Page 1
8/10/2019 about frostig and art
http://slidepdf.com/reader/full/about-frostig-and-art 1/8
A
Follow-Up
Evaluation
of
the
Fr
ostig
Visual-P
erceptual
Train
ing Progra
m
DURIN
G the 1965-66
school year
a
n evaluation
of
th
e Frostig Visual-Perc
eptual
progr
am was conducted.
S
ubjects
were in
prekinderga
rten,
k
indergarten,
and first
grade.
In gener
al,
it wa
s found that
children
taking the F
rostig
p
rogram did
better on th
e
Frosti
g Visual-Per
ceptual
Test than did con
trols. Lar
gest gains were
regis
tered by
first
grader
s. After initial differen
ces in Fros
tig
scores
were can
celled
out,
no si
gnificant d
if
fere
nces in
end-of-year
reading readiness
scores
were
found among
kin
dergarten
pu
pils
in Frostig
and non-Frostig
classes. Finally
,
subtest analyses were
made andrevealed
that
pr
actically all
the differen
ces
in Frosti
g total
test
sco
res noted abov
e
were
due
to
one
sub-
test; i.e., Form Co
nstancy.
Readers
who
are
not
acquainted
with the
Frostig
V
isual-Perceptua
l pro
gram or the in
strume
nt, are
referred to
the
original
study
or to
t
he Frostig
manual
.-
The
purpose
of
th
is study
is
to re
plicate
as much as possible
the forme
r study a
nd
to follow up the
children
in
the study t
o deter
mine th
e
pr
edictive v
alidity of the Frost
ig
tests as
well as to study
the cumulat
ive
James N Jacobs.
An
Evaluati
on of
the
Frostig Visual-
Perception Training Pr
ogram E
5 4): 33
2-40; January
196
8.
effects
of the Frostig
program
upon
future
re
ading achievement.
Specifically
, this study
is de
signed to
answer
the
follo
wing q
uestions:
1 Do
children
i
n p
rekindergarten,
kind
er
garten, and
grade one
who take th
e
Frost
ig pro
gram achieve
better tha
n controls
on the Frostig
Visual-P
erception Tes
ts? Is
the
b
enefit unifo
rm
among grades tested?
2. Do
kindergarte
n Frostig
children
achieve
high
er on
the
Metr
opolitan
Reading
Readiness Test
at
the
end of their
kindergarten
year
as compared to
controls?
3.
What
is
th
e predic
tive
validit
y of the
- M. Frostig, P.
Maslow, W.
LeFever,
an
d
J R
. B. Whitt
lesey. The M
alo
Alto, Cali
fornia: Consulting Psy
chologists Press,
1964.
(Stand
ardization.)
Errors
in
generalization
are
common
to
studies of this
type.
B
efore
practical de
cisions re
gardin
g
th
e
use
of the treatmen
t are made, replica
ti
on studies
should
be
made.
J
hio, a
169
Page 2
8/10/2019 about frostig and art
http://slidepdf.com/reader/full/about-frostig-and-art 2/8
Frostig test a
s comp
ared
t
o
the
Metropol
itan
Readin
g Readiness T
est?
4. Do Fro
stig trained
children achieve bet
ter on read
ing tests as
compared
to co
ntrols?
5.
Do childr
en
exposed
to two years
of
the
Frostig
p
rogram
achieve
better on rea
ding tests
as compa
red to one-year Frostig
ch
ildren and
controls?
That
is
is
there
a
cumulative
effect
due
to
F
rostig inv
olvement in reading
achieve
ment?
Th
e first two qu
estions
were pursued
in
the initial
study and are parti
ally
r
eplicated
in
this study. The
last two questions
had
to
be delayed
until
F
rostig pr
ogram chil
dren
last
year
ad
vanced in
grade.
The
pupils
used
in
this study
were
iden
tified as
ex
perimental
(Frostig)
or
control.
They
attended
five
dif
ferent schoo
ls
a
nd w
ere
in
intact
classes
. For most com
parisons,
con
t
rols were drawn from
the same
schools as
experimen
tal
cla
sses. Further
, the pupils
may be
co
nsidered
disadv
antaged since
they
attend schools
d
esignated as
primary t
arget
s
chools unde
r
the
Education
Act.
