Primary children's print awareness and self-correction behavior as predictors of reading achievement

MISSING IMAGE

Material Information

Title:
Primary children's print awareness and self-correction behavior as predictors of reading achievement
Physical Description:
viii, 113 leaves : ; 28 cm.
Language:
English
Creator:
Harlin, Rebecca Potter, 1949-
Publication Date:

Subjects

Subjects / Keywords:
Reading (Primary)   ( lcsh )
Language awareness   ( lcsh )
Self-actualization (Psychology)   ( lcsh )
Genre:
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1981.
Bibliography:
Includes bibliographical references (leaves 107-111).
Statement of Responsibility:
by Rebecca Potter Harlin.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 000295689
notis - ABS2037
oclc - 07937488
System ID:
AA00002189:00001

Full Text














PRIMARY CHILDREN'S PRINT AWARENESS
AND SELF-CORRECTION BEHAVIOR AS
PREDICTORS OF READING ACHIEVEMENT








By

REBECCA POTTER HARLIN


A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF
THE UNIVERSITY OF FLORIDA
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY


































Copyright 1981

by

Rebecca Potter Harlin
























my fami ly


Digitized by the


Internet Archive


in 2011 with funding from
University of Florida, George A. Smathers Libraries with support from LYRASIS and the Sloan Foundation
















ACKNOWLEDGMENTS


In conducting this study


I wish to acknowledge and thank the


following for their support and help:

Dr. William Powell, chairperson, who encouraged and guided the


study from


ts inception to the final form.


Lawrence


Smith, member of the doctoral committee, who


contributed valuable insight into the research problem.


Dr. Stephen Olejnik, member o

patience, kindness, and understanding


f the doctoral committee


as well as hi


for his


research expertise.


Friends, who shared thi


time and encouraged my efforts.


My parents, who taught me to set goals and to strive for them.

And to Jim and Amy, who loved me when I was unlovable.


















TABLE OF CONTENTS


Page


ACKNOWLEDGMENTS


ABSTRACT


CHAPTER


INTRODUCTION


Statement'


t ofthe Problem 4 4 4 *
L of the Problem ............


Significance of the


Study


S 4 5 4 4 S3


Implications . .
Apple ications . . . . .

Definition of Terms . . . .


Hypotheses

Limitations


* 4 5 4 4 4 0 4 5 4 4 4 4 4 4 4 4

* S 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4


Organization of the Study

REVIEW OF THE RELATED LITERATURE


Children's Perceptions of the


Reading Act .

Early Readers .

Print Awareness .

Print Awareness and Achievem'

Self-Correction Behavior














CHAPTER


Page


The Instruments . . . . .

Self-Correction Rates . . . .

Collection of Data . . . . .

Treatment of Data . . . . .


ANALYSIS


OF THE DATA


S 66


Findings Related to the Hypotheses

Findings Related to the Questions


U U U U S U

U S a a a a


SUMMARY


, CONCLUSIONS, AND RECOMMENDATIONS


a S S S S 590


summary


Impl ications

Recommendations . . . . .


APPENDIX:


DATA FOR INDIVIDUAL SUBJECTS


S .. . 98


BIBLIOGRAPHY


BIOGRAPHICAL SKETCH

















Abstract of Dissertation Presented to the Graduate Council
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy



PRIMARY CHILDREN'S PRINT AWARENESS
AND SELF-CORRECTION BEHAVIOR AS
PREDICTORS OF READING ACHIEVEMENT

By


Rebec


Potter Harlin


August 1981


Chairperson:


William R


Major Department:


Powell


Instructional Leadership and Support


This study examined the relationship of two print awareness in-


dicators, the Concepts about Print Test


Sand, and the average


self-


correction rates or oral reading errors, to subsequent reading


achievement.


The following questions were ra


1) I


there a


difference in the development of awareness (a


(b) among grade


among age


, (c) between prereaders and readers?


evel


(2) Is


there a pro


across


self-co


gress


a) book 1


ion in the development of a


, (b


rrection to errors


grade


estimate the


self-monitoring reader


Does the ratio of


evel of reading achievement?


(4) Does the


of print awareness


estimate the level of reading


achievement?












asses


of first graders, and all Title I second and third grade


children.

All subjects were tested with the Concepts about Print Test,


Sand


by the researcher.


In addition, oral reading errors were recorded.


Both indicators


, the Sand and the average self-correction rates, were


used to predict subsequent reading achievement.

The data were analyzed using Pearson-product-moment correlations,


of variance


, linear and multiple regression anal


data indicated that the


Sand


scores were effective predictors of read-


ing achievement for the Titl

correction rates were not.


I children, but the average


However,


self-


it was found through multiple


regression analysis that the two indicators when combined did not


effectively predict reading achievement.


cores on the Sand were


found to correlate significantly


with the Comprehensive


ests


of Basic


Skill


, Levels B, C, and I; and the Metropolitan Readiness T


ests,


Level


I and II.


Linear trends were found for the Sand scores and the


subjects


age and grade


evels.


Self-correction rates appeared to


fluctuate according to the


evel of the basal reader.


The conclusion drawn from the data is that indicators of


print awareness, such


as the Sand, are effective predictors of reading


achievement for primary grade children.

















CHAPTER I
INTRODUCTION


Each year


as the first day of school begins, millions of


children enter classrooms for the first time in their 1i


ves


Many of


these children


expect to learn how to read (Brumbaugh


, 1940)


However,


what they expect to learn, what they actually will


earn, and how well


they will learn may be dependent upon factors other than intelligence


and home background


Their succ


n learning to read may be closely


related to their mastery of specific


tasks before they entered school


These task


may be an under


standing of the reading process itself,


linguisti


awareness,


and familiarity with print concepts.


It ha


come apparent from the research in the area of reading readiness that


the aforementioned tasks are often overlooked


n determining readiness


for reading. More research is needed to determine the extent of

children's cognitive confusion in learning to read and the effect of

that confusion on subsequent reading achievement.


ever


studio


of beginning read


(Day


Spicola, and


Griffin


, 1979) hav


indicated that there i


a relationship


between linguistic awareness


and subsequent su


access


in learning to


read


however


, the strength and the nature of the relationship has not






2





The lack of consistence among these related studies may be


attributed to their differences in sampling and methodology


investigating linguistic awareness


Researchers


have often depended upon small


samp


to test their hypotheses (Downing


1967


Reid, 1966)


In addi-


tion, their sample


consisted primarily of white, middle-class


subjects which limited the generalizability of their conclusions to a


population which differed


socioeconomic


status and race


, as well


geographic area (Holden and MacGinitie, 19


results of these studies may al


Differences in the


explained by their varying method-


ology


In some


studi


interviews were employed to obtain insights


into the


child's knowledge of the reading process


while


in others,


tasks or informal tests were used.


Researchers also differed in their


criteria for


success


ful readers.


To some, the criterion was a score


on a standardized silent reading test. To others, it was the number or

kind of miscues from an oral reading sample. Still others used the


f-correction rates of errors


of the above discrepancy


as their criterion (Clay, 1969)


have prevented the relationship between


linguisti


awareness and


success


in reading from being clearly defined


or widel


accepted.


This study was concerned with the relationship of linguistic


awareness to subsequent reading achievement


The major focus was on


measuring the changes in the development of print awareness across age


and grade


as well


as time.


It was al


intent of this


study


v






3





Statement of the Problem


The primary purpose of this


study was to observe the development


of print awareness in kindergarten and first grade students as well as


in Title I second and third graders during the course of the year.


This


study was meant to relate this development to later reading achievement.


A second purpose of thi


study was to observe the development of


a self-monitoring


stem in Title


readers through a longitudinal


record of their oral reading errors and subsequent self-correction.


Answers to the following questions were


sought:


Is there a difference in th


development of print


awareness


Among age


levels


b. Among grade levels

c. Between pre-readers and readers?

Is there a progression in the development of a

self-monitoring reader across


Book

Grad


evels

level


Does the ratio of self-corrections to errors estimate


evel of reading


Does the


achievement?


eveT of print awareness estimate the


of reading achievement?


-r *.












was inconsistent.


Through this study


as well as subsequent ones, the


stages may be more clearly defined and their relationship to reading


achievement established, for normal


as well as for disadvantaged readers.


A second implication for future research was the use of the

Concepts about Print Test, Sand, with older disadvantaged readers.


Through its use in this


study as well


as those of future studies,


reliability for older


students may be substantiated.


From the oral reading sample


their predictive value for


and the self-correction rates,


estimating the reading progress of educa-


tionally disadvantaged students was demonstrated.


For the predictive


value to be substantiated, more longitudinal research in this area,

using below average readers may be necessary.


Applications


f studies such as thi


awareness exists in


one help demonstrate that linguistic


stages and that these stages can be taught, there


are several applications for practitioners.


However


the results


from this research study would only apply to the population from which

the sample was drawn.


From future research and from the data gathered in this study


be found that definition of the stag


it may


of linguistic awareness may be


useful in diagnosing potential


reading failure.


Secondly, if it were


possible


to move


students from one


stage to the next through classroom






5





terms of sequence or duration, methods of teaching the disadvantaged

reader may change.


There have already been a few studio


which correlate the results


of the Concepts about Print Test, Sand, with


established


standardized


instruments (Day and Day, 1979a; Johns, 1980)


The results of this


study as well


as others may encourage teachers


in classrooms and clinics


to use the Concepts about Print Test a


part of their diagnosis


in l ieu


, or in conjunction with widel


used standardized tests.


of the


Concepts about Print Test may save


time, effort, and money since


invol


a ten-minute period, a life-like situation, and requires only


a manual and a children


s book entitled Sand.


It is


easy


to administer,


score, and interpret.


This study has yielded additional information on


behavior


self-correction


as it relates to text difficulty and the reading level of the


disadvantaged


It may help classroom and remedial teachers to estab-


lish acceptable criteria to gauge the progress


of the individuals with


whom they work.


The criteria may or may not be the same


average-progress readers.


since


as that for


self-correction behavior may be noted


daily


as part


f the reading lesson,


it does not involve


e an additional


record-keeping burden for the teacher.


With uniform criteria, it would


be possible to exchange information concerning a student


progress with


ocal


education agency teachers, Titl


staff, supervisors, and parents.


r I J~ -


I' r ~P


n r~ r













Basal reader:


A basal reader is one text of a coordinated


series of material


that is designed to provide for developmental,


sequentia


stemati


reading proficiency


Concepts about Print T


est, Sand:


The Concepts about Print Test


Sand, is an informal test designed to measure a subject's knowledge of


significant concepts about printed language, such as word

function of space between words, and uses of punctuation.


of a


, letter,

It consists


storybook entitled Sand and a manual which lists 24 questions


asked by the examiner


as he reads to the child


Directionality:


Directionality i


the orientation to print as


it relates to the reading process, consisting of beginning at the top

left corner of the page, movement along the line of print to the right

margin, and return to the left margin at the beginning of the next


ower


ine of print, continuing until one reaches the bottom of the


page.


Error or miscue:


An error or miscue is any deviation from the


printed text, including substitutions, omissions, unknown words, mis-


pronunciations


and transpositions.


Linguistic awareness:


Linguistic awareness is a child's under-


standing and knowledge of the similarities and differences between


written and spoken language such as


segmenting sentences into words,


identifying word boundari


in print by the


space between words, the


ii a.. II a -I1 I1 I






7





one grade level below that in which the Title I students are currently


placed.


Scores for these lower ability, Title I students are determined


from a set of norming tabi


created especially for this purpose.


Such


instruments are used to obtain a more accurate estimate of the student

actual achievement than would be possible using an appropriate grade

level test.


Print awareness


Print awareness is used synonymously with lin-


guistic awareness in this study.


Self-correction:


elf-correction i


the student's


spontaneou


recognition of an oral reading error and hi


her subsequent


ubstitution


of the appropriate response to the printed stimulus.


Word boundaries:


Word boundaries are the physical limits of


a printed term designated by the space preceding its initial letter and


that following its final


better


Hypotheses


The following hypotheses were tested at the 0.05


evel of sig-


nificance:


There are no differences in the level of print aware-


ness, as measured by the tota


score on the Concepts


about Print Test, Sand, among the age level


There are no differences in the leave


ness,


tested.


of print aware-


as measured by the total score on the Sand


2mnnn 4-ka nrA 1 cd -oc a


r tarf all






8





students and the Title I second and third grade

students.

There are no differences in the gain scores on the

Sand between (a) first grade students and Title I


second grade


students, (b) first grade


Title I third grade


students,


students and


c) Title I second and


Title


I third grade


students


and (


d) first grade


students and Title I second and third grade students.


There


no relationship between the total score on


the fall-administered Sand and the total reading score


on the appropriate grade


evel test of the Comprehensive


ests of Basi


Skill


(CTBS) for the non-Title I first,


Title I second


There i


, and Title I third grade students.


no relationship between the total score on


the Sand and the total


score on the Metropolitan


Readines


s Tests for the kindergarten and first grade


students.


V Ii


There is no relationship between the total score on


the Sand and the total reading score on the appropriate

level of the Metropolitan Achievement Tests (MAT) for

the Title I second and third grade students.


VIII.


There is no relationship between the average


correction rate on the oral reading samples and the






9




There is no relationship between the total

score on the Sand, the average self-correction


rate on the oral reading samples


and total


reading score on the MAT for the Title I second


and third grade


There i


students.


no relationship between the total


score on the Sand and the average


self-correction


rate on the oral reading samples for the

Title I second and third grade students.

There is no relationship between the average

self-correction rate and the oral reading samples


and the total reading s

level CTBS for the Titl


.core on the appropriate


I second and third grade


students


There is no relationship between the total score


on the Sand


, the average self-correction rate on


the oral readings sample


, and the total reading


score on the CTBS for the Title I second and

third grade students.


