Children learning to read later catch up to children reading earlier

Full text

Early
Childhood
Research
Quarterly
28 (2013) 33–
48
Contents
lists
available
at
SciVerse
ScienceDirect
Early
Childhood
Research
Quarterly
Children
learning
to
read
later
catch
up
to
children
reading
earlier
Sebastian
P.
Suggatea,∗, Elizabeth
A.
Schaughencyb, Elaine
Reeseb
aUniversity
of
Regensburg,
Germany
bUniversity
of
Otago,
New
Zealand
a
r
t
i
c
l
e
i
n
f
o
Article
history:
Received
30
March
2011
Received
in
revised
form
5
April
2012
Accepted
22
April
2012
Keywords:
Reading
age
Reading
achievement
Phonemic
awareness
Reading
comprehension
Early
reading
a
b
s
t
r
a
c
t
Two
studies
from
English-speaking
samples
investigated
the
methodologically
difficult
question
of
whether
the
later
reading
achievement
of
children
learning
to
read
earlier
or
later
differs.
Children
(n
=
287)
from
predominantly
state-funded
schools
were
selected
and
they
differed
in
whether
the
read-
ing
instruction
age
(RIA)
was
either
five
or
seven
years.
Study
1
covered
the
first
six
years
of
school
following
three
cohorts
across
a
two-year
design.
Analyses
accounted
for
receptive
vocabulary,
reported
parental
income
and
education,
school-community
affluence,
classroom
instruction,
home
literacy
envi-
ronment,
reading
self-concept,
and
age.
The
earlier
RIA
group
had
initially
superior
letter
naming,
non-word,
word,
and
passage
reading
but
this
difference
in
reading
skill
disappeared
by
age
11.
In
Study
2,
the
decoding,
fluency,
and
reading
comprehension
performance
of
83
additional
middle
school-age
children
was
compared.
The
two
groups
exhibited
similar
reading
fluency,
but
the
later
RIA
had
generally
greater
reading
comprehension.
Given
that
the
design
was
non-experimental,
we
urge
further
research
to
better
understand
developmental
patterns
and
influences
arising
from
different
RIAs.
© 2012 Elsevier Inc. All rights reserved.
1.
Introduction
The
age
at
which
children
enter
school
and
learn
to
read
is
intuitively
an
important
factor
in
later
reading
achievement.
For-
mal
reading
instruction
typically
begins
comparatively
early
in
the
English-speaking
world,
around
age
five,
with
there
being
a
recent
trend
for
other
countries—and
early
education
institutions
within
countries—to
lower
the
age
when
children
begin.
In
the
US,
for
example,
the
call
for
early
intervention
to
address
risk
for
reading
failure
was
repeatedly
made
at
hearings
for
the
National
Reading
Panel
(2000),
with
the
publication
of
this
report
credited
as
con-
tributing
to
initiatives
such
as
Reading
First,
focusing
on
reading
instruction
in
the
early
grades
so
that
children
become
compe-
tent
readers
by
the
end
of
third
grade,
and
Early
Reading
First,
intended
to
promote
reading
readiness
skills
among
preschool-age
children
(Shanahan
&
Lonigan,
2010).
Similarly,
in
England,
liter-
acy
is
included
among
the
early
learning
goals
for
the
preschool
years
(Quick,
Lambley,
Newcombe,
&
Aubrey,
2002),
with
many
preschool
(i.e.,
reception)
classrooms
providing
a
Literacy
Hour
by
the
end
of
the
school
year
(Quick
et
al.,
2002).
This
practice
is
not
without
controversy,
however:
An
independent
review
of
English
∗Corresponding
author
at:
Lehrstuhl
für
Schulpädagogik,
University
of
Regens-
burg,
Universitätsstraße
31,
93040
Regensburg,
Germany.
Tel.:
+49
941
943
1750;
fax:
+49
941
943
1993.
E-mail
addresses:
sebastian.suggate@ur.de,
sebastian.suggate@gmail.com
(S.P.
Suggate).
elementary
education
criticized
the
early
start
to
formal
schooling
and
recommended
re-examining
school
starting
age
(Cambridge
Primary
Review,
2009).
Moreover,
a
report
on
reading
instruction
for
the
House
of
Commons
concluded
that
current
evidence
does
not
provide
a
clear
answer
to
the
question
of
what
is
the
best
age
to
begin
formal
teaching
(House
of
Commons
Education
and
Skills
Committee,
2005).
1.1.
Previous
research
Contemporary
researchers
have
typically
considered
the
ques-
tion
of
whether
preschool
or
kindergarten
children
are
too
young
to
learn
to
read
(e.g.,
Cunningham
&
Carroll,
2011a;
Ehri,
Nunes,
Stahl,
&
Willows,
2001).
The
answer
to
this
question
appears
to
be
“no”
for
many
children.
Evidence
from
meta-analysis
suggests
that
preschool
and
kindergarten
children,
including
those
at-risk,
can
be
taught
decoding
skills
(e.g.,
Bus
&
van
Ijzendoorn,
1999;
Ehri,
Nunes,
Stahl
et
al.,
2001;
Suggate,
2010).
Moreover,
single
studies
(Feitelson,
Tehori,
&
Levinberg-Green,
1982)
and
compar-
ative
international
studies
(Seymour,
Aro,
Erskine,
&
COST
Action
Network,
2003)
suggest
that
younger
children
can
learn
to
read
from
the
age
of
four.
However,
in
this
paper
we
turn
to
a
somewhat
different
ques-
tion:
Is
there
a
long-term
advantage
in
reading
for
having
learned
to
read
earlier?
Given
that
the
reading
skills
gained
in
elementary
school
predict
subsequent
educational
and
occupational
pathways
for
children
(Savolainen,
Ahonen,
Aro,
Tolvanen,
&
Holopainen,
2008),
coupled
with
the
poignant
reality
that
many
children
fail
0885-2006/$
–
see
front
matter ©
2012 Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.ecresq.2012.04.004

34 S.P.
Suggate
et
al.
/
Early
Childhood
Research
Quarterly
28 (2013) 33–
48
to
master
these
skills
(Adams,
1990;
National
Early
Literacy
Panel
(NELP,
2008)
;
National
Reading
Panel,
2000;
Snow,
Burns,
&
Griffin,
1998),
this
question
has
educational
significance.
Also,
given
that
reading
involves
the
weaving
together
of
pre-reading,
decoding,
reading,
and
language-related
skills
(Dickinson,
Golinkoff,
&
Hirsh-
Pasek,
2010;
Whitehurst
&
Lonigan,
2001),
skills
that
may
be
differentially
influenced
by
formal
and
informal
learning
experi-
ences
(Cunningham
&
Carroll,
2011a),
examination
of
the
effect
of
reading
instruction
age
(RIA)
potentially
enhances
our
understand-
ing
of
reading
development.
We
recognize
that
skilled
reading
typically
comprises
the
development
of
pre-reading,
decoding,
reading,
and
reading
com-
prehension
skills,
each
of
which
is
influenced
by
a
variety
of
preschool
and
school
experiences
(Adams,
1990;
Biemiller,
2006;
Dickinson,
McCabe,
Anastasopoulos,
Peisner-Feinberg,
&
Poe,
2003;
Neuman,
2006;
Storch
&
Whitehurst,
2002;
Whitehurst
&
Lonigan,
2001).
Because
it
is
widely
recognized
that
reading
comprises
two
main
strands
of
text-related
and
language-related
skills
(Gough
&
Tunmer,
1986),
discussions
of
the
sub-components
of
pre-reading,
reading,
and
reading
comprehension
skills
typically
include
both
language-
and
text-related
skills.
Given
that
we
seek
to
investigate
whether
the
earlier
introduction
of
formal
reading
instruction
and
concordant
earlier
mastery
of
reading
leads
to
later
advantage
in
reading,
we
are
left
with
a
definitional
dilemma:
On
the
one
hand,
in
framing
our
research
question,
we
do
not
believe
it
conceivable
to
test
whether
pre-literacy
skills
should
be
either
taught
in
their
entirety
or
removed
altogether.
There
is
a
solid
research
base
sup-
porting
the
importance
of
oral
language
and
phonemic
awareness
skills
in
reading
development
(e.g.,
NELP,
2008);
yet
on
the
other
hand
we
seek
to
study
the
role
of
formal
reading
instruction.
Accordingly,
we
adopt
a
similar
distinction
to
the
Simple
View
of
Reading
(SVR;
Gough
&
Tunmer,
1986)
and
apply
this
to
pre-literacy
and
early
reading
development.
Thus,
here
we
define
decoding-
related
skills
as
those
that
require
text,
including
alphabetic
awareness,
decoding
skills,
and
reading
fluency.
We
call
our
second
strand
of
skills
language
comprehension
skills,
consistent
with
the
SVR
(Gough
&
Tunmer,
1986),
including
phonemic
awareness
and
phonology,
vocabulary,
syntax,
semantics,
pragmatics,
and
think-
ing
skills
used
in
understanding
written
and
spoken
language.
Although
this
definition
is
not
perfect
because
many
decoding-
related
skills
also
require
language,
it
captures
the
essence
of
the
distinction
between
early
instruction
of
decoding-related
skills
while
acknowledging
that
oral
language
and
phonemic
awareness
belong
to
pre-literacy
and
reading
development.
Finally,
we
define
skilled
reading
as
being
able
to
read
accurately
and
fluently,
and
to
understand
what
is
read.
Herewith,
we
also
adopt
the
view
that
many
important
pre-
literacy
experiences
constitute,
in
some
respects,
the
beginning
of
learning
that
leads
to
competent
reading
skills
(Whitehurst
&
Lonigan,
2001).
However,
the
question
we
ask
here
is
whether
the
early
development
of
decoding-related
skills
per
se
is
important
in
later
skilled
reading?
We
define
early
reading
instruction
as
that
intended
to
lead
to
the
development
of
decoding-related
skills,
be
they
more
focused
on
teaching
the
explicit
relations
between
graphemes
and
phonemes
or
between
deciphering
graphemes
with
the
help
of
contextual
cues.
In
doing
so,
we
do
not
relegate
the
role
of
language
in
reading
(Dickinson
et
al.,
2010),
but
rather
seek
to
isolate
the
role
of
early
reading
instruction
in
the
later
development
of
skilled
reading.
Knowledge
of
decoding-related
skills,
including
letter
knowl-
edge,
letter-sound
correspondences,
and
word-reading,
has
proven
to
be
a
robust
predictor
of
later
reading
in
correlational
stud-
ies
(Snow
et
al.,
1998).
For
example,
preschool
or
kindergarten
non-word
decoding
correlates
highly
with
later
reading
(r
=
.72,
Lonigan,
Schatschneider,
Westberg,
&
NELP,
2008).
Importantly,
because
a
variety
of
third
factors
are
associated
with,
and
may
well
determine
and
drive,
both
early
and
later
reading
skill—such
as
instruction
(National
Reading
Panel,
2000),
home
environment
(Niklas
&
Schneider,
2010),
genetic
(Olson
&
Gayan,
2001),
and
lan-
guage
factors
(Dickinson
et
al.,
2003)—correlations
between
early
decoding-related
skills
and
later
achievement
do
not
fully
answer
the
question
of
how
early
versus
later
readers
fare
long-term.
A
second,
important
line
of
research
arises
from
controlled
read-
ing
intervention
studies.
Many
intervention
studies
have
shown
that
early
decoding-related
skills
can
be
boosted
in
the
short-
term
(Bus
&
van
Ijzendoorn,
1999;
Ehri,
Nunes,
Stahl
et
al.,
2001;
Ehri,
Nunes,
Willows
et
al.,
2001;
Suggate,
2010).
However,
sel-
dom
do
studies
include
long-term
data
beyond
12
to
18
months
post-intervention,
preventing
investigation
of
whether
the
early
advantage
in
skilled
reading
later
washes
out.
Moreover,
such
inter-
vention
studies
usually
compare
children
receiving
one
form
of
decoding-related
instruction
with
another
form.
Thus,
even
the
control
children
in
intervention
studies
in
kindergarten
or
grade
1
learned
early
decoding-related
skills
or
received
early
instruction.
Clearly,
to
understand
whether
early
reading
instruction
leads
to
long-term
advantage
in
reading,
two
groups
differing
in
when
they
receive
reading
instruction
need
to
be
compared
over
a
long
period
of
time.
Few
studies
directly
compare
the
long-term
effects
of
having
an
earlier
versus
later
reading
instruction
age
(RIA).
Findings
from
two
international
studies
suggest
that
children
with
a
later
school
entry
age—which
can
be
expected
to
correlate
highly
with
RIA—achieve
similarly
in
long-term
reading
achievement
(Elley,
1992;
Suggate,
2009).
However,
by
virtue
of
being
cross-sectional,
these
studies
do
not
account
for
development
at
the
beginning
of
primary
school
and
conflate
international
language
differences
with
RIA.
Importantly,
the
countries
with
the
earlier
RIAs
also
speak
English,
which
has
a
uniquely
complex
orthography
and
spelling-to-sound
structure
(Seymour
et
al.,
2003).
International
studies
do
not
rule
out
the
possibility
that
an
earlier
RIA
is
needed
to
offset
the
added
difficulty
in
learning
to
read
English.
We
consider
that
a
stronger
test
of
whether
there
is
a
long-term
relative
advantage
in
reading
achievement
from
earlier
reading
instruction
would
involve
comparisons
of
children
beginning
for-
mal
reading
instruction
at
markedly
different
ages
with
a
follow-up
assessment
occurring
at
a
chronologically
equivalent
age.
There-
with
children
beginning
at
five,
the
typical
age
of
beginning
reading
instruction
in
English-speaking
countries,
could
be
contrasted
with
children
beginning
at
age
six
or
seven,
which
more
closely
approx-
imates
the
typical
age
of
beginning
reading
instruction
in
other
highly
literate
countries,
such
as
those
in
Scandinavia.
Moreover,
given
English’s
unique
orthographical
features,
a
strong
test
would
involve
English-speaking
samples
and,
given
the
difficulty
of
ran-
dom
assignment,
control
for
key
variables,
and
capture
learning
longitudinally.
The
ethics
of
conducting
this
methodologically
ideal
study
cur-
rently
precludes
random
assignment
because
evidence
from
less
invasive
non-experimental
research
would
first
need
to
be
com-
pelling.
In
a
cross-sectional
study
in
the
United
States,
children
randomly
selected
from
a
waiting
list
to
either
enter
or
have
entry
declined
to
a
Montessori
school
were
compared
(Lillard
&
Else-
Quest,
2006).
In
Montessori
schools,
children
often
end
up
learning
to
read
sometime
soon
after
age
three,
perhaps
because
of
the
individual
support
offered
them
(Edwards,
2007).
Accordingly,
the
Montessori
children
had
an
advantage
in
reading
over
the
public-
schooled
children
at
age
five
but
not
at
age
12.
An
older
study
(i.e.,
Durkin,
1974–1975)
also
investigated
reading
longitudinally
but
had
a
small
sample
size,
did
not
have
random
assignment,
focused
on
IQ
as
the
main
potential
confound,
and
covered
only
two
years
of
the
primary
school
period.
Interestingly,
the
findings
of
this
study
were
consistent
with
those
of
Lillard
and
Else-Quest
(2006)
and
the
trends
from
international
studies
(Elley,
1992;
Suggate,
2009):
After

S.P.
Suggate
et
al.
/
Early
Childhood
Research
Quarterly
28 (2013) 33–
48 35
two
years
of
primary
school,
the
children
whose
parents
opted
for
entry
into
the
early
school
program
had
lost
their
earlier
advantage
in
reading
skill.
A
more
recent
study
conducted
in
the
UK
capitalized
on
schooling
differences
between
Steiner
and
standard
state
cur-
ricula
(described
more
fully
below)
to
investigate
whether
the
younger
state
children
would
acquire
beginning
decoding-related
skills
more
slowly
than
the
older
Steiner-educated
children
(Cunningham
&
Carroll,
2011a).
Decoding-related
development
was
followed
across
the
first
year
of
school
for
both
the
state-
educated
(n
=
31)
children
aged
four
years
and
nine
months
and
the
Steiner-educated
children
(n
=
30)
aged
seven
years
and
10
months.
Follow-up
data
were
collected
after
two
years
of
school
on
a
smaller
sample
(n
=
19
for
Steiner
educated
and
n
=
25
for
state
educated).
Key
findings
were:
(a)
across
the
initial
two
years
of
decoding-
related
instruction,
there
were
no
systematic
differences
in
reading
performance,
(b)
the
state-educated
children
had
an
advantage
in
spelling,
and
(c)
the
state-educated
children
received
more
read-
ing
and
phonics
instruction.
Moreover,
Cunningham
and
Carroll’s
(2011a)
study
does
not
explore
the
question
of
whether
the
ear-
lier
mastery
of
decoding-related
skill
by
the
state-educated
sample
would
lead
to
a
later
advantage
in
skilled
reading
on
a
chronolog-
ically
matched
sample
for
two
main
reasons.
First
and
foremost,
the
study
was
not
conducted
sufficiently
long-term
to
allow
for
age-matched
comparisons.
Second,
importantly,
given
the
associ-
ation
of
socioeconomic
factors
(Chiu
&
McBride-Chang,
2006)
and
parental
education
(Boyce
&
Snow,
2009)
to
reading
achievement,
Steiner
education
in
the
UK
is
privately
funded
and
there
appeared
to
be
marked
differences
in
parental
education
between
the
two
samples
in
Cunningham
and
Carroll’s
study.
Therefore,
given
dif-
ferences
in
parental
education
and
financial
constraints
on
school
attendance,
the
extent
to
which
factors
related
to
the
cultural
cap-
ital
within
families
contributed
to
reading
outcomes
is
uncertain.
1.2.
Background
to
the
current
studies
To
address
the
question
of
whether
early
RIA
is
associated
with
later
advantages
in
reading
achievement,
we
present
the
results
of
two
studies,
one
employing
three
pairs
of
longitudinal
samples
and
the
other
cross-sectional,
conducted
in
New
Zealand
(NZ),
spanning
the
first
six
years
of
school,
for
students
with
an
RIA
of
five
versus
seven
years.
By
spanning
the
first
six
years
of
school,
the
design
captured
a
diverse
period
of
reading
acquisition,
enabling
consid-
eration
of
the
antecedents
of
reading,
fluency
development,
and
later
reading
comprehension
(Leppänen,
Aunola,
Niemi,
&
Nurmi,
2008).
Moreover,
NZ
presents
an
opportune
context
for
investigat-
ing
the
role
of
the
RIA
because
the
same
two
educational
options
as
in
Cunningham
and
Carroll
(2011a,
2011b)
studies
are
available,
which
similarly
differ
in
RIA,
but
with
one
crucial
difference:
Gov-
ernment
funding
allows
parents
to
enroll
children
in
either
type
of
school,
often
without
significant
differences
in
parental
financial
contributions.
1.2.1.
Reading
instruction
in
state
schools
Children
who
attend
state
schools
in
NZ
begin
school
on
their
fifth
birthday
(unless
this
date
occurs
during
a
holiday)
and
receive
formal
reading
instruction
from
that
day.
As
described
in
Ministry
of
Education
teacher
resource
materials
(Ministry
of
Education,
2007),
literacy-related
activities
should
pervade
the
school
day
in
the
first
year
of
school.
The
typical
reading
session
is
described
as
involving
shared
and
guided
reading
in
a
language-rich
text
(Ministry
of
Education,
2007).
Context-based
word
identification
strategies
have
influenced
approaches
to
reading
instruction
in
NZ
state
schools
(Clay,
1998;
Smith
&
Elley,
1997;
see
Greaney,
2002,
for
discussion),
although
Ministry
of
Education
documents
also
discuss
recommendations
for
greater
attention
to
word-level
decoding-related
skills
as
an
aspect
of
reading
acquisition
(Ministry
of
Education,
1999,
2003b,
2010)
along
with
development
of
a
sight
reading
vocabulary
(Ministry
of
Education,
2003a,
2010).
Although
the
approach
in
NZ
classrooms
may
not
always
include
explicit
instruction
in
phonemic
awareness
and
phonics
as
recom-
mended
in
recent
syntheses
such
as
that
from
the
National
Reading
Panel
(2000),
scores
on
leveled
readers
indicate
that,
by
age
six,
most
children
can
read
connected
text
consisting
of
about
300
words
with
one
storyline
or
topic
with
an
accuracy
of
over
90
per-
cent
(McNaughton,
Phillips,
&
MacDonald,
2000).
This
normative
standard
of
reading
at
level
12
(Green)
in
the
Ready
to
Read
series
of
leveled
readers
has
been
set
as
the
benchmark
for
achievement
in
national
achievement
standards
being
implemented
in
2010
(Ministry
of
Education,
2009),
with
the
expectation
that
children
who
are
achieving
below
the
15th
percentile
should
be
provided
with
additional
reading
support,
such
as
through
one-to-one
reme-
dial
programs
(Ng,
2006).
Our
previous
research
in
NZ
suggests
that
many
children
entering
school
at
age
five
are
just
beginning
to
develop
alphabetic
skills
(e.g.,
correctly
naming,
on
average,
approximately
12
letters
in
one
minute,
Suggate,
Schaughency,
&
Reese,
2011);
therefore
we
consider
it
justified
to
assume
a
formal
RIA
of
five
years
for
state-curriculum
children,
while
acknowl-
edging
that
some
children
may
have
acquired
literacy,
and
even
reading,
skills
in
preschool
or
at
home.
Moreover,
given
that,
by
age
six,
many
children
meet
expectations
for
text
reading
(e.g.,
using
punctuation
marks
to
produce
smooth
phrasing
in
oral
reading),
we
consider
also
that,
in
addition
to
receiving
early
reading
instruction,
many
NZ
children
accurately
read
connected
text
with
expression.
1.2.2.
Reading
instruction
in
Steiner
education
Conversely,
children
in
the
Steiner
curriculum
enter
school
in
the
year
of
their
seventh
birthday,
typically
spending
the
prior
two
years
in
Steiner
kindergartens
instead
of
state
schools.
In
these
kindergarten
years,
written
language—be
it
in
the
form
of
letters,
words
or
stories—is
excluded
from
the
kindergartens
to
encourage
the
development
of
oral
language.
Instead,
children’s
activities
cen-
ter
on
play,
activities
(e.g.,
painting,
drawing,
cooking),
singing
or
oral
story
telling
from
the
kindergarten
teacher
(Edwards,
2007).
Therefore,
reading
instruction
does
not
occur
in
Steiner
kinder-
gartens.
However,
we
recognize
that
the
kinds
of
language
activities
taking
place
may
well
develop
a
strong
foundation
for
later
skilled
reading
(Dockrell,
Stuart,
&
King,
2010),
even
though
they
do
not
involve
text.
Unlike
in
state
schools
in
NZ
where
most
children
enter
school
when
they
are
precisely
five
years
old,
Steiner
children
enter
at
the
beginning
of
the
academic
year
when
they
turn
seven.
Accord-
ingly,
they
have
an
average
chronological
age
of
6.5
years,
hence
the
difference
in
RIA
with
state
school
children
is
around
18
months.
Reading
instruction
begins
in
the
first
year
of
school
through
the
artistic
introduction
of
the
letters
and
their
corresponding
sounds
through
story,
one
per
week
(Steiner,
1924),
which
lasts
until
the
middle
of
grade
1.
Once
children
have
acquired
a
number
of
con-
sonants
and
vowels,
they
begin
exploring
how
words
comprise
smaller
phonemic/graphemic
components,
with
a
focus
on
seg-
menting
as
opposed
to
blending
(Burnett,
2007).
Children
also
gain
experience
with
the
letters
through
writing
and
forming
the
letters
they
know
into
words
(Burnett).
Towards
the
end
of
grade
1
and
in
grade
2,
children
begin
learning
to
read
connected
text.
Read-
ing
activities
early
in
the
Steiner-curriculum
have
been
described
as
following
an
analytic
phonics
approach,
as
this
tends
to
focus
on
word
recognition
through
learning
of
initial
sound
letter
correspon-
dences
and
sight
word
reading
(Cunningham
&
Carroll,
2011a).
This
less
synthetic
and
systematic
approach
to
phonics
in
the
Steiner
schools,
coupled
with
sight
recognition
training
and
use
of
oral
language
to
identify
words,
gives
rise
to
more
similarity
between
the
NZ
state
and
Steiner
curriculum
than
is
the
case
in
the
UK.

36 S.P.
Suggate
et
al.
/
Early
Childhood
Research
Quarterly
28 (2013) 33–
48
Moreover,
similarities
between
the
NZ
state
and
Steiner
curricula
exist
after
the
initial
acquisition
of
decoding-related
skills.
1.3.
The
current
studies
Capitalizing
on
these
RIA
differences,
the
first
study
follows
reading
development
across
the
first
six
years
of
school,
which
corresponds
to
the
typical
length
of
elementary
school
in
NZ.
The
second
study
extends
the
first
by
focusing
(a)
on
an
older
Year-7
state
school
(i.e.,
beginning
middle
school)
sample
that
is
closer
in
age
to
the
Year
6
end-point
Steiner
sample
in
the
first
study,
(b)
on
reading
comprehension,
and
(c)
at
an
age
when
children
are
expected
to
be
reading
for
meaning
(Chall,
1987;
Ministry
of
Education,
2009).
To
begin
to
explore
the
extent
to
which
curricula
in
state-
and
Steiner-curricula
RIAs
covary,
we
include
a
measure
of
classroom
instruction.
We
also
thought
it
important
to
account
for
a
host
of
family
and
child
characteristics,
to
evaluate
whether
the
Steiner
and
state
curricula
children
were
comparable
on
variables
sug-
gested
by
previous
research
to
be
related
to
reading
achievement.
Finally,
a
third
feature
of
the
first
of
the
current
studies
is
that
we
build
upon
the
conclusions
and
recommendations
for
future
research
made
by
members
of
the
NELP
(Shanahan
&
Lonigan,
2010).
Recognizing
that
young
children
change
over
development,
we
followed
reading
trajectory
development
at
regular
inter-
vals,
investigating
conventional
pre-literacy,
decoding,
and
reading
skills
(i.e.,
decoding,
oral
reading
accuracy
and
fluency,
and
read-
ing
comprehension)
identified
in
the
National
Reading
Panel
(2000)
and
NELP
(2008)
reports.
By
assessing
development
of
decoding-
related
skills,
we
are
able
to
compare
groups
as
a
function
of
age
and
exposure
to
schooling
and
different
educational
experiences,
thus
extending
the
work
of
Cunningham
and
Carroll
(2011a,
2011b).
2.
Study
1
Study
1
longitudinally
follows
three
pairs
of
elementary
samples
to
investigate
the
long-term
development
of
reading
skill
as
a
func-
tion
of
RIA.
We
further
investigated
for
differential
development
of
specific
reading
skills,
expecting
to
see
initial
greater
decoding-
related
skills
for
the
early
RIA
sample,
given
previous
research
(Cunningham
&
Carroll,
2011a;
Suggate
et
al.,
2011).
Moreover,
we
included
both
comprehension
and
reading
outcome
measures
for
the
older
children,
to
ensure
that
reading
skill
development
was
not
confined
to
more
constrained
measures
(Paris,
2005),
which
could
potentially
lead
to
ceiling
effects,
masking
group
differences.
Participants
in
Study
1
were
children
in
their
first
(junior
cohort),
third
(middle
cohort),
or
fifth
(senior
cohort)
year
of
pri-
mary
school
at
the
beginning
of
the
study
and
followed
across
two
years,
thus
providing
the
opportunity
to
observe
reading
perfor-
mance
across
the
period
of
elementary
education
in
New
Zealand.
Crucially,
the
key
independent
variable
of
RIA
was
determined
by
whether
children
were
recruited
from
state
schools
(RIA
of
five
years)
versus
Steiner
schools
(RIA
of
seven
years).
Table
1
presents
the
overall
design
of
the
study,
including
the
cohorts
and
measure-
ment
schedule.
2.1.
Method
2.1.1.
Participants
Participants
were
recruited
from
three
government-supported
Steiner
curriculum
(n
=
111)
and
three
state
curriculum
schools
(n
=
234).
Technically,
two
Steiner
schools
were
“state
integrated,”
receiving
full
governmental
funding
but
allowed
to
retain
their
special
character.
The
third
school
was
an
“independent
school”
which
also
received
some
governmental
funding,
but
at
a
lower
rate
(not
exceeding
30%).
This
independent
Steiner
school
(n
=
21)
was
included
in
this
research
to
bolster
sample
size
and
to
improve
representation
of
lower
SES
families,
as
this
school
drew
its
children
from
a
community
with
a
low
to
moderate
SES.
Government-
reported
indices
suggested
no
state/Steiner
differences
in
estimates
of
the
economic
affluence
of
school
communities
(see
discussion
of
school
decile,
below),
increasing
the
likelihood
that
the
samples
were
matched
on
SES.
During
the
course
of
the
study,
14.96
percent
of
state
and
9.91
percent
of
Steiner
children
left
the
schools,
with
the
differ-
ence
between
school
types
not
being
significant
according
to
a
Fisher’s
exact
test
(two-tailed,
p
=
.24).
One
of
these
children
left
the
study
near
its
completion,
such
that
these
data
were
still
able
to
be
included
in
the
analyses.
The
reasons
for
leaving
the
schools
appeared
to
be
due
to
a
residential
address
change,
according
to
teacher
reports.
A
further
13
children’s
data
were
eliminated
because:
(a)
either
there
was
too
much
missing
data,
defined
as
three
or
more
missing
data
points,
or
(b)
the
children
were
given
the
wrong
test
materials,
usually
due
to
their
having
been
in
a
lower
or
higher
academic
year
than
that
tested,
with
this
place-
ment
changing
in
the
course
of
the
study
(due
to
staggered
school
entry
according
to
birthdates,
some
children
enter
Year
1
directly,
whereas
others
enter
the
year
previously
denoted
as
Year
0).
This
left
190
state
and
97
Steiner
pupils,
13
of
whom
first
entered
state
schooling
and,
therefore,
had
an
RIA
of
five
even
though
they
later
attended
steiner
schooling.
The
reasons
why
13
chil-
dren
in
our
Steiner
school
samples
had
changed
from
a
state
school
are
not
known
in
every
instance,
but
include
dissatisfac-
tion
with
the
former
school,
and
waiting
for
an
available
place.
Accordingly,
two
variables
ensue,
either
(a)
curriculum,
with
chil-
dren
either
in
the
Steiner
versus
state
curriculum
during
the
study,
or
(b)
RIA,
with
children
either
beginning
school
earlier
in
the
state
school—regardless
of
whether
they
transferred
into
a
Steiner
school—or
later
and
entering
formal
schooling
for
the
first
time
in
the
Steiner
curriculum
around
age
seven.
The
early
RIA
children
began
school
at
60.09
months
(SD
=
.44)
and
the
later
RIA,
on
average,
19
months
later
(M
=
78.64,
SD
=
4.02).
Forty-one
of
the
80
Year
1
children
attended
school
in
2006,
since
their
fifth
birthday
was
prior
to
2007
(in
NZ,
school
entry
typi-
cally
occurs
on
the
day
of
the
fifth
birthday),
such
that
they
entered
school
at
the
normal
age
of
5
in
2006,
but
entered
the
study
when
a
little
older
at
the
beginning
of
2007.
Participation
rates
were
esti-
mated
from
Ministry
of
Education
and
school
roll
data
as
83.72
percent
for
children
in
the
Steiner-curriculum
and
59.77
percent
in
the
state
curriculum.
2.1.2.
Measures
A
description
of
which
measures
were
administered
and
when
appears
in
Table
1.
2.1.2.1.
School
decile
and
curriculum.
In
NZ,
the
Ministry
of
Edu-
cation
calculates
a
decile
rating
for
each
school
based
on
the
proportion
of
students
attending
from
low
socio-economic
status
communities.
The
decile
scores
(each
representing
10
percentiles)
are
integers
ranging
from
one
to
10,
with
ten
indicating
greater
affluence.
School
curriculum
indicated
whether
the
school
followed
the
state
or
Steiner
curriculum.
2.1.2.2.
Classroom
instruction.
In
both
2007
and
2008,
teachers
were
asked
to
complete
a
self-report
questionnaire
of
class-
room
reading
instruction
practice.
The
questionnaire
contained
six
domains
derived
from
research
on
classroom
activities
asso-
ciated
with
learning
to
read
(Foorman
&
Schatschneider,
2003).
The
domains
were:
oral
language,
learning
about
books,
learn-
ing
about
sounds,
learning
about
meaning,
reading
of
text,
and
writing.
Teachers
were
asked
to
describe
the
kinds
of
activities
engaged
in
during
class
time,
providing
an
accompanying
time

S.P.
Suggate
et
al.
/
Early
Childhood
Research
Quarterly
28 (2013) 33–
48 37
Table
1
Study
design
including
measures,
cohorts,
and
testing
points.
Cohort
Testing
point
Calendar
year
tested
State
school
measures
Steiner
school
measures
Junior-Elementary
School Year
1/Class
I
1 2007,
beginning ISF,
LNF,
PPVT-IV ISF,
LNF,
PSF,
NWF,
PPVT-IV
2 2007,
middle
ISF,
LNF,
PSF,
NWF,
HLEQ
ISF,
LNF,
PSF,
NWF,
HLEQ
3
2007,
end
LNF,
PSF,
NWF
LNF,
PSF,
NWF
Year
2/Class
II
4
2008,
beginning
LNF,
PSF,
NWF
LNF,
PSF,
NWF
5
2008,
middle
PSF,
NWF,
ORF,
RSC
PSF,
NWF,
ORF,
RSC
6
2008,
end
PSF,
NWF,
ORF,
RSC,
WI,
WA
PSF,
NWF,
ORF,
RSC,
WI,
WA
Middle-Elementary
School Year
3/Class
III
7 2007,
beginning
NWF,
ORF,
PPVT-IV
NWF,
ORF,
PPVT-IV
8
2007,
middle
ORF,
HLEQ
ORF,
HLEQ
9
2007,
end
ORF
ORF
Year
4/Class
IV
10
2008,
beginning
ORF
ORF
11
2008,
middle
ORF,
RSC
ORF,
RSC
12 2008,
end
ORF,
RSC,
WI,
WA,
ORF,
RSC,
WI,
WA
Senior-Elementary
School Year
5/Class
V
13 2007,
beginning ORF,
PPVT-IV ORF,
PPVT-IV
14
2007,
middle
ORF,
HLEQ
ORF,
HLEQ
15 2007,
end
ORF,
WI,
WA
ORF
Year
6/Class
VI
16
2008,
beginning
ORF
ORF
17
2008,
middle
ORF,
RSC
ORF,
RSC
18
2008,
end
ORF,
WI,
WA,
PC,
RSC
ORF,
WI,
WA,
PC,
RSC
Note:
ISF,
initial
sound
fluency;
LNF,
letter
naming
fluency;
PSF,
phonemic
segmentation
fluency;
NWF,
nonsense
word
fluency;
ORF,
oral
reading
fluency;
PPVT-IV,
Peabody
Picture
Vocabulary
Test
IV;
HLEQ,
home
literacy
environment
questionnaire;
RSC,
reading
self-concept;
WI,
word
identification;
WA,
word
attack;
PC,
passage
comprehension.
estimate.
Although
self-report
measures
may
be
subject
to
bias
(e.g.,
response,
desirability),
teacher-report
provides
a
broad
per-
spective
of
classroom
instruction
from
a
social
validity
perspective
(Erchul
&
Sheridan,
2008;
Gersten,
Baker,
Haager,
&
Graves,
2005).
Preliminary
evaluation
of
this
measure
indicated
it
captured
the
transition
from
code-based
to
comprehension
instruction
across
the
grades
(Barnes,
2008).
2.1.2.3.
Peabody
Picture
Vocabulary
Test
IV
(PPVT-IV).
The
PPVT-IV
was
administered
in
the
middle
of
2007
to
assess
receptive
vocab-
ulary,
which
is
strongly
linked
to
both
reading
and
verbal
skill
(Hodapp
&
Gerken,
1999;
Sénéchal,
Ouellette,
&
Rodney,
2006).
This
measure
(a)
has
good
supporting
technical
adequacy
evidence
(Dunn
&
Dunn,
2007);
(b)
is
widely
used
in
research
(Sénéchal
et
al.,
2006);
and
(c)
performs
well
in
NZ
(Cleveland
&
Reese,
2008;
Reese,
Jack,
&
White,
2010;
Reese
&
Read,
2000).
Standardized
scores
are
presented.
In
some
instances
children
were
absent
for
initial
PPVT
testing,
thus
this
was
administered
at
the
next
possible
testing
point.
2.1.2.4.
Home
literacy
environment
questionnaire
(HLEQ).
The
HLEQ
was
employed
to
account
for
differences
in
children’s
HLE
and
was
adapted
from
a
previously
validated
measure
(Griffin
&
Morrison,
1997)
to
encompass
the
NZ
state
and
Steiner
contexts
and
increases
in
technology
used
in
homes
since
the
questionnaire’s
inception.
There
were
nine
literacy-related
questions.
Factor