‘7! l .
23 it}
In. 1. £313....“
In! .. ....
y f.

‘ ,tr.
21%.}:
V ‘ III.
a»; . .
. z
4‘ 3’9.
2 .. .

:3éks‘
3.
, a...
maﬁa)“
«...."m‘mw... _ “5%.. ‘
E73!!! .v‘ﬂ.“
.9
. a.
, 3%
1‘

V.
.1.

a. 1:1

...W ;
...
«e...

L

. ....

9.33..

.1 a
I:

 

at: ,
.‘Dl :
littz
. FT.

1. .\‘

a.

.2
7.617197%]
....‘ixaui
ikn‘ti‘
...xlcahﬂnuuil...
yin-3...: $153.st krill.“
fix... . 3%.”er naﬂnrgzhnn
..2 11:22.... it:
‘ :3
arm
'n
...

w... . Jag“! ..

2r.

 

.: J... a fit...
.... ... T -2 .t..r;"~.n..v.._.

: :5 .3: 1.

 

“q LIBRARY
1m 18 Michigan State
,; 3°! University

 

 

 

This is to certify that the
dissertation entitled

PATTERNS IN AND PREDICTORS OF ELEMENTARY STUDENTS'
READING PERFORMANCE: EVIDENCE FROM THE DATA OF
THE PROGRESS IN INTERNATIONAL READING LITERACY
STUDY (PIRLS)

presented by
YONGHAN PARK

has been accepted towards fulﬁllment
of the requirements for the

Ph.D. degree in Educational Psychology &
Educational TechnologL

2/ WM 73W MML

Major Professor’s Signature

IL/CCI /.L00¥

 

 

 

 

Date

MSU is an Afﬁrmative Action/Equal Opportunity Employer

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5/08 K:lProj/Acc&Pres/ClRC/DaleDue.indd

PATTERNS IN AND PREDICTORS OF ELEMENTARY STUDENTS’ READING
PERFORMANCE: EVIDENCE FROM THE DATA OF THE PROGRESS IN
INTERNATIONAL READING LITERACY STUDY (PIRLS)

By

Yonghan Park

A DISSERTATION

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

DOCTOR OF PHILOSOPHY
Educational Psychology & Educational Technology

2008

ABSTRACT
PATTERNS IN AND PREDICTORS OF ELEMENTARY STUDENTS’ READING
PERFORMANCE: EVIDENCE FROM THE DATA OF THE PROGRESS IN
INTERNATIONAL READING LITERACY STUDY (PIRLS)
By
Yonghan Park
This dissertation consists of two manuscripts reporting on patterns in and
predictors of fourth-grade students’ reading performance. By analyzing the data from the
Progress in International Reading Literacy Stuay (PIRLS), the studies in this dissertation
focused on student-level and classroom-level information in order to explore signiﬁcant
predictors of US. students’ reading performance. The multilevel model was the main
analytical approach of the studies in this dissertation, which involved students nested
within classrooms. The studies drew on a complex view of the development of reading.
In the ﬁrst study of this dissertation, I explored predictors of student performance

on two different purposes of reading, that is, informational reading and literary reading.
The inspiration for this study came from the fact that US students have a large
performance gap favoring literary reading over informational reading on average
according to the reports from PIRLS 2001 and 2006. In order to understand this gap, the
study examined differential contributions of literacy-related variables to informational
versus literary reading performance on a country-level and within selected countries
including the United States. The results indicated that several genre-related activities and
curriculums were signiﬁcantly related to fourth-grade children’s reading performance on

either literary or informational reading and, in some cases, across both of these genres. In

the US, for instance, more instruction with informational text in classroom predicted
higher reading performance in informational reading. The implication of this study is that
substantial opportunities to learn with both literary and informational text and also
opportunities to learn text style and structure should be provided in primary grades for all
children.

The second study focused on the variables of reading motivation, based on their
strong relationships to student reading performance shown in the ﬁrst study. The
purposes of the study were to explore underlying factors in reading motivation measured
in PIRLS and to investigate the relationships between those factors and reading
performance. Exploratory and conﬁrmatory factor analyses revealed that motivational
items in PIRLS consisted of four different factors which also could be clustered as two
contrasting facets (internal orientation and external orientation) under two different
dimensions (attitudes toward reading and reading self-concept). Further analysis
discovered interaction effects between internally oriented motivation and externally
orientated motivation to predict better reading performance. Although the relationship of
one motivational facet to reading performance was different depending on the levels of
its contrasting facet, internally oriented motivation was always a positive predictor of
student reading performance. Extrinsic motivation under the dimension of attitudes
toward reading, for example, positively related to reading performance if a student had
high intrinsic motivation. However, it was a negative predictor of reading performance
for those students with low intrinsic motivation. Finally, four motivational factors

together explained 17% of the student-level variance in overall reading performance.

To Juwon, Jungwon, Eun-I, and my Parents

iv

ACKNOWLEDGEIVIENTS

It gives me great pleasure, with the completion of this dissertation, to offer my
thanks to a number of individuals. Although this dissertation is a product of ﬁve-years of
graduate studies, I believe it will be a cornerstone of my upcoming professional life
beyond graduate school. This work would not have been possible without the support of
many individuals, from the very onset of my studies to the ﬁnal writing-up, and I have
been fortImate to receive their generous intellectual, emotional, and ﬁnancial support.

I would like ﬁrst to express my deepest gratitude to my advisor Nell K. Duke.
She has proven to be a supportive advisor, a serious scholar, and also a warm-hearted
friend. She generously shared her many insights into my research and provided inspired
guidance at every step along the way. Thank you, Nell. You are my hero and the role
model for my academic life.

I also would like to extend my thanks to my other committee members: Barbara
Schneider, Evelyn Oka, and William Schmidt : Barbara for serving as my dissertation co-
director and for providing me with thoughtful advice in regard to data analysis and
writing, Evelyn for helping me with the second study for this dissertation and without
whose intellectual guidance I would never have been able to come up with the solution of
the model, and Bill for his inspiring scholarship and for helping me formula my research
to address issues in a global arena.

Another person I wish to thank is the late Michael Pressley, my ﬁrst-year advisor.
Although he was only here for the ﬁrst three years of my program, the inﬂuence of his
gifted expertise and intellectual passion on my work is incalculable, and I will always

miss his humor and insight.

I am also indebted to my many ﬁiends at Michigan State University for their
friendship and warm support. In particular, I would like to mention Wonpyo Hong, J iwon
Son, Minjung Bae, Sooyong Byun, Taeseop Shin, Gayoung Ahn, Shenglan Zhang, Juliet
Halladay, Alison Billman, Wooyeong J 00, and Kwangsu Mok, all of whom shared their
ideas and feelings with me while I worked on this dissertation research.

The ﬁnal completion of my dissertation writing took place while I was conducting
research at the Center on Teaching and Learning at the University of Oregon. As this
opportunity provided me with precious experiences that broaden my view and deepen my
knowledge, I would like to express my special gratitude to the director of the center,
Edward J. Kame’enui. Ed provided me with this great opportunity at the very start of my
academic career and I would like to thank him for his warm and thoughtﬁrl support of my
work and research. I would also like to thank Hank Fien of the Center for being my
valued colleague while working closely with me on my research here as well as for his
companionship on thoughtful jogs along the McKenzie River. His support was
instrumental in succeeding with my research at the Center and, without him, I would
never have been able to ﬁnish my dissertation in such a timely fashion.

I also wish to express my gratitude to other researchers and colleagues at the
center for their encouragement and friendship: Scott Baker, Mike Stoolmiller, Jorge
Preciado, Erin Chaparro, Williex Menitt, Deni Basaraba, and Helena Reid, to name a few,
all who gave me support in various ways.

Last but not least, I want to thank my family for their support of me on my
academic journey—there are not enough words to express my sincere gratitude towards

them and their undying encouragement, trust, and patience which contributed

vi

immeasurably to my journey’s success to date. Special thanks go to my two wonderful
kids, Juwon and Jungwon, to whom I would like to say: I am sorry that I could not have
been with you for so much of the time but please know that I was always thinking of you

and I love my two little ones both very much.

vii

TABLE OF CONTENTS

LIST OF TABLES ................................................................................... ix
LIST OF FIGURES ................................................................................. xii
INTRODUCTION ................................................................................... 1
International Studies of Reading ............................................................... 2
Overview of the Dissertation ................................................................... 7
Tables ............................................................................................ l 1
References ....................................................................................... l6

STUDY ONE: PREDICTORS OF ELEMENTARY STUDENTS’ READING
PERFORMANCE IN INFORMATIONAL AND LITERARY READING

.................................................................................................................... 1 8
Abstract .......................................................................................... 18
Introduction ..................................................................................... 19
Theoretical & Empirical Background ...................................................... 21
Purposes of the Study ........................................................................... 29
Methods ........................................................................................... 30
Results ............................................................................................ 36
Discussion ........................................................................................ 50
Tables ............................................................................................. 5 9
References ....................................................................................... 76

STUDY TWO: HOW MOTIVAITONAL CONSTRUCTS INTERACT TO PREDICT

READING PERFORMANCE .. .............................................................. 82

Abstract .......................................................................................... 82
Introduction ..................................................................................... 84
Theoretical & Empirical Background ...................................................... 85
Purposes of the Study ........................................................................... 95
Methods ........................................................................................... 95
Results ............................................................................................ 98
Discussion ..................................................................................... 107
Tables ........................................................................................... 113
Figures .......................................................................................... 124
References ..................................................................................... 126
APPENDICES .................................................................................... 132
A Description of the Variables Used for the Study ..................................... 133
B Analysis Results of the Data from Denmark .......................................... 142
C Analysis Results of the Data from England ........................................... 149
D Analysis Results of the Data from Taiwan ........................................... 156

viii

LIST OF TABLES

TABLES FOR INTRODUCTION
Table 1.1 US. Students ’ Reading Performance in the IEA Reading Literacy Study .. . ..11
Table 1.2 US. Adults ’ Reading Performance in the International Adult Literacy Survey
(L4LS) .................................................................................... 12
Table 1.3 US. Adults ’ Reading Performance in the Adult Literacy and Life Skills Survey
(ALL) ................................................................................... 13
Table 1.4 US. I 5 - Y ear-Olds ’ Reading Performance in the Programme for International
Student Assessment (PISA) .............................................................. 14
Table 1.5 (IS. F ourth-Grade Students ’ Reading Performance in the Progress in
International Reading Literacy Study (PIRLS) ...................................... 15
TABLES FOR STUDY ONE
Table 1.1 Average Reading Scores of US. Students by Reading Purpose and Gender in
PIRLS 2001 and 2006 ................................................................ 59
Table 1.2 Means, Standard Deviations, and Intercorrelations of Country-level Variables
and Reading Performance ........................................................... 60
Table 1.3 Multiple Regression Analysis Summary for Country-level Variables
Predicting Country-level Reading Performance .................................. 62
Table 1.4 Means and Standard Deviations of Country-level Reading Performance by
Country-level Curriculum Emphasis at 4'” grade on Reading to Acquire
Information and Reading for Literary Experience ................................. 64
Table 1.5 Means, Standard Deviations, and Intercorrelations of US. Classroom-level
Variables and Reading Performance ................................................ 65
Table 1.6 Means, Standard Deviations, and Intercorrelations of US. Student-level
Variables and Reading Performance ............................................... 67
Table 1.7 US. Results ﬁ'om HLM Analyses Predicting Informational Reading
Performance ............................................................................ 69
Table 1.8 US. Results ﬁ'om HLM Analyses Predicting Literary Reading Performance

.......................................................................................... 70

ix

Table 1.9 Means and Standard Deviations of US. Average Raw Responses in Reading
Amount and Reading Instruction Amount Related Variables .................... 71

Table 1.10 Country-level and Within-US. Classroom/Student-level Predictors of
Reading Performance in Informational Reading and Literary Reading ... ..72

Table 1.11 Descriptive Comparisons of Country-level Characteristics in the Selected
Countries ﬁom PIRLS 2006 ........................................................ 73

Table 1.12 Average Reading Performance Scores by Informational-Literary Reading
Performance Gap Size in Selected Countries .................................... 74

Table 1.13 HLM Coefﬁcients for Major Predictors of Reading Performance in Selected

Countries .............................................................................. 75
TABLES FOR STUDY TWO
Table 2.1 Motivational Variables in PIRLS ................................................ 113
Table 2.2 Descriptive Statistics and Intercorrelations of Motivational Variables,
Background Variables & Reading Scores ........................................ 114
Table 2.3 Extracted Factors ﬁom Exploratory Factor Analysis .......................... 116

Table 2.4 2X2 Reading Motivational Constructs of PIRLS Derived ﬁom Factor Analysis
........................................................................................... l 17

Table 2.5 Descriptive Statistics and Intercorrelations among New Motivational Factors,
Background Variables, & Overall Reading Scores ............................. 118

Table 2.6 Comparisons of the Fit Statistics between the 2 Factor Model and the 2X2
Factor Model ﬁ'om Conﬁrmatory Factor Analysis ........................... 119

Table 2.7 Interactions between Intrinsic Motivation (1110 and Extrinsic Motivation (EM
in Predicting Overall Reading Score .............................................. 120

Table 2.8 Interactions between Self-referenced Perceived Competence (SPC) and
Peer--referenced Perceived Competence (PPC) in Predicting Overall Reading
Score .................................................................................. 121

Table 2.9 Results ﬁom HLM Analyses Predicting Overall Reading Performance
........................................................................................... l 22

TABLES FOR APPENDICES

Table B]

Table 3.2

Table B.3

Table B.4

Table C.1

Table C.2

Table C.3

Table C.4

Table D.1

Table D2

Table D3

Table D4

Means, Standard Deviations, and Intercorrelations of Classroom-level

Variables and Reading Performance (Denmark) ................................. 143
Means, Standard Deviations, and Intercorrelations of US. Student-level
Variables and Reading Performance (Denmark) ................................. 145
Results ﬁ'om HLM Analyses Predicting Informational Reading Performance
(Denmark) .............................................................................. 147
Results ﬁom HLM Analyses Predicting Literary Reading Performance
(Denmark) .............................................................................. 148

Means, Standard Deviations, and Intercorrelations of Classroom-level
Variables andReading Performance (England) 150

Means, Standard Deviations, and Intercorrelations of US. Student-level
Variables and Reading Performance (England) .................................. 152
Results ﬁom HLM Analyses Predicting Informational Reading Performance
(England) .............................................................................. 154
Results ﬁ'om HLM Analyses Predicting Literary Reading Performance
(England) .............................................................................. 155
Means, Standard Deviations, and Intercorrelations of Classroom-level
Variables and Reading Performance (Taiwan) ................................... 157
Means, Standard Deviations, and Intercorrelations of US. Student-level
Variables and Reading Performance (Taiwan) ................................... 159
Results ﬁom HLM A nalyses Predicting Informational Reading Performance
(Taiwan) ................................................................................ 161
Results from HLM A nalyses Predicting Literary Reading Performance
(Taiwan) ................................................................................ 162

xi

LIST OF FIGURES

FIGURES FOR STUDY 2
Figure 2.1. Interaction of extrinsic and intrinsic motivation ............................ 124

Figure 2. 2. Interaction of self-referenced and peer-referenced perceived competence
........................................................................................ 125

xii

INTRODUCTION

The ability to read is critical in and out of school from childhood through
adulthood in modern society. Still, many students in the US. are not developing
appropriate reading skills in spite of a long history of national efforts to improve
children’s reading abilities. According to a recent report ﬁ'om the National Assessment of
Educational Progress (NAEP) 2007, for instance, 34% of fourth-grade students and 27%
of eighth-grade students read below basic level which is deﬁned as “partial mastery of
prerequisite knowledge and skills that are ﬁmdamental for proﬁcient work at a given
grade” (Lee, Grigg, & Donahue, 2007). Given this, how can we improve children’s
reading abilities? This is the main research question that I have been examining
throughout my graduate studies.

Reading is not a simple process and thus it requires more complex approaches
that consider a variety of components including decoding, language comprehension,
memory, speed, vocabulary, world knowledge, type of texts, motivation and so on
(Pressley et al., 2008). This complexity implies that better reading programs should
reﬂect a comprehensive model of reading involving reader, text, activity and context in
order to fully develop children’s reading abilities (RAND Reading Study Group, 2002).

Drawing on the less simple perspective for reading and reading instruction, my
ﬁrst research interest during my graduate program focused on a rather neglected topic in
the ﬁeld of reading research, that is, reading motivation. As a part of my practicum, I
conducted a study to develop an instrument to measure elementary students’ volition in
reading (Park & Duke, 2005). In this study, volition was deﬁned as an individual’s

character and behavior for an effortful pursuit of goals, which enables him/her to

overcome obstacles and preserve intention despite adverse circumstances in the process
of reading. The concept of volition was distinguished from that of motivation, in that the
latter concerns initial commitment that provokes goal-directed learning activities while
the former concerns the maintenance of the initial motivation to achieve the goal. The
instrument found six factors under volitional character and volitional strategy use to
measure children’s volition in reading.

A second strong interest during my graduate career has been international
reading studies. Large-scale international studies on student academic performance
provide useful information which researchers and policy makers can analyze to see how
their nations are doing compared to other countries and to explore what factors
within/across countries explain their academic achievement in fundamental subject areas.
Several international studies have been conducted for examining literacy achievement,
especially focusing on reading abilities.

International Studies of Reading

This section describes some signiﬁcant international studies of reading
performance in which the US. has participated since 1990. These are the IEA Reading
Literacy Study, the International Adult Literacy Survey (IALS), the Adult Literacy and
Life Skills Survey (ALL), the Programme for International Student Achievement (PISA),
and the Progress in International Reading Literacy Study (PIRLS). Brief descriptions of
each study are given below.

IEA Reading Literacy Study
In 1991, the International Association for the Evaluation of Educational

Achievement (IEA) conducted the Reading Literacy Study to assess the reading skills of

fourth and ninth graders in 32 countries or jurisdictions around the world. This study
deﬁned reading as “the ability to understand and use those written language forms
required by society and/or valued by the individual” (Binkley & Rust, 1994, p. 103). It
measured student reading performance with three different types of text that students
oﬁen encounter in school and in everyday life: narrative prose, expository prose, and
documents. The study deﬁned narrative prose as text in which the writer tells a factual
or ﬁctional story. Expository prose was deﬁned as text in which the writer conveys
factual information or opinion. Documents referred to information displays such as
charts, maps, tables, graphs, lists, or sets of instructions. Reading scores were scaled
with an international mean of 500 and a standard deviation of 100. US. fourth graders
placed second (Mean = 547) among the participating countries on a combined score. U.S.
ninth-grade students were ranked ninth (Mean = 535) on a combined score among the
participating countries. Eight countries scored higher than the US, but only one country
among them showed a statistically signiﬁcant score difference with the US. The mean
scores and relative ranks of US. students on three reading subscales are shown in Table
1.1 (Elley & Schleicher, 1994). It is noticeable that US. fourth graders had a large score
difference between narrative and expository prose reading, which was the second biggest
achievement gap favoring narrative over expository prose reading among the 32
countries. This is consistent with the result ﬁ'om PIRLS 2001 in which, as discussed later,
U.S. fourth graders had the largest achievement gap favoring literary reading over
informational reading. In 2001, nine of the initially participating countries in the IEA
Reading Literacy Study replicated the study using only fourth graders to look for any

changes in their children’s reading performance. Overall reading scores of US. students

in 2001 did not change signiﬁcantly from 1991 (Martin, Mullis, Gonzalez, & Kennedy,
2003).
International Adult Literacy Survey

The International Adult Literacy Survey (IALS) was the ﬁrst comparative study
of adult literacy skills ever undertaken internationally. Over 75,000 adults aged 16 to 65
from 22 countries or jurisdictions including the US. were interviewed and tested in their
homes in 15 languages between 1994 and 1998. The main purposes of this study were to
ﬁnd out how well adults use information to ﬁmction in society and to investigate the
factors that inﬂuence literacy proﬁciency at home, at work, and across countries
(Tuijnman, 2000). IALS deﬁned literacy as “the ability to understand and employ
printed information in daily activities, at home, at work and in the community—to
achieve one’s goals, and to develop one’s knowledge and potential” (OECD & Statistics
Canada, 2000, p. x). It assessed adult literacy in three categories: prose literacy,
document literacy, and quantitative literacy. Prose literacy assessed the knowledge and
skills needed to understand and use information ﬁ'om continuous texts including
editorials, news stories, brochures, poems, and ﬁction. Document literacy tested the
knowledge and skills required to locate and use information contained in various non-
continuous texts, including job applications, payroll forms, transportation schedules,
maps, tables, and charts. Quantitative literacy evaluated the knowledge and skills
required to apply arithmetic operations to numbers embedded in written text. IALS
measured literacy proﬁciency for each domain on a scale ranging ﬁ'om 0 to 500. As
shown Table 1.2, the average scores of US. adults were 273.7 on prose literacy, 267.9 on

document literacy, and 275.2 on quantitative literacy. Those average scores were ranked

10th on prose literacy, 15th on document literacy, and 13th on quantitative literacy among
the 22 participating countries (OECD & Statistics Canada, 2000).
Adult Literacy and Life Skills Survey

The Adult Literacy and Life Skills Survey (ALL) was an international study
conducted in 2003 to provide participating countries or regions with comparative
information about reading and other abilities of their adult populations. This study
measured adults’ life skills in four domains: prose literacy, document literacy, numeracy,
and problem solving, although the US. did not ﬁeld the problem solving domain. Two
literacy domains, prose literacy and document literacy, were adapted from IALS
described above. Although similar to quantitative literacy in IALS, numeracy in ALL
was deﬁned more broadly and thus considered a different domain. The subjects were the
adults aged 16 to 65 ﬁom seven participating countries. The average scores of US.
adults in two literacy domains were 268.6 (5th ranked) on prose literacy scale and 269.8
(5m) on document literacy scale (see Table 1.3). Compared to the average scores in the
IALS, the US. average score on prose literacy in ALL decreased signiﬁcantly, while
there was no signiﬁcant change on document literacy (Statistics Canada & OECD, 2005).
Programme for International Student Assessment

The Programme for International Student Assessment (PISA) is an international
comparative study that has been conducted by the Organisation for Economic Co-
operation and Development (OECD) every three years beginning in 2000, to assess 15-
year-old students’ knowledge and skills in reading, math, science and problem solving.
The numbers of the participating countries or jurisdictions were 43 in 2000, 41 in 2003

and 56 in 2006 (OECD, 2006). This study deﬁned reading as “the ability to understand,

use and reﬂect on written texts in order to achieve one’s goal, to develop one’s
knowledge and potential, and to participate effectively in society” (OECD, 2003, p. 108).
The items in reading assessment were constructed considering text formats (continuous
texts and non-continuous texts), reading processes (retrieving information, interpreting
texts, and reﬂection and evaluation), and reading situations (reading for personal use,
reading for public use, reading for work, and reading for education). Scores were sealed
with a mean of 500 and a standard deviation of 100 (OECD, 2003). The average overall
reading scores of US. 15-year-old students were 504 (15th) in 2000 and 495 (18th) in
2003 (OECD, 2004). There was no statistically signiﬁcant difference between these two
scores. PISA 2000 provided reading performance scores on different text formats and
different reading processes. US. average reading performance scores on different text
formats were 504 (13th) on continuous texts and 506 (14th) on non-continuous texts. In
relation to different reading processes, US. average reading scores were 499 (15th) for
retrieving information, 505 (15th) for interpreting texts, and 507 (11th) for reﬂection and
evaluation (see Table 1.4).
Progress in International Reading Literacy Stuay

The Progress in International Reading Literacy Stuafv (PIRLS) is a 5-year cycle of
international comparative reading assessment that was ﬁrst carried out in 2001. It was
conducted under the supervision of the International Association for the Evaluation of
Educational Achievement (IEA). The assessed children were typically in the fourth grade
and the average age was around 10 in most countries including the United States (Mullis,
Martin, Gonzalez, & Kennedy, 2003). The numbers of participating countries or

jurisdictions were 35 in 2001 and 45 in 2006. PIRLS provided separate reading scores

for overall reading, literary reading, and informational reading. Those scores were sealed
with an international mean of 500 and a standard deviation of 100.

The results from PIRLS 2001 and PIRLS 2006 indicate that US. fourth-grade
students read better than the international average in both literary and informational
reading. The average scores of US students in 2001 were 542 (9th highest) on overall
reading proﬁciency scale, 550 (4th) on literary reading scale, and 533 (13th) on
informational reading scale. In 2006, those scores were 540 (186‘) on overall reading
proﬁciency scale, 541 (18*) on literary reading scale, and 537 (19th) on informational
reading scale. These average scores of US. students did not show a statistically
signiﬁcant change between 2001 and 2006 (see Table 1.5; Mullis, Martin, Gonzalez, &
Kennedy, 2003; Mullis, Martin, Kennedy, & Foy, 2007).

Overview of the Dissertation

This dissertation was formatted in an alternative style (Duke & Beck, 1999) in
order to report the results ﬁom my dissertation research effectively and move the
dissertation to publication efficiently. Following this part of overall introduction, two
independent but related studies I conducted using the data from PIRLS 2006 are
presented in two stand-alone manuscripts prepared for submission to academic journals.

The purpose of my dissertation research was to examine important predictors of
elementary student reading performance in broader contexts. Using the data from PIRLS
2006, I was able to explore various literacy-related home/school background variables of
fourth-grade students and examine reading performance for literary reading and
informational reading separately. In addition, the result from the US. could be compared

with those from other countries.

In the ﬁrst study, I explored various predictors of student performance in different types
of reading to ﬁnd out whether and, if so, how those variables differentially predict
student performance on informational reading and literary reading. The inspiration for
this study came ﬁom the fact, as noted earlier, that US fourth-grade students had a large
performance gap favoring literary reading over informational reading. The study
addressed the following questions: First, what are the signiﬁcant country-level predictors
of countries’ average reading performance in informational and literary reading?
Seconds, within the US, what variables from inside and outside classrooms are related
to student performance in different genres? Third, are there any similarities or
differences between the US. and other countries regarding the predictors of student
reading performance? By analyzing the country-level data and within-country data from
PIRLS 2006 using multiple regression and hierarchical linear modeling, this study
showed that genre-related in and out-of-school activities and curriculums were
signiﬁcantly related to student reading performance in a speciﬁc genre and, in some
cases, across different genres, within countries and also across countries. The
implication of this study is that we should provide substantial opportunities to learn with
both literary and informational text and also opportunities to learn text style and
structure in primary grades for all children.

As the ﬁrst study showed that reading motivation was a consistently signiﬁcant
predictor of student reading performance, in the second study, I focused on motivational
variables. As noted earlier, reading motivation has been a research interest of mine
throughout my graduate studies. The purposes of this study were to explore underlying

factors in reading motivation measured in PIRLS and to investigate the relationships

between those motivational factors and students’ reading performance. A special interest
was given to interactions between different motivational facets to predict better reading
performance. The study analyzed the US. data from PIRLS 2006 using
exploratory/conﬁrmatory factor analysis methods and hierarchical linear modeling. This
study ﬁrst showed that PIRLS motivation items clustered under two motivational
dimensions: attitudes toward reading and reading self-concept. Within each dimension,
two contrasting facets represented students’ internal or external orientation of reading
motivation. Speciﬁcally, student attitudes toward reading consisted of extrinsic
motivation and intrinsic motivation, while student reading self-concept was composed of
self-referenced perceived competence and peer-referenced perceived competence.
Second, internally oriented motivational facets and externally oriented motivational
facets interacted with each other in relation to student reading performance. Although
this interaction means that the relationship of one motivational facet to reading
performance is dependent on the levels of its contrasting facet, this study found that
intemally-oriented motivational facets were always more positive predictors of student
reading performance than externally-oriented motivational facets. Extrinsic motivation
under the dimension of attitudes toward reading, for example, positively related to
reading performance if a student had high intrinsic motivation. However, it was a
negative predictor of reading performance for those students with low intrinsic
motivation. Third, reading motivation was a strong predictor of reading performance
even after other literacy-related variables were statistically controlled. Four reading
motivational facets in this study explained 17% of the student-level variance in students’

reading performance. This study conﬁrms that reading motivation has a

multidimensional and multifaceted nature and that internally oriented motivation is
especially strongly linked to better reading performance. Without support ﬁ'om internally
oriented motivation, externally oriented motivation has a negative relation to
performance. The study suggests that researchers should consider interacting
relationships of different motivational constructs.

In summary, the studies in this dissertation offer important implications for genre
and reading motivation in relation to the development of reading abilities. They also
provide another demonstration of how data ﬁ'om large-scale international studies can be

meaningfully used for reading research.

10

Table 1.1

US. Students ’ Reading Performance in the IEA Reading Literacy Study

 

 

 

Grade Performance Overall Narrative Expository Documents
4th Mean (SE) 547 (2.8) 553 (3.1) 538 (2.6) 550 (2.7)
Grade Rank 2nd 2nd 3rd 3rd
9th Mean (SE) 535 (4.8) 539 (4.9) 539 (5.6) 528 (4.0)
Grade Rank 9th 6th 5th 14th

 

Note. A total of 32 countries or jurisdictions participated in the IEA Reading Literacy Study.

11

Table 1.2

US. Adults ’ Reading Performance in the International Adult Literacy Survey (IALS)

 

 

Quantitative
Performance Prose literacy Document literacy
literacy
Mean (SE) 273.7 (1.6) 267.9 (1.7) 275.2 (1.7)
Rank 10th 15th 13th

 

Note. A total of 22 countries or jurisdictions participated in the IALS.

