length is a measure of linguistic
complexity. O'Hare (1973), Mellon (1969), Crowhurste
(1980), Hunt (1965), and Loban (1976) state that a

relationship exists between the length of 'I-units and

linguistic complexity. Hunt termed this relationship
"syntactic maturity," and Mellon called it "syntactic
fluency."

As Myers indicates, the T—units are used to show
complexity, but do not show the utilization of

ubordination. To measure subordination, a further step

 

as taken. This was quantifying the ratio of clauses to

36

Highly Organized Knowledge (3)

Superordinate concepts——higher class category

e.g., schizophrenia-—"One of a group of severe mental
disorders"

Definitions——precise meaning
e.g., schizophrenia——"a psychotic disorder
characterized by withdrawal from reality
including behavioral disturbances."

Analogies-—substitution or comparison for a literal
concept of expression
e.g., schizophrenia——"schizophrenia is a god-demon"

Linking——connecting one concept with another
e.g., schizophrenia—-"schizophrenia is like living in
two worlds because . . ."

Partially Organized Knowledge (2)

Examples--equal class, but more specific
e.g., schizophrenia~-"split personality"

Attributes-—subordinate to large class
e.g., schizophrenia--"character disorder"

Defining characteristics-—defines a major aspect of the
concept

e.g., schizophrenia-—"withdrawal from reality"
Diffusely Organized Knowledge (1)

Associations—-peripheral cognitive links
e.g., schizophrenia——"Jekyll and Hyde"

Morphemes-—echoes smaller unit of meaning such as
prefixes, suffixes, or root words
e.g., schizophrenia-—"Schizoid"

Sound alikes-—similar phonemic units
e.g., schizophrenia-—"Lotte Lenya"

Firsthand experiences——periphera1 responses based on
recent exposure
e.g., schizophrenia—-"crazy--1ike in the movies"

.Figure 1. Levels of organization and response
categories: Schizophrenia.

(Source: Newell & MacAdam, 1987)

 

37

T-units in order to find out how often subordinate
clauses were added to main clauses. Myers (1985) defines
a T—unit as “the main clause and all of its modifiers"
(p. 30). To measure the students' writing complexity,
the raters were given guidelines for T—unit counts
originated by Hunt (1965), used by Mellon (1969), and
illustrated by Meyers (1985). Using those guidelines,
raters were trained to count every T—unit in the essays;
they also counted the words. The mean number of words
per T—units was obtained by dividing the number of words
in the essay by the number of T-units in the same essay.
It is probably appropriate to indicate here that despite
the criticism that T—unit length has been conceived of as
a measure of syntactic complexity, the T-unit is still
widely used in measuring native and nonnative speakers'

syntactic ability.

Amount of Subordination

As Beaman (1984) indicated, amount of
subordination is also an indicator of linguistic
complexity. The amount of subordination is measured by
obtaining the ratio of main clauses to all clauses. This
ratio is rueasured by counting all clauses (subordinate
nd main) in an essay and counting all the main clauses
'n the same essay then dividing the number of all clauses

y the number of main clauses.

 

I'JU

38

Revision Strategies

 

Revision is pausing to go forward. Its
importance comes from the notion that if a writer does
not go back, that writer cannot successfully carry on the
composition in an orderly fashion. Revision is an
important component in written discourse strategies that
helps the writers to clarify the meaning of the text they
are writing. Taylor (1981) defines revision. as "that
crucial point in the process when discovery and
organization come together, when writers refine and
recast what they have written and shape it into a
coherent written statement" (p. 7).

According to Murray (1978). there are two levels
of revision: internal and external. He defines internal
revision as a concentration on the content to rework the
subject, information and argument, and structure.
External revision pertains to the audience and relates to
style and tone, language, and mechanics. This step is
important as Perl (1979) contended to translate
explicitly the tacit sense which would result in the
improvement of the form and the product of the students'
writing.

In light of these definitions, a simple counting
of internal revision strategies, external revision

strategies, and total revision strategies evolved.

 

 

39

Global Coherence
According to Cooper (1988), Langer (1984),
Carrell (1982), and Witte and Faigley (1981), the
cohesive ties provided in Haliday and Hasan (1976) do not
measure the quality of coherence. Cooper said, "Writing
exercises must go beyond syntax to include the complete
written text" (p. 356). And Langer states, "(a) more
coherent text would not necessarily have more cohesive
ties than a less coherent one" (p. 33). Rather,
coherence is determined by the degree to which cohesive
chains interact with each other. Langer also contends
that 23 more coherent text "would necessarily have more
interaction among the cohesive chains“ (p. 33) than would
‘ a less coherent text.
Coherence in this study means the extent to which
a text is globally coherent. A recent definition of
coherence provided by Phelps (1985) and modified by
Cooper (1988) is:
Coherence is a property of intentional global
relatedness that readers ascribe to textual
meaning and thus receive their own integration
or are strongly correlated with the intentions
of the writer. Just as coherence is the
semantic and pragmatic integrity discovered by
readers in textual meaning, cohesion is broadly

the verbal relatedness of the texts as a cuing
system (p. 354).

 

So coherence is related to readers, while

ohesion is related to texts. A recent Holistic

4O

Coherence Scale provided by Bamberg (1984) of the
University of Southern California measures the extent to
which the text seems coherent to the reader.

To measure coherence, Bamberg's (1984) five—point

coherence scale was' used. In this scale, number 5
denotes "fully coherent" and number 1 denotes
"miscellaneous." This measure will be discussed in more

detail in Chapter Three.

Involvement

Chesky (1984) defined involvement as the degree
to which students like their writing, find it easy, and
believe they truly have something to say to someone. The
involvement measurement in this study was designed by the

research based on Chesky's definition of involvement and

based on Krashen's (1984) effective filter hypothesis.

Summary

These studies have shown that readers' previous
knowledge about a passage helps them construct the
totality or scope of the writer's message and arrive at
full comprehension. This research recognizes the equal
importance of both the reader and the text as well as the
necessity of the interaction between. then] in. order for
comprehension to be achieved.

These results encouraged researchers to conduct

studies to investigate the degree to which writers'

41

previous knowledge is important to their writing. The
results on the whole have indicated significant
relationships. However, the research has left some other
questions, such as the role of background knowledge in
writing of ESL students.

In this review of the literature, we have a also
seen that studies about the relationship between grammar
and language proficiency and writing of nonnative
speakers is not so substantial as was once assumed.

Included in this chapter is also a review of the
measurement of prior knowledge and some writing elements,
such as T—unit length, amount of subordination, revision
strategies, global coherence, and involvement with the
texts. We have discussed the state of the art of the

measurement of each writing element.

 

CHAPTER III

METHODOLOGY

The main purpose of this study was to examine the
relationship among prior knowledge, language proficiency
and writing among nonnative speakers. This chapter
presents a description of all variables, samples,
materials, and procedures followed to administer, score,
and analyze the tasks performed by both the subjects and
the raters in the collection of the data. A brief
description of the statistical procedures used to analyze

the data is also provided.

The Study's Variables
The prior knowledge and overall language
proficiency which constitute the two rnajor independent

variables of this study were represented by several

components. The sub—components of prior knowledge are
overall knowledge, fluency, organization, and
combination. The sub-components of overall language

proficiency are a subset of those language skills
measured by the language proficiency test of the English

Language Center of Michigan State University. Those

42

 

 

43

skills are grammar, vocabulary, reading, listening, and
writing.
The measures of the students' writing which
constitute the dependent variables are:
1. Linguistic complexity represented by:
a. T-unit length
b. Amount of subordination
2. Writing proficiency scores (as measured by
the ELC)
3. Revision strategies:
a. Internal revision
b. External revision
c. Total revision
4. Content length
a. Number of words
b. Number of ideas
5. Global coherence
6. Text sophistication

7. Involvement with writing.

Subjects

The accessible population was a group of students
who were studying English at the advanced levels in the
English Language Center in the spring of 1988. The study
was carried out in accordance with the requirements for

human subjects and the English Language Center

 

 

 

44

regulations. After requests had been approved, all
students in levels D and E were asked to participate
voluntarily in the study (see Appendix A). However, only
65 out of 102 students attending these levels agreed to
participate 1J1 the study. Sixteen students were later
excluded from the study because of incomplete responses
to one of the study's two tasks (the prior knowledge task
and the written discourse task). Only 49 students were
found to have complete responses to both 'tasks. This
number constituted the final number of students

participating in the study.

Materials

Topics.-—The researcher began with the following
ten topics which were reviewed by the guidance committee:
eclipses, snoring, condominiums, hurricanes, how to fix
an automobile, high blood pressures, diving, and
photography. Of these topics, photography, diving, and
eclipses were considered most appropriate in conjunction
with the instructional progrmn of the English Language
Center.
Selecting the Prompts for
the Prior Knowledge Test

The researcher selected the three prose passages
about photography, diving, and eclipses from the

Britannica Encyclopedia. Those passages were given to

 

 

 

 

45

three raters who selected seven prompts from each
passage. The prompts were the words or phrases that best
represented the central ideas of each topic. Before
making their selections, the three raters were provided
with guidelines for selecting the prompts provided by
Newell and MacAdam (1987) (See Appendix G).

Each rater" was asked to rank independently the
order of their seven prompts from the most important to
least important. The five most important prompts on
which all three raters agreed were chosen. Disagreement
among raters was resolved through discussion in a meeting
moderated by the researcher.

A pilot study was conducted to train the raters
on how to score the subjects' prompts and on how to score
the writing elements. In the training session held
during the pilot group, the researcher met with the three
raters to discuss the results and to help resolve any
differences of opinion.

Prior Knowledge Test and
Background Questionnaire

After the problems that arose during the pilot
study were overcome, the researcher met with the subjects
in special sessions. In those session, the subjects were
given booklets which contained both the background
information questionnaire and the prior knowledge tests

(see Appendices B and C). In the background

 

 

 

46

questionnaire, the subjects were asked to provide
information about their ages, gender, English Language
Center's proficiency scores, level of education, native
language, and the amount of reading they did.

After the subjects had completed the background
information questionnaire, they were guided to the prior
knowledge test section. Together with the researcher,
they went over the examples provided in the prior
knowledge test to make sure they knew how to perform the
test. After the researcher felt comfortable that the
subjects understood the task, the subjects were given
forty-five minutes to write associations to the fifteen
prompts representing the topics of photography, diving,
and eclipses. Three minutes were allowed for each
prompt. To insure the subjects' comfort during the test
performance, they were told this test would be utilized
for research purposes only, and no scores would be
provided. Furthermore, they were also assured that their
names would not be revealed to any one other than the
researcher.

Scoring the Subjects' Associations
in the Prior Knowledge Test

 

As was done during the pilot study, the three
raters used the procedures outlined by Newell and MacAdam

(1987) (see Appendix D) to score the associations to the

 

 

 

 

 

 

47
prompts. The raters scored the associations on three
bases:
1. Fluency: A score for the total number of
responses to the prompts. The assumption is made that a

larger number of responses indicates a greater amount of
knowledge about the topic.

2. Organization: A score for organization or
quality of knowledge. To score this category, raters
examined the responses to each prompt and gave a score.
To determine a student's total score for organization of
knowledge of a topic, the highest level of responses to
each topic was averaged across all five prompts. For
example, if the student's prompts are scored 2, 2, 3, 1,
and 2, they total 10 points. These 10 points are then
divided by the number of prompts (five), yielding an
average score of 2 (lo/5 = 2). Thus, the student received
an average score of 2 for the organization of knowledge or
the quality of his/her' knowledge about the topic (high
quality = 3, moderate = 2, low = 1).

3. Combination: A score that combines partial
and high organization. In other words, the scores of high
knowledge and moderate knowledge are combined to represent
the combination scores, and the low (chance) scores were

excluded. For example, the above—mentioned student would

 

 

 

 

48

receive a score of (9) for combination because one of the
prompts received a low score (1) which was omitted.

Two goals were achieved as a result of' scoring
these students' associations. First, the students' prior
knowledge scores were obtained. Second, after analyzing
those scores, it was possible to know which topic of the
three had the largest spread of knowledge. Consequently,
eclipses was chosen to be the topic for our subjects'
writing. Interrater reliability on the test for fluency,
organization, and combination of this topic eclipses was
.86, .82, and .91, respectively.

Writing Performance and
Involvement QueStionnaire

 

 

Since eclipses happened to be the topic with the
largest spread of knowledge, the subjects were asked in an
in-class session to write a forty—minute composition about
eclipses. In the same session and after the subjects
finished writing their compositions, they were asked to
spend five tninutes to reSpond to a short questionnaire
about their involvement in writing about the topic (see

Appendix E).

Writing Measurement
The students' compositions were then given to the
three raters who were trained during the pilot study. The

compositions were scored for writing measures in the

 

 

 

49

categories of T—unit length, amount of subordination,
writing proficiency revision strategies, content length,
global coherence, and text sophistication.

T-Unit Length and Amount
of Subordination

 

T—units length and amount of subordination are two
measures of linguistic complexity. To measure the mean T—
unit length, the three raters were instructed to count
only T—units which were either error—free or those
containing minor errors which did not interfere with the
"readability" of the text. Second or third degree errors,
as they were classified by Nas (1975), would be the kind
of errors that might hinder the readability of the text,
and were not counted. After omitting the T—units that
contained major errors, raters counted the remaining
number of T—units and the number of words in each of the
compositions . The mean number of words per T-unit was
obtained by dividing the number of words in the essays by
the number of T-units in that essay (Hunt, 1965).

Concerning the second indicator of linguistic
complexity which is the existence of subordination, a
ratio of main clauses to all clauses was quantified. This
was done by (a) counting all clauses (subordinate and main
clauses), (b) counting only the main clauses, and (c)
dividing the number of all clauses by the number of main

clauses.

 

 

 

50

The three raters were asked to work independently.
After all compositions were rated by all the raters, a
meeting was arranged to discuss and overcome any
discrepancies among the raters (see Appendix F for

Linguistic Complexity Measures).

Writing Proficiency Scores

 

The scoring procedures outlined by the English
Language Center of Michigan State University were followed
to score the writing language proficiency of the students'
compositions. The three raters were teachers at the
Center and were experienced in applying the Center's
writing scale which was used in this study. This scale
measures clarity of the texts, language errors, vocabulary
choice, and organization (see Appendix G for English
Language Center's Writing Scale).

For scoring the subjects' compositions, the raters
read each of the compositions and assigned it a score on a
35-100 scale. The final score of each essay was the
average of the three raters‘ scores. The interrater
reliability was quantified by Michigan State University's

SPSS-X Computer Program and was found to be .82.

Revision Strategies
Revision strategies included counting internal and

external revision strategies. Using samples from the

 

 

 

 

 

51

pilot study, the researcher and the raters met several
times to discuss the way of counting the revision
strategies and to discuss differences in opinion. After
the raters had enough information about how to score the
revision strategies, they were asked to count the revision
strategies which occurred in the compositions of the
subjects (see Appendix F for categories of internal and
external revision) The count included the changing of
sentences, phrases, and words; cutting; adding;

reordering; embedding; and word choice.

Content

The length of the texts was quantified by both the
number of words and the number of central ideas contained.
The method previously used for counting the words to
measure the T-unit length was used again in this section
to determine the length of the texts. Furthermore, the
raters were asked to count the number of central ideas to
determine how fluent the text was. Then, in a three-point
scale, the raters rated each essay test as having high
fluency (3), moderate fluency (2), or low frequency (1).
Both the number of words and the number of ideas were
utilized to quantify the length of the texts in this

study. Content measure are included in Appendix F.

 

 

 

 

52

Global Coherence

The global coherence scale of Bamberg (1984) was
utilized to measure the global coherence of the texts in
this study. The lowest category in this scale is
"miscellaneous" with the highest category being “fully
coherent" (see .Appendix (3 for global coherence scale).
The three raters independently scored each text's global
coherence with the guidance of ‘this scale. The 'three
raters's scores of each text were averaged to figure the
final coherence score of each text. Interrater

reliability of the three raters was .90.

Text Sophistication

 

Before raters scored the text sophistication in
this study, the researcher met with them to discuss the
way text sophistication was to be scored. Examples from
the pilot study were used to illustrate this process. The
meetings of the researcher and the raters were also
utilized to resolve any differences in opinion among the
raters.

The text sophistication variable was measured
through raters' decisions concerning how sophisticated
each text was. Their decisions were made and quantified
by answering questions about the essays which were

concerned with the complexity and sophistication of the

 

 

 

 

 

 

 

53

terms used in the text, and the amount of definitions
explanations, and repetitions.
The measurement of this variable is contained in

Appendix F.

Involvement

 

Chesky (1984) defined involvement as the degree to
which students like their writing, find it easy, and
believe they truly have something to say to someone. The
involvement measurement in this study was designed by the
researcher and based on Chesky’s definition and Krashen's
(1982) affective factor' hypothesis. Participants ‘were
asked questions about the topic immediately after writing
their texts. The questions concerned the writers'
anxiety, confidence, interest, liking, and whether they
would write again about this topic if they had a chance.
The questions were intended to reflect their feelings
about the topic while writing. This measurement was used
in the pilot study and worked well. The obtained
reliability was a relatively high reliability of .84. The

involvement measurement is included in Appendix E.

Language Proficiency and Writing

The participants' language proficiency level was
determined by their scores on the English Language
Center's Language Proficiency Test. The participants were

asked in the questionnaire (Appendix B, Item 4) to provide

 

 

54

the latest scores they achieved on the test. Those scores
were used to examine the relationship between language
proficiency and written discourse components mentioned and

explained above.

Data Analysis

After data were collected from both students and
raters, responses were coded, perfectness and reliability
were examined. Also the interrater reliability among the
three raters was checked. The researcher used coding
procedures and computer programming assisted by of
colleagues who were experienced in this field.

The mean scores of the three raters were
quantified and used to represent the measurement of the
three prior knowledge measurements——fluency, organization,
and combination. These scores were also used to represent
the written discourse measures: T—unit length,
subordination, writing global language proficiency,
revision strategies, number of words, number of ideas,
coherence, and text sophistication.