Educati
on
A
ct servic
es,
which
may bias result
s, are con
sid
ered
to
be
present in eq
ual force with
experimental
and
control children. Thus,
no
systemat
ic bias
is
exp
ected.
In
answering
the
fi
rst two
questions, the
Fro
stig tests and
the Metrop
olitan Readi
ng
Readin
ess Test were
admi
nistered
to
expe
ri
mental
a
nd
co
ntrol children in
May onl
y.
The
initial s
tudy
inclu
ded
beginning-of-y
ear
Frostig testin
g as well as end
-of-year
testing.
T
his
feature cou
ld
not
be
repli
cated
this past
year. The
analysis
consisted simpl
y
of
a two
way
analysis
of var
iance of gra
de by
treat
men
t
u
sing t
otal Frostig test
score
as
the
crite
rion. For
reading readines
s comparisons
only
kinderga
rten
childre
n were teste
d. Thu
s,
a one-way
analysis of variance
was applied.
All
ex
perimental children
in this phase of the
study took
the Frostig program
one year
only.
The
last
three
qu
estions were
studied
by
administering the
Gates-Ma
cGinitie test
to
first-grade children
in May and the
Stanford
Primary
I at the en
d
of g
rade two.
Since
th
e
Gates
produces two su
btest scores, i.e.,
vo
cabula
ry and comp
rehension,
separate analy
ses on
these two
tests were
made.
The firs
t
gr
ade experim
ental children tested
with the
Gat
es we
re
su
bdivided into two
groups:
on
g
roup having
ta
ken the Frostig
program
dur
ing the first
grade only;
and the second gro
u
who
had taken the Frostig
program
during
their kindergarten
year
as well
as
during th
first grade. Thes
e two
groups were in
turn
compared
with
co
ntrol
children who
did no
have the Frostig pro
gram
either
year.
T
he
qu
estion of
the effects of th
e Frosti
program
on rea
ding achievem
ent was studie
in anot
her
context.
Second-
grade childre
n
i
1966 6
7 who
were in
volved
i
n
th
e initia
l ex
perimen
t
as
first graders in
1965-66 wer
iden
tified and com
pared with
respect
to
end
of-year (M
ay) reading achieve
ment using
th
Stanford Primary
I
battery. The
pupils use
in this
phase had taken
the Frostig pr
ogram
only
during the first grade
last year as
com
p
ared
to
c
ontrols who
did
not have the Frosti
progra
m.
summ
ary
of th
Frostig
test results
is
seen in
Table
1
Th
children in
the exper
imental group took th
Frostig progra
m only
for
the
given ye
ar .
Results of the analysi
s of
varianc
sh o
wed
sig
nificant
differences be
tween trea
m
ents and a
mong grades,
but no
si
gnifican
inte
raction
betw
een
treat
ments
a
nd grade
Thu
s, we observe an
average differen
ce o
2.73 in
favor
of
experimen
tal pupils
o
n th
Fr
ostig test. In
the
ab
sence of
interaction w
conclude
further that th
e benefits ac
cruin
from the Frostig
program
on
the Frostig te
score are
eq
ual
to prekinderga
rten, kinde
garten,
and firs
t
grade
pupils.
These
results
bear
out the initi
al
stud
with
one exception:
significant
interact
io
was
noted
in the i
nitial study, with
first-grad
pupils show
ing
sig
nificantly mor
e gain th
a
prekinder
garten
or
k
indergarten childre
This
same
result
o
ccurred but la
cked stati
st
cal signi
ficance. Most
important,
howeve
r,
dete
rmination th
at
the
Frostig program
doe
Page 3
8/10/2019 about frostig and art
http://slidepdf.com/reader/full/about-frostig-and-art 3/8
ro
du
ce
hig
he
r
Fro
sti
g s
cor
es
in
c
om
par
iso
n
o
co
ntr
ols
.
The
d
iffe
re
nce
no
te
d in
int
era
c
ion
pro
ba
bly
is
th
e res
ul
t
o
f
the
fac
t t
hat
a
ov
aria
nc
e a
nal
ysi
s w
as
a
pp
lied
in
the
in
itia
l
t
udy
u
sin
g
Fr
os
tig p
re-
tes
t s
cor
es as
a
con
l me
asu
re
w
hi
le t
his
c
ou
ld no
t be
d
one
in
hi
s
inv
est
iga
tio
n. S
in
ce
th
ere
is
litt
le re
as
on
o
believe
that
the
experimental
pupils began
e
y
ea
r at
a hig
he
r p
re-
tes
t l
eve
l th
an
co
n
rol
s, w
e m
ay
int
erp
re
t
t
he
hig
he
r
po
st-t
est
ore
s
a
s
r
esp
on
din
g to
the
Fro
sti
g
pro
gra
m.