Limitations


These limitations outlined the boundary


The sample


of the study:


population was restricted to primary












The Title


sample population was restricted to second


and third grade students.


Results from this study can


be applied only to disabled readers of these grade


level


The instrument for measuring print awareness was

limited to the Sand.

The oral reading samples of Title I children used in

this study were continued to the following basal reader

levels:

a. First pre-primer

b. Second pre-primer

c. Third pre-primer

d. Fourth pre-primer

e. Primer

f. First grade reader


Organization of the Study


In Chapter I


described.

of the study


the study is introduced and the research procedure


Chapter II reviews the research that is related to the forms

. Chapter III describes the collection and treatment of


the data.


Chapter IV reports the analysis of the data.


Chapter V


includes a summary of the study as well as implications and recommenda-

tions for future research.

















CHAPTER II
REVIEW OF THE RELATED LITERATURE


Studies involved with linguistic awareness and its relationship


water achievement in reading have not been numerous


a few studies


There have been


concerned with children's perceptions of the reading act,


usual


sampling preschool or kindergarten children


A few studi


have


been


concerned with the reading processes of the self-taught early


reader.


This review was concerned with the above topics


as well


as the


development of print awareness in primary grade children


, the relation-


ship of this awareness to later reading achievement, and the use of self-


correction rates


indicators of reading progress.


Children's Perceptions of the Reading Act


One of the earliest studies of children'

reading act was done by Brumbaugh (1940). Over


in thirty-one


expectations of the


700 kindergarten students


classrooms of the New York City public schools were


interviewed by


student teachers.


The interviews, which took place


one month before promotion to first grade, concerned the children's


perceptions of first grade learning opportunities, particularly


reading.


When


asked what they thought they would do in first grade, onl


50 per-


cent


said that they


expected to read.


Most responded that their mothers,












reading would be hard for them, while


33 percent though it would be


easy.


Although this study was not a model of scientifically designed


research,


its conclusions are worth noting.


The study proposed that


perhaps emotional factors are more responsible


for suc


cess


in reading


than was previously


believed.


Given that half of the sample did not


expect to read in first grade


it would have


been interesting to inves-


tigate whether or not these


addition


children actually did learn to read.


, some follow-up of those children who thought reading would


be difficult


as well a


those who thought it would be easy, would have


helped define the relationship between the children


expectations and


their subsequent reading performance.


In Great Britain


employed by earl


, McKinnon (1959) studied the reading strategies


readers in their cooperative efforts to read


unfamiliar material for the first time.


From the observations recorded


during the


session


ns, valuable


information about children's approaches


to breaking the code and an


indication of their conception of the reading


act were gained.


material


Through the


control over both


use of the Richards-Gibson reading


better and word intake was possible


Simple sentences were paired with line drawings which indicated the


sentence content


At first


, the children relied heavily upon the pictures for


meaning.


Although they


were able to read the accompanying sentences,












between parts of the text, the words these


of words


ves, and component parts


Although the British subjects used in the study were of


varying


of ability, all were abl


to mak


contributions to, and


earn from, the group


In follow-up testing


McKinnon found that, for most subjects,


there was a transfer of their knowledge to other reading situations.


success


problem


which group members experienced when they solved a new


seemed to excite them and to spur them on to further explora-


tions of reading


When errors did occur, it was noted that there were


three consistent types of errors--those due to reverting to a previous


sentences, those semanticall


acceptable with the illustration, and


those semantically


acceptable


but untrue.


These error patterns were


ater substantiated


Francis


, 1977)


During the first week in


school in 1962


, Weintraub and Denny


1965


individually interviewed five


and taped their responses.


A tota


sses


sample


of beginning first graders


of 108 children of widely


divergent socioeconomic status was interviewed.


represented in the study were rural


The social strata


all Negro, middle class, and


ower cl


Students were


asked,


"What i


reading?"


Their


responses were categorized in two ways--object related


which referred to the material


one reads, or


cognitive related which


referred to reading


as a cogniti


act.


About


33 percent of the


responses were object related;


0 percent were cognitive related






14





who come to first grade having no understanding of what the reading


process actually is.


As in Brumbaugh's study, it would have been bene-


ficial to interview these same children at the end of the year to deter-

mine if their definition of reading evolved to one resembling the


teacher


so, a correlation of their reading achievement with their


initial response would have been useful


In a much smaller study


notions about reading


Reid


children had at t


1966) attempted to study the

he beginning of school and how


those beliefs changed during the


year.


From a cl


of 40 in an


Edinburgh school


Reid randomly


selected a sample of twelve five-year-


olds--seven boys and seven girl


The children represented various


socioeconomic


strata.


Each child was interviewed three times during the


first year of school.


The results of the first interview indicated


that almost all of the children were aware that they could not read


However, many


children did not know precisely what reading was.


There


was little


indication that the children were aware that words were


composed of letters although some children were aware of what writing


was.


Reid indicated that, at this point, the children had not learned


the difference between pictures and written symbol


that all language, written and spoken,


By the end of the


nor did they realize


is composed of words


second and third interviews, progress toward


organizing the material and the technical terms of reading instruction


was indicated; however, that prog


ress


was uneven and did not move












depended upon whether or not he had the vocabulary which would help


him understand and grasp the schemata of hi


s language.


One of the


problems with Reid


s study


beside the small sample, has been the


requirement that the children verbalize somewhat difficult concepts

without concrete aids. However, her study was significant because it


was one of the first to


explore children'


grasp of the technical vo-


cabulary of reading, such


as "letter" and "word."


In a replication of Reid'


study, Downing (1967) endorsed


Reid


original research.


The five-year-old children in thi


sample


of 13 English


students were beginning to read their basal


One of the


important findings of Downing'


research was that young readers were


still having difficulty understanding the purpose of written language.


Secondly


the children had onl


a vague idea of how people read


They


also indicated some difficulty understanding abstract linguisti


terminology


In this replication, there were two differences from Reid


original


study


First, the children were given a book without pictures


to answer the question,


"What part of


a book do adults look at when


read?"


econdl


y, pictures and concrete objects were used when


asking the


children to identify the learner car and the correct bus


for a given destination.


It was found that the children could achieve


more with the


concrete objects


as aids during the interviews than in


the more abstract verbal situation of Reid'


study


However


the chil-












In a second experiment involving the same children,


Downing


asked the children to discriminate between auditory stimuli on the


basis of whether or not the stimulus was a word.


A second task involved


discriminating a


sound from a non-sound.


The results indicated that


the children had difficulty both in discrimination of words from all


stimuli and in sounds from non-sounds


In this study of the


Hemel


Hemstead children


, Downing found that concrete aids helped the children


demonstrate their understanding of the concepts related to what adults


do when they read.


Downing's replication of Reid's research further


underscored the confused state many young learners are in when reading


instruction i


initiated


Both studies demonstrated the importance


of understanding the technical


Using a sample of


terminology of reading.

Georgia pre-schoolers, G. Mason (1967


explored their perceptions


of the reading act.


The sample of three-


four-


, and five-year-old


was a racial


socioeconomically, and


intellectual 1


stratified one.


During individual interviews, the


children were


asked whether they


iked to read


Almost all of the


children responded that they did like


to read.


Of those


who responded


affirmatively, almost 90 percent


said they could read all by them-


ves.


Since al


of the children in this sample had not yet entered


school


, it is obvious that many who responded that they could read,


actually could not.


what reading


This


actually entail


indicated that they may not have understood


Mason concluded that perhaps one of












Similar results were obtained by Johns (1972) in a longitudinal


study of fourth, fifth, and sixth grad


sample


students.


Unlike G. Mason, the


interviewed included subjects who were already reading.


Both


able and disabled readers indicated their


ack of understanding of the


process in their response to the question


"What i


reading?"


Johns,


like other researchers, hypothesized that this lack of understanding


may contribute to reading failure.


encouraged better communication


about the reading pro


cess


between researchers and teachers,


as well


between teachers and students,


as a way to better utilize the availabi


knowledge.


Studi


of minority children have helped highlight the miscon-


ceptions of what the reading act entails.


Oliver (1975)


studied the


development of the concepts of letters, reading, and writing, in Yakima


Indian children.


From his sample


of 78


three-


four-


, and five-year-


olds


, he found that many confused writing with drawing, reading with


looking


and letters with numerals.


When asked if they could read,


most of the three-year-olds responded affirmatively


while the five-


r-old


indicated that they could not read.


The educational level


and the reading ability of the parents appeared to have a strong


cor-


relation with the d


development of print and language concepts.


Children


from homes where little


difficulty responding ve


reading was done by parents had the most

rbally to questions and appeared to be unaware


of the purpose of written language (Durkin, 1966; Clark, 1976)












reading was supposed to make sense.


It appeared that the absence


of a match between the reading material and the


experiential level of


the student was responsible for the misconception of reading (Sider-


bergh


977).


To extend knowledge regarding literacy behavior and sociocul-


tural factors


, Downing, Ollila, and Oliver (1976) sampled 92 non-


Indian and


ndian children of kindergarten age in Canada.


Both


groups were given the Canadian Reading Readiness battery which consisted


of five


subtests.


The Indian children


who had


experience with


iteracy behavior and speech due to their culture,


scored significantly


ower than their non-Indian counterparts on four of the five subtests.


Their concepts of the function of reading and writing,


as well as their


understanding of the terms "word" and "letter," were immature and


inadequate


as measured by this test


From this study


the researchers


concluded that children's

by sociocultural factors.


Concept of literacy skills i


As in other studi


influenced


already discussed,


follow-up of these subjects would have been useful to determine whether


or not

ences.


school


are successful in compensating for these initial differ-


This study may cause one to be somewhat


skeptical of its results


because


the instrument used


n measuring


literacy behavior may be


biased


n its instructions in favor of the non-Indian


children


Tovey (1976) sampled children from arad


one through


six in


a small midwestern community in order to gain


night into children's













Tovey was interested, among other things, in determining whether or


not children perceived reading


as a silent process.


Another interest


was to find how much the children knew about the process itself and how


they approached the reading task.


Each


child in the sample was inter-


viewed on three separate occasions.


When asked to read a book, it was


noted that only six children chose to read it silently


y although the


researcher did not specify whether it should be read orally or


silently.


This amounted to 11 percent of the 90 possib


opportunity


It was


noted that none of the first or


silently.

they read


second graders chose to read


When the children were asked what they thought they did when


percent indicated that reading had something to do with


meaning, while 43 percent thought reading had to do with looking at or


pronouncing words


Subjects were al


every word when they read.


asked whether they looked at


A minority, 17 percent,


said that they


did not look at every word while


percent


said that they did.


Follow-


ing these


responses,


subjects were asked whether or not they thought


it was necessary to


ook at


every word.


About 30 percent felt it was


unnecessary, while


percent


d it was.


From the results of these


interviews, it is noteworthy that


children viewed reading not a


a silent process


as an oral one


Is this an indication of th


type of instruction they are receiving


Since a majority of students reported that they still look at every


LL. L r, 2j.


~~1inrrI~ hoc' r4li Li I 1 rc Hnrr 1n r .i ffffffV


-- ---------


~II~ rTJ k Or


nnh m71r I.lnH~n~ 7~n11C C ir n







20





The importance of dealing with children's misconceptions of


reading cannot be underplayed.


have indicated


As the studies cited in this section


, these early misconceptions left unquestioned, appear


to further complicate reading acquisition in older students, particularly


minority and bilingual


students


Important questions concerning the


role of the teacher and the utilization of group problem solving (Mc-


Kinnon, 1959) have been raised.


How much does


the teacher facilitate


or hinder?

hypotheses b


Does


the text used for reading lend itself to testable


children?


Where is the match between the learner


experience and the


text he's


expected to read?


How much parent involve-


ment in the reading


process


should be encouraged


Early Readers


As adults become more proficient in their ability to read,

this sophistication interferes with their ability to recall the


processes


by which they


learned to read.


In retrospect, reading does


not appear to be a difficult task.


An exception to those adults who


have forgotten their struggle with reading was Jean-Paul Sartre.


account of his exposure to the world of books, Sartre (1


his fascination with print and his eventual mastery of it.


In an


recalled


His recol-


election was a vivid description of the successful experience we wish


all children would


share in their encounters with print:


There, perched on a folding bedstead, I pretended


.1-- r ~ I ~ I II


i













catechumen; I even gave myself private lessons: I
climbed on to my folding bedstead with Hector Malot's
Sans Famille, which I knew by heart, and half-


reciting


, half-deciphering it, I went through every


page, one


after another:


when the 1


turned, I knew how to read.


One of the most frequently


page was


194)


cited researchers in the area of


readers


was Dolores Durkin.


Her longitudinal studi


of early


readers in California, fr

attempted to identify the


read (1966)


characteristic


om 1958-1964


processes


It was found that the


, and in New York, from 1961-1964,

y which young children learn to


early readers shared many common


Among these were an interest in printing,


spelling,


sounds of letters, meanings of words, and identification of written


words,


letters, and numbers (Hildreth


, 1936; Lavine, 1977


Help with


these activity


was usually given by the mother whenever the child


expressed an interest in one of them.


Mothers al


were important to


early readers by serving


as role model


of adult readers and by reading


to their children.


To gain insight into the reading


processes


of a self-taught


reader


, Torrey


973) conducted a


case


study of a black five-year-old boy


during the 1969


school year.


The child


, who came from a poor family,


had taught himself to read by watching television commercial


attending closely to them

standardized instruments,


and by


When the researcher tested him, using


he was not found to be exceptional in


verbal ability based on hi


test


scores.