To tabulate the demographic information of the
participants, the frequency analysis was used. The
correlation coefficient was used to compute the
relationship between the participants' demographic
information, amount of reading, prior knowledge, and

language proficiency and all the written discourse

 

 

 

 

 

55

measures. Regression was also used to explain the effects
of demographic information, amount of reading, prior
knowledge and overall language proficiency on the written
discourse measures.

Multivariate analysis of variance and Analysis of
Variance were used only to compare differences in written
discourse components according to various native-language
groups. The Tukey test was also used to compare the means
of each of the two subgroups in the written discourse
measures.

The SPSS-X statistical package was used in the
Michigan State University's computer to derive and analyze
the results. The results and the analysis are reported in

the next chapter.

Summer

This chapter lays out a road map of what was to be
accomplished in the study. The fundamental goal of this
study was to find the relationship among prior knowledge,
overall language proficiency, and ESL written discourse
measures. 'Da carry out this goal, 49 advanced students
attending the English Language Center in Spring, 1988,
participated in the study. "Eclipses" was the topic
chosen and prior knowledge was first measured. Prior
knowledge is composed of fluency, organization,

combination, and overall knowledge. A standard method was

 

 

 

 

 

 

56

followed to construct a prior knowledge test; three raters
were trained to score it. In measuring the students'
language proficiency, the English Language Center Language
Test was used. 'The students were asked to write on the
topic of eclipses. This topic was chosen from three
topics because it was found to be the topic with the
largest spread of knowledge. Subjects were also asked to
respond to a questionnaire measuring their personal
background, language proficiency, amount of reading, and
involvement. The subjects’ writing performance was
analyzed by three trained raters who scored the essays for
writing proficiency, content, text sophistication, global
coherence, and revision strategies. Both prior knowledge
and overall language proficiency were utilized to assess
their effects on the subjects' writing performance. The
effects of the subjects' demographic characteristics and
the amount of reading they do on the subjects' writing

were also assessed.

 

 

 

L________

 

 

CHAPTER IV

RESULTS

The main goal of this study was to investigate
the relationship among prior knowledge, overall language
proficiency, and written discourse of nonnative speakers
of the English language. This study also investigated
the effects of both students' background information
(age, gender, etc.) and the amount of reading they do on
their writing.

This study attempted to address the following
questions:

1. Is there any difference among the
participants' writing related to their
personal characteristics and the amount of
reading they do?

2. What is the relationship between prior
knowledge about a specific topics and the
nonnative speakers' written discourse?

3. What is the relationship between the overall
language proficiency and written discourse of

nonnative speakers?

57

 

58

4. What is the additive effect of both prior
knowledge and overall language proficiency on
nonnative speakers' written discourse?

These are important issues in teaching written
discourse to nonnative speakers and very crucial to the
field of writing. A prior knowledge test, the raters'
evaluation. of the writing' measures, the participants'
writing on the topic eclipses, and a questionnaire were
the means utilized to elicit information to answer these

questions.

Hypotheses
In this study it was hypothesized that:

Hypothesis 1: There is In: relationship between the
nonnative speakers' prior knowledge about. a
specific topic and their scores in (a) a writing
test, (b) content, (c) text sophistication, (d)
global coherence, (e) linguistic complexity, (f)
revision strategies, and (g) involvement.

 

fiypothesis 2: There is In) relationship between the
nonnative speakers' overall language profiC1ency
and (a) writing proficiency scores, (b) content,
(c) text sophistication, (d) global coherence,
(e) linguistic complexity, (f) reViSion
strategies, and (g) involvement.

 

rm: test these hypotheses, the advanced students
attending the English Language Center in Spring Term of
1988 were asked to participate in the study. As a result
of their participation, data related to prior knowledge,
overall language proficiency, and written discourse

measures were collected.

 

prior
partici
used t
knowle<
partic
Tukey
prior

the S'

chapi

59

Three independent raters were hired to score the
prior knowledge and the written discourse of the
participants. Correlation and regression techniques were
used to quantify the relationship between both prior
knowledge and overall language proficiency and the
participants' written discourse and MANOVA, ANOVA, and
Tukey test were used to measure the differences in the
prior knowledge and the written discourse measures among
the subgroups as follows:

The findings of this study are presented in this
chapter as follows:

I. The descriptions of the participants and
their writing performance. Samples of prior
knowledge tests and writing are provided in
Appendix I.

1. Characteristics of the participants

2. Characteristics of the sample performance
a. The sample's prior knowledge task
b. The characteristics of the sample's

essays

II. The results:

3. The relationship of demographic
information and amount of reading with
writing

4. Writing proficiency scores

 

 

 

par

age

ou-
ma.
0f
of
10
be

pa

60

5. Content
a. Number of words
b. Number of ideas
6. Text sophistication
7. Global coherence
8. Linguistic complexity
a. T—unit length
b. Subordination
9. Revision strategies

10. Involvement

Characteristics of the Participants

Part of the study was designed to elicit the
participants' background information, such as gender,
age, level of education, native language, and ELC scores.

As indicated in Table 1, the data disclose that
out of 49 participants, the majority 29 (59.2%) were
male, and 20 (40.8%) were female. The data show that out
of the 49 participants, 5 (10.2%) were less than 20 years
Of age, 18 (36.7%) were between 20 and 24 years of a age,
10 (20.4%) were 25-29 years of age, 8 (16.3%) were
between 30—34, 5 (10.2%) were between 35—39, and only 3
(6.1%) were 40 years of age or above.

The level of education data show that 12 (24.5%)
participants were holding diplomas from high school , 23

(46.9%) were holding degrees from college, 13 (26.5%)

61

 

 

 

Table 1. Gender, Age, Level of Education, and Native
Language
Variable F Percent
Gender
Male 29 59.2
Female 20 40.8
Age
Below 20 years 5 10.2
20—24 18 36.7
25—29 10 20.4
30-34 8 16.3
35—39 5 10.2
40 and above 3 6.1
Level of Education
High School 12 24.5
College 23 46-9
Masters Degree 13 26.5
other 1 2.5
Native Language
Arabic 6 12-2
Chinese 10 20'4
Korean 1 2-0
Spanish 3 6-1
Japanese 23 46:9
Indonesian 4 8'2
Turkish 2 4‘0

 

 

 

 

62

participants were holding master's degrees, and one
participant did not indicate the degree he was holding.

Concerning the participants' cultural background,
6 (12.2%) were Arabic-speaking students, 10 (20.4%) were
Chinese, 1 (2.0%) was Korean, 3 (6.1) were Spanish, 23
(46.9%) Japanese, 4 (8.2%) were Indonesian, and 2 (4.0%)
Turkish.

Table 2 illustrates the participants' ELC scores.
The scores reflect the average of grammar, reading,
writing and vocabulary of the official test of the ELC.
As it is summarized in Table 2, 7 (14.3%) of the
participants had scores between 65-70, 8 (16.3% had
scores between 71-75, 19 (38.8) had scores between 76—80,
10 (20.4%) had scores between 81—85, while 5 (10.2%) had

scores between 86-89.

Table 2. ELC Scores of Participants

 

 

ELC Scores F Percent
65—70 7 14.3
71—75 8 16.3
76—80 19 38.8
81—85 10 20.4

86-89 5 10.2

 

 

 

 

 

 

63

Characteristics of the Sample Performance

The Sample's Prior
Knowledge Task

This task was aimed at measuring the

 

participants' prior knowledge. As it was indicated in
the previous chapter, these raters were hired to score
the prompts representing each topic. As a result of this
procedure, scores of fluency, organization, and
combination, the three measures of the prior knowledge
were obtained. The results showed first that the
interrater reliability in rating the fluency,
organization, and combination are .71, .79, and .81,
respectively. The scores of the three raters were
averaged, and the results shown in Table 3 indicate that
for fluency, the minimum score of the sample was 1.00 and
the maximum score was 3.0 with a Inean of 1.95 and a
standard deviation of .61. For organization, the minimum
score is 6.00, the maximum score is 14.00 with a mean of
10.31 and standard deviation of 2.06. For combined
knowledge, the minimum score is 2.00, the maximum score
is 14.00 with a mean of 9.04 and standard deviation of

3.00.

The Characteristics of the

Sample's Essays

To provide background information about the

 

participants' writing ability, some characteristics of

 

 

 

64

Table 3. Characteristics of the Sample's Prior Knowledge

 

 

Task
Prior Knowledge Minimum Maximum Mean S.D.
Fluency 1.00 3.00 1.95 .61
Organization 6.00 14.00 10.31 2.06

Combination 2.00 14.00 9.04 3.00

 

their essays are explained. Among those characteristics
are the number of words, the number of T-units, the
number of clauses, the number of topics, revision
strategies (internal, external, and total), the number of
words per T—unit, the ratio of all clauses to main
clauses, coherence scores, and language proficiency
scores.

As indicated in Table 4, the data reveal that a
the minimum number of words in the essays is 82 words,
while the maximum number is 445 words with a Kean of
265.55 and 23 standard deviation of 93.68. The minimum
number of total T—units is 7, while the maximum number is
43 with a mean of 25.02 and a standard deviation of 8.46.
The essays have a minimum number of 7 for the central T—
units, maximum number of 42, with a mean of 24.69 and a
standard deviation of 8.34. But for the peripheral T—
units, the essays have the minimum number of 1 and the

maximum number of 5 with a mean of 1.10 and a standard

 

 

 

 

65

Table 4. Characteristics of the Sample's Essays

o

 

Variables Minimum Maximum Mean S.D.
No. Of Words 82 445 265.55 93.68
Total of T—units 7 43 25.02 8.46
No. of Central T—units 7 42 24.69 8.34
No. of Peripheral

T—units l 5 1.10 .59
No. of Clauses 10 59 33.20 11.02
No. of Main Clauses 5 36 21.06 7.26
No. of Topics per Essay 1 7 4.00 1.40
No. of Internal Revision 0 9 2.10 2.23
No. of External Revision 0 72 14.61 13.39
No. of Total Revision 0 76 16.71 14.13
Linguistic Complexity 7 15.45 10.68 1.84
Subordination 1.20 2.22 1.61 .24
Global Coherence 2 4 3.40 54
Writing Proficiency

Scores 66.33 83 75.32 3.72

 

 

 

 

 

 

66

deviation of .59. Concerning the number of all clauses,
the essays have a [minimum number of 10 and a Inaximum
number of 59, with a mean of 33.20 and standard deviation
of 11.02. For main clauses, the minimum number of main
clauses is 5, and the maximum number is 36, with a mean
of 21.06 and standard deviation of 7.26. The minimum
number of topics in each. essay is 1, and the Inaximum
number is 7, with a mean of 4.00 and standard deviation
of 1.40.

The revision variable is divided into three
components: internal revision, external revision, and
total revision. The data show that the minimum number of
internal revision is 0 (zero), and the maximum number is
9, with a mean of 2.10 and standard deviation of 2.23.
In external revision, the minimum number is also 0
(zero), while the maximum number of total revision is 76,
with a mean of 16.71 and standard deviation of 14.13.

The linguistic complexity computation shows that
the minimum number of words per T—unit is 7, and the
maximum number per T—unit is 15.45, with a mean of 10.68
and standard deviation of 1.84. In subordination, the
minimum ratio of all clauses to main clauses is 1.20, and
the maximum number is 2.22, with a mean of 1.61 and a
standard deviation of .24. Coherence was rated

separately, and the minimum score of coherence is 2,

 

 

 

 

 

 

67

while the maximum coherence score is 4 with the mean of
3.40 and standard deviation of .54.

The sample's essays were rated for the writing
proficiency scores. The data show that the minimum score
for this variable is 66.33, and the maximum score is 83
with a mean of 75.32 and a standard deviation of 3.72.

Relationships of Demographic Information
and Amount of Reading with Writing

 

 

Schema theory generally does not emphasize the
rules of age, level of education, gender, and amount of
reading on reading comprehension” nor have the studies
applying this theory to writing also included these
variables. It, therefore, seemed important to have these
variables contained in our model. The question addressed
by this inclusion is: What is the relationship among the
participants' demographic information, amount of reading,
and writing discourse measures?

Concerning the relationship between the
participants' demographic information and written
discourse measures, Table 5 shows that there was a
significant positive relationship between age and number
of words (.279, p<.03), number of ideas (.309, p<.02),
and involvement (.331, p<.Ol). This means that the
greater the age, the longer the participants' texts, and

the more the participants were involved with their texts.

 

 

 

 

 

 

68

Table 5. The Relationship of Age, Level of Education,
and Amount of Reading with Written Discourse

 

 

 

 

 

 

Measures

Writing Measure Age Level of Gender Amount of
Education Reading

Linguistic Complexity
T—unit Length .077 .014 .189 .005
Subordination -.019 —.181 .065 .193
Writing Proficiency
Scores .068 -.052 —.047 .027
Revision Strategies —.131 —.042 .046 .341**
Content
No. of Words .279* -.042 .046 .341**
No. of Ideas .309* .072 —.142 —.010
Global Coherence .088 —.010 -.156 -.012
Text Sophistication .137 -.049 -.128 -.060
Involvement .331** .241* -.240 .341**

 

*Significant at p<.05.

**Significant at p<.01.

 

 

 

 

69

The table also shows that there was a positive
significant relationship between the level of education
and involvement (.241, p<.05). That is, the higher the
participants' educational degrees, the more involved they
were with their texts.

Concerning the relationship between the amount of
reading and written discourse measures, Table 5 shows
that there was a significant positive relationship
between the amount of reading and revision strategies
(.341, p<.05) and involvement (.341, p<.008).

The multiple regression reveals that together
age, level of education, gender, and amount of reading do
not significantly account for the variance in any of the

written discourse measures.

Comparisons by Native Langpage

Is there any difference among the participants'
writing related to their native language?

MANOVA and ANOVA were used to determine if a
difference exists between the population means of writing
variables among the subgroups (Arabic, Chinese, Japanese,
Spanish, and Indonesian). The writing variables are T—
unit length, subordination, global coherence, text
sophistication, writing proficiency scores, involvement,
number of words, and number of ideas. As the ANOVA

results showed, as indicated in Table 6, only the mean of

 

 

 

70

Table 6. MANOVA. and ANOVA. Results for the .Difference
Between the Population Means Writing Measures
by Native Language

Variables ANOVA—F ANOVA—P MANOVA—P

Writing Variables .02*

T—unit Length 2.212 .08
Subordination 4.610 .004**
Writing Proficiency Scores 1.03 .39
Number of Words .625 .64
Number of Ideas .909 .46
Global Coherence 1.428 .24
Text Sophistication .806 .52
Involvement 2.448 .06

 

 

*Significant at < .05.

**Significant at < .01.

 

 

7l

subordination was found to be significantly different at 5
= .01 according to group native language. The MANOVA P—
value shows that the native language has a significant
effect on writing because the overall difference among the
means of all variables according to group native language
is significant.

Analysis of variance p—values and the means of the
variables which were found to be significantly different
are shown in Table 7. Tukey's Test indicates that Spanish
subjects have a higher organization and combination means
than those of the Chinese subjects. The results also show
that the Arabic and the Spanish subjects had higher
subordination than the Chinese and the Japanese. ANOVA P-
values and the means of the variables that were found not

significant are shown in Table 8.

Writing Proficiency Scores

Writing Proficiency Scores
and Prior Knowledge

Hypothesis 1.A: There is no relationship between the
nonnative speakers' prior knowledge and writing
proficiency scores of their essays.

Table 9 shows that there was a positive
significant relationship between overall prior knowledge
and writing proficiency scores (.485, p<.001) for all the

groups. In. addition, there was ea positive significant

relationship shown between writing proficiency scores and

 

 

 

 

 

 

72
Table 7. Means of Organization, Combination, and
Subordination by Native Language.
Group N Organization Combination Subordination
Arabic 6 11.83 11.11 1.85
Chinese 10 9.70 8.30 1.47
Japanese 23 9.89 8.52 1.56
Spanish 3 12.66 11.88 1.85
Indonesia 4 11.58 11.25 1.66
.02* .04* .004*

ANOVA P VALUE

 

*Significant at p<.05.

 

 

73

 

 

Table 8. Means of Writing Proficiency Scores, Number of
Words, NUmber of Ideas, Text Sophistication,
Global Coherence, T—Unit Length, Revision
Strategies, and Involvement of Writing
Components by Native Language.

Group Arabic Chinese Japan— Spanish Indon— ANOVA

ese esian Values

N 6 10 23 3 4

Lan-

guage

Profi-

ciency 73.05 76.93 75.37 75.88 75.83 .39

Number

of Words 282 293 248 268 294 .64

Number

of Ideas 2.16 2.60 2.26 2.66 2.75 46

Text

Sophis—

tication 12.83 12.40 11.73 12.66 12.75 .52

Global

Coherence 3.33 3.76 3.30 3.22 3.50 .25

T-unit

Length 10.04 11.12 10.41 13.36 10.45 .08

Revision

Strate—

gies 17.7 14.8 17.8 10.6 27.5 .06

Involve—

ment 14.50 13.00 10.47 12.33 15.25 .06

 

 

 

74

 

 

 

 

 

 

 

 

 

 

Hoo. axes
Ho. max
mo. 64
how. omv. *xoww. mam. mfiowflmaawomflz
«ssmhh. *xromn. *mwm. *rrmmb.. wwaMQMb
mHm. th. *NHO. hvm. wwwGHSU
*NMP. row. xmwh. *hwh. MHQMH¢
*sxwmv. xxxbmv. #rhmm. «*xmmv. mQSOHO Ham
mmHOUm meOUm WQHOUW \w WOHOUW
xocwflOHMOHm hocwflOHmOHm hocwfloflwonm ozwwmwmwmm mQDOH
msfluflyz msHUHHZ msHDHMB \mmpomecm w
\soflwmsflnfioo \coflpmmflcmmuo \mocosHm HHoHo>O

 

monoom mocoflOEMOHm msflufluz can ompmHBOsM soflnm soospom soflpmHoHMOU

.m manna

 

 

75

fluency (.357, p<.01), organization (.457, p<.001), and
combination (.494, p<.001).

For the Arabic subgroup, there was a positive
significant relationship between overall prior knowledge
and writing proficiency scores (.767, p<.05). There was
also a positive significant relationship between writing
proficiency scores and fluency (.748, p<.05),
organization (.80, p<.05), and combination (.732, p<.05).