12

Table 1.3

US Adults ’ Reading Performance in the Adult Literacy and Life Skills Survey (ALL)

 

 

Performance Prose literacy Document literacy
Mean (SE) 268.6 (1.3) 269.8 (1.5)
Rank 5th 5th

 

Note. Only seven countries or jurisdictions participated in the ALL.

13

Table 1.4

US. 15- Year-Olds ’ Reading Performance in the Programme for International Student

 

 

 

Assessment (PISA)
Text formats
Performance Overall
Continuous texts Non-continuous texts
Mean (SE) 504 (7.1) 504 (7.1) 506 (7.2)
Rank 15th 13th 14th

 

Note. The scores and ranks above are based on the report in Reading for Change: Performance and
Engagement across Countries (OECD, 2002) which conducted the analysis with a set of 31 countries or
jurisdictions from PISA 2000.

14

Table 1.5
(1.5. F ourth—Grade Students ’ Reading Performance in the Progress in International

Reading Literacy Study (PIRLS)

 

 

 

Year Performance Overall Literary Informational
Mean (SE) 542 (3.8) 550 (3.8) 533 (3.7)
2001
Rank 9th 4th 13th
Mean (SE) 540 (3.5) 541 (3.6) 537 (3.4)
2006
Rank 18th 18th 19th

 

Note. A total of 35 countries or jurisdictions participated in 2001 and 45 in 2006.

15

References

Binkley, M., & Rust, K. (1994). Reading literacy in the United States: Technical report
of the US. component of the IEA Reading Literacy Study. . Washington, DC: US.
Government Printing Office.

Binkley, M., & Williams, T. (1996). Reading literacy in the United States. Findings
from the IEA Reading Literacy Stuay. Washington, DC: US. Government
Printing Office.

Duke, N. K. & Beck, S. W. (1999). Education should consider alternative formats for the
dissertation. Educational Researcher, 28, 31-36.

Elley, W. B., & Schleicher, A. (1994). International differences in achievement levels. In
W. B. Elley (Ed.), The IEA Study of Reading Literacy: Achievement and
Instruction in Thirty-Two School Systems (pp. 35-63). Great Britain: Pergamon.

Lee, J ., Grigg, W., & Donahue, P. (2007). The nation ’s report card: Reading 2007
(N CES 2007—496). National Center for Education Statistics, Institute of
Education Sciences, US. Department of Education, Washington, DC.

Mullis, I. V. S., Martin, M. O., Gonzalez, E. J., & Kennedy, A. M. (2003). PIRLS 2001
international report: IEA ’s study of reading literacy achievement in primary
school in 35 countries. Chestnut Hill, MA: International Study Center, Boston
College.

Martin, M. O., Mullis, I. V. S., Gonzalez, E. J ., & Kennedy, A. M. (2003). Trends in
children's reading literacy achievement 1991-2001. Chestnut Hill, MA:
International Study Center, Boston College.

Mullis, I. V. S., Martin, M. 0., Kennedy, A. M., & Foy, P. (2007). PIRLS 2006
international report: IEA ’s progress in intenational reading literacy study in
primary schools in 40 countries. Chestnut Hill, MA: International Study Center,
Boston College.

OECD, & Statistics Canada. (2000). Literacy in the information age: Final report of the
International Adult Literacy Survey. Paris: Organisation for Economic Co-
operation and Development.

OECD. (2003). The PISA 2003 assessment ﬁamework. Paris: Organisation for Economic
Co-operation and Development.

OECD. (2004). Learning for tomorrow '3 world: First results ﬁom PISA 2003. Paris:
Organisation for Economic Co-operation and Development.

OECD. (2006). Assessing scientiﬁc, reading and mathematical literacy: A framework for
PISA 2006. Paris: Organisation for Economic Co-operation and Development

16

Park, Y., & Duke, N. (2005, December). An assessment of upper elementary students ’
volition for reading. Paper presented at the annual meeting of the National
Reading Conference. Miami, FL.

Pressley, M., Duke, N. K., Gaskins, I. W., Fingeret, L., Halladay, J., Hilden, K., Park, Y.,
Zhang, S., Mohan, L., Reffrtt, K., Bogaert, L. R., Reynolds, J ., Golos, D., Solic,
K., & Collins, S. (2008). Working with struggling readers: Why we must get
beyond the simple view of reading and visions of how it might be done. In T.
Gutkin & C. R. Reynolds (Eds.), The Handbook of School Psychology, Fourth
Edition. New York: Wiley.

RAND Reading Study Group (2002). Reading for understanding: Toward a research
and development program in reading comprehension. Santa Monica, CA: RAND.

Statistics Canada, & OECD. (2005). Learning a living : First results of the Adult
Literacy and Life Skills Survey. Paris: Organisation for Economic Co-operation
and Development.

Tuijnman, A. (2000). International Adult Literacy Survey - Benchmarking adult literacy
in America: An international comparative study. Ottawa, Canada: Statistics
Canada.

17

STUDY ONE
PREDICTORS OF ELEMENTARY STUDENTS’ READING PERFORMANCE

IN INFORMATIONAL AND LITERARY READING

Abstract

In increasingly information-rich societies, the ability to acquire and use
information from text effectively is becoming more important. This informational
reading, however, has received little attention in US. primary-grade classrooms. As
might be expected then, according to the results from the Progress in Informational
Reading Literacy Study (PIRLS), U.S. fourth-grade students have a large performance
gap favoring literary reading over informational reading. In order to explore predictors
of reading performance in these two different types of reading, this study analyzes the
data from PIRLS 2006. In addition to general demographic predictors, several genre-
related predictors were examined in relation to student reading performance in
informational reading and literary reading across countries and within selected countries.
This study found that in- and out-of-school genre-related activities and curriculum
signiﬁcantly predicted children’s reading performance in a speciﬁc genre and even
across different genres. The results suggest that we should provide substantial
opportunities to learn with both informational and literary text and also opportunities to

learn text style and structure in primary grades for all children.

18

Introduction

Narrowing the gaps in students’ academic achievement has long been one of the
crucial issues for educational researchers and policy-makers in the United States
(Stedman & Kaestle, 1987). But the effort to eliminate those gaps seems not yet
successful, as they have not substantially narrowed. In relation to reading achievement,
for example, Klecker (2006) argued that there continued to be a gap in reading
achievement between females and males, favoring females over males, according to the
longitudinal analysis of the National Assessment of Educational Progress (N AEP)
reading scores. Themstrom and Themstrom (2003) highlighted a steady reading
achievement gap among different racial groups from the NAEP results showing the
relatively low achievement of Black and Hispanic students compared to that of White
and Asian students. Lemke and associates (2005) pointed out another reading
achievement gap from the US. data of the Programme for International Student
Assessment (PISA) 2001, ﬁnding that US. reading achievement was signiﬁcantly
differentiated by socioeconomic status (SES) as well as gender and ethnic groups.

Although many have raised issues regarding gaps in reading achievement of US.
students as seen above, their focus has been mainly on gaps due to the characteristics of
the readers, which is related to the question of who reads. There seems to have been
little interest in gaps surrounding the questions about what they read and why they read.
The RAND Reading Study Group (2002) has contended that reading comprehension——
“the process of simultaneously extracting and constructing meaning through interaction
and involvement with written language” (p. ll)—consists of three elements: 1) reader,

2) text, and 3) activity or purpose for reading. Since critical views regarding the reader

19

element in various reading achievement gaps have been already expressed by many
researchers as mentioned above, this study focuses on the gaps related to the other
elements: speciﬁcally, text and purpose for reading. These two elements can be reduced

to the concept of genre in literacy (Kirk & Pearson, 1996; Pare & Smart, 1994; Swales,

1990)

In the Progress in International Reading Literacy Study (PIRLS), which is an
international comparative study of fourth graders’ reading performance, students have
been tested with texts of different genres and, more speciﬁcally, their reading scores
have been reported in two different reading purposes: literary reading and informational
reading. According to reports ﬁ'om PIRLS (Mullis, Martin, Gonzalez, & Kennedy, 2003;
Mullis, Martin, Kennedy, & Foy, 2007), US. fourth-grade students showed the largest
performance gap between these two reading purposes favoring literary reading over
informational reading among the participating countries in 2001 and the gap still existed
in 2006, although the size of the gap was smaller than before.

So why is US. informational reading performance relatively lower than its
literary reading performance? And what might be done in order to improve children’s
performance in different types of reading? These are questions with serious implications
because, in modern society, it is becoming increasingly important to have good
informational reading abilities (Benson, 2002; Doyle, 1994; Eisenberg, Lowe., & Spitzer,
2004). Moreover, it is noteworthy that the subjects of PIRLS are students in the fourth
grade, a time when students are believed to experience the so-called fourth-grade slump
(Hirsch, 2003). Some researchers have argued that the fourth-grade slump may be linked

to the relative expansion of informational text at that grade (Chall, Jacobs, & Baldwin,

20

1990). Other studies have shown that US children in primary grades have very little
experience with informational text (Duke, 2000; Pressley, Rankin, & Yokoi, 1996; Yopp
& Yopp, 2006). Accordingly, there may be some connections among large reading
performance gaps between different genres, students’ experience with texts of different
genres, and their academic difﬁculties at the fourth grade. This provides a compelling
argument for the importance of studying the predictors of student reading performance
in different text genres.

Theoretical & Empirical Background
Genre

A discussion of the second and third elements in the RAND Reading Study Group
(2002) model — text and activity/purpose — can be found in genre theory, which
emphasizes the relationship between language and its social context (Halliday, 1978).
From the perspective of genre theory, our oral and written language is the result of the
social situation in which we are communicating. Thus, people use different types of
language in different social contexts in which they have different purposes. In short,
context, purpose, and type of communication build the basis of genre theory
(Stamboltzis & Pumfrey, 2000).

The Oxford English Dictionary deﬁnes genre as “a particular style or category of
works of art; especially a type of literary work characterized by a particular form, style,
or purpose.” In the ﬁeld of literacy, Duke and Purcell-Gates (2003) deﬁned genre as
“patterns in the situations in which a text is used and patterns in the features of that text
including its language, format, structure and content” (p. 31). They further deﬁned it as a

set of identiﬁable linguistic forms that are dynamic and ﬂuid to serve different social

21

purposes that are situated within socio-cultural contexts (Purcell-Gates, Duke, &
Martineau, 2007). Therefore, genre is not about independent text structures but about the
interactions between texts we use and social contexts in which we live (Littlefair, 1992).

In general, genre is considered to include two aspects of text: the purposes for
which it is written (or spoken) and the features with which it is constructed (Chapman,
1999; Kirk & Pearson, 1996). According to Littlefair (1991), the purpose of text is a
primary determinant of genre as authors write with different purposes such as promoting
particular values through a story, conveying information, entertaining with a poem,
teaching someone to read, and so on. These different purposes, in turn, have an effect on
the second aspect of genre, that is, the text features, which refer to the ways in which
words, sentences, graphics and other text elements are patterned and organized. The
same applies to reading. A reader often picks up a certain text with a certain purpose,
and the texts chosen for different purposes usually contain different features.

Genre can help us better understand the complex relationships among texts,
readers, and authors (Epstein-Jannai, 2005). It arouses expectations in readers by
organizing what is read as a signiﬁcant construct and allowing readers to anticipate what
is coming. Additionally, genre is conceptualized as a framework one can use to
understand and interpret a text. That is, it acts as a cultural ﬁ‘amework that shapes
meaning and fulﬁlls social and cultural goals while enabling an examination of the
possible ways of creating meaning as a kind of social exchange (Cope & Kalantzis,
1993)

Genre Terminology

22

In the ﬁeld of literacy, many different terms have been used to refer to speciﬁc
text genres and there is no complete consensus on the deﬁnitions of these terms.
Nevertheless, most would agree that texts can be categorized broadly into two types:
ﬁction and nonﬁction (e.g., Cooper, 1995; Harkrader & Moore, 1997; MacLean &
Chapman, 1989). This distinction was initially used in bibliographic classiﬁcation and
subject analysis for librarians in the early 1900s and, since then, has remained almost
unquestioned in discussion of text genres (Beghtol, 2001). Another popular distinction is
narrative and non-narrative (e. g., Caswell & Duke, 1998; Georgakopulou & Goutsos,
2000), narrative texts being those that involve a series of events constrained by the
temporality and the causality of the successive action and non-narrative texts being those
that do not (Labov, 1972). These above-mentioned distinctions are related to different
aspects of text. That is, the ﬁction versus nonﬁction distinction deals with the content of
text whereas the narrative versus non-narrative distinction concerns the form of text. In
fact, there can be narrative ﬁction, narrative nonﬁction, non-narrative ﬁction, or non-
narrative nonﬁction by these dimensions of text genre.

Many other experts in literacy have differentiated text types or genres using other
terminologies. The distinction between narrative texts and expository texts is another
popular categorization in use (e.g., Abadiano & Turner, 2002; Saenz & Fuchs, 2002).
Saenz and Fuchs (2002) deﬁned narrative texts as stories written to entertain while
expository texts as materials written to communicate information to help readers learn
something new. There were also other categorizations with more than two genres. For

example, Littlefair (1991) categorized four different text types including literary,

23

expository, procedural and reference, and Arnold (1992) suggested three different text
genres including narrative, autobiographical, and informational.

In recent years, narrative texts and informational texts have become a popular
distinction between genres (e.g., Duke, 2000; Kletzien & Dreher, 2004), although this is
similar to the distinction between narrative texts and expository texts. The contrast
between narrative and informational texts can be difﬁcult to determine in many instances,
because those two terms refer somewhat to different text properties. The word
informational is related to the content or purpose of text, while narrative often refers to
the style of text. Thus, as Duke (2000) indicated, we can ﬁnd informational texts which
have the styles of narrative texts and poems, so-called narrative-informational texts and
informational-poetic texts. To avoid this confusion, it might be more advisable to use the
terms literary texts and informational texts, both of which refer to the purposes of
reading or purposes of text. This differentiation between literary and informational
reading is being used in large-scale literacy assessments such as NAEP and PIRLS
(Campbell et al., 2001; National Assessment Governing Board, 2006). These various
ways of categorization among text genres in literacy, however, share many common
features, in that all try to contrast something more informative and/or expository with
narrative and/or literary texts.

Several researchers have provided evidence that different text genres affect the
process of reading differently. Kirk and Pearson (1996), for instance, conducted a small
experimental study with 20 children in ﬁrst and second grade, and found that children
read a narrative text with greater accuracy and showing more predictions when they

were asked to suggest how the passage might end, while children read an informational

24

text with a higher level of engagement in its meaning in that they made more meaningful
substitutions for unknown words when reading an informational text. Wolfe (2005)
investigated how text genres moderate the effect of semantic association on memory in
reading and found that semantic association played a larger role in memory with
expository texts than in memory with narrative texts. Kucan and Beck (1996) studied the
differential effects of narrative and expository texts on the reading process with fourth
graders. They found that readers made more inferences, predictions, and interpretations
to synthesize and integrate whole text information by extracting important ideas when
reading narrative texts, whereas they drew more on personal knowledge and experience
focusing on local text information when reading expository texts. These and other data
offer a glimpse of genre effects on the reading process and thus imply that we should
consider text genre as an important factor when we teach children to read. All genres
have something to offer the developing reader (Stamboltzis & Pumfrey, 2000). Thus,
each of them can be used to develop particular aspects of reading in literacy education.
U.S. Performance in Informational Reading

Informational reading is getting more attention from educators and researchers,
because contemporary society is attaching more importance to the ability to ﬁnd,
understand, and manage information from text in this information-rich environment. In
her discussion of the effect of genre on literacy, Duke (2000) has termed knowledge of
socially-valued genres semiotic capital, a form of cultural capital. This term denotes that
ﬂuency in a type of discourse valued in a particular social context can be a signiﬁcant
power. In turn, this capital affects the quality of one’s life in communities, schools, and

workplaces (Bourdieu, 1991; New London Group, 1996), as do other types of capitals

25

such as cultural capital and economic capital. However, there has been relatively little
large-scale research evidence on how this new semiotic capital, that is, informational
reading, is distributed to children in our society.

As noted earlier, PIRLS is one of the few large-scale literacy studies from which
students’ reading performance in literary and informational reading can be seen
separately. When comparing the average informational reading score of each country
with its average literary reading score in PIRLS, many countries performed relatively
better or worse on one type of reading than on the other. The results ﬁ'om PIRLS 2001
and 2006 indicate that US. fourth grade students read better than international averages
with both literary and informational texts (Martin, Gonzalez, & Kennedy, 2003; Mullis,
Martin, Kennedy, & F oy, 2007). The average overall reading scores of US. students
were 542 in 2001, the 9th highest among 35 countries or jurisdictions, and 540 in 2006,
the 18th highest among 45. However, there are notable differences between informational
and literary reading achievement in the US The average reading scores of US students
on the two genres had a signiﬁcant gap favoring literary over informational reading.
More speciﬁcally, in 2001 the average reading scores of US. students were 550 (4th
highest) in literary reading and 533 (13*) in informational reading. The score gap
between two genres was 17, which is the largest gap in favor of literary reading among
35 participating countries or jurisdictions in 2001. This gap narrowed in 2006, with 541
(18*) in literary reading and 537 (19th) in informational reading, but US. fourth graders
still scored lower in informational reading than in literary reading at the .05 level of
statistical signiﬁcance. (According to PIRLS, the average US. scores in overall, literary,

and informational reading did not change between 2001 and 2006 in a statistically

26

signiﬁcant manner, although there was a relatively large drop in the literary reading
score and rank.) This pattern was consistent across genders, with both girls and boys
scoring lower on informational reading than literary reading (and girls scoring
signiﬁcantly higher than boys in both types of reading; see Table 1.1).
Factors Related to (1.5. Performance in Informational Reading

So why do US. students struggle with informational text? And under what
conditions do they have greater success with such texts? The answers to these questions
are not yet readily apparent. One possibility is relative inattention to informational text
in the primary grades. In the United States, a large proportion of the literacy experience
children have in and outside of school in their early years is devoted to narrative and
literary text genres (Caswell & Duke, 1998). Yopp and Yopp (2006) analyzed read-
aloud titles at school and in the homes of children in early grades, and their ﬁndings
suggest that children have far less exposure to informational texts (less than 10% of
read-alouds) than narrative texts (more than 75% of read-alouds). Pressley, Rankin, and
Yokoi (1996) surveyed primary-grade teachers and found that only 6% of the reading in
their classrooms was of expository genres while 73% of the reading was of literary and
narrative genres. Duke (2000) also showed that informational texts were scarce in ﬁrst-
grade classrooms, particularly in classrooms in low-SES settings. This study supports the
argument that semiotic capital is being distributed unequally to children depending on
their socio-economic status.

Reporting the ﬁndings ﬁ'om a three-year longitudinal study conducted with
children in the United Kingdom from age 8 to 13, Chapman (1987) suggested that many

primary-grade children (especially poor readers) would have reading problems later in

27

their upper grades when they meet texts of less familiar genres. Chall, Jacobs, and
Baldwin (1990) also suggested, as mentioned earlier, that the fourth-grade slump in
children’s literacy achievement might come from their difﬁculties with informational
text reading. Thus, lack of experience with informational texts in early grades has been
identiﬁed as a likely negative inﬂuence on children’s literacy development.

In spite of the importance of experiencing various types of reading in early grades,
why is there an imbalance between genres in primary classrooms, i.e. why is there the
scarcity of informational texts compared to the prevalence of literary texts? The lack of
informational reading in the early grade classrooms may depend on widespread, but
largely unfounded, beliefs that young children cannot handle informational texts, do not
like informational texts, and should ﬁrst learn to read and then read to learn (Duke,
Bennett-Armistead, & Roberts, 2003). These beliefs, however, have little research
evidence to support them. In contrast, researchers have recently reported some evidence
that refute those unsupported beliefs. Hall, Sabey, and McClellan (2005), for instance,
investigated the effectiveness of an instructional program designed to teach reading
comprehension with expository texts in second-grade classrooms and showed that young
children were able to be taught successﬁrlly with expository texts to improve their
reading comprehension. Reutzel, Smith, and F awson (2005) also found that it was
beneﬁcial to teach second-grade students comprehension strategies using science
informational texts. In her dissertation research, Jacobs (2004) conducted an intervention
study with at-risk ﬁrst grade students in low-SES, low performing urban schools. She
provided them with 18 twenty-minute lessons designed to explicitly teach precursory

skills for expository text comprehension. Children in treatment groups performed better

28

than children in control group on several reading measures. Mohr (2006) asked ﬁrst-
grade students to select one book of their preference from a set of books which were
different in their attributes such as genres, genders, ethnic groups, and so on. The
overwhelming majority (84%) of the students selected nonﬁction books rather than
ﬁction books, and almost a half of the students (46%) selected informational books in
particular. These and other studies suggest that children in primary grades can handle
informational texts successfully through adequate instruction and that many actually
have positive attitudes toward informational texts.

However, there is still a lack of empirical evidence about what kinds of literacy
experiences are differentially related to student performance in different types of
reading: informational reading and literary reading. Most research on predictors of
reading achievement has not distinguished different types of reading, even in the rare
studies that have assessed reading performance separately for different types of reading
(e.g., NAEP; Perie, Moran, & Lutkus, 2005).

Purposes of the Study

The main purpose of this study is to explore various predictors of student
performance in different types of reading to ﬁnd out how those variables differentially
predict informational reading performance and literary reading performance. To do this,
the data from PIRLS 2006 were analyzed to address the following questions. First, what
predicts country-level performance in different types of reading? Second, within the US,
what kinds of student-level and classroom-level variables are related to student
performance in informational reading and literary reading. Third, are there any

similarities or differences between the US, a country with a performance gap favoring

29

literary reading over informational reading, and other countries with different country-
level patterns in two types of reading performance. Looking at the data through this lens
will provide explanations about the relatively lower performance of US. children in
informational reading over literary reading and will suggest possible ways in which
children’s literacy in different genres could be improved—avenues to investigate in
future research.
Methods

The design of this study is a secondary analysis, the reanalysis of existing data
(Sales, Lichtenwalter, & Fevola, 2006), using the data collected for PIRLS in 2006. To
explore genre-related predictors, various literacy-related information from the PIRLS
background questionnaires was collected and scaled, if necessary, for further analysis.
Next, by using ordinary least square multiple regression, country-level predictors were
explored in relation to country-level average reading performance with different genres
and country-level percentage of students with a large gap between informational reading
and literary reading. Then, by using hierarchical linear modeling with student-level and
classroom variables, US data were analyzed to investigate signiﬁcant predictors of US.
student performance on different types of reading. Finally, the data from selected
countries were analyzed by the same analysis method and then compared with the results
from the analysis of US. data.
Progress in International Reading Literacy Study (PIRLS)

The PIRLS is a 5-year cycle of international comparative reading assessment
which has been conducted under the supervision of the International Association for the

Evaluation of Educational Achievement (IEA). It assesses the reading performance of

30

children in the upper grade of the two adjacent grades that contain the largest portion of
9-year-olds at the time of testing. These children were typically in the fourth grade and
the average age was around 10 in most countries including the United States (Mullis,
Martin, Gonzalez, & Kennedy, 2003). The ﬁrst data collection took place in 2001 with
subjects from 35 countries or jurisdictions and the second in 2006 with subjects ﬁom 45
countries. PIRLS not only assessed children’s reading performance but also obtained a
variety of literacy-related information, such as reading motivation, out-of-school literacy
experience, classroom literacy instruction, school culture, and national literacy
curriculum, from children, parents, teachers, school administrators, and national
coordinators of the assessment.

PIRLS deﬁned reading as “the ability to understand and use those written
language forms required by society and/or valued by the individual. Young readers can
construct meaning from various texts. They read to learn, to participate in communities
of readers in school and everyday life, and for enjoyment” (Mullis, Kennedy, Martin, &
Sainsbury, 2006, p. 3). Based on this deﬁnition, PIRLS tried to build a thorough and
theoretically cohesive framework of reading in order to develop a reading assessment
with strong validity. In general, the framework consisted of two broad aspects of reading
comprehension: processes of comprehension and purposes of reading. Processes of
comprehension were reﬂected in the assessment through four different types of
comprehension questions—questions that require students to: (a) focus on and retrieve
explicitly stated information; (b) make straightforward inferences; (c) interpret and
integrate ideas and information, and (d) examine and evaluate content, language, and

textual elements. Purposes for reading were reﬂected in the types of reading materials,

31

in other words, text genres, used in the assessment. PIRLS focused on two reading
purposes that account for most of the reading done by young students both in and out of
school: (a) reading for literary experience and (b) reading to acquire and use information
(Campbell, Kelly, Mullis, Martin, & Sainsbury, 2001). These two broad purposes for
reading are associated with certain text genres that bring different reading experiences.
When reading for a literary purpose, a reader becomes involved in imagined events,
settings, actions, consequences, characters, atmospheres, feelings, and ideas, and the
reader brings his or her own experiences, feelings, appreciation of languages, and
knowledge of literary forms to the text, which is usually narrative ﬁction. On the other
hand, when reading for information, a reader engages informational or expository texts
not with an imagined world but with a real universe.
Measures and Variables

The reading assessment of PIRLS 2006 was comprised of 10 different texts, ﬁve
for each of the two reading purposes. Each text was accompanied by approximately 12
test items with half in the multiple-choice format and half in the constructed-response
format. Students participating in PIRLS received one of the 12 test booklets, each of
which included 2 of the 10 prepared texts. Their reading performance scores were scaled
for overall reading, literary reading, and informational reading separately with an
international mean of 500 and standard deviation of 100 using the Item Response Theory.
The median Cronbach’s alpha reliability coefﬁcient of this reading assessment across the
booklets was .88 on average across all countries (Mullis, Martin, Kennedy, & Foy, 2007).

To identify the factors related to children’s reading performance, PIRLS

developed ﬁve background questionnaires: Student Questionnaire, Home Survey,

32

Teacher Questionnaire, School Questionnaire, and Curriculum Questionnaire. The
Student Questionnaire asked participating students about their general demographic
information, educational resources, literacy-related activities in and outside of school,
and reading motivation. The Home Survey asked a parent of each participating student
about his or her literacy-fostering activities with child, home educational resources, and
other home literacy information. This parent questionnaire was not administered in the
US. and thus was excluded ﬁ'om the analysis in this study. The reading teachers of
participating students completed the Teacher Questionnaire about the teacher’s
demographic information, class characteristics, instructional materials and technologies,
instructional strategies and activities, homework assignments, and so on. The principals
of the sampled schools completed the School Questionnaire about school characteristics,
school policy and curriculum, school environment and resources, literacy resources,
community relations, and school climate. This principal questionnaire was excluded
from the within-country analysis of this study because the information ﬁ'om the principal
questionnaire was less closely related to the focus of this study. The national research
coordinators of the participating countries completed the Curriculum Questionnaire that
asked about the nature of the development and implementation of a nationally deﬁned
reading curriculum in addition to national demographics, emphasis on literacy,
governance and organization of education system, curriculum characteristics, and so on.
Among the information collected ﬁom the PIRLS questionnaires, of particular
interest in this study are the variables related to the development of children’s general
and genre-speciﬁc literacy. The country-level variables of this study included not only

the country-level information from the Curriculum Questionnaire but also the national

33

average values or percentages of students providing particular responses on other
questionnaires. The within-country variables consisted of the information obtained ﬁom
the Student Questionnaire and the Teacher Questionnaire, because the information in
these two questionnaires best represented students’ literacy-related activities and
resources in and out of classroom. The variables included for the analysis in this study
were selected on the basis of the preliminary inspection of correlations between the
variables in PIRLS and student performance and on the basis of ﬁndings from prior
large-scale research such as PIRLS, PISA, and NAEP that studied contextual predictors
of reading performance.
Participants

A total of 45 countries or jurisdictions participated in PIRLS in 2006 (see

http://pi1‘lS.bC.Cdu/pirls?.006/ for more details). In each country, a representative sample

 

of students was selected using a two-stage stratiﬁed cluster sample design, with schools
as the ﬁrst stage and one or two classes within each sampled school as the second. In the
U.S., a total of 5,190 fourth-grade students from 253 classrooms ﬁ'om 183 schools
participated in the ﬁnal sample in 2006. Preliminary analyses were conducted to identify
signiﬁcant variables associated with children’s reading performance with the country-
level data across the countries and also with the class- and student-level data of the
United States.

After arranging a set of key variables, country-level analyses were conducted
using the data of 41 out of 45 countries or jurisdictions in PIRLS. Four countries—Hong
Kong, Kuwait, Luxembourg, and Morocco—were excluded from the analyses because

of substantial country-level missing information. Then within-country data were

34

analyzed, with listwise deletion of missing values. In addition to the U.S., data were
analyzed for three other countries——Denmark, England, and Taiwan—whose average
overall reading scores were not statistically different from the US. score. Those three
countries, however, were different in their performance patterns between literary and
informational reading. Like the U.S., Denmark had a statistically signiﬁcant gap
favoring literary over informational reading, whereas Taiwan had a gap favoring
informational over literary reading. The gap was not statistically signiﬁcant in England.
More details about these countries will be discussed later in this article.