Torrey's conclusions were












and what does that print


The child in this


case


study expected


the print to


print accordingly

high verbal abili


something understandable to him and reacted with

. Another key point made by the researcher was that

ty and cultural privilege may stimulate reading, but


are not necessarily crucial


as evidenced by her own


case


study subject


as wel


as other self-taught readers.


From thi


case


study, one may


wonder how thi


child intuitively understood the


conventions of print


while


many of hi


peers in other classrooms did not.


Many


imilariti


in results were found in the study of young


readers by Clark (1976)


Like Durkin's and Torrey


subjects, her


samp


represented a variety of backgrounds


, family si


and income


Her subjects al


expressed an early interest in writing and


reading.


While mothers again played a significant rol


>y giving help


upon request, participation of fathers in the reading process


increased.

Stammer (1


As in Hiebert's research


0), the children, especially


) and that of Dewitz and


had become aware


of print through product logos and television advertisements.


addition


, Clark reported that some of the


explanations of terms such


children could give verbal


as "letter" and "word"''


however, most of the


children were able to demonstrate their knowledge of these concepts.


Unlike other


studies


, the importance of libraries was emphasized in


conjunction with the development of earl


readers, primarily


as providers


stimiil atini rrpai inn matpr ial


J I1







23





native language), using the methodology outlined in Glenn Doman's


book, How to Teach Your Baby to Read.


experimental methodology


stressed the importance of allowing the child


herself, to discover


the codes of both written and spoken language.


At the beginning of the


experiment, however, the girl was


taught to proceed hori


zontall


y from left to right, from the top of the


bottom of the page, and to turn the pages right-left.


Soderbergh'


results were similar to those of Clay (1969) in terms of voice-print


match, search-and-check behavior


self-correction, and the need for


linguistic repetition in the learning process.


Both studio


emphasized


the importance of meaningful reading and of opportunity


to apply


one's


knowledge.


Like McKinnon (1959), Sbderbergh believed that with


simple


syntax and control of letter and word intake, it was possible


for the reader to discover the code on her own.


Each of the above

of the beginning reader.

of answering children's q


studies contributed to better understanding

All of the studies emphasized the importance


questions about print and writing and of


attempting to satisfy their natural


curiosity.


success


with which


the self-taught and earl


readers broke the written code should prompt


a closer examination of the strategies taught to beginning readers.


Are these strategic


impeding or expediating the process?


Can the


acquisition of reading occur in a more naturalistic, unhurried way in a


ccrhnnl rcl

Hnw ran thp tpachahlp mnmpnt for reading be


liii'


rl1h r







24





Print Awareness


Perhaps the first


study of the stages of initial reading was


that of Gates and Boeker (1


They observed that th


perceptual


learning of children resembled that of adul


Chinese characters.


who were trying to learn


Although the children lacked the same degree of


skill


as the adult


both studied the word or character until a detail


was identified which helped the learner remember the whole word or


har-


acter.


For the children, the detail was a feature such


as the tail in


the word "monkey," or the similar beginning and ending of "window."


Learning to


see words


as separate entities was identified as the most


difficult task in


earning to read.


Therefore


, the researchers pro-


posed more studies in the area of word perception in developing readers.


Similarly


nation stra

graders).


tegi


Marchbanks and Levin (1965) compared the word recog-

of non-readers (kindergarteners) with readers (first


Their investigation focused on configuration, first letter,


last letter, and medial letters)


and quingrams.


as cues to recognition of trigrams


The results indicated that both groups of subjects


followed a similar pattern.


A majority of children used the first


letter


as the most important


cue; the last letter was second most


important.


cue.


Shape of the word (configuration) was designated the weakest


The researchers hypothesized that the use of the first and last


letters was probably a function of the white


space


bordering either


A~~~~~~~\ 1l w ka 04n1 YI nl~c Qrnnl fln77 nn 4n


latt ar I\ 1


nrn1~~1 anrl Mn nm~nn


af ar rfllrl


I


1













who had been in first grade for two and a half months.


Within approxi-


mately a ten-minute period, there were nine separate tasks for the


child to complete.


From the child


performance, th


researchers


were


cooking for three kinds of information.


First, how does this


child define a written word, and what are word boundaries (according


to this child)


Second


y, can the child discriminate between numbers,


letters, and/or words?


Third, does


the child grasp the concept that a


whole


may be made up of


several


parts?


From their resul


The first was that there i


the researchers stated two conclusions

a sequence of development in the child


concept of word boundaries before the child actually uses the space


between words


as a boundary.


At the beginning of this sequence, the


child believes that letters are words.


At the end of the sequence


child uses space as a boundary unless there i


a tall


better in the middle


of the word.


The second implication was that the use of certain


reading materials by the children may result in their having incorrect


cues to word boundaries


For example


, the researchers found that in


the reading seri


used by this sampi


, the word boundari


were often


defined by tall letters


either at the


beginning or end of the word.


This may hav


added to the children's misconception that word boundaries


are defined by tall letters.


children who were th


The study also involved 1


poorest readers in their grad


second grad


Of these chil


A2Jt 'I A. t. .













and Figa (1972)


Forty-five first graders were randomly


selected from


150 in a rural northeast Goergia school, representing upper, middle,


and lower classes.


Subjects were


assigned to one of three conditions--


basal reader two words/sentence to six

approximation to English using nonsense


letters


tion where word


words/sentence; second order


words having the same number


in the basal reader condition; and, adult novel condi


ength was longer but the number of words per sentence


was the same


as in the other two conditions.


All sentences were typed


on strips


of paper.


The subjects were


asked to cut the strips wherever


they thought a word occurred.


The results indicated that the most


common error was that of combining two or more words.


There were no


differences in the results between the first and second treatment con-


editions.


Perhaps the


reason for this is that the word length was the


same in both conditions.


However, the researchers did not find the


developmental trend


as had Meltzer and Herse.


In contrast, a later


replication by Evans, Taylor, and Blum (1979) did support Meltzer and

Herse.


What is a word?


The answer to this question was found to vary


in children aged four to ten


Papandropoulou and Sinclair


years


1974)


In an experiment conducted by


the differences in their answers


indicated that the children were at various level


of "metalinguistic


competence.


Four level


of awareness were designated within the age


rnnnp nf thp


pYnprimpntal Ramnlp


At the vnunaPSt 1


oii l"


,rnac fniir tn


cv uuL~ UUI ~V














ages five to seven, that words are conceived of both


abels.


as comments and


Children begin to talk about the number of letters in words.


From ages s

and to act


ix-and-a-half to eight,

as elements of larger un


words were believed to carry meaning

its such as sentences or stories.


The highest level


, from ages eight to ten, was the stage in which


word


were recognized


as meaningfu


units


, usually with


letters.


this level


grammatical

conclusions


, children were also beginning to think of words in terms of


structure, but had not accepted this aspect of words.


of their research with Swiss children was that


, although


children had great facility with their language, their capacity to think


about their language and to


express their conceptions, developed


slowly


(Chappell


Pick


Unze


, Brownell, Drozdal, and Hopmann, 1978


Templeton and


vey,


980)


Similar


a study proposed by Sulzby and


Templeton (1980


has begun to further explore metalinguistic development.


Perhaps some of the most important research in the area of


linguisti


awareness was that done by Clay (1969).


Marie Clay's work


is an improvement over the aforementioned


studies in methodology


sampling

involved


, length of time, and implications.


00 five-year-olds in New Zealand


In her study


sample


which


of reading behavior


were taken week


. The samp


were obtained


as the subjects read


orally with their classroom teacher.


During the prereading phase, the


child passed through


several


1 sta


while trying to find print to


- I


I !


inlr n rrn flil h Cn flttf rc~r r+nnr r'r 'r. ,r ,,. 'IrflllLnl I,- ,rt


.I _- *


I i 1 II


A 1


I


I


I" ll







28





eliminated the awkwardness of the tasks and interviews in the pre-


viously mentioned studies.

on New Zealand youngsters,


in Europe and the United Sta


since this instrument was initially normed


it has been undergoing validation for us


This instrument was partially vali-


dated in the United States through the work of Day and Day (1980b).


Sand has provided researchers with a uniform way to measure somewhat


difficult and ambiguous concepts.


During the


1967-68 school


year


Holden and MacGinitie


1972)


interviewed 84 children at the end of their kindergarten year


sample


which was white and middle class, was composed of 47 boys and


girls.


In the first experiment, the children were tested indi-


vidually


The task was to tap one poker


chip for


each word in a tape-


recorded sentence played to the subject.


The results indicated that


the children often confused function words and usually combined them

with the following content word.


In the second experiment


, using


of the original 84 subjects,


the child was


asked to


segment a sentence he heard orally.


Then he


was asked to indicate which of the four sentences on a card had the


same number of word


subjects were


he had heard on the tape.


shown sentences written in nonsense


During thi


words


phase,


This


resulted in confusion for many


training.


children because they had had some phonics


The remaining 24 subjects were shown sentences containing


thp arntIal wnrrd thpv had hoard nn f-ano


ThP rP~cl rlfE tn fhi' ovinri-







29





to teach the children about this convention, using each subject's


first and last names.


There were onl


y five subjects who were able to


perform thi


task


consistently after the teaching.


This study indi-


cated a need for further study of children's conceptions of word


boundaries,


as well


as better methods of teaching the print convention


of space and better-designed


experimental tasks


In a study of word boundaries a


perceived by first graders,


Mickish conducted a study similar to Holden and MacGinitie


samp


used by Mickish (1974)


included 117 white children, primarily


from upper- and middle-cla


of their first


suburban backgrounds


ear of reading instruction.


who were at the end


Their task was to mark


between the words of a sentence which had no spaces to designate word


boundaries.


There were si


words in each sentence.


The words them-


came from the reading seri


read by the children during that


school


year.


The results were that almost 50 percent of the subjects


were unable to correctly mark the


and familiarity with them in their rea


words in spite of their exposure

ding. The results of this study


emphasize the cognitive uncertainty children have concerning word


boundaries even after a year of reading


Although thi


instruction (McNinch, 1974).


study was an improvement over that of Holden and


MacGiniti


in terms of the task, one trial per subject i


insufficient


to judge performance fairly


flr,a.n; n n n nrAfl ,'v


0l7A\ t-.n ;rln nnl;r ,nA nnrv~lnr rnn~r'


r i


F







30





and an auditory rather than a visual retraining task, an experiment


was conducted to determine the subjects


ability to discriminate a


single word from other auditory stimuli


Preschoolers


, kindergarteners,


and first graders--14 children


from


each age group--as


to respond "yes" when they heard a single


word.


The results confirmed


Downing's previous findings


that young children do not have an


adequate


concept of what constitutes the


spoken word.


In a replication of Downing and Oliver's study, Johns


1977)


attempted to determine whether the results obtained for the British


Columbian


children would be true for American children


as well.


Johns'


sample included


0 children who were predominant


white and middle-


to upper-cl


ass.


There were


0 boys


0 girl


included in each of


three age ranges--


to 6.5


6.6 to


0, and


.1 to 9.5 years.


Downing and Oliver study had onl


contained the first two ranges; Johns


added the third group to investigate whether the confusion of


with words continued.


earlier study--young child


llables


results generally confirmed those of the

en, even at the early reading level, do not


seem to have grasped the concept of a spoken word.


In this study


almost 40 percent of the beginning readers, who ranged in age from


6 to 6.5


, consistently failed to recogni


a spoken word


as a word.


Although they were cons


istently ab


to distinguish phrases and sen-


tences from words, most of the children in the


.1 to 9.5 range


J^ J J L1 ..


1~11







31





of his results, the methodology of both experiments has no established


reliability or validity.


Perhaps the use of auditory stimuli was


too confusing to the children involved.


Many


children seemed to


experience confusion because they identified


as words, considering these


syllables and phonemes


as words they had not yet learned


In a study attempting to confirm Reid


s (1966) and Downing's


(1967


earlier


research, Franci


(1973) sampled 50 children, 20 boys


0 girls, who were from fiv


to seven years old


The children


were described


as having good


social background and above average


ability.


The children were tested at six-month intervals for a total


of four testing


sessions.


In each


session, the children were asked


to demonstrate their understanding of the concepts of letter, word


and sentence,


initially through questioning and then with a visual


recognition task.


Her results indicated that the concept of


better


presented the least overall difficulty and was learned before word.


The concept of word preceded the learning of sentence


Both word and


sentence were


children

language


earned while


, when questioned


the children were already reading.


did not relate these concepts to spoken


but, instead to spelling, reading, and writing.


Both Reid's


and Downing's findings that children's concepts of letter


word


sentence are vague and confused


were confirmed by this


study


Francis


felt that the children


s confusion was a reflection of the fact that


thP rnnr~ n "nrl" i nnt \,iarv wP11 r1 -Finnr1


Ilnl iL nnwn inn Prnnr c






32




The use of context in identifying word boundaries was explored


in a study by Klein, Klein


included both fourth and


, and Bertino (1974).

ixth grade students.


Their subjects

Subjects were asked to


perform two tasks for determining word boundary


information and one without it.


es--one with contextual


The results of the word boundaries in


context task indicated a developmental trend from fourth grade to

sixth grade in the ability to utilize contextual information.

The question of whether or not context-free words are easier to


identify than context-dependent words was also


asked in a 1977 study


Ehri (1979) trained students to discriminate real from nonsense words


using a situation resembling a game.


The Mr. Silly-Mr


Sense game was


constructed for this task.


children were instructed to point to


Mr. Silly for nonsense words and to Mr. Sense for real words.


students received pretraining on this task.


graphed


When the results were


a bimodal distribution of word discrimination scores appeared,


instead of

preted the


the


expected normal distribution.