For the Chinese subgroup, there was a positive
Significant relationship only between fluency and writing
proficiency scores (.612, p<.05).

For the Japanese subgroups, there was a positive
Significant relationship between overall prior knowledge
and writing proficiency scores (.755, p<.001). There was

also a positive significant relationship between writing

proficiency scores and fluency (.363, p<.05),
Organization (.730, p<.001), and combination (.772,
p<.001).

For the miscellaneous group, there was a positive
Significant relationship only between fluency and writing
proficiency scores (.696, p<.01).

On the other hand, looking at the same table, we
conclude that there was not any significant relationship
between prior knowledge as a total and writing
prOficiency scores for Chinese and miscellaneous

subgroups, between organization and writing proficiency

76

scores for Chinese and miscellaneous subgroup, or between
combination and writing proficiency scores for Chinese
and miscellaneous subgroups.

Writing Proficiency Scores
and Overall Language

Proficiency
Hypothesis 1.B: There is no significant relationship
between overall language proficiency and writing
proficiency scores.
Table 10 shows that there was a positive
significant relationship between overall language

proficiency and global language proficiency (.333, p<.01)

for all the group.

 

 

 

Table 10. Correlation Between Overall Language
Proficiency (OLP) and Writing Proficiency
Scores

Groups OLP/Writing Global Proficiency

All Groups .333**

Arabic .279

Chinese .289

Japanese .335*

Miscellaneous .401

*p < .05

**p < .01

 

 

77

For the Japanese subgroup, there was also a
positive significant relationship between overall
language proficiency and writing proficiency scores
(.335, p<.05).

On the other hand, looking at the same table, we
find no relationship between overall language proficiency
and writing proficiency scores for the Arabic, the

Chinese, and the miscellaneous subgroups.

The Additive Effect

 

The question related to the additive effect is:
What is the additive effect of prior knowledge and
overall language proficiency on writing proficiency
scores?

The results of using regression analysis
indicated in Table 11 show that prior knowledge was a
Significant predictor for writing proficiency scores at
the p<.01 level. The results also show that 29.2% of the
total variance in writing proficiency scores was
accounted for by both variables. Stepwise regression
analysis show that 23.5% of the total variance in the

writing proficiency scores was accounted for by the prior

knowledge alone. The overall language proficiency,
however, is a marginal predictor for the writing
proficiency scores at the p<.0591) level. So, the

results of the relationship between prior knowledge and

 

 

78

Table 11. Regression Results of Writing Proficiency
Scores (n = 49)

 

 

Variables Coefficient Significance
Overall Knowledge .2968 .0013*
Language Proficiency .1579 .0591
Constant 56.7116 .0000**
Multiple R = .5408

R-SQR = .2925

 

*p < .01; **p < .001

writing proficiency scores show that there was a
significant positive relationship between all prior
knowledge scores and writing proficiency scores for
almost all the group and all the subgroups.

The very few instances where there is lack of
significant relationship between prior knowledge and
writing proficiency scores do not keep us from concluding
that the relationship between prior knowledge and writing
proficiency scores is significant.

Concerning the results of the relationship
between overall language proficiency and writing
proficiency scores, we find that they were significantly
related for the whole group and for the Japanese subgroup

only.

 

 

79

However, prior knowledge accounts for more
variance on the writing proficiency scores than overall
language proficiency.

Content: Number of Words and Number
of Ideas

 

Content and Prior Knowledge

 

Hypothesis 1.B: There is no relationship between the
nonnative speakers' prior knowledge and the
content: number of words and number of ideas.

Number of Words

Table 12 shows that there was a positive
significant relationship between overall prior knowledge
and the number of words (.406, p<.01) for all the groups.
There was also a positive significant relationship
between the number of words and fluency (.410, p<.01),
organization (387, p<.01), and combination (.387, p<.01)
for all the groups.

For the Arabic group, there was a positive
significant relationship between overall prior knowledge
and the number of words (.694, p<.05). There was also a
positive significant relationship between the number of
words and organization (.720, p<.05), and combination
(.676, p<.05) for the Arabic subgroup.

For the Chinese subgroup, there was a positive
significant relationship between prior knowledge and the

number of words (.719, p<.05). There was also a positive

 

 

 

8O

 

HOO. v asks

 

 

 

 

HO. v firs
mo. v a.
w¥mmh. NQH. xsmwh. VON. ixowh. NVN. kxhow. HaN. mSOwCMHHOUmHZ
«sovm. «mow. rimmv. «hmm. MNN. «#QHV. «*NNm. «hmv. OWOCQQMU
«mow. i«mmh. smnm. *«QHN. ««Hmw. «VNQ. «mam. «mHh. mmOCHLU
«NHQ. «WPQ. sromw. «ONF. smmh. Omw. immm. «v00. UflQflH4
«ssmmm. 445mm. «*«hom. «sham. «soflv. ¥«OH¢. «irmvm. *xoov. QDOHO Had
mmmpH mpuoz mmwvH mpuoz mmoUH mpuoz mmopH mpuog
msouo
coflHMCHQEOU COflUMNHcmmHO kocosam mmpoa3ocM Hfimum>o

 

mmmpH m0 qu532

pcm mono: Mo quEsz

msouo :omm can mmsouo Ham now

.ucoucoo pcm ompmazocx Hoaum coozuom cowumHoHHoo .NH manna

 

 

 

 

81

significant relationship between the number of words and
fluency (.624, p<.05), organization (.716, p<.05), and
combination (.728, p<.01) for the Chinese subgroup.

For the Japanese subgroup, there was a
significant positive relationship between prior knowledge
and the number of words (.427, p<.05). There was also a
significant positive relationship between the number of
words and fluency (.416, p<.01), organization (.387,
p<.05), and combination (.403, p<.05) for the Japanese
subgroup.

On the other hand, looking at Table 12 we find
that there was not any significant relationship between
fluency and the number of words for the Arabic subgroup.
We also found no significant relationship between prior
knowledge as total or as components and the number of

words for the miscellaneous subgroup.

The Number of Ideas

Table 12 also shows that there was a significant
positive relationship between overall prior knowledge and
the number of ideas (.542, p<.001) for all the group.
There was also a positive significant relationship
between the number of ideas and fluency (.410, p<.01),
organization (.507, p<.001), and combination (.553,

p<.001) for all the group.

 

 

 

82

Table 13. Correlation Between Overall Language
Proficiency (OLP) and Content: Number of
Words and Number of Ideas

 

 

 

 

Groups OLP/Content
Number of Words Number of Ideas
All Groups .298* .166
Arabic .284 .629
Chinese .776** .439
Japanese .348* .177
Miscellaneous .201 .250
*p <.05
**p < .001

For the Arabic subgroup, there was a significant
positive relationship between overall prior knowledge and
the number of ideas (.828, p<.05). There was also a
positive significant relationship between the number of
ideas and fluenCy (.759, p<-05). organization (.850,
p<.01), and combination (.812, p<.05) for the Arabic
subgroup.

For the Chinese subgroup, there was a positive
significant relationship between overall prior knowledge
and the number of ideas (.519, p<.05). There was also a

significant relationship between the number of ideas and

 

 

 

 

 

83

fluency (.731, p<.01), organization (.573, p<.05), and
combination (.498, p<.05) for the Chinese subgroup.

For the Japanese subgroup, there was a positive
significant relationship between overall prior knowledge
and the number of ideas (.522, p<.01). There was also a
positive significant relationship between the number of
ideas and organization (.495, p<.01), and combination
(.546, p<.01) for the Japanese subgroup.

For the miscellaneous subgroup, there was a
positive significant relationship between overall prior
knowledge and the number of ideas (.807, p.<01). There
was also a positive significant relationship between the
number of ideas and fluency (.760, p<.01), organization
(.763, p<.01), and combination (.785, p<.01) for the
miscellaneous subgroup.

On the other hand, looking at Table 12, we found
that there was not a significant relationship between
fluency and the number of ideas only in the Japanese

subgroup.

Content and Overall Language Proficiency
Hypothesis 2—B: There is no relationship between
overall language proficiency and content: number
of words and number of ideas.
Number of Words

Table 13 shows that there was a positive

significant relationship between overall language

 

 

 

 

 

 

 

 

84

proficiency and number of words (.298, p<.05) for all the
groups. There was also a positive significant
relationship between overall language proficiency and
number of words (.776, p<.01) for the Chinese subgroup.
A positive significant relationship also existed between
overall language proficiency and number of words (.348,
p<.05) for the Japanese subgroup.

On the other hand, there was no significant
relationship between overall language proficiency and the
number of words for the Arabic subgroup, and the

miscellaneous subgroup.

The Number of Ideas
Table 13 shows that there was no significant
relationship between overall language proficiency and the

number of ideas for all the group or for the subgroups.

The Additive Effect

The question related to the additive effect is:
What is the additive effect of prior knowledge and
overall language proficiency on content: number of words
and number of ideas?

The results of regression indicated in Table 14
for the number of words show that overall knowledge is a
significant predictor for the number of words at p<.01,
while the overall language proficiency is not. Both

variables account for 21.2% of the total variance in the

 

 

 

 

 

 

85

Table 14. Regression Results of the Content: Number of
Words (n = 49)

 

Variables Coefficient Significance

 

 

Overall Knowledge 6.1842 .0098**
Language Proficiency 3.6465 .0993
Constant —.0011 .3717
Multiple R = .4614

R—SQR = .2129

**p < .01.

number of words. A stepwise regression analysis shows

that prior knowledge alone accounts for 16.4% of the
total variance in the number of words.

The results of regression indicated in Table 15
for the number of ideas show that the overall prior
knowledge as a significant predictor for the number of
ideas at p<.001. Both variables account for 29.7% of the
total variance in the number of ideas. A stepwise
regression analysis shows that the overall prior
knowledge variable alone accounts for 29.4% of the total
variance ill the number of ideas. The overall language
proficiency was not a significant predictor for the
number of ideas.

So, the results of the relationship between prior

knowledge and content indicate that prior knowledge had a

 

 

 

 

 

 

86

Table 15. Regression Results of the Content: Number of
Ideas (n = 24)

 

 

 

Variables Coefficient Significance
Overall Knowledge .0726 .0001***
Language Proficiency .0075 .6473
Constant

Multiple R = .5451

R-SQR = .2971

***p < .01

strong positive relationship with content: number of
words and number of ideas. The overall prior knowledge

was also found to be a significant predictor for both
number of words and number of ideas.

On the other hand, the overall language
proficiency had no relationship with content. It was not

a significant predictor for content either.

Text Sophistication

 

Text sophistication and prior knowledge:
Hypothesis 1-C: There is no relationship between the
nonnative speakers' prior knowledge and text
sophistication.
Table 16 illustrates that there was a pmsitive
significant relationship between overall knowledge and

text sophistication (.283, p<.05) for the whole groups.

There was also a positive significant relationship shown

 

 

 

 

87

Ho. v Q44
VQ¥

 

 

mo.

«mmw. «mam. ««mmn. *Hmm. msoocmHHoomflz

SN. mom. mmo. SN. omocmaae

NHq. omv.| HHa.1 smm.| omocHLU

mmw. mmm. smmw. qmm. oflhmud

Lam. 2.5.. mam. «SN. 3:05 2a
cofluoOHumHsQom :oHumOHpmflcmom :oflumoflumflsmom coflpmoflumflsmwm

uxoe uxoe uxoe u H masono
\coHumcH Eou coaummacmmuo \xoco: \ompoa30cx
. .n \ . . Hm Hamuo>o

 

How cofluMOHumflsmom uxoe pom mmpoazocx uoHum cooBuom COHumaouuou

dsouo comm ocm mmoouo HH<

.wH THEME

 

 

88

between text sophistication and organization (.271,
p<.05) and combination (.281, p<.05) for all the groups.

For the Arabic subgroup there was a positive
significant relationship between fluency and text
sophistication (.689, p<.05).

For the miscellaneous subgroup, there was a
positive significant relationship between overall prior
knowledge and text sophistication (.661, p<.05). There
was also a positive significant relationship between text
sophistication and fluency (.729, p<.01), organization
(.617, p<.05), and combination (.626, p<.05).

Looking at Table 16, we see that there was no
significant relationship between fluency and text
sophistication for all the groups. For the Arabic
subgroup, there was no significant relationship between
text sophistication and overall knowledge, organization,
or combination. For the Chinese and the Japanese
subgroups, there was no significant relationship between
prior knowledge total or any components with text

sophistication.

 

 

89

Text Sophistication and Overall
Language Proficiency

Hypothesis 2—C: There is no relationship between

overall language proficiency and text
sophistication of the nonnative speakers' written
discourse.

Table 17 shows that there was a positive
significant relationship between overall language
proficiency and text sophistication (.394, p<.05) for the

whole group.

Table 17. Correlation Between Overall Language
Proficiency and Text Sophistication

 

 

 

Groups OLP/Text Sophistication
All Groups .394*
Arabic .863**
Chinese —.633*
Japanese .098
Miscellaneous .137

*p < .05
**p < .01

For the Arabic subgroup, there was also a

positive significant relationship between overall
language proficiency and text sophistication (.863,
p<.01). However, for the Chinese subgroup, there was a

negative significant relationship between overall

90

language proficiency and text sophistication (—.633,
p<.05).

On the other hand, for the Japanese and the
miscellaneous subgroups, there was no significant
relationship between overall language proficiency and

text sophistication.

The Additive Effect
The question related to the additive effect is:

what is the additive effect of prior knowledge and
overall language proficiency on text sophistication?

The results of regression as indicated in Table
18 show that overall knowledge is marginally related to
text sophistication at p<.0528. Looking at the same
table, we notice that 8.05% of the total variance on the
text sophistication was accounted for by overall
knowledge. The overall language proficiency does not
significantly contribute to the prediction of text

sophistication.

 

 

 

91

Table 18. Regression Results of Text Sophistication

 

 

 

Variables Coefficient Significance
Overall Knowledge .0445 .0528
Language Proficiency .0420 .8979
Constant 3.2280 .0018**
Multiple R = 2838
R-SQR = .0805
**p < .01.

Also there was a positive significant

relationship between overall language proficiency and
text sophistication for all the groups, and the Arabic
subgroup.

For the Chinese subgroup, there was a negative
significant relationship between overall language
proficiency and text sophistication.

The additive effect results show that overall

knowledge must be considered a marginal predictor of the

text sophistication.

Global Coherence

Global Coherence and
Prior Knowledge
Hypothesis 1—D: There is no significant relationship
between the nonnative speakers' prior knowledge
and global coherence of their essays.

92

Table 19 shows that there as a positive
significant relationship between overall knowledge and
global coherence (.273, p<.05) for all the group. There
was also a positive significant relationship between
global coherence and fluency (.307, p<.05) and
combination (.268, p<.05) for all the group.

This is also exactly the same for the Japanese

There was a positive significant relationship

subgroup.
between overall knowledge and global coherence (.391,
p<.05) for this subgroup. There was also a positive

significant relationship between global coherence and
fluency (.413, p<.05) and combination (.377, p<.05) for
the Japanese subgroup.

On the other hand, looking at the same table, we
find that there was no significant relationship between
organization and global coherence in all the groups. For
the Arabic, Chinese, or miscellaneous subgroups, there
was no significant relationship between global coherence
and overall knowledge, or its three components. For the
Japanese subgroup, only organization did not have a
significant correlation with global coherence.

Global Coherence and Overall
Lppguage Proficiency

Hypothesis 2—D: There is no significant relationship
between overall language proficiency and global
coherence in the nonnative speakers' written

discourse.

 

 

 

93

 

 

 

 

 

 

 

 

 

 

 

 

mo. v as
hem. Ham. Ham. «mm. msoocoaaoomflz
seem. cam. emae. *Hmm. omocmmmb
mva. NNH. mmfi. oma. omoCHsU
owe. mom. 0mm. Ham. oanmud
«mmm. mmm. .som. .msm. mnsouo sea
mucouosoo
mucouosoo oocouocou ooconocou
\cofluoCHQEoo \cofiummflcmquo \mososam \omwwmwmmw mesouw

 

msouo comm use
mQSOHU Ham How mononosoo HoQon can ompoazocm Hoflhm coozbom coaumaonuoo .mH wanna

 

 

94

Table 20 shows that there was no significant
relationship between overall language proficiency and
global coherence the whole group, the Arabic, the

Chinese, the Japanese, or the miscellaneous subgroups.

Table 20. Correlation Between Overall Language
Proficiency and Global Coherence

 

OLP/Global Coherence

 

Groups

All Groups .099
Arabic —.316
Chinese .196
Japanese -.034
Miscellaneous .432

 

The Additive Effect
The question related to the additive effect is:

What is the additive effect of prior knowledge and

overall language proficiency on global coherence?

The results of regression indicated in Table 21
show that neither prior knowledge nor overall language
proficiency were a significant predictor for global

coherence.

 

 

95

Table 21. Regression Results of Global Coherence

 

 

 

(n = 49)
Variables Coefficient Significance
Overall Knowledge .0144 .0753
Language Proficiency .0136 .7509
Constant 1.0460 .0209
Multiple R = .2763
R-SQR = .0763

 

So, the results of the relationship between prior
knowledge and global coherence show that these two
variables are significantly related for all the groups
and for the Japanese subgroups with the exception of
organization. The prior knowledge and global coherence
are not significantly related for the Arabic, the
Chinese, and the miscellaneous subgroups.

Overall language proficiency, on the other hand,
did not have a significant relationship with global
coherence for all the groups or with any of the
subgroups.

None of the prior knowledge or overall language
proficiency scores was a significant predictor for global

coherence.

96

Linguistic Complexity: T—Unit
Length and Subordination

Linguistic Complexity and
Prior Knowledge

Hypothesis 1-E: There is no relationship between the
nonnative speakers' prior knowledge and T—unit
length and subordination in their essays.

 

Table 22 shows that there was a positive,
significant relationship between overall prior knowledge
score and T-unit length only for the Chinese subgroup
(.662, p<.05). Also, the prior knowledge are broken down
into the three components of fluency, organization, and
combination there is a statistical significant
relationship between each component with T—unit length.
That is, T—unit length is significantly correlated with
fluency (.776, p<.01), organization (.611, p><.05), and
combination (.633, p<.05).