M
ost
of th
e c
hil
dre
n
at
th
e k
ind
erg
ar
ten
lev
el
w
ho
se
F
ro
stig
sco
res
a
re
rep
ort
ed in
T
ab
le
also
took
the
Metropolitan Reading Readiness
T
est
at
th
e
en
d of
t
he sc
ho
ol
ye
ar.
T
he
m
ea
n
r
aw
sc
ore
m
ade
on
the
re
adi
ng
re
adi
nes
s
t
est
for
ex
pe
rim
en
tal
pu
pil
s
wa
s
3
7.2
2 an
d
a
s
tan
da
rd
d
ev
iati
on
of
14.
4.
Kin
de
rga
rte
n
c
on
tro
l pu
pils
ha
d a
m
ea
n
r
aw s
co
re
o
f
29
.9
0
o
n
the
rea
din
g
rea
din
es
s an
d
a sta
nd
ard
de
v
iat
ion
o
f
15
2
Th
e di
ffer
en
ce of
7.
32
fav
or
in
g
th
e ex
per
im
ent
al
p
upi
ls w
as
fo
un
d to
be
sig
nif
ica
nt
t=
2.
63,
d
f.
112
).
Es
sen
tia
lly
t
he
sam
e
co
m
par
iso
n w
as
made
in the
initial
study.
The
results
showed
the
exp
er
ime
nt
al
g
rou
p to
be
h
ig
her
tha
n
co
n
tr
ol
bu
t d
id
n
ot
re
ach
a
st
ati
stic
all
y
sig
ni
fi
ca
nt
d
iffe
ren
ce
. In
vie
w of
thi
s
rep
lic
ati
on,
it
see
ms
saf
e to
say
th
at
k
ind
er
gar
ten
p
up
ils
w
ho
ta
ke
the
F
ros
tig
pro
gra
m
gen
era
lly
wi
ll
sho
w hi
gh
er
M
etro
po
lita
n R
ea
din
g
R
ead
ine
ss
sco
res
.
S
inc
e
pu
pils
w
ho
tak
e t
he F
ro
stig
pr
o
gra
m
ac
hie
ve hi
gh
er o
n
b
oth
th
e F
ro
stig
test
a
nd
o
n
the
M
etr
opo
lita
n
Re
adi
ng
R
ea
din
ess
T
es
t
in co
mp
ari
son
to
c
on
tro
l, it
see
ms
r
ea
s
on
abl
e
to
c
on
clu
de
tha
t th
ere
i
s a
cor
rel
ati
on
between
the
two
tests. The correlation
co
eff
ici
ent
o
bt
ain
ed
bet
we
en
t
he
Fro
sti
g
te
st
s
cor
e
and
the
M
e
tro
pol
itan
R
ea
din
g
R
ea
di
n
es
s
raw
sco
re w
as
.55
6
a
mo
ng
t
he e
xp
eri
m
en
ta
l
pu
pils
a
nd
.65
9
am
ong
c
ont
rol
s.
T
hes
e
c
orr
ela
tion
s a
re
n
ot s
ign
ific
an
tly
dif
f
ere
nt
a
nd
r
efl
ect
a
m
o
der
ate
d
eg
ree
of
r
ela
t
ion
shi
p b
etw
ee
n t
he
tw
o
tes
ts.
ha
t i
s t
h
e
fir
st-g
rad
e pu
pil
s u
se
d
in
th
e
init
ial
stu
dy
we
re
in
the
sec
ond
gra
de
in 19
66-
67
.
Th
e
Fro
sti
g
sc
ore
s o
f
the
se
pu
pil
s
ob
tai
ned
i
n
S
ep
tem
be
r
o
f
th
e f
irs
t g
ra
de a
nd
aga
in
in
M
ay
of
t
he
fir
st g
rad
e we
re
b
oth
co
rre
lat
ed a
gai
nst
St
an
for
d
Pri
ma
ry
I r
ead
in
g
su
bte
sts
.