The researchers inter-


graph to mean that the ability to distinguish word


non-words is not


a sing


from


kill as was previously thought; rather it is


a composite of


skills.


Mason (1980) also studied preschool


in an attempt to


determine whether or not a hierarchy existed in the development of


word meaning.


From her sample


of four-year-olds,


she explored their


comnPtence with 1 Pttpr and wnrrc nvur nin-mnnnh nrni nrl [Qc













with trying to read signs.


At the second level


, visual recognition,


the children were trying to use


better names or sounds to guess at


common three-letter words


as well


as some uncommon ones.


At the


highest


and utili


, letter sound analysis


the children tried to organize


etter-sound knowledge to identify word


The kinds of


strategic


children used in attempting the tasks were also noted


by the researcher.


In addition


, parents of the children were asked to


complete questionnaires regarding their child


interest in letters


and word


as well


as their own involvement in helping the


child learn


to read


Although the samp


was homogeneous with regard to home back-


ground and age


, the children did not progress at the same rate.


One of the weakness


of J


Mason


(1980)


study


according


to Dewitz and Stammer


980)


was the use of parent survey


forms


method of data collection.


specifically,


In order to resolve the question of what


the children were responding to when reading signs and


logos,


Dewitz and Stammer designed a study of their own.


They sought


to di


cover whether children were responding to the print itself, the


visual

logo.


characteristics of the logo


or the total context of the sign


Using a sample of 36 middle- and lower-cl


representing three-


their subject


four-


children, equally


, and five-year-olds, the researchers tested


'recognition of 16 common logo


including McDonald


Kelloggs, and


Stop,


n one of four contexts.


Each subject saw all of


fho lnnnc


hnii/QWQV^ i nn 1 rnn 1 c13 cnnn 4 n mn,-n +knn nnn rnn iy +nJ+ in


Ylh







34





In addition to identifying the logos under one of the four contexts,

subjects were also asked to write their name, to identify upper- and


lowercase letters, and to read


common nouns.


children grow older, the results indicated


, their ability


to identify logos in their natural context,


as well


as in isolation,


increases.


All age groups had low performance ratings when presented


with logos in a misleading context


ogos written in standard manusc


while none were able to identify

ript. In this last context, older


children more frequently recognized that the logo was a word, although


they were unable


to read it.


In their discussion of results, Dewitz and Stammer stated


that reading skill appeared to develop along parallel lines.


identified the first of these


They


as being mainly context dependent, such


as recognition of a stop sign by its shape and color, rather than by


the print alone.


The second


ine appeared to be a combination of logo


reading and letter recognition. In

subjects were context dependent and


their research sample, the younger

still unable to identify words or


letters.


The older subj


ects


appeared to be


learning this skill


evidenced by their ability to identify most of the letters presented


as well as several of the words.


fired J. Mason's (1


Thus


980) developmental sequence


, their results partially con-


Hiebert, 1978; Doehring


and Aull


1979)


DccrIIlfriiik +c ~ CL i ihrrk nrnn-4t tA r4-r A 4-. rt.nPr 4- an 4-..4 r n, ,4 t4- :







35





children's print-related knowledge in terms of concepts and skills.


By using a variety of concrete tasks


as well


as traditional reading


readiness measures


, the differences among three-


four-


and five-


year-old preschoolers were explored.


Although there were differences


between age groups, particularly three- and five-year-olds


within


individuals


each age group varied in their proficiency in dealing with print.


The data suggested that there was an


interrelationship between skills


and concepts of print awareness in young children; that print awareness


was a unified process


reading level


rather than a sequence


of skills at the pre-


No fixed order in the acquisition of


and con-


cepts was substantiated b


the data.


Hiebert offered two explanations


for the differences between


her finding


and those of previous


Weintraub, 1963; G. Mason


studies (Reid


, 1967; Oliver,


975)


, 1966; Denny and

The primary difference


was in the tasks employed in the


studio


In Hiebert's study, concrete


objects


such as books


and a cardboard town


, were present when ques-


tioning of her subjects about the uses and function of print took


ace.


In contrast, during her


interviews Reid (1966


used no print


stimulus at all, while


textualized setting."


Denny and Weintraub used print in a "decon-

Secondly, Hiebert indicated that there were


differences in the research


Hiebert's sample


sample


in terms of time and place.


was composed of predominantly


urban, middle class


children


Hpr CamnlP p inlik

& k n C


N


Smay havp rpfloprifped thp


Il" III llrl













availability of children's books in recent decades.


In addition,


her sample shared


common nursery school experience.


Denny and


Weintraub's sample was composed of a different generation as

a different socioeconomic state, including rural, all-Negro


well as


lower-


and middle-cl


subjects (G. Mason, 1967


Oliver, 1975)


Also their


sample would not hav


had the benefits of the


ming or the abundance of quality children


difference


ame television program-


s books.


was that of the countries from which the sampi


Another important


were


drawn.


Reid's sampi


of Scottish


children may have been exposed to


cultural influences which affected their awareness of print (Oliver,


1975; Downing, Ollilia, and Oliver


1976


American Indian children in the latter


Thus the Scottish and


studies may not have


shared


common experiences with the samp


of Denny and Weintraub and of


Hiebert.

Although researchers investigating the development of print

awareness in young children did not agree upon the extent of the


awareness or the pro


cesses


by which it i


developed, there are several


questions which need further investigation.


helped to develop an awareness of print


How can children be

cally and naturally?


What experiences would be the most beneficial in developing thi


ability


How does the nature of the tasks


children are asked to per-


form in pres


school programs affect their acquisition of the concepts of


II ~ ,r "1r n* +o~n l I1 ff ~I


C'n ~nf Y k~ri++'?rl +I,e~rI be~n irA am ~Ar Lwna A.. raIJ


'F













literature

awareness?


, and television programs to facilitate development of print

What provisions need to be made for the disadvantaged or


culturally different child?

by the aforementioned studies


These questions which have been raised

warrant the time and effort required


to explore their answers.


Print Awareness and Achievement


How do

achievement?


es


an awareness of print relate to future reading


In several studies the nature of this relationship has


been explored


The effectiveness of print awareness


as a predictor of


reading achievement has been


studio


investigated, especially in several recent


The degree of relationship between the two factors may facili-


tate the acceptance of print awareness


as a better indicator of reading


read


ness


and reading achievement then traditional reading readiness


tests.


Do readers


earn to read better because of their awareness of


print or does the reading act increase their lexical awareness?


Ehri


conducted two experiments


in an attempt to answer thi


question


investigation of the relationship between learning to read and lexical


awareness was conducted b


Ehri (1975).


Prereaders' performance was


compared to that of readers on tasks such


as sentence


segmentation


and sentence comparison--identification of the word which differed in


two almost identical sentences.


Her results indicated that the first






38




The results of the first study were not considered conclusive


because the age variable


was confounded with the reading variable.


The readers may have demonstrated better performance because they were

older than the prereaders and had had more linguistic experience; thus,


the second study was undertaken.


experiment conducted in 1976


was designed to eliminate the possibility that factors associated with


age might have


been responsible


prereaders (Ehri, 1979).


In thi


for the differences between readers and


study kindergarten readers and pre-


readers were matched in age and included in the sample, along with


a group of first graders.


Ehri


hypothesis


was that the readers


should be abl


to learn words faster than prereaders due to the readers'


superior word consciousness


subjects were given five


orally


pronounced words to associate with five visual nonsense figures


as a


paired


assoc


iate learning task.


Her results indicated that the hypothe-


was valid.


Learning to read


not age, was


responsible


for the


development of word consciousness


The design of this study was an


improvement over her earlier study in that it did not control for the age

variable.


Unlike Ehri who used white


middle-cl


subjects


Ryan,


McNamara, and Kenney (19


sample in one of their experiments.


) included a predominantly black and ma


It was the intent of the re-


searchers to study the relationships between certain aspects of


linguistic awareness and early reading ability.


In their first


ex-













readers,


as evidenced by an above-median


score


on the Slosson Oral


Reading Test, performed better than the below average readers on the

word discriminating and multigrammatical function word tasks. No sig-

nificant differences were found on the sentence comparison task. The


word tapping task did not favor the better readers.


In their second


experiment, Ryan, McNamara


and Kenney (1977


varied both the tasks and their sample


From third and fourth grade


asses,


a sample of


children was drawn, all of whom were receiving


remedial reading instruction.

basal reader placement, as ha

first grade to third grade.


The children were identified by their


ving reading levels ranging from high


Of the sample


were black and 17 were


male.


Two of the same tasks


from th


first


experiment, word tapping


and word discrimination


tion


, were administered to the subjects.


the children were given two written tasks


In addi-


The first of these


was a sentence segmentation task in which the subjects were presented


stimulus sentences without spaces between words


, then asked to divide


the sentences into words.

paragraphs with deletions


A written cloze task


, consisting of two


was the second written task.


The results


indicated that the children with reading


superior to those be


above the median were


ow the median on word segmentation, written seg-


mentation, and written cloze.


in the first study


, word tapping did


not differentiate the better readers from the poorer ones.


failure


nt~~~~~~~h- --I- C~:U -l----l-L





1













One aspect of linguistic awareness


use of intraword patterns


to compare words, was investigated by Pick, Unze


Brownell


, Drozdal,


and Hopmann (1


978)


Children from first, third, and fifth grade were


compared on their ability to differentiate which pair of words, from


the two pairs that were shown, were the most alike.


of either three or four letters, had the


The words, composed


same beginning, same ending, or


same ending


which rhymed.


The same pairs of words were then presented


orall


and the


child


asked to decid


upon their similarity or difference.


Their results indicated that the first graders, unlike the older sub-


jects


used the beginnings of the words they read in deciding upon


similarity or difference.


However, when the first graders heard the


words


, they based their judgments on the end of the word.


The older


children based their judgment of word


words more often than nonrhyming words.


milarity on the ends of rhyming

Their results and conclusions


tended to support previous findings that beginning readers tend to


rely on first letters of words in visual recognition tasks


and Levin


Does


ment?


Marchbanks


, 1965).


knowledge about print predict subsequent reading achieve-


An investigation aimed at answering this question was conducted


by Evans, Taylor


and Blum


979).


Their sample included middle and


lower class first grad


children.


A battery of seven tasks


measuring


knowledge about the


I


r9 I


written language code was administered to the


I. 1 .. .I? 1


"Il n n V. vIn I~~n fl Ct n t'n1 In .% nfl .n r. r *~.t r7t "C C ~v ~ '" C.-.. ,., C.m .,l, ~ ." n r~ C'v


n~













assroom teachers.


In May, the comprehension section of Metropolitan


Achievement T


ests,


Primary I, Form F, was administered by the researchers.


The coefficients of correlation between the written language awareness

battery and the Metropolitan Readiness Tests ranged from 0.46 to 0.85.

According to the investigators, this indicated that both measures were


tapping some of the same domain of reading readiness


skills.


battery equaled the Metropolitan Readines


s Tests


in predicting scores


on the comprehension subtest of the Metropolitan Achievement Test,


with r


Several conclusions based on these results were worth noting.


First, the researchers found that those


children who were more


success-


ful on


inguistic awareness tasks in September were better readers at


the end of the year


Thus, these tasks may be representative of the


skills necessary to learn to read.


Secondly


, the written language


battery was proposed


as a more useful instrument diagnostically than


the Metropolitan Readiness T


ests


(MRT)


The researchers reasoned that


the skill


of the battery tapped the reading acquisition tasks better


than the MRT.


ince validation of Clay


Concepts about Print Test (CAP)


Sand, was needed in the United States


Day and Day (1979a) conducted a


correlation


study of the CAP, C1


ay's


Record of Oral Language (ROL)


modified for American usage and the Metropolitan Readiness T


ests.












high correlation between the CAP


, the ROL, and the Metropolitan


Readiness T


ests


which was administered to the children during the


first weeks


of first grade.


A follow-up


study of these children at the


end of first grade was done to determine whether or not the linguistic

awareness measured by the CAP predicted reading achievement as


measured by the


Iowa Tests of Basi


Skill


(Day


Spicola, and


Griffin


, 1979)


There was a high correlation among the separate


administration


from 0.6


of the CAP and the test of reading achievement,


for the first administration of the CAP to 0


ranging


.77 for the fourth.


A further analy


of their data


yielded


ome surprising results


Day and Day, 1980b).


Approximately


first grade with considerable


percent of the children entered


knowledge of book orientation


print


direction rules


, and letter and word concepts.


Conversely, 20 to


per-


cent of the children were still


confused about print-direction and


etter-word concepts and entered first grade reading instruction without


having grasp of them.


also noteworthy that there was a high


correlation between the initial score on the CAP at the beginning of

kindergarten and the reading achievement score at the end of first


grade.


Perhaps th


underscores the importance of an adequate lin-


guistic conceptua


base


to reading achievement.


In this sample, girls


experienced a


lightly earlier development of orthographi


c linguistic


awareness than did the boys (Day and Day,


979b).


In an


expanded study


. Day


Dav. Hnllinanwnrth. and Mrclplland


a











scores of these samples.


In all three samp


the researcher found


that girls consistent


answered more items correctly than boys.


One of the latest studies of children's knowledge of the


reading


process and print convention was done by Johns (1980).


From a


population of primarily white, middle


and upper-income first graders,


60 children were


selected.


Above average


, average, and below average


readers were equally


represented.


At the end of first grade, each


child was individually given the Concepts about Print Test, Sand.


scores


indicated that the above-average readers


' concepts about print


were significantly


higher than those for average and below average


readers.


The above-average readers were


superior in print-direction,


etter-word and advanced print concepts.