Analysis Procedures

The main analytic methods of this study were multiple regression for the
country-level analysis and the two-level hierarchical linear modeling for the within-
country analysis. As mentioned above, PIRLS collected the data on multilevel bases and
the current study focused on the literacy-related information from students and teachers.
Because the actual sampling units of PIRLS were classes and students within sampled
classes participated in the study, the data had a hierarchical structure with students
nested within classes. The hierarchical linear model is a recommended statistical method
when analyzing the data with this nested structure (Raudenbush & Bryk, 2002). SPSS
15.0 and HLM 6 were used as the tools for the statistical analyses of this study.

Many variables in this study were scaled by the Rasch model (Rasch, 1980;
Wright, 1997) using WINSTEPS 3.63.0, when the variables were constructed ﬁom
multiple questions addressing the same construct. The classical techniques of the
variable construction by summing or averaging the responses have several limitations

(Bond & Fox, 2007; Smith, 2000). First, these variables are usually measured on the

35

ordinal scale, which is weaker than the interval scale in its interpretation and statistical
application. In many cases, the response options in the PIRLS questionnaires were
presented using the Likert scale: for example, 1) Every day or almost everyday, 2) Once
or twice a week, 3) Once or twice a month, and 4) Never or almost never. This Likert
scale itself is not an interval but an ordinal scale and, when summing or averaging this
kind of responses ﬁom several questions, it is difficult to interpret what the values mean
practically. The second limitation of the classical techniques is that they cannot consider
the characteristics of persons and the characteristics of items simultaneously. When the
responses of multiple questions are summed or averaged for one variable, it is assumed
that each question contributes equally to the variable, which is not always true. In many
cases, some questions among them contribute more to the variable, while the others
contribute relatively less. The Rasch model can overcome these limitations. Thus, using
the Rasch model, the variables were standardized with the person mean of 0 and the
person standard deviation of l with non-extreme responses (Bond & Fox, 2007; Smith,
2000).
Results

In order to investigate the predictors associated with US. students’ performance
gap favoring literary reading over informational reading, various country-level variables
were explored in relation to children’s performance in informational and literary reading
and the gap between the two across nations using the data from PIRLS 2006. Then, these
variables were more closely examined with US. data to study their relations with US.
students’ reading performance in different genres. The analyses were focused especially

on opportunities for literacy learning afforded to children in and outside of classrooms.

36

Finally, the analysis results of the data from several selected countries were compared in
detail with those ﬁom the US. to see whether there are any similarities or differences
among the countries in predictors of student achievement in informational and literary
reading.

Predictors of the Country-level Reading Performance

The country-level data were analyzed ﬁrst, in order to explore the predictors and
patterns of literacy performance on different types of reading across 41 participating
countries or jurisdictions in PIRLS 2006. This analysis focused on ﬁve country-level
outcome variables: (a) informational reading score, (b) literary reading score, (c) overall
reading score, ((1) the percentage of students with a large gap favoring literary over
informational reading, and (e) the percentage of students with a large gap favoring
informational over literary reading. In this study, the ‘large gap’ between informational
and literary reading was deﬁned operationally as above or below one standard deviation
of score difference between those two types of reading (difference = ‘informational
reading score’ minus ‘literary reading score’, international mean = 1.06 , SD. = 32.27).
Detailed descriptions of all predictor variables are presented in Appendix A. Table 1.2
presents the means, standard deviations, and intercorrelations of the country-level
variables of interest.

Early emphasis—before fourth grade—of school curriculum on genre features
such as text style and structure predicted positively the country-level informational
reading performance as well as literary reading performance. Meanwhile, at fourth grade
neither teaching with informational texts at least weekly nor teaching text style/ structure

at least weekly signiﬁcantly predicted any types of country-level reading performance.

37

Regression analyses in Table 1.3 indicate that the higher national average scores on
informational, literary, and overall reading were all predicted by lower infant mortality
rate, less formal reading instructional time at the fourth grade, more students taught with
literary texts at least weekly, and more students in schools with a major emphasis of
reading curriculum on text style and structure before the fourth grade. In addition, a
country’s average class size negatively predicted its literary reading score.

In relation to the reading score gap between different genres, the percentage of
students with a large score gap favoring informational reading over literary reading
correlated with several demographic variables (see Table 1.2). A country’s higher
percentage was signiﬁcantly predicted by its higher infant mortality rate and higher
average class size (see Table 1.3). In other words, more developed countries have fewer
students with a large reading performance gap favoring informational over literary
reading. According to these country-level analyses, the infant mortality rate best explains
the variance in country-level reading performance on overall, literary, and informational
reading as well as the variance in country-level percentage of students with a large
performance gap favoring informational reading over literary reading. This is not
surprising in that infant mortality rate has been considered as one of the best predictors
of state failure and life quality (King & Zeng, 2001). The percentage of students with a
large score gap favoring literary reading, however, showed no signiﬁcant correlations
with any variables of interest and its regression model was not signiﬁcant.

Country-level reading curriculum emphasis on different reading purposes at
fourth grade is related to country-level reading performance. A country with more

emphasis of their national reading curriculum on one reading purpose scored higher on

38

that type of reading performance as well as on the other type of reading performance.
Table 1.4 shows the means and standard deviations of the cOuntry-level reading
outcomes according to the degrees of emphasis in the national reading curriculum at the
fourth grade on each of the two reading purposes, that is, informational reading and
literary reading. Among 41 countries or jurisdictions, 25 responded that their fourth
grade curriculum had strong emphases on both informational reading and literary
reading. Although the mean differences were not analyzed statistically due to the small
number of the participating countries, visual inspection of results suggested that the
countries with more emphasis on each of the reading purposes had higher reading
performance scores than those with less emphasis. This tendency was also more salient
for the emphasis of informational reading purpose than for literary reading purpose. For
example, the score differences in reading performance—informational, literary, and
overall—between the countries with strong emphasis and the countries with moderate
emphasis was about 20 points for the emphasis on informational reading purpose but
less than 7 points for the emphasis on literary reading purpose.

The relationships were less clear, however, between the national curriculum
emphasis on reading purposes and the large gaps in reading performance between two
types of reading. Table 1.4 shows that the countries with major emphasis on each of the
two reading purposes have smaller percentages of students with a large performance gap
favoring literary reading over informational reading than the countries with some
emphasis. On the contrary, the countries with major emphasis of their national fourth-
grade reading curriculum on each of the two reading purposes have slightly larger

percentages of students with a large performance favoring informational reading over

39

literary reading than the countries with some emphasis on average. No statistical mean
differences were examined because of the small number of countries in each emphasis
category, and thus conclusions are based only on visual inspection.

Predictors of (1.5. Students ’ Literacy Performance

In order to explore the predictors and patterns of students’ reading performance
in different genres in more detail, this study looked into the US. data focusing on
classroom-level and student-level variables. The classroom-level variables included
students’ literacy experience and resources at school, while the student-level variables
explained their literacy experience and resources outside of school (see Appendix A for
the variable descriptions). After missing data deletion, the data from 4,508 students in
249 US. classrooms were analyzed. Table 1.5 and Table 1.6 present the means, standard
deviations, and intercorrelations of the variables on the classroom-level from the teacher
questionnaire and on the student-level ﬁ'om the student questionnaire respectively.

On the classroom-level (see Table 1.5), the class mean overall, informational,
and literary reading scores correlated positively with the class mean SES and negatively
with the proportion of students with reading difﬁculties in class. Teachers’ computer use
for reading instruction and their instruction for informational reading had signiﬁcantly
positive correlations with the class mean informational reading score, and they did not
correlate signiﬁcantly with class mean literary reading performance. Reading homework
assignment by teachers had a signiﬁcantly negative correlation with class mean literary
reading performance. There were also more students with a large score gap between the
two genres favoring literary over informational reading in class, when the class had less

instructional time for reading and had fewer students with reading difﬁculties. These are

40

bivariate correlations, though. When class-level variables were considered together to
predict student reading performance scores, many of those variables did not have a
signiﬁcant prediction and thus were excluded from the analysis models.

On the student-level (see Table 1.6), many variables of interest correlated
signiﬁcantly with students’ reading performance scores. For example, students’ overall,
informational, literary reading scores had positive correlations with gender (as being
girls), books at home, student SES, attitudes toward reading, and reading self-concept,
whereas they had negative correlations with students’ reported home/ school computer
use and abnormal amount of reading (a great deal of reading and no reading) outside of
school. In addition, students’ informational reading outside of school correlated
negatively with their reading scores, while their literary reading outside of school
correlated positively with the reading scores. Finally, the students with a large gap
favoring literary reading were more likely to be better readers, while those with a large
gap favoring informational reading were more likely to have lower reading performance
scores. These student-level variables were still signiﬁcant predictors of reading
performance even when considered together in the hierarchical linear models.

Two-level HLM analyses were conducted to explore the signiﬁcant predictors of
students’ informational reading performance and their literary reading performance
respectively. When predicting informational reading performance scores, 26% of the
variance was between classrooms (intraclass correlation = .26). The student-level
variables of this analysis in Model I explained 26% of the within-class variance in
students’ informational reading performance, while the classroom-level variables

accounted for 32 % of the between-class variance in students’ informational reading

41

scores (see Table 1.7). When predicting literary reading performance scores, 22 % of the
variance existed between classrooms (intraclass correlation = .22). Thus, there was 4
percentage points more between-class variance in informational reading performance
than in literary reading performance. When looking at literary reading performance, 27%
of the within-class variance was explained by the student-level variables in Model 1,
while 31% of the between-class variance was accounted for by the classroom-level
variables in Model 2 (see Table 1.8).

Reading with informational texts played different roles in the prediction of
student reading performance depending on whether it is led by teachers in classrooms or
is self-initiated by students out of school. Teachers’ use of informational texts for
reading instruction positively predicted the class mean informational reading
performance. Speciﬁcally, a class in which the teacher used informational texts for
reading instruction one standard deviation more often than other teachers scored 3.66
points higher in informational reading with other variables held constant (Beta = 3.66, p
< .05; see Table 1.7). However, any genre-related classroom variables including
informational reading instruction and literary reading instruction did not predict the class
mean literary reading performance signiﬁcantly as shown in Table 1.8 (p > .10). Among
the genre related student-level variables, students’ literary reading outside of school
positively predicted their reading performance in both types of reading (Beta = 4.45, p
< .001 for informational reading; Beta=5.40, p < .001 for literary reading), whereas
students’ informational reading outside of schools negatively predicted their reading

performance (Beta = -7.22, p < .001 for informational reading; Beta = -8.65, p < .001 for

42

literary reading). This last ﬁnding is surprising and will be discussed further in the
Discussion section below.

Gender has been considered to be an important predictor of reading achievement
and so was it in this study. According to the report ﬁ'om PIRLS 2006, the average
reading scores of girls were higher than those of boys in both genres in most countries
including the U.S., although in a few countries there was no statistically signiﬁcant
difference between girls and boys in reading performance. Boys did not outperform girls
on average in both literary and informational reading in any participating countries or
jurisdictions (Mullis, Martin, Kennedy, & Foy, 2007). When other variables were
considered, however, the relationship between gender and reading performance was only
marginal in the United States as Table 1.7 and Table 1.8 show. Notably, the gender
effect was less salient in informational reading performance, in which the gender
coefﬁcient was not statistically signiﬁcant (Beta = 2.55, p > .10; see Model 1 of Table
1.7), than in literary reading, in which it was (Beta = 4.45, p < .10; see Model 1 of
Tablel .8). However, the gender variable interacted with the class mean SES in
predicting students’ informational reading performance (Beta = -7.37, p < .10; see
Model 2 of Table 1.7). Speciﬁcally, girls outperformed boys in informational reading in
the classrooms with low to medium average SES while boys outperformed girls in those
classrooms with high average SES. This interaction was not statistically signiﬁcant in
relation to literary reading performance (p > .10).

Another variable that differentially predict informational reading performance
and literary reading performance in the US. was whether students reported that they did

not read at all outside of school. Students who reported no reading outside of school

43

scored signiﬁcantly lower in informational reading (Beta = -6.07, p < .01) than those
who reported at least some reading outside of school, but it was not signiﬁcant in literary
reading (Beta = -3.95, p > .10). This may imply that student performance on
informational reading is more sensitive to the amount of out-of-school reading than
performance on literary reading.

The remaining variables in Table 1.7 and Table 1.8 were related to student
reading performance in the same way for informational reading and literary reading.
Positive predictors of literary and informational reading performance were class mean
SES on class-level, and books at home, student SES, attitudes toward reading, and
reading self-concept on student-level. Negative predictors were proportion of students
with reading difficulties on class-level, and abnormally large amount of reading outside
of school students reported and the frequency of home/ school computer use on student-
level.

Because most genre-related variables were standardized with a mean of 0 and
standard deviation of 1 by the Rasch scaling, they do not provide an actual amount or
frequency of reading teachers and students reported. They only provide the relative
amount or frequency of reading. Table 1.9 shows the means and standard deviations of
US. average raw responses in major variables for the analyses above. Classroom
teachers answered that they used different types of informational texts and literary texts
more often than once a month but less than once a week on average for their reading
instruction (see Table 1.9 for more detail). Students also reported that they read various
types of informational and literary texts outside of school and use the computer in and

out of school more often than once a month but less than once a week. According to the

standard deviations in this table, there was slightly more variation for informational
reading than for literary reading across classrooms, while more variation was found for
literary reading than for informational reading outside of the classroom. Students’
reported average reading amount outside of school was between one hour and three
hours per day, but had a relatively large variation. However, these means and standard
deviations should be interpreted with caution, because the response options were not
interval scales but ordinal scales.

In sum, as shown in Table 1.10, US. student informational reading performance
was positively predicted by class mean SES, more reading instruction with informational
texts, gender (speciﬁcally, girls in classroom with low to medium mean SES), books at
home, student SES, attitudes toward reading, reading self-concept, and the ﬁ'equency of
literary reading outside of school, when other predictors are controlled. Most of these
predictors also predicted student literary reading performance in the same pattern except
two variables: more instruction with informational texts and gender. That is, students
taught by teachers who use more informational texts had higher informational reading
performance scores on average, whereas any class-level genre variables such as
informational reading instruction and literary reading instruction did not predict student
literary reading performance signiﬁcantly. Also, boys were found to be better readers in
informational reading if they were in high SES classrooms, while girls were always
better readers in literary reading regardless of class mean SES.

Comparisons with Other Countries
Data from three additionally selected countries were analyzed in the same way

that the US. data were analyzed in order to explore any similarities or differences

45

among those countries. In PIRLS 2006, there were 12 other countries whose average
overall reading scores were not statistically different from the US. overall reading
performance score. For reasons explained in the Methods section, Denmark (with higher
literary reading on average), England (with no signiﬁcant difference between
informational and literary reading on average), and Taiwan (with higher informational
reading on average) were selected out of those 12 countries according to their reading
performance patterns between literary and informational reading. Table 1.11 presents
basic descriptive statistics about these countries including the United States. Compared
to other countries in this table, the US. had more formal reading instruction hours per
week, more students in poverty schools, and more teaching of informational and literary
texts.

Among the selected countries only in Denmark, a country with a signiﬁcant gap
favoring literary reading over informational reading on average, was the average overall
reading performance score lower for the students with a large score gap favoring literary
reading than for the students with a large score gap favoring informational reading. In
three of four countries, students with a large gap favoring informational reading over
literary reading had lower overall reading scores. Table 1.12 shows students’ average
performance scores in informational, literary, and overall reading in each country by the
size of gaps between informational and literary performance. When looking at literary
reading scores by the groups, students who had a large gap favoring informational
reading had lower average literary reading scores in all the countries than students who
had a small gap or a large gap favoring literary reading. For informational reading scores,

however, the students with a large gap favoring informational reading had the lowest

46

average informational reading scores in the US and Taiwan, while the students with a
large gap favoring literary reading had the lowest average informational scores in
Denmark and England. As seen in Table 1.12, results failed to provide a systematic
relationship between country-level reading performance gap and within-country
grouping by informational-literacy performance gap.

Several key ﬁndings are worth noticing when the results from separate HLM
analyses of within-country data for each country are compared with those from the
analysis of the US. data. Appendix B, Appendix C, and Appendix D present these
results ﬁ'om Denmark, England, and Taiwan respectively, and Table 1.13 displays the
coefﬁcients of predictors ﬁom all selected countries for comparison.

Compared to the intraclass correlations of the US. data (.26 for informational
reading performance and .22 for literary reading performance), the intraclass correlations
of other countries were small. Taiwan had the lowest intraclass correlations (.09 for both
informational and literary reading performance) and Denmark was next (.14 for both
informational and literary reading performance). The intraclass correlations of England
were .22 for informational reading performance and .21 for literary reading performance.
Therefore, the US. had the largest between-class variance in student reading
performance among the four countries, whereas Taiwan had the smallest between-class
variance in predicting student reading performance.

Only in the US. did more informational reading instruction in the classroom
better predict students’ informational reading performance (Beta = 3.66, p < .05). Any
genre related class-level variables such as informational reading instruction and literary

reading instruction did not predict student reading performance signiﬁcantly in all other

47

countries. For genre related student-level predictors, more literary reading outside of
school predicted higher performance on literary and informational reading performance
in all four countries. The coefﬁcient of literary reading out of school ranged from 2.08
(Denmark) to 7.48 (England) for informational reading performance and from 2.79
(Denmark) to 8.73 (England) for literary reading performance with one standard
deviation change in literary reading out of school. But informational reading outside of
school negatively predicted student reading performance in all four countries. The
coefﬁcient of informational reading out of school ranged from -7.22(U.S.) to -10.66
(England) for informational reading performance and ﬁ’om -8. 12 (Taiwan) to -13.49
(England) for literary reading performance with one standard deviation change in
students’ informational reading outside of school.

SES was a consistently signiﬁcant predictor of students’ performance on
informational reading and literary reading both at the student-level and at the classroom-
level in all those countries. The coefficients of student-level SES ranged ﬁ'om 2.52
(Taiwan) to 6.42 (U .S.) for informational reading performance and from 3.65 (Taiwan)
to 6.82 (Denmark) for literary reading performance. In all four countries, however,
student-level SES had a curvilinear relationship with student reading performance and
thus this relationship between SES and reading performance was especially larger for
lower SES children. The coefficients of class mean SES ranged from 8.90 (Denmark) to
28.47 (U .S.) for informational reading performance and from 8.62 (Denmark) to 28.54
(U .S.) for literary reading performance with one standard deviation change in class mean

SES.

48

When other variables were considered, student gender was a signiﬁcant predictor
of literary reading performance in the US. (Beta = 6.37, p < .05), Denmark (Beta = 8.33;
p < .01), and Taiwan (Beta = 4.49, p < .05), but not in England. Gender was also a
signiﬁcant predictor of informational reading performance in the US. (Beta = 4.27, p
< .10) and Denmark (Beta = 6.68, p < .001), but not in England and Taiwan. While the
gender variable interacted with class mean SES in predicting informational reading
performance in the U.S., in Denmark regardless of class mean SES girls performed
better on average.

Students’ reading motivation, measured by their attitudes toward reading and
their reading self-concept, was a consistent predictor of student performance on
informational and literary reading. The coefﬁcients of attitudes toward reading ranged
ﬁ'om 3.92 (England) to 6.50 (U .S.) when predicting informational reading performance
and from 2.98 (Denmark) to 8.30 (U .S) when predicting literary reading performance
scores by one standard deviation change in students’ attitudes toward reading. The
coefﬁcients of reading self-concept ranged from 8.54 (Taiwan) to 22.88 (Denmark) for
informational reading performance and ﬁom 9.42 (Taiwan) to 21.62 (England) for
literary reading performance with one standard deviation change in students’ reading
self-concept.

Among the remaining variables in the HLM analyses, abnormal amount of
reading out of school students reported (no reading or a great deal of reading) and
students’ reported home/school computer use predicted students’ informational reading
performance scores negatively (see Table 1.13). These negative relationships were also

true for literary reading performance except in the US. (insigniﬁcant coefﬁcient for no

49

reading) and in Taiwan (positive relationship below the average amount of students’
reported home/school computer use but negative above the average). The proportion of
students with reading difﬁculties in class was also a negative predictor of class average
informational and literary reading performance in all countries exceptTaiwan in which it
was a statistically insigniﬁcant predictor. More books at home, on the contrary, always
predicted better reading performance on informational reading and literary reading in all
selected countries.
Discussion

In this study, various literacy related variables in and out of classroom were
examined in relation to student reading performance in two different genres-—
informational reading and literary reading, by analyzing the data from PIRLS 2006. The
purpose was to explore whether and, if so, how those variables differentially predicted
student informational reading performance and literary reading performance in order to
understand the US. reading performance gap favoring literary reading over
informational reading. Country-level data were analyzed across 41 countries and also
within-country data were analyzed using two-level Hierarchical Linear Modeling with
students nested in classrooms. The main ﬁndings and implications are as follows.

This study strongly supports the contention that reading is genre-speciﬁc (Duke,
2005). Student reading performance and its relations to other variables showed different
patterns depending on genres--informational reading and literary reading. When looking
at student reading performance gap between informational reading and literary reading,
for example, more than 20% of students in the four selected countries of this study

showed over 30 score differences between two genres. Among those students, some had

50

a large score gap favoring literary reading over informational reading, while others had a
large gap favoring informational reading over literary reading. These different
performance patterns between different genres were also true for country-level average
reading performance. As mentioned in the beginning of this paper, the US. average
reading performance in PIRLS had the largest gap favoring literary reading over
informational reading in 2001 and this pattern was repeated in 2006 although the gap
narrowed and it was not the largest gap any more among participating countries. Other
countries also showed different patterns in reading performance between informational
reading and literary reading. In addition, cross-national and within-country analyses of
the relationships between literacy related variables and reading performance indicated
that some variables predicted literary reading performance and informational reading
performance in different ways. The variables, such as country-level class-size, classroom
informational reading instruction, and student gender, for example, predicted only
literary or informational reading, but not both in some countries.

This genre-speciﬁcity of reading has been often underrecognized. For instance,
many reading assessments widely used across schools do not specify genre and do not
separate results for different genres. Certainly, PIRLS and other studies of reading
achievement should continue to assess student reading performance in different purposes
or genres. Through this, they can provide better assessments of reading that are more
informative for children, teachers, and policy makers.

It is important to provide elementary students with more opportunities to learn
different genres for reading in school. The opportunity to learn (OTL) has been

considered to be an important indicator for interpreting student performance and for

51

evaluating the quality of education (McDonnell, 1995; Schmidt & McKnight, 1995).
According to Chapman (1987) and Chall, Jacobs, and Baldwin (1990), less opportunity
to learn with different genres in reading might cause later academic and reading
problems. The present study provided the evidence showing that more genre-speciﬁc
opportunity to read in classroom is related to higher reading performance. In the
country-level analysis, for example, country-level average performance on both literary
and informational reading was higher as more students were taught with literary texts at
least weekly in classroom and as more students were taught at the schools whose reading
curriculum put major emphasis on text style and structure before fourth grade. As well,
more emphasis of national reading curriculum on each of two different reading
purposes—reading to acquire information and reading for literary purpose—were
positively related to country-level performance on both information reading and literary
reading. These results suggest that reading curriculum and instruction should consider
provision of more opportunity to learn various genres to students for better reading
performance.

Although previous results imply that children’s experience with the genre-
speciﬁc opportunity to read is important for reading development, it seems that all
experiences are not the same but they are more likely context-dependent. For instance, as
noted earlier, this study showed that, in the U.S., fourth-grade children who get more
instruction with informational texts in the classroom had higher reading performance on
informational reading. More informational reading out of school, on the contrary, was
not related to higher reading performance of those children. In fact, they had a negative

relation (this was also true for all other selected countries). It is not clear ﬁ'om the

52

analysis of this study why more out-of-school informational reading is negatively related
to student reading performance. Among other things, the analysis in this study did not
specify information about what kinds of informational texts students were reading and
how different kinds of informational texts were differentially related to their reading
performance. This may be an important factor. In a study with ﬁfth grade students by
Anderson, Wilson, and Fielding (1988), for example, time spent in reading
newspapers/magazines and reading mail out of school, which can be considered as out-
of-school informational reading, predicted student’s reading related abilities including
comprehension, vocabulary, and speed negatively or nonsigniﬁcantly. Further analysis
will be necessary to consider more speciﬁc kinds of texts and their relations to reading
performance. Another piece of information we have ﬁ'om the PIRLS dataset that may be
relevant is that students’ out-of-school informational reading was more related to the
ﬁ'equency of computer use students reported than out-of-school literary reading. The
computer use negatively predicted student reading performance as well. This might
indicate that, although students report more out-of-school informational reading, this
does not necessarily mean they read with the types of informational texts and in the
types of ways that translate to achievement in informational reading as measured by the
PIRLS. It is noteworthy that research shows a relatively low relationship between online
reading comprehension skill and ofﬂine reading comprehension skill (Hartman, Morsink,
& Zheng, in press). Possibly, only certain types of informational text may be associated
with success on the types of informational reading required in assessments like PIRLS.
Another explanation for opposite relationships between reading and performance

depending on whether the reading happens in or outside of classroom is that a more

53

systematic approach like classroom reading instruction might be necessary for
informational reading and thus simply reading more informational text is not enough to
develop informational reading skills. Students have relatively little instruction with
informational texts in the primary grades and may not have naturally developed
appropriate skills or strategies through out-of-school reading to understand informational
texts well. Researchers have suggested that more explicit instruction of reading skills
and strategies helps struggling students to develop their reading abilities and facilitate
reading development (e.g., Pressley, 2006; Pressley & Wharton-McDonald, 1997). This
explicit instruction may be necessary for novice informational text readers.

The relationships between literary reading in and outside of school and literary
reading performance were also different according to the context of reading, whether
more literary reading happened in classroom or out of school. Although more literary
reading in the classroom was not signiﬁcantly related to higher reading performance,
more literary reading outside of school was a signiﬁcant predictor of higher reading
performance. Most possibly, this is because literary reading is common across US.
fourth-grade classrooms and the amount of literary reading is relatively consistent across
classrooms. As shown in Table 1.11, about 92% of US. students receive reading
instruction with some kinds of literary texts from their teachers at least weekly. Positive
relationships between out-of-school literary reading and student reading performance
should be understood in the context of the fact that more out-of-school literary reading
of US. students was positively related to more books at home, higher motivation, and
amount of reading outside of school. These relationships were not seen for informational

reading outside of school. In addition, the variance in literary reading outside of school

54

was larger than that of informational reading outside of school. More research is needed
to clarify this context-dependence of the relationship between different types of reading
and reading achievement.

Girls are widely lmown to be better readers than boys on average (e. g., Klecker,
2006). This female advantage in reading, however, seems to be more evident in literary
reading than in informational reading. For instance, the analysis of US. data in this
study showed that, when other predictors were controlled, reading performance of girls
were clearly better than boys in literary reading on average. In informational reading,
however, boys outperformed girls in the classrooms with high mean SES. When
comparing average reading scores of girls and boys in 45 countries or jurisdictions
participating in PIRLS 2006 (Mullis, Martin, Kennedy, & F oy, 2007), average scores of
girls were signiﬁcantly higher than those of boys in all participating countries except
only one country for literary reading. For informational reading, there was no
statistically signiﬁcant difference in performance scores between boys and girls in ﬁve
countries. Gender difference between two genres might come from differences in their
preference or choice of text. Coles and Halls (2002) asked 10- to 14-year-old U.K.
children about their book reading habits, ﬁnding that girls generally read more books
than boys but more boys described themselves as non-ﬁction readers than girls. From a
study with second through ﬁfth grade US. children, Watson (1985) also found that girls
read more books about family stories and romance while boys read more books about
science and animals. Additional studies are needed, however, especially regarding why
gender and class-SES interact for reading performance as shown in the analysis of the

US. data.

55

The results of the present study support that SES, quantity of books at home, and
reading motivation are strong predictors of reading performance, which is cOnsistent
with prior studies (e.g., Aikens & Barbarin, 2008; Weinberger, 1996; Wigﬁeld &
Guthrie, 1997). However, the ﬁnding of a negative relationship between reading
performance and student home/ school computer use is inconsistent with some past

research, though consistent with other past research. Neevdal (2007), for example,

investigated this relationship with Norwegian adolescents and found that time spent
working on the computer at home positively predicted students’ achievement in English
at school. This positive relationship was more evident for poor readers and girls. On the
other hand, O’Dwyer, Russell, Bebell, and Tucker-Seeley (2005) argued that the
relationship between computer use and achievement would be dependent on the
purposes of computer use. Their study with US. fourth grade students showed that,
when controlling for prior achievement and SES, more computer use of students at
school was positively related to higher English/language arts test scores if it was used to
edit papers but not if it was used to prepare presentation. Also, students’ recreational use
of computer at home was a negative predictor of student achievement. In the present
study, teachers’ reported frequency of computer use for reading instruction had a
positive correlation to student performance in informational reading in the U.S., though
this correlation was not statistically signiﬁcant when other variables were controlled.
The student-level variable for computer use was a Rasch scaled composite score of
students’ reported ﬁ'equency of computer/Intemet use at home and at school in order to

use it in regression analyses. Further research will be necessary to clarify this issue.