There is also a statistically significant
relationship between organization and T—unit length
(.282, p<.05) only for the whole group. There is also a
negative significant relationship between prior knowledge
and subordination (-.819, p.<05) in the Arabic subgroup.

On the other hand, looking at the same table we
can conclude that there was not any statistical
Significant relationship between the two components of
linguistic complexity and the overall knowledge as a
total or the prior knowledge components for all the group

or for any of the subgroupS-

 

97'

 

 

 

 

 

 

 

 

 

HO. v Qua
mo. v a.

Low. Ham. mmm. mam. woo. mam. mQN. wam. moomcmaaoomaz
smo.s Ham. aqo.u mv~. coo. «06.- smo.u ova. omocmamo
mam.- «mmo. Ham. .Hao. 6mm. 446ss. mam. «moo. mmmcaro
.mam.- mma. ms~.- can. Hmo.n mmm. #65.) non. cabana
was. mmm. ese. .mmm. ems. men. doe. mom. mosouo Ada

:Oaumcflouonsm ”NWMWM sewum2wouonsm ”NWHWM :oflumcfipuonsm ”wmmwm :oHumsaouonsm ”HWMWN mm20u
. o

:anncaneoo :ofiumnacmmuo >ocosam ompoazocx aaouo>o

 

can :uocoq uflcste

msouo comm can masouo Ham uOu :OAumceouonsm

u>nwxoameou oHumaso:MA use ompmazosx newum coozuom codenaouuou

.NN OHQMB

 

 

 

98

Linguistic Complexity and Overall
Language Proficiency

 

Hypothesis 2-E: There is no significant relationship
between overall language proficiency and
linguistic complexity.

Table 23 shows that there was a positive

significant relationship between overall language

Table 23. Correlation Between Overall Language
Proficiency (OLP) and Linguistic Complexity

 

Linguistic Complexity

 

 

 

Groups
OLP/T-unit Length OLP/Subordination
All Groups .438*** .172
Arabic .136 —.792*
Chinese -.021 -.513
Japanese .514* .406*
Miscellaneous .491 .287
*p < .05

***p < .001

proficiency scores and T—unit length (.438, p<.001) for
all the groups. There is also a negative significant
relationship between overall language proficiency and
subordination (—.792, p<.05) for the Arabic subgroup.
For the Japanese subgroup, there is a positive

significant relationship between overall language

 

 

99

proficiency scores and T—unit length (.514, p<.05) and
subordination (.406, p<.05).

Looking at the same table, we find that there was
not any relationship between overall language proficiency
and T—unit length for the Arabic, the Chinese, and the
miscellaneous subgroups. There was not any relationship
between overall language proficiency and subordination
for all the groups, the Chinese, thee miscellaneous

subgroups either.

The Additive Effect

The question related to the additive effect is:
What is the additive effect of prior knowledge and
language proficiency on the linguistic complexity: T—
unit length and subordination?

The results of regression for T-unit length are
indicated in Table 24. The table shows that overall
language proficiency was a significant predictor for T-
unit length at p<.01, while the overall knowledge is not.
Regression results also show that 22.4% of the total
variance in the T—unit length was accounted for by both
variables. A stepwise regression analysis shows that
19.2% of the total variance in the T—unit length. was

accounted for by the overall language proficiency alone.

 

 

100

Table 24. Regression Results of the T—Unit Length
(N= 49)

 

 

Variables Coefficient Significance
Language Proficiency .1278 .0041**
Overall Knowledge .0621 .1714
Constant -.5866 .8572
Multiple R = .4738

R—SQR = .2245

 

**Significant at p < .01.

Table 25 shows that overall knowledge and overall
language proficiency were not significant predictors for
subordination.

So, the results of the relationship between prior
knowledge and linguistic complexity show that there was a
significant relationship between prior knowledge as a
whole and as components and T-unit length only for the
Chinese subgroup.

The results of the relationship between overall
language proficiency and linguistic complexity show that
there was a significant relationship between overall
language proficiency and T—unit length for the whole
group, and for the Japanese subgroup. There was also a
significant relationship between overall language

proficiency and subordination for the Japanese subgroup,

 

 

 

 

 

 

101

Table 25. Regression Results of Subordination

 

 

(N= 49)
Variables Coefficient Significance
Overall Knowledge .0058 .3267
Language Proficiency .0061 .3753
Constant 1.0060 .0390
Multiple R = .2152
R-SQR = .0436

 

but a significant negative relationship for the Arabic
subgroup.

The results of the effect of the interaction
between prior knowledge and overall language proficiency
on linguistic complexity show that only overall language
proficiency was considered to be a significant predictor

for T—unit length.

Revision Strategies

 

Revision Strategies and

Prior Knowledge

Hypothesis 1-F: There is no relationship between the
nonnative speakers' prior knowledge and the
number of revision strategies found in their
essays.

 

Table 26 shows that there was a positive

significant relationship between revision strategies and

 

 

102

 

 

.mo. v a.

 

 

whm. Hmm. mom. 0mm. msoocoaaoomflz

ooo. omo. mos. ooo. mmmcmomo

ofia.- mos.- owe.) mos. mmocaao

loos. «moo. mom. Hos. oenmua

oos. hos. ono. one. mosouo Ana
cofimfl>om coflmfl>om soflmfl>om :oflmfi>om

Hobos Hobos Hobos m Hobos ozone
\oOaumoanaoo \coanmuncmono \soowsnn \6 amazocx
HHMHo>O

 

Ham now moflmmuonum coamfl>om can

msouu comm pom masouw

ompoazocx Hoflum cooKuom soaumaouuoo .mm oases

103

organization (.769, p<.05), and combination (.740, p<.05)
for the Arabic subgroup.

No other significant relationship was found
between prior knowledge as total or between the prior
knowledge components and revision strategies for all the
groups or for any other subgroup.

Revision Strategies and Overall
Language Proficiency

 

 

Hypothesis 2—F: There is no relationship between
overall language proficiency and revision
strategies.

Table 27 shows that there was a significant
positive relationship between overall language
proficiency and revision strategies (.873, p<.01) for the

Arabic subgroup only.

Table 27. Correlation Between Overall Language .
Proficiency (OLP) and Revision Strategies

 

 

 

Groups OLP/Revision Strategies
All Groups .122

Arabic .873**

Chinese .439

Japanese -.017

Miscellaneous .212

 

**p < .01

104

No other significant relationship was found
between overall language proficiency and revision

strategies for all the group or for any other subgroup.

The Additive Effect

 

The question related to the additive effect is:
What is the additive effect of prior knowledge and
overall language proficiency on revision strategies?

The results of regression indicated in Table 28
shows that neither prior knowledge nor overall language
proficiency were a significant predictor for revision
strategies.

So, the results of the relationship between prior
knowledge and revision strategies show that there was a
positive significant relationship between revision
strategies and organization and combination for the
Arabic subgroup.

The results of the relationship between overall
language proficiency and revision strategies show that
there was a positive significant relationship between
revision strategies and overall language proficiency only
for the Arabic subgroup.

Neither prior knowledge nor overall language
proficiency are significant predictors for revision

strategies.

 

 

105

Table 28. Regression Results of Revision Strategies (n =

 

 

 

 

9)
Variables Coefficient Significance
Overall Knowledge .3928 .3101
Language Proficiency .2249 .5370
Constant —9.1572 .7427
Multiple R = .1925
R-SQR = .0370
Involvement

Involvement and Prior Knowledge

Hypothesis 1-G: There is no significant

 

relationships between the nonnative speakers'
prior knowledge and their involvement to writing.

Table 29 shows that there was a positive
significant relationship between overall knowledge and
involvement (.272, p<.05) for the whole group. There is
also a positive significant relationship between
involvement and organization (.264, p<.05) and
combination (.274, p<.05) for all the groups.

Looking at the same table we notice that there
was no significant relationship between prior knowledge

as a whole or as components and involvement for any

subgroup.

 

 

106

 

 

 

mo. v a.

moo. Hem. mmo. eve. mooocmaaoOmHz

oma. NNH. omo.- sag. mmmcmomo

Hem. mam. mom. Nmm. omoCch

HmH.I one.) omH.| mNH. mahnu<

.osm. .oom. oma. .Nom. monouo Has
usoEw>Ho>cH ucoeo>ao>sH bcoEo>Ho>cH ustoWHome

\coflDMCHQEOO \coflbmuflcmmno \wocosHm \ Wawwo>w masonw

 

moouo room can
mmsomw Ham MOM usoEo>Ho>cH can ompoaaocm Hoosm coosuom GOHuMHoHMOU .mm wands

 

 

 

107

Involvement and Overall
Language Proficiency

Hypothesis 2—G: There is no relationship between
overall language proficiency and involvement.

Table 30 shows that there was no significant
relationship between overall language proficiency and
involvement for all the groups or for any of the

subgroups.

Table 30. Correlation Between Overall Language
Proficiency and Involvement

 

 

Groups Involvement
All Groups .275
Arabic —.062
Chinese .485
Japanese .104
Miscellaneous —.141

 

The Additive Effect
The question related to the additive effect is:

What is the additive effect of overall prior knowledge
and overall language proficiency on involvement.

The results of' regression indicted in 'Table 31
show that neither overall knowledge nor overall language

proficiency significantly contribute to the prediction of

involvement.

 

 

 

108

Table 31. Regression Results of Involvement

 

 

Variables .. Coefficient Significance
Overall Knowledge .1950 .0734
Language Proficiency .0232 .8179
Constant 6.0294 .4384
Multiple R = 2740

R—SQR = .0751

 

So the results of the relationship between prior
knowledge and involvement show that these two variables
were not significantly related for all the groups with
the exception of fluency. Fluency did not have a
significant relationship with involvement for any of the
groups. No other significant relationship was found
between prior knowledge and involvement for any of the
subgroups.

The overall language proficiency also had no
significant relationship with involvement. FUrthermore,
neither prior knowledge nor overall language proficiency

were a significant predictor for involvement.

Summary
Concerning the relationship between the
demographic information (age, gender, and level of

education) the results of the correlational analysis

demonstrated a statistically significant relationship

 

 

109

between age and content and involvement and between level
of education and involvement. The results also show a
significant relationship between amount of reading and

involvement. This means that the older the students, the

more they write and the more they are involved with their

writing. Also, the more they read, the more they were

involved with their writing.
Concerning the relationship between prior
knowledge and writing measures, we conclude that there was

a significant relationship between prior knowledge and

writing proficiency scores and content. There was also a

significant relationship between prior knowledge and text

sophistication for all the groups and for the

miscellaneous subgroup.

The results also showed that there was a

relationship between prior knowledge and global coherence,

revision strategies, and involvement. The relationship

between prior knowledge and T—unit length for the Chinese

subgroup was significant, but weak for the rest of the

subgroups. Finally, the relationship between prior

knowledge and amount of subordination was not significant
for the whole group and all of the subgroups.

Concerning the relationship between overall

language proficiency and writing, the results showed that

there was a significant relationship between overall

language proficiency and writing proficiency scores for

 

 

 

110

the whole group as a heterogeneous group, text
sophistication for the whole group and for the Arabic
subgroup, T—unit length for all the group and for the
Japanese subgroup, subordination for the Japanese
subgroup, and revision strategies for the Arabic subgroup.
There was no significant relationship between overall
language proficiency and involvement.

With the exception of the T-unit length, prior
knowledge had more effect on all the writing measures than
overall language proficiency. It is also obvious from the
results that prior knowledge accounted for more variance
in writing than overall language proficiency did.

Finally, the results also show that there was a
difference among the participants' writing related to
their native language. How that difference was caused by

their native languages is beyond the scope of this study.

CHAPTER V
CONCLUSIONS

Introduction

This study, as discussed earlier, was designed to
investigate the relationship among prior knowledge,
overall language proficiency, and nonnative speakers'
written discourse. Furthermore, this study has addressed
several questions regarding the relationship among
demographic information, amount of reading, the subjects'
native language and writing measures.

To review, the research questions regarding the
nonnative speakers' written discourse were:

1. Is there any difference among the
participants' writing related to their
personal characteristics and the amount of
reading they do?

2. What is the relationship between prior
knowledge about specific topics and the

nonnative speakers' written discourse?

lll

 

 

112

3. What is the relationship between overall
language proficiency and the nonnative
speakers' written discourse?

4. What is the additive effect of prior
knowledge and overall language proficiency on
nonnative speakers' written discourse?

Forty-nine advanced students attending levels D

and E 111 the English Language Center at Michigan State
University in Spring, 1988, wrote on the topic of
Eclipses, a topic on which the students showed a large
spread of knowledge. Before writing the essay, the
participants took free-association and prior-knowledge
tests on the topic. After writing the composition, they
filled out a questionnaire, which reflected the degree of
their involvement with the topic. Trained raters scored
the participants' compositions on writing proficiency
scores, content (the number of ideas and the number of
words), text sophistication, global coherence, linguistic
complexity (T-unit length and subordination), and
revision strategies. The raters also scored the
knowledge test for the degree of fluency and
organization. Their background information, amount of
reading, language proficiency scores, and involvement
were self-reported on a questionnaire. Composition

ratings, and responses to the questionnaire were

 

 

 

113

correlated using the Pearson—Product Moment Correlation
coefficient. Regression analysis was used to quantify
the additive effect between prior knowledge and overall
language proficiency on writing measures. MANOVA, ANOVA,
and Tukey Test were used to compare the subgroups'
writing scores. Chapter IV presents the results in

detail.

Personal Characteristics and Writing

 

1. Is there any difference among the
participants' writing related to their
personal characteristics and the amount of
reading they do?

Correlation analysis was used to assess the
relationship among the participants' gender, age, level
of education, and the nonnative speakers' writing.
Analysis of variance was used to assess the variation in
the participants' writing associated with their native
language.

This study revealed that older students were
significantly xnore likely to write longer compositions
and to discuss more ideas than younger students. Average
older students were also more involved with their essays
than younger students. That is, older students tended to
like the topic more, found it more interesting, and

experienced less anxiety while writing. Writing

 

 

 

 

 

 

 

114

proficiency scores, text sophistication, global
coherence, linguistic complexity, and revision strategies
were not significantly related to the students' age.

Level of education might be expected to play a
significant role on the students' compositions;
surprisingly, none of the writing measures except
involvement was affected by the level of education. The
study revealed that the higher the students' level of
education, the more they were involved with their
compositions. Involvement was also affected by how much
students read. The amount of reading also could be
expected to play a significant role on the quality and
quantity of the students' compositions. Krashen (1984)
and Stotsky (1983) said that ability to write was not
taught, but rather was acquired through extensive
reading, and therefore, better writers are necessarily
better readers. This study suggested otherwise. Writing
proficiency scores, content length, global coherence,
linguistic complexity, and revision strategies, which
appeared to represent writing ability, were not
significantly related to the amount of reading the
students reported that they did. Therefore, Krashen's
notion of reading as the source of writing skill was not

confirmed in this study.

 

 

 

 

 

 

 

115

Participants' Native Language

 

This study disclosed that there were differences
among the means of writing measures that can be
associated with the subgroups‘ native language. For
instance, the Arabic and Spanish subgroups outperformed
the rest of the subgroups in the amount of subordination.
Why or how elements of these subgroups' native language
caused the difference in the subjects' writing

performance is beyond the scope of this study.

Prior Knowledge and Writing Performance

 

2. What is the relationship between prior
knowledge about specific topics and the
nonnative speakers' written discourse?

The relationship between prior knowledge and each

of the writing measures are more fully discussed below.

Writing Tests

This study expands the previous studies' findings
to the ESL students' writing. The findings of this study
revealed that great background knowledge of the topic
about which they were writing affected the nonnative
speakers' writing at the quantitative and qualitative
levels.

Specifically, the nonnative speakers with high
prior knowledge scored higher in aa writing proficiency

test similar to the writing holistic scores in the

 

 

 

 

 

116

previous studies of Langer and Chesky. On both scales,
the students' compositions were scored holistically in
terms of stating and supporting a position, passage
relevance, organization, details, sentence structure,
word choice, and mechanical errors. This finding was
supported by Hilgers (1982) who also found that students
familiar with content pay more attention to style,

arrangement, grammar, and mechanics.

Content Length
The students with high knowledge were able to

present and discuss many more ideas and to write longer
essays than the students with low knowledge. In this
study, there were some students with low knowledge who
wrote relatively long essays, but their ideas were not
relevant. They either discussed culturally related ideas
or presented legends related to the phenomenon of
Eclipses. Langer (1984), Chesky and Hiebert (1987), and
Stroethoff (1988) contended that knowing much about the
topic enabled the students to include and present more
ideas, details, experiences, and concepts. This finding
is also consistent with Perl's (1979) study which
revealed that students write more information when they
write about topics familiar to them, such as personal

topics.

 

 

 

 

117

It is obvious that if writers know nothing about
the topic, they will write very little. Some subjects in
this study expressed their unwillingness to write on the
topn: of eclipses because they said they did not know
anything about it. A few of them agreed, but wrote only
one sentence, "I know nothing about Eclipses," and handed
in their papers at the end of the session. This polite
refusal to participate appeared to reflect their agony at
being asked to write (x1 a topic about which they knew

nothing.

Text Sophistication

 

Text sophistication was measured by looking at
the existence of elements, such as terminology, jargon,
references, definitions, and explanations which
presumably more characterized texts addressed to more
sophisticated audiences. Concerning the relationship
between prior knowledge and text sophistication, this
study revealed different findings from those of
Stroethoff (1988). In this study, the relationship
between prior knowledge and text sophistication was
apparent. In Stroethoff's study, the relationship
between prior knowledge and text sophistication was not
consistent. in the three topics on which. she asked her
subjects to write. Two reasons may account for these

different findings. One is that Stroethoff asked her

 

 

 

 

118

subjects to write on three topics and the variation of
the subjects' knowledge about the three topics might be
responsible for the inconsistency of the effect of prior
knowledge on these two variables. Another reason is the
difference in age and analytical ability between her
subjects and the academically more sophisticated subjects
of this study Inight account for the subjects of this
study outperforming hers in these two variables. Langer
(1984) and Chesky and Hiebert (1987) did not investigate
their subjects' text sophistication.

