The Keating-Manis Phoneme Deletion
Test
Patricia A. Keating (UCLA Linguistics Dept.)
Frank Manis (USC Psychology Dept.)
This test was prepared under NIH grant R01 HD29891, "Perceptual,
Linguistic, and Computational Bases of Dyslexia", to Frank Manis. Here we give information for test-users, including classroom
teachers.
1. Introduction
The Keating-Manis
Phoneme Deletion Test is a new deletion and elision
test of phonemic awareness. Phoneme deletion tests
go back to at least Rosner (1974, in The Reading Teacher). Our test differs from other phoneme deletion tests in
that the test items were constructed systematically, in accord with discussions
of linguistic difficulty by Yopp (1988, in the Reading Research
Quarterly, and more particularly by Stahl and Murray (1994, in the Journal of Educational Psychology), and other linguistic considerations. Also, our test is administered via a tape recording, so
that every child hears the test items pronounced the same way. The teacher does not have to pronounce any isolated phonemes.
The items in
the test are in order of increasing difficulty, based on errors made by children
in a pilot study.
2. Instructions to the test-giver
First part of test: real word items
1) The test should be administered using headphones
if at all possible. This is better for the child;
however it means that you will not be able to hear the test items unless
you have a "splitter" for the tape recorder output so that two headphones
can be used together. If you use a loudspeaker, be
sure the area of the room used for the test is reasonably quiet. (That is, can you yourself hear the tape well?) Adjust the playback volume if necessary.
2) Tell the child that this is a game about turning
one word into another word. For each item, first
the talker asks the child to repeat a word. There
is a short pause on the tape for this. Then the talker
asks the child to say the word but leaving out part of it (which makes a
different word). There is another pause on the tape
for this. Play the practice items on the tape. Give instructions again and/or repeat practice items if
necessary.
3) Play the test items. If
the child hesitates before repeating a word, pause the tape. If the child repeats a test word incorrectly, stop the
tape and ask the child to listen more carefully. Then
rewind and repeat. (It's important, for the test
to be meaningful, that the child have heard the test word correctly; that's
why we have the child repeat the word, so that you can check this.)
4) When the child gives the deletion response,
mark score sheet as right or wrong. If the child self-corrects,
score the last response. If the child cannot respond
to an item, mark it wrong.
5) Stop the test if the child makes 5 errors in
a row.
Second part of test: nonsense word items
(You may decide not to give this part of the test
if the child completes only a few of the items in the first part.) The instructions for this part of the test are the same
as above, except here, the game is to turn one space-alien word into another
space-alien word.
3. Test items and answers (make
copies of this page and next to use for scoring)
Some words as pronounced here can
correspond to more than one written word. These are
indicated here. They are given only for your information;
since this is an oral test, the spellings of the test words and answers are
not part of the test. For example, in item 3, you
won't know whether the child is thinking about the word "buy" or the word
"by" -- and it doesn't matter. In 3 items, however,
the pronunciation of English permits alternative right answers, and these
are marked "either is right".
First part: Real-word items
Practice
items:
SAY:
delete: correct answer:
a. baby bay bee
b. picnic nic pick
c. meal /m/ eel
d. beam /m/ bee
Test
items:
1. mat /m/ at
2. shone/shown /sh/
own
3. bike
/k/
by/buy
4. trade /tr/ aid
5. boil /b/ oil
6. same /m/ say
7. float /fl/ oat
8. pile /l/ pie
9. steal/steel /st/ eel
10. wild /ld/ why/Y
11. film /m/ fill
12. snow /s/ no/know
13. wind (long i)
/d/
wine/whine
14. steer /s/ tear/dear (either
is right)
15. blue /l/ boo
16. mist/missed /t/
miss
17. act
/k/
at
18. first /st/ fur/fir
19. locks /ks/ law/la
20. fly
/f/
lie
(cont. on next
page)
21. sneeze /n/ sees/seas
22. sweat /w/ set
23. bring /b/ ring
24. roast /s/ wrote/road (either
is right)
25. field /l/ feed
Second part: Nonsense items
Practice
items:
a. sab
/s/ ab
b. smig
/sm/ ig
c. nald
/d/ nal
Test
items:
1. mag
/m/ ag
2. stib
/st/ ib
3. leet
/t/ lee
4. narp
/p/
nar
5. yits
/s/
yit
6. rinch /ch/ rin
7. grem
/g/
rem
8. snup
/s/
nup
9. floke /f / loke
10. blick /l/ bick
11. stome /t/ soam
12. kimp /m/ kip
13. snool (long u) /n/
sool
14. bilp /l/ bip
15. doast /s/ dote/dowed (either is right)
Interpreting
the results:
Count the number correct on the two parts together. Then relate this score to the reading grade level as measured
by the Woodcock Word Identification test, for a sample of 181 children in
grades 1-3 who took both tests (and who spanned a wide range of reading
ability). (Note that this is not a normed test! As our study continues, we will refine this relationship,
and we will be acquiring data on older children. But
at least this gives a starting point.) This will give
you an idea of how the child's phonemic awareness, as measured by this test,
compares with the average for that grade-level achievement
in reading.
Number correct Reading grade level
1 < 1
2 "
3 "
4 "
5 "
6 "
7 "
8 "
9 "
10 "
11 "
12 1.0
(1st grade scores by children who can delete all
13 1.2
initial consonants - singles and blends -
14 1.4
and some finals)
15 1.7
16 1.8
17 2.0
(2nd grade scores by children who also delete
18 2.4
first consonant in initial blend, and last consonant
19 2.6
in final blend, but mostly only in real words)
20 2.7
21 2.9
22 3.0
(3rd grade scores by children who delete some
23 3.2
consonants at the beginning of a word-final blend)
24 3.5
25 3.7
26 3.9
27 4.0
(4th grade scores by children who do as above,
28 4.2
but more reliably)
29 4.5
30 4.8
(our data so
far don't tell us about scores above this)