56

Another limitation is that the variables this study explored in relation to fourth
graders’ reading performance were mainly ﬁ'om the information about students’ literacy
experience at the fourth grade in and outside of classroom. This study had only a little
information available about students’ experience in earlier grades. Literacy experience in
early grades certame inﬂuences later literacy performance and literacy development of
children (e.g., Chapman, 1987; Stanovich, 1986). In order to examine the predictors of
fourth graders’ reading performance, it will be good to know what kinds of literacy
experience those students had when they were in primary classrooms. For instance, in
the U.S., informational reading is scarce in primary grades (Duke, 2000) though it is
more common in fourth grade classrooms. Thus, the difference of informational reading
experience in earlier grades might tell us more about a possible source of the difference
in fourth graders’ informational reading performance.

Finally, this study is not an experimental study but a correlational study. In this
study, there is no strong evidence for causal inference about whether or not more
instruction with texts in a speciﬁc genre actually improves reading performance in that
genre. Future research will be necessary to conﬁrm this causality and to provide policy
makers and educators with more compelling results.

In summary, this study found that genre-related activities and curriculum were
deﬁnitely related to children’s reading performance in a speciﬁc genre and across
different genres. Because of the US. performance gap favoring literary reading in
PIRLS, this study has focused more on informational reading and the results have shown
that more instruction with informational text in US. fourth grade classrooms makes a

signiﬁcant difference in student performance on informational reading. This does not

57

mean, however, that we can disregard another type of reading, that is, literary reading. In
fact, students’ literary reading out of school and literary text instruction on country-level
were signiﬁcant predictor of both literary and informational reading. In addition, student
with higher literary reading performance over informational reading were better readers
on average in three of the four selected countries. The results of this study suggest that
we should provide substantial opportunities to learn with both literary and informational
text and an early emphasis on text style or structure for all children. This may help shape

children’s reading experiences with genres they will encounter throughout their lives.

58

Table 1.1

Average Reading Scores of US. Students by Reading Purpose and Gender in PIRLS

 

 

 

 

2001 and 2006
. PIRLS 2001 PIRLS 2006
Readrng
purpose
Girls Boys All Girls Boys All
Literary 558 (4.2) 542 (4.6) 550 (3.8) 547 (3.6) 534 (4.1) 541 (3.6)

Informational 541 (4.1) 525 (4.3) 533 (3.4)

Overall 551 (3.8) 533 (4.9) 542 (3.8)

542 (3.1) 532 (4.4) 537 (3.4)

545 (3.8) 535 (4.4) 540 (3.5)

 

Note. The scores are based on the PIRLS international reports (Mullis, Martin, Gonzalez, & Kennedy,

2003; Mullis, Martin, Kennedy, & Foy, 2007).

( ) Standard errors appear in parentheses.

59

Table 1.2

Means, Standard Deviations, and Intercorrelations of Country-level Variables and

Reading Performance (N =4 1 )

 

 

 

- Zero-order correlations
Country level M SD
Variables 1 2 3 4 5
National Demographic Variables
1. Infant mortality rate 10.22 1 1.19 l
2. Net enrollment rate (%) 95.95 5.39 -57” l
3. Hours of formal reading ***
instruction 2.56 .91 .30 -.55 1
4. Average class size 24.10 4.80 .36‘. -.02 -.15 l
(5%Students 1n poverty SChOOlS 33.97 20.27 .72!!!" -.58‘1" .47“- .29 1
6. Achievement gap between .. " .. .. ......
students in low and high 33.18 28.50 .54 -.21 -.12 .63 ,5 5
poverty schools
Genre-related Variables
7' Teaching ”0mm“ “”“S 62.1] 16.21 .01 .05 .31 .02 .14
weekly (/o)
8. Teaching literary texts 9'
weekly (%) 86.51 11.25 -.01 -.14 .16 -.34 -.10
9. Teaching text style or *t " "
structure weekly (%) 52.71 22.94 .26 -,44 .50 .12 .51
10. Major emphasis on text "
style and structure before 4‘h 39.63 20.25 -.29 .32 .17 .02 -.04
grade at school (%)
1 1' Emphas‘s °“ “mus .425 1.36 -.02 .07 -.23 .26 -.10
readmg purpose
Reading Performance
. . ### it t it!
12. Informatlonal reading score 511.38 54.43 -.72 _43 -.25 -39 -.68
. . tlt it #t it.
13. therary readrng score 511.56 57.04 -_73 .44 -.24 -.45 -.69
. #tt 1* it tit
14. Overall readlng score 51 1.90 57.21 -_73 .43 -.25 -,42 -.69
15. Students with large score
gap favoring literary reading 15.06 6.35 .11 -.09 .22 -.28 .13
(%)
16. Students with large score ...... n n .. 1.
gap favoring informational 13.41 7.64 , 57 -.43 .12 , 5 5 .49

reading (%)

 

60

Table 1.2 Continued

 

 

 

Country- Zero-order correlations
level
Variables 6 7 8 9 10 1 l 12 13 14 15
National Demographic Variables
l.
2.
3.
4.
5.
6. I
Genre-related Variables
7. .07 l
a:
8. -.ll _31 l
9. .16 .13 .04 1
it

10. -.02 .42 .01 .29 1
11. .28 .17 .11 -.33’ -.23 1
Reading Performance

it t #t
12. -.50 .15 .35 -.O6 .42 -.01 1

1“ t C. lit
13. -.53 .14 .37 -.06 .43 -.05 .99 1

it t it it! *t#
14. -.52 -l4 .36 =06 .43 -.03 .99 .99 1
15. -.11 -.03 .26 .20 .03 -.22 -.12 .00 -.O6 1

*t it t tit ii
16. .45 .03 -.17 .16 -.29 .25 -,49 -.60 -55 -,44

 

# it tit
p<.05; p<.01; p<.001

61

Table 1.3
Multiple Regression Analysis Summary for Country-level Variables Predicting Country-

level Reading Performance (N =4 1 )

 

 

 

 

Informational Literary
Country-level variables B readmgcore B readmgscore
(SE) Beta t (SE) Beta t
446.56 "" 45455 at.
Constant (60.59) 7.37 (58.50) 7.77
General National Variables
. -2.55 "W -2.62 5""

1. Infant mortality rate (.57) -.52 .4_49 (.55) -.51 .4,73
2. Net enrollment rate (%) T T T I T T
3. Hours of formal reading -l7.57 * -l7.83 *
instruction (7.06) ‘29 '2.49 (682) '28 -2.62

. -2.04 -2.70 *
4. Average class srze (1.37) -.18 -l.50 (1.32) -.23 -205
5. Students in poverty schools 1 T T l 1 T
(%)
6. Achievement gap between
students in low and high I l l l T ’1
poverty schools
Genre-related Variables
7. Teaching informational -.01 -.08
texts weekly (%) (.39) "0° "03 (.37) "02 "22
8. Teaching literary texts 1.54 H 1.67 **
weekly (%) (.51) '32 3-05 (.49) '33 3-41
9. Teaching text style or .39 .39
structure weekly (%) (.28) '16 "39 (.27) '16 1'43
10. Major emphasis on textm .78 .. .87 "
style and structure before 4 ( 32) .29 2,45 ( 31) .31 2.82
grade at school(%) ' '
ll. Emphasis on various 2.18 1.43
reading purpose (4.52) '05 '48 (4.36) '03 '33

it. it.
F F(8, 32)=1 1.77 F(8, 32)=14.58
Adjusted R2 .68 .73

 

 

 

 

62

Table 1.3 Continued

 

 

 

 

 

 

o .
Overall 41:82:23: :mh % “Students “’“h
r 1' score ﬁavorin g p large score gap favoring
Country-level g l' t Elm informational reading
variables 1 erary rea g
B B B
(SE) Beta t (SE) Beta t (SE) Beta t
450.80 "W 2.94
C°“S‘am (60.80) 7-42 ' ‘ ' (10.92) '27
General National Variables
-2.65 "- .24 t
2. T T T _ _ _ T T T
-l8.23 *
3. (7.09) -.29 -257 - - - l T T
-2.47 .51 *
4, (1.37) '.2] '1.80 ' ' ' (.24) '32 2.11
5. T T T _ _ _ T T T
6. T T T _ _ _ T T T
Genre-related Variables
-.05 .05
7. (‘39) -.02 -.14 - - - ('07) .10 .69
1.65 .. .08
8. ('51) .32 3,24 - - - (.10) -.12 -.84
.39 .05
9. (.28) .16 1.40 - - - (.05) .16 1.13
.85 .. -.09
10. (.32) .30 2.67 - - - (.06) -.24 -1.55
1.80 .98
11. (4.54) .04 .40 - - - (.85) .18 1.16
"It tit
F F(8, 32)=l3.3l n.s. F(7, 33)=5.42
Adjusted R2 .71 . - .44

 

 

 

 

n.s. Not signiﬁcant statistically

T These variables were eliminated from the analysis due to multicollinearity.

It it
p<.05; p<.01;

ﬁt

up
P

< .001

63

Table 1.4
Means and Standard Deviations of Country-level Reading Performance by Country-level
Curriculum Emphasis at 4“ grade on Reading to Acquire Information and Reading for

Literary Experience (N =4 I )

 

Country-level curriculum emphasis on reading purposes

 

 

 

 

 

 

 

 

Reading performance Statistics To acquire information For literary purpose
Strong Moderate Strong Moderate
. Mean 517.38 496.63 517.44 511.52
Informational
reading score
(S.D.) (52.29) (59.66) (54.30) (40.62)
_ Mean 516.57 497.68 516.23 514.06
Literary
reading score
(S.D.) (56.19) (59.75) (58.81) (39.77)
Mean 517.53 497.21 517.39 513.25
Overall
reading score
(S.D.) (55.71) (61.27) (58.15) (41.08)
% of students with Mean 14.13 15.93 13.79 17.80
large score gap favoring
1"er reading (SD) (5.47) (7.15) (5.23) (8.53)
% of students with Mean 13.89 13.11 14.24 12.48
large score gap favoring
Informational reading (SD) (7.86) (7.28) (8.33) (6.42)
Number of countries 27 13 27 1 1

 

 

Note I . Twenty-ﬁve countries reported strong emphasis on both informational and literary reading. The
U.S. reported strong emphasis on both reading to acquire information and reading for literary purpose.
Note 2. A few countries reported little emphasis on either purpose. One country (Hungary) reported little
emphasis on reading to acquire information, while three countries (Poland, Qatar, and Belgium-French)

reported little emphasis on reading for literary purpose.

64

Table 1.5

Means, Standard Deviations, and Intercorrelations of U.S. Classroom-level Variables

and Reading Performance (24 9 classrooms)

 

 

 

Class-level Mean S D Zero-order correlations

variables (U .S.) 1 2 3 4

General Classroom Variables

1. Class mean SES .21 .47 1

2. Weekly instructional time 8.88 4.30 _.07 1

for reading

3. Proportion of students with "r *

reading difﬁculties 25 -17 -.27 .14 1

4. Reading comprehension

strategy instruction .31 1.21 -.05 .12 .10 1

5. Computer use for reading ***

instruction -.13 1.11 .11 .12 -.08 ,23

6. Autonomy support in 3-

reading instruction .02 1.04 .05 .11 -.02 .16

7. Social activities in reading -.03 1.09 -.01 .14. .12 .331."

8. Homework for reading -.06 1.17 -.16‘ .19” .10 .19"

Genre-related Classroom Variables

9. Instruction for ......

Informational reading .00 .99 -.02 .08 -.11 ,37

r1 Ensign-“Chou for literary '00 .99 .02 .14" .02 37*"

Reading Performance

11. Class informational *** it"

reading score 532.91 39.84 .56 -.05 -.43 .00
IMHO! till!

12. Class literary reading score 536.00 41.90 , 55 -.05 -_44 -.01
*1!!! #1.

13. Class overall reading score 535.27 42.07 .56 -.06 -,43 -.01

14. %ofstudents inclass with u ......

large score gap favoring 12.90 8.52 .04 -.16 -.18 -.O2

literary reading

15. % of students in class with

large score gap favoring 7.40 6.65 -.10 -.01 .09 -.05

informational reading

 

65

Table 1.5 Continued

 

Class- Zero-order correlations

 

level
Variables 5 6 7 8 9 10 11 12 13 14

 

General Classroom Variables
1 .

2.

5. 1

**

6. .20 I

lit #**

7. .29 .30 1

it

8. .03 .20 .12 1

Genre-related Classroom Variables

tit 3“ *1. ﬁ

9- .39 .28 .33 .14 1

it! ##t it. ** tit

10- .26 .31 .34 .19 .35 1

Reading Performance

* t

11. .14 .09 -.03 -.12 .13 .03 l
* it!
12. .12 .10 -.03 313 .11 .01 .99 l
t it! tit
13. .13 .10 -.03 -. 12 .12 .02 _99 .99 1
it t
14. -.09 .01 -.09 -.11 -.05 -.11 .09 .20 .14 1
it tit t# ﬁt!
15. -.02 .04 .02 .05 .05 .08 -.13 -,27 -22 -.25

 

t it it.
p<.05; p<.01; p<.001

66

Table 1.6

Means, Standard Deviations, and Intercorrelations of U.S. Student-level Variables and

Reading Performance (4,5 08 students)

 

 

 

Student-level Mean S D Zero-order correlations
variables (U .S.) ' ' 1 2 3 4 5
General Student Variables
1. Gender (O-Boy, l-Girl) .51 .50 l
2. Books at home (O-Few Mt
books, l-Many books) '65 '48 '06 l
3. Amount of reading "I. ......
outside of school 1 — No .27 .44 -,12 -_ 10 1
reading (O-No, l-Yes)
4. Amount of reading
outside of school 2 — A u... u an
great deal of reading (0- '11 '31 '07 '04 ‘21 1
No, l-Yes)
5. SES .23 1.18 .07.... 22"" -.13‘" .02 1

. Attitud t d at" on also no it!
Sewing es 0w” -11 1.21 .17 .18 -.27 .10 .10
7. Reading self-concept .54 1.48 .03 .18." 407*" .07". .14":
8. Computer use -.14 131 .051!!! .04!!! -.l I... .06ttt .13“.
Genre-related Student Variables
9. Informational reading u :m mu use
outside of school "14 1'20 -04 '01 '23 22 .05
10. Literary reading it. a" "a ”a we
outside ofschool 26 1.81 .19 .20 -.30 .11 .15
Reading Performance
11. Inf ti l ”at: "a "a nuts ##4-
reading 1:33.60“ 539-19 63.03 .07 .25 -.10 -.09 .25
$025,136me reading 542-82 69.78 .08m .24". -.08m -.10m .23m
5:363va reading 542-11 67-76 .07‘" 25‘" 209*” -.09*" .23."
14. Large score gap ..
favoring literary reading .14 .342 .03 .01 .01 -,04 .00
(O-No, 1-Yes)
15. Large score gap ...... t ..
favoring informational .07 .26 -.06 .01 -.03 .04 -.01

reading (O-No, l-Yes)

 

67

Table 1.6 Continued

 

 

 

Student- Zero-order correlations
level
variables 6 7 8 9 10 ll 12 13 14
General Student Variables
1.
2.
3.
4.
5.
6. 1
tit
7. .25 1
it!
8. .07 .01 1
Genre-related Student Variables
it. t it.
9. .25 .04 .27 1
it! til *t* *tt
10. .52 .22 .10 .22 1
Reading Performance
it! **# tit it! ttt
11- .28 .39 -.12 -.16 .32 1
it! tit it! it. tit it!
12- .29 .39 -.13 -.17 .32 .93 1
it. til tit it! it! til ﬁtt
13. .28 .39 -.13 -.17 .32 .95 .95 1
it! t t #t it it! tit
14. .07 .04 -.03 -.04 .04 .01 .24 .12 1
iii it! it it it. it! it. it! it.
15. -.06 -.09 .04 .04 -.05 -.07 -.26 -.18 -.11

 

t it ﬁt!
p<.05; p<.01; p<.001

68

Table 1.7

U.S. Results from HLM Analyses Predicting Informational Reading Performance

 

 

 

 

 

 

 

Model 1 Model 2
Level B S.E. t B 8.13. t
Intraclass Correlation from the one-way ANOVA model = .26
Fixed Effect
Class mean score
tit it!
Intercept 529.81 3.25 163 , 1 8 533.76 4.09 13042
Class mean SES 28.47 4.31 6.60""
Class reading dzﬁiculties -3903 9.99 -391”
Informational reading *
Instruction 3°66 1-75 2.10
Literary reading
Instruction -.89 1.76 -.51
Gender
Intercept 2.55 2.32 1.10 4.27 2.38 1,79'r
Class mean SES -7.37 4.02 -1331.
Books at home 8.29 2.10 395"" 7.66 2.08 3,53"
Amount of reading outside ...... "-
of school — No reading '6'22 2'03 '3-06 ‘6-07 2'02 -3.01
Amount of reading outside I.“ 1‘"
of school - A great deal of -15.25 3.41 -4.48 -14.94 3.39 .44]
reading
tit it.
SES 7.18 1.23 5.83 6.42 1.24 5,13
2 1“ it.
(SES) -2.67 .44 -6.08 -2.60 .44 -5.96
Attitudes toward reading 6.42 1.25 5.16". 6.50 1.25 5,19".
Reading self-concept 12.52 .64 19.43"" 12.45 .64 19.39""
Computer use -6. 14 .78 4.33"" -6. 19 .78 4.94""
2 iii it.
(Computer use) -1-39 -33 -5.70 '1-38 .33 -5.69
Informational reading -7.48 .90 -330." -7.22 .90 804."
Literary reading 4.57 .63 7,23’" 4.45 .63 7,02”
Random Effect Variance df 12 Variance df 12
Class mean 855.60 246 952.31” 579.63 242 740,13”
Gender slope 151.63 246 307,41" 142.64 245 304,19"
Student-level effect 2566.73 2562.61
Variance explained (R2) At student-level: .26 In class mean: .32
Conditional ICC .25 .18

 

Note. The square terms in SES and computer use denote the curvilinear relationships between these

variables and the outcome.

T

it

t it #
p<.10; p<.05; p<.01; p<.001

69

Table 1.8

U.S. Results ﬁom HLM A nalyses Predicting Literary Reading Performance

 

 

 

 

Model 1 Model 2
Level B S.E. t B 8.15. t
Intraclass Correlation from the one-way ANOVA model = .22
Fixed Effect
Class mean score
##t ...
Intercept 531.51 3.07 17328 535.54 3.99 13430
Class mean SES 28.54 4.82 5,92m
Class reading as
D iﬂicu I ties -39.29 10.78 -3.64
Informational Reading
Instruction 3.12 1.89 1.65
Literary reading
Instruction -1.37 1.87 -.73
Gender ..
Intercept 4.45 2.32 1,92’r 6.37 2.83 2,25
Class mean SES -7.86 5.21 -1.51
*
Books at home 7.96 2.31 3,44 ’ 7.17 2.30 3,12"
Amount of reading
outside of school — No -4.16 3.14 -1.32 -3.95 3.10 -1.27
reading
Amount of reading . u g * 1.
outside of school — A -l9.76 3.58 -552 -19.29 3.56 -5_42
great deal of reading .
tit tit
SES 7.17 1.21 5,93 6.24 1.20 5,22
2 it. 1..
(SES) -2.43 .54 .453 -2.35 .53 .442
it! it.
Attitudes toward reading 8.20 1.17 7_o3 8.30 1.16 7,13
Reading self-concept 14.04 .69 20,23." 13.95 .69 2029".
Computer use -6.98 .95 4.34"" -705 .95 4.40""
2 it! ##t
(Computer use) '230 38 -6.06 -2-28 38 -6.06
t t
informational reading -8.98 1.02 -316" ' -8.65 1.03 -841" ‘
Literary reading 5.56 .69 307"" 5.40 .69 7.34""
Random Effect Variance df X 2 Variance df X2
Class mean 853.42 246 787,15” 584.45 242 527,61”
Gender slope 135.98 246 23521" 126.67 245 232,29l
Student-level effect 3301.1 1 3294.93

 

Variance explained (RT At student-level: .27

In class mean: .31

 

Conditional ICC

.21

.15

 

Note. The square terms in SES and computer use denote the curvilinear relationships between these

variables and the outcome.

1.

it

i ii
p<.10; p<.05; p<.01;

#
p< .001

70

Table 1.9
Means and Standard Deviations of U.S. Average Raw Responses in Reading Amount and

Reading Instruction Amount Related Variables

 

 

Reading amount related variables Mean SD. Min. Max.
Class-level
Informational reading instruction 1.60 .60 .33 3.00
Literary reading instruction 1.45 .45 .00 2.75
Student-level
Amount of out-of-school reading 1.38 1.26 .00 4.00
Out-of-school informational reading 1.53 .61 .00 3.00
Out-of~school literary reading 1.72 .99 .00 3.00
Computer use in and outside of school 1.62 .66 .00 3.00

 

Note. Except the ‘amount of out-of-school reading’ variable, the items in each variable above were coded
as 0 — Never or almost never, 1 — Once or twice a month, 2 — Once or twice a week, and 3 — Everyday or
almost everyday. For the ‘amount of out-of-school reading’ variable, one question was given to students
regarding the daily amount of book or magazine reading and it was coded as 0 — No time, 1 — Up to an

hour, 2 — One to three hours, 3 — Three to ﬁve hours, 4 — Five hours or more.

71

Table 1.10
Country-level and Within-U.S. Classroom/Student-level Predictors of Reading

Performance in Informational Reading and Literary Reading

 

Relationships to reading performance
Variables

 

InformatiOnal reading Literary reading

 

Country-level predictor
Infant mortality rate - -
Hours of formal reading instruction - -

Average class size n.s. -

% of students receiving informational text
instruction at least weekly

% of students receiving literary text
instruction at least weekly

% of students receiving instruction on text
style and structure at least weekly n.s. n‘s'
% of students receiving major emphasis to
text style and structure before 4'” grade

Classroom-level predictor (U .S.)

11.8 11.8.

Class mean SES + +

% of students with reading difficulties - -

Frequency of Informational text
instruction

Frequency of literary text instruction n.s. n.s.
Student-level predictor (U .S.)

+
Gender (Girls) +

( - in high SES classes)
Books at home + +

No reading outside of school - n.s.
A great deal of reading outside of school - -
SES +
Attitudes toward reading +
+

+++

Reading self-concept
Frequency of home/school computer use - -
Informational reading outside of school - -

Literary reading outside of school + +

 

+ Positive predictor ; - Negative predictor ; n.s. Not signiﬁcant

72

Table 1.11

Descriptive Comparisons of Country-level Characteristics in the Selected Countries

 

 

 

from PIRLS 2006
S l ted tri
Country-level variables 6 cc con” es
Denmark U.S. England Taiwan
Average Reading Performance
Overall reading score 546.03 539.92 537.04 535.37
Informational reading score 541.39 537.16 534.82 538.26
Literary reading score 547.09 540.66 536.11 530.44
Difference
(Informational _ Literary) -5.70 -3.50 -1.29 7.82
5min“ ‘1’“ large ““3"“ gap 16.89 13.55 16.76 7.53
favoring literary reading (%)
Students with large score gap
favoring informational reading (%) 7'89 7'69 13'68 ”'43
General National Variables
Infant mortality rate 4 7 5 5
Net enrollment rate 100 93 100 99
Hours of formal reading
Instruction 1.70 4.80 1.80 1.00
Average class size 20.40 23.00 27.40 31.70
Students in poverty schools (%) 9.60 62.40 39.00 9.10
Achievement gap between
students in low and high poverty 22.76 38.78 57.63 24.77
schools
Genre-related Variables
Teaching informational texts
at least weekly (%) 36.80 85.80 54.90 30.10
Teaching literary texts
at least weekly (%) 89.50 91.60 74.70 66.90
Teachmg text Style °’ mm" a‘ 39.10 71.40 74.50 55.20
least weekly (%) .
Major emphasis on text style and
structure before 4th grade (%) 35.80 75.58 82.77 12.78
Degree Of emphaSis on yanous -.52 .56 .56 .56
reading purpose in curriculum
Curriculum emphasis on Some Major Major Major
informational reading purpose at ha . h . h . has'
4m grade emp SIS emp aSis emp aSis emp is
Curriculum emphasis on literary Some Major Major Major
reading purpose at 4th grade emphasis emphasis emphasis emphasis

 

Note. The national average reading performance scores in this table are the population estimates with the

original samples in PIRLS.

73

Table 1.12
Average Reading Performance Scores by Informational-Literary Reading Performance

Gap Size in Selected Countries

 

 

 

 

 

 

. Informational-literary Average reading performance
Countries reading performance 11 %
(N) gap Informational Literary Overall
reading score reading score reading score

Large gap with "I. ”
higher literary 619 16 506.87 553.11 532.60
reading score

Denmark

( 3,786) Small gap 2,897 77 552.10 553.71 554.57
Large gap with u u,
higher informational 270 7 568.84 52344 545.62
reading score
Large gap With it! tit
higher literary 609 14 539.51 583,02 560,79
reading score

U.S.

( 4, 508) Small gap 3,579 79 538.60 539.86 540.66
Large gap with ..
higher informational 320 7 524.03 478.84 500.02
reading score
Large gap with .
higher literary 632 17 528.56 577.07 554.32
reading score

England

(3,829) Small gap 2681 70 537.83 537.65 539.55
I'fargc gap With . a taint tau
higher mfonnatlonal 516 13 545.42 493.74 521.60
reading score
Large gap with ..." ......
higher literary 305 7 543.66 587.15 562.70
reading score

Taiwan

(4,207) Small gap 3,193 76 543.54 540.03 543. 19

Large gap With . axis-1t him '1”
higher Informational 709 17 527.40 479.24 504.75
reading score

#

Note. Mean differences were statistically analyzed in comparison with the means of the small gap groups.
Shaded scores are the lowest scores among the scores of the three groups by information-literary reading

performance gap size for each of informational, literary, and overall reading performance scores.
.

t it t t
p<.05; p<.01; p<.001

74

Table 1.13

HLM Coeﬂicients for Major Predictors of Reading Performance in Selected Countries

 

 

 

Coefﬁcient for Coefﬁcient for
Variables informational reaﬂgperformance Literary reading Erformance .
Den U S Eng Tai Den U S Eng Tai
mark ' ' land wan mark ' ' land wan
Class-level predictor
for Class mean score
Class mean SES 8.90 28.47 15.61 21.79 8.62 28.54 17.83 26.52
Class reading
difﬁculties -66.94 -39.03 -91.58 n.s. ~63.21 -39.29 -100.54 n.s.
Informational reading
. . n.s. 3.66 n.s. n.s. n.s. n.s. n.s. n.s.
instruction
Literary reading
instruction n.s. n.s. n.s. n.s. n.s. n.s. n.s. n.s.
Student-level predictor
Gender 6.68 4,27a n.s. n.s. 8.33 6.37 n.s. 4.49
Books at home 19.88 7 .66 19.03 20.60 16.62 7.17 20.39 21.12
Amount of reading
outside of school - No -10.62 -6.07 -15.98 -14.63 -9.10 n.s. -13.73 -11.79
reading
Amount of reading
outside of school -15.56 -14.94 -19.75 -8.99 -10.79 -19.29 -10.85 n.s
- A great deal of reading
SES 5.55 6.42 6.40 2.52 6.82 6.24 3.96 3.65
(SES)2 -1.42 -2.60 -1 .66 -3.15 -1.45 -235 -1.22 -2.16
Attitudes toward reading 5.27 6.50 3.92 5.26 2.98 8.30 4.97 4.72
Reading self-concept 22.88 12.45 20.42 8.54 19.57 13.95 21.62 9.42
Computer use -1 . 15 -6.19 -5.52 -1.60 -2.15 -7.05 -6.28 .43
(Compute, usef n.s. -1.88 -3.05 -3.07 n.s. -2.28 -3.43 -3.11
Inf°imu°nal readmg -8.93 -7.22 -10.66 -8.01 -10.05 -8.65 -13.49 -8.90
ou181de of school
”em" ”mg “‘5‘“ 2.08 4.45 7.48 3.22 2.79 5.40 8.73 5.00
of school
Inmc'm ”mm” .14 .26 .22 .09 .14 .22 .21 .09

(ICC)

 

 

Note. The square terms in SES and computer use denote the curvilinear relationships between these

variables and the outcome.

a . . . . .
This coefﬁc1ent assumes that class mean SES is zero because of interaction between gender and class

mean SES.

n.s.p>.10

75

References

Abadiano, H. R., & Turner, J. (2002). Reading expository text: The challenges of
students with learning disabilities. The New England Reading Association
Journal, 38, 49-55.

Anderson, R. C., Wilson, P. T., & Fielding, L. G. (1988). Growth in reading and how
children spend their time outside of school. Reading Research Quarterly, 23,
285-303.

Arnold, H. (1992). Diagnostic reading record. London: Hodder & Stoughton.

Beghtol, C. (2000). The concept of genre and its characteristics. Bulletin of the American
Society for Information Science and Technology, 27, 17.

Benson, V. (2002). Shifting paradigms and pedagogy with nonﬁction: A call to arms for
survival in the 21 st century. The New England Reading Association Journal, 38,
1-6.

Bond, T. G., & F ox, C. M. (2007). Applying the Rasch model: Fundamental
measurement in the human sciences. Mahwah, NJ: Lawrence Erlbaum.

Bourdieu, P. (1991). Language and symbolic power. Cambridge, MA: Harvard
University Press.

Campbell, J. R., Kelly, D. L., Mullis, I. V. S., Martin, M. O., & Sainsbury, M. (2001).
Framework and speciﬁcations for PIRLS assessment 2001. Chestnut Hall, MA:
International Study Center, Boston College.