In text sophistication, the high knowledge
subjects were able to use specialized terminology,
jargon, and references which were only accessible to a
highly educated and sophisticated audience. On the other
hand, the low knowledge subjects tend to provide simple
explanations of the notion of Eclipses. Interestingly,
some high knowledge subjects drew some pictures of the

sun and moon Eclipses and why and how they occur.

Global Coherence

This study revealed that prior knowledge was
significantly correlated with global coherence. The
texts of the students with high knowledge more clearly
identified the topic, orient the readers, employ cohesive
ties, plan and organize details, and followed the

discourse structural rules. In contrast, the data

 

 

 

 

 

119

suggested that students with low knowledge tended to
write fragmentary and haphazard texts. Those students
tended not to sense where each idea fit, nor did they
have a plan to organize or plan the details of their
compositions. This study's finding was consistent with a
similar study conducted by McCutchen (1986) who found
that students with high knowledge about football were
able to bring out and discuss important ideas of football
more coherently than the students with low knowledge.
This finding was also supported by Britton (1978) who
asserted that high prior knowledge writers know which
ideas are related and how all the parts are to be put
together. Langer (1984) and Chesky and Hiebert (1987)
found a positive significant relationship between prior

knowledge and global coherence.

Involvement

Regarding involvement, this study revealed that
high prior knowledge subjects were more involved in their
compositions while writing them. The findings of this
study suggest that it is important for writers to have a
positive attitude toward the topic about which they are
writing. This kind of attitude may create confidence,
motivation, and lessen anxiety which will allow the flow
of ideas and permit learning from writing to take place.

Chesky (1984) indicates that high prior knowledge

 

 

 

 

 

120

subjects were more involved with their writing. As a
result of their involvement, they wrote better and found

the writing easier.

Linguistic Complexity

This study and all previous studies of this type
did not reveal any relationship between prior knowledge
and linguistic complexity. One suitable explanation of
this finding was provided by Chesky (1984) who indicated
that the mean T-unit length constitutes a problem as a
method of measuring the quality of writing. The problem
comes from the fact that T—unit length does not reflect
the ability of students to carry out and construct
meaning.

Another problem with T—unit length as a measure
of linguistic complexity is that it is not clear whether
our subjects deliberately choose not to write long T-
units, or were not able to write long T—units. This
problem constitutes a major drawback of the mean of T-
unit length as a measure of linguistic complexity.
Furthermore, Witte and Faigley (1981) indicated the
inappropriateness of using T—unit length as an indicator
of writing quality. They contended that T-unit length
and clause length have very little influence on the
quality of writing. Although the T-unit measurement has

received criticism, it is still the widely used

 

 

 

121

measurement for the syntactic complexity of the native

and nonnative speaker's written discourse.

Revision Strategies

Revision strategies are an important stage in
writing which helps student writers to construct clearly
the intended message. The data of this study did not
show a significant relationship between prior knowledge
and revision strategies. One explanation might be that
students spend most or all of the 45 minutes writing the
compositions, not revising them. On the other hand,
those who did not write long compositions may have spent
their time thinking about new ideas, not revising them.
This finding is consistent with Raimes' (1987) study
which demonstrated that ESL student writers do not pay
attention to their mistakes because they are language
learners. Most of this study's subjects only pay
attention to the writing down of ideas, leaving the
mistakes to their teachers to correct.
Prior Knowledge and Writipg
Performance of the Subgroups

when students are grouped by native language, the
relationship among prior knowledge, writing proficiency
scores, content length, text sophistication, global
coherence, linguistic complexity, revision strategies,

and involvement become much more complex.

 

 

122

For instance, the relationship between prior
knowledge and writing proficiency scores was evident for
the Arabic and Japanese subgroups only. Prior knowledge

also consistently related to the length of the essays of

all the subgroups. In the miscellaneous subgroups, text
sophistication appeared to have a significant
relationship with prior knowledge. No other subgroup's

writing measures appeared to have a significant
relationship with prior knowledge.

This inconsistency of the relationship between
prior knowledge and the subgroups' writing measures may
be attributed to either or both of the following factors.
One factor might be that the number in each subgroup was
small. Another factor might have been an effect of the
subgroups' native languages. Investigating this factor
lies beyond the scope of this study.

Overall Language Proficiency and
Writing Performance

 

 

3. What is the relationship between overall
language proficiency and the nonnative
speakers' written discourse?

This study revealed that there was a significant
positive relationship between overall language
proficiency and writing proficiency scores, text
sophistication, and mean T-unit length for the whole

group. No significant relationship was found between

 

 

 

 

123

overall language proficiency and length of the essays,
global coherence, amount of subordination, revision
strategies, and involvement for the whole group.

For the homogeneous subgroups, the results of the
relationships between overall language proficiency and
writing measures were also very complex. For instance,
for the Arabic subgroup, there was a significant positive
relationship between overall language proficiency and
revision strategies and also between overall language
proficiency and text sophistication. For the .Japanese
subgroups, there was also a significant positive
relationship between overall language proficiency and
mean T-unit length and amount of subordination. No other
significant relationships were found between overall
language proficiency and any of the writing Ineasures.
Hence, it seems that placing students in writing classes
according to their language proficiency scores alone may

be a mistake.

Additive Effect
4. What is the additive effect of prior
knowledge and overall language proficiency on
nonnative speakers' written discourse?
The results of the study revealed that prior
knowledge accounted for more of the variability in

writing measures than overall language proficiency.

 

 

 

 

 

124

Specifically, prior knowledge accounted for more
variability in writing proficiency scores, number of
words, and number of ideas than overall language
proficiency did. The overall language proficiency, in
contrast, accounted for more variability on only mean T-

unit length than prior knowledge did.

Implication
The findings of this study demonstrated that

familiarity with the topic affected writing proficiency
scores, content length, text sophistication, coherence,
and involvement with the text. Based on the findings of
this study, writing instructors might consider the
following:

1. Choose interesting topics and ask students to
read and discuss materials related to those 'topics in
prewriting activities.

Instructors should ask their students to write
about topics that are interesting to them. Once the
instructors make sure that the topic interests them, they
then ask the students to read some materials related to
that topic in order to activate. either their existing
schemata about that topic or to provide them with
information that would enable them to write well.
Another alternative is dividing the classroom into groups

and asking them to discuss collaboratively the topic. In

A

 

 

 

 

 

125

this way students could exchange information about the
topic students are writing.

2. Prior knowledge should take its place with
language proficiency in the writing class

This study revealed that prior knowledge was more

strongly related to students' writing than language
proficiency was. Therefore, in writing classes, language
should not be overemphasized since it does not account
for the majority of the variability in the students'
writing performance.

3. Necessary jargon and vocabulary must be
provided to students

During the administration of the writing

assignment, students appeared to have a lot of questions
regarding the topic about which they were to write. Some
questions related to clarifying facts, others related to
the vocabulary needed to express specialized facts.
Therefore, if the topic requires explanation of abstract
facts or requires jargon, the vocabulary should be
provided.

4. Ideas relevant to the issues being discussed
in the topic must be highlighted in
prewriting activities

This study revealed that one problem of ESL

student writers lay in recalling ideas irrelevant to the

issue they were discussing. One way to solve this

:-

 

 

 

126

problem is by utilizing the prewriting session to
highlight the relevant ideas of the topic. Doing so will
help the students to learn which ideas fit what issue,
and will, in turn, help them generate well organized and
coherent texts.

Carrell (1983) said that ESL students do not
utilize prior knowledge properly when they read. This
finding seems to apply to some writers since some of
their essays contained interesting cultural ideas about
eclipses, but these ideas did not fit or support the
issues they were discussing. Consequently, it is
important for the students to learn how to separate
scientific facts from cultural legends. For example, if
ideas are unrelated to the issues being discussed, they
should not be included in the essay.

5. The students' native language could be

initially used to generate ideas

Lee (1986) and Swaffar (1988) indicated that
native language has an advantage over a second language
in the process of recalling information about the text to
be read. Therefore, students might be asked to jot down
ideas in their first language and then, with the
assistance of their instructors translate what they have
written into the second language.

6. Grades should not be emphasized as the

ultimate goal of writing

 

 

127

Based on observation during the conduct of this
study, we believe that grades should not be emphasized
because students usually fear grading. Therefore,
students should be told that they are not writing to
attain grades. The agony of writing for grades, as
Chesky (1984) pointed out, may keep the students from
freely expressing their thoughts and from trying
different styles. Of course, grades encourage and
motivate students, especially the good students, but
grades may also deter others.

7. The prewriting activities should not be

understated or overstated

Prewriting activities must be wisely planned and
organized to avoid wasting time and effort. Spending too
little or too much time preparing the students to write
may aggravate more than help the preparation process.

This matter is actually a judgmental one.

Suggestions for Further Research

 

For further research the following suggestions

could be considered:

1. Academic progress, study skills, and
intelligence should be studied along with
prior knowledge and language proficiency

We have learned while conducting this study that

there could be a third variable in addition to content

128

and language proficiency. Perhaps the good writers were
the ones who had good study skills and who happens to
know more about the topic. In addition, they might also
be the kind of students who would be disciplined enough
to be good writers. So, it may be that when this study
measures skills, it not only measured writing and
background knowledge, but it measured something else that
really governed both of them. Variables, such as
academic progress, study skills, academic sophistication,
and intelligence, may not only cause students to write
well, but may also allow them to be more knowledgeable.
So, it may be that it is not the students' prior
knowledge that affects their writing, but rather,
something else such as the above—mentioned variables.
Consequently, we suggest that any subsequent study take
into account the effect of those variables on the
students' writing performance.

2. students' native language should be used for

testing the students' prior knowledge

As indicated earlier, the students' native
language may provide an advantage over their second
language in expressing the knowledge they have about the
assigned topic. Therefore, we suggest that subsequent
studies ask students to express their knowledge in their
first language. In this way, a full picture of students'

knowledge about the topic will be obtained.

 

 

129

3. students' linguistic levels should be
measured in advance

In this study, we do not know whether the
linguistic complexity reflected in the students'
compositions was a result of the students' knowledge
about the topic or a production of their actual
linguistic level. To control this variable, the
students' level of linguistic complexity should be tested
beforehand. As a result of this testing, a group of
students whose proficiency is the same should be chosen
as a sample for any subsequent similar studies.

4. To measure revision strategies, a designated

time for revisions should be allowed

Students could be given an allotted time for
revisions. Most of the students expended the given time
to generate meaning, but not to clean up their
compositions. We speculate that the lack of relationship
between prior knowledge and revision strategies was due
to the lack of using time to revise. Whether this is
true or not would be proven by asking the same subjects
to write a second draft which focuses on clarifying the

meaning they have generated in the first draft.

 

 

130

Summary

This chapter has presented a summary of the
conclusions of this study, a discussion of the findings,
implications of the findings, and suggestions for further
studies.

Prior knowledge apparently seems to be related to
writing proficiency scores, content length, text
sophistication, global coherence, and involvement. T—
unit length, amount of subordination, and revision
strategies appear to be unrelated to prior knowledge.

Overall, language proficiency seems to have a

relationship with writing proficiency scores, text
sophistication, and linguistic complexity. Content
length, global coherence, revision strategies, and

involvement were not related to overall language
proficiency.

This study is consistent with similar studies in
finding that prior knowledge is crucial to writing.
Accordingly, we suggest that writing instructors ask
their students to write about topics interesting to them
and to de-emphasize the teaching of language in writing
classes. Further, we suggest that instructors provide
jargon and specialized terminology, highlight the issues
of the topic, encourage students not to discuss
peripheral ideas, and not overstate testing for preparing

their students for writing assignments.

 

 

131

For subsequent research, we suggest the
following: first, research should take into account the
students' academic progress, study skills, and academic
sophistication. Second, they should use the students'
native language in testing their prior knowledge. Third,
testing the students' linguistic complexity in advance.
And finally, since revision strategies cannot be judged
from the first draft, students should be asked to write a

second draft focusing on revision strategies.

 

 

 

APPENDICES

 

APPENDIX A

MISCELLANEOUS CORRESPONDENCE

:_~”, —.

 

 

April 6, 1988

Dr. James Stalker
English Language Center
Michigan State University
East Lansing, MI 48823

Dear Dr. Stalker:

I am a Ph.D. student in the English Department of
Michigan State University, and in the process of
conducting research for my dissertation on the effect of
nonnative speakers' prior knowledge on their expository
writing.

This letter is to request permission to use the
English Language Center as a setting for the research.
Students :u1 D- and E-levels will be asked to write an
essay during one of their regular class periods. They
will also be asked to respond to a questionnaire, take
tests, and demonstrate their prior knowledge of three
topics during two extra special sessions at the ELC.

Taking into account the students' time and effort,
an arrangement with their teachers will be sought to make
the students' participation consistent with the
teachers' endeavors to improve the students' writing.

As a result of the students' taking part in this
study, they will benefit by getting four extra test
scores (which they value), and doing the writing
assignment. The study will also contribute to the field
of teaching writing, to TESOL, and the program at ELC.

I would also like to request sixty copies of the ELC
grammar, vocabulary, and reading sample tests for the
purpose of measuring the students' language proficiency.

I appreciate your kind attention to this matter.

Sincerely yours,

Ali S. K. Al—Ghamdi

 

 

MICHIGAN STATE UNIVERSITY

 

ENGLISH LANGUAGE CENTER EAST LKNSING ' MICHIGAN ' 48824-1055
CENTER FOR INTERNATIONAL PROGRAMS

April 11, 1988

Mr. Ali 5. K. Al-Ghamdi

2309 E. Jolly Road

Lansing, MI 48910

Dear Mr. Al-Ghamdi:

Your request to conduct research in the D and E Writing classes at the English
Language Center is hereby granted. I think your research will be of use to our
instructional staff in our continuing curricular considerations, and your

research should improve the quality of instruction for future students.

You may begin your research as soon as we have a copy of your proposal and the
permission of the University Committee on Research on Human Subjects on file.

Sincerely yours,

10,...ch

James C. Stalker
Director

JCS:rhc

cc: Dr. Barry Gross, Department of English

.WSL' u an Affirmative Action/Equal Oppnrlumty IVISHIIIHOH

 

 

136

April 11, 1988

MEMORANDUM

TO: D and E Writing Teachers: Boyd, Lazarowicz,
Leonhardt, Milton, Novak, Wolf

FROM: James C. Stalker, Director

RE: Ali Al-Ghamdi's Research

Ali Al-Ghamdi, who is completing his doctorate in the
English Department, has my permission to conduct research
in the writing classes at the ELC. He is currently
awaiting official approval of the University Committee on
Research on Human Subjects. As soon as he receives that,
he will be able to move forward.

He will be coming to you to explain his research and tell
you what help he will need, but essentially he will want
to hand some material out in our class and have our
students do some writing for him during one of your class
periods. He will, of course, schedule his needs around
your requirements. If you have any questions, please see
me.

JCS:rhc

cc: Mr. Ali—Al-Ghamdi

 

 

137

MICHIGAN STATE UNIVERSITY

 

DEPARTMENT OF ENGLISH EAST LANSING O MICHIGAN ' 488244056
MORRILI. HALL

April 6, 1988

John Hudzik, Chair
UCRIHS
206 Berkey

Dear Dr. Hudzik:

Attached is the documentation requested and a full copy of
Ali S.K. Al-Ghamdi's research proposal which has been approved by
his Guidance Committee.

We request exemption from full committee review on the basis

of 1C, page 133 of UCRIHS guidelines. We appreciate your kind
attention to this matter as soon as possible.

Siesrely yours,
/,,-’> .

Paul E. Munsell

Associate Professor

PElefh
Enclosure

MIC-'1 an Affirmative Aztlan/Equal Opportunuy Inmluunn

”:94 c...

 

 

MICHIGAN STATE UNIVERSITY

 

UNIVERSITY COMMITTEE ON RESEARCH INVOLVING EAST LANSING ' MICHIGAN 0 48824-1111
HUMAN SUBJECTS (UCRIHS)

206 BERKEY HALL

1517) 553-9738

April 20, 1988

Ali S.K. Al—Ghamdi
Dept. of English
Morrill Hall

Dear Mr. Al—Ghamdi:

Subject: "THE EFFECT OF PRIOR KNOWLEDTGE ON NON-NATIVE
SPEAKERS EXPOSITORY WRITING #88-120"

The above project is exempt from full UCRIHS review. I have reviewed
this project and approval is granted for conduct of this project.

You are reminded that UCRIHS approval is valid for one calendar year.
If you plan to continue this project beyond one year, please make
provisions for obtaining appropriate UCRIHS approval prior to April
20, 1989.

Any changes in procedures involving human subjects must be reviewed by
the UCRIHS prior to initiation of the change. UCRIHS must also be
notified promptly of any problems (unexpected side effects,
complaints, etc.) involving human subjects during the course of the
work.

Thank you for bringing this project to our attention. If we can be of
any future help, please do not hesitate to let us know.

 

ohn K. Hudzik, Ph.D.
hair, UCRIHS

JKH/sar

cc: P. Munsell

"5L 1: .m :V/Irmum-e .dl'rtlm/Equul ()ppnrmmlv Insutumm

.1“

 

 

139

Dear Participant:

I am a Ph.D. student in the English Department at
Michigan State University. I am conducting research for
my dissertation on nonnative writing.

Your participation is highly important to the completion
of the study, and will also assist you in improving your

writing. You will write one essay and respond to a
questionnaire and to three short probes of your
knowledge. In addition, and as ea benefit to you, you

will also take three English Language Center (ELC) type
tests and get the results. You will also get an ELC—type
score for your essays.

Your responses will be completely confidential; please do
not write your name anywhere on your questionnaire,
essays, or tests. Participation in the study is
voluntary; without penalty, you may choose not to answer
certain questions, or not to participate at all.

Thank you very much for your time and effort in helping
me complete this study.