T
he
M
etro
po
lita
n
Re
adi
ng
Re
adi
nes
s
Te
st
sc
ore
s,
ob
tain
ed
by
ess
en
tia
lly
the
sa
me
pup
ils
at
th
e
end
of
t
he k
in
der
ga
rten
ye
ar, w
er
e al
so
co
r
r
ela
ted
w
ith
the
S
ta
nfo
rd re
ad
ing
su
bte
sts
.
The
results
of
these correlations
are seen in
Ta
ble
2
se
e
p
age
1
72
).
It
s
ho
uld
be
bo
rne
i
n
m
in
d th
at
bo
th
Fr
ost
ig
t
est
s w
ere
gi
ven
o
n
da
tes
c
o
th
e adm
in
ist
rat
ion
of
th
e
cr
iter
ion
re
ad
ing
tes
ts th
an
w
as
th
e M
etro
po
lita
n
R
ead
in
g
4
3 1 5
4 0
42
Page 4
8/10/2019 about frostig and art
http://slidepdf.com/reader/full/about-frostig-and-art 4/8
Page 5
8/10/2019 about frostig and art
http://slidepdf.com/reader/full/about-frostig-and-art 5/8
high
est
gro
up of correlations
was between
the Metropolitan
Reading Readin
ess
Test and
the re
ading test within the
control group, one
may con
clude that the Metrop
olitan tests
seem
to
be better predictors
of reading
ach
ievement
th
an the Fro
stig tests.
This
question was pursued by studyi
ng the
av
erage reading
scores of
th
e
same
pupils
on
whom the
co
rrelations in
Table 2 were
based.
It
should be bo
rne
in mind
that
these
reading
tests were
given
at the end of gra
de
two
while
the
experimental
pup
ils we
re
defined
as
pupils
having
one y
ear
of
the Frostig pro
gram
in the
f s compared to con
trols who did not
take
the
Frostig program
in
the
first grade.
Summary
of
the results
is
shown in Table
3
The
data
shown in
Table 3
bear
out
what was
learned previously, i.e., th
at there
is
little relationsh
ip
betwe
en Frostig
scores
and readin
g
achievem
ent. Thus we n
ote that,
while
the
Frosti
g sco
res of
the experimental
gro
up exceed
those
of the
controls by 3.55,
ne
vertheless the
reading subtest
scores
are
consistently
in
f
avor of the control
group.
Two of the three
subtests
show
sig
nificant
differences favoring the control
group.
In
all
probab
ility the reading d
ifferences favoring
co
ntrols
are
due to
initial
ly
h
igher levels
of
ach
ievement by
the control grou
p.
Thus
it
is apparent that
high
Frostig
test
per
form
ance is no
guarantee
of higher
reading
achievem
ent.
irs
t-grade
Frostig childre
n
in
1966-67
were
divided
into
two
groups, one group wh
o had taken the
Frostig program
as
kindergarteners
and a
second group
who were
be
ginning the
Frostig
progr
am for the first time in
the first grade.
End of
first-grade Frostig scores were
com
pared
between
t
he two-year an
d one-year
Fr
ostig pupils as well
as
against
controls. In
additio
n, this first-grade group
was given
the
Gates-M
acGinitie test at
the
end o
f
their first
grade, thus
per
mitting compariso
n of reading
achievement
of two-year Frostig
pupils, one-
year
Frostig program pup
ils, and
controls.
These results are sh
own
in
Table 4
Inspection
of Table 4 reve
als immedi
ately
that
the
two-year Frostig gr
oup, i.e.,
pupils w
ho took the Frostig program
in kin
dergarten a
n grade
one perform
ed higher
on
the
Frostig test
51.57)
than did
pupils
who
took
the Frosti
g
prog
ram in grade one
onl
y 48.68). Both
of
the
Frostig groups in
turn scored hig
her
on
the
Frostig
test
tha
n
did
the
control pupils
who had
a
mean Frostig
score of 44.06. Thus,
it
is apparent
that the
Frostig program does
have cumulative effect
s
o
n
the
Frost
ig test s ore the longer a pupil
is
in the program, th
e higher is his
Frostig
test score.