Caswell, L. J., & Duke, N. K. (1998). Non-narrative as a catalyst for literacy
development. Language Arts, 75, 108-117.

Chall, J. 8., Jacobs, V. A., & Baldwin, L. E. (1990). The reading crisis: Why poor
children fall behind. Cambridge, MA: Harvard University Press.

Chapman, L. J. (1987). Reading: From 5-11 years. Milton Keynes, England: Open
University Press.

Chapman, M. L. (1999). Situated, Social, Active: Rewriting genre in the elementary
classroom. Written Communication, 16, 469-490.

Coles, M., & Hall, C. (2002). Gendered readings: learning from children’s reading
choices. Journal of Research in Reading, 25, 96-108.

Cooper, J. (1995). Children reading non-ﬁction for pleasure. Literacy, 29, 15-21.

76

Cope, B., & Kalantzis, M. (Eds). (1993). The powers of literacy: A genre approach to
teaching writing. Pittsburgh, PA: University of Pittsburgh Press.

Doyle, C. S. (1994). Information literacy in an information society: A concept for the
information age (Report No. IR-97). Syracuse, NY: ERIC Clearinghouse on
Information & Technology.

Duke, N. K. (2000). 3.6 minutes per day: The scarcity of informational texts in ﬁrst
grade. Reading Research Quarterly, 35, 202-224.

Duke, N. K. (2005). Comprehension of what for what: Comprehension as a non-unitary
construct. In S. Paris & S. Stahl (Eds.), Current issues in reading comprehension
and assessment (pp. 93-104). Mahwah, NJ: Erlbaum.

Duke, N. K., Bennett-Armistead, S., & Roberts, E. M. (2003). Filling the nonﬁction
void: Why we should bring nonﬁction into the early-grade classroom. American
Educator, 27, 30-34, 46.

Duke, N. K., & Purcell-Gates, V. (2003). Genres at home and at school: Bridging the
known to the new. The Reading Teacher, 5 7, 30-37.

Eisenberg, M. B., Lowe, C. A., & Spitzer, K. L. (2004). Information literacy: Essential
skills for the information age (2nd ed.). Westport, CT: Libraries Unlimited.

Epstein-Jannai, M. (2005). Using the concept of genre in a writing workshop:
Description of a learning environment. L1 -Educational Studies in Language and
Literature, 5, 57-74.

Georgakopoulou, A., & Goutsos, D. (2000). Revisiting discourse boundaries: The
narrative and non-narrative modes. Text, 20, 63-82.

Hall, K. M., Sabey, B. L., & McClellan, M. (2005). Expository text comprehension:
Helping primary-grade teachers use expository texts to full advantage. Reading
Psychology, 26, 211-234.

Halliday, M. A. K. (1978). Language as social semiotic: The social interpretation of
language and meaning. Baltimore, MD: University Park Press.

Harkrader, M. A., & Moore, R. (1997). Literature preferences of fourth graders. Reading
Research and Instruction, 36, 325-339.

Hartman, D.K., Morsink, P.M., & Zheng, J. (in press). From print to pixels: The
evolution of cognitive conceptions of reading comprehension. In B. Baker (Ed.),

Multiple perspectives on new literacies research and instruction. New York:
Guilford.

77

Hirsch, E. D. (2003). Reading comprehension requires knowledge of words and the
world: Scientiﬁc insights into the fourth-grade slump and stagnant reading
comprehension. American Educator, 27, 10, 12-13, 16-22, 28-29.

Jacobs, P. R. (2004). Expository text comprehension instruction for at-risk ﬁrst grade
children. Unpublished doctoral dissertation, Columbia University, New York.

King, G., & Zeng, L. (2001). Improving forecasts of state failure. World Politics, 53,
623-658.

Kirk, L., & Pearson, H. (1996). Genres and learning to read. Reading, 30, 37-41.

Klecker, B. M. (2006). The gender gap in NAEP fourth, eighth, and twelfth-grade
reading scores across years. Reading Improvement, 43, 50-56.

Kletzien, S. B., & Dreher, M. J. (2004). Informational text in K—3 classrooms: Helping
children read and write. Newark, DE: International Reading Association.

Kucan, L., & Beck, 1. L. (1996). F our fourth graders thinking aloud: An investigation of
genre effects. Journal of Literacy Research, 28, 259-287.

Labov, W. (1972). Language in the inner city: Studies in the Black English vernacular.
Philadelphia: University of Pennsylvania Press.

Lemke, M., Sen, A., Johnston, J. S., Pahlke, E., Williams, T., Kastberg, D., et a1. (2005).
Characteristics of U.S. 15-year-old low achievers in an international context:

Findings from PISA 2000 (N CES 2006-010). Washington, DC: National Center
for Education Statistics.

Littlefair, A. B. (1991). Reading all types of writing: The importance of genre and

register for reading development. Milton Keynes, England: Open University
Press.

Littlefair, A. B. (1992). Let's be positive about genre. Reading, 26, 2-6.

MacLean, M., & Chapman, L. J. (1989). The processing of cohesion in ﬁction and non-
ﬁction by good and poor readers. Journal of Research in Reading, 12, 13-28.

McDonnell, L. M. (1995). Opportunity to learn as a research concept and a policy
instrument. Educational Evaluation and Policy Analysis, 1 7, 305-322.

Mohr, K. A. J. (2006). Children's choices for recreational reading: A three-part
investigation of selection preferences, rationales, and processes. Journal of
Literacy Research, 38, 81-104.

Mullis, I. V. S., Kennedy, A. M., Martin, M. O., & Sainsbury, M. (2006). PIRLS 2006
assessment ﬁamework and specifications (2nd Edition). Chestnut Hall, MA:
International Study Center, Boston College.

78

Mullis, I. V. S., Martin, M. O., Gonzalez, E. J., & Kennedy, A. M. (2003). PIRLS 2001
international report: IEA ’5' study of reading literacy achievement in primary
school in 35 countries. Chestnut Hill, MA: International Study Center, Boston
College.

Mullis, I. V. S., Martin, M. 0., Kennedy, A. M., & Foy, P. (2007). PIRLS 2006
international report: IEA ’s progress in international reading literacy stuoy in
primary schools in 40 countries. Chestnut Hill, MA: International Study Center,
Boston College.

National Assessment Governing Board. (2006). Reading Framework for the 2007
National Assessment Educational Progress. Washington, DC: U.S. Government
Printing Ofﬁce.

Neevdal, F. (2007). Home-PC usage and achievement in English. Computers &
Education, 49, 1112-1121.

New London Group. (1996). A pedagogy of multiliteracies: Designing social futures
Harvard Educational Review, 66, 60-92.

O’Dwyer, L. M., Russell, M., Bebell, D., & Tucker-Seeley, K. R. (2005). Examining the
relationship between home and school computer use and students’
English/language arts test scores. Journal of Technology, Learning, and
Assessment, 3, 4-45. Retrieved September 13, 2008, from
http://escholarship.bc.edu/jtla/

Paré, A., & Smart, G. (1994). Observing genres in action: Towards a research
methodology. In A. Freedman & P. Medway (Eds.), Genre and the New Rhetoric
(pp. 146-154). Bristol, PA: Taylor & Francis.

Perle, M., Moran, R., & Lutkus, A. D. (2005). NAEP 2004 trends in academic progress:
Three decades of student performance in reading and mathematics (NCES 2005-
464). U.S. Department of Education, Institute of Education Sciences, National
Center for Education Statistics. Washington, DC: Government Printing Ofﬁce.

Pressley, M. (2006). Reading instruction that works: The case for balanced teaching (3“
edition). New York: Guilford Press.

Pressley, M., Rankin, J. L., & Yokoi, L. (1996). A survey of instructional practices of
primary teachers nominated as effective in promoting literacy. The Elementary
School Journal, 96, 363-384.

Pressley, M. & Wharton-McDonald, R. (1997). Skilled comprehension and its
development through instruction. School Psychology Review, 26, 448-466.

Purcell-Gates, V., Duke, N. K., & Martineau, J. A. (2007). Learning to read and write
genre-speciﬁc text: Roles of authentic experience and explicit teaching. Reading
Research Quarterly, 42, 8-45.

79

Rasch, G. (1980). Probabilistic models for some intelligence and attainment tests.
Chicago: University of Chicago Press.

RAND Reading Study Group (2002). Reading for understanding: Toward a research
and development program in reading comprehension. Santa Monica, CA: RAND.

Raudenbush, S. W., & Bryk, A. S. (2002). Hierarchical linear models: Applications and
data analysis methods (2nd ed.). Thousand Oaks, CA: Sage Publications.

Reutzel, D. R., Smith, J. A., & F awson, P. C. (2005). An evaluation of two approaches
for teaching reading comprehension strategies in the primary years using science
information texts. Early Childhood Research Quarterly, 20, 276-305.

Saenz, L. M., & Fuchs, L. S. (2002). Examining the reading difﬁculty of secondary
students with learning disabilities: Expository versus narrative text. Remedial
and Special Education, 23, 31-41.

Sales, E., Lichtenwalter, S., & F evola, A. (2006). Secondary analysis in social work
research education: Past, present, and future promise. Journal of Social Work
Education, 42, 543-558.

Schmidt, W. H., & McKnight, C. C. (1995). Surveying educational opportunity in
mathematics and science: An international perspective. Educational Evaluation
and Policy Analysis, I 7, 337-353.

Smith, E. V. (2000). Understanding Rasch measurement: Metric development and score
reporting in Rasch measurement. Journal of Applied Measurement, 1, 303-326

Stamboltzis, A., & Pumfrey, P. (2000). Reading across genres: A review of literature.
Support for Learning, 1 5, 58-61.

Stanovich, K. E. (1986). Matthew effects in reading: Some consequences of individual
differences in the acquisition of literacy. Reading Research Quarterly, 21, 360-
407.

Stedman, L. C., & Kaestle, C. F. (1987). Literacy and reading performance in the United
States, from 1880 to the present. Reading Research Quarterly, 22, 8-46.

Swales, J. (1990). Genre analysis: English in academic and research settings.
Cambridge: Cambridge University Press.

Themstrom A., & Themstrom S. (2003). No excuses: Closing the racial gap in learning.
New York: Simon & Schuster.

Watson, M. (1985). Diyferences in book choices for reading pleasure between second

though ﬁﬁh grade boys and girls (ERIC Document Reproduction Service No.
ED259304).

80

Weinberger, J. (1996). A longitudinal study of children’s early literacy experiences at
home and later literary experiences at home and later literacy development at
home and school. Journal of Research in Reading, 19, 14-24.

Wigﬁeld, A., & Guthrie, J. T. (1997). Relations of children’s motivation for reading to
the amount and breadth or their reading. Journal of Educational Psychology, 89,
420-432.

Wolfe, M. B. W. (2005). Memory for narrative and expository text: Independent
inﬂuences of semantic associations and text organization. Journal of
Experimental Psychology: Learning Memory and Cognition, 3 I , 359-364.

Wright, B. D. (1997). A history of social science measurement. Educational
Measurement: Issues and Practice, 1 6, 33-45, 52.

Yopp, R. H., & Yopp, H. K. (2006). Informational texts as read-alouds at school and
home. Journal of Literacy Research, 3 8, 37-51.

81

STUDY TWO
HOW MOTIVATIONAL CONSTRUCTS INTERACT

TO PREDICT READING PERFORMANCE

Abstract

The purposes of this study were to explore underlying factors in reading
motivation measured in the Progress in International Reading Literacy Study (PIRLS)
and to investigate the relationships between those motivational factors and students’
reading performance. A special focus was given to interactions between different
motivational facets in predicting reading performance. The study analyzed the U.S. data
ﬁom PIRLS 2006 using factor analysis and hierarchical linear modeling. Result revealed
ﬁrst that PIRLS motivation items clustered under two motivational dimensions: student
attitudes toward reading and student self-concept as readers. Within each of these
dimensions, there were two contrasting facets based on their internal or external
orientation: student attitudes toward reading consisted of extrinsic motivation and
intiinsic motivation, while student reading self-concept was composed of self-referenced
perceived competence and peer-referenced perceived competence. Second, different
facets under the same motivational dimensions interacted with each other when
predicting students’ reading performance. Although the relationships of one motivational
facet to reading performance were diﬂerent depending on the levels of its contrasting
facet, this study found that internally-oriented motivational facets were always more
positive predictors of students’ reading performance than extemally-oriented motivational

facets. Extrinsic motivation under the dimension of attitudes toward reading, for example,

82

positively related to reading performance if a student had high int1insic motivation.
However, it was a negative predictor of reading performance for those students with low
int1insic motivation. Third, reading motivation was a strong predictor of reading
performance even when other literacy-related variables were controlled. Four
motivational facets constructed in this study explained 17% of the student-level variance
in reading performance. The results of this study conﬁrm the multidimensional and
multifaceted nature of reading motivation with emphasis on the predictive power of
intemally-oriented motivation. The study suggests that researchers consider interacting

relationships of different motivational constructs.

83

Introduction

Educational researchers agree that learning involves not only various
cognitive/metacognitive processes for knowledge construction and application, but also
motivational/emotional aspects of human nature. This has been supported by the fact that
struggling students with low academic performance usually display maladaptive patterns
of motivation to learn, for instance, low interest, less persistence, and negative
competence beliefs (e. g., Chapman, 1988; Gans, Kenny, & Ghany, 2003). Teachers have
also expressed great concern about how to stimulate and maintain their students’
motivation to learn (O’Flahavan, Gambrell, Guthrie, Stahl, Baumann, & Alvermann,
1992). In the classroom context, they have been especially interested in how to make
instruction more attractive to students, how to build individual and collaborative learning
goals, and how to teach things that are worth learning (Brophy, 2004).

Reading researchers and educators have also paid attention to the role of
motivation in improving students’ reading abilities over the past couple of decades,
concomitantly with the rising issue of subject-speciﬁcity in academic motivation (Eccles,
Wigﬁeld, & Schiefele, 1998; Pintiich, 1994). Nevertheless, motivation has been
neglected in several studies and reviews that have signiﬁcantly inﬂuenced reading policy
and the ﬁeld of reading research (e.g., Gough & Tunmer, 1986; NICHD, 2000). This is
unfortlmate, because motivation certainly plays an important role in the development of
reading. For example, motivation often distinguishes between reading that is superﬁcial
and shallow and reading that is deep and internalized (Schiefele, 1999). In addition,
motivated students read more books and spend more time reading, which is likely to

result in better reading performance (Wigﬁeld & Guthrie, 1997).

84

According to a report ﬁom the Progress in International Reading Literacy Study
(PIRLS) conducted in 2006 (Mullis, Martin, Kennedy, & Foy, 2007), U.S. fourth-grade
students have relatively lower reading motivation than students in other countries. Only 6
out of 45 countries or jurisdictions participating in PIRLS had lower percentages of
students with high attitudes toward reading than the United States. Also, there were only
four countries or jurisdictions which had higher percentages of students with low reading
self-concept than the U.S.

This and other evidence provides a solid basis for the necessity of more research
on motivation in the ﬁeld of reading in the U.S. Thus, this study focuses on the
motivational predictors of children’s reading performance using the data ﬁom PIRLS
2006, and explores an alternative perspective on the complex association between
motivation and reading performance. In particular, drawing on the multidimensional and
multifaceted characteristics of subj ect-speciﬁc motivation, this study investigates the
interactive relationships among diﬁerent facets under diﬂ‘erent motivational dimensions
and reading performance of U.S. students.

Theoretical & Empirical Background

In this section, I ﬁrst will review various deﬁnitions of motivation and some key
issues in motivation research. Then, because this study is based on motivation data ﬁ'om
the Progress in International Reading Study (PIRLS) in which extrinsic/intrinsic
motivation and self-concept were the main motivation variables among multiple
motivation dimensions, it will brieﬂy explore those two dimensions generally and also in
relation to reading. Finally, this section will present an introduction to PIRLS and how it

assessed students’ reading motivation.

85

Motivation

Motivation is a set of complex constructs, and therefore it has many slightly
different deﬁnitions from different perspectives. Wittrock (1986) deﬁned motivation as
the process of initiating, sustaining, and directing human behaviors. Maehr and Meyer
(1997) expressed it as the investment of personal resources such as time, energy,
knowledge, and skills toward a certain direction with a certain degree of intensity,
persistence, and quality for resulting outcomes. Guthrie and Wigﬁeld (1999) also
described motivation as a multifaceted set of goals and beliefs that guide behavior. What
is consistent in different deﬁnitions is that motivation broadly refers to something that
initiates and guides a person’s inclination toward goal-directed behaviors. Several key
issues regarding the conceptualization of motivation have appeared in the ﬁeld of
motivation research: multidimensionality, contrasting multifacets, subject-speciﬁcity,
developmental change, and directionality.

Multidimensionality. Because there are different aspects of human motivation to
learn, no single concept can fully explain its complex dynamics (Bong, 1996). Thus,
motivation should be understood as a multidimensional concept with multiple
constituents. In their extensive review of motivation literature, for instance, Murphy and
Alexander (2000) summarized many different motivational terms into four representative
dimensions based on their conceptual association: goal or what learners want to achieve
(Wentzel, 1989), intrinsic/extrinsic motivation or whether a learner performs a task
because of the task itself or because of something outside of the task (Whang & Hancock,
1994), interest or the processes by which the underlying needs of learners are energized

(Alexander, Murphy, Woods, Duhon, & Parker, 1997), and self-schema or personal

86

knowledge about oneself (Pintrich & Schunk, 1996). Given this, understanding
motivation requires multidimensional models.

Contrasting multzfacets. Under any one of the motivational dimensions, different
people may not only have different amounts of that motivation, but also different types of
motivation (Ryan & Deci, 2000). Therefore, one motivational dimension can be
constructed by diﬁerent facets that share a central concept but are contrasted in a certain
aspect of the quality. For example, the goal dimension of motivation can be dichotomized
into mastery goals and performance goals as a ﬂinction of how competence is deﬁned:
learners with a mastery goal orientation value the process of learning itself and aim to
develop new skill or knowledge, whereas learners with a performance goal orientation
concern outperforming others and showing the evidence of their abilities (Ames & Archer,
1988). The goal dimension can also be diﬁerenﬁated into approach and avoidance as a
function of valence: learning through an approach goal is directed by a positive outcome
or expectancy, while learning activities through an avoidance goal are directed by a
negative outcome or expectancy (Elliot, 1999). Interest as a motivational dimension, for
another instance, is often categorized into individual interest as engagement led by a
person’s relatively enduring predisposition and situational interest as focused attention
triggered by environmental stimuli, according to what tiiggers the affective reaction to
learn (Ainley, Hidi, & Bemdorff, 2002; Hidi & Renninger, 2006). Other motivational
dimensions also involve several contrasting facets as will be seen later.

Subject-speciﬁcity. It is now regarded as highly important to examine the
separate dimensions of academic motivation in different domains or subject areas,

because students’ motivation develops with different patterns in different subject areas

87

(e. g., Wigﬁeld et al., 1997). In research on competence belief or self-concept, Shavelson
and his associates modeled the structure of self-concept as a multidimensional and
hierarchical structure (e. g., Shavelson, Hubner, & Stanton, 1976; Byme & Shavelson,
1986). That is, students have subject-speciﬁc self-concepts for different subject-matters
under a more comprehensive academic self-concept. Further, academic self-concept
together with a non-academic self-concept composes a general self-concept. Using this
model, Byme and Worth Gavin (1996) exhibited empirical evidence that students in
Grades 3, 7, and 11 had subject-speciﬁc self-concepts for English and math. These
studies suggest that research on motivation should approach the topic in a subj cot-speciﬁc
manner.

Developmental change. Research has shown that children’s learning motivation
changes over time. Unfortunately, positive motivations such as intrinsic motivation,
mastery goal orientation, and perceived competence are known to signiﬁcantly decrease
as children get older (e.g., Eccles, Wigﬁeld, & Schiefele, 1998; Leeper, Corpus, &
Iyengar, 2005; Unrau & Schlackman, 2006). Wigﬁeld (2000) provided two possible
explanations about this negative development of motivation. First, as children get older
and experience more social comparison in and outside of school, they become much
better at understanding and interpreting evaluative feedback about themselves and their
performance. Second, the classroom environment changes signiﬁcantly in that evaluation
and competition become more conspicuous to students.

Directionality. Although there is little experimental research on the causality
between motivation and achievement, several correlational studies have proposed a

bidirectional reciprocal relationship between the two (e. g., Guay, Marsh, & Boivin, 2003;

88

Marsh, Trautwein, Lﬁdke, Koller, & Baumert, 2005). Morgan and Fuchs (2007) reviewed
15 studies which tested a relationship between students’ reading motivation and their
reading abilities in primary grades and found support for the hypothesis that children’s
early reading difﬁculties and low motivation interact to undermine their later reading
growth. This also provides a testimony of negative “Matthew eﬂ'ects” in the development
of reading abilities (Stanovich, 1986, p. 381). The early reading diﬁ'rculty children
experience discourages them from reading and this again leads to poorer reading ability.
Intrinsic/Extrinsic Motivation

The intrinsic/extrinsic dimension has received the most empirical and theoretical
attention among various dimensions of academic motivation (Ryan & Deci, 2000;
Vansteenkiste, Timmermans, Lens, Soenens, & Van den Broeck, 2008). The two facets of
this dimension concern different loci of causality about what leads someone to be
engaged in learning (deCharm, 1968): internal vs. external locus of causality. Intrinsic
motivation refers to the activities in which a learner gets involved because of the activity
itself, whereas extrinsic motivation refers to the activities that are pursued for something
that comes from outside the learner.

In the past, researchers conceived of motivation as falling on a continuum
between intrinsic motivation on one end and extrinsic motivation on the other end,
suggesting that those two are negatively correlated (e. g., Harter & Jackson, 1992). From
this perspective, a person is either extrinsically or intrinsically motivated toward an
object by a bipolar dichotomy. Recent studies, however, propose that intrinsic and
extrinsic tendencies can be better understood as two independent motivation facets and

thus should be represented on separate continua (e. g., Covington & Mueller, 2001; Lin,

89

McKeachie, & Kim, 2003). This perspective assumes that a person may possess multiple
motivations, that is, both intrinsic and extrinsic motivation, at the same time.

Although learners can be motivated for both extrinsic and intrinsic reasons, many
motivation researchers argue that intrinsic motivation is more beneﬁcial to learning than
is extrinsic motivation (Deci & Ryan, 1985). This is because intrinsically motivated
learners are more likely to be deeply engaged in their work and also make use of various
eﬂ‘ective strategies for learning than extrinsically motivated students. As noted earlier,
many researchers have reported that there is a signiﬁcant shift in students’ motivation
ﬁ'om intrinsic motivation to extrinsic motivation toward school learning, especially
during early adolescence (e. g., Eccles & Midgley, 1990; Harter, 1981). Thus, study of
motivation at different ages should consider both intrinsic & extrinsic motivation.
Self-Concept

Learners who hold different beliefs about themselves demonstrate different
levels of cognitive, affective, and social engagement in learning (Bong & Skaalvik, 2003).
In motivation research, these beliefs are often referred to as self-concept, but are
sometimes mentioned as perceived competence, ability beliefs, and so on. Self-concept is
broadly deﬁned as the totality of the individual’s thoughts and feelings, having reference
to oneself as an object particularly in relation to one’s ability (Rosenberg, 1979). This
perceived competence is formed through one’s experience with his or her environment,
and it is inﬂuenced especially by evaluations from signiﬁcant others, reinforcements, and
attribution (Shavelson, Hubner, & Stanton, 1976).

Self-concept sometimes has been confused in motivation research with self-

effrcacy, but those two constructs should be distinguished. Self-efﬁcacy refers to beliefs

9O

in one’s abilities to organize and execute a series of actions required to attain one’s goal
in a given situation (Bandura, 1977). In their review of research on self-concept and self-
efﬁcacy, Bong and Skaalvik (2003) showed how those two constructs could be
systematically differentiated. According to their study, self-concept represents relatively
stable perceptions of the self that are past oriented, whereas self-efﬁcacy denotes
malleable and future-oriented conceptions of the self and one’s potential.

Similarly to the distinction between intrinsic and extrinsic motivation by
different loci of causality, self-concept can be decomposed depending on diﬁerent loci of
comparison, that is, diﬁ‘erent frames of reference: external comparison and internal
comparison (Marsh, 1986). By external comparison, people compare their perceptions of
their own competence in a certain domain with the perceived abilities of other people and
use this comparison as a basis of their self-concept in that domain. This is contrasted with
internal comparison by which people compare their perceived competence in one domain
with their perceived abilities in other domains. Studies have shown that these two facets
work together to build one’s academic self-concept through different processes (Bong,
1998; Marsh, 1986).

Relationships between Motivation and Achievement

Although reciprocal relationships are assumed between motivation and
achievement as mentioned above, many motivation researchers have been especially
interested in motivation as a predictor of academic achievement in reading and other
subject matters. This might be because the evidence of that causality direction can
provide more useful implications to educators for the development of motivationally

more desirable classroom instruction and environments.

91

There have been numerous studies about how various motivational dimensions
facilitate learners’ academic performance and achievement. In reading research, for
instance, Guthrie, Wigﬁeld, Metsala, and Cox (1999) studied the relationships among
reading comprehension, intrinsic motivation, and reading amount with elementary
through high school students. They found that reading motivation signiﬁcantly predicted
students’ reading amount and the reading amount predicted reading comprehension
signiﬁcantly when other related variables were controlled. Chapman, Tunmer, and
Prochnow (2000) also showed that children’s reading-related skills and performance at
school in early grades were strongly predictive of their academic self-concept and this
self-concept signiﬁcantly predicted subsequent reading performance.

Most studies have examined a single motivation variable regarding the
relationship between motivation and achievement. There has been relatively little
research on how different motivational dimensions and facets work together to predict
academic performance. Moreover, when available studies on the motivation—achievement
relationship are reviewed, results have been mixed. For example, researchers have long
reported interactive relationships of intrinsic motivation and extrinsic motivation to
achievement saying that offering people extrinsic rewards for performing an intrinsically
motivated activity tends to decrease their intrinsic motivation for the activity and thus
results in a negative relationship with performance (e. g., Lepper, Green, & Nisbett, 1973;
Vansteenkiste et al., 2008). In their meta-analytic study, Deci, Koestner, and Ryan (1999)
concluded that extrinsic motivation has a substantial and reliable undermining eﬂ‘ect on
intrinsic motivation. On the contrary, several studies have argued that extrinsic

motivation can enhance intrinsic motivation under certain circumstances and, in fact, it is

92

positively related to performance (e. g., Cameron & Pierce, 1994; Covington, 2000;
Eisenberger & Cameron, 1996). The meta-analysis by Cameron and Pierce (1994), for
example, indicated that the detrimental effects of extrinsic rewards occur only under
highly restricted, easily avoidable situations and the positive effects of extrinsic
motivation are easily observed when extrinsic rewards such as verbal rewards and
tangible rewards depending on performance quality are used in appropriate ways. Thus,
this issue of the relationship of extrinsic versus intrinsic motivation to performance is still
controversial. For the self-concept related facets, Marsh (1986, 1990) proposed that
students’ achievement in one domain was related positively with both external and
internal comparison based self-concepts in that domain, while self-concepts in different
domains are uncorrelated with each other (but see Bong, 1998). Still, there is not enough
research about the mechanism by which different facets under the same motivation
dimensions are interactively related to achievement.

Progress in International Reading Literacy Study (PIRLS)

The Progress in International Reading Literacy Study (PIRLS) is an international
comparative reading assessment which has been conducted every ﬁve years since 2001
under the supervision of the International Association for the Evaluation of Educational
Achievement (IEA). It assesses the reading performance of students in the upper grade of
the two adjacent grades that contains the largest portion of 9-year-olds at the time of
testing. These students were typically in the fourth grade and their average age was
around 10 in most countries including the United States (Mullis, Martin, Kennedy, & Foy,
2007). The ﬁrst PIRLS assessment took place in 2001 with students from 35 countries or

jurisdictions, and the second assessment had students from 45 countries or jurisdictions in

93

2006.

PIRLS not only assessed students’ reading performance but also obtained a
variety of literacy-related information from children, parents, teachers, school
administrators, and national coordinators of the assessment. Reading motivation is one of
the important areas in which PIRLS collected information ﬁom students. It constructed
two dimensions of reading motivation: student reading attitudes toward reading and
student reading self-concept. In student questionnaires, six questions were asked to assess
student attitudes toward reading, and four questions were asked for student reading self-
concept (see Table 2.1). Attitude is generally deﬁned as “a learned predisposition to
respond in a consistently favorable or unfavorable manner with respect to a given object”
(Fishbein & Ajzen, 1975). In research on academic learning, it has been often used as a
general motivation variable which represents a composite of several motivational
dimensions (e.g., McKenna, Kear, & Ellsworth, 1995; Wallbrown, Levine, & Engin,
1981), but there has been no clear consensus about which motivational dimensions
should be included in the attitude construct. In PIRLS, student attitudes toward reading
involved the questions related to reasons for reading and enjoyment of reading. The
questions concerning student reading self-concept asked about how students perceived
their reading abilities.

Table 2.2 displays the means and standard deviations of the motivation variables
in PIRLS and the zero-order correlations between those variables and some other PIRLS
variables including reading performance scores and background information. In this
correlation matrix, it is noticeable that different motivation variables under the same

dimensions show somewhat different patterns in their correlations with other variables.