Sincerely,

Ali S. K. Al—Ghamdi
Ph.D. Candidate

 

 

 

APPENDIX B

BACKGROUND INFORMATION QUESTIONNAIRE

 

 

141
Rater No. l ( ) Rater No. 2 ( ) Rater No. 3 ( )
Student :
I - Demographic Items:
1 — Your Age:
( ) less than 20 years
( ) 20— 24 years
( ) 25— 29 years
( ) 30—34 years
( ) 35—39 years
( ) 40—44 years
( ) 45—49 years
( ) 50 years or above
2 — Please indicate your level of education you have
completed:
( ) High school
( ) College
( ) Master's degree
( ) Other (please specify )
3 - Gender:
( ) Male
( ) Female
4 — As well as I recall, my average in the last ELC
tests was
5 - My native language is
6 — How much reading do you do?

I do very much reading
I do much reading

I do little reading

I do no reading at all

AAAA
VVVV

APPENDIX C

WORD ASSOCIATION/PRIOR KNOWLEDGE TEST

 

 

143

II - PRIOR KNOWLEDGE:

TOPIC NO. 1, Photography:

 

 

The following words/phrases represent the central ideas
of the topic, photography

Please write what you know about these words/
phrases. The examples below illustrate how you should
respond to these words/phrases.. (Remember that you will
not be identified nor will you receive any grade for
doing this exercise.)

Example 1, Trees:
Prompt No. 1, Evergreen trees:
Associations for the phrase "evergreen trees":

-— The trees that do not lose their leaves in fall, but
remain green all winter.

-- The trees with slender and sharp-pointed leaves.

-- The trees with leaves that stay on the trees for
several years before dying

-- Trees that do not change colors.

-- Trees with short branches at the top, and long
branches at the bottom.

-- Come from cold weather places.

-- Protect themselves from freezing weather and
winter storms.

Examples 2, Weather:

Prompt No. 1, Summer season:

I will write the phrase "Summer season” on the board, and
we will practice together writing associations. Think
about what you know about "Summer season" and raise your
hand if you have something to say. This exerc1se 15 to
make sure that you understand how you should write

associations.

 

 

 

144

Now, start writing associations to the following prompts
of the topic Photography.

Prompt No. 1, Camera:

Prompt No. 3, Negatives:

Prompt NO. 4, Focusing:

145

146

Prompt No. 5, Flash:

147

Topic No. 2, Underwater Diving:

 

The following words/phrases represent the central ideas
of the topic, underwater diving:

Prompt NO. 1, Skin diving:

Prompt No. 2, Scuba Diving"

148

Prompt No. 3, Underwater Swimming:

Prompt No. 4, Diving Equipment:

149

Prompt No. 5, Oceanography:

150

Topic No. 3, Eclipses:

 

The following words/phrases represent the central idea of
the topic, Eclipses.

Prompt No. 2, Shadow of Eclipses:

Prompt No. 2, Solar Eclipses:

Prompt No. 3, Lunar Eclipses:

Prompt No. 4, Astronomers:

151

152

Prompt No. 5, Eclipses Prediction:

APPENDIX D

SCORING THE PROMPTS

 

154

Rater #:
Student:

 

 

III - Scoring the prompts:
A - Fluency:

( ) high fluency
( ) moderate fluency
( ) low fluency

B — Organization:

Write scores for all the prompts:
( ) ( ) ( ) ( ) ( )

Total =
Average =

( ) High Organized
( ) Partially organized
( ) Diffusely organized

C — Combination:

Average of the scores of high organized and
partially organized

( ) Highly combined
( ) Moderately combined
( ) Low

Topic No.

1

155

Rater #: . Topic No.
Student: .

 

 

III - Scoring the prompts:

A — Fluency:

( ) High fluency
( ) Moderate fluency
( ) Low fluency

B - Organization:

Write scores for all the prompts:
( ) ( ) ( ) ( ) ( )

Total
Average

( ) Highly organized
( ) Partially organized
( ) Diffusely organized

C - Average of the scores of high organized and
partially organized

( ) Highly combined
( ) Moderately combined
( ) Low

 

156

Rater #:

 

. Topic No.
Student: .

 

III - Scoring the prompts:

A — Fluency:

( ) High fluency
( ) Moderate fluency
( ) Low fluency

B - Organization:

Write scores for all the prompts:
( > ( ) ( ) ( ) ( )

Total =
Average =

( ) Highly organized
( ) Partially organized
( ) Diffusely organized

C — Average of the scores of high organized and
partially organized

( ) Highly combined
( ) Moderately combined
( ) Low

3

APPENDIX E

THE WRITING TASK AND

THE INVOLVEMENT QUESTIONNAIRE

158

Rater No. 1 ( ) Rater No. 2 ( ) Rater No. 3 ( )
IV - Writing the essay:

You have forty-five minutes to write about the
topic: Eclipses

You should assume that you will be writing this
essay for well-educated readers. You can write as
much as you can. (Please use pens for writing.)

 

 

159

Involvement:

Please put a (

) beside the appropriate answer as

it applies to you:

Liking the topic:

(
(
(

( )

Level

Vvv

AAA“

Level

I like this topic very much

I like this topic

I don't have a firm opinion about liking or
disliking it

I don't like it.

of difficulty as writing about this topic:
It
It
It
It

is very easy to write about
is easy to write about this
is hard to write about this
is very hard to write about

this topic
topic
topic
this topic

of interest and excitement while writing on

this topic:

was highly interested
was interested

was somewhat interested
was not interested

HHHH

of anxiety:

not nervous while writing on this topic
somewhat nervous while writing on this

I was
I was
topic
I was
I was
topic

nervous while writing on this topic
very nervous while writing on this

of confidence:

I felt very confident while writing on this
topic . .

I felt confident while writing on this topic
I felt somewhat confident while writing on
this topic . .

I didn't feel very confident while writing on
this topic

6

160

— If there is a chance I will write about this topic
again:

strongly agree
agree

somewhat agree
don't agree

HHHH

(
(
(
(
— I have enjoyed writing about this topic:

strongly agree
agree
somewhat agree

(
E
( don't agree

HHHH

- Who did you have in mind as your reader when you

were writing?

( ) An expert reader in this topic

( ) An average reader with only some knowledge
( ) A reader with little knowledge

( ) A reader with no knowledge

APPENDIX F

SCORING THE PARTICIPANTS' WRITTEN DISCOURSE

TEXT SOPHISTICATION QUESTIONNAIRE

 

Rater #

Student

VI — Level of English proficiency:
ELC Score

VII — Linguistic Complexity:

A — T—Unit length
Mean of number of words per t—unit:
Number of words in the essay =
Number of T-units in the essay =

B — Subordination
Ratio of main clauses to all clauses
Number of all the clauses:
Number of main clauses:

VIII - Writing global language proficiency score:
Score:

IX — Global coherence:

Miscellaneous
Incomprehensible
Incoherent
Partially coherent
Fully coherent

XI

XII

163

- Revision Strategies:

— Internal revision
( ) Number of marks in subject
( ) Number of marks in information and arguments
( ) Number of marks in structure

- External revision:
( ) Number of marks in tone and style
( ) Number of marks in language
( ) Number of marks in mechanics
- ( ) Total revision
- Content:
- Number of words
— Fluency (central ideas)
( ) High fluency
( ) Moderate fluency
( ) Low fluency

— Text sophistication:

- The complexity of terminology and/or jargon of
essay:

Highly complex
Complex

Less complex
Not complex

Vvvv

The level of sophistication of references:

( ) Very sophisticated

( ) Mature but simple

( ) Very general

( ) Child like (very simple)

Definitions and Explanation:

Not found at all
Very few

Few

Very many

AAA“
VVVV

 

4

8

164

- Repetitions:

( ) No repetitions

( ) Very few repetitions
( ) A few repetitions

( ) A lot of repetitions

- Is there any mentioning of the audience to which
they are addressing their writing?

Examples:

- Is there any indications that show their awareness
of the audience?

 

Examples:
- False start Yes . No
- Stray thoughts Yes . No

 

Examples

 

APPENDIX G

PROCEDURES FOR SELECTING PROMPTS
PROCEDURES OF SCORING KNOWLEDGE ORGANIZATION
ENGLISH LANGUAGE CENTER WRITING SCORING SCALE

BAMBERG COHERENCE SCALE

 

166

PROCEDURES FOR SELECTING PROMPTS
(From Newell & McAdam, 1987)

Care should be taken not to select words/phrases that
are semantically close, such as antonyms (hot—cold)
and synonyms (large—enormous).

When words/phrases in the content structure are
accompanied by modifiers (for example, large, white
automobile) that do not change the meaning of the
concept, the modifiers are excluded. However, if the
modifier(s) does affect the meaning of the word/
phrase (for example, welfare program), the modifier
is included.

Proper nouns are not selected unless they are
critical to understanding the passage. For example,
in a passage about Stonehenge, Stonehenge would be
selected but not England.

When the passage includes both a technical term

(deficit spending) and its synonym (spending money
one does not have), the technical term should be

selected.

 

167

PROCEDURES OF SCORING KNOWLEDGE ORGANIZATION
(From Newell & McAdam, 1987)

Each association is rated separately and
independently at only one level of knowledge
organization.

Responses are rated according to how the concepts are
used 1J1 the selected passages (definitions for each
tenn are available to raters). For example, if the
concept "elements" was selected from a passage on the
chemistry of compounds and mixtures, responses
related to "weather" would have to be rated as
diffusely organized knowledge.

Incomplete definitions and imprecise word use may be
"defining characteristics" (partially organized)
rather than definitions (well organized). For
example, "elements" defined as "things that make up
other things" indicates 3partially organized
knowledge.

Attention to level of specificy is important in
rating a response. A response may appear to be an
example, attribute, or defining characteristic of a
concept but too broad or general and therefore
peripherally linked only to the concept (for example,
the response "money" for the prompt "financial
rewards").

When the prompt or its root is used as part of the
response with little or no elaboration (for example,
"Neutrality: to remain neutral"), such a response is
rated "diffusely organized" unless there is further
elaboration in the response indicating a higher level

of organization.

 

 

 

168

When the prompt consists of two or more words (for

example, "Internal Affairs: matters concerned with
the inner workings of government such as collection
of taxes"), the response should be rated as

representative of a high level of organization only
if it indicates understanding of the prompt as a
concept; responses that define only one of the words,
such as "affairs inside a country," would be rated as
partially organized.

In cases where the students have not indicated any
separation between their string of responses, raters
should use the following guidelines for determining

what qualifies as a single response: (a) phrases
"join the army," "fight for the country," (b) clauses
"because of the need for food," (c) sentences that

are clearly marked by a period, (d) single words in a
string of responses.

 

100

95

9O

85

8O

75

169

ENGLISH LANGUAGE CENTER SCORING SCALE

Very clear. Native grammatical and vocabulary
usage. Few if any misspelled words or punctuation
problems. About 2 pages or more.

Between 90 and 100.

Very clear. Few grammatical errors of any kind.
Generally smooth. Good command of vocabulary with
not more than several inaccurate or unnatural
usages. May have a few misspellings or punctuation

problems. Well organized. Clearly developed
paragraphs.

Between 80 and 90. Apparent that writer is not a
native speaker. Good organization and paragraph
development. Typical of a very advanced ELC
student.

Very clear and understandable with minor obscure
parts. Generally accurate use of grammatical
structures at all levels with some problems with
advanced structures. Mostly smooth sentences with
a few choppy or run-on constructions. Largely
accurate use of vocabulary with some problems with
advanced words and some unnatural usage. A few
misspelled or distorted words, and a few
punctuation problems. Adequate attention to
organization and development.

Clear with possibly one or two parts that require
some mental editing. Generally accurate on
elementary and intermediate grammatical
structures, with occasional serious errors on
advanced structures. Generally smooth, with a few
choppy or poorly connected or run—on clauses or
sentences. Some use of advanced vocabulary with
several incorrectly or unnaturally used words.
Several misspelled or distorted words and several
punctuation problems. Some attention to overall
organization and development. About 1—1/2 pages
long.

 

7O

65

6O

55

45

170

Clear with minor editing but may have some parts
requiring major editing or guessing. Frequent
minor grammatical errors, but also accurate or
nearly accurate use of a wide range of structures
and tenses. Often smooth but with several poorly
connected or choppy or run—on clauses or sentences.
Often smooth but with several poorly connected or
choppy or run-on clauses or sentences. Vocabulary
largely elementary, but often advanced words are
used, although. not always accurately. Possibly
many misspelled words and up to several
distortions. Sentences usually clearly marked but
with several punctuation problems.

Probably some unclear parts. Use of a wide range
of grammatical structures but with numerous errors,
minor and serious, especially on the more advanced
structures. Several choppy, run-on, or poorly
connected clauses or sentences. Vocabulary range
in the elementary and intermediate ranges but with
frequent errors at all levels, especially with

advanced words . Quite a few unnatural usages .
Probably many misspelled words or distortions.
Punctuation may show frequent problems. Usually I

to 1—1/2 pages.

Probably becomes unclear or confusing when
expressing some complex ideas. Sometimes accurate
at the elementary level but with serious
inaccuracies at all levels. Many choppy, run-on or

poorly connected clauses or sentences. Many
serious errors at all levels, in vocabulary, though
fair use of elementary vocabulary. Frequent

unnatural or apparently translated or invented
words. Probably many misspellings, distortions, or
mispunctuations. About a page or more.

Probably be unclear in several parts or may
possibly attempt to express only very elementary

ideas. Few accurate sentences or natural and
adequate use of vocabulary. Rarely smooth. Many
apparent translations or invented words. Many

misspellings, distortions, and nuspunctuations.
About one page.

May be largely unclear or overly simple. Few
accurate or complex sentences. Inaccurate use of
vocabulary, even at elementary level. Usually less
than one page.

 

 

 

171

35 Name plus attempts at one or two sentences. Barely
recognizable as English.

30 Name only; no attempt to write.

172

HOLISTIC COHERENCE SCALE

4 = Fully Coherent
Writer clearly identifies the topic
Writer does not shift topics or digress

Writer orients the reader by creating a context or
situation

Writer organizes details according to a discernible plan
that is sustained throughout the essay

Writer skillfully 'uses cohesive ties such. as lexical
cohesion, conjunction, reference, etc. to link
sentences and/or paragraphs together

Writer often concludes with a statement that gives the
reader a definite sense of closure

Discourse flows smoothly-~few or no grammatical and/or
mechanical errors interrupt the reading process

3 = Partially Coherent

If writer does not explicitly identify the topic, s/he
provides enough details so that readers can
probably identify the specific subject

Writer has one main topic, but there may be minor
digressions

Writer provides some reader orientation, either by
briefly suggesting the context or by directly
announcing the topic

Writer organizes details according to a plan, but may not
sustain it throughout or may list details in part
of the essay

Writer uses some cohesive ties such as lexical cohesion,
conjunction, reference, etc., to link sentences
and/or paragraphs together

173

Writer does not usually conclude with a statement that
creates a sense of closure

Discourse generally flows smoothly although. occasional
grammatical and/or mechanical errors may interrupt
the reading process

2 = Incoherent

Some of the following prevent the reader from integrating
the text into a coherent whole:

Writer does not identify the topic and the reader would
be unlikely to infer or guess the topic from the
details provided

Writer shifts topics or «digresses frequently from 'the
topic

Writer assumes the reader shares his/her context and
provides little or no orientation

Writer has no organizational plan in most of the text and
frequently relies on listing

Writer uses few cohesive ties such as lexical cohesion,
conjunction, reference, etc. to link sentences
and/or paragraphs together

Writer creates no sense of closure

Discourse flow is irregular or rough because mechanical
and/or grammatical errors frequently interrupt the
reading process

1 = Incomprehensible

Many of 'the following' prevent the reader from Inaking
sense of the text:

Topic cannot be identified

Writer move from topic to topic by association or
digresses frequently

Writer assumes the reader shares his/her context and
provides no orientation

174

Writer has no organizational plan and either lists or
follows an associative order

Writer uses very few cohesive ties such as lexical
cohesion, conjunction, reference, etc. and
sentences do not seem connected or linked together

Discourse flow is very rough or irregular because writer
omits structure words, inflectional endings and/or
makes numerous grammatical and :mechanical errors
that continuously interrupt the reading process

0 = Unscorable/Miscellaneous

Source: Bamberg (1984).

 

APPENDIX H

THE STUDY'S RAW DATA

 

 

13:?