C
omparison of
the Gates readi
ng
score
s
shows
a
picture
favorable to the Frostig
groups.
In b
oth
the voca
bulary and com
prehension subtests
b
oth
Fro
stig
g
roups
scored higher than
did
the
control
s.
One-
y
ear
and
two-year Frostig
progr
am
p
upils
te
nd to
score similarly on
both
vocabulary
and com
prehension
while both
score
com-
Page 6
8/10/2019 about frostig and art
http://slidepdf.com/reader/full/about-frostig-and-art 6/8
pa
ratively
higher than
controls.
For
both
subtests howeve
r
an
alysis of
varian
ce
showed no signifi
cant difference among the
th
ree
groups
on
either the vocabulary
or the
comprehension subtests.
The lack of signifi
cant
difference revealed by the analysis of
variance was a result of
lar
ge variation
within each
of
the
three
groups.
I
n view of the comparati
ve achievement
of Frostig and control
pupils
on
reading
tests
at
the
end of grade
tw
o
re
ported earlier),
one
is inclined to
accept
th
e
analysis
of vari
ance at face value and
conclude
that
there
is no signi
ficant difference in the achieve
ment of
Frostig
pro
gram pupils
and
controls
on
reading tests
at
the end of
grade
one.
The results of
this
study
may
be
sum
marized briefly
as
follows:
1
Pupils who take the Frostig
program in
prekindergarten kindergarten or first grade
tend to show
higher
le
vels
of
visual-perceptual
performance on the Frost
ig
test
in comparison
to control
pup
ils.
2.
There is a moderate degree of corre
la
tion between Frostig scores and
the Metropolitan
Reading Readiness
sco
res resulting
in
the
fact
that
higher scoring Frostig pupils tend
to
be
higher
sco
ring pupils in terms of reading readi
ness
te
sts.
3. The predictiv
e
validit
y of
both
the
Fros
tig
test
and
the Metropolita
n Reading
Readiness
Test s
quite
low using standardized
reading tes
ts
as
criteria.
4. Pupils who take the
Frostig program
seem
to
have no particular
adva
ntage
as far as
future
reading achievement is concerned as
compared to pupils who do not take
the Frostig
program.
5. Pupils who take the Frostig program
fo
two con
secutive years
achieve
higher on the
F
rostig test compared to pupils who take
the
Frostig
program
for one
year
who
in turn
achiev
e
better
th
an
pupils who do
not
take
the
Frostig
program in
either
grade.
One other finding not discussed in
the
body of
this
report relates
to
subtest
analy
ses. In the original
study
the only
subtes
which show
ed
a
significant difference favor
in
g Frostig pupils
was
the Form Constancy
subtest. This
finding
was disturbing
in tha
we could
not explain
why
this should
happen
In the present
study however
we found
tha
experimental pupils in general performed
higher
on a f the five
subtests
than
did the
controls.
The differences favoring the exp
eri
mental
Frostig group
on
each test were as
fol
lows: Eye-Motor
Coordination
+.81;
Figure
Ground + .61;
Form
Cons
tancy
+.62
Position
in Space
+.8
2;
Spatial Relations
.35 . These
results are much
mor
e reason
able and highlight the
im
portance of rep
lica
tion
in
studies of
this
type.
O
ne could argue that the
differences in
mental ability of
the comparison groups coul
account for
any
observed
differences in
cri
terion
mea
sures. In this regard
it
is inter
es
ting
to
note
that
the mean
IQ's
as measure
by
the
Pintner Cunningham
Mental
Abilit
Tests
of
the groups shown in Table 4 wer
92
for the
two-year
Frostig
g
roup 91 for th
one-year Frostig group
and 88
for the con
trol group. Since the
se
ment
al tests wer
gi
ven at the
end
of the year or at
the sam
time
tha
t the criterion tests were
adminis
tered one might argue that
the difference
in
IQ were a respons
e
to
the Frostig program
One could als
o
argue that
these difference
in
mental
ability while small may
accoun
7
Page 7
8/10/2019 about frostig and art
http://slidepdf.com/reader/full/about-frostig-and-art 7/8
Page 8
8/10/2019 about frostig and art
http://slidepdf.com/reader/full/about-frostig-and-art 8/8
Copyright © 1968 by the Association for Supervision and CurriculumDevelopment. All rights reserved.