It was stated that differences


in reading ability were reflected in the


scores on the Sand.


One of


the weakness


of this study was the u


of two different instruments


to measure the reading achievement of the subjects--the Metropolitan


Achievement T


ests,


Primary I, and the Gates-MacGinitie Reading Tests,


Primary A.


Averaging the grade equivalent scores for each group for


comparison was not a statistically sound procedure, given that each

test was based on a different norming sample and had different items,


format, and


scale.


From the above studies, it appeared that there was a relation-


ship between print awareness and subsequent reading achievement.


ever,


How-


the strength of that relationship varied with the methodology of


~~LL.~t .LJ1 -,.. --.4--


-L.












(1977) study

specified as


were not found in any of the other studies and were not


tasks measuring print awareness per


Self-Correction Behavior


As the child becomes aware of the spaces between words


begins to realize that there should be a match between the number of


word


he pronounces and that of the text he is reading


This realiza-


tion may be one of the first steps toward becoming a


reader


elf-monitoring


His subsequent error-correction behavior may pass through


several


stages before he is able to


self-correct individual words


Within the


following section, the relationship between self-correction


behavior and reading progress is discussed.


Evidence of


f-correction behavior was noted by McKinnon


959) in his study of beginning reading


He noted that, initially,


the control group subjects who read individually made fewer errors


and had more


self-corrections than the


experimental subjects who read


the same text in a group


situation.


However,


as time passed, th


experimental group decreased the number of errors and increased their


self-correction rate appreciable

experimental group scored sign

During the testing, the children


and receive credit for their corrections.


In the follow-up testing, the


ficantly better than the control group.


in were allowed to correct their errors


As a result, the experimen-


tal group was shown to have


a greater tendency


to self-correct than






45





number of times a student self-corrected his own errors was noted.

From the results of this study, it was found that successful readers


develop a way of learning from their errors.


The child who learned


to search and check and to pay attention to more detail in print


did become progressive


better at self-correction.


indicated


that the students who learned to monitor their own responses to print

would continue to improve not only in accuracy but in detecting and

correcting their own errors.


Each of the progress group


in the study--High


High-Middle,


Low-Middle, and Low--was


compared by plotting the relationship of


their error rates to self-correction rates.


to have a low-error


The High Group was found


, high-self-correction ratio, believed to be the


result of efficient processing cues.

high-error, low-self-correction ratio


Conversely, the Low Group had a

and fewer responses than the


next higher group


. A self-


correction rate of one


n four errors was


considered good.


Clay, 1979b)


if it did not meet this idea


This was later amended to a rate of between 1


It was noted that self-correction behavior, even


ratio, was considered a sign of progress.


If there were no


evidence of


self-correction, it was interpreted


as a


danger sign that perhaps the child was unaware of the cues he should


be processing or did not know how to use those cues.


It was presumed


by the


resear


cher that the


children were correctly placed in their


basals


according to each child'


instructional reading level


One


One






46





At the prereading stage, Clay found that the child passed


through five


(iii
(iv


stages in his error-correction behavior:


From page matching, in which the child repeated
a memorized text for the page without locating
any detail in the print;
To line matching in which the child repeated a
memorized line of print, locating that line as a
whole;
To locating some words within a memorized line;
To "reading the spaces" and thus coordinating
visually located word patterns, speech impulses
and the spaces between words
This led to movement-speech mismatch when there
were too few or too many spoken impulses for the
number of patterns available, or a speech-vision
mismatch when a word failed to coincide with its
known visual pattern during the co-ordinating


process.


(Clay, 1969


Thus, error-correction behavior began much earlier than actual book

reading.

When all four groups in the study reached the book reading stage,

evidence of their progress was reported by their self-correction rates.


For the High group, the ratio was


High-Middle,


'is'-,


the Low-


Middle


and the


Low, 19


Thus, the two highest groups corrected


approximately


11 to 14 percent of their errors, while


the two lower


groups corrected


to 35 percent


In a study completed in 1964


Elder (1971


also noted the


presence or absence of self-correction behavior.


His study compared


the oral reading achievement of Scottish second graders to American


second and third graders.


His results indicated that the American


LI' A -






47




pronouncing words than the Americans, had a lower percentage of self-


correction.


Since this was primarily a one-shot study, it would have


been more valuable to study the changes over time in the self-


correction rates of both groups


Also, the use of standardized oral


reading paragraphs does not indicate whether the ratios would be the

same for more familiar material or text resembling that found in the


basal series being read by each group.


There was no differentiation in


number of self-corrections occurring at each subject


instruc-


tional


evel in the discussion of this study


s results.


Hood and Kendall (1975) investigated the differences between


reflective and impulsive


second graders in the


number and category


of oral reading errors and their self-correction.


The sample


for the


study included


25 children of each cogniti


style, impulsive


ective.


No differences were found in the proportions of contex-


tually appropriate errors, but differences were found in the proportion


of self-corrections.


Impulsive subjects corrected proportionately


fewer errors that were appropriate to the preceding, but not to the


following context.


However, it was found that over-error correction


by the refl


ectiv


group did not improve


their comprehension


instrument used to designate


reflective and the impulsive groups


should be further examined to determine its effectiveness in measuring

this trait.

A longitudinal study of two first grade classes was conducted






48





non-instructional material from a tradebook series with approximately


ame vocabulary as the basal.


Tape recordings were made of each


individual while


oral reading took place.


Self-correcting behavior,


which had been practically non-existent in November, increased


month by month


teadily


For the good readers, self-correction on instructional


material ranged from 0.005 in November to a high of 0.1 in May,

followed by 0.17 in June; on non-instructional materials, the November


rate was 0, and the June rate 0.30.


In comparison, th


poor readers'


performance on instructional material


in June; on non-instructional materials


the June rate 0.02


ranged from 0 in November to 0.04

, the November rate was 0 and


Thus, it was evident that the good readers


steadily increased their self-correction rate; while


there was i ttle


evidence that the poor read


were actively engaged in thi


behavior.


It was interesting to note that the good readers'


correction rate was better on the non-instructionaj material than on


the basal reader material.


readabil

given.


it


However, no information concerning the


the selections on non-instructional material


was


They may or may not have been of equal difficulty to those of


the basal reader.


Cohen proposed that the "increase in


self-corrections


reflects a growing ability to selectively sample


better arrangements.


More of th


errors


are made by Good Readers who appear more capab


success


explorations of a word rather than remaining with a first


decision" (


Cohen, 1975, pp


. 646-47).


Thus, it appeared that


self-


self-






49





Is there a developmental trend for self-correction behavior


across grade level


One study which


explored the answer to this ques-


tion was that completed by Recht (1976)


A sample of 47 children in


grades two, three, four, and si


was chosen.


Equal numbers of high,


medium, and low abilities were included in the group from each grade


evel.


There was also an equal number of males and females.


After


orally reading a


0-word passage


each subject was given a


loze test


on the passage.


scores.


If-correction rates were compared to comprehension


A positive relationship between comprehension and self-


correction rate was found.


In addition


, the rate of self-correction


was found to increase with the


ability, comprehension, and grade level


of the reader.


However, the


increase in self-correction rate was found


to level off at the sixth grade level


the plateau may be


According to the researcher,


explained by the decreased opportunity for oral


reading at the


sixth grade


of the trend was that, at thi


Another explanation for the


evel


, the reader may have been


evening


self-


correcting silently rather than orally


similarly, Clay (1979a) found


self-correction behavior a more useful sign of progress within the


first three years of instruction.


Later, it appeared that readers


solved their problems silently rather than orally.

Recently, Kendall and Hood (1979) also investigated the

relationship between self-correction rates and comprehension for two


types of disabled readers.


Subi


ects


for thi


study were selected from


1 '


"






50




identified as High Comprehension-Low Word Recognition (HiC-LoWR) sub-


jects


On the Iowa Tests of Basic


kills, their comprehension


scores


were in the highest 75 percent


while word recognition was in the lowest


percent.


The other half of the samp


was identified


as Low Com-


prehension-High Word Recognition (LoC-HiWR)


Their scores were in the


lowest


percent in comprehension and the highest


recognition.


In the HiC-LoWR group were 1


percent in word


subjects; the LoC-HiWR group


consisted of


subjects.


(Each group lost


seven subjects as a result


of taping problems during oral reading data collection.)


Each student read two


selections--one of high third grade


readability


words in


ength), and one of fifth grade readabability


words)


based on Botel


criteria.


After reading each


selection,


subjects were asked ten literal questions .


Errors,


self-corrections,


and comprehension


scores


were averaged for each gorup


When the


data


were analyzed according to passage difficulty, the HiC-LoWR subjects


corrected approximately the same proportion of errors


group.


as the LoC-HiWR


Furthermore, the HiC-LoWR group had better comprehension than


the LoC-HiWR group, in spite of their having made twice


as many word


recognition errors


as the latter group


superiority of their


comprehension


scores


demonstrated that the HiC-LoWR group was actively


processing and remembering the text, although their error rate would

have indicated that they were having difficulty with individual words


Do learning disabled readers self-correct


as efficiently






51





explore the similarities and differences in the extent and kind of

self-correction employed by learning-disabled and normal students.


In addition


, differences between inner-city and affluent suburban


backgrounds were explored


Learning disabled and normal students were


selected from both


socioeconomic backgrounds.


All 50 inner-city


participants were black; therefore, to remove scoring biases, all Black


English pronunciations were removed from the analysis.


The inner-city


and suburban groups were similar


, in terms of age, grade, and


sex.


suburban participants were white


leve


It was found that the achievement


of the learning disabled white suburban student was similar to


that of the black


nner-city norma


student.


To control for performance level differences among the groups,


the test passage was rewritten on eight readability levels


to the Spache and Dale-Chall criteria.


according


Each version contained the


same information and ranged from


evel which


to 348 words


would generate an error rate of


assign a story


to 10 percent, a three-


step procedure was used.


Errors were scored by three experimenters


the passage was being read, then later compared to an audio tape. The

data revealed that the LD children made more errors than the non-LD

children, even though reading level had been controlled.

Self-correction behavior was examined in a variety of ways.

In frequency, the LD readers made more self-corrections than the normal


readers.


When the self-corrections were examined proportionately






52




instructional reading level, the "LD readers were found to correct


these


significantly les


effectively than the non-LD readers"


(Pflaum and Bryan, 1980


p. 256).


Finally


both the LD readers and


the inner-city subjects made more multipi


attempts to correct them-


selv


than did the suburban normal readers.


children had


ower reading achievement overall


Although the urban


there were few differ-


ences in oral reading strategic


From the limited number of studies in thi


area, it appeared


that self-correction behavior has been


supported


as a sign of reading


progress, especially for primary-aged children


kindergarten through


third grade), and for disabled readers.


Valuab


information about


the reading process and the child'


interaction with print can be


gained from an analysis of his self-correction behavior.


Within the


studies previously cited, there were several differences in methodology


worth noting


First, there was one study in which the material was


familiar and in which preteaching appeared to have taken place, prior


to oral reading (Cohen, 1975)


would have depressed the error


rate


since the


students would have had previous exposure to the text.


In other studies (Hood and Kendall


, 1975,; Pflaum and Bryan, 1980


Recht,


1976


passages were unfamiliar to the subjects; thus the error


rate may have been inflated.


behavior were diagnosis


If th


then th


purpose of obtaining self-correction


latter procedure would be advisabi


since it approximates that of the


informal reading inventory.


A more






53





Summary


Within the studies cited, the credibility of print awareness


as a predictor of reading achievement has been discussed.


Across the


studies, it was evident that each researcher varied widely in method-


ology


sampling


and task requirements.


Some studies involved small


sample


for short periods of time, while others employed larger sampi


in longitudinal research.


In the short-term studies, it was learned


that children have obvious misconceptions about the reading process.


From the


longitudinal studies, it was learned, for example, that


correction rates were better indicators of progress for primary and


disabled readers than for older, intermediate grade readers.


Also,


self-correction behavior and print awareness appeared to change over


time at varying rates, according to the child


s reading ability level


Thus, different information has been gained from each type of study.

Differences in task requirements clearly affected the conclu-


sions drawn by the researchers.


In some studies


, when young children


were asked to explain abstract concepts such


as "letter" and "word,"


their explanations were usually inadequate or incorrect.


when the same questions were


However,


asked and the children given a concrete


object upon which to perform the task, or demonstrate the concept, they


were more successful


print?


Which task actually measured their awareness of


The answer depends upon your own orientation.
















CHAPTER III
COLLECTION AND TREATMENT OF DATA


The discussions in Chapters I and II indicated a need to examine

the degree of predictability of different linguistic awareness tasks

and instruments for primary grade children in an effort to determine

which one of these indicators offers the best explanation of the strate-


gies needed to understand the reading process.


The purpose of this


study was to determine the degree and predictability of two indicators,


performance on th


reading sample


Sand and the average self-correction rate on oral


for reading achievement.


The Sample


All participants in the study were selected from one Volusia


County elementary


children.


choo


With a tota


whose population consisted of lower and middle


enrollment of 527, the school was con-


sidered average in size (the district mean for elementary


school


525)


For the purposes of the study,


25 subjects from primary grade


classrooms were


chosen.


Since the Concepts about Print Test, Sand, was originally

intended for five- and six-year-olds, the inclusion of subjects from


kindergarten and first grade was justified


Their performance on the






55




classrooms--two classrooms each from kindergarten and first grade.

The use of classroom groups provided a balance in that sample of race,

sex, and ability level since the students were administratively


assigned to heterogeneous classrooms.