20—24
25-29
30—34
35—40
40 or

UlnthH
II II II II II

L.E. (Level of

IbLAJNF‘

G.R. (Gender)
Male

176

years
years
years
years
over

Education)

M.A. degree

B.A. degree

High school degree
Others

Female

ELC Score (Language Proficiency Scores (35—100))

R (Amount of Reading)

1
2
AFL (Fluency)

1 LOW
2
3

High

Little reading
Much reading

Moderate

AOG Organization (Scale 5—15)

ACM Combination (Scale 0-15)

LCP (T—Unit Length)

Sub (Subordination)

WLP (WritingiGlobal Language Proficiency)

WRD (Number of Words)

IDS (Number of Ideas Scale: 1—3)

 

 

 

 

 

 

 

 

177

TNT 4(Total of T-Units)

 

 

 

 

 

 

 

GC (Global Coherence Scale: 1—4)
COH (Cohesion Scale: 1-4)

TS (Text Sophistication)

AA (Audience Awareness Scale: 1—4)
INV (Involvement)

ER (External Revision)

IR (Internal Revision)

TR (Total Revision)

 

 

 

 

 

 

178

Table H—l

eonm 11H.1163m mncm<.m not cons

 

 

Idln

 

 

 

 

 

 

zmnH<m rmoocmom 50m HEN. on non w >2. 500 so: How moo zoo 2.5 How 5.25. on no: em >5 :2 m» Ho .3
N N H NN N N N.NN N.NN N.NN N.00 mm Hmb H H0 N.NN N.NN

.5365 N N H No N N Ho. NN N.NN H0.NN N .3 NN NNN N NN N.oo N .00 W w H HM w HM
N N H N0 N N HN.NN HN.NN N.00 H.m0 NN NN5 N NN N.NN N.00 H5 N Hm HN H Hm
N N H Nm N N HH.NN H0.NN H0.00 N.05 N5 NNO N NN 5.00 5.00 H0 N Hm HN 0 H
N N H N0 H N HN.00 HN.NN HN.NN H.00 NN N00 N Na N.NN N.00 Hm N H0 H5 0 NW
N N H NH H N H5.00 H5.00 H0.NN H.mN N5 N00 N Nm N.00 N.00 H5 5 H5 H0 0 N0

orHsmmm N N N No H N N.NN N.NN H0.0H H.m0 00 N50 N NN 5.00 5.00 Hm N HN H5 N HN
N N N 2. N N 8.8 N.NN HH.NN HNN NN 58 N No 5.00 N.NN HN N Hm No H NH
m N H NN N N HN.00 HN.NN HN.05 H.0H N0 NNo N No 5.00 5.00 HN N HN N H N
5 N H mm N N HH.NN H0.NN HH.5H H.5m NN Nmm N NN 5.00 5.00 HN N NN 0 N HH
N N N NN H N Poo N.NN HH.No H.No No NNN N No N.NN N.NN HH N Ho H o H
5 N H No N N 0.00 0.00 H0.00 H.Nm Nm Nmo N Nm N.NN N.NN HN N Hm H 0 H
N N N 00 N N H0.00 N.NN HN.00 N.NN N5 NNH N NN N.00 N.00 HH H N H0 H N0
5 N N 0N H H N.00 0.00 N.NN H.Nm Nm Nmm N No 5.00 N.NN HH N H0 HN H HN
5 N H N0 H H N.NN N.NN N.NN H.0m NN HNm N N0 N.NN N.NN Hm N N m N m

omomommo N N H N0 H H HH.NN H0.NN H0.NN H.5m Nm NNN N NN N.NN N.NN HH N HN HN o HN
N N H NN H N 0.00 0.00 H0.NN H.N5 N5 50N N NN N.00 N.00 N H HH N5 0 N5
N H N 0N N N H0.NN H0.NN HH.H0 N.0N NN NNN N N0 N.00 5.00 H0 H N m m H0
N N H NN H N HH.NN HH.00 H0.N0 H.5H 00 NmN N N5 5.00 5.00 HN N Ho m N m
5 N N 00 N N HN.NN HH.NN HN.HN H.mN 00 Now N N0 5.00 5.00 HN N H0 Nm 0 Nm
N N N N0 N H N.NN HN.NN N.NN H.Nm mm HoN H HN N.00 H.mN HH N HN m 0 m
N H H NN H N N.NN N.NN 0.05 H.5N NN NNm N Nm 5.00 N.NN HN N m HN 0 HN
m N H 00 N N H0.NN N.NN HH.00 H.mN mH Nmm N N5 5.00 5.00 Hm N Hm NN 0 NH
H H N NN H H N.NN 0.00 0.00 H.Nm NN 00 H H0 N.00 N.00 HH N H0 N 0 N
N H H NN H N N.NN N.NN HN.00 H.5N NN NH5 N No 5.00 5.00 HN N H0 H0 N H0
5 N H No H H H0.NN N.NN H0.5m H.Nm NN NmN N Nm N.00 N.00 HN N 0 N0 m N5
N N N Na H H N.NN N.NN H5.mm H.N0 mm H00 H HH N.00 H.0N H0 N 0 5m H 50
5 N N 0N N N HN.NN HN.00 H0.00 H.N0 00 50m N Na N.NN N.NN HN 5 H0 Hm N Hm
N N H mm H N HH.NN H0.NN HH.NN H.N0 Nm NNN N Nm N.00 N.00 H5 5 HH NH N N5
N N N 00 H N H0.NN H0.00 HH.NN H.m5 N0 N50 N N0 N.NN 5.00 H0 H N H0 0 H0
N 5 H mN N N HN.NN HN.NN N.NN H.NN N5 NmH N No N.00 N.00 HH N Ho HH 0 HH
N H N 0N N N 0.00 N.00 HN.NN N.00 NN Nmm N NH N.00 N.NN HH N HH 0 N HH
N H H N0 H N HH.00 H0.NN N.NN H.mN Nm H50 N HN N.NN N.00 HN H m H H N
N H N NH H N H0.00 N.NN 0.H0 H.NN NH NH H H0 N.00 N.00 0 H N N 5 N
N H H N0 N N HH.NN H0.NN N.NN H.Nm N0 N05 N NN 5.00 5.00 H5 N H5 Hm N HN
N H N N0 H H 0.00 0.00 HH.NN H.mm Nm NN5 N H0 N.00 N.NN Hm N H0 H5 N Hm
N N H N0 H H N.NN N.00 N.Nm H.5m Nm NOH N N0 5.00 N.NN H0 H m NN m NN
N N H 00 H N N.NN N.NN N.NN H.5H Na NHH N 50 5.00 5.00 H0 H N 5m m m0
H H N N5 H N 0.00 N.00 N.NN H.wm 0N 505 H NN N.NN N.00 HH N H0 0 0 0

momon: m N N mm H N HN.00 HN.NN HH.H5 H.NN NN NHN N NN N.00 N.00 HN N Hm m 0 m
H H 5 N0 H N H5.00 H5.00 Hm.5m N.NN N5 HNo N HH N.00 N.NN HN H H0 HN 0 HN
H H H 00 H N HH.00 0.00 HN.No H.mN 00 NN5 N N5 N.NN N.mN HN N HN HN H HN

HancommHo: H H N 00 H N HN.00 N.00 H0.0N H.NN NN 5NN N 5N 5.00 5.00 H5 N HN NN 5 No
m N N 00 N N HN.00 HN.00 HN.50 H.N5 N0 N50 N Nm 5.00 N.NN HN H Hm 0 0 0
w N H N0 N H H0.00 N.NN 0.00 H.mm NN HNm N Hm N.00 N.NN HN N HH N m 0
a N H NN N N HH.NN HH.NN H0.NN H 0H N5 NHH N No N.00 N.00 HN N N0 H5 N Ho
5 N N. mNI, M 111:-rIINHQQIIIJHHHHIIIHHWHHIiz:NWQQI--Nflsll as H N chc1 u.cc HMI H1. N m H m
N N N NN N H N.NN 5.NN H0.NN H.Na N5 55m H 5N N.00 N.NN H0 H H5 N N m
N N H NH H N H0.NN N.NN N.NN H.N0 Nm :3 N H0 N.NN N.NN HN H HH H o H

 

 

 

APPENDIX I

PRIOR KNOWLEDGE TEST AND WRITING SAMPLES

 

 

 

 

180

STUDENT 1

Personal Characteristics

 

Age: 40-44

Level of Education: M.A.
Gender: Male

Amount of Reading: Much reading

Prior Knowledge Test

Fluency: High
Organization: High
Combined: High

Writing Performance

Overall Language Proficiency Score: 73
Holistic Writing Score: 71

Number of Words: 311

Number of Ideas: 3

Text Sophistication: Moderate

Global Cohererce: Partially coherent
Involvement: Highly involved

Total Number of T—units: 30

Mean T—unit Length: 10.37

Amount of Subordination: Partial
Total Number of Revisions: 16

students' Associations to the Prompts
Topic No. 3: Eclipses:

The following words/phrases represent the central ideas
of the topic, Eclipses.

Prompt No. 1, Shadow of Eclipses:

--Comes to the earth

—-Will dark in the earth where the shadow comes
--Some time make someone's afraid

—-The plant will not process its activities
-—The place will get dark a moment

-—The place where the shadow comes like at night

 

 

181

Prompt No. 2, Solar Eclipses:

--The sun doesn't look clearly

--There is a total eclipses and part eclipses

—-There is a story about this eclipses among the
people

--The story about the giant eat the sun

-—The earth will get dark a moment

'--Happen in the day

--There is an effect to the tea

--Many people to watch it by glasses

Prompt No. 3, Lunar Eclipses:

--The people will not see the moon clearly
--Happen in the night

--The people get unhappy because they are afraid
—-The moon looks a part or total

--Happens in the full moon

--There is an effect to the sea

Prompt No. 4, Astronomers:

--The persons who expect in the astronomy

—-They make account to the star

--They work with instruments

--They are graduate from the engineering faculty

Prompt No. 5, Eclipses Prediction:

--Made by astronomers

——The dark can be predicted

--Sometime a year happen twice or once
--Be counted by detail numerical point
-—Can't be found by commoner

 

 

182

2.4554 awe...“ é/eumz? ‘M- @MWAN/ M7
4% M/oawém WWW“ 1;; ”Mae.
fl» (M 444/? £954, 74%«2/‘4/5‘12 «xv-4H; cod/ea WK
49mm Mac/Ha ,Waz Hoes/Ha W @114; 517/47»; A244.
0.4» «565“ W M/f% ”Hz/HQ "
Jaw ALA ’4 Hag/52,637. wad 75465., %%mA/N
(AZ. Has/NH Nam/H ma. /% 45mm. J‘Léz/érmfl
N.NN/{NM ”'46; M,M.;r MM... M/Aé72/flr 45....
7%xwrm),éfiv%e m £9,262.34 [Hex/64; «ya/m» 74/4
4mm. W ,4. $455.. aߢ4aam WWW f
442;”... H£Hw M fl¢M, 72;; dokxaz/éé
é/fiw m%cz/W%m a 1744/. éM/éac. %.W
Mflwém /WW 4.; 44-.-MHH7, WM- W _
4...“, 4i flLHf/WWH W2 £744: 7W ”(é/14; oa/m
Zd%.ao’e‘ W W ”MW/7&W 4'72 71/7,: mm
.  M35. .242... ywwamwfl HN.NN/4...... H456,
4.;me HA7, W44 MPH/A7 aél/im Hag/He c2». 4;. -
flNnaeg/zxa ng ”H, 4726 M,/;,452 flM,/&a¢¢a€

 

 

 

183

, é/N- flZWé W ifxfo/ 4% v‘zm M .72.:- m

4. A? A 4274 W, rnu{ gflésy / 1,4/ 4% flamevn744
£5,752 (Ha/55¢ ”dra—m4: 4;. 74227.41 H NH. we; fi/Heé
Mm. ‘ 4...». W745 Q.WN//€MMN (“é/gm 4.; Ha-
flhfigh .55 {449,41 /’m/( .444.”sz :71 ”Cum -w—ELLH 47%;944144
4M .XN-éflen. /’//z:( 4% %.(' 4‘44 .- mm . Z22 Jwy/Q/V/fomz
44%“ /%’é:wa.m toaé.'.e WWI/n70; 4%,»? 7Vx49z<¢Lz
M/eax 2: 7;...2/ 72/46 5w 1.2%; «nu/HoHH/amy’as
f/fe My; M/H‘N , 542104144? ./é& £44,,“ zap/7&4 Ham” {
HWLWae 25 75/52 flmrn meél.
AXH//% [544,44 mxc 65444 14675.5( exp-N cod-W4 //trv7é.¢yz

A47 //7/;——;.v KémH/v , éw/cif f’ém/ 566% 4m/( fivfléaée‘

 

 

 

184

STUDENT NO. 2

Personal Characteristics

 

Age: 20—24

Level of Education: Graduate
Gender: Female

Amount of Reading: Little

Prior Knowledge Test
Fluency: High

Organization: Partial
Combined: Partial

Writing Performance

 

Overall Language Proficiency Score: 88
Holistic Writing Score: 79

Number of Words: 340

Number of Ideas: 3

Text Sophistication: Moderate

Global Cohererce: Partially coherent
Involvement: Low

Total Number of T—units: 29

Mean T-unit Length: 11.72

Amount of Subordination: Moderate
Total Number of Revisions: 10

students' Associations to the Prompts

Topic No. 3: Eclipses:

The following words/phrases represent the central ideas
of the topic, Eclipses.

Prompt No. 1, Shadow of Eclipses:

--Shadow sometimes hide the sun or the moon because
the earth is moving all the time

185

Prompt No. 2, Solar Eclipses:

--You can't see the sun, because the moon stays
between the sun and the earth and it (the moon)
(hides the light from the sun

Prompt No. 3, Lunar Eclipses:

--You can't see the moon, because the earth stays
between the sun and the moon and it (the earth)
prevents the light that shines the moon.

Prompt No. 4, Astronomers:

--The study of the earth, and other planets or
stars

Prompt No. 5, Eclipses Prediction:

--The long time ago, when the reason of solar
eclipses and the lunar eclipses are not known,
people thought it will be the end of the earth,
because the sun (or the moon) has disappeared
suddenly.

186

WU?! WY! TWO INKS cl: €C-11F$QS)SO(CIC‘ €C(.‘FSC and (war
QdCPse. The (chm. b coma-med uhtk twee Flam-ab) the 5mm)
‘tf‘n: C’N‘Hn Cmd “the moon.

Th6 $c(0kr eclibe Occures when “We: MOON acme)
5<W<ih ‘W'i Sun and ‘Hdi Cox-HA. The, moon, dib‘IHPM T“:
scant {mm The sum "to TM: (aviary) 30 m some, far-{- 61,; The,
Gaff”) PQOFW (CV1 ”‘9‘— See‘ “4(- SV‘VW ‘FOT’ C'\ (.m‘ch‘Nm amouvfi‘
g-C ‘hmz.

Oh The other-1146\nd., Tl“ (war crime. Occur-{5
'JC'MA Wt COKF‘HV COWS bfi‘i‘w‘iem (‘4: Sum (And The moon.
LCC can see The moon, when ‘Hm’. Sum“ ts 914.».ch rt) >51-
wiw ”on: su‘Jmcrtnn of [unar desc ) ‘Hm: eom—tw chm-urn
‘Hfl'c [:SKT ‘From The Sun ‘1'th i3 §H§nmj The, moan.

\w scmz (Doer 0+ The worm.) We can not 3ee the meow;
becowxse, ct this.

Wage, 'WO' kmds a”? Calu‘PSeS O'CCLLNS) beccwkSe. Tke.
K’QFTV‘ wad +149. (moon, rg Mex/(mg around the 5mm.

U 1'37. t"3\F{"€r\) 'OhLi \'V\ '1 Let—fawn ‘Hmr, , de. H" "'9 hOT‘

0x S+W~AR~L I‘Mbj To Malayan. Howumr) HA the, (Cw-1.3 dog;

 

 

187

0;. 0‘“ “771°” ; “MC“ The. W‘mgg M the plOmet- one, nei-
Nfll—fixowmmcd and mo‘l~ decu- —G.— most of: W
poem at That ‘trmc , flung Théwgtrfl‘ H- is A' k—Md 6?- ;
cm 6de 0‘“ Wm fleddowhem {Mi QCUPSiS OCCMNd.
EX(‘.O.M$€, The $Lm bide among 03? ‘tWL sokar QNMPSE) peopke/
chfi‘th TM») NM never Set The Sun again/J wag/‘g‘gmwd
of CC'vxseiwugnuzé {'14ch UJTH (Lamar When W6 5010“" 3"?“le
fixcR 04mg. firmmu‘g) These fincngs 0N ixammed .bg
suavx’n‘s’vs W4?:¢Mmfid 3.0312,) awwb bacamc dear .

HOLWS ‘ (VIN QMIKCK Rouwx osbcud' 90(04- (($939,
and \var «1va vol/Rn Thea am in, flag, (Lennard-m3}
gghccfl ) so no one‘ia‘fmx‘d a? echyses ow») more. Become,

Et OCUM‘QS n01~ Mtg cfi-m/ )APQOFN -&— (001cm

”(Mex 4““? )
Per—Nara To seems ‘tws whevx (TA out/um}.
wt“ ‘
...}:—

\"V is 955‘: A ow: consewu. (If ox no~+wm\
bU~T if 1‘7 C0“$id€l‘<d WM dt‘tfreriflfj m We (ms-”(— Md .‘v.
I‘v; PV‘QSNM‘t ,B \ Think {-MQ yum ’ ‘H’h‘: (MCCn and ‘H’Ue.
earth m4 creosred om .mkmst‘nR "Ric? M W“ “SWIS-
\J

< 30 mex‘Rs Nah-v3 >

\)

 

188

STUDENT NO. 3

Personal Characteristics

Age: 20—24

Level of Education: Graduate
Gender: Male

Amount of Reading: Little

Prior Knowledge Test

 

Fluency: Partial
Organization: Partial
Combined: Partial

Writing Performance

 

Overall Language Proficiency Score: 71
Holistic Writing Score: 71
Number of Words: 159

Number of Ideas: 2

Text Sophistication: Moderate
Global Cohererce: Coherent
Involvement: Partial

Total Number of T-units: 19
Mean T-unit Length: 8.37
Amount of Subordination: Low
Total Number of Revisions: 1

students' Associations to the Prompts

Topic No. 3: Eclipses:

The following words/phrases represent the central ideas
of the topic, Eclipses.

Prompt No. 1, Shadow of Eclipses:

——Two kinds: Solar eclipses 8 lunar eclipses
—-In that situation we can't see sun or moon.

189

Prompt No. 2, Solar Eclipses:

—-We can't have sunshine
--Makes dark

--We can see the sun with the help of specific
instruments.

--It occures once in a year.
Prompt No. 3, Lunar Eclipses:
--We can't see the moon
--It's not dangerous
--It occures once in a year
Prompt No. 4, Astronomers:
--A people who works about astronomy.
--They try to figure out mysterie of space.

Prompt No. 5, Eclipses Prediction:

-—It's not dangerous that's why we don't need
serious prediction.

190

(71:71:: ,, .9 4w people know: domdkgf 51:54!- eah'pses Just
(Gite. m€-T“~cre. are. {No has of gum... '. 55.....— ail/psts 5M4
Zumr alfpses. «1er {he éormr‘ 2.5.0992. Was, {he mean comes
Beb¢% Hum AMA worH-A-L Hack Pandapkme. sonnL Psi/é
0.? earth Jaesn‘Q- MJL sonskaeflaemusc. Whom Mocks .soAslmfnLIt
mew... q (“4... mmmflwaokk 3.....— QJNN P541... N.NN- ....
41x2. 54/6117“.