94

Under student attitudes toward reading, for example, ATRl , ATR4, and ATR6 have higher
correlations with reading performance scores than do other attitude variables, while those
variables have lower correlations with students’ informational reading outside of school
than others. Under student reading self-concept, RSCl and RSC3 have positive
correlations with students’ informational reading outside of school, whereas RSC2 and
RSC4 have negative correlations with the variable. This provides a possibility that those
motivation dimensions in PIRLS can be broken down into smaller factors. A closer look
at each question of the motivation variables also conﬁrms this possibility. For instance, ‘I
enjoy reading’ and ‘I need to well for my future’ under student attitudes toward reading
can be regarded as very different motivational orientations: one as intrinsic motivation
and the other as extrinsic motivation.
Purposes of the Study

The present study aimed to clarify underlying factors within reading motivation
data from PIRLS and to explore the relationships between those factors and students’
reading performance. A special focus was given to the interactions between contrasting
motivational facets in predicting reading performance. To do this, I analyzed the relevant
U.S. data from PIRLS 2006 and asked the following questions: First, what factors
constitute reading motivation in the PIRLS motivational variables which were originally
constructed as student attitudes toward reading and student reading self-concept? Second,
are there any interactions between different motivational factors or facets in relation to
reading performance? Third, how does reading motivation predict students’ reading
performance aﬁer other background variables are controlled?

Methods

95

The design of the present study is a secondary analysis, the reanalysis of existing
data (Sales, Lichtenwalter, & Fevola, 2006), using the U.S. data collected for the Progress
in International Reading Literacy Study (PIRLS) in 2006. The dimensions and facets
underlying the motivation items of PIRLS were constructed and conﬁrmed using factor
analysis. In addition, the relationships between different motivational facets and reading
performance were investigated using a multilevel modeling technique.

Participants

In PIRLS, the U.S. sample was designed to be representative of all fourth-grade
students in the 50 states and the District of Columbia (Baer, Baldi, Ayotte, & Green,
2007). The representative sample of students was selected using a two-stage stratiﬁed
cluster sample design, with schools sampled in the ﬁrst stage and one or two classes
within each sampled school in the second. In the U.S., a total of 5,190 fourth-grade
students from 253 classrooms in 183 schools were included in the ﬁnal sample for 2006.

The current study only included the data from U.S. students who provided their
responses to all the motivational variables in the PIRLS questionnaires. This sample
consisted of 4,826 students (2,421 females, 2,402 males, 3 unreported) from the same
number of classrooms and schools. SES of the students in the sample is reported in
Tables 2.2 and 2.5; no information about student racial/ethnic background is provided in
PIRLS.

Measures and Variables

PIRLS had two different sources of information: one from the reading

assessment and the other from the background questionnaires. The PIRLS reading

assessment tested students’ reading performance on two different types of reading, that is,

96

informational reading and literary reading. Thus, students’ reading performance scores
were separately scaled for overall reading, literary reading, and informational reading,
with an international mean of 500 and standard deviation of 100 (Mullis, Martin,
Kennedy, & Foy, 2007). In addition to the reading assessment, PIRLS administered ﬁve
background questionnaires: a student questionnaire for participating students, a home
survey for one of the student’s parents, a teacher questionnaire for their reading teachers,
a school questionnaire for students’ school principals, and a curriculum questionnaire for
the national research coordinators. All these questionnaires asked the respondents about
literacy related activities and resources that students experienced in and outside of school
as well as basic demographic information (Mullis, Kennedy, Martin, & Sainsbury, 2006).

In order to explore the relationships between motivation and reading
performance, the present study only used the data from the reading assessment, student
questionnaire, and teacher questionnaire. In particular, it focused on the motivational
variables from the student questionnaire, originally constructed into two dimensions by
PIRLS as shown in Table 2.1. The items for other background variables included in the
analysis of this study were presented in Appendix A. These background variables were
chosen because they are known to be associated with students’ reading performance
according to the previous studies. At a student-level, those variables included gender (e. g.,
Klecker, 2006), SES (e.g., Bowey, 1995), and the amount of literary and informational
reading outside of school (e.g., Anderson, Wilson, & Fielding, 1988). At a classroom-
level, the variables were class mean SES (e.g., NCES, 2002) and proportion of students
with reading difficulties.

Analysis

97

Factor analysis and multilevel regression were two main analytical methods for
this study. In order to address Factor analysis was used to derive more meaningful factors
for reading motivation in PIRLS. Both exploratory and conﬁrmatory factor analytic
methods were applied. In general, exploratory factor analysis deals with the question like
“what are the underlying processes that could have produced correlations among the
variables?” while conﬁrmatory factor analysis is related to questions like “are the
correlations among variables consistent with a hypothesized factor structure?”
(Tabachnick & Fidell, 2001, p. 585). Factor scores were obtained for each motivational
construct and then were used for further analyses.

Multilevel regression based on the Hierarchical Linear Modeling (Raudenbush &
Bryk, 2002) was a main statistical method to test the relationship between reading
motivation and reading performance. The PIRLS data inherently had a multilevel
structure with students nested within classes, for which the hierarchical linear modeling is
recommended as a more adequate statistical method. The study ﬁrst explored the
interacting relationships between different motivational facets and then looked into the
role of motivation in predicting reading performance when other background variables
were controlled. SPSS 15.0, AMOS 6.0 and HLM 6.0 were the statistical software
packages used in this study. Additional information about analyses procedures is
presented in the Results section.

Results

Results of this study identiﬁed a two-by-two four-factor model for reading

motivation by analyzing the data ﬁ'om the motivational items in the PIRLS student

questionnaire, and demonstrated that all these four factors were well-predictive of reading

98

performance even after controlling other literacy-related factors. In addition, this study
found that intemally-oriented motivational factors were more predictive than externally-
oriented factors. These results are reported below in three sections - the ﬁrst on factor
structure of reading motivation in PIRLS, the second on the interaction between
motivational factors when predicting reading performance, and the third on the prediction
of reading performance by reading motivation and other literacy-related background
information.

What Factors Constitute Reading Motivation in PIRLS?

Exploratory factor analysis. The present study ﬁrst performed exploratory factor
analysis in order to examine the underlying constructs of reading motivation in PIRLS.
Principal axis factoring was used as an extraction method with oblique rotation, because
multivariate normality is not assumed due to several negatively skewed variables and also
some correlations were assumed between different motivation constructs (Fabrigar,
Wegener, MacCallum, & Strahan, 1999). Four factors were assumed, drawing on the
scree plot which showed the four factors on the steep line before it meets the gradual
trailing line.

Two basic assumptions were tested by the Kaiser-Meyer—Olkin (KMO) measure
of sampling adequacy and the Bartlett’s test of sphericity. The KMO measure should be
greater than at least .50 to be grouped into a smaller set of latent factors, and the Bartlett
test should be statistically signiﬁcant to support the assumption that the variables are
correlated highly enough to provide a reasonable basis for factor analysis (Leech, Barrett,
& Morgan, 2005). In this study, the KMO measure was .75 and the Bartlett test was

statistically signiﬁcant (chi-square=10050.95, df=45, p<.001), which satisﬁes both

99

assumptions. In addition, 65.9% of the total variance was reported to be explained by the
ﬁrst four factors. The internal consistency reliability coefﬁcient of 10 motivational
variables in PIRLS was .71 by Cronbach’s alpha.

Table 2.3 displays the extracted four motivational constructs in addition to the
factor loadings of each variable. The ﬁrst factor was composed of the three variables in
column 1: ATR4 (I think reading is boring), ATR6 (I enjoy reading), and ATRl (I read
only if I have to). The second factor comprised the two variables in column 2: RSC2 (I do
not read as well as other students in my class) and RSC4 (I read slower than other
students in my class). The third factor was composed of the two variables in column 3:
RSC3 (When I am reading myself, I understand almost everything I read) and RSCl
(Reading is very easy for me). Finally, the fourth factor consisted of the three variables in
column 4: ATR2 (I like talking about books with other people), ATR3 (I would be happy
if someone gave me a book as a present), and ATRS (I need to read well for my ﬁlture).
Thus, these results suggested that each of the two PIRLS motivation dimensions, reading
attitudes and reading self-concept, comprised two components each. Inspection of the
items in each factor disclosed that those components formed the contrasting motivational
facets of each dimension. For example, factor 1 and factor 4 under the reading attitude
dimension represented intrinsic motivation and extrinsic motivation respectively. The
items of factor I asked students about their individual interest in reading, while those of
factor 4 were related to the external reasons for reading, namely socialization and
utilization. Factor 2 and factor 3 under the reading self-concept dimension also could be
labeled as peer-referenced perceived competence and self-referenced perceived

competence respectively. The concept of peer-referenced perceived competence for factor

100

2 was almost the same as that of self-concept based on external comparison which Marsh
(1986) described, in that the items asked students to rate their reading abilities
normatively in comparison with their peers. However, self-referenced perceived
competence for factor 3 was rather different from Marsh’s internal-comparison-based
self-concept. Intemal-comparison—based self-concept was deﬁned as students’ perceived
competence in one domain compared with their competence in other domains, whereas
self-referenced perceived competence in this study is viewed more generally as students’
perception of their own reading abilities independently of normative comparison with
others. In sum, this study found two contrasting motivational facets under each of two
reading motivation dimensions in PIRLS, as shown as tow-by-two reading motivational
constructs in Table 2.4. One of the two facets under each dimension represented the
intemally-oriented construct (i.e., intrinsic motivation and self-referenced perceived
competence), while the other facet implied the extemally-oriented construct (i.e.,
extrinsic motivation and peer-referenced perceived competence).

Factor scores and their correlations. From the exploratory factor analysis above,
this study derived factor score estimates using the Bartlett method for further analysis.
The Bartlett method uses least squares procedures to minimize the sum of squares of the
unique factors over the range of variables and it ensures unbiased factor score estimates
(Bartlett, 193 7). As Marsh (1986) did, the current study planned to use factor score
estimates for analytical purposes rather than to use unweighted scale scores such as sums
or means of the variable values. Table 2.5 displays the zero-order correlations between
factor score estimates of each derived factor and other related variables in PIRLS.

Contrasting facets with different internal or external orientations under the same

101

motivational dimensions had moderately positive correlations with each other (.21
between intrinsic and extrinsic motivation, while .20 between self-referenced and peer-
referenced perceived competence). Interestingly, the facets with the same internal or
external orientation under different motivational dimensions also had moderately positive
correlations (.25 between intrinsic motivation and self-referenced perceived competence,
while .21 between extrinsic motivation and peer-referenced perceived competence).
However, very small correlations were found between different facets with different
orientations under different dimensions (.08 between intrinsic motivation and peer-
referenced perceived competence, while -.06 between extrinsic motivation and self-
referenced perceived competence). This supports the conceptual organization of the
motivational facets and dimensions as given in Table 2.4.

Table 2.5 also shows that students’ performance scores correlated positively with
intrinsic motivation, self-referenced perceived competence, and peer-referenced
perceived competence, and the scores had no signiﬁcant correlations with extrinsic
motivation. In relation to background variables, student gender had positive correlations
with intrinsic/ extrinsic motivation but not with se1f-referenced/peer-referenced perceived
competence. That is, girls had relatively higher intrinsic and extrinsic motivation than
boys, but there were no differences between girls and boys in their self-referenced and
peer-referenced perceived competence on average. Under the reading attitude dimension,
both intrinsic and extrinsic motivation had moderately positive correlations with students’
literary reading outside of school. Extrinsic motivation was, however, more positively
related to student SES and informational reading outside of school than intrinsic

motivation. Under the dimension of student reading self-concept, peer-referenced

102

perceived competence correlated more positively with students’ background variables
than self-referenced perceived competence. For instance, students’ informational reading
outside of school correlated positively with peer-referenced perceived competence but
negatively with self-referenced perceived competence.

Conﬁrmatory factor analysis. Conﬁrmatory factor analysis was conducted to
assess the adequacy of the new factor structure for reading motivation constructed from
exploratory factor analysis in the present study. The new factor structure (2X2 factor
model: extrinsic/intrinsic motivation under one dimension and self-referenced/peer-
referenced perceived competence under the other dimension of reading motivation) was
compared with the original two factor model (reading attitudes and self-concept) of
reading motivation in PIRLS, using several ﬁt statistics. Table 2.6 shows this comparison
of the competing models by the chi-square/degrees of freedom ratio, the Root-Mean-
Square Error of Approximation (RMSEA), the Goodness—of-Fit Index (GFI), the Normed
Fit Index (NF 1), the Comparative Fit Index (CPI), and the Tucker-Lewis Index (TLI).

According to suggestions from prior research on conﬁrmatory factor analysis
(e. g., Hu & Bentler, 1999; Kline, 2005), smaller rather than larger values indicate a good
ﬁt for the chi-square tests. For RMSEA, values less than .05 indicate a close approximate
ﬁt, values between .05 and .08 are a reasonable error of approximation, and values
over .10 suggest a poor ﬁt. The remaining indices such as GFI, NFI, CFI, and TLI are
also considered as a reasonable good ﬁt if values are greater than roughly .90. In sum, the
ﬁt of the new 2X2 factor model of reading motivation was improved in comparison with
that of the original two factor model -- the new model proved to be superior to the old

model by these criteria.

103

Interaction Eﬂects of Motivational Facets

After the 2X2 factors for the PIRLS motivational variables were constructed and
conﬁrmed, this study explored how these factors work together in order to predict
students’ reading performance. In particular, the current study focused on the interactions
between the contrasting facets that comprised the same motivational dimensions, because
recent studies on academic motivation have suggested that motivation is a construct with
multidimensions and multifacets and that people have different motivational dimensions
and facets concomitantly toward an object or an activity (e.g., Bong, 1996; Ryan & Deci,
2000). It should be ﬁrst noticed that preliminary analysis indicated that extrinsic
motivation had a curvilinear relationship with reading performance scores showing an
inverted U-shaped function. That is, more extrinsic motivation predicted higher reading
performance until a certain point of extrinsic motivation, but it predicted lower reading
performance after that point. This curvilinear relationship was included as a square term
in following analyses (Aiken & West, 1991).

Table 2.7 demonstrates how intrinsic motivation and extrinsic motivation interact
to predict students’ reading performance scores from the multilevel regression output.
Model 2 showed an improved level-l variance explained from 10% to 12% by adding the
interaction terms (IM‘EM and IM "' EMz) to model 1, and the coefficient of one of the
interaction terms (IM*EM) was statistically signiﬁcant (t=7.95, p<.001). This conﬁrmed
the assumed interactions between intrinsic and extrinsic motivation, indicating that the
prediction of students’ reading performance by one of these motivational facets would
depend on the level of the other motivational facet. Figure 2.1 displays this relationship

graphically. The levels of motivation were divided into high (factor scores = +1 SD),

104

medium (factor scores = 0), and low (factor scores = -1SD). As seen in this graph,
extrinsic motivation had a generally negative relationship with reading performance for
students with low intrinsic motivation, whereas it had a positive relationship for students
with high intrinsic motivation. However, the high level of extrinsic motivation was
negatively related to reading performance for all levels of intrinsic motivation. Also,
students with higher levels of intrinsic motivation had better reading performance on
average regardless of their levels of extrinsic motivation. In summary, although extrinsic
motivation predicted reading performance for some students with high or medium levels
of intrinsic motivation, the results suggests that intrinsic motivation is important for all
students for better reading performance. Of special concern are students with low
intrinsic motivation, as any levels of extrinsic motivation always predicted reading
performance negatively for them.

Similarly to the interaction between intrinsic and extrinsic motivation, Table 2.8
demonstrates how self-referenced perceived competence and peer-referenced perceived
competence interact to predict students’ reading performance. By adding an interaction
term (SPC*PPC), model 2 explained slightly more variance in student’ reading
performance than did model 1 (from 16% to 17%), and this interaction was conﬁrmed by
the statistically signiﬁcant coeﬂicient of the interaction term (t=6.79, p<.001). Figure 2.2
displays this interactional relationship between self-referenced and peer-referenced
perceived competence in predicting students’ reading performance. According to this
graph, although peer-referenced perceived competence positively predicted reading
performance on the whole, this positive prediction was more salient for those students

with higher levels of self-referenced perceived competence. This suggests that how

105

students perceive their reading abilities not only in relation to others but especially in
relation to themselves is very important for reading performance.
Reading Motivation as a Predictor of Reading Performance

The last analysis was conducted to see how the four motivation variables and
interaction terms contribute, in concert with other various literacy related variables, to the
prediction of students’ reading performance. As explained in Methods above, literacy
related variables were chosen ﬁ'om the information in the student questionnaire and the
teacher questionnaire, and they include classroom SES, proportion of students with
reading diﬂiculties in the classroom, student gender, student SES, students’ informational
reading outside of school, students’ literary reading outside of school. After the listwise
deletion of missing data, the data from 4,668 students in 249 classrooms were analyzed
for this purpose, using Hierarchical Linear Modeling.

Table 2.9 shows the analysis results from three models. In the ﬁrst model, the
information only from classrooms was included in the analysis. The class mean SES and
the proportion of students with reading diﬂiculties in the classroom together accounted
for 41% of the between-class variance in students’ overall reading performance scores. In
the second model, the analysis added individual variables such as gender, SES, student’s
informational reading outside of school, and student’s literary reading outside of school.
Thirteen percent of the within-class variance in students’ overall reading performance
was explained by these student-level variables. Finally, the third model entered students’
four reading motivation facets into the analysis with their interaction terms, and these
motivation variables explained 17 % more of the within-class variance in students’

reading performance. This implies that students’ reading motivation, measured by

106

intrinsic/extrinsic motivation, self-referenced/peer-referenced perceived competence, and
their interaction terms, accounts for quite a large portion of the variance in students’
reading performance in PIRLS. In this model, other statistically signiﬁcant predictors of
students’ reading performance were classroom mean SES, student SES, and student’s
literary reading outside of school (positive predictors), and the proportion of students
with reading difﬁculties in class and student’s informational reading outside of school
(negative predictors) when the other variables were held constant. Gender, however, did
not predict reading performance signiﬁcantly when the other variables were considered in
this analysis. This contradicted the wide-spread belief in girls’ superiority over boys in
reading performance on average (e. g., Gates, 1961), which suggested that some variables
possibly mediated the relationship between student gender and reading performance.
Discussion

The present study investigated the multidimensional multifaceted nature of
reading motivation using the motivational variables ﬁom PIRLS. It also tested the
interactions between different motivational facets in predicting reading performance, and
how motivational variables hold up when other common predictors of reading
performance are entered into the model. The main ﬁndings are as follows:

First, this study found ﬁom exploratory factor analysis that each of the two
motivational dimensions hypothesized by PIRLS, reading attitudes and self-concept, can
bee divided into two contrasting but related facets by their internal or external
orientations. This brought intrinsic and extrinsic motivation facets under the reading
attitude dimension and self-referenced and peer-referenced perceived competence facets

under the reading self-concept dimension. Confrrmatory factor analysis validated the ﬁt

107

of this newly constructed reading motivation structure. This result is consistent with the
argument of recent studies that motivation is a multifaceted construct with multiple
constituents (e. g., Guthrie et a1, 2007; Murphy & Alexander, 2000). Thus, reading
motivation is not a simple construct but should be understood in the multidimensional
multifaceted context.

Second, the study demonstrated the interacting relationships of different
motivational facets with reading performance. Self-referenced perceived competence and
peer-referenced perceived competence interacted with each other when predicting
students’ reading performance. Peer-referenced perceived competence predicted reading
performance more positively when students had higher self-referenced perceived
competence in reading. The study also found an interaction between intrinsic motivation
and extrinsic motivation and a curvilinear relation between extrinsic motivation and
reading performance. Extrinsic motivation was positively related to students’ reading
performance only when their intrinsic motivation was high but extrinsic motivation was
not so high. Extrinsic motivation always had a negative relationship with reading
performance if students had low intrinsic motivation. This result is consistent with the
ﬁnding of Lin, McKeachie, and Kim (2003) which showed that college students in two
countries achieved better when they had medium extrinsic motivation coupled with
higher intrinsic motivation. These results also aﬁirm recommendations that educational
attention should be paid more to internally-oriented motivation such as intrinsic
motivation and self-referenced perceived competence than to extemally-oriented
motivation (e.g., Cameron & Pierce, 1994; Deci & Ryan, 1985; Wigﬁeld & Guthrie,

1997).

108

The above mentioned ﬁndings also offer one alternative explanation about
extrinsic motivation. In the ﬁeld of motivation research, the role of extrinsic motivation
has long been controversial (e.g., Reiss, 2005), especially concerning the undermining
effect of extrinsic motivation on intrinsic motivation. The current study suggests that this
undermining effect of extrinsic motivation depends on the levels of intrinsic motivation
and extrinsic motivation that students have toward an activity or an object. When students
had higher intrinsic motivation, moderate extrinsic motivation did not have a negative
relation but a positive relation with reading performance.

Finally, the study showed how motivation predicts students’ reading performance
when all four motivational constructs of this study and their interaction terms are present
in the model with other common predictors of reading performance. Speciﬁcally, the
background variables controlled in the analysis are student SES, student’s literary reading
outside of school, student’s informational reading outside of school, class mean SES, and
the proportion of students with reading diﬁiculties in classroom. The four motivational
facets and the interactions between contrasting motivational facets accounted for 17% of
the within-class variance in students’ reading performance scores above and beyond those
background variables. Similarly, Guthrie et al. (2007) also showed that reading
motivation of fourth-grade students explained between 3% and 22% of the variance in
their reading comprehension when prior reading performance was controlled. These
suggest that motivation may be a very important component of reading performance.

The motivational dimensions and facets of this study were constructed only from
the data available in PIRLS 2006. PIRLS originally collected the information about

students’ reading motivation within two dimensions, attitudes toward reading and reading

109

self-concept, and the current study conﬁrmed these two dimensions. However, motivation
in reading also concerns many other dimensions and factors such as goal orientation,
social goal, and interest, as Murphy and Alexander (2000) argued. Thus, the small
number of items and the narrow breadth in assessing reading motivation are the
limitations of this study as well as PIRLS. Nevertheless, this study oﬂ°ers important
implications for PIRLS and other reading studies by showing that even a simple
motivational dimension can be decomposed into diﬁ‘erent facets. Future studies will have
to include broader dimensions of reading motivation while paying attention to their
underlying facets at the same time. .

Among the motivational factors constructed in this study, the items under
extrinsic motivation may be controversial. In this study, the extrinsic motivation factor
was composed of three items associated with socialization and utilization in reading
(ATR2 —- I like talking about books with other people; ATR3 — I would be happy if
someone gave me a book as a present; ATRS — I need to read well for my future). The
item ATR2 is related to socialization in reading, while ATRS is close to utilization. The
item ATR3 involves both utilization (present) and socialization (someone gave me).
These three items are viewed as extrinsic in this study because they concern something
outside of reading itself as of the motivator for reading and the source of reading
enjoyment. Although utilization can be easily regarded as extrinsic motivation, there
would be less agreement about whether socialization is intrinsic or extrinsic in motivation.
Social motivation has been sometimes considered as a positive and intrinsic predictor of
learning (e.g., Anderman & Anderman, 1999). Some researchers have also

conceptualized social motivation as a separate motivational dimension which is distinct

llO

ﬁ'om academic motivation but is related to academic achievement (e.g., Urdan & Maehr,
1995; Wentzel, 1999). However, many researchers still categorize social motivation as
extrinsic motivation. In reading research, for example, Unrau and Schlackman (2006) and
Wang and Guthrie (2004) have included social motivation into extrinsic motivation in
their structural modeling of motivation and have shown their negative or null
relationships with reading achievement. Drawing on these studies, socialization in this
study was considered as a component of extrinsic motivation. More research will be
necessary to clarify the extrinsic or intrinsic role of social motivation in learning,
however. It was the limitation of this study as a secondary analysis that we are limited to
the items as written, without the modiﬁcation or addition of items.

This study also raised a question about the relation between gender and
motivation as a topic for future research. When the motivational variables were entered in
the multilevel regression model, the coefficient of student gender in predicting reading
performance diminished ﬁ'om 4.05 (p<.10) to 2.39 (n.s.). This suggests that motivation
might mediate the relationships between gender and reading performance. However, this
issue goes beyond the scope of the current study and more research is needed to clarify
these relationships.

In conclusion, the present study supported the important role of reading
motivation in relation to students’ reading performance. It also supported the
multidimensional and multifaceted nature of reading motivation. Those different
motivational facets interacted together to predict students’ reading performance. It
implies that reading motivation should be understood not as a simple direct predictor of

reading performance but as a complex system in which various motivational components

111

work reciprocally. The ﬁnal implication of this study is that internally-oriented
motivational facets should be emphasized in the development of reading. Although
externally-oriented motivational facets, such as extrinsic motivation and peer-referenced
perceived competence, positively predicted students’ reading motivation in some cases in
this study, their positive contribution depended on the level of students’ intemally-
oriented motivational facets. Intemally-oriented motivational facets, such as intrinsic
motivation and self-referenced perceived competence, were solid predictors of better
reading performance. These interacting relationships between motivational constructs
have been rarely studied because it brings about analytical diﬁiculties to include
interaction variables in a statistical model. As this study showed, however, those
neglected interaction eﬁects of motivation provide us with alternative and very important

interpretations about the role of different motivational constructs.

112

Table 2.1

Motivational Variables in PIRLS

 

 

Statement in the PIRLS questionnaire Variable
name
Student attitudes toward reading SATR a
I read only if I have to. ATRl ®
I like talking about books with other people. ATR2
I would be happy if someone gave me a book as a present. ATR3
I think reading is boring. ATR4 ®
I need to read well for my future. ATR5
I enjoy reading. ATR6
Student reading self-concept SRSC a
Reading is very easy for me. RSCl
I do not read as well as other students in my class. RSC2 ®
When I am reading by myself, I understand almost everything I read. RSC3
I read slower than other students in my class. RSC4 ®

 

Note. The ATR5 variable was eliminated when PIRLS constructed the SATR variable.

Reverse coded for the analysis

3 These derived variables were coded as O—Low, l—Medium, 2—High based on the responses to the

questions under each dimension. All other variables were coded as 0—Disagree a lot, l—Disagree a little, 2—

Agree 3 little, 3—Agree a lot with the exception of reverse-coded variables, in which the variables were

coded as 3—Disagree a lot, 2-Disagree a little, 1—Agree a little, 0—Agree a lot.

113

Table 2.2
Descriptive Statistics and Intercorrelations of Motivational Variables, Background

Variables & Reading Scores 0V =5, 1 90 students)

 

 

 

 

Zero-order correlations
Variables M SD Student attitudes toward reading
ATRl ATR2 ATR3 ATR4 ATR5 ATR6 SATR
Motivation
variables
ATRl ® 1.39 1.18 1
ATR2 1.33 1.09 ,10’ 1
ATR3 2.03 1.05 .19' .32' 1
® t It It
ATR4 2.08 1.11 .38 .23 .42 l
ATR5 2.45 .90 .02 ,20’ ,21’ ,13’ 1
ATR6 2.21 1.02 .31’ .34" .52’ .65’ .26’ 1
t t It I t It
SATR 1.26 .68 .55 .51 .64 .73 .22 .75 1
t t t t t t #
RSCl 2.41 .79 .10 .10 .14 .16 .07 .23 .18
t t t It t
RSC2® 1.67 1.14 .20 -.01 .06 .17 .00 .12 .15
$ t t t I! It It
RSC3 2.45 .81 .09 .14 .15 .16 .14 .25 .20
RSC4 ® 1.86 1.10 .19' -.01 .01 ,16' .00 09* ,12’
t It It II III 1|: It
SRSC 1.46 .58 .19 .05 .10 .20 .06 .20 .20
Background
variables
Gender at a a a a a a
(04’0” 1_ girls) .50 .50 .11 .06 .11 .13 .05 .18 .16
It t t t t It
SES .22 1.22 .10 .07 .03 .01 .08 .06 .07
Informational .- .. * .. ,, ... ...
reading outside --09 1.25 -.04 .30 .23 .10 .18 .21 .19
of school
Lita?” reading a a a a a a a
outsrde of .24 1.82 .32 .29 .37 .38 .16 .47 .49
school
Reading scores
Informational 534.06 64.89 .33. -.03 ,12* .28. .10‘ .22‘ .26“
, a a a a a a a
Literary 537.29 71 .98 .38 -.04 .10 .30 .09 .23 .27
t III # t t t t
Overall 536.62 70.04 .35 -.03 .11 .29 .10 .22 .26

 

 

 

114

Table 2.2 Continued

 

 

 

 

Zero-order correlations
Variables Student reading self-concept
RSCl RSC2 RSC3 RSC4 SRSC
Motivation variables
ATR] ®
ATR2
ATR3
ATR4 ®
ATR5
ATR6
SATR
RSCl 1
® *
RSC2 .35 1
RSC3 .36' .12’ 1
® II t #
RSC4 .31 .52 .16 1
:1- s- a a
SRSC .60 .71 .49 .68 1
Background variables
Gender "'
(0.130”, l-girls) .01 .03 .02 .01 .03
it It i t #
SES .1] .09 .ll .07 .13
Informational reading * * "
outside of school '06 "04 '12 ‘02 '02
Literary reading * * * * *
outside of school .19 .13 .20 .10 .19
Reading scores
_ a a s a a
Informational .27 .33 .24 .26 .36
. at t t t t
Literary .28 .33 .23 .25 .36
t 1 $ t *
Overall .28 .33 .24 .25 .36

 

 

Note. Pairwise deletion was used for missing data.

a:
Statistically signiﬁcant (p < .05)

® Reverse coded for the analysis

115

Table 2.3

Extracted Factors ﬁvm Exploratory Factor Analysis (N =4, 826)

 

Factor (Pattern Matrix)

 

 

 

Variable Mean SD Cronbach

a
1 2 3 4

ATR4 ® 2.09 1.11 .825 .71

ATR6 2.21 1.02 .561 .341

ATRl ® 1.40 1.18 .457

RSC2 ® 1.68 1.14 .727

RSC4 ® 1.87 1.10 .668

RSC3 2.45 .81 .577

RSCl 2.41 .79 .568

ATR2 1.33 1.09 .547

ATR3 2.04 1.04 .204 .545

ATR5 2.45 .90 .358

Eigenvalues (Rotated) 1.941 1.374 1.360 1.659

 

Note. Loadings < .20 were omitted.