All Titl


I subjects were


selected on the basis of their


scores


on the Comprehensive T


ests


of Basi


kills (April 1980 administration)


The uniform Volusia County Titl


I selection criterion applied to the


sample


as well


as to students in


eligible public


elementary schools


as well


as participating parochial school


The criterion was designated


by a score at or below the 40th perc

subtest or the total reading test.


e, on either the comprehension


Within the five second and third


grade classrooms, all students who met this criterion became Titi


target


students and were included in the study


For the purposes of


the study


, the selection criterion for all subjects was the ability


to respond


verbally


or by pointing, to items on the Sand.


No subjects


were


excluded from the study on thi


basi


As the school year progressed, there were some changes in the


original sample.


children


As a result of


the kindergarten sample continued to include


five of the 34 first grade subjects moved during the year.


student transfers, the Title I sample became racially


unbalanced


, having slightly more black students than white.


From the


64 Title


subj


ects,


it was possible


to obtain oral reading samples


from all but fiv


subj


ects.


When the study was completed, the three






56





The Instruments


Four instruments were used to evaluate the performance of sub-


jects on several


variables.


They included the following


The Metropolitan Readiness Tests, Level I, Form Q,


and Level II


Form P


1976)


The Metropolitan Achievement Tests


Primary I,


Form H, and Primary II, Form H (1971)


The Comprehensive Tests of Basic


Skill


s, Level B,


Form


Level C


Form


and Level I


Form


(1974)


The Concepts about Print Test, Sand


1972)


Use of the Metropolitan Readiness Tests (MRT) had several justi-


fications.


Primarily, the MRT was


selected because it was a standard-


ized instrument, widely used with both kindergarten and first grade


students


, for the purpose of predicting subsequent success in prereading


and reading activity


high--0.9


plit-half reliability for both levels was


for Level I and 0.94 for Level II (prereading composite


score)


econdly


MRT was administered to the research sample dur-


ing district-wide testing periods.


with other schools.


research sample


would permit later comparisons


Third, the use of the MRT with two groups of the


would allow comparisons to be made both within and


across grade


evels.


Finally, the MRT had been used in previous re-


search which compared performance on the MRT with that on the Sand.


Day and Day


1979a) found correlations between the Metropolitan Readi-









57



Because of their district-wide use in the Title I program, the


Metropolitan Achievement Tests (MAT) were


selected as measures of reading


progress.


The MAT has a reported split-half reliability coefficient of


0.96 and a K-R 20 coefficient of 0.97


The MAT was administered to the


Title I subjects


as an out-of-level test.


This meant that each Title I


subject was given a test which wa


intended for the grade level below


his/her current grade placement.


Thus, the second grade


student


received Primary I, Form H, while the third grade Title I student was


given Primary II


Form H


. Use of the MAT allowed for comparisons to be


made within and across the grad


level


included in the research sample,


as wel


as with other Titl


I schools in the district


Another reason


for the use of the MAT was the use of thi


s instrument in a previous


study


which compared the reading scores with the performance on the Sand


1980)


Johns,


Finally, comparisons of the correlations between the Sand and the


MAT with that of the Sand with the CTBS were possible.


The Comprehensive Tests of Basic Skills (CTBS) were


selected


because of their district-wide use


as measures of reading achievement.


The CTBS has a reported K-R 20 coefficient of 0.96 and a validity coef-


ficient of 0.9


The CTBS


standard scores facilitated comparisons of


reading achievement between the first graders in the sample with the


Title


students.


only, performance on the CTBS had not been cor-


related with the Sand in the previous research


As a standardized, but informal, measure of children'


C',, ,.


print


rL. r..


~I~It=n t'nnr


e~, r,,.,, c,, c.,,, c n,:, c ThrC






58





been employed in research studies in the United States for the same pur-


poses


Day and Day, 1979a; Johns, 1980).


The Sand has a reported K-R 20


reliability coefficient of 0.95 and a validity coefficient of 0.79


instrument consisted of a 24-item checklist of questions asked while the


storybook, Sand, was being read to the child.


Among the concepts checked


were those of letter


word


, directionality in reading


and uses of punc-


tuation.


It was preferred over other techniques such as the interview


because it (the Sand) involved a natural


while tapping the subject'


situation and a concrete object,


knowledge of abstract concepts (Dewitz and


Stammer


1980


Hiebert,


978).


Self-Correction Rates


For Title I children


have been the best


therefore


, reading achievement test scores may not


indicator of reading achievement or performance;


, it was decided to gather longitudinal information about their


reading strategic


n the form of oral reading


tampl es.


Each Title I


subject


except the five previously


mentioned


, was tape recorded


individually as s/he read the last


selection in the basal reader in its


entirety


The only exception to using the final


selection was the primer


level book in which the second-to-last selection was read.


last selection was a play


(The


while all others were narrative.)


sections were read and recorded before the student had been exposed


to the new vocabulary of the selection or


introduced to its content.


~LII 2 J J I .. Ut I I t1 I I V






59





the Holt Basic Reading System were used because this series was con-

sistently used in most of Volusia County's Title I units.


From the oral reading samp


the self-correction rates were


of interest to this


study


It was intended that the self-correction


rates, in addition to the


of print awareness


scores


on the Sand, would estimate the level


of the Title I students.


Both indicators were


examined

together


to be reliable


(Clay, 1979a; Recht,


as predictors of reading achievement when used separately and


In previous studies, self-correction rates had been shown


indicators of reading progress in primary grade children


976).


Collection of Data


In September


, the Concepts about Print Test


Sand, was admin-


istered individually to each first, second, and third grade subject by


the researcher.


Testing took place in an empty classroom with only


the researcher and the


storybook, Sand, was


subject present.


At the beginning of the session,


introduced to the child.


It was explained


that the researcher would read the book and ask the child questions


throughout the story


Following the procedures outlined in Clay


Detection of Reading Difficulties, the 24-item checklist was


given.


Scoring for each item was completed


as the child responded.


Points were awarded on the basis of one point for a correct response


and zero for an incorrect or no response.


No partial credit for any


7. 4- "1 n -% -: .. .* *l -. in n ~ 4 .AI A Ar C h n L n a %


r












appropriateness of his/her response


In January, all kindergarten


subjects were tested by following the same procedure and testing con-


editions


In late March and April


all first grade and Title I second


and third grade subjects were retested following the same procedure.


For the first grade


subjects, the Metropolitan Readiness Tests,


Level II


, Form P, were administered by their classroom teachers.


battery was given during a four-day period during the second week of


school.


Subjects were tested in groups of ten to fifteen.


Results of


the battery were obtained from the classroom teachers after they com-


pleted the


scoring.


In January


, the kindergarten students were given


the MRT, Level I


Form Q, by their cl


assroom teacher who also scored


the tests


Like the first graders, the kindergarteners were given the


battery during a four-day testing period.


For the Titi


I subjects


, the appropriate form of the Metro-


politan Achievement Tests, Form H


and another Title I teacher.


was administered by the researcher


During the one-week fall testing period


designated by the Title I director, the subjects took the


test in groups


of five to


students.


conditioned, portable e


All tests were administered in air


classrooms with few distractions.


Testing took


place on two separate days with one subtest administered each day.


second grade


ects


received Primary I


while the third graders were


given Primary II. Both the researcher and the Title I teacher hand-

scored the test booklets. This Drocedure was rpnpatnd durinn thp Anril






61





All first, second, and third grade subjects, along with their


peers, were given the Comprehensive T


ests


of Basi


Skills


Form


their classroom teachers.


subjects were te


the district week-long testing period. The

received Level I were tested in late March.

graders were given Level B, while the second


listed in groups of 15 during

Third grade subjects who

In mid-April, the first


Id graders received Leve


All test booklets were machine-scored.


Test results for


each subject


were obtained from the school printout.


Treatment of Data


For all Titl


I subjects, the oral reading samples were scored


after the taping had been completed.


Omissions,


ubstitutions


includ-


ing mispronunciations)


unknown words (those pronounced by the examiner),


transpositions, and insertions were the only scorable errors.

repeated on any word were not counted after the second miscue.


Errors

The


self-correction rate for each passage was computed by dividing the


number of self-corrections by the number of errors.


An average


self-


correction rate


ASCR) was


calculated by adding the total number of


self-corrections for all passages read by each subject and dividing by


the total number of errors


across


all passages.


All scoring of errors


was done by the researcher.


Each subject's level of print awareness


, indicated by his/her


total raw


score


on the Sand, was entered.


Scores


on both the fall and


~ ~ ~ --


-






62





Metropolitan Readiness Tests was entered for comparison with the Sand.

From the reading subtests of the Comprehensive Tests of Basic Skills,

the total standard score for each first, second, and third grade sub-


ject was noted


as a measure of reading achievement.


In addition, each


Title


I subject's total reading standard score on the Metropolitan


Achievement T


ests


was entered.


Following the recording of the above data for each subject,


several statistical procedures were employed to test the 1


of the study


hypotheses


Pearson-product-moment correlations were calculated to


demonstrate the strength of the relationship between scores on the Sand


and the subject


age and grade level.


Correlations between the Sand


and the measures of reading achievement, the MAT and the CTBS, as well

as the measures of reading readiness, the MRT were also computed. For

the first, second, and third grade subjects, scores on the fall and

spring administrations of the Sand were correlated. Finally for the


Title


I subjects, Pearson-product-moment correlations were calculated


for the average self-correction rates (ASCR) and the measures of

reading achievement, the MAT and the CTBS.

To determine the relationship between scores on the Sand


dependent variable) and the subject


age (independent variable),


linear regression analysis


was used.


In order to


examine performance


differences on the Sand for the non-Titi


a one-way analysis


I and the Titi


of variance was first calculated


I students,


followed by a


.


w












were the variabi


used in the analysis of variance.


From the analysis


of variance data, one complex and three pairwise contrasts of group

means were calculated.


For the Title I subjects


scores


a multiple regression analysis


on the Sand and the average self-correction rates (indepen-


dent variables


variable) wa


across the standard


completed


scores


on the MAT (dependent


Multiple regression anal


was also used for


the Sand and the ASCR


independent variable


es) across the standard scores


on the CTBS (dependent variable).


Finally, computation of self-


correction rates across book 1


evel


the ideal range of 0.20 to 0


was completed for comparison with


according to Clay


1979b).


The hypotheses tested were

I. There are no differences in the level of print


awareness,


as measured by the total score on the


Concepts about Print Test


Sand, among the age


levels tested


There are no differences in the level of print


awareness,


as measured by the tota


score on the


Sand among the grade level


tested


There are no differences in the means of the Sand


between the non-Title I kindergarten and first


grade students and the Titi


I second and third


grade students.












Title


second grade students, (b) first grade


students and Title I third grade students,

(c) Title I second and Title I third grade students,


and (d


first grade students and Title I second


and third grade students


There i


no relationship between total score on the


fall-administered Sand and the total reading score

on the appropriate grade level test of the Comprehen-


sive Tests of Basic Skill


CTBS) for the non-Title I


first, Title I second, and Title I third grade

students.


There is no relationship between the total


score on


the Sand and the total score on the Metropolitan

Readiness Tests for the kindergarten and first grade

students.


There is no relationship between the total score


on the Sand and the total reading score on the

appropriate level of the Metropolitan Achievement


Tests (MAT) for the


Title I second and third grade


students


VIII.


There is no relationship between the average


self-


correction rate on the oral reading samples and


the total reading


score on the appropriate












oral reading samples


and tota


reading score on


the MAT for the Titl


I second and third grade


students.

There is no relationship between the total score on


the Sand and the average


self-correction rate (ASCR


on the oral reading samples for the Title I second

and third grade students.


There i


no relationship between the average


self-


correction rate on the oral reading samples and the


total reading


for the Titl


There i


score on the appropriate level CTBS

I second and third grade students.


no relationship between the total score


on the Sand, the average self-correction rate on


the oral reading samp


es, and the total reading


score on the CTBS for the Title I second and third

grade students.

Chapter IV explains the results of the tests of the hypotheses.

















CHAPTER IV
ANALYSIS OF THE DATA


This study was concerned with determining the effectiveness of


print awareness and self-correction rates


achievement.


as predictors of reading


Primary grade students were given the Concepts about Print


Test, Sand, by the researcher.


Second and third grade Title I


students'


oral reading was tape recorded at predetermined places in their basals.

The data were analyzed using Pearson-product-moment correlations,


ysis


of varian


ce, and both linear and multiple1


regression analysis,


available from the packaged Statistical Analysi


stem (SAS


The following questions were raised in the study


there a difference in the development of print


awareness


Among age levels


Among grade


evels


Between pre-readers and readers?


Is there a progression in the development of a


self-


monitoring reader across


Book


Grade


evels


evels


4-LaL1 4.,.,~:, 2, 1 .t .2- .


n,,, LL,






67





Findings Related to the Hypotheses


All hypotheses were tested at the 0.05 level of significance.


Hypothesis I:


There i


no significant relation-


ship between the level of linguistic awareness, as

measured by the total score on the Concepts about


Print Test


, Sand, among any of the age level


tested.


A simple linear regression of Sand scores on age was computed.

The estimated regression slope was computed to equal 0.305 with a


standard error of 0.026.


The R2


was equal to 0.


5 meaning that 52.5


percent of the variability in Sand was shared with age.


The computed


F statistic for the overall regression equation equaled 129.6 with a


probability of 0.0001, given no relationship


exist


between Sand and age.


since the probability of the computed F statistic was


than the


0.05, the chosen criterion for statistical significance, the null

hypothesis was rejected and it was concluded that a significant linear


relationship exists between scores on the Sand and age.

ship is positive meaning that scores increase with age.


The relation-


Hypothes i


There i


no difference in the


average level of linguistic awareness, as

measured by the total score in the Concepts about


Print Test


, Sand, among any of the grade


evel


tested.


The means for each group on the Sand were computed


For the


For the












deviation of


5.23.


For the


second graders, the fall mean was 16.66


with a standard deviation of


3.27.


The fal


mean for the


third


graders was 20.00 with a standard deviation of 3.18.



TABLE 1
MEANS FOR THE FIRST ADMINISTRATION
OF THE SAND BY GRADE


Standard
Grade N Mean Deviation

Kdgn 32 7.60 3.61

First 34 8.94 5.23

Second
Title I 27 16.66 3.27

Third
Title I 25 20.00 3.18


The group means were


compared using the


one-way analysis of


variance strategy (see Table 1)


The computed F statistic equaled 63.40


which had the probability of 0.0001, given the population means were


equal


since the probability


of the computed test statistic was less


than the 0.05 level set


as the criterion for statistics


significance,


it was


concluded that the average


scores


on the Sand for the


students in


the grade


studied were not equal


The scores increased both by


the age and the grade


evel of the subject.






69




total score on the fall Sand, between the non-

Title I kindergarten and the first graders, and

the Title I second and third graders.


Hypothes i


III was tested by constructing a confidence interval


for the difference in the average means of the Title I


average means of the non-Title I students.


students with the


The Bonferroni procedure,


which controls the error rate per family, was used in constructing the


confidence interval


between Titl


The value for the


estimate of the difference


I and non-Title I subjects equaled -10.03


A 95 percent


confidence interva


for the population parameter ranged between -10.77


and -9.29.


Since 0 was not included


n the interval


, the null hypothesis


was rejected.


There was a difference in the level of linguistic awareness


between the prereaders, the non-Title I kindergarten and first grade

students, and the readers, the Title I second and third graders. Th


average level of


inguisti


awareness was greater for the readers.


Hypothesi


There is no difference in the


evel of linguistic awareness


gained


as measured


by the tota


score on the Concepts about Print


Test, Sand, between the fall and


spring admin-


istrations for the (a) first grade students and


Title I second grade


students


(b) first grade


students and Title I third grade students;


(c) the Title I


second and Title I third orade






70





A one-way analysis of variance was first computed for the gain


scores


to test the hypothesis


that al


grades gained equally over the


school year.


The mean gain for the first graders on the total


score on


the Sand was 9.07 with a


standard deviation of 3.95; for the Titl


second graders, the average gain from the fall to the spring was


with a standard deviation of


students, the mean gain was


and for the third grade Title I


.21 with a standard deviation of 2.10


TABLE


AVERAGE GAIN IN TOTAL SCORE ON SAND
FROM FALL TO SPRING ADMINISTRATION


Mean Standard
Grade N Gain Deviation

First 29 9.07 3.95

Second
Title I 24 2.83 2.62

Third
Title I 24 2.21 2.10


Using the mean gain


scores


each group reported in Tabl


the overall computed F statistic was


given that the groups had gained equally


64 with a probability of 0.001,


since the probability of the


test


tatisti


was


than the 0.05 leve


set as the criterion for


statistical


significance, the null hypothesi


was rej


ected.


It was con-











computed using the Bonferroni procedure.


hypothesis


To answer part (a) of the


, the value for the estimate of the difference between grade


means, of the first grade and Title I second graders equaled 5.68.


95 percent confidence


interval for the population parameter ranged


between 4.87 and 6.49


Since 0 was not included in the interval


there


was a difference between the gain scores of the first graders and the


Title I second graders.


The mean of the first grade gain scores was


significantly higher than that of the Title I


second graders


In comparing the first graders with the Title I third graders


[part


a confidence


interval was also constructed.


The value for


the estimate of the difference between grade means equaled 6.35.

95 percent confidence interval for the population parameter ranged


between 5.54 and 7.16.

was a difference betwe


Since 0 was not included in the interval, there


;en the gain scores of the first graders and the


Title I third graders. The fi

significantly higher than those


rst graders' gain


of the Titl


scores on the Sand were


I third graders


When the second and third grade Title I students were compared


[part (c)], a confidence


interva


was also constructed.


The value for


the estimate of the differences between grade means equaled 0.67.

95 percent confidence interval for the population parameter ranged


between


-0.18 and 1.5.


Since 0 was included in thi


interval, it was


concluded that there was no significant difference in gain scores


between the two grade levels of Title


students.










grade means equaled 6.02.


A 95 percent confidence


interval for the


population parameter ranged between


5.33


and 6.71


Since 0 was not


included in the interval, there was a difference in the gain


scores of


the two groups.


It wa


found that the first graders had significantly


higher gain scores on the Sand than the Title I


second and third graders.


Hypothesi


the level of


There i


inguisti


no relationship between

awareness, as measured by


the fall total


score on the Concepts about Print


Test, Sand, and the level of reading achievement as

measured by the total score on the Comprehensive


Tests of Basi


kills for the first graders and


the Title I second and third graders.

To test this hypothesis, a Pearson-product moment correlation


for each grade was calculated.


When the Pearson-product-moment correla-


tion was computed for the first graders


, the coefficient was 0.803


with


a p value of 0.0001


; for the second grade Title I students, it was 0.570


with a p value of 0.045; and for the third grade Title I students, it


was 0.797 with a p value of 0.0001.


ince the probability of each


coefficient for all grade level


than the 0.05


set as the


criterion for statistical significance, the overall null hypothesis


was rejected


It was concluded that the Sand had the highest correla-


tion with the CTBS, Level B, for the first grade, followed by CTBS,


Level I, for the third grade


with CTBS, Level C


, for the second grade,


having the


owest coefficient


. However,


in al


-,J- -


cases,


the Sa nd corre-












TABLE


CORRELATION OF THE FALL SAND WITH THE SPRING
SCORE ON THE COMPREHENSIVE TESTS
OF BASIC SKILLS


Grade Test r p

First CTBS/B 0.803 0.0001

Second
Title I CTBS/C 0.570 0.0450

Third
Title I CTBS/I 0.797 0.0001


Hypothesi


There is no relationship between


the level of linguistic awareness, as measured by

the total score on the Concepts about Print Test,

Sand, and the pre-reading composite score on the

Metropolitan Readiness Tests for kindergarten


and first grade


students.


The Pearson-product-moment correlation was computed


this hypothesis


Level I


to test


The correlation coefficient for the Sand and the MRT


, Form Q (kindergarten) was 0.508 with a p value of 0.003, while


for the MRT, Level II


Form P


first grade), it was 0.767 with a p value


of 0.0001


Thus,


it appeared that the Sand had a higher correlation


with the first grade test than for the kindergarten level

correlation of the Sand with the MRT (both levels) was 0.


The overall


616 with a p












strong, positive relationship between the leve


of print awareness


as measured by the Sand


score


and the level of reading readiness


measured by the Metropolitan Readiness T


ests.


Hypothes i


There is no relationship between


evel of linguistic awareness,


as measured by


the total


score


on the Concepts about Print Test.


Sand, and the tota


reading score on the Metropoli-


tan Achievement T


grade Ti tl


for the second and third


I students.


To test thi

was calculated. Co


hypothesis


a Pearson-product-moment correlation


irrelations for both the


tions of the Sand for each grade


as wel


fall and spring administra-

as the entire Title I sample,


were computed.


For both Titl


I groups, the correlation between th


fall

When


scores

scores


on the Sand with the MAT was 0.660 with a p value of 0.0001.

on the fall Sand were correlated with those of the MAT for


each grade level


, the resulting coefficients were 0.612 with a p value


of 0.0025 for the


second grade Primary I test, and 0


of 0.0103 for the third grade Primary II test.


When


.523


with a p value


scores


for the


spring administration of the Sand were compared with the MAT, the fol-


lowing coefficients were calculated:


for both Titl


groups


0.683


with a p value of 0.0001; for


second grade Primary I


0.515 with a p


value


of 0.0042


; and for third grade Primary II, 0


.672


with a p value


of 0.0001


Since


the probability of each coefficient was


than the












Hypothesis VIII:


There is no relationship


between the average self-correction rates on

the oral reading sample and the total reading

achievement score on the MAT.

Pearson-product-moment correlations were computed to test the

strength of the relationship between the two variables designated in the


hypothesis.


For each grade, separately and combined, Pearson-product-


moment correlations were computed for the average self-correction rates


ASCR


and the total reading standard score on the MAT.


For the Primary


I second grade test, the correlation coefficient was 0.716 with a p


value of 0.0001; for the third grade Primary II test


t was 0.247


with


a p value of 0.1790.


5, with a p value


For both level

of 0.0019. Si


of the test, the coefficient was


nce the probability of the correla-


tion coefficients for each grade was not less than the 0.05 level


designated


hypothesis


as the criterion for statistical significance


was not rejected.


the null


The self-correction rates were more close-


ly associated with the second grade


test, Primary I, but not with


the third grade test, Primary II.


Hypothesi


There is no relationship between


of linguistic awareness, as measured by


the total score on the Concepts about Print Test,

Sand, the average self-correction rates on the oral

reading samples, and the score on the MAT for the












variable to be made using two or more independent variabi


This


sis also facilitated measurement of the strength of the relation-


ship between the dependent variable,


core on the MAT, and a single


independent variabi


either the Sand or the ASCR


, while controlling the


level of another independent variable


When reviewing the results of two previous hypotheses


Hypothe-


ses VII and VIII, the relationship of each independent variable to the

MAT differed. The score on the Sand correlated significantly with the

MAT for both grade levels and for the total Title I sample while the

strength of the correlation of the average self-correction rates to the


MAT was 1


strong for the total group, moderately strong for the


second graders, and


significant for the third graders.


Thus, the


effectiveness of the two variable


as predictors needed to be explored


careful


For the 42 Title I students who had both an average


correction rate and a fal


self-


score on the Sand, a multiple regression


anal

Tabl


ysis strategy was used to test the null hypothes


the R2


As shown in


was equal to 0.437 meaning that 43.7 percent of the


variability


in the score on the MAT Primary I and II was shared with


the fall Sand score and the average self-correction rate.


The computed


statistic for the overall reg


ress


ion equation equaled 15.93 with a


probability of 0.0001, given no relationship


exists between the inde-


pendent variables


the fall Sand score and the ASCR


. and the denpndent


J













grade

the R2


evel are reported in Table 4.


For the 21 Title I second graders,


was equal to 0.470 meaning that 47 percent of the variability in


the score on the MAT was


shared by the fall Sand and the ASCR.


The com-


puted F stat


stic for the regression equation equaled 8.43 with a proba-


ability of 0.0024, given no relationship exists between the independent


variable


es, the fall Sand and the ASCR


, and the dependent variable, the


evel of reading achievement on the MAT Primary I.


For the 21 Title I


third graders


, the R2


was equal to 0.085 meaning that


5 percent of the


variability in the score on the MAT Primary II was shared by the fal


Sand and the average self-correction rate.


The computed F


statistic was


0.88 with a probability of 0.4306, given no relationship


exists between


independent variabi


the fall


Sand and the ASCR


, and the dependent


variable, the level of reading achievement on the MAT Primary II.


the probability of each F statistic wa


designated


Since


not less than the 0.05 level


as the criterion for statistical significance, the null


hypothesis


was not rejected.


Thus, it was found that the average


self-


correction rates and the


scores on the fall Sand, combined


were not


significant predictors of the reading achievement score on the Metro-


politan Achievement T


ests.


It was also noted in the multiple regression analy


ses,


that the


scores on the Sand and the average self-correction rates did not con-


tribute


equally to the prediction of the MAT score when entered


t nn n +hnr 4n4- +kc v'.nnr-' rtn, ~ A rk ,. -n n 4 b" r. r; 4- I. nn .I 1 4; .


c, e 1,,,,,,,,~ 1.,






78




relatively small with the exception of that for the Sand for the total


Title I sample, 24.19 with a p value of 0.0001.

exceeded the 0.05 level set as the criterion fc


All other F statistics


,r statistical significance.


TABLE 4
PREDICTION OF SCORE ON THE MAT BY THE COMBINED
FALL SAND SCORE AND THE AVERAGE
SELF-CORRECTION RATE


Source F R2 p

MAT Primary I
and II by
Sand and ASCR 15.93 0.437 0.0001

Sand 24.19 0.0001

ASCR 0.80 0.3749

MAT Primary I
by Sand and
ASCR 8.43 0.470 0.0024

Sand 3.65 0.0712

ASCR 3.41 0.0804

MAT Primary II
by Sand and
ASCR 0.88 0.085 0.4306

Sand 1.36 0.2580

ASCR 0.34 0.5658


From the data, it appeared that both independent variables were not












their effectiveness as predictors of reading achievement.


This rela-


tionship was explored by Hypothesi


Hypothesi


There i


no relationship between


the level of linguistic awareness, as measured

by the Sand and the average self-correction rate.

This hypothesis was tested by computing a Pearson-product-moment


correlation of the Sand with the average self-correction rates.


For each


grade, separately and combined, correlations between the ASCR and both


the fal


and spring administrations of the Sand were calculated.


the second graders


, the correlation coefficient for the fall Sand with


the ASCR was 0.581 with a p value of 0.0046


for the spring Sand with


the ASCR


t was 0.686 with a p value of 0.000


For the third graders,


the correlation coefficients were computed as 0.042 with a p value of


0.85 for the fall Sand and the ASCR and 0.17


with a p value of 0


for the spring


Sand and the ASCR.


When both groups were combined


correlation coefficients were 0.346 with a p value of 0.0215 for the

fall Sand and 0.353 with a, p value of 0.0060 for the spring Sand.

Since the probability of each correlation coefficient for each grade


was not 1


than the 0.05 level designated as the criterion for sta-


tistical significance, the null hypothesis was not rejected.