The 0418.9.” LTAJ cg afifpsz :‘3 Last. uL‘PS-e, \K/‘nm “we.

cum comes 10¢.th mooq MA wet-1A , some. PJ‘- 0‘ 141$wa
mn'+ dz!" moonlifht 'm Hag racial: In “151 dried-tar) dun Made.
moon-agfb‘- lunar- “flame. Lon‘lf'nuzi an Covete_ c}?- me 12,5-

A5 far 51.: CL anl net'erf‘ aoncv" zclfp>¢ nOf‘ awrxr

eCH-QfPsa m clawmé - frkcay OCLur" Once— 7A .1 yeflf‘ Tn NAN?»
OOmC— {Dcopfle Tn ("a-e4 t NT‘H\ :41er (Duffy tk; colt“? Le ! (50m 4;
{909ng try +0 See. NAN .‘1 “PM? - (They 603 aéd‘r‘olz‘ay fiCA‘l‘cr's

&AQ1 M use. +¢1€5Ckopes- "\ c'O’H'JV-‘H'OA 1"" Cam be. tfi+¢re¢+fiy

(

I; You out. Potash-l ‘(A ash—01:307.-

 

 

 

REFERENCES

 

REFERENCES

Anderson, R. C. (1978). Schema-directed processes in
language comprehension. In A. M. Lesgold, J. W.
Pellegrino, S. D. Fokkema, and R. Glaser (Eds.),
Cognitive Psychology and Instruction, (pp. 67-82).
New York: Plenum Press.

Anderson, R. (1984). Role of the reader's schema in
comprehension, learning, and memory. In R.
Anderson, J. Osborn, & R. Tierney, eds., Learning to
read in American schools: Basal readers and content
texts. Hillsdale, N.J.: Lawrence Erlbaum
Associates, Publishers.

 

Anderson, R. C., Reynolds, R. E., Schallert, D. L., and
Geotz, E. T. (1977). Frameworks for comprehending
discourse. American Educational Research Journal,
44, 367-382.

 

Applebee, A. N. (1982). Writing and learning in school
settings. In M. Nystrand, ed., What writers know:
The language_process and structure of written
discourse (pp. 365—381). New York: Academic Press.

 

 

 

Applebee, A. N. (1984). Contexts for learning to write:
Studies of secondaryyschool_instruction. Norwood,
N.J.: Ablex.

 

 

Bamberg, B. (1978). Composition instruction does make a
difference: A comparison of the high school
preparation of college freshmen in regular and
remedial English classes. Research in the Teaching
of English, 12, 47—59.

 

 

Bamberg, B. (1984). Assessing coherence: A reanalysis
of essays written for the National Assessment of
Educational Progress, 1969-1979. Research in
Teaching of English, 18, 305-319.

 

 

Beaman, K. (1984). Coordination and subordination
revisited: Syntactic complexity in spoken and
written narrative discourse. In D. Tannen,
Coherence in spoken and written discourse. Norwood,
NJ: ABLEX.

192

 

193

Black, J. B., Wilkes-Gibbs, D., and Gibbs, R.W. (1982).

What writers need to know that they don't know they
need to know. In M. Nystrand, ed., What writers
know: The language, process, and structure of
written discourse. New York: Academic Press.

Britton, J. (1978). The composing process and the
functions of writing.

In C. R. Cooper and L. Odell
(Eds.), Research on Comppsing (pp. 13-29).

Urbana,
IL: National CounciI'of Teachers of English.

Britton, J., Burgess, T., Martin, N., McLeod, A., and
Rosen, H. (1975).

The development of writing
abilities (pp. 11-18). London: Macmillan
Education.

Brown, G., and Yule, G. (1983).

Discourse analysis.
Cambridge:

Cambridge University Press.
Calkins, L. M. (1983).

 

 

Lessons from a child: On the
teaching_and learning_of writing. Portsmouth:
Heinemann Educational Books.

Canale, M. (1983). From communicative competence to
communicative language pedagogy- In J. Richard and

R. Schmidt (Eds.), Language and communication (pp.
2—27). New York: Longman.

Carrell, P. L. (1982). Cohesion is not coherence.
TESOL Quarterly, £9! 479-488.

Carrell, P. L. (1983). Some issues in studying the role

of schemata, or background knowledge in second
language comprehension. Reading in a Foreigg
Language, 1, 81—92.

Carrell, P. (1986). "Second language pedagogy in the
light of schema theory."

Paper presented at the 6th
Annual Midwest Regional TESOL Conference, Ann Arbor,
Michigan, November 6—8.

Carrell, P. L. (1987). Content and formal schemata in
ESL reading.

TESOL Quarterly, lg, 461—481.

Chesky, J .A. (1984). The effects of prior knowledge
and audience on writing.

Unpublished doctoral
dissertation, University of Kentucky, Lexington.

 

 

194

Chesky, J. A., and Hiebert, E. H. (1987). The effects
of prior knowledge and audience on high school
students' writing. The Journal of Educational
Research, 89, 304-313.

 

Chiesi, H. L., Spilich, G. J., and Voss, J. F. (1979).
Acquisition of domain-related information in
relation to high and low domain knowledge. Journal
of Verbal Learning and Verbal Behavior, 18, 257—273.

 

Chomsky, N. (1959). Review of Skinner's verbal
behavior. Langgage, 22, 26—58.

Clarke, M., and Silberstein, S. (1977). Toward a
realization of psycholinguistic principles in the
ESL reading class. Langgage Learning, 21, 135-154.

 

Connor, U. (1984). A study of cohesion and coherence in
English as a second language students' writing.
Papers in Linguistics: International Journal of
Human Communication, 11, 301-316.

 

 

Cooper, A. (1988). Enhancing coherence through
cohesiveness. Written Communication, g, 352-267.

 

Coulthard, M. (1977). An introduction to discourse
analysis. Longman.

 

 

Crowhurst, M. (1980). Syntactic complexity and
teachers' quality ratings of narrations and
arguments. Research in the Teachinggof English, 14,
223-231.

 

Cziko, G. A. (1980). Language competence and reading
strategies: A comparison of first- and second-
language oral reading errors. Language Learning,
3g, 101-114.

 

DeGroff, L. (1986). The influence of prior knowledge on
process-approach writing tasks. ERIC Ed 269758.

Devine, J. M. (1980). Developmental patterns in native
and nonnative reading agguisition. Unpublished
doctoral dissertation, MiChigan State University,
East Lansing, MI.

 

 

Donovan, T., and McClelland, B. (1980). Eight
approaches to teaching composition. Urbana, IL:
National Council of Teachers of English.

195

Eamon, D. (1979). Selection and recall of topical
information in prose by better and poorer reaaers.
ReadingResearch_guarterly, 44, 244-257.

Elley, W., Barham, 1., Lamb, H., and Wyllie, M. (1976).
The role of grammar in a secondary school
curriculum. Research in the Teaching of English,
10, 5-21.

Eskey, D. E. (1988). Holding in the bottom: An
interactive approach to the language problem of
second language readers. In P. Carrell, J. Devine,
and D. Eskey, Interactive approaches to second
language reading (pp. 93-100). Cambridge;
Cambridge University Press.

 

Farr, R. (1969). What can be measured? Newark, DE:
International Reading Association.

 

Fish, S. (1980). Is there a text in this class?
Cambridge, MA: HarvardiUniversity Press.

Flower, L. S., and Hayes, J. R. (1977). Problem—solving
strategies and the writing process. College
English, 39, 449-461.

Flower, L. S., and Hayes, J. R. (1981). A cognitive
process theory of writing. College Composition and
Communication, g3, 365-387.

 

 

Goodman, K. (1967). Reading: A psycholinguistic
guessing game. Journal of the Reading Specialist,
4, 13—26.

Goodman, K., and Goodman, Y. (1982). Learning about
psycholinguistic processes by analyzing oral
reading. In H. B. Allen and M. Linn, ed., Readings
in Applied English Linguistics. New York:
Appleton—Century-Crofts.

Graves, D. H. (1983). Writing: Teachers and children
at work. Exeter, NH: Heinemann.

Haliday, M. A. K., and Hassan, R. (1976). Cohesion in
English. London: Longman.

Hanania, E., and Shikhani, M. (1986). Interrelation-
ships among three tests of language proficiency:
Standardized ESL, Cloze, and Writing. TESOL
Quarterly, 49, 97-109.

 

196

Hilgers, T. L. (1982). Experimental control and the
writing stimulus: The problem of unequal
familiarity with content. Research in the Teaching
of English, 44, 381-390.

 

Hoetker, J. (1982). Essay examination topics and
students' writing. College Composition and
Communication, 43, 377-392.

 

 

Hudson, T. (1982). The effects of induced schemata on
the "short circuit" in L reading: Non-decoding
factors in L2 reading pefformance. Langgage
Learning, 34, 1-31.

Hunt, K. (1965). Grammatical structures written at
three grade levels. Research Report No. 3.
Champaign, IL: National Council of Teachers of
English.

 

Iser, W. (1978). The act of reading. Baltimore: John
Hopkins University Press.

 

Johns, A. M. (1986). Coherence and academic writing:
Some definitions and suggestions for teaching.
TESOLQuarterly, 4Q, 247-265.

 

Johnson, P. (1981). Effects on reading comprehension of
language complexity and cultural background of a
text. TESOL Quarterly, 4;, 169-181.

 

Jones, C. S. (1985). Problems with monitor use in
second language composing. In M Rose (ed.), When a
writer can't write: Studies on writer's block and
other composing process problems (pp. 969—118). New
York: Guilford Press.

 

Jones, C. S., and Tetro, J. (1987). Composing in a second
language. In A. Matsuhashi (Ed.), Writing in real
time: Modelling the production_processes (pp. 34-
57). Norwood? APLEX.

 

Judy, S. (1980). The experiential approach: Inner
worlds to outer worlds. In D. Foster, ed., Eight
approaches to teaching comppsitions: Theories,
theorists, issues, problems (pp. 37-51). Boynton:
Cook Publishers.

 

197

Kellogg, R. T. (1987). Effects of topic knowledge on
the allocation of processing time and cognitive
efforts to writing process. Memory and Cognition,
42, 256—266.

Kincaid, G. L. (1953). Some factors affecting
variations in the quality of students' writing.
Unpublished Ed.D. Dissertation (Michigan State
College) Michigan State University.

Kitao, S. K. (1987). Schema theory: Reading and second
language learning. Unpublished dissertation.
Department of Communication, Michigan State
University.

Krashen, S. (1982). Principles and practice in second
language learning. Oxford: Pergamon.

Krashen, S. (1984). Writing: Research, theory, and
application. Oxford: Pergamon.

Langer, J. A. (1982). Facilitating text processing:
The elaboration of prior knowledge. In J. A. Langer
and M. T. Smith-Burke (Eds.), Reader meets Author
/Bridgingithe4gap (pp. 149—162). Newark:
International Reading Association.

Langer, J. A. (1984). The effects of available
information on responses to school writing tasks.
Research in the Teaching of English, 22, 22—44.

Langer, J. A., & Nicolich, M. (1981). Prior knowledge
and its relationship to comprehension. Journal of

Reading Behavior, 4, 373-379.

Lay, N. D. S. (1982). Composing processes of adult ESL
learners: A case study. TESOL Quarterly, lg, 406.
Lee, J. F. (1986). The effects of three components of

background knowledge on second language reading.
The Modern Language Journal, 19, 350-354.

Light, R. L., Xu, M., and Mossop, J. (1987). English
proficiency and academic performance of
international students. TESOL Quarterly, 22, 251-
261.

Littlewood, W. (1984). Foreign and second language
learning: Language acquisition research and its
implications for classroom. Cambridge: Cambridge
University Press.

 

198

Loban, W. D. (1976). Language development:
Kindergarten through grade twelve. Research Report,
18, Urban, Ill: National Council of Teachers of
English.

McCutchen, D. (1986). Domain knowledge and linguistic
knowledge in the development of writing ability.
Journal of Memory and Language, 431—444.

McKay, S. (1984). Some limitations in teaching
composition. On TESOL '83, pp. 187—194.

McLaughlin, B. (1988). Theories of second—language
learning. London: Edward Arnold.

Mellon, J. C. (1969). Transformational sentence
combining. A method for enhancing the development
of syntactic fluency in English composition.
Research Report No. 10. Champaign, IL: National
Council of Teachers of English.

Meyers, B. F. (1985). Prose analysis: Purposes,
procedures, and problem. In B. K. Britton and J. B.
Black (Eds.), Understanding expository text: a
theoretical and practical handbook for analyzing
explanatory text (pp. 11-64). N.J.: Lawrence
Erlbaum.

Mohan, B. A., and Lo, W. A. (1985). Academic writing r
and Chinese students: Transfer and development
factors. TESOL Quarterly, 22, 515-534.

Mosenthal, P. (1984). The influence of classroom
ideology on children's production of narrative text.
American Educational Research Journal, gg, 1—28.

Mosenthal, P., Conley, M. W., Collella, A., and Davidson—
Mosenthal, R. (1985). The influence of prior
knowledge and teacher lesson structure on children's
production of narratives. The Elementary School
Journal, §§, 621—634.

Murray, D. M. (1978). Internal revision: A process of
discovery. In C. Cooper and L. Odell (Eds.),
Research on composing (pp. 85—103). Urbana, IL:

 

National Council of Teachers of English.

Nas, G. (1975). Determining the communicative value of
written discourse produced by L2 learners. Utrecht,
The Netherlands: Institute of Applied Linguistics.

 

199

Newell, G. E. (1984). Learning from writing in two
content areas: A case study/protocol analysis.
Research in the Teaching of English, 22, 265—287.

 

Newell, G. E., and MacAdam, P. (1987). Examining the
source of writing problems: An instrument for
measuring writer' topic-specific knowledge. Written
Communication, 4, 156—174.

 

O'Hare, F. (1973). Sentence combining: Improving
§tudent writing without formal_grammar instruction.
Research Report No. 15. Urbana, IL: National
Council of Teachers of English.

 

Ostler, S. E. (1987). 2_study in the contrastive
rhetoric of Arabic, English, Japanese, and Spanish.
Unpublished dissertation, Faculty of the School of
Education, University of Southern California.

 

 

Pearson, P., Hansen, J., & Gordon, C. (1979). The
effect of background knoweldge on young children's
comprrpeehension of explicit and implicit
information. Journal of Reading Behavir, 22, 201-
209.

 

Perl, 3. (1979). The composing processes of unskilled
college writers. Research in the Teaching of
English, 22, 317-336.

 

Phelps, L. W. (1985). Dialectics of coherence: Toward
an integrative theory. College English, 41, 12-29.

 

Raimes, A. (1985). What unskilled ESL students do as
they write: A classroom study of composing. TESOL
Quarterly, 22, 229—258.

 

Raimes, A. (1987). Language proficiency, writing
ability, and composing strategies: A study of ESL
college student writers. Language Learning, 22,
439-467.

 

Reid, J. (1987). ESL composition: The expectations of
the academic audience. TESOL Newsletter, 4, p. 34.

 

Rosenblatt, L. (1978). The reader, the text, and the
poem. Carbondale, IL: Southern Illinois University
Press.

 

 

 

 

200

Rumelhart, D. E. (1975). Notes on a schema for stories.
In D. G. Bobrow, and A. M. Collins (Eds.),
Representation and understanding studies in
cognitive science. New York: Academic Press.

Rumelhart, D. E. (1977). Introduction to human
information processing. New York: Wiley.

 

Rumelhart, D. (1980). Schemata: The building blocks of
cognition. In R. J. Spiro, B. C. Bruce, and W. E.
Brewer (Eds.), Theoretical issues in reading
comprehension. HilISdale, N.J.: Lawrence Erlbaum
Associates.

 

 

Selinker, L., Todd-Trimble, M., and Trimble, L. (1978).
Rhetorical function-shifts in EST discourse. TESOL
Quarterly, 22, 311-320.

 

Sharon, A. (1972). English proficiency, verbal
aptitude, and foreign students success in American
graduate schools. Educational and Psychological
Measurement, 22, 425-431.

 

 

Skinner, B. F. (1957). Verbal behavior. New York:
Appleton-Century—Crofts.

Sommers, N. (1980). Revision strategies of student
writers and experienced adult writers. College
Composition and Communication, 22, 378—388.

 

Steffensen, M., Joag-Dev, C., & Anderson, R. (1979). A
cross—cultural perspective on reading comprehension.
ReadingQuarterly, 22, 10—29.

Stosky, S. (1983). Research on reading/writing
relationships: A synthesis and suggested
directions. Language Arts, 29, 627-642.

 

Stroethoff, J. (1988). The effect of prior knowledge,
prior interest, and learning on writing.
Unpublished Doctoral Dissertation, Michigan State
University, East Lansing, MI.

 

Swaffar, J. K. (1988). Readers, texts, and second
languages: The interactive process. The Modern
Langunge Journal, 22, 123-150.

 

Taylor, B. (1979). Good and poor readers' recall of
familiar and unfamiliar text. Journal of Reading
Behavior, 22, 326-329.

 

 

 

201

Talor, B. (1981). Content and written form: A Two—way
street. TESOL Quarterly, 22, 5-12.

Thomas, G. P. (1986). Mutual knowledge: A theoretical
basis for analyzing audience. College English, 42,
580—594.

Voss, J., Vesonder, G. T., and Spilich, G. J. (1980).
Text generation and recall by high-knowledge and
low-knowledge individuals. Journal of Verbal
Learning and Verbal Behavior, 22, 651—667.

Waltman, J. L. (1983). Grammar competency and business
communications. Paper presented at the meeting of
th Southwest Division of the American Business
Communication Association. Houston, TX, March 10—
12. ED 227505.

Wiseman, S., and Wrigley, J. (1958). Essay—reliability:
The effect of choice of essay-title. Educational
and Psychological Measurement, 22, 129-138.

Witte, S. P., and Faigley, L. A. (1981). A comparison
of analytic and synthetic approaches to the teaching
of writing. TWRG Research Report No. 1. Austin:
University of Texas. Department of English. ED

209677.

Zamel, V. (1982). Writing: The process of discovering
meaning. TESOL Quarterly, 22, 195—209.

Zamel, V. (1983). The composing processes of advanced

ESL students: Six case studies. TESOL Quarterly,
2_7_, 165-187.

 

HICHIGQN STQTE UNIV. LIBRARIES

I 1 \HHlle l

42fl2

    

’2

312939 78