® Reverse coded for the analysis

116

Table 2.4

2X2 Reading Motivational Constructs of PIRLS Derived ﬁom Factor Analysis

 

Facets under motivational dimensions

Motivational dimensions

 

 

Intemally-oriented Externally-oriented
Attitudes toward reading Intrinsic motivation Extrinsic motivation
Self-referenced Peer-referenced
Reading self-concept
perceived competence perceived competence

 

117

Table 2.5
Descriptive Statistics and Intercorrelations among New Motivational Factors,

Background Variables, & Overall Reading Scores (N =4, 826 students)

 

Zero-order correlations

 

 

Variables M SD
1M EM SPC PPC

Factor scores estimates

Intrinsic motivation (1M) 0 1.15 1

Extrinsic motivation (EM) 0 1.34 .21!‘ 1

3:88:38?“ 0 1

fjf;;§‘:’§§§§f,3§’°”"ed o 1.38 .08‘ 21* .20' 1
Background variables

Gender (0-boys, l-girls) .50 .50 .16]t .11‘ .02 .02

SES .25 1.19 .01 ,07' ,09’ ,13'

Litﬁgfgggﬁdmg -.09 1.25 05* 35* 407" .12'

18.33:? reading outside of .2 6 1.82 .414 39* . l 1* 22*
Reading scores

Informational 536.44 64.19 .32. .03 .33. .27‘

Literary 539.76 71.21 ,35' .01 ,34' .27"

Overall 539.15 69.18 .33. .01 ,34' 37"

 

Note. Pairwise deletion was used for missing data.

I
Statistically signiﬁcant (p < .05)

118

Table 2.6
Comparisons of the Fit Statistics between the 2 Factor Model and the 2X2 Factor Model

ﬁ'om Conﬁrmatory Factor Analysis

 

 

Model Chi-square RMSEA GFI NFI CF I TLI
1289.96
2 factor model .09 .94 .87 .88 .83
(df=34)
2X2 factor 655.32
.06 .97 .94 .94 .91
model (df=32)

 

119

Table 2.7
Interactions between Intrinsic Motivation (Ill/0 and Extrinsic Motivation (EM) in

Predicting Overall Reading Score

 

 

 

 

 

Model 1 Model 2
Level B S.E. t B S.E. t
Fixed Effect
1*. iii
Intercept 545.81 2.38 229.27 543.87 2.34 232.33
til ##I
Intrinsic motivation (1M) 15.57 .85 18.35 18.19 1.11 16.46
tit
Extrinsic motivation (EM) -3.14 .76 -4.12 -l .37 .82 -1.68
tit tit
EM2 -397 .38 -10.36 -3.96 .43 -9.26
*t*
IM * EM 4.77 .60 7.95
2
IM "' BM -.61 .36 -l.73
. 2 . 2
Random Effect Vanance df X Vanance df x
tilt it.
Intercept 1092.31 252 1657.05 1032.74 252 1607.68
Student-level 3614.1 1 3553.44
Level-1 variance
.10 .12
explained
‘1‘
p < .001

120

Table 2.8

Interactions between Self-referenced Perceived Competence (SPC) and Peer-referenced

Perceived Competence (PPC) in Predicting Overall Reading Score

 

 

 

 

 

Model 1 Model 2

Level B SE. B S.E t
Fixed Effect

it! tit
Intercept 538.42 2.44 220.57 537.15 2.41 222.61
Self-referenced

tit tit
Perceived Competence 10.30 .66 15.54 11.05 .66 16.65
(SPC)
Peer-referenced

*1!!! *t*
Perceived Competence 15.87 .81 19.53 15.84 .80 19.59
(PPC)

tit
SPC * PPC 3.86 .57 6.79
Random Effect Variance df Variance df )8
It!!! *tt
Intercept 1273.54 252 1983.13 1227.60 252 1941.31
Student-level 3371.67 3331.64
Level-1 variance explained .16 .17
ﬁt.
p < .001

121

Table 2.9

Results from HLM Analyses Predicting Overall Reading Performance (249 classes 4, 668

students)

 

Level

Model 1

 

B

S.E. t

 

Intraclass Correlation (ICC) from the one-way ANOVA model = .26

Fixed Effect
(Classroom variables)
Intercept
Class mean SES
Class reading drﬁiculties
(Individual variables)
Gender
SES
Informational reading
Literary reading
(Motivation variables)
IM
EM
EM}
1M "' EM
1M * 15M2
SPC
PPC
SPC "' PPC

547.10
39.20
-67.59

at
4.02 135.94
aaa

4.66 8.41
12.42 -5,44*

it

a

 

Random Effect
Classroom-level
Gender slope
SES slope
Informational reading slope
Literary reading slope
Student-level

Variance
800.67

3964.56

df x2
*I
246 1164.81

t

 

Variance explained (R2)

At classroom mean: .41

 

Conditional ICC

.17

 

Table 2.9 Continued

 

 

 

 

 

 

 

 

Model 2 Model 3

Level B S.E. t B S.E. t

Fixed Effect

(Classroom variables)
Intercept 539.92 3.66 14751” 543.09 3.44 157.77”
Class mean SES 24.46 4.09 5.93"" 25.09 3.83 6.56"”
.31-gagging -53.48 10.67 -501." -43.25 9.92 4.36".

(Individual variables)
Gender 4.05 2.23 1,311 2.39 2.13 1.12
SES 5.52 1.03 533"" 4.06 .95 4.27""
fﬁt‘m‘ -11.30 .95 -1190: -9.28 .90 40.32:"
Literary reading 10.14 .65 15,51 5.14 .67 7,67

(Motivation variables)
IM 9.94 1.11 3.94m
EM -221 .84 .2453"
EM2 -2.87 .40 -720".
IM * EM 3.05 .54 562""
IM * EM2 .21 .33 .65
SPC 9.69 .65 14,31”
PPC 10.81 .82 13.19""
SPC * PPC 2.70 .57 4.71"”

Random Effect Variance (if x2 Variance df 17
Classroom-level 733.36 243 606,73” 604.19 243 639.41”
Gender slope 169.67 245 281,701 179.16 245 296.00'
SES slope 41.51 245 29731" 22.72 245 290.49‘
ﬁlﬁgﬁﬁ 30.21 245 329,15” 16.04 245 31591"
512;? reading 7.58 245 266.17 7.28 245 230,24'r
Student-level 3429.95 2789.06

Variance explained (R2) At student-level: .13 At student-level: .30

Conditional ICC .18 . 18

1- * it iiﬁ
p<.10; p<.05; p<.01; p<.001

123

 

 

 

 

 

lntrlnalc
Motivation
“‘ ”High IM
560~ ----- Medium IM

~-*-Low 1M

2

o

0

w ..............................

35404 ......................

13 -------

u

o

a:

i - -----------------------

0 “* ~.

5 520-

500-
I I ‘1
Low Medium High
Extrinsic Motivation

Figure 2.1. Interaction of extrinsic and intrinsic motivation

124

 

580'" Self-referenced

Percelved
Competence

— High SPC
----- Medium SPC
--- Low SPC

U1
0)
O

L

 

Overall Reading Score
E
C.’

O
u
a
I
O
a
a
O
o
a
O
O
O
.0
o
O
O
0
I
'9
I

.o
d"
f

01

M

O
l

500-

 

 

 

Low Medium Hi‘gh
Peer-referenced Perceived Competence

Figure 2. 2. Interaction of self-referenced and peer-referenced perceived competence

125

References

Aiken, L. S., & West, S. G (1991). Multiple regression: Testing and interpreting
interactions. Newbury Park, CA: SAGE Publications.

Ainley, M., Hidi, S., & Berndorﬁ', D. (2002). Interest, learning, and the psychological
processes that mediate their relationship. Journal of Educational Psychology, 94,
545-561.

Alexander, P. A., Murphy, P. K., Woods, B. S., Duhon, K. E., & Parker, D. (1997).
College instruction and concomitant changes in students’ knowledge, interest, and

strategy use: A study of domain learning. Contemporary Educational psychology,
22, 125-146.

Ames, C., & Archer, J. (1988). Achievement goals in the classroom: Students’ learning
strategies and motivation processes. Journal of Educational Psychology, 80, 260-
267.

Anderman, L. H., & Anderman, E. M. (1999). Social predictors of changes in students’
achievement goal orientations. Contemporary Educational Psychology, 25, 21-37.

Anderson, R. C., Wilson, P. T., & Fielding, L. G. (1988). Growth in reading and how
children spend their time outside of school. Reading Research Quarterly, 23, 285-
303.

Baer, J., Baldi, S., Ayotte, K., & Green, P. (2007). The reading literacy of U.S. fourth-
grade students in an international context: Results ﬁom the 2001 and 2006
Progress in International Reading Literacy Study (PIRLS) (N CES 2008—017).
National Center for Education Statistics, Institute of Education Sciences, U.S.
Department of Education. Washington, DC.

Bandura, A. (1977). Self-efficacy: Toward a unifying theory of behavioral change.
Psychological Review, 84, 191 —2 l 5 .

Bartlett, M. S. (1 93 7) The statistical conception of mental factors. British Journal of
Psychology, 28, 97—104.

Bong, M. (1996). Problems in academic motivation research and advantages and
disadvantages of their solutions. Contemporary Educational Psychology, 21, 149-
165.

Bong, M. (1998). Tests of the intemal/extemal frames of reference model with subject-
speciﬁc academic self-efﬁcacy and frame-speciﬁc academic self-concepts.
Journal of Educational Psychology, 90, 102-110.

Bong, M., & Skaalvik, E. M. (2003). Academic self-concept and self-efﬁcacy: How
diﬂerent are they really? Educational Psychology Review, 15, 1-40.

126

Bowey, J. A. (1995). Socioeconomic status difference in preschool phonological
sensitivity and ﬁrst-grade reading achievement. Journal of Educational
Psychology, 87, 476-487.

Brophy, J. (2004). Motivating students to learn (2nd ed.). Mahwah, NJ: Lawrence
Erlbaum Associates.

Byme, B. M., & Shavelson, R. J. (1986). On the structure of adolescent self-concept.
Journal of Educational Psychology, 78, 474-481.

Byme, B. M., & Worth Gavin, D. A. (1996). The Shavelson model revisited: Testing for
the structure of academic self-concept across pre-, early, and late adolescents.
Journal of Educational Psychology, 88, 215-228.

Cameron, J ., & Pierce, W. D. (1994). Reinforcement, reward, and intrinsic motivation: A
meta-analysis. Review of Educational Research, 64, 363-423.

Chapman, J. W. (1988). Cognitive-motivational characteristics and academic
achievement of learning disabled children: A longitudinal study. Journal of
Educational Psychology, 80, 357-365.

Chapman, J. W., Tunmer, W. E., & Prochnow, J. E. (2000). Early reading-related skills
and performance, reading self-concept, and the development of academic self-
concept: A longitudinal study. Journal of Educational Psychology, 92, 703-708.

Covington, M. V. (2000). Intrinsic versus extrinsic motivation in schools: A reconciliation.
Current Directions in Psychological Science, 9, 22-25.

Covington, M. V., & Mueller, K. J. (2001). Intrinsic versus extrinsic motivation: An
approach/avoidance reformulation. Educational Psychology Review, I 3, 157-176.

deCharms, R. (1968). Personal causation. New York: Academic Press.

Deci, E. L., Koestrrer, R., & Ryan, R. M. (1999). A meta—analytic review of experiments
examining the effects of extrinsic rewards on intrinsic motivation. Psychological
Bulletin, 125, 627-668.

Deci, E. L., & Ryan, R. M. (1985). The general causality orientations scale: Self-
determination in personality. Journal of Research in Personality, 19, 109-134.

Eccles, J. S., & Midgley, C. (1990). Changes in academic motivation and self-perception
during early adolescence. In R. Montemayor, G R. Adams, & T. P. Gulotta (Eds.),
From childhood to adolescence: A transitional period? (pp. 134-155). Newbury
Park, CA: Sage.

Eccles, J. S., VVrgﬁeld, A., & Schiefele, U. (1998). Motivation to succeed. Hoboken, NJ:
John Wiley & Sons.

127

Eisenberger, R., & Cameron, J. (1996). Detrimental effects of reward: Reality of myth?
American Psychologist, 51 , 1 153-1166.

Elliot, A. J. (1999). Approach and avoidance motivation and achievement goals.
Educational Psychologist, 34, 169-189.

Fabrigar, L. R., Wegener, D. T., MacCallum, R. C., & Strahan, E. J. (1999). Evaluating
the use of exploratory factor analysis in psychological research. Psychological
Methods, 4, 272-299.

F ishbein, M., & Ajzen, I. (1975). Belief attitude, intention and behavior: An introduction
to theory and research. Reading, MA: Addison-Wesley.

Gans, A. M., Kenny, M. C., & Ghany, D. L. (2003). Comparing the self-concept of
students with and without learning disabilities. Journal of Learning Disabilities,
36, 285-293.

Gates, A. I. (1961). Sex differences in reading ability. The Elementary School Journal, 61 ,
43 1-434.

Gough, P. B., & Tunmer, W. E. (1986). Decoding, reading, and reading disability. RASE:
Remedial & Special Education, 7, 6—10.

Guay, F., Marsh, H. W., & Boivin, M. (2003). Academic self-concept and-academic
achievement: Developmental perspectives on their causal ordering. Journal of
Educational Psychology, 95, 124-136.

Guthrie, J. T., Hoa, A. L. W., Wigﬁeld, A., Tonks, S. M., Humenick, N. M., & Littles, E.
(2007). Reading motivation and reading comprehension growth in the later
elementary years. Contemporary Educational Psychology, 32, 282-313.

Guthrie, J. T., & Wigﬁeld, A. (1999). How motivation ﬁts into a science of reading.
Scientific Studies of Reading, 3, 199-205.

Guthrie, J. T., Wigﬁeld, A., Metsala, J. L., & Cox, K. E. (1999). Motivational and
cognitive predictors of text comprehension and reading amount. Scientiﬁc Studies
of Reading, 3, 231-256.

Harter, S. (1981). A new self-report scale of intrinsic versus extrinsic orientation in the
classroom: Motivational and informational components. Developmental
Psychology, 17, 300-312.

Harter, S., & Jackson, B. K. (1992). Trait vs. nontrait conceptualizations of
intrinsic/extrinsic motivational orientation. Motivation and Emotion. 16, 209-230.

Hidi, S., & Renninger, K. A. (2006). The four-phase model of interest development.
Educational Psychologist, 41, 111-127.

128

Hu, L., & Bentler, P. M. (1999). Cutoff criteria for ﬁt indexes in covariance structure
analysis: Conventional criteria versus new alternatives. Structural Equation
Modeling, 6, 1-55.

Klecker, B. M. (2006). The gender gap in NAEP fourth, eighth, and twelﬁh-grade reading
scores across years. Reading Improvement, 43, 50-56.

Kline, R. B. (2005). Principles and practice of structural equation modeling (2nd ed.).
New York: Guilford Press.

Leech, N. L., Barrett, K. C. & Morgan, (1 A. (2005). SPSS for Intermediate Statistics:
Use and Interpretation (2“! ed. ). Mahwah, NJ: Lawrence Erlbaum Associates.

Lepper, M. R., Corpus, J. H., & Iyengar, S. S. (2005). Intrinsic and extrinsic motivational
orientations in the classroom: Age differences and academic correlates. Journal of
Educational Psychology, 97, 184-196.

Lepper, M. R., Green, D., & Nisbett, R. E. (1973). Undermining children’s intrinsic
interest with extrinsic reward: A test of the ‘overjustiﬁcation’ hypothesis. Journal
of Personality and Social Psychology, 28, 129-137.

Lin, Y., McKeachie, W. J ., & Kim, Y. C. (2001). College student intrinsic and/or extrinsic
motivation and learning. Learning and Individual Diﬂerences, 13, 251-258.

Maehr, M. L., & Meyer, H. A. (1997). Understanding motivation and schooling: Where
we’ve been, where we are, and where we need to go. Educational Psychology
Review, 9, 371-409.

Marsh, H. W. (1986). Verbal and math self-concepts: An intemal/external ﬂame of
reference model. American Educational Research Journal, 23 , 129-149.

Marsh, H. W. (1990). Inﬂuences of internal and external ﬂames of reference on the
formation of math and English self-concepts. Journal of Educational Psychology,
82, 107-116.

Marsh, H. W., Trautwein, U., Lﬁdtke, 0., Killer, 0., & Baumert, J. (2005). Academic
self-concept, interest, grades, and standardized test scores: Reciprocal eﬂects
models of causal ordering. Child Development, 76, 397-416.

McKenna, M. C., Kear, D. J., & Ellsworth, R. A. (1995). Children’s attitudes toward
reading: A national survey. Reading Research Quarterly, 30, 934-956.

Mullis, I. V. S., Martin, M. 0., Kennedy, A. M., & Foy, P. (2007). PIRLS 2006
international report: IEA Is progress in international reading literacy stuay in
primary schools in 40 countries. Chestnut Hill, MA: International Study Center,
Boston College.

Murphy, P. K., & Alexander, P. A. (2000). A motivated exploration of motivation

129

terminology. Contemporary Educational Psychology, 25, 3-53.

National Center for Education Statistics. (2002). The condition of education 2002 (N CES
2002-025). Washington, DC: U.S. Government Printing Ofﬁce.

National Institute of Child Health and Human Development (NICHD). (2000). Report of
the National Reading Panel. Teaching children to read: An evidence-based
assessment of the scientiﬁc research literature on reading and its implications for
reading instruction: Reports of the subgroups (NIH Publication No. 00-4754).
Washington, DC: U.S. Government Printing Office.

O’Flahaven, J. O., Gambrell, L. B., Guthrie, J., Stahl, S., Baumann, J. F., & Alvermann,
D. E. (1992). Poll results guide activities of research center. Reading Today, 10,
12.

Pintrich, P. R. (1994). Continuities and discontinuities: Future directions for research in
educational psychology. Educational Psychologist, 29, 137-148.

Pintrich, P. R., & Schunk, D. H. (1996). Motivation in education: Theory, research, and
applications. Englewood Cliffs, NJ: Merrill Prentice-Hall.

Raudenbush, S. W., & Bryk, A. S. (2002). Hierarchical linear models: Applications and
data analysis methods (24“ ed.). Thousand Oaks, CA: Sage Publications.

Reiss, S. (2005). Extrinsic and intrinsic motivation at 30: Unresolved scientiﬁc issues.
Behavior Analyst, 28, 1-14.

Rosenberg, M. (1979). Conceiving the self New York: Basic Books.

Ryan, R. M., & Deci, E. L. (2000). Intrinsic and extrinsic motivations: Classic deﬁnitions
and new directions. Contemporary Educational Psychology. 25 , 54-67.

Sales, E., Lichtenwalter, S., & Fevola, A. (2006). Secondary analysis in social work
research education: Past, present, and future promise. Journal of Social Work
Education, 42, 543-558.

Schiefele, U. (1999). Interest and learning ﬂom text. Scientiﬁc Studies of Reading, 3,
257-279.

Shavelson, R. J ., Hubner, J. J ., & Stanton, G C. (1976). Self-concept: Validation of
construct interpretations. Review of Educational Research, 46, 407-4-41.

Stanovich, K. E. (1986). Matthew eﬂ‘ects in reading: Some consequences of individual
differences in the acquisition of literacy. Reading Research Quarterly, 21, 360-
406.

Tabachnick, B. G., and Fidel], L. S. (2001). Using Multivariate Statistics (4“ ed. ). Boston:
Allyn and Bacon.

130

Unrau, N., & Schlackman, J. (2006). Motivation and its relationship with reading
achievement in an urban middle school. Journal of Educational Research, 100,
81-101.

Urdan, T. C., & Maehr, M. L. (1995). Beyond a two-goal theory of motivation and
achievement: A case for social goals. Review of Educational Research, 65, 213-
243.

Vansteenkiste, M., Timmermans, T., Lens, W., Soenens, B., & Van den Broeck, A. (2008).
Does extrinsic goal ﬂaming enhance extrinsic goal-oriented individuals’ learning
and performance? An experimental test of the match perspective versus self-
determination theory. Journal of Educational Psychology, 100, 387-397.

Wallbrown, F. H., Levine, M. A., Engin, A. W. (1981). Sex differences in reading
attitudes. Reading Improvement, 18, 226-234.

Wang, J. H. & Guthrie, J. T. (2004). Modeling the effects of intrinsic motivation, extrinsic
motivation, amount of reading, and past reading achievement on text
comprehension between U.S. and Chinese students. Reading Research Quarterly,

39, 162-186.

Wentzel, K. R. (1989). Adolescent classroom goals, standards for performance, and
academic achievement: An interactionist perspective. Journal of Educational
Psychology, 81, 131-142.

Whang, P. A., & Hancock, G. R. (1994). Motivation and mathematics achievement:
Comparisons between Asian-American and non-Asian students. Contemporary
Educational Psychology, 19, 302-322.

Wigﬁeld, A. (2000). Facilitating children’s reading motivation. In L. E. Baker, M. J. E.
Dreher & J. T. E. Guthrie (Eds.), Engaging young readers: Promoting
achievement and motivation (pp. 140-158). New York: Guilford Publications.

Wigﬁeld, A., Eccles, J. S., Yoon, K. S., Harold, R. D., Arbreton, A. J. A., Freedman-Dean,
C., et a1. (1997). Change in children’s competence beliefs and subjective task
values across the elementary school years: A 3-year study. Journal of Educational
Psychology, 89, 451-469.

Wigﬁeld, A., & Guthrie, J. T. (1997). Relations of children’s motivation for reading to the
amount and breadth or their reading. Journal of Educational Psychology, 89, 420-
432.

Wittrock, M. C. (1986). Cognitive science in technical training and basic skills education.
Advances in Reading/Language Research, 4, 291-305.

131

APPENDICES

132

APPENDIX A

Description of the Variables used for the Study

133

A1. Country-level variables

1.

Infant mortality rate: The number of infants who die before reaching on year of
age per 1,000 live births in a given year.

Net enrollment rate (%): Percentage of children of ofﬁcial school age who are
enrolled in school to the population of the corresponding ofﬁcial school age
based on the national education system.

Hours of formal reading instruction: Average hours per week spent on formal
reading instruction, which was answered by teachers.

Average class size: National average class size answered by teachers.

Students in poverty schools (%): Percentage of students who attend the schools
with more than 25% of economically disadvantaged children, which was
answered by school principals.

Achievement gap between students in low and high poverty schools: Average
score difference in overall reading performance between the students attending
schools with 25% or less of economically disadvantaged children and the
students attending schools with over 25% of economically disadvantaged
children.

Teaching informational texts weekly (%): Percentage of students whose teachers
answered they taught informational texts at least weekly. (Description and
explanations about things, people, or events; instructions or manuals about how

things work; and charts, diagrams, or graphs)

134

8. Teaching literary texts weekly (%): Percentage of students whose teachers
answered they taught literary texts at least weekly. (Short stories; longer books
with chapters; poems; and plays)

9. Teaching text style or structure weekly (%): Percentage of students whose
teachers answered they asked them to describe text style or structure at least
weekly.

10. Major emphasis on text style and structure before 4’” grade (%): Percentage of
students attending schools whose reading curriculum gives a major emphasis on
text style and structure before fourth grade according to school principals’
response

11. Emphasis on various reading purpose: Rasch score of the responses ﬂom
national research coordinators to the following ﬁve questions.

How much emphasis does the reading curriculum place on the following
purposes for reading?

a. Reading to improve reading
b. Reading for literary experience
c. Reading to acquire information
d. Reading for social awareness/civic duty
e. Reading for enjoyment

A2. Classroom-level variables

1. Class mean SES: Average SES of the students in the same class (ﬂom the

students questionnaire).

135

2. Weekly instructional time for reading: Hours for reading instruction per week
according to teacher responses.

3. Proportion of students with reading diﬂiculties: The number of students who
need remedial reading instruction in a class divided by the number of all students
in that class according to teacher responses.

4. Reading comprehension strategy instruction: Rasch score of the responses ﬂom
teachers to the following seven questions.

How often do you ask the students to do the following things to help develop
reading comprehension skills or strategies?

a. Identify the main ideas of what they have read

b. Explain or support their understanding of what they have read

c. Compare what they have read with experiences they have had

d. Compare what they have read with other things they have read

e. Make predictions about what will happen next in the text they are reading

f. Make generalizations and draw inferences based on what they have read

g. Describe the style or structure of the text they have read

5. Computer use for reading instruction: Rasch score of the responses ﬂom
teachers to the following three questions.

How often do you have students do the following computer activities?
a. Look up information on the intemet
b. Read stories or other texts on the computer

c. Use the computer to write stories or other texts

136

6. Autonomy support in reading instruction: Rasch score of the responses from

teachers to the following two questions.

a.

When you have reading instruction and/or do reading activities, how
often do students work independently on a goal they choose themselves?
When do have reading instruction and/or do reading activities with the
students, how often do you give students time to read books of their own

choosing?

7. Social activities in reading instruction: Rasch score of the responses from

teachers to the following ﬁve questions.

a.

When you have reading instruction and/or do reading activities with the
students, how often do you ask students to read aloud in small groups or
pairs?

When you have reading instruction and/or do reading activities, how
oﬁen do you create same-ability group?

When you have reading instruction and/or do reading activities, how
often do you create mixed-ability group?

As students have read something, how often do you ask them to talk with
each other about what they have read?

How often do you have students use computers to email or chat with

other students about what they are learning?

8. Homework for reading: Rasch score of the responses ﬂom teachers to the

following two questions.

a.

How often do you assign reading as part of homework (for any subject)?

137

b. In general, how much time do you expect students to spend on homework
involving reading each time you assign it?
9. Instruction for informational reading: Rasch score of the responses ﬂom
teachers to the following six questions.

When you have reading instruction and/or do reading activities with the
students, how often do you have the students read the following types of text?
a. Descriptions and explanations about things, people, or events (non-

ﬁction)

b. Instructions or manuals about how things work

c. Charts, diagrams, graphs

When you have reading instruction and/or do reading activities with the
students, how often do you use the following resources?

(1. Children’s newspapers and/or magazines

e. Reading material on the Internet (Web pages)

f. Materials ﬂom other subjects

10. Instruction for literary reading: Rasch score of the responses from teachers to

the following four questions.

When you have reading instruction and/or do reading activities with the
students, how often do you have the students read the following types of text?
a. Short stories (e. g., fables, fairy tales, action stories, science ﬁction,

detective stories)
b. Longer books with chapters (ﬁction)

0. Poems

138

(1.

Plays

A3. Students-level variables

1.

2.

Gender: Dichotomous variable on gender (0 — boys, 1 — girls)

Books at home: Dichotomous variable on the quantity of books at students’

home (0 — 25 or fewer books, 1 — more than 25 books)

Amount of reading outside of school I : Dichotomous variable on students’ daily

amount of book or magazine reading outside of school (0 - Others, 1 — No

reading)

Amount of reading outside of school 2: Dichotomous variable on students’ daily

amount of book or magazine reading outside of school (0 — Others, 1 — A great

deal of reading(5 hours and more))

SES: Rasch score of the responses ﬂom students to the following nine questions.

Do you have any of these things at your home?

a.

b.

Computer (do not include TV/video game stations)

Study desk/table for your use

Books of your very own (do not count your school books)

Daily newspaper

Your own room

Your own mobile phone

Video or digital Camera (U .S), automatic dishwasher (Denmark), your
own television (England), parent’s car for private use (Taiwan)

More than one car (US), reference books (Denmark), any musical

instruments (England), private teacher (Taiwan)

139

i. More than one bathroom (U .8), Internet connection (Denmark), your own
CD and/or DVD player (England), place to read with no disturbances
(Taiwan)
6. Attitudes toward reading: Rasch score of the responses ﬂom students to the
following six questions.
What do you think about reading? Tell me how much you agree with each of
these statements?
a. I read only if I have to
b. I like talking about books with other people
c. I would be happy if someone gave me a book as a present
(1. I think reading is boring
e. I need to read well for my future
f. I enjoy reading
7. Reading self-concept: Rasch score of the responses ﬂom students to the
following four questions.
How well do you read? Tell me how much you agree with each of these
statements?
a. Reading is very easy for me
b. I do not read as well as other students in my class
c. When I am reading by myself, I understand almost everything I read
(1. I read slower than other students in my class
8. Computer use: Rasch score of the responses ﬂom students to the following three

questions.

140

a.

b.

C.

How often do you use a computer at home?
How often do you use a computer at school?

How often do you use the Internet to look up information for school?

9. Informational reading outside of school: Rasch score of the responses ﬂom

students to the following six questions.

How often do you do these things outside of school ?

a.

b.

6.

f.

I read to ﬁnd out things I want to learn
I read books that explain things

I read magazines

I read newspapers

1 read directions or instructions

I read brochures and catalogues

10. Literary reading outside of school: Rasch score of the responses ﬂom students to

the following two questions.

How often do you do these things outside of school ?

a.

b.

I read stories or novels

I read for fun outside of school

141

APPENDIX B

Analysis Results of the Data ﬂom Denmark

142

Table B.1
Means, Standard Deviations, and Intercorrelations of C lassroom-level Variables and

Reading Performance (205 classrooms in Denmark)

 

 

 

Class-level Zero-order correlations
variables (Denmark) Mean SD'
1 2 3 4

General Classroom Variables
1. Class mean SES .91 .53 1
2. Weekly mstructronal time 4.57 2.94 .02 1
for readmg
3. Proportion of students with an
reading difficulties -15 -08 -.30 -.10 1
4. Reading comprehension
strategy instruction '12 1'18 -03 ~09 -. 10 1
5. Computer use for reading us
instruction .02 1.09 -.04 .27 -.08 . 10
6. Autonomy support in n- as:
reading instruction -.02 1.09 .09 .15 -,13 '20

. . .. . . tit # ttt
7. Socral actrvrtres in reading .00 1.00 .00 .26 -, 17 .39
8. Homework for reading -.02 1.09 -.09 19" .00 ,13'

Genre-related Classroom Variables

9. Instruction for u- ..
Informational reading "07 1'07 -06 .22 -.07 .13
10. Instruction for literary :- a am"
reading ~00 -10 .14 -12 -.14 .35
Reading Performance
11. Class informational "a tea
reading score 544.11 30.10 .46 .05 -,37 -.05

. . *tﬁ tit
12. Class literary readrng score 550.06 28.11 .48 .04 -.37 -.05

_ it. ﬁt!

13. Class overall readmg score 548.76 29.01 .47 .06 ~ -,37 -.05
14. % of students in class with
large score gap favoring 16.83 9.02 -.12 -.05 .09 .02

literary reading

15. % of students in class with
large score gap favoring 7.15 6.22 -.13 .05 .00 -.03
informational reading

 

143

Table B.1 Continued

 

 

Class- Zero-order correlations
level
variables 5 6 7 8 9 10 ll 12 13 14

 

General Classroom Variables

l.
2.
3.
4.
5. 1
iii
6. .35 l
tit #
7- .18 .18 l
8. .04 .05 .09 l

Genre-related Classroom Variables

it! it. it!

9- .39 .28 .25 -01 1
it! it t
10. .14 .33 .24 .13 .15 1
Reading Performance
11. -.03 .04 -.12 -.08 -.01 .07 1
ﬁt!
12. -.04 .04 -.13 -.08 -.02 .05 .97 l
’6‘ ill
13. -.04 .04 -.12 -.08 -.02 .06 .93 .99 1
**# it! ttt
14. -.02 .03 .03 .07 .00 -.12 -,42 -,25 -33 1
iii
15. .00 -.12 -.02 -.01 .00 .02 .06 -.09 -.02 -,3o

 

# it

t t t
p<.05; p<.01; p<.001

144

Table B.2
Means, Standard Deviations, and Intercorrelations of Student-level Variables and

Reading Performance (3, 786 students in Denmark)

 

 

 

Student-level Mean S D Zero-order correlations

variables (Denmark) ' ' 1 2 3 4 5
General Student Variables

1. Gender (O-Boy, l-Girl) .52 .50 l

2. Books at home (O-Few

books, 1-Many books) .75 .43 .02 1

3. Amount of reading as... ......

outside of school 1 — No .20 .40 -.16 -,11 1

reading (O-No, l-Yes)

4. Amount of reading

outside of school 2 — A m

great deal of reading (0- '03 '18 '03 '00 "09 1

No, l-Yes)

5. SES .95 1.41 .03 30"" -.ogm -.03 1

6. Aﬂ'tud t d an an an ”a as:
reading es mm .08 1.12 .17 .15 -.31 .14 .11

7. Reading self-concept .58 1.40 -.01 ,16’" -.10." .03‘. .151."
8. Computer use -.27 1.29 .,03"' .01 -.05" .02 .05"
Genre-related Student Variables

9. Informational reading 4'" an

outside of school -.37 1.34 -.01 -.03 -20 .14 .00
10, Literary reading an art-1: can tun ate
outside of school -40 1-63 .19 .13 -.33 .13 .12
Reading Performance

11. Informational sat an" an" a an:
reading we 546.94 66.07 .08 .30 -.15 -.04 .26
ﬁfgrlgmmmd’“ 552.39 62.48 .10’" .28’" -.13’" .0! .23’"
]3_ O era“ di a" that sea an ace
scorcv m "g 551.38 64.78 .10 .30 -.14 -.04 .27
14. Large score gap "I. ”

favoring literary reading .16 .37 .01 -.06 .05 .01 -.02
(O-No, l-Yes)

15. Large score gap H u
favoring informational .07 .26 -.03 .00 .03 .04 -.06

reading (O-No, l-Yes)

 

145

 

 

 

Table B.2 Continued

Student- Zero-order correlations
level
variables 6 7 8 9 10 1 1 12 l3 14
General Student Variables
1.
2.
3.
4.
5.
6. 1

Ill
7- .28 1

it

8. .01 .05 1
Genre-related Student Variables

it. # lit
9. .26 .03 .18 1

it! til lit ##1
10- .55 .23 .06 .27 1
Reading Performance

tit it! t*# tit
11- .26 .57 --02 -.15 .23 1

til it. it iii 1“ tit
l2. .22 .51 -.04 -.19 .21 .91 1

it. it! ii! iii it! it!
13- .24 .54 --02 -.18 .22 .94 .95 1

it! tit it t it til tit
14. -.08 -.15 -.05 -.04 -.05 -.28 --01 -. 14 1

it **$ t. tit tit it! tit

15- .05 .06 .05 .08 .01 .09 -.13 -.03 -.12

 

# it it!
p<.05; p<.01; p<.001

146

Table 3.3

Results from HLM A nalyses Predicting Informational Reading Performance (Denmark)

 

 

 

 

 

Model 1 Model 2
Level B S.E. t B S.E. t
Intraclass Correlation ﬂom the one-way ANOVA model = .14
Fixed Effect
Class mean score
Intercept 510.32 3.65 139.82*** 513.06 5.80 3344""
Class mean SES 8.90 3.33 2,67"
Class reading N
dill?“ I ties -66.94 21.36 3.13
Informational reading
Instruction -.46 1.59 -.29
Literary reading
Instruction .90 1.73 .52
Gender 6.63 2.05 3,23" 6.68 2.05 3.25"
ﬁﬁt it.
Books at home 20.23 2.41 341 19.88 2.40 8,27
Amount of reading outside .... N
of school — No reading "1059 3'09 '3-43 4092 3-08 -3.45
Amount of reading outside .. ..
of school - A great deal of -15.54 6.22 -2.50 -15.56 6.23 -250
reading
it! ﬁt!
SES 5.95 .86 6.94 5.55 .86 6.44
2 it ‘*
(SES) -1 .40 .42 -3.37 -1.42 .42 -3,41
Attitudes toward reading 5.24 1.14 4.60"" 5.27 1.14 4.63"”
Reading self-concept 22.91 .89 25.72” 22.88 .89 25.68""
Computer use -1.13 .77 -1.46 -1.15 .77 -1.49
Informational reading .904 .83 40.35“" -8.93 .83 40,74'"
Literary reading 2.13 .98 2,17‘I 2.08 .98 2.13T
Random Effect Variance df x2 Variance df x2
Class mean 439.35 204 332,33'" 370.26 200 728.85***
Student-level effect 2653.77 2653.65

 

Variance explained (R2)

At student-level: .38

In class mean: .16

 

Conditional ICC

.14

.12

 

Note. The square term in SES denotes the curvilinear relationship between these variables and the outcome.

T

t it t
p<.10; p<.05; p<.01;

#

*
p<.001

Table B.4

Results ﬁom HLM A nalyses Predicting Literary Reading Performance (Denmark)

 

 

 

 

 

 

 

Model 1 Model 2
Level B S.E. t B S.E. t
Intraclass Correlation ﬂom the one-way ANOVA model = .14
Fixed Effect
Class mean score
ii. iii
Intercept 517.03 3.67 140.73 519.49 5.55 9355
Class mean SES 8.62 3.20 2,69M
Class reading "
difficulties -63.21 20.80 -3.04
Informational reading
Instruction -.40 1.53 -.26
Literary reading
Instruction .25 1.60 .16
it. ##ﬁ
Gender 8.28 2.00 4_ 14 8.33 2.00 4.16
.t# ﬁt!
Books at home 17.00 2.66 6.38 16.62 2.66 6,25
Amount of reading outside N H
of school — No reading -9‘08 2‘78 '3-27 ‘9' 10 2-77 -3.28
Amount of reading outside
of school — A great deal of -10.77 5.92 -1 321. -10.79 5.91 -1.83T
reading
tit #tﬁ
SES 7.23 .96 7,57 6.82 .96 7,11
2 iii ***
(SES) -l.44 .39 -3,71 -1.45 .39 -3,75
Attitudes toward reading 2.94 1.17 2.51’ 2.98 1.17 255'
Reading self-concept 19.61 .88 2226"" 19.57 .88 22,13."
Computer use -2.14 1.14 -1 137'1 -2.15 1.14 4.901
Informational reading -10.17 .88 -1154". -10.05 .88 -1 L38”.
Literary reading 2.83 .80 3,53” 2.79 .80 3,49"
Random Effect Variance df X2 Variance df X2
Class mean 378.29 204 759.02“ 316.42 200 663,64”
Student-level effect 2593. 10 2593.23
Variance explained (R2) At student-level: .34 In class mean: .16
Conditional ICC .13 .ll

 

Note. The square term in SES denotes the curvilinear relationship between these variables and the outcome.

1.

t it it!
p<.10; p<.05; p<.01; p<.001

148

APPENDIX C

Analysis Results of the Data ﬂom England

149

Table C.l
Means, Standard Deviations, and Intercorrelations of Classroom-level Variables and

Reading Performance (1 6 7 classrooms in England)

 

 

 

Class-level Zero-order correlations
. Mean SD.

variables (England) 1 2 3 4
General Classroom Variables
1. Class mean SES .43 .46 1
2. Weekly mstructronal time 3.55 2.98 ,17‘ 1
for reading
3. Proportion of students with "
reading difficulties '19 '17 "22 '05 l
4. Reading comprehension "
strategy instruction .38 1.34 -.03 .24 .01 1
5. Computer use for reading ...-... "a
instruction .03 1.06 .09 .33 -.02 .36
6. Autonomy support in ..
reading instruction -.08 1.14 .00 .13 -_22 .13
7. Social activities in reading .00 1.00 -.17. .28.." -.03 .35."
8. Homework for reading -.33 1.43 .00 .04 -.05 -.07
Genre-related Classroom Variables
9. Instruction for "a "t
Informational reading '01 l '01 '04 '35 '00 '32
10. Instruction for literary ...... 4"
reading .03 1.04 -.05 .22 .00 .22
Reading Performance
11. Class informational 4" 4"
reading score 534.43 44.18 .41 -.09 -_ 51 -.09

. , are sea
12. Class hterary reading score 535.60 45.81 .40 -.09 -.53 -.09
13. Class overall reading score 536.60 45.62 .41". -.09 -52". -.09
14. % of students in class with
large score gap favoring 16.66 9.32 .04 -.01 -.09 -.05
literary reading
15. % of students in class with In
large score gap favoring 13.26 8.52 -.10 .01 .20 -.07

informational reading

 

150

Table C.l Continued

 

 

 

Class- Zero-order correlations
level
variables 6 7 8 9 10 11 12 13 14
General Classroom Variables
1.
2.
3.
4.
5. 1

t
6. .16 1

*** **#
7- .35 .27 1
8. .13 .04 .02 1
Genre-related Classroom Variables

##t tit lit
9. .51 .33 .40 .10 1

* tit
10. .18 .08 .05 .00 .44 1
Reading Performance

* it
11. -.03 .16 -_21 .00 -.11 -.01 1
it ***
12. -.02 .14 -,21 -.01 -.10 -.01 .98 1
it ##t tit

13. -.02 .15 -.21 .00 -.11 -.Ol _99 .99 1
14. .02 -.12 -.07 -.11 -.11 -.07 .09 20*" .15 1
15. -.03 .02 .01 .01 -.04 .08 .03 -.09 -.04 -.13

 

I it lit
p<.05; p<.01; p<.001

151

Table C.2
Means, Standard Deviations, and Intercorrelations of Student-level Variables and

Reading Performance (3,829 students in England)

 

 

 

Student-level Mean S D Zero-order correlations
variables (England) ' ° 1 2 3 4 5
General Student Variables
1. Gender (O-Boy, l-Girl) .50 .50 l
2. Books at home (0-Few "9
books, l-Many books) '74 '44 '08 1
3. Amount of reading ......- "t
outside of school 1 — No .23 .42 -,21 -.13 1
reading (O-No, l-Yes)
4. Amount of reading
outside of school 2 - A an on
great deal of reading (0- '06 '24 -06 '00 --14 1
No, l-Yes)
it! it. it?

5. SES .46 1.31 .07 .18 -,13 .01 l
6. Attitud t d an an mu u a"
reading 68 ow“ -14 1.19 .23 .22 -.28 .11 .07

. ii. iii tﬁi ﬁt *tt
7. Readrng self-concept .30 1.37 .08 .19 -.12 .05 .10
8. Computer use -.05 113 .06.! 01 ..O7#** .07ttt ozottt
Genre-related Student Variables
9. Informational reading an an an area
outside of school "19 1'22 '05 "02 “-22 30 ~11
10. Literary reading that ass aaa ate all-a
outside ofschool .28 1.69 .25 .26 -.29 .12 .08
Reading Performance
11. Informational ail-a ail-a ail-a ail-a use
reading score 538.73 77.75 .09 .33 -.18 -.07 .18
12, L'te din an can an a an
score! my ma g 53959 83.10 .12 .32 -.17 -.04 .14
13. 0 11 d. an as: an as: "a
scorevm tea mg 534.50 83.93 .11 .33 -.16 -.08 .15
14. Large score gap "m n "m u
favoring literary reading .17 .37 .07 _05 -.03 .07 -,05
(0-No, l-Yes)
15. Large score gap H i "
favoring informational . 13 .34 -_0 5 -.02 -.04 -.03 .05

reading (O-No, l-Yes)

 

152

Table C.2 Continued

 

 

 

 

Student- Zero-order correlations
level
variables 6 7 8 9 10 11 12 13 14
General Student Variables
1.
2.
3.
4.
5.
6. 1
##t
7- .29 1
ﬁt! #
8- .06 .04 1
Genre-related Student Variables
tit t# it
9. .21 .06 .25 1
it! tit ##t *t*
10. .62 .28 ..06 .17 1
Reading Performance
It! tit *tt ti* *1!
l 1. .31 .46 -.07 -.17 .36 1
1“ ii! iii it. tit tit
12- .32 .46 -.09 -.19 .37 .92 1
it. til 1*! it! tit ttt ***
13- .31 .45 -.09 -.19 .36 .94 .94 1
*t t it it! it ﬁtt ***
14- .05 .04 --02 -.04 .09 -.04 .22 .08 1
t * tit # ii! iii tit
15. --02 -.04 .04 .07 --01 .04 -.22 -.10 -.18
i it ##l
p<.05; p<.01; p<.001

153

Table C.3

Results ﬁ'om HLM A nalyses Predicting Informational Reading Performance (England)

 

 

 

 

 

 

 

Model 1 Model 2
Level B S.E. t B S.E. t
Intraclass Correlation from the one-way ANOVA model = .22
Fixed Effect
Class mean score
ii. iii
Intercept 521.35 3.48 149.89 531.88 4.82 11032
Class mean SES 15.61 4.89 3,19”
Class reading "*
dlﬂiculties -91.58 13.75 -6.66
Informational reading
Instruction -3.37 2.20 -l .53
Literary reading
Instruction 1.25 2.17 0.58
Gender .07 2.66 .03 -.20 2.64 -.07
iii ﬁt.
Books at home 19.07 2.70 707 19.03 2.69 7,03
Amount of reading outside "- N
of school _ No reading -16.30 3.96 -4.11 -15.98 3.97 4,02
Amount of reading outside .. ... u
of school — A great deal of -l9.57 5.27 -3,72 -19.75 5.26 -3,75
reading
00* *“
SES 6.83 1.08 6.33 6.40 1.08 5,95
2 ii! iii
(SES) -1.65 .44 -3,75 -1.66 .44 -3,79
Attitudes toward reading 3.86 1.40 275" 3.92 1.40 230"
Reading self-concept 20.61 .90 23.00"” 20.42 .89 2295'"
Computer use -540 1.72 -3,14’ -552 1.72 -322’
2 it. *tt
(Computer use) '2-99 -63 -4.74 '3-05 ~63 -4.83
Informational reading -10.62 1.45 -734." -10.66 1.44 -738."
Literary reading 7.41 1.04 7,14’" 7.48 1.04 722""
Random Effect Variance (if X2 Variance df XZ
Class mean 791.09 166 913,61” 459.36 162 624,13'"
Student-level effect 3671 .33 3670.75
Variance explained (R2) At student-level: .32 In class mean: .42
Conditional ICC .18 .11

 

Note. The square terms in SES and computer use denote the curvilinear relationships between these

variables and the outcome.

*1

t it
p<.05; p<.01;

at
p<.001

154

Table C.4

Results from HLM Analyses Predicting Literary Reading Performance (England)

 

 

 

 

 

 

 

Model 1 Model 2
Level B S.E. t B S.E. t
Intraclass Correlation ﬂom the one-way ANOVA model = .21
Fixed Effect
Class mean score
tit tit
Intercept 517.68 3.93 13132 528.86 4.80 110_ 14
Class mean SES 17.83 5.03 3,54"
3513;111:231“ -100.54 13.95 4.21""
Km‘t’gfzm “mg -301 2.31 -1.30
Literary reading
Instruction .88 2.26 .39
Gender 4.09 2.65 1.55 3.86 2.64 1.46
Books at home 20.42 2.85 7.17"" 20.39 2.84 7.19""
Amount of reading outside "It "'1'
ofschool—No reading -l4.ll 3.14 .4,49 -l3.73 3.14 .4,37
Amount of reading outside ... ...
of school - A great deal of -10.74 5.04 -2,13 -10.85 5.02 -2_16
reading
it. it
SES 4.46 1.11 4,01 3.96 1.13 3,51
l #
(31.33)2 -1.20 .49 -246 -1.22 .48 -252
Attitudes toward reading 4.91 1.38 3,56" 4.97 1.37 3,63"
Reading self-concept 21.84 1.01 21,67." 21 .62 .99 21,74”.
Computer use -6.14 1.05 -533'" -6.28 1.05 45.00""
2 It!" 1"".
(Computer use) '3 38 -72 -4.71 -3.43 -71 -4.81
Informational reading -1340 1.43 9.34"" -13.49 1.44 9.33""
Literary reading 8.67 1.02 3.43"" 8.73 1.02 356'"
Random Effect Variance df x2 Variance df 12
Class mean 844.85 166 362,515" 445.71 162 553,64”
Student-level effect 4212.88 421 1 .29
Variance explained (R2) At student-level: .33 In class mean: .47
Conditional ICC .17 .10

 

Note. The square terms in SES and computer use denote the curvilinear relationships between these

variables and the outcome.

# it ﬁt.
p<.05; p<.01; p<.001

155

APPENDDC D

Analysis Results of the Data ﬁ'om Taiwan

156

Table D.1
Means, Standard Deviations, and Intercorrelations of Classroom-level Variables and

Reading Performance (150 classrooms in Taiwan)

 

 

 

Class-level Mean S D Zero-order correlations
variables (T alwan) 1 2 3 4
General Classroom Variables

1. Class mean SES .02 .33 1

2. Weekly instructional time *

for reading 2.10 1.56 .20 1

3. Proportion of students with '

reading difﬁculties '12 '10 "22 '10 1

4. Reading comprehension

strategy instruction .07 1.10 .12 .11 -.09 l
.5' C°mP“‘°’ “5° f0” ’cadmg -.15 1.03 -.09 -.09 -.04 ,31’"
instruction

6. Autonomy support in * _ ......
reading instruction .07 1.15 ,17 .03 .08 .22

7. Social activities in reading -.04 1.06 -.01 .10 .03 ,43'"
8. Homework for reading -.07 1.23 .08 .07 -.04 .12
Genre-related Classroom Variables

9. Instruction for "It
Informational reading -.10 1.13 .06 .10 .03 .38
{1; gym” f‘” "”3er -.04 1.06 ,20" .ll -.08 ,40’"
Reading Performance

11. Class informational WI *

reading score 537.80 20.59 .59 .12 -,21 .00
12. Class literary reading score 529.98 23.88 ,61‘" .10 -.17’ .03
13. Class overall reading score 534.95 22.84 50*" .10 -.20‘ .00
14. % of students in class with i. In
large score gap favoring 7.45 5.09 .17 -.05 -.03 _22
literary reading

15. % of students in class with 1m:

large score gap favoring 17.49 7.91 -.30 .04 -.09 -.11

informational reading

 

157

Table D.l Continued

 

 

 

 

158

Class- Zero-order correlations
level
variables 6 7 8 9 10 1 l 12 13 14
General Classroom Variables
l.
2.
3.
4.
5. l
6. -.04 1
it!
7. .34 .06 1
8. .05 .03 .04 l
Genre-related Classroom Variables
till it Ililt #
9. .37 .24 .39 .17 1
till I!!! tilt # tilt
10- .41 .21 .35 .19 .42 1
Reading Performance
11. -. 10 .08 .05 . 13 .04 .06 1
It“!
12. -.12 .05 .05 .13 .05 .06 .97 1
##t tit
13. -.10 .07 .05 .13 .05 .07 .98 .98 l
ttﬁ *t
14. .03 -.05 .05 -.Ol .09 .05 .18 .31 .24 1
it! tit tt# *1.
15. .04 .07 -.O9 -.14 .01 -.O9 -,45 -,59 -.52 -35
It ﬁt tit
p<.05; p<.01; p<.001

Table D2
Means, Standard Deviations, and Intercorrelations of Student-level Variables and

Reading Performance (4,207 students in Taiwan)

 

 

 

Student-level Mean S D Zero-order correlations
variables (Taiwan) ' ' 1 2 3 4 5
General Student Variables
1. Gender (O-Boy, l-Girl) .49 .50 1
2. Books at home (O-Few *
books, 1-Many books) 60 '49 ~04 1
3. Amount of reading ...... "m
outside of school 1 — No .24 .43 -.18 -20 1
reading (O-No, l-Yes)
4. Amount of reading
outside of school 2 — A it" Mt
great deal of reading (0- '05 '22 '03 '08 "13 1
No, 1-Yes)
tit it! it. ti!

5. SES .06 1.08 _05 _23 -.15 .08 1
6, Attitud t d ant-a anu- aaa eta all-a
reading es ow“ .21 1.28 .20 .25 -.34 .17 .21
7. Reading self-concept .40 1.46 .10." .27." -.23*" .13“. .23".
8. Computer use -.03 1.04 -.01 .06." -,11"* .03 ,19'"
Genre-related Student Variables
9. Informational reading "It ...... t" in
outside of school "20 1'26 '02 '15 “25 '18 '21
10, Literary reading at: ant-a all-a um- nuns
outside of school .19 1-51 .17 .26 -.33 .17 .21
Reading Performance
11. Informational "It It" It" an
reading score 541.78 52.10 .03 ,3 5 -.26 .02 ,17

, L' readi eta an au- a mu
$620,;me ”g 534.12 61 .80 .12 .32 -.22 .19
13. O ] ad' tat-a III": an” a an"
gamma! re mg 539.07 58.12 .10 .34 -.25 .04 .18
14. Large score gap ml .. n "I.
favoring literary reading .07 .26 .07 .04 .00 .05 .07
(O-No, l-Yes)
15. Large score gap ...... u I.“
favoring informational . 17 .37 -.09 ._04 .03 .00 -_07

reading (O-No, l-Yes)

 

159

 

 

 

 

Table D2 Continued

Student- Zero-order correlations
level
variables 6 7 8 9 10 l 1 12 13 14
General Student Variables
l.
2.
3.
4.
5.
6. 1

ti*
7. .45 1

tﬁt it
8. .06 .04 l
Genre-related Student Variables

it! it! tit
9- .31 .19 .20 1

ti‘ 0*! ltt *tt
10- .46 .36 .11 .41 1
Reading Performance

til it! t it!
11- .32 .40 .04 -.02 .26 1

til it. ﬁt ii. iii
12. .29 .37 .05 .00 .27 .90 1

it it! it t** tit tit
13. .31 .40 .04 --01 .27 .94 .94 1
# tit ﬁt! 11* iii
14- -03 .03 -03 .07 .08 .03 .26 .13 I
lit it tit tit tit tit 01.
15. -.01 -.O7 -.04 .01 -.O6 -.14 -.41 -.28 -.13
t it it!

p<.05; p<.01; p<.001

160

Table D3

Results from HLM A nalyses Predicting Informational Reading Performance (Taiwan)

 

 

 

 

 

Model 1 Model 2
Level B S.E. t B S.E. t
Intraclass Correlation from the one-way ANOVA model = .09
Fixed Effect
Class mean score
Intercept 533.63 2.54 210.47*** 533.32 2.94 181 .70***
Class mean SES 21 .79 4.08 534*“
$3,522?“ -1.05 12.79 -.08
ﬁzﬁ‘mfgﬁm’mdm .56 1.21 .47
fgﬁggfding - l .62 1 .24 -l .30
Gender -1.52 1.92 -.80 -1.39 1.92 -0.73
Books at home 21.10 1.91 11.06*** 20.60 1.93 1068*"
ﬁggffﬁgﬁgide -14.82 3.10 -4.78** -14.63 3.09 -4.74**
Amount of reading outside
of school — A great deal of -925 4.92 -l 88* -8.99 4.91 -l.83*
reading
SES 3.11 .82 3.77“" 2.52 .82 3.07"
(SES)2 -3.13 .48 4549*" -3.15 .48 45.55".
Attitudes toward reading 5.21 .83 627*" 5.26 .83 6.35“”
Reading self-concept 8.51 .85 10.01 *" 8.54 .85 1009*“
Computer use -1 .63 .79 -2.06* -1 .60 .79 -2.02*
(Computer use)2 -3.05 .53 -5.73*** -3.07 .53 -5.74"‘**
Informational reading -8.00 .87 -9.25*** -8.01 .86 -9.29"'**
Literary reading 3.34 .94 3.54" 3.22 .94 3.43"
Random Effect Variance (if X2 Variance df X2
Class mean 153.19 149 442.39*** 113.22 145 366.841*"'*
Student-level effect 2205.03 2203.1 5

 

Variance explained (R2)
Conditional ICC

At student-level: .24

.06

In class mean: .26

.05

Note. The square terms in SES and computer use denote the curvilinear relationships between these

variables and the outcome.

1.

t it
p<.10; p<.05; p<.01;

**

at
p < .001

161

Table D4

Results ﬁ'om HLM Analyses Predicting Literary Reading Performance (Taiwan)

 

 

 

 

 

 

Model 1 Model 2
Level B S.E. t B S.E. t
Intraclass Correlation from the one-way ANOVA model = .09
Fixed Effect
Class mean score
tilt *tt
Intercept 520.16 2.56 203.56 518.99 3.11 16711
Class mean SES 26.52 5.13 5,17."
Class reading difﬁculties 5.37 15.94 .34
Informational reading .
Instruction .75 1.55 .48
Literary reading
Instruction -2.08 1.59 -1.31
Gender 4.34 1.96 2,21" 4.49 1.96 229"
Books at home 21.71 2.43 3.95"" 21.12 2.44 3,55‘"
Amount of reading outside of "'1' *M
school _ No reading -12.02 2.79 .432 -l 1.79 2.81 .420
Amount of reading outside of
school - A great deal of -5.90 4.54 -l .30 -5.61 4.55 -1.23
reading
lit it
SES 4.34 1.11 3.93 3.65 1.12 325
it it
(SES)2 -2.13 .61 -343 -2.16 .61 -352
Attitudes toward reading 4.67 1.09 429*" 4.72 1.09 4,34.“
Reading self-concept 9.38 1.24 759'" 9.42 1.24 7150'"
Computer use .40 1.04 .39 .45 1.04 .43
2 ill #4".
(Computer use) '3-09 ~66 -4.71 -3.1 1 -65 -4.75
Informational reading -8.90 1.10 -3,10*" -8.90 1.10 $.12".
Literary reading 5.14 .81 635"" 5.00 .81 620'"
Random Effect Variance (if x2 Variance df X2
Class mean 233.65 149 44543” 173.48 145 369.84***
Student-level effect 3320.19 33 18.3 1
Variance explained (R2) At student-level: .20 In class mean: .26
Conditional [CC .07 .05

 

Note. The square tenns in SES and computer use denote the curvilinear relationships between these

variables and the outcome.

t

II it t *
p<.05; p<.01; p<.001

162