ED 243 964
AUTHOR
TITLE
INSTITUTION
PUB DATE
NOTE
AVAILABLE FROM
PUB TYPE
EDRS PRICE
DESCRIPTORS
IDENTIFIERS
ABSTRACT
DOCUMENT RESUME
TM 840 287
Bliss, Leonard B.
Item Analysis and Report of Student Skills of
Secondary School Students. Technical Report #2,
Public School Basic Skills Achievement Survey.
-
College of the Virgin Islands, St. Thomas. Caribbean
Research Intt.
Jun 82
103p.; For Technical Report No. 11 see ED 227 150.
Caribbean Recaz.,.h Institute, College of the Virgin
Islands, St. Thomas, Virgin Islandt 00801 ($4.00).
Reports
Research/Technical (143)
MF01/PC05 Plus Postage.
*Academic Achievement; *Achievement Tests; *Basic
Skills; Difficulty Level; Elementary Secondary
Education; *Item Analysis; Language Skills;
Mathematics Skills; Reading Skins; *School
Districts; Test Items
Stanford Achievement Tests; Stanford Test of Academic
Skills; *Virgin Islands
A sample of slightly over 1,500 students from
even-numbered grades in public schools of the U.S. Virgin Islands
were given the 1973 edition of the Stanford Achievement Test (in
grades 2, 4, 6, and 8) and the Test of Academic Skills (grades 10 and
12) in an attempt to assess student academic achievement
in the basic
skill areas of mathematics, reading, and English Language. This
report describes the data analysis which involved a detailed item
analysis of each item on each test given to the sample of
students in
grades 8, 10, and 12 as well as a summary of
student skills based on
their achievement along objectives provided by the test publisher and
keyed to individual test items. Measures of
item difficulty and item
discrimination were calculated for the entire territorial
school
system and for the individual school districts. (Author)
***********************************************************************.
*
Reproductions supplied by EDRS are the best that can be made
*
*
from the original document.
*
********************************************************),*************
Virgin Islands of the United State's
Public School Basic Skills
Achievement Survey
Technical Report # 2
Item Analysis and Report of Student
Skills of Secondary School Students
Caribbean Research Ingtitute; College of the Virgin Islands
Leonard B. Bliss, Ph.D. - Principal Investigator
June 1982
Contents
1
Abstract
2
Introduction
3
Background
5
Item analysis
6
Difficulty indices
6
Discrithination indices
7
Summary of student skills
11
Availability of data
13
8th Gi.'ado (Advanced Level)
14
Vocabulary
16
Reading Comprehension
18
Mathematics Concepts
29
Mathematics Computations
32
mathortics Applications
36
Spelling
40
Language
43
10th Grade (TASK Level I)
51
Reading
53
English
63
Mathematics
67
12th Grade (TASK Level II)
73
Reading
75
English
84
Mathematics
88
Discussion
96
References
100
4
Abstract
A sample of slightly over 1500 students from even
numbered grades in public schools of the U.S. Virgin Islands
were given the 1973 editiOn. of the Stanford Achievement Test
(in grades 2,
4,
6, and 8) and the TeSt of Academic Skills
(grades IQ and 12) in an attempt to assess student academic
achievement in the.basic skill areas of mathematics,
reading;
and English Language.
This report deSdribeS Phase II of the
data analysis which involved a detailed item analysis
of each
item
each test giv6n to the sample of students
in grades
;--_
8, 10; and 12 as well as a summary of student skills based on
their achievement along objectives provided by the test
publisher and keyed to individual test items;
Measures. of
item difficulty and item discrimination were calculated
for
the entire territorial Sdhbbl system and for the individual
school districts
At the request of the US.
Virgin Islands Department of
8-du-cation and the iioard. of Trustees Of the College of the Virgin
Islands, the Caribbean Research Institute embarked
of basic skills achieVeMent i
It soon became clear that
virgin islands
any Strategy designed to
skills achievement needed to Start Off with
on a study
public schools,
improve baSid
_
\
a fairly detailCd
description of current achieVeMent levels of student:s in
territorial public schools
It Was founa that this was not
available.
The task force set up to design
the study decided that the
Most efficient way to obtain information on levels of basic
skills achievement was
to administer a Standardized achievement
test to a representative sample of students and to analyze
the
1982)
results of this test.
Technical_ Repo_r_t_l (Bliss,
describes, in detail, the process used to choose an appropriate
test;
Finally, the Stanford Achievement Test
(1973 version)
was chosen as the instrument of choice.
Briefly, the reasons
for choosino this adhieVetent test battery were that 1)
it covered
the grades K-12, 2)
it seemed, on initial
oCServation, to be a
good match with subject content taught in the schools,
3)
it was
technically sound given the population on which it was standardiZed
(a seemingly representative sample of mainland U.S.
students); and
4)
it would report out criterion referenced results;
Due to various organizational
and fiscal contraints only
students in even numbered grades were tested.
This seemed
acceptable since many of the objectives tested by the Stanford
Test carry across adjacent levels of the test an-,
no reason to suspect that the patLerns of academic
of students in odd numbered gra
s were difforen
in even numbered grades.
These constraints are
in Technical Report #1.
of students to be tested was done using a
technique with classes as the cluster unit.
Ti-ii Ciet;
17=iplino procedure can be found in Technical lepOi-t
discussion of the effects of using this sami,
On the information obtained.
Table I presents a
awn of the sample as it finally emerged;
Table 1
U.S. Virgin Islands Sample Sizes
Test Level
Total
St.Thomas
System
St. John
District
St.
C.
1)i st. r
TASK II
129
74
TASK I
254
167
/
Advanced
345
173
1 7
Intermediate
II
227
146
Primary TIT
346
186
Primary I
234
143
ji
Total
1535
889
(.16
3
'sting was done
at the grade level recommended by the
`c.. c. i-,UbliSher during the Week of October ni 1980 in the
r:MiS/St. John district and dUting the week of December 1,
in thO St. Croix district. Testing materials and completed
-et.-0 collected; answer Sheets checked to deterMine
with marking instructions;
and answer sheets sent
to ')o machine scored.
BackgrbUnd - Technical
Report 42
is the socond of four reports that Will deal with
thc :_:osults or the basic skills assessment deSeribed above.
Tc!chnil Report U.
detailed the procedures used in test
selection
SaMpling; and test administration;
More importantly;
it estblished empitically; the content validity and the
rilLibility of the Stanford AchieVeMent Test when-it was
a,Thihistered to a Sample of
b;S. Virgih Islands bdbli6 school
This iS particularly important since there exists
i:JcLr(lizod test Of academic achievement which includes
t(idonts in its Standardization group.
report examines the scores of secondary school students
0; and 12) and presents
:
I)
al: item analysis of each item on each test of the
battery which includes indices of item difficulty
and discrimination
and
2)
a summary of student skills based on
their scores on
the itOM8 keyed to specific objectives.
6
ii-em _Analysis
Dfficulty indite8
The difficulty index of an item is found by taking the
number of examinees scoring
correctly on an item and diViding
it by the total number of students taking the test.
In shOrt,
it is the proportion of examinees who scored correctly on an
item and has a range from 0
(no students SCoting correctly)
to
1 (all students scoring correctly).
BecauSe of the fact it
is
a proportion, it is often designated in the
literature as
"p" (e.g. p=.75).
It may be worthwhile to point out that the
term "difficulty indo" may be somewhat of a misnomer since
items with high difficulty indices are actually leS8 difficult
than items with low diffitulty indices-.
NevertheleSS, the term
and its definition have become standard in the area Of psycho-
metrics throughout the United States.
Difficulty indices for each item on
each test were reported
out by the test scoring service.
In addition; difficulty indiC88
for examinees in the standardization group in the same grade as
local examinees at approximately
the same time of year are
reported.
The test scoring service
used a Chi-squared test
for
proportions Oh each difficulty index to test the hypotheses
that
the proportions of local
students scoring correctly on individual
items in greater or less than the proportions of examinees in
the standarditatiOn group scoring correctly at the
.05 level
of significande.
Significant differences
in either direction
were reported out.
9
7
DiscriMinatien incicos
The itern Oicrr1natLon :ndex indicates the degree to
Which resp6nse:, Ch
ltaM
related to responSeS On Other
items on
he teeL.
ThO
itic indicates whether a person
who does well on thc? tt
Whole (that is
a person who iS
presumably high On the
trait betricj measured) is more likely
to -9iet the particular iern correct than a person who 6OeS
poorly on the test as a whole.
In other words; the item
diSCriMinatic:i
whethet ail item discriminates
between thdSe who lo
an
those who do poorly on the test
as a whele.
Taking the item difficulty and the item discrimina-
tion indek ibtb COnSideratioru
the developerS of tests desire
to construct tests WhiCh discriminate well among examinees with
varying' levels of a trait,
The item disc:rtMination
index is calculated by the formula
,
L
Where
= thL:newiher
Ohb have total test scores
tOlal test scores and who also
have the it.pr; cnrrct.
L = the numb'el: of examineos Who have total test scores in
the lbwy range of total test 'scores and who also have
the item coi-ect
N - the
in the upper or low range
C) 1.
the test score,.
8
By definition; d is the difference between the proportion of
high scoring examinees who got the item correct and the proportion
Of loW scoring examinees who got the item
correct;
Upper and
lower ranges generally are defined as the upper and lower 10% to
33% of the sample, With-examinees ordered on the basis of their
total test score.
When total scores are normally distributed,
using the upper and lower 27% produces the best estimate of d
(Kelly, 1939).
If the distribution of total test scores is flatter
than the normal curve, the optimum percentage is larger and approaches
33% (Cureton, 1957).
However; Allen and Yen (1979) fOund that;
for most applications, any percentage between 25 and 33 will yield
similar estimates of d.
In this study; 27% was used as the upper
and lower percentage because examination of selected distributions
Of actual test scores revealed nearly normal distributions.
The theoretical range of d is between -1 and ±1.
However;
maximum discrimination is likely to occur when p=.50:
When
p=.50 the Varian-C-6 in item scores, which is p(1-p), is maximized.
As an item becomes more difficult, it is less
likely that any student
will score correctly on it.
As it becomeS less difficult it
is
more likely that any student will get it correct.
This could lead
to the suggestion that all items should have p=.50, but the useful-
ness of this suggestion is mitigated by interrcorrelations among
items.
In an extreme case; if the items on a test all interrcorre-
latod Perfectly and had difficulties of .50i
half the examinees would
receive a total test grade of zero and the other halt would have
perfect test scores.
Hence, there would be no
fine discrimination
between examinees' levels of achievement or whatever trait
Was being measured; _.,
general, test designers tend
to try
to choose items with a range of diffitUltieS that average around
.50.
IteMs of particularly low diffietlty are often included
in tests (usually among the earlier
items) for motivational
reasons.