It was


found that there was a moderately strong positive relationship between

scores on both administrations of the Sand and the average self-


correction rates for the second araderS


hut nnt fnr thp third nrad rc


.






80




in the second graders, but not in the third graders, or it may be a

result of improper book placement for the third grade students which

would affect the average self-correction rate reported for them.

It should be noted that the low correlational value of the


coefficients for the ASCR and the Sand may have


affected the results of


Hypothesi


score


, interfering with effect


on the MAT by both variables.


and of their effect


veness


prediction of the reading


The relationship of both variables


as predictors was explored in Hypothesis XII.


The relationship of the average self-correction rates to reading


achievement on the CTBS was explored in Hypothesis


Hypothesis XI:


There is no relationship between


the average self-correction rates and the total


reading


core on the Comprehensive Tests of Basi


Skill


To test this hypothesis


Pearson-product-moment correlations


were computed for both Title I groups, separately and combined.


second graders, the correlation was moderate


For the


trong--0.64 with a p


value of 0.0002; howev


for the third graders


, the coefficient wa


low--O.07 with a p value of 0.6755.

the correlation coefficient for the


When both groups were combined,

ASCR and the CTBS reading achieve-


ment score was 0


each coefficient was not


with a p value of 0.


Since the probability of


than the 0.05 level designated


as the


criterion for statistical significance, the overall null hvoothesis was






81





moderately strong positive relationship for the second grade Title I

students, but not for the third graders.


Hypothesi


There is no relationship between


evel of linguistic awareness, as measured by


the total score on the fall Sand and the average

self-corrections rate on the oral reading samples,

and the reading achievement score on the CTBS


A multiple


regression analysis


strategy was used to test this


hypothesis.


Multiple regression analysis allowed predictions for one


variable, total reading standard score on the CTBS, to be made using


two independent variables, the fall Sand score and the average


self-


correction rates.


analysis


also facilitated measurement of the


strength of the relationship between the dependent variable, the CTBS


score, and a single independent variable


ASCR.


either the Sand score or the


The results of the analyses for the total Title I group and for


each grade level separately are reported in Table 5.


For the 42 Title I students who had both an average


correction rate and a fall score on the Sand, the R2


self-


was equal to 0.474


meaning that 47.4 percent of the variability in the score on the CTBS


Levels C and I was shared with the fall Sand score and the ASCR.


computed F


statistic for the


overall


regression equation equaled 18.47


with a probability of 0.0001


given no relationship


exists between


the independent variables, the fall Sand score and the average


self-






82





The results of separate multiple regression analysis of the

average self-correction rates and the fall Sand on the CTBS score for


each grade level are reported in Table 5.


For the


21 Title I second


graders, the R


was equal to 0.41


meaning that 41


percent of the


variability


in the score on the CTBS/C wa


shared by the fall Sand


and the ASCR. The

equation equaled 6


computed F statistic for the overall regression


with a probability of 0.0058, given no relation-


ship exists between the independent variabi


ASCR


the fall Sand and the


, and the dependent variable, the level of reading achievement on


the CTBS/C.


since the probability of the test statistic was less than


the 0.05 level designated as the


criterion for significance, there was


a relationship between the fall Sand, the ASCR and reading achievement.


For the 21 Title I third graders, the R


meaning that


was equal to 0.357


5.7 percent of the variability in the score on the CTBS/I


was shared by the fall Sand and the ASCR.

for the overall regression equation equaled


The computed F statistic


with a p value of


0.0151, given no relationship exists between the independent variable


the fall


and and the average self-correction rates, and the dependent


variab


the 1


evel of reading achievement on the CTBS/I


Since the


probability of the test statistic was


than the 0.05


evel desig-


nated


as the criterion for significance, there was a relationship


between the fall Sand, the ASCR and reading achievement.


since the probability of each test statistic was


. '-- ..


than







83





graders, separately and combined, the average self-correction rates

and the fall Sand, were significant predictors of the total reading

standard score on the CTBS.



TABLE 5
PREDICTION OF SCORE ON THE CTBS BY THE COMBINED
FALL SAND SCORE AND THE AVERAGE
SELF-CORRECTION RATE


Source F R2 p

CTBS/C and I
by Sand and
ASCR 18.47 0.474 0.0001

Sand 30.38 0.0001

ASCR 0.25 0.6215

CTBS/C by Sand
and ASCR 6.82 0.418 0.0058

Sand 2.67 0.1185

ASCR 3.05 0.0969

CTBS/I by Sand
and ASCR 5.28 0.357 0.0151

Sand 10.50 0.0043

ASCR 0.01 0.9263


It should be noted in Table 5 that the scores on the Sand and

the ASCR did not contribute equally to the prediction of the CTBS score


--






84





however, was statistically significant for the total group and for the


third graders with p values 1

designated for statistical si


than the 0.05 leve


gnificance.


set as the criterion


These results confirm those of


earlier hypotheses, specifically Hypotheses V, X, and XI.


Findings Related to the Questions


Question 1


Is there a difference in the development of

print awareness


Among age levels

Among grade levels

Between pre-readers and readers?


When the data for the Sand were regressed across the age level

tested, it was found that there was a difference in print awareness.


From a confidence interval using the point estimate for the


data appeared to have a positive linear trend.


slope


As the age of the sub-


ject increased, the score on the Sand also increased.

To answer the second part of the question concerning the rela-


tionship of print awareness to the subject's grade level


a one-way


of variance was computed.


From this analysis


it was deter-


mined that there was a difference among grade levels in print awareness.

When the change in print awareness in the form of gain scores was com-

pared between the first graders and the Title I subjects, the first






85





graders appeared to have gained a greater awareness of print concepts


during the year than the Title I


students.


How do pre-readers and readers compare on their level


awareness?


of print


When the Sand scores for the kindergarteners and first


graders were combined and compared to those of the Title I second and


third graders, a difference in print awareness wa


found


In a complex


contrast of Sand scores, it was found that the Tit


significantly


students performed


better on the first administration of the Sand than the


kindergarten and first grade subjects.


Thus, in thi


comparison, the


readers had a higher


evel of print awareness than the pre-readers


possessed.


Question


Is there a progression in the development of a

self-monitoring reader across


Book


Grad


evels


level


The self-correction rates for 59 Title I students were computed


for each book


When compared with the 0.20 to 0.33 range, each


book


evel had the following percentages of


students having self-


correction rates within that range--first pre-primer, 0 percent; second


pre-primer


percent


third pre-primer,


42 percent


fourth pre-primer,


percent


primer, 26 percent; and first reader,


28 percent


Thus, the






86





compiled, the following percentages were found--first pre-primer,

50 percent; second pre-primer, 29 percent; third pre-primer, 35 per-

cent; fourth pre-primer, 55 percent; primer, 54 percent; and first


reader


, 39 percent.


The large percentage


of students at


some book


level


having


elf-correction above the desired range may indicate one


of two possibilities--either the book placement wa


easy,


inappropriate, i.e.,


or the student may have become more efficient at processing


print, thereby increasing his/her self-correction rate.


For each grade level


, the self-correction rates were examined.


Within the second grade Titl


I sample, 61 percent of the


28 subjects


were within the desired range of 0.20 to 0


for at least one book


level, compared to 42 percent of the 31 third grade subjects. This

indicated efficient processing of print by the correction of one in


every three to five errors.


The lower percentage of third grade


students falling within this range may have indicated that they were

incorrectly placed in a basal reader which was too easy for their

instructional reading level, or they were becoming more proficient in


correcting their own word recognition errors.


students who never attained a


Although there were some


self-correction rate within the 0.20 to


0.33 range, all students, with the exception of one third grader,

showed evidence of developing a self-monitoring system of error correc-


tion.


Some students


showed evidence of error correction which exceeded


the desired range, indicating greater efficiency in the processing of












52 percent or 16 of the 31 third graders.


Thus, it appeared that


the development of a self-monitoring reader became more evident as the


student grew older


Not only did the two grade levels differ on the percentages of

students within and above the desired range, but also the book level for

the higher percentage of within-range corrections was not the same for


both groups


The peak percentage of second graders having within-range


self-corrections occurred at the second pre-primer level with 50 percent

of the students having self-correction rates between 0.20 and 0.33, com-


pared to the third pre-primer level with


percent of the third graders


self-correcting within the desired range.


Question


Does the ratio of self-corrections to errors

estimate the level of reading achievement?

The answer to this question was found by examination of the

Pearson-product-moment correlations of the average self-correction

rates with the measure of reading achievement, the Metropolitan Achieve-


ment Tests or the Comprehensive Tests of Basi


Skills.


Also, linear


regression analysis


of the above variables yielded confirming informa-


tion about the nature of that relationship


The ASCR-MAT correlation


was higher for the second graders, 0.649, than for the third graders,


0.078.


The correlation was


statistically significant for the second






88





statistically significant for the second graders at both the 0.05 and


0.0l


evels, but not at all for the third graders.


Thus, for the total


Title I sample


the average self-correction rates were not found to con


sistently predict reading achievement.


of reading achievement for one grade


The ASCR was a better predictor


, the second grade, but not


for the other.


Question 4


Does the


evel of print awareness estimate


evel of reading achievement?


Through two different anal


yses,


the answer to this question was


found.


Using the Pearson-product-moment correlation, the coefficient


for the fall administration of the Sand with the CTBS was 0.686 for all


subjects; and 0.


for the spring administration of the Sand.


Both correlations


were


significant at the 0.01 and 0.05 level


coefficient for the fall Sand with the MAT was 0.660

spring Sand was 0.638 for the Title I subjects. Aga


statistically


while that for the


in, both were


significant at the 0.01 and 0.05 levels.


When the Pearson-product-moment correlations were calculated


for each grade level for the


fall Sand and the CTBS, the following


coefficients resulted--first grade, 0.803


second grade Title I, 0.570;


and third grade Titl


the 0.05


I, 0.79


evel, with those


All were


statistically significant at


for the first and third grade Titl


I students












Primary II


0.672 for the spring Sand and Primary II (third grade).


All correlation coefficients were significant at the 0.05 level.


Thus


,for the first and third grade subjects


, the fall administration


of the Sand was more highly


correlated with


scores


on the grade 1


reading achievement test, the CTBS, Levels B and I, respectively


than


for the

subjects


second grade subjects who received Level C.


the fal


For the Titl


Sand had a higher correlation with the second grade


test MAT Primary I (the out-of-level measure of reading achievement)


than for the third grade test, Primary II.


subjects


When the data for the Title


' fall administration of the Sand were combined with those


of the MAT and CTBS in a linear regression analysis, both achievement

measures followed a positive linear trend with the Sand.

















CHAPTER V
SUMMARY, CONCLUSIONS, AND RECOMMENDATIONS


Summary


This study examined the relationship of two print awareness

indicators, the Concepts about Print Test, Sand, and the average self-

correction rates or oral reading errors, to subsequent reading achieve-


ment.


The following questions were raised:


there a difference in the development of


awareness


Among age levels


b. Among grade levels

c. Between prereaders and readers?

Is there a progression in the development of

a self-monitoring reader across

a. Book levels

b. Grade levels

Does the ratio of self-corrections to errors

estimates the level of reading achievement?


Does the leve


of print awareness estimate the


evel of reading achievement?













All subjects were tested with the Concepts about Print Test,


Sand, by the researcher.


In addition, oral reading errors were recorded.


Both indicators, the Sand and the average self-correction rates, were

used to predict subsequent reading achievement.

The data were analyzed using Pearson-product-moment correlations,


of variance


, linear and multiple regression analysis.


data indicated that the Sand


scores


were effective predictors of read-


ing achievement for the Titi


I children


, but the average


self-


correction rates were not.


However, it was found through multiple


regression analysis that the two indicators when combined did not


effectively predict reading achievement.


Scores on the Sand were


found to correlate


significantly with the Comprehensive T


ests


of Basic


Skills


Level


, and I; and the Metropolitan Readiness Tests,


Level


I and


Linear trends were found for the Sand scores and the


subj


age and grade


evels.


Self-correction rates appeared to


fluctuate according to the


evel of the basal reader.


The conclusion drawn from the data is that indicators of


print awareness, such


as the Sand, are effective predictors of


reading achievement for primary grade children.


Implications


Upon examination of the study


results


it was learned that


the Concepts about Print Test, Sand, was an effective predictor of












(Day, Day, Spicola, and Griffin,


1979; Johns, 1980).


The Sand was


also demonstrated to correlate highly with a widely accepted reading


readiness measure


, the Metropolitan Readiness T


ests


Furthermore


the correlation of the Sand with the MAT, Primary I for the Title I

second graders, 0.61, approached that of the MRT with the MAT, 0.70.


The significant correlations of the Sand with the


CTBS further demon-


stated its effectiveness


fore, these results


as a predictor of reading achievement.


support the


There-


use of the Sand for young children from


kindergarten through third grade, and especially


like the Titi


for disabled readers


subjects.


While the average self-correction rates were not as highly


correlated with reading achievement


scores


on the Sand, they were


important indicators of reading progress.


but one of th


By the end of the year, all


Title I subjects began to monitor their own reading.


Secondly, only four of the 59 Titi


I subjects failed to attain or


surpass


a self-correction rate of 0


0 to 0.33.


self-correction


rate when averaged for the entire Title I sample


0.36, exceeded the


upper


imit of the range,


indicating that most


subjects were well


on their way toward becoming self-monitoring readers.


corrections for the Title I sample


The range of self-


was 0.00 in November to 0.83 in


April, compared to that of the first graders in Cohen'


(1975) study


whose


range for non-instructional material


was 0.00 in November to


0.30 in June.