Itefii discrimination indices were
calculated in this study
to provide indications that items may be flawed when used with
U VI students:
Such flaws are ambiguity;
the presence of
clues; the presence of more
than one correct answer,
and other
technical defects;-
If none was found upon examination of the
iteM,
and it was determined that the item did; indeed; appear
to measure the objective it was
intended to; the i-6-2:M was
indlUded in the overall analysis
of results.
Any item that
discriminate positively can make a contribution to the measure-
ment of pupil achievement and low indices Of discrimination
are frequently Obtained for reasons other than
iteM defects;
Standardized
achievement tests are designed to measure
several different types of learning outcomes (e.g.
knowledge;
understanding; application, etc.).
Where this is the case,
test items that represent an area receiving relatively little
emphasis will tend to have pbbr discriminating POw6r.
For
example, if a test has forty item measuring kh6isi16d4e
of
specific facts and ten items measuring understanding;
the latter
items can be expected to have low disatimination indices.
12
This
is because
the items measuring understandihg have less
representation in the total test score and there is typically
.a low correlation between measures of knowledge and measures
of unde-standihg.
LOW discrimination indices here merely
irldicate that these items are measuring something different_
from what the major part of the test is measuring;
Removing
such items from the test would make it a more homogenous
measure of knowledge outcomes;
but it would also damage the
content validity of the test because it would no longer measure
objectives in the understandihg area.
Since achievement test
batteries need to measure a wide variety of objectives in a
reasonably short period of time; they tend to be fairly
heterogeneous in nature and moderately
low discrimination indices
tend to be the rule rather than the exception.
To summarize, a low discrimination index alerts test users
to the possible presence of defects in test items but doeS not
cause them to discard these items if they appear to be
funetiOh-
ing as they should.
A well constructed achievement test will
-
of necessity; contain items with low discriminating power and
to discrard them would result in a test which is less;
rather
than more; valid;
Dthe to these considerations, in this
study
items were examined if they had diSeriminz:tion indices lower
thah .20.
This is a rather conservative criterion since items
that diSCtiminate as low as this may provide useful information;
but given the unknown test taking charactcristics of USVI
students; it was decided to be particularly cautious in the item
analysis.
13
11
Summary of Student Skills
The items on the tests of the various levels of the
Stanford AdhieVetent Test battery were keyeii3 tti behaviorally
stated instructional Objectives.
These objectives are
grouped into a series Of item groups..
Tables are aVailable which
present objectives within item groups and the difficulty and
discrimination
ildiceS for the item or items which evaluate
those objectives;
TheSe may be obtained from the Caribbean
Research Institute;
In addition, the difficulty index for
each item for the examinees in the Standardization group
is
available.
This standardization group consists of examinees who
were in the same grade at approximately the same time during the
school year as the U.S. Virgin Island8 sample;
The national p values are used not as a means
to compare
U-S. Vitain Islands students with mainland US
examinees.
Historital and cultural
differences ti-otweon these two groups
of examinees makes
this comparison an inappropriate one.
Philosophital tonsiderations
aside, hoticievet, such comparisons
are of little use to the people who make curricular decisions
in schools.
What these people need to knOW are the particular
levels of skills of students as
measured aqaitist well defined
objectives, not hOW well their students achieved thse skills
as compared to other students.
Nevertheless, since the skills and knowledges taught in
schools are seldoM taught once, but are dealt with at a
number
of grade levels where they are reinforced and broadened,
the
level of achievement= on
specific obje tives should_be expected
12
to change from grade to grade for a particular student or group
of students.
The publishers of the Stanford Achievement Test
take this factor into consideration by testing particular
objectives across a number of levels (grades) of the test;
this study the national p valuea are used to indicate the
relative level of difficulty of an item by which the performance
of the local sample nay be judged
For instance, it would be
foolish to be dissatistified if 20% of the USVI students have indicated
that they have reached an objective when only 18% of comparable
students in the standardization group had reached that same
objective.
What is more likely the case is that this is a
difficult and complex objective .t.hat had just been taught
recently and would be retaught and elarged upon at a later
time.
Thetefbte; the following criteria were used in summarizing
student skills.
Skills are described as "adequate" if the propor-
.tions of local examinees scoring correctly on items measuring
those skills are not significantly higher or lower than the
proportions of the standardization group scoring correctly or,
if significantly higher or lower; the proportions correct are
within 10% of the standardization group proportion correct as
reported by the test scoring service;
Skills are described
as "strong" if the proportions of local examinees scoring
correctly on items measuring those skills are significantly
higher and more than 10% greater than the standardization group
proportion correct as reported by the test scoring
are described, a8 "weak", if the proportion of 15-dal
examinees scoring cOrrectly'on items measuring those skills are
13
significantly 16Wet and more than 10% less than the standiza-
tion group proportion correct.
The 10% standard was set*in
the realization that some differences,
while statistically
significant may be educationally trivial and it was noted that
most differences indicated as significant exceeded 10 %.
Findlly, in cases where the scores of examinees from
both the St. Thomas/St; John and the St.
Cr6iX districts were
the same based OA the criteria stated above summaries were
based on the entire USVI sample; Where differences were noted,
the skills of e:-<atinee4 within
each district were summarized
separately
Avallabil_l_ity of Data
Tables listing each objective
on every test keyed to
test items with national diffidUlty
indices, difficulty
indices for the entire USVI sample and the individual district
subsamples as well as item discrithinatitin indices for the
whole sample and each subsample are available and may be
obtained frbt the Caribbean Research Institute,College of
the Virgin IslandS.
Grade 8 -_Advanced Level
The Advanced Level of the Stanford Achievement Test
Basic Battery consists of seven tests with the number of items
listed below:
1.
Vocabulary
50
items
2;
Reading Comprehension
-
74 items
3;
Mathematics concepts
35
items
4.
Mathematics Computation
-
45 items
5.
Mathematics Applications
-
40 items
6.
Spelling
==
60 items
7.
Language
=
70 items
The means; standard deviations; and reliability estimates
for theSe tests
determined previously
(see BlisS; 1982) are
presented below:
14
'Test
Mean
Raw Score
Standard
DeviatiOn
KR720_
Reliability
Standard
Error
Total USVI Sample_
Vocabulary
21.0
7;2
.81
3.15
Reading Compre-
hension
31.5
15.4
.95
3.45
Math Concepts
15.3
5;8
.74
'
2.65
Math Computation
23.0
7.6
.85
2.95
Math Application
16.9
6.6
.83
2.72
Spelling
31;8
12.3
.93
3.27
Language
35;6
12.1
.88
4;20
17
15
St.
Thomas/St.
John Sample
Vocabulary
20.7
.78
3.10
Reading Compre-
hension
32.6
17.4
.97
3.02
Math Concepts
16.4
6;0
.81
2.61
Math Computation
23;3
1.1
.83
2;92
Math Applications
17.7
6.6
;83
2;72
Spelling
32.8
11.8
;92
3;33
Language
35.8
10.6
;84
4.23
St.
Croix Sample
Vocabulary
21.2
8.5
.87
3.05
Reading Compre-
hdrIsion
30.5
13.1
.93
3.46
Math Concept
14.2
5.3
.75
2.67
Math Computation
22.8
8.2
.87
2.94
Math Applications
16.1
6;5
.82
2.76
Spelling
30;8
12.9
;93
3.40
Language
34;6
13.6
.91
4;07
Following are the item analyses and summa -ies of student
aChievement on the objectives evaluated by each of the tests;
16
Vocabulary
I_tet_ALnadvsis
All items showed acceptable discrimination indices with
the following exceptions:
Item #16 - This item was an easy. one (p---.72) for USVI examinees
regardless of overt.11 achievement;
Item #32
Thib item was particularly difficult for USVI students
(p=.16).
It Called for them to select a synonymn for
"poneer;"
Many atudehtt Chbad "beat" over the correct
responoe "think."
Item #39
This item was particularly difficult (1)=.17); but
there seemed *-o be no particular pattern of incorrect
responses;
Item #40
This item simply did not discriminate well between
high and low achievers; although difficulty
was moderate
(p=.37).
It required students to select a synonym for
"contemplative" and many examinees chose the distractor
"cautious" over the correct response "thoughtful."
Item #43 - This item diSdriMinated very poorly between high
and low achievers;
High adhievers often choose "dejected"
as a synonym for "exhilarated" rather than the correct
response "stimuIated;"
Item #47 - This item discriminated very poorly in spite of a
moderate difficulty
index (p -.29) and there appeared to
be no pattern of incorrect responses.
17
Item # 29 - ThiS item shOWed very little
discriminatibil With
high achievers equally likely to answer the item correctly
as were low achievers with a moderate difficulty indeX
(p=.52).
IteM # 34 = This item showed
low diSdkimination between high and
LOW achievers.
The difficulty index
(p=.31) was not partidUlarly
loW nor was there a noticeable pattern of incorrect responseS0
Item
#
1 = Thi8 item was very easy for USVI students (p=.96).
Item
#
30 - This item tended to be difficult for USVI students
(p=.21)
with no patterns observed in incorrect responses.
Item
#
31 - This item was a difficult one for USVI students
(p=.21).
Itet
#
44 - This item discritinated satisfactorily
for St. Crbix
School district student80 but poorly for St.
Thomas/St.' John
diStrict examinees.
These latter examinees
in the high
adhieving group chose "dangerbUt" very often as a synonym for
"anomolbus" over the correct response.
This pattern was not
evident among St. Croix school diStridt students.
Item # 41 - ThiS item discriminated satisfactorily
for St. Croix
district students, but did not discriminate at all among high
and low achievers in the St. Thomas/St. John district;
These
examinees tended to choose-"active" as a synonym for "audacious"
over the correct retpbnse;
"bold."
No such pattern was
noticeable among St. CrOix district students.
In general; this
;tended to be a difficUlt item for students in both St. Thomas/
St. John and St. Croix districts (p=;19 and
.21,respectively).
IteM # 19 - This moderately diffidtilt (p=.40) item tended to
diSdriMinate poorly among high and low achievers.
There appeared
to be no ambiguity in the
item and no observable pattern of
incorrect response.
1C
Summary of- Student S
1is
In general, USVI students tested seem to be weak in tht area
f standard EngliSh vocabulary recognition.
Most items seem to
function well with most poorly discriminating items being the result
of either extremely high or low levels of
difficulty;
Since few
of
e incorrect response patterns
(i.e. examinees choosing a
particular incorrect item with a high frequency) which typically
mark ambiguous items were presenti it is unlikely tAat these scores
were the result of technical difficulties in the items themselves.
Of the fifty items on the test; a
significantly smaller proportion
of USVI students were able to respond correctly when compared to the
standarditatiOn group on 39 of these items.
Beyond the question
of statistical Signigicance, the proportions of USVI students who
successfully answered many items was substantially lower than the
the standardilation sample.
The differences in item mean difficulty
values are a clear indication of this problem;
Of particular note is the single item on which the USVI
sample outperformed tne standardization
sample (but only in the
St. Thomas/St. John district)
This item asked examinees to
chocLe
the synonym for "capsized" by stating, "A capsized boat has been -.=="
-
With the correct response choice being "overturned."
Since such
an item may allude directly to personal experiences of USVI
students;
thiS is a phenomenon that bears examining in looking at other
items.
Reading Comprehension
Item Analysis
19
All items showed satisfactory discrimination indices except
the following:
Item #12 - High and Iow achieving
students seemed to be
equally likely to correctly answer this moderately
difficult (p=.30) question in the St. Croix district.
In the St. Thomas district; low achievers appeared more
likely to be able to correctly answer the item than were
high adhieverS.
High achieving students seemed to choose
distractor #4 very often.
The item requires examinees
to choose the most appropriate title for a.passage they
have read;
Distractor # 4 refers to a particular section
of the passage and uses names and ideas
in that section.
However; it does not refledt the overall idea of the
passsage.
Students choosing this distractor are not
aware of this overall
(or "global") ideaf.
IteM
# 54 = This item seemed to discriminate Satisfactorily
for examinees in the St; Croix distridt; bUt less well
for those in the St. Thomas/St; John district.
This was
'a fairly diffictilt question (p=.18) which required examinees
to choose the main idta in a short poem by Walt
Whit:Man.
A
large number of StUdentS from both districts seem to have
omitted this item which appeared toward the end of a 74 item
examination: ThisjwiI1 be 'commented on
in the summary. Item
#54; 60; and 74 display the
same phenomena as item #54,
showing many omits on what well to be difficult items;
Itet #3 = This appeared to be an easy item for most students
(p=.90)
IteM #57 = ThiS item appeared to
be a difficult one for students
in
22
20
both districts (p=.18).
However, it should be noted that
48% of students tested omitted this item and this `could account
for the low difficulty and discrimination indeces.
Item #
73
This item seemed to exhibit the same phenomenon as
item # 57
In this case 59% of students tested omitted this
item;
Item # 41 - In this fairly difficult item (p=.28) high achieving
students were distracted by choice # 4;
In this question
students read a paragraph about home mortgages and were
asked to choose a statement indicating the value of a mortgage
to a lender although this advantage was not explicitly stated.
Choice # 4 indicated that the value was that the lender
would "eventually own the property," which the paragraph
indicated was the consequence to default on the part of the
borrower.
The choice which was considered correct was
his
money earnsinteresti" and the fact that the borrower pays
interest was also explicitly stated in the passag.
Since
either of these situations could be considered as "value"
by a thoughtful examinee, it may be that the question was
defective in that it had two answers that could be considered
correct by students who lacked experience with mortgage.
Item # 72 - This item appeared to be very difficult for
USVI
examinees.
However, 59% of examinees omitted the item and thiS
could account for the low difficulty and discrimination
irdices
which were observed in the St. Thomas/St. John examinees
responses.
Item # 65, 67, 69 = TheSe three iters showed satisfactory discrimina-
tion in the St. Ckoix diSttitt, 'Alt poor discrimination in the
23
2
St. Thomas/St. John district.
One phenomenon that was
noted was that while the numbers of high achievers scoring
correctly on these items was fairly constant between districts
the number of lOW achievers scoring correctly WAS greater
in St;
'homas /Stt JOhh than in St.
Croix;
AIS6
the
proportion of omits was higher in the St; Thomas/St. John
district (63,64, and 66%; re;pectively vs 52;51; and 52%
on St. Croix) and this could ad-count for the differences
item discrimination power that was observed.
Item # 23
This moderately difficult
item 4.34) discriminates
little; if at all; between high and lOW achieving
.tideEts;
In general; LOW achieving students seem just as
likely to
obtain correct scores on this item as high achieving
students ;
There does; hOWeVer, seem to be differenceS in the patterns
of incorrect responses between these two groups of examinees;
The item asked Students who had read a passage which dealt
with animals supported by the British governMent which included
cat who was the "OLtidial Rat Catcher" in
the storeage
cellars of London's government offices;
The passage indicated
that the cats food cost the government the equivalent Of
per week while neglecting to give any information on
the cost
of upkeep for any of the other animals in the passage;
Lhe
question; "the writer probably mentions exactly what it casts
to feed Peta [the cat)
because he thinkS it is
---;" should
elicit "amusing" as the correct
response.
Most examinees
failed td See this as being amusingi with the low achievers
tending to select "too
and the higher achievers
tendifi4
22
to select "unfair" as their answers.
The latter seems to
be more of a moral judgement and it may be significant that
this was the choice of the higher achieving students.
It is
also interesting to note that the Technical _Data Re ort
published by the testmaker shdWs this item to discriminate
less well than most other items (although much better than on
the USVI sample of examinees) when the test was administered
to the standardization groups.
Item # 29
This Moderately difficult item (p=.36) discriminated
poorly between high and low achievers.
It required students
to infer something about the climate of Tropical Africa from
reading a passage concerning rivers in the region and their
great potential for the
production of hydroelectrioal power.
Most students concluded that in this region there was "sun-
shine most of the time" rather than the conclusion drawn
from the passage that there was "quite a bit of rainfall"
What may be happening is that examinees are draWing from
their own previously learned concept: about the nature of the
African climate and/or their personal experiences concerning
the climate in a tropical zone instead of using the informa-
tion provided in the written passage they were to read.
Items # 58, 61; 63; 66 - These items appeared to be very difficUlt
for USVI examinees (p=.15i .141
.17i \and .12i respectively).
Howeveri as was the case"in other items- owa d the end of
the test; the number of students omitting these items was very
high (52
56
59
and 63%i respectively in the St. Thomas/
St: John district and 49; 52
54! and 56% in the St. Croix
23
district).
These high proportions of omitted responses
could account for the low diffieulty and discrimination
indices which were obtained;
Summary of Student Skills
Table 2 shOWS the proportions
(in percents) of omitted
responses on each item in
the Reading Comprehensidn
test.
From
this table it is apparent that students in this sample did
not have the opportunity to attempt all of the items on the test
in the time allowed to them.
This is an important observation.
If students could not attempt items, they could not possibly score
correctly on them and this would
tend to deflate;the difficulty
indices of items toward the end of the test; making scores of USVI
students Seeth lower than they should actually be.
In addition,
since lesS than half the students were
able to. attempt items
60
through 74, we would expect lower discrithihation
indices than would
be obtained had all students been able to complete
the test. ThiS
phenomenon seems to be unique
to the Reading Cdthprehension test.
On all other tests
in the Advanced Level battery examinees
appear to have ample time to finish within the time provided by
the test publishers..Any attempt to explain why the USVI group
required more time than the
standardization groups 14-bad only be
speculation;
but further research should be done to investigate
the posSibilities that such factors as a more deliberate reading/
I
style (resulting in longer periods of time being required for
students to read each passage),
fatigue, a and differences
in
attention span might be contributing to this phenomenon.
Given theSe observad.Oni it might be well to be 6icii-oti(JE
24
TabIe 2
Proportions of the Eighth Grade Sample
Omitting Reading Comprehension Test Items
(Percents)
ITEM#
USVI
STTIST3
STX
0
3
4
1
1
C
5
1
1
6
0
0
0
1
8
9
0
o
10
1
11
1
2
12
13
0
1
2
14
1
1
15
1
16
1
17
0
18
1
1
0
19
1
2
"--,
'20
21
1
27
Table 2 Continued
22
1
23
6
1
24
1
2
U
25
5
3
26
5
2
27
4
28
4
5
29-
-7-
4
5
30
5
5
5
31
6
5
6
32
8
8
8
33
8
8
9
34
7
6
8
35
8
8
8
36
9
7
11
37
11
9
13
38
13
11
14
39
14
.
13
15
40
14
12
15
41
15
14
17
42
18
17
19
43
19
19
19
44
20
20
20
45
21
20
21
46
'23
23
23
47
29.
31
27
48
30
31
28
25
Table 2 Continued
34
38
38
30
30
11
49
50
51
32
34
35
52
39
44
34
53
41
45
37
54
42
45
40
55
44
48
40
56
47
49
49
57
48
52
44
58
49
52
46
59
50
-
53
48
60
51
55
48
61
52
56
47
62
53
58
48
63
54
59
56
64
55
60
51
65
57
63
52
66
56
63
50
67
58
64
51
68
57
64
51
69
59
66
52
70
59
6.7
.51
71
58
65
52
72
59
63
54
73
59
64
59
74
58
63
53
26
27
.drawing any conclusions from the results of items 50-74 where more
than a third of the examinees omitted the items in question.
Re8tridting the observations to items 1 thrOugh 49, however,
the fbllowing generalizations can be made;
USVI students tested seem to be somewhat weak in the area of
determining the general idea of the
passages they read, although
this must be a somewhat tenuous judgement based upon the small
number of items (five) from whidh thi8
judgement could be made
because of the large number of examinees
omitting later items of
the test.
However, in three of these fiVe items, local students
had a significantly smaller proportion of correct answers than did
the students in the standardization groups.
In the area of determining meanings explicitly stated in a
reading passage, USVI students
appeared to have achieved well.
In
most of the items which most students seem to have had the opportunity
to answer, the proportidn of local students answering correctly
was as high asor only slightly lower than the standardization
groups proportions correct.
Implicit meaning
refers to ideas in a piece of reading which
are not specifically in the reading, but must be deduced from facts
USVI examinees
ear to have difficulty in determining explict meanings
in
pa sages they have read.
In ten out of thirteen items dealing
with
is skill that most students were able
to attempt to answer
in the time allowed, the students
in the local sample scored
correctly a significantly lower proportion, f the time than
standardization sample,
Further, the mea
proportion of students
or other information presented in the test.
The
28
scoring correctly on these itemE was particularly low (59% vs:70%
for the standardization group);
Often a reader must obtain information from a text by obtain-
ing clues from what is explicitly stated within the text.
The
USVI examinees had great difficulty in determining meaning from
written passages on the basis of contextual clues.
In seven out
the nine items that most examinees were able to attempt within
the time allowed, the proportion of local students answering
items
correctly was significantly lower than the proportion of examinees
in the standardization sample answering the -same items correctly.
Again, these proportions were particularly lbw with a mean of 38%
for tt local examinees versus a mean of.52
for the standardization
group.
The
students in the USVI sample had great difficulty
in
tasks which required them to make inferences or draw conclusions
from passages they had read.
The proportion of USVI examinees
answering items correctly was significantly lower than
the proportion
of ,examinees in th'e standardization sample who answered the same
items correctly for all of the eleven item; which most examinees
in the USVI sample were able to attempt'in the alloted time;
Determining inferential meaning appeared tc, be the reading com--
prehension skill which gave USVI students the most
difficulty with
a mean correct proportion on the above mentioned eleven items of
28% versus 46% for the standardization group.
An observation that may be significant. is that USVI students
.showed strength in the area of determing explicit meaning from
material which they read while showing weaknesses in aohivement
31
29
in other areas of reading-comprehension.
The taxonomy of
educaticnal objectives developed by Bloom (1956) may be a good
theoretical framework from which to view this phenomenon.
Being
able to derive meanings that are explicit in readings inVOlVeS
objectives at the comprehension leVel of the
taxonomy while
the other objectives types (deriving global meening; implicit
meaning, inferential meaning, and meaning
frbt context) involve
higher level (and, therefore, more diffidUlt) objectives in the
area of analysis and synthesis.
It appears that the sample of
USVI students may have, as a group,
achieved reading comprehen-
sion objectives at the comprehension level, but have not, as yet,
fully mastered many of the higher level objectives in thi8 skill
area.
Mathamatics_Concepts
Item_Analysis
All items discriminated s4tisfadtbrily except the
Itet
6 - This was-a)iparticularly diffidUlt item
for USVI
students (P=-15)-
One distractor operated tj cause many
students to choose it over the correct response.
Specifically,
the item asked students to complete the aritAmetid sentence:
6+12=.
Most students, regardless of general leVel of athieve-
ment, chose "2" as the correct answer
rather tha
11".
This
indicates that the item was functioning properly, but that
examinees had not mastered the objective being tested.
32
30
Item #
4 -
This proved to be a relatively easy item for examinees
(p=.75); but it did not discriminate satisfactorily in
the St. Thomas/St. John district.
This was primarily because
low achieving-stude
s were almost as likely to answer correctly
as were high achiev_
students.
Item # 16 -
On this moderately difficult item (p=.36) high achieving
students were not much more likely to answer correctly than
were low achieving students;
There appeared to be no pattern
observable in the incorrect responses;
Item # 29 -
In this moderately difficult item (p=.35) high achieving
students were not much more likely to answer correctly tha..
low achieving students.
The item asked students to determine
the effect on a factor if the quotient remained constant and
the other factor decreased.
Many students from both achieve-
ment level groups chose the response that indicated that
the factor decreased rather than the correct response indica-
ting that it would increase.
The item appears to be operating
properlyi howeveri since thij; error is a common one and
indicates a lack of understanding of the concept.
Item # 33 -
On this difficult item (p=.27) most students in both
achievement level groups scored incoriectiy.
The lack of
discrimination was very likely due to the difficulty of the
item rather than to any technical fault in its writing..
No
response patterns were observed in the incorrect answers.
Item #
35 -
This was an extremely difficult item for all students
(p=.22) resulting in a low level of discrimination..
There
33
31
appeared to be no pattern of incorrect responses.
Summary of Student- Skills
Number
The itet grouping "Number" refers to the concepts which
deal with the relationships betWeen different types of numbers
(e.g. primes;
oddS vs evens;
fractions vs whole).
USVI students
showed no particular strengths in this area
Concept areas that
appeared to be particularly weak were:
A)
the concept of fraction and its relationship to other
types of numbers.
b)
the' concept Of odd and even numbers
c)
the concept of Signed numbers:
d)
the concept of "modulo"
Notation
"Notation" refers to the technique by which numbers are
expressed in written
form;
SUCh forms include expanded notation,
differing base Values, decimals;
and fractions.
The students in
the USVI.sampIe
seemed to have
a strong grasp of
the.concopt
of exponential notation. They seemed to be weakest in the following
areas:
A)
.use of expanded notation
b)
decimal notation
c)
translating arable numbers into word f)rm
Operations
The term "operations" refers to the manipulation of numbers
based on the use of a set of previously defined fanctions (e.g.
addition; subtraction;
multiplication).
The students in the local
32
sample Showed strenghts in the area of multiplication of decimal
fractiont, but seemed to be weak in the following areas:
a)
solving inequalities
b)
identifying multiplicative identity
c)
identifying the uses of the associative property
d)
the use of sighed numbers in multiplication and division
e)
identifying formulas fbr qualities of geometric functions
.f)
estimation in multiplication and division
Geometry and measurement
This appeared to be the atea,bf math
concepts in which the
USVI sample exhibited strongest levels of achievement;
Examinees
performed well in the following areas:
a)
metric conversion
b)
the use of Venn diagrams
c)
graphing
Areas in which weaknesses were observed were:
a)
set theory
b)
estimation of angle sizes
c)
properties of geometric figures
d)
probability
Mathematics Computation
Item Analysis
All items discriminated satisfadtbrily with the exception of
the following.
J i5
33
Item # 36
This item appeared to be vary difficult for examinees
in the USVI sample (p=.10).
The question required students
to find the average of 12 and 6.
In the low achieving
group there appeared to be no pattern of incorrect responses,
but high achievers were noted to choose the response
"18" a
large proportion of the time. This response would occur if
examinees summed to two numbers; but failed to follow through
on the second part of the averaging operation and did not
diVide.
Item # 45 -
This Appeared to be a difficult item for the examinees
in the local sample (p=.24).
It should be noted that the
item was also difficult and discriminated at low levels
(d=;29) for the standardization sample.
An examination of
the item; however; indicates that it is quite clearly worded
and the task seems to be simply presented.
The lack of
diScritiliriatinq power is most likely due to the
general
difficulty of the item.
Summary of Student SkIlla
Knowledge of Primary Facts and Solution of
Simple Mathematical Sentences
On these skills there appears to be some substantial differehdeS
between the achievement leVelS of students in the
St; Thomas/St.Johh
and the St; Croix district.
Because of this it would be useful to
summarize achievement charAdteriStieS inch district; individually.
In the St. Thomas/St.John district students showed strong
skills in the following areas.:
a)
use of the associative property of multiplication
36
34
b)
multiplication and diViSibn with integers; fractions;
and mixed numbers
c)
addition; subtraction; and renaming of fractions
d)
use of the distributive
property
Weaknesses were observed in these areas:
A)
multiplication with carry facts
b)
operations with negative terms
c)
renaming exponential numerals
In the St. Crbix district
examinees exhibited strong achieve-
ment in these areas:
a)
multiplication and diVisibti of integers; fractions and
mixed numberS
b)
use of the distributive property
d)
appropriate use of the distributiVe property
Weaknesses in the following areas were obsetVed:
a)
use of the associative property in multiplication
b)
operations involving negative terms
multiplidatic)n with carry facts
Addition and Subtraction_A1orithms
Objectives in thit item grouping
deal with the mechanics
of addition and substraction.
Students in ti.e USVI sample showed
good achievement in this area on
all objectives=
Multipl_icatiah and Division Algorithms
This item grouping tested objectives dealing with the
mechanics of multiplication and diviSiOn.
Students showed good
achieVement levels on all objectives
in thiS area with the exception
i
of the ability to find the quotient in a division problem with a
35
two digit diVisor and a zero in the ones column of the quotient.
majority of students missing the item which tested thiS
objective indicated "NH" (not tiOre) as their choice leading the
researchers to suspect that they came up with an answer through
faulty use of the algorithm;
-rnmiTnn FractiOti
St. Thomas/St. John district studentt showed particular
strengthS in finding common denominators and in working problems
where common denominators must be found.
They achieved adequately
on the objective of finding the common fraction of a whole
number;
However, they appeared weak in the area of fraction
multiplication.
St; Croix dittriCt students appear to haVe achieved
adequately in the area of fraction multiplication, but were weak
an_objectives which required them to find and work with common
denominators;
Students in both schbol districts appeared to have
particular difficulty in the area of division by iiaC16ns;
Other nparatiOnal Models
All USVI students achieved strongly in the objectives
involving exponential notation and prime factbrization.
Their
achievement seemed adequate on objectives dealing with
decimal
notation and number lines.
Weak areas were On objectives having
to do with finding averages and working with percentages.
St. Thomas/St. John district studentS exhibited good
achievement levels in 'all other areas.
St. Croix students showed good achievement levels in other
aleas with the exception of objectives dealing with common and
simpae algebraic manipulation.
Mathematics applications
Item_Analvs1s
All items discriminated adequately except the following.
Item #
11 - This was a particularly difficult item for USVI
students (p=.20).
High achieving students were only slightly
more likely to score correctly on this item than were low
achieving students.
There appeared to be no pattern in the
distribution of incorrect answers, except that choice number
4
(which indicated that a shaded area of a square took up
the entire square when it obviously did not) was chosen very
seldomly.
Item # 35 - This appeared to be a difficult item for the local
sample students (p=.29);
It was the only problem dealing
with metric units on the test.
Howeveri the metric equiva-
*
lents were given in the stem of the item; so a student,
knowing nothing of the metric system could have determined
the correct answer.
A large number of students missing
this item either left in blank or indicated
the
choice
"NH" (not here);
it is possible
thz.t these students were
intimidated by the metric units and did not even attempt the
item and that this phenomenon carried through both high and low
39
37
achieving students
Item # 37 = This
item discriminated poorly.
It tended to be a
difficult item (p=;21)
and asked students to determine
the sum of six angles marked off in triangle with
a perpehdicUlar drawn from one Of the angles: to the
opposite Side.
The student seledtitig the proper
answer needed ,to put together the fadt8
that -the sum
of the angles Of a triangle is 1800
and that a
perpendicular drawn to a side produces two angles of
90°
.
The incorrect choice chosen most frequently
was
"180°" which indicated that examinees may have
been aware of this first fact, but were unaware of
the second or were unable to use the
two facts together;
Item # 40 - This was a
very difficult item for USVI
students
(p=;14).
Thi8 less than chance ptcportion of students
scoring correctly on this item and the large proportion
of omits (20%) evokes the possibility that; although
it is partof the OUrriculumi the objective was
never really taught in the scho61s, and/or that being
the last item on the test; a substential number of
Students never had the chance to attempt to answer its
Summary- of
Skills
Anal sis and Dr.velomeht of a Solttion
Design, Selection-of a Solution SentPnce.
andAdequacT of Data
St. Thomag/St.
John district examinees exhibited
adequate skills in solving problems which required them to:
40
a)
divide to determine shares
b)
add and divide in multistep sequences
F)
translate verbal problems into mathematical
sentences
d)
determine correct change in a two step sequence
They displayed weaknesses on items requiring them to:
a)
determine the adequacy of given data
for solving
a problem
b)
express verbally the process for solving a
problem
c)
estimate the ratio of two'numbers greater
than
1,000
St; Croix examinees showed adequate achievement on
none of the objectives in thit item grouping.
They showed
particular weaknesses on items requiring them to:
a)
divide to deterMine shares
b)
translate verbal problems into mathematicai
sentences
.
c)
determine the adequacy of given data for solving
a problem
d)
determine correct change in a
two step sequence-
e)
express verbally the process for solving a
problem
f)
estimate the ratio of two numbers greater than
1,000
Date, Scale; and Percent
The USVI sample showed adequate achievement.only on
41
39
the objective of deterMining a Specific rate when one rate is
expresses as an average;
Particularly marked weaknesses were
fOthd on items which required the examinee to:
a)
express part of a total as a fradtion
b)
determine specific rate when no average rate.is
given
c)
interpret map scales
d)
find a total When a part and its fractional
representation are known
e)
determine percents
Measurement
Examinees in the USVt sampledetOnstrated adequate
adhieVetent in solving problems whiCh required them to:
a)
find the volume of a rectangular solid
b)
find the area of a rectangle
c)
find the sum of six angles
They showed weaktie88e8- it solving problems Which required them
to:
a)
calCUlate the area of a shaded region
b)
convert Standard masures of weight in order to
divide
,
c)
add negative numbers
_
d)
match EngliSh unit with metric approximations
e)
find and compare area
of squares'and parallelograms
f)
identify finishing time when given starting and
elapsed times
40
Students in the USVI sample demonstrated adequate achieve
ment in identifying a function in a mathematical
sentence With
two variables and choosing the solution set for the sentence.
They showed Weaknettet in items dealing with the following
objectives:
a)
choosing the graph of a solution set
b)
using line graphs
Statistics; Avera and Probabilit
St. Thomas/St; John district students demonstrated
ado irate.
achievement in the ability to compute an average.
They showed
great weakness in their ability 4o determine the probability
of the occurance of a given event;
_
St. CrOix examinees exhibited weaknesses across both
of these objectives.
Spelling
temAnalysis
All items discriminated at a satisfactory level except
the
following;
Item # 10 - This was a very difficult item. for USVI students
(p=;18);
Both high and low scoring studentt appeared to
find it equally difficult;
Summary of Student Skills
Homophones
HbitibphOhet are words which sound alike
but have different
meanings
and
are
spelled
differently;
On
these
items,
43
41
student8 were presented with four homophones used in context
and were requited to cnoose the phrase in which the homophone
was used inappropriately.,
StudehtS in, the S
Thomas/St; John distriet showed
particularly strong ability to recognize the distinctibn between
the'homophones "choral" and "coral" when used in context.
They
showed particular weakneases in distinguishing missuse of the
following homophones in context:
a)
cereal/serial
b)
straight/strait
c)
miner/minor
S
d)
arc/ark
Oh other homograph recognition pairs these students
performed at an adequate level;
St. CroiX district students showed particular weaknesses
in distinguishing missuse of the following homophones in context:
a)
cereal/serial
b)
bored/board
el
wrap/rap
d)
straight/strait
e)
miner/minor
They showed adequate achievement in dealing with all
other homophoneS in context.
Phonics
The students in the USVI sample showed a particularly
strong grasp Of the use of phonics in spelling.
Specific
objectives where local students achieved very highly were where
42
misspelled wordS had situations where
a) /1e/was spelled e-e instead of ee
b)
final /e/ was spelled ey instead of y
c)
schwa sound was omitted
Objectives on which local students showed weaker
achievement were tested by items where the misspelled words had
situations where
a)
/k/ was spelled c instead of ch
b)
/kt/ was spelled ck instead of ct
c)
/e/ was spelled i instead of
d)
/e/ was spelled e instead of ea
e)
/ch/ was spelled d instead of t
f)
/e/ was spelled e instead of a
It might be useful to note here that difficulties
appeared to be of two' types.
FirSt, in the use Of vowels,
particularly of those having the sound /e/.
This sound is
represented in a great number of ways in the English language,
ver. y ofte
quite arbitrarily;
As a result, it would be difficult
to apply phonetic rules to
its use;
Second, in words that are
generally mispronounced by the population at large.
We tend to
say "strickIy" rather than "strictly" and "congradulatione
instead of "congratulations."
Again; applying phonetic ruleS
to such words could cause difficulties.
Word Building
Word building skills refer to those which are involved
in adding suffixes and prefixes to
root words.
USVI students
showed very'satisfactory achievement
in this area ..
The few weak
43
areas if achievement are those in.WhIch students were required
to identify misspelled words where
A)
/ens/ is spelled ense instead of ence
b)
a final 1 is omitted when adding ly
Laaa.6.8a-s
Item_ Analysis
All it
discriminated at a satisfactory level except the
following.
Item
#
16 - This item was an extremely difficult item for USVI
students (p=.08).
Low achieving students were just as
likely to score correctly on it as high achieving,studetns.
Incorrect answers were spread in no particular pattern
except that the incorrect answer was chosen lest often
than could be expected by chance
(there were four choices).
Only 2% of students tested omitted the item;
However, it
was noted that high achieving students responded in a
particular way.
The item required students to fill in the
blanks in the sentence,
"Beddbet goon saw that Humphrey was
as good as
in fact, Humphrey was the --- of the two."
The choices were 1) he-best,
2) him-better; 3) he-better;
and 4) him-best; with choice #
3 litted as the correct
answer.
High achieving students tended to choose response
1 and 4 over the other two responses,
indicating that they
believed the Superlative form of "good" was the corrcet
response.
cooper
(personal communication)
points
4s'
Item #
out that persons who are
considered to be exemplars
of the use of formal standard Englith have adopted
thit usage.
He suggests that it will beCOMO the
fottally accepted usage within ten years.
Given this
situation, the item may have been ambiguous since it
could be thought of as having two correct answers.
- This item proved to be very difficult for students
in the VSVI sample (p=.16).
It tested students'
knowledge of both the capitalization conventions for
school subjects and the use of commas to offset words
in a series;
Most studentb Chete choices # 2 (Latin,
Greek, Math and Science) and 4 3 (latin, greek, math
and science) with the farmer indicating they were not
aware of the capitalization convention and the latter
that they were unaware of both the capitalization
convention and the comma usage.
Choice #2 was, by far;
the most popular.
High achieving student§ who missed
the item chose 4 2 in almost every
instance.
Item # 61 - This was a difficult item for examinees
in the local
sample (p=.23).
There appeared to be no pattern of
incorrect responses either for
the high achieving or
low achieving group.
The item gave a dictionary entry
for the word "versatile"
including a pronounciation
guide.
Examinees had to be able to determine form
the pronounciation guide that the last syllable of the
word contained no vowel sound.
In writing, and often
in common pronounciation the word 116.8 an "i" in the
45
final syllablet but the letter is not sounded;
What
may have contributed to the difficulty here
that
students attempted to answer the item without using
the pronounciatiOn guide as thdy were instructed to
do and thust tended to rely on their own pronountiation
of the word be-Cat:Se they could or would not make use
of the pronounciation guide.
Summary of Student Skil s
Capitalization Conventions-
in the St.Thomas/St; John school district students
seemed to have a good grasp of capitalization conventions tested
except those used when naming school subjects.
In this latter
case they tended to capitalize the names of
11 school subjettS.
Students in the St; Croix distritt had difficulty in
dealing with the following capitalization conventions:
a)
"Sir" as an honorific
(e;q. Sit Walter Raleigh)
b)
the name of a play
c)
school subjects
d)
names of religions
e)
names of institutions and cities
Oh other capitaIizatitin conventions tested, these
students appear to have achieved Adequately;
Punctuation Conventions
Students in the St;
thomaS/St. Croix school diStrict
seem to have difficulty in the use of commas in
certain Situations;
They seem to be
adequately familiar with comma'usage under the
following circumstances:
48
4
a)
in dates
b)
after an introductory participial phrase
c)
in setting off appositive phrases
d)
in Setting off direct quotations
e)
after introductory adverb clauses
f)
before "and" when it joins two .independent tlaUSOS.
g)
between adjectives modifying the sate noun
h)
after introductory prepositional phrases
These students do seem to have difficulty in the use of commas
under these circumstances.
a)
setting off a parenthetical expression
b)
when they are not used to se
off indirect
quotations
c)
Setting off direct quotationfi
d)
setting off
nonessential cli,uses
In the St. Croix Sdhool district examinees appeared
to be adequately famil_
'7 with the use of commas in these situations:
a)
in dates
b)
in setting off appositive phrases
c)
after introductory adverb,cliuses
d)
when they are not used to se
off indirect
quotations
e)
before "and" when it joins tlo independent.clauses
These students seemed to show weaknesses it
the use of
commas
under the following circumstances:
a)
after an introductory participial phrase
b)
Setting off a parenthetical Expression
)
separating items in a series;
19
di
separating adjectives modifying the same noun
e)
setting off direct quotations
f)
setting off
nonessential clauses
g)
after introductory prepositional phrases
47
Usage Conventions
Usage conventions refers to the choice of particular
forms of a word which are appropriate in a given context.
Eamples
of thiS indltde the formation of past and present participles
and the choide of appropriate verb tense.
StUdentS in the USVI sample seemed to have a good
grasp on the fOrtation and appropriate use of the past participle
of irregular verbS, but showed some weakneSSet in determining the
past participles of the verbs, "to write," "to begin," and "to
wear."
The latter verb only appeared to be troublesome to
St. Thomas/St. John district students;
,
Examinees in both dittridtS showed weaknesses
in
achievement on the following objectives:
a)
choosing a verb form to
avoid double negatives
b)
choosing reflexive pronouns
c)
choosing the appropriate verb when singular
subjects are joined by "nor."
The three item groupings noted above (i.e;
capitalization;
punctuation, and usage conventions) form Patt:A of the language
test and constitute
the-first 31 items of the test.
This section
is timed separately and it was noticed that after item #25,
over 15% of the examinees tended to omit items.
Twenty one percent
Of
the examinees omitted item # 31.The items
in thiS section
50
48
of the test require examinees to carefully read one or two
sentence passage and choose 'the
pprOptiate words to fill one
or two blanks in the sentence or
sentences.
It was noted in
the Reading Comprehension test that local studentS appear to
take more time to. read material presented to them with the result
that they have difficUlty getting to the later items on tests
requiring much reading.
It is possible that the same phenomenon
is occurring on this section of the Language test,
alti;ough on
a smaller scale; with the result that a substantial number of
examinees are not able to attempt the last few items on the test.
This would have the effect of deflating difficulty indices and
Making the achievement of studentS in the
local sample appear to
be lower on the objectives tested by the-Se items, and on the test
as Whble.
It is suggested that the resUltb obtained on these
later-items be viewed with caution when interpretations are made.
Distinguishing_Fra_gments
Sentences, and Run,0nSentences
Students in the St, Thomas/St;
John school district
appear to have little difficulty recognizing complete sentences
and sentence fragments.
On the Other hand; they do seem to be
weak in recognizing a run -on sentence when it is PieSeiited to
them.
Students in the St. CrOiX school district appear to be
able to recognize sentence
fragments and run -on sentences,
but
Seem to have difficulty recognizing complete sentences;
English Skills
This section of the test examir.es student achievement
51
49
in areas such as the use of dictionaries and other reference
works, and in basic literary usage.
In the area of dictionary skills, students in the
USVI sample shOWed adequate skills in using dictionary entries
to determine the appropriate meanings of words used'in context.
They seemed to be weak in other areas of dittionary use such as
use of the
pronoundiatiOn guides and use of guide words;
Students in bOth districts had diffibUlty identifYing
the Reader's _Cuitle to Periodical L terature as a source for
finding the titles of magazine articles when the author was
known and St. Croix district students exhibited difficulty in
identifying an atlas as the reference one would use
to find
maps.
thd St. Thomas/St. John district students achieved
adequately on the following
objectives:
a)
recognizing a word formed by a prefix or suffix
b)
identifying meanings of morphemes
c)
selecting the modal in the present tense
d)
choosing sentences which have identical meanings
These students showed lOW achievement on these objectives:
a)
identifying literary concepts
b)
recognizing determiners that are used incorrectly
C)
indicating reasons why given sentences
are
ungrammatical
d)
identifying irrelevant words in sentences.
In'the St. Croix district students achieved adequately
on the following objectives:
a)
identifying meanings of morphemes
50
These students showed weaknesses in achievement on the
other objectives listed above;
51
Grade 10 - TASK I Level
The Stanford Achievement TeSt - Test of Academic Skillsi
1,6V-el I consists of three
tests with the number of items
listed beloW:
1.
Reading
78 items
2.
EngliSh
69
items
3.
Mathematics
48 items
The means; standard deviatiOns
and the reliability
estimates obtained for raw scores on these tests; as previously
reported (see Bliss;
1982), are presented below
Test
Mean
Raw Score
Standard
KR-20
Deviatibn
Reliability
Standard
Error
Total USVI Sample
Reading
45.6
14.0
.93
3.70 s'_
,.)
English
32.0
14.6
.98
2.07
Mathematics
48.0
12.0
.92
3.38
St. Thomas St. John Sample
Reading
43.6
14.2
..54
3.48
English
32.i
.6.9
.99
1.69
Mathematics'
47.6
10.8
.90
3.41
St. Croix Sample
Reading
48.5::
14.3
.94
3.50
English
-31.5
8.6
.89
2.85
MathematiCS'
49.0
14.0
.95
3.70
54
Following are the item analyses and summaries of student
achievement on the objectives
evaluated by each of the tests.
1
53
R'gAdia
Item Analysis
Most items seem to discriminate at a satisfactory level.
The exception
to this statement will
be examined shortly.
Before doin4 this; however; it will be useful-to note that; as
in
(
the Readinq\Comprehension test on the Advanced Level given to
eighth grade Stwients;.large:numbers of students seem to have
been unable to attempt all the items on the test in the time
allowed.
Table 3 indicates the proportion of omits on each item
in the Reading test.
One difference from the eighth grade
sample which is readily apparent is that the proportion of students
omitting items in the St.
Thomas/St. John district
i8 considerably
higher than that for the Students in the St; Croix district.
Also, the difficulty indidet on items where a, great number of
omits are present seem to be higher for St. Croix
studentt (i.e.
more St. Croix students scored correctly on these items); whereas
the proportions of students scoring correctly on items where
there few omits do not differ greatly between districts;
From thiS it is clear that students in the St.
Croix district
are more likely to be able to attempt to answer later items on
the Reading test than are students in the St. ThOMaS/St. John
di-EiiCt and that thiS phenomenn results in generally higher
scores on this test fbt St. Croix students than for St. Thomas/
St. John students (mean raw scores are 48.5 and 43.6, respectively).
54
Reading
Item Analysis-
Table 10 presents the difficulty and disetiMination indices
for items on the Reading test.
Most items seem to discriminate
at a satisfactory level.
The exceptions to this statement will
be examined shortly.
Before doing this; however; it will be
useful to note that; as in the Readitig Comprehension
test on the
Advanced Level given to eighth grade students;
large numbers of
students seem to have been unable to attempt all the items on the
test in the time allowed.
Table 11 indicates the proportion of
omits on each item in the Reading test.
One difference from
the eighth grade sample which is readily tppatorit is that the
proportion of students omitting items
in the St. Thomas/St. John
distriet it considerably higher than
that for the Students in the
St. CrbiX ditttiet., Also; the difficulty indidet on items
where a great number of omits
are present seem to be higher
for St. Croix students (i.e. more St;
Croix students scored
correctly on these items); whereas the proportions of students
scoring correctly on items where there we--e few omits do not differ
greatly between districts.
From this it is clear that studehtS
the St. Croix district
are more likely to be able to attempt to Anter later items on
the Reading test than are students
in the St. Thomas /St. John
diStriet and that this phenomenon results
in generally higher
scores on this test for St.
Croix students than fbt St.
Thomas/
St; Jc- students (mean raw scores are
48.5 and 43.6, retpectively).
57
Table 3
Proportions of the Tenth Grade Sample
Omitting Reading Test Items
(PercentS)
55
ITEM #
US VI
STT/STJ
STX
1
0
4
1
1
1
5
0
6
4
3
2
1
10
11
12
13
4
14
1
0
15
1
1
2
16
1
0
17
2
2
18
2
19
1
20
Table 3
Continued
21
2
2
0
22
0
23
24
2
25
6
3
26
5
7
1
27
6
8
2
28
8
II
29
14
18
30
14
18
6
31
15
21
32
17
23
5
33
20
27
8
34
21
29
8
37
31
41
14
38
30
41
10
39
32
42
14
40
38
47
20
41
37
48
16
42
39
50
18
43
27
35
11
44
28
35
13
45
30
38
14
46
35
43
20
47
35
44
21
48
37
46
21
49
37
46
21
56
57
Table
-5
Continued
50;
39
48
23
51
41
49
26
52
5
7
1
53
9
11
6
54
8
10
6
55
8
10
5
56
9
10
9
57
14
14
'
15
58
9
9
8
59
14
14
14
60
14
14
14
61
18
19
16
62
13
14
13
63
18
19
16
64
20
21
17
65
17
19
14
66
20
21
15
68
19
19
18
69
21
22
20
70
20
22
16
71
23
25
20
72
24
26.
21
73
25
28
20
74
-24
27
20
75
28
31
58
22
76
30
34
22
77
30
34
22
78
33
38
24
Any attempt to explain these differences would be speculative
at this time; but there are some pOSSible explanations whidh
come to mind.
The most obvious is
that St.' Croix district
students who took thi8 test are more
generally academically
competent than their St. Thomas/St. John counterparts.
This
seem Unlii:ely since other tests in the TASK battery do not
exhibit this difference;
Proportions of omits and raw score
means are very similar (see Blibp;
1982);
Another possibility is that St, Croix counselors gave students
time to finith the test and this violated the standard diteCtions
of the test.
There is some evidence that this may have been
the case;
The examinees who were supervised by one counselbt
on St. Croix consistently left out fewer items toward the end
Of the examination than did the studentS supervised by the other
counselor.
In fact; the proportion; of students omitting iterit
in the latter class was approximately the same
(
or. slightly
higher) than the proportion Omitted by the entire St.
Thomas/
St. John group of examinees.
WhateVet the explanation;
dedision
makers Would do well to
view the results of the Reading test as somewhat tentative;
particularly on items where more than 1/3 of the students omitted
the item,
The following discussiOn will limit itself to those
items which were answered by 2/3
or more of the examinees.
All items showed satisfactory diterimination indices except
the following:
60
Items # 1, 3, 4, 22 - These proved to be very easy ite:is for
)
USVI students
(p..931 .95i .93, .94, reSpedtively), and
low achieving students were very likely to score correctly
on these items;
Item # 64
This item seemed to discriminate satisfactorily among
St. Croix district examinees but poorly among St.
Thomas/
St. John examinee-S.
The item required students to match
the word "aspire" to the
one Word among five that was
most closely related to it.
This was a moderately difficult
item for St. Thomas/St. John students (p==.26) and for
some reason low achievers were just at likely to get the
item correct in this district as were high achievers.
No
patterns of incorrect answers could be observed.
1111
PO- t Skills
Glabal_Meaming
Students in the USVI sample showed weaknesses in
their
_ability to determine the main idea of a reading passage.
In
the St. ThomaS/St. John district the proportion of students
_
correctly responding
to the fOtt items with omit proportions
under 33%
was significantly and considerably lower than the
proportions of the standardizatidn group scoring correctly on
all items (mean p for the St. Thomas/St. John group was
.41
versus .55 for the standardization 4roup).
In the St; Croix district the proportion of students
correctly responding to the six items -leasuring global meaning
was significantly and considerably lower than the standardization
63
61
group in 4 of the items;
The mean p for the St. Croix group
was .44 versus .56 for the standardiZation group.
Explicit Detail
The USVI sample of examinees appearg 'to gave achieved well
in the .area of detertining explicit details from reading pas-
sages. Ox a third of the items (six out of eighteen) the proportion
of local examinees scoring correctly was significantly higher
than that of the StandardizatiOn group.
On only four of the
eighteen items did local.
stUdent8 score significantly below the
standardization group.
Implicit Detail
IMplidit meaning refers
to'inforMation which is present
in a body of text but is not explicitly Stated..
Students in the St. Thomas/St. John diStridt seem to do
well on items requiring them to determine implicit details from
a reading passage.
Exceptions to this rule seem to be possibly
related to the nature Of the
reading passage.
For instance they
tended to do poorly (p=.36 and .45 vs 60
and .59 for the
Standardization group) on items
from a very long passage on
sail
boat safety.
It is possible that reading passage length may
be a faCtOr here since other paSSages were considerably shorter.
St. CrOix district examinees tended to score similarly to
St. ThomaS/St. John examinees with the exception cf the fact
that they seemed to have less trouble with one of the two items
listed above where 52% of the sample scored correctly compared
to 59% of the standardization group and only
45% of the St; Thomas/
64
62
St. John sample.
Inference and Logical Analysis
This section of the test evaluated students'
abilities
to draw logical conolusiOnt ftbit materials they have read
and
to make inferences concerning SituatiOn8 beyond the content
of the rradings; Students in the USVI sample appear
to be
adequately proficient in these skills with a Mean group p
value of ;58 versus ;55 for the standarizatiOn group.
Meanlng_from_Context
These items appeared at the end of Part A Of the test.
BedauSe Of this, the proportions of omitted responses
were
extremely high indicating that many students did not have
sufficient time to attempt to respond to these items (the
mean proportion of omits for the entire sample on these items
was over 34%).
Thit phenomenon makes any attempt to interpret
tnis data extremely suspect and,
-consequently, no attempt will
bE made here;
Word Meaning
This portion of the test evaluated standard English
vocabulary.
In genera
,
the USVI sample of examinees
showed weak
achievement in this area.
St; Thomas/St. John district students'
proportions correct were significantly lower than the correspond-
ing proportions on the same items in all bUt One of the 27 items
on this subtest.
In many cases the differences were quite
substantial with a mean p value of
.51 for the
St. Thbilie.s/
65
63
St. John students compared to .69 for the standardization group.
In the St. Croix district examinees did somewhat better;
}Jilt Still showed weaknesses in vocabulary achievement.
On 17
out Of 27 items on this subtext
the proportion of St. Croix
diStritt StUCent correctly responding was significantly below
that of the StAdardization group;
Again; differences wer
quite substahtial in many cases;
The mean p value for the
St. Croix district studentS on these items was .58 compared to a
value of .69 for the standardiZation group.
English
Item Anajas
111 items
showed satisfadtbry discrimination except the
following:
IteM #
4
This item discriminated satisfactorily between high
and low achievers in the St. Croix diStrict, but not in
the St. Thomas/St. John district.
ThiS was a very easy
iteM for studentS in both districts (p =.94 in bOth) with
the result that most studenti correctly answered it.
This
accounts for the low diStrimination:values obtained.
Item 4
12 - This it zti. did net discriminate well on St. Croix.
fact; more students in the low achieving group scored
correctly on it than stUdehtt in the high achieving group.
It is a moderately easy
item (0....8o on St; Croix);
However;
ah examination of
the incorrect response pattern failed
66
E4
to show any reason for this phenomenon.
The item
asked the student to choose the situation when the
term "love ya" would not be appropriately used;
It
provided the following choices:
1)
in conversation with a friend
2)
in conversation with an older person
3)
in a letter to a friend
4)
in a letter of application for a job
Choice # 4 was considered the correct answer.
It Was
noted that high achieving students who responded incorrectly
were no more likely to choose one distractor over another.
Item #
30
This item discriminated poorly for examinees in
the St. Croix district.
It was a very easy item for these
students (p =.94) and this should account for the low
discrimination index obtained;
item # 48 -
This relatively easy item (p.,,;87) did not thscriminate
satisfactorily between high and low :Achieving students in
the St. Croix oistrict;
From the three word; "askt";
"trouple"; "srtion
;
and 'exspect" examinees were asked
to choose the number of words spelled correctly.
Low
achieving students who scored incorrectly on the item
showed no incorrect response pattern, but high achieving
students who missed this item all chose "3" as the
number of correctly spelled words.
It seems that there
is one of these two misspelled words that was not obviouSly
misspelled even to students who spell well.
67
65
Itemt #53 and 54
were very easy for St. Croix studentt
(1)=.97 and .90; respectively)
and this should account
for the LOW discrimination levelt observed for these
items;
Summary of Student_Skills
Learning Skins_
St. Thomas/St. John schddl district examinees showed
weakhettet in the following areas of dictionary usage skills:
a) identifying a word that contains a given vowel sound
b) the thahqo in principal stress in two
forms of a word
They appeared to have achieve- adequately on the other
dictionary skills which were
tested.
St. CroiX examinees achieved adequately on
all dittionaty
use skills tested except the ability to identify the change in
the principal stress in two forms of a word.
While USVI students soon to be abIe to identify a reference
to aid vocabulary development, they seem to have great difficulty
in identifying a reference source for articles in
periodical:S.
The Students in the usvi sample showed. satisfactory
achieVeMent in objectives
dealing With appropriate English
usage;
They were particularly strong in objectives having to
do with identifying the appropriate usage of slang exprestions
and with identifying the action wordt
in a sentence.
Usage Conventions
Examinees in the USVI sample thOWed weakness
in their
knowledge of the correct capitalization conventions
for seasons
66
and common nouns.
St; Croix school district students exhibited adequate
achievement in the choosing of appropriate Conuna usagei but
St. Thomas/St. John examinees showed weakneSSPS in the proper
use of commas between items in a series and Of the use of commas
after introductory words.
St. ThOMat/St. John examinees showed some weakneSSe8 on
items which measured the appropriate use of verb tenses; but
showed particular strength in dealing with objectives concerning
-;
proper use of pronouns and adjectives.
St. Croix school district students
showed adequate achieve-
r.
Ment in all the areas listed above.
Spelling
USVI students showed particular
strength in spelling;
On all but two of the 15
spelling itemsi the proportion of
USVI studentS scoring correctly was not signifiCantly different
or was significantly higher than that of the standardization
group
The mean p value for spelling items of .82 compares
favorably with the value of .81 obtained by the Standardization
grour
Sentence Sensitivity
item. c-:aIuating'sentence sensitivity
2.erate by
presenting t
students with four similar sentences which
eiffe: pri.mrilyin terms of punctuation and syntax; The student
it .77.e.od th choose the sentence which 'expresses the idea
LeSt." USVI examinees showed strong achievement along this
62
objective with a mean p value of ;82 on these items
compared.
.81 obtained by the standardization group.
Paragraph Arrangement
This section of the Ehi4ligh test evaluates the examinee's
ability to organize a logical paragraph by ordering
four
sentences which are presented into their proper sequence.
USVI
students showed adequate ability to accomplish this task with
a mean p value of .73 compared to a mean p value of .77 on
these items when attempted by the standardization group;
,-
Matheirtics
All items appear to discriminate adequatel
with the follow-
ing exceptions.
Items
1, 2,
5"
and 30 == These were all very (=,t;ty items for the
students in the USVI sample
(p=;92
.94,
.92, and .90,
respectively).
As a result of this;
all studentt regardless
of general mathematiCs achievement level, were able to score
correctly on these items.
IteM
.
11
This,was a moderately difficult item (p=.63)
which
asked students to choose the name which "best desoribet
a lbSed figure with four congruent sides and four con=
gruent angles."
The choices were,
it.ectangIe"i
"rhombus",
.
.
_
"quadrilateral";-and "square." It was notec that examine-et,
particularly
the high achieving group, 'who missed this
item tended tb choose "quadrilateral" rather than
"square"
as the correct answer.
The described figure is, indeed,
70
68
a quadrilateral.
It is also a rectangle and a rhoMbus.
It is obviously the intent of the author of the item that
examinees realize.that the four sides are equal
(congruent), the angles are equal
and
must each measure
90 degrees,. and
therefore,the opposite sides must be
parallel.
The tequitement that the student choose the
"best" descriPtibh may be ambiguous.
ft is possible that
some students may
not know A-711at "best" is in thit case.
The item author obviUtly means, "most pretAse", but
"best"
Co!..ild also mean ''most generalizable ", which would make
"quadrilateral"
the correct answer.
It Is not certain
whether
a student choosing
"quadrilateral" did so because
heir she was unable to go through the thought process
described above or because he or shE defined "best"
in
a different manner
than the author.
The item may be
ambiguous be-catise of this and it is uncertain what
information concerning student achievement it provides;
Sumr:iaty of Student SkiiIs
Numbers, Symbols, andSets
Students in the St.
Tnomas/St: John school diStrict
Showed adequate achievement levels in the following objectives:
a) use of expanded notation
b) converting standard numerals to word form
t) determining place value
d) defining "prime number"
69
They showed weaknesses in
a) determining relationships between two sets
b) determining the value of a linear expression with two
unknowns' when given the values of
the unknowns
t) interpreting Venn diagrams.
St
CroiX school district examinees demonstrated adequate
achievement in all the objectives listed above with the
exception ,L identifying the relay
Jnip between two sets.
Number Properties and Operations
Whole Numbers
St
Thomas/St; John schbbl district students showed
adequate achievement levelS oh the following objectives:
a) multiplication of common
fractions
b) reducing fractions tc
their lowest terms
C) locating decimal points in products
d) subtracting
decimal fr7actibhS
These examinees showed weaknesses in achievement on
the
Following objectives:
a) finding decimal equivalents ofsfraCticn
b) addiI
common fractions with reducing
c) subtracting
common fractions
d) determining relatiVe values of fracticns
e) adding mixed nUMbert
St. Croix school distriet students demonstrated adequate
achievement on the following objectives:
a) multipl]cation of common
fractions
70
b) reducing fractions to lowest terms
c) locating decimal points in products
d) subtracting decimal fractions
They showed weaknesses in achievement in the following
areas:
J.)
finding decimal equivalents of fractions
b) adding and subtracting fractions
c) determining relative values of fractions
d) adding mixed numbers
Nmbex_Properties -an-d- Oper=ations-
Integers and Exponents
St. Thomas/St. John students exhibited adequate achievement
on all objectives in this section of the test except determining
the distance between two points on a number line.
St. Croix school district examinees showed adequate
achievement on all objectives on this section of the test except
the ability to add negative integers.
Mathematical Sentences
USVI students showed adequate achievement on all objectives
evaluated in this section of the test;
eagnremPnt
St; Thomas/St;-John school district examinees showed
weaknesses in determining the meaning of metric prefixes while
Sty Croix district examinees achieved adequately on both
objectives in this section of the test
Ratio and Percent
USVI students tested achieved adequately on all objectives
dealing with ratio and percent except on the objectiVe of
finding a total when a part and its percent of the total
is
given;
f,raphsi probability; and Statistics
Students in both USVI school diSttiCtS showed adequate
achievement on the following objectives:
a) identifying percents represented on a circe graph
b) finding averages of groups of nu dbegs.
These students showed weaknesses on the following
objectives:
a) computing percents represented on circle graphS
b) computing the amounts represented by percents on
circle graphs.
Mathematical Reasoning
St. Thomas/St; John district examinees showed adequate
achievement levels on the following Objectives:
a) solving three step problems involving money
b) determing missing data needed to
solve a problem.
They showed weaker achievement on theSe objectives:
a) stating the solution of a
word problem as a
mathematical expre:;sion
b) SOlving word problems by deduction
St. CroiX ;JiStriCt students achieved at adequate levels
on all of'the objectives listed above with the exception of the
ability to state solutions to word problems as mathematical
expressions;
75
72
Grade 12 - AASK II
Levol
73
The Stanford AchieVetent Test
Test of Academic Skills;
Level II consists of three tests With the number of items
listed below:
Reading
-
78
items
2.
English
-
69
items
3.
Mathematics
-
48
items
The means, standard deviations; and reliability estimates
for the raw scores On these tests, as previously reported
(see Bliss; 1982) are pre_ tinted below.
Test
Mean
Standard
KR-20
Standard
Raw Score
Deviation
Reliability
Error
Total USVI Sample
Reading
43.9
13.8
.93
3.65
English
25.3
8.3
.91
3.01
Mathematics
46.8
11.2
.87
3.36
St.__Thomas/St-_.__John Sample
Reading
40.6
12.3
;91
3.69
English
24.6
7.4
;90
2.98
Mathematics
45.6
10.9.
;84
3.45
St. Croix Sample
Reading
48.2
15.2
.95
3.40
English
26.3
9.4
.91
2.97
Mathematics
48.4
11.5
.90
3.44
aV
Following are the item analyses and summaries of student
ze"Eievnent on the objectives evaluated by each Of the tests.
77
74
75
Reading
Item Analysis
As'on the tenth and eighth grade levels;
it was noted that
a high proportidh of studentsomitted later items on the reading
test;
Table 4 presents the proportion of examinees omitting
each item;
It should be noted that items 52 through 78 make up
Part B
of the test and that these items are timed separately;
This accounts for the drop in omits leginning with item 52;
Again; this phenomenon very likely indicates
that large numbers
of U.S. virgin Islands students did not\have time to attempt
later items on the test and the Vafldity Of\the resUlts on
these Iatei items is definitely qUeStionablABecause
Of this;
\
_
only items which less than 33% of the examinees\dmitted will be
considered in the item analysis and summary of student skillS.
All items whiCh 67% or more-of the examinees did not omit
discriminated at a satisfactory level except the following items.
Item # 1; 3;
9,
11, and 15 - These were particularly easy
items for the USVI sample (p=.95; ;84,
.94, .83; and ;97,
respectively) and low achieving studentS were just as
likely to score correctly on these as were high achievers.
ThiS accounts for the loW discrimination
indices.
76
Table _4_
Ptopottions_of the Twelfth Grade Sample
Omitting Reading Test Items
(Percents)
ITEM #
USVI
STT/STJ
S TX
9
5
6
7
10
11
12
13
14
15
16
17
18
19
20
21
22
1
1
2
2
0
0
0
0
0
0
0
1
0
0
0
0
1
0
0
0
0
2
77
Table
4
Continued
5
7
23
24
8
4
25
7
5
9
26
6
5
27
9
9
28
10
9
11
29
14
15
13
30
14
15
13
31
14
15
13
32
16
16
16
33
16
15
16
34
18
16
20
35
18
16
20
36
23
23
24
37
29
27
33
38
29
24
35
39
30
26
36
40
35
34
36
41
33
28
38
42
34
30
40
28
30
25
44
31
31
31
45
32
31
33
45
32
31
33
46
34
34
35
47
37
35
40
80
Table
Continued
37
38
38
39
35
34
34
35
78
40
44
44
44
48
49
50
51
52
5
53
10
14
5
54
6
9
2
55
9
11
5
56
8
11
4
57
7
9
58
6
9
2
59
7
II
60
9
12
61
11
16
62
10
16
63
12
19
64
10
15
4
65
13
19
66
17
24
67
15
23
4
68
16
24
5
69
16
24
70
16
24
7
71
19
24
13
72
15
23
4
73
16
24
4
81
TablO
4
Continued
79
74
16
24
4
75
19
26
11
76
17
26
5
77
19
26
11
78
18
28
4
80
1:e:7
18
This moderately difficult item (p=.63) showed
elrticui,;Iiy IOW discrimination in the St;
Thomas/St.
,L.)hn saran:
(d=.05 With p=.57) .
The item involved "subject"
as it was used in reading Passage about music;
the
response 'object" was chbeh by most of the high
achieving students who missed thiS item rather
than
"theme;" 'which was the correct response.
Low achievers
who scored incorrectly in the St. Thomas/St.
Jbhn sample
answered more or less randomly;
High achieVing students
who were not Certain Of the correct answer may have been
distracted by the altorr,itive "object" since it is
used
often in lanQuage in connection with the term "subject."
Low achieving students were likely act to hae perceived
this reIationShib and to have
answered more randomly
with some students chobSing the correct response
hy
chance;
This could account fbt the low discrimination
observed in this item.
This was a very diffiCult it 'm for St: Thomas /St.
Jah:i students (p=;07); but was soM4hat easier for
St
Croix school districtstudents
(0=.22).
The poorer
discrimination in the St. thOnia. JOhn district resulted
from the fact that fewer high achieiing students in that
district = cored correctly on this item.
The proportions
of loW aChieVing students scoring
correctly was approxi-
mately the same in both distracts;
The item asked
examinees to ChObSe ar eamp,e of a "geographic building
8J
m-Fiterial"; defined in the reading as a material which
"uses local resources
.
.
.
and suits the local climate."
Choices were:
a) stone for a pyramid in Egypt
b)
snow for an igloo in Northern Alaska
c)
adobe for a museum in New York City
d)
a steel girder for an office building in Pittsburgh
The response considered correct was "b."
Mbtt St, Thomas/
St. Johh dittriCt students chose "c" as the correct response
while thb-se St. Croix examinees in the high achieving group who
s)red incorrectly chose more or less evenly between "a" and "O."
Adobe is the only material specifically
identified F'7 a geographic
.buiIding material in the passage (although only
ir
cf
once to
to a community in New Mexico and the "hot; dry scut.
,:ern
part of the United States." Students may have seen adobe in
the passagc acrd chosen response "C" on the basis of this;
disregarding the location of the museum.
On the other hand; they
may not have been aware of New Y:.rk City geography and climate;'
Neither explanation takes into account the differences observed
in the behavior, of the St; Croix sample.
In 9enetal
May be a poor item for a number of reasons.
requireS examinees
to haiie a strong knowledge of
.lecigralbhic facts n--)t given in the
readings.
This may make it a good geography test item but it
is uncertain whet:ler a student a:swering incorrectly did so
due to lack pf geographical %nowledge or lack of reading skilIG.
84
82
Tn addition; the item may_ really haVe tWo correct answers.
one for a pyramid might well be considered a geographical
building material since it is natural to Egypt and suits the
local climate there.
He:
e, the results on this item should
be examined with caution.
Sumary of Student Skills
lobal. Meaning
Students in the USVI sample showed some weaknesses in
the area of
meaning determination.
However, they seemed
to_score well on earlier; easier items.
Examinees Seem to
score better, in absolute terms and in relation to the standard-
iation groan, on earlier items dealing with global meaning.
On the fii;ur
teems with OMit rates of less 'than 33 %, the roan
p value wczs
.56 compared to .65 for th
standardization group.
Meaning of Explicit
USVI students appear to have achieved quite weid on
objectives dealing with determining
explicit meaning from
readings.
The mean p
an
value for the
It.z,ms measuring these
Skills gas .74 compared to
.75 for tl, standardization group.
On three items the proportion of USVI students scoring correctly
was significantly greater
than the proportion Of the
staiiJard-
ization group.
meaning of Implicit Detail
The studentS In the USVI sample seen to have achieved
adequately in objectives dealing with tne determinatioh Of
Tracts which are iMplitit in
readings;
On items which an
85
83
appropriate number of :ittia-ents were able to attempt to respond
to tne mean p values
,
.67 in the St. ThOmas St. John
district and .74 in tFie St. Ci-oix district comared to .74
in
the
grcup;
Again. it was observed that local
students tend to do less well on items ocurring later in
the
test.
Inference and Logical Analysts_
USVI students ShOWe-1/4.1
weaknesses in the area of
inference and analySiS,
Cf the items
were appropriate
t o eyamine; the rrle,
p va3.ues were .511f.).7- the St;
Thomas/
St. John sc1lool
and
for
C.:oix school
district compared to .64 for tne Ftandaiaizaton group.
While
the St. Croix grols auhie ement level could
be Said to be
adeclU
there shculd be some concern ,Ibdut the level of
achluvent,:Tit7
n the
i
Thomas/St. John distr_let.
Wurd Mean is
Wdrd mean involves vocabulary skins.
Pere, market,
iifferences were noted between examinees In the
two school
stricts.
St
CrOiX district students showed c:onsiderably
higher achrevement "mean p value E.
.68) than Lite St. ThOM6S/
St. John district examinees (p,i..52).
It was noted that the
Thorna/gi--; John examinees
rig tcd items considerahIy more
often than St; Croix examinees,
For example,
11
of thi.
St. Croix examinees -.)mitted
the last item
#78) on the test
compared t
28% in C:ie St. Thomas/St. John district.
This
difference seems to be due to one St.
Thomas/St. John
class
8f;
of the examins omitted the item compared to l4
it
the othet.
Sinte scores on the other parts of the battery are
not greatly different between these two clasa,=, it is possib-1_e
that he was a
lem in test adminis
in the former
claSs;
Such thin
s inad-quate instruct!on on this part of
test or error in timing could be the cause of this
pheholion
At any rate: the St. Thomas/St; John scores valiai
are oper to question.
The St. Croix examinees show good achievement on the
items with their Mean p value (.68) comparing favorably with
the standardization mean p value of .70.
English
Iter Analysis
All ite:As discriminated adeqUate17? eci:tit the
following:
3;
3i; 37; 42; 46; and 52
These wGre particularly
it,is for the USVI sanple of examl:.ees
(p-=,
.96.
.93,
.96,
.96;
6; and .92, respecitive:
aG.i:unt for the io
aisciimination
This
,17.-ould
observed on thG:.
_;like low achieing students wer_ almost as ltkely
to score correctly on these items as were high achieving
:Ltd n
c
L
.)eau:,e they were Frasy.
tem
18
Thi' iLet
Auat
tor the St. thai6
st.
J(i.nn fchool district &.kanin-ci: , bu
poorly. for '0:16
t.c:-oix examinees. This was due to the fact that St; Cro;'
examinees found ikie item much easier
(p=55) than the
St. Thcmas/St. John examine-6S (p=.86) and this produced
the ldi.4 level of discriminatien typically found on
items with high difficulty indie6S.
It
54 - This was a moderately easy item for USVI examinees
(p=;83):
It was difficult to see why this item disoriminare(i
as poorly as i i
did.
There appeared
be no pattern
of incorrect responses
and no extraneous clueS in the
item itscl.
The item appeared to be clear and worded
Simply.
Summary Of Student Skills
Fearning Skills
On Objectives concerned with dictionary usage HSVI
studes '6.-sted appeared to have difficulty
using the rn-mutiti-ciAtL n
key to intify words
containing
vowel sounds;
8E: CrbiX.
examinees
Fhowed some weakneSS
their ability to locate
entry words in a dictionary.
On all other.
dictionary skills
tOrtt-ed;
tIte USVI sample
Showed adequate achicvemen
Students in the St.
Crei:;: sampL showed exceptionally
s rbhg achievement o- objectives dealing with the use
of
reference sources;
In tact
students in this simple had
proportions answering correctly significantly higher than the
standardiatiOn group on all
tii6e object
St.
'1'h-Otiia8/ct.
to
-r.hieved
aduqw:teiy C
ai_ objecLives
de7'ilind with the use of referene,
sources except the ability to identify the use of a hhesaUtus.
86
On
bjectives dealing with knowledge of the use and
structure of English USVI
examinees showed strong achievement
on the following:
a)
idenLifi-ati^n of a prepositional phiage
b)
identification of an adverb forming morpheme
They deminstrated weakne'sSeS OP the following objectives:
a)
identifying an appropriate place
for slang expressions
b)
identifying an appropriate Closing for a formal
letter
_
C)
identif:2,ng a standard EncT1111 sentence
d)
ideotifving a sentence containing a prepositional
phrase.
usage convritions
CapitLlization
Examlneet in both districts shoWed weaknesses in
their
knowledge of ca:pitalization conventions for common nouns;
In
qenerali they tended to
capitalize commor r uos
that should
have been written &,:ing leer case Initt,1 letters.
St. Thomas /St. John
district stuu.1 is showed weakness;_ _
in the knowledge that natiOnaiities shou7d be capitalized
While St. Croix district students appeand to have achieved
aleqoat,. y along tj..s
oojective.
EXaM_Te(:..; in both distiiCtS aLpeartd to have little
diffiedltV recooni7ing that dayS Ot the t--k should
hr
- Puncr,.s'Lion
.XEth
-he
...YT
-mple 5h6wnj aden,late
achievement
87
on all items
testing the appropriate use of punctuation.
F--,:age Con ent
,s
Parts o
-,ech
St. Thomas/St; John schobl dis-Aitt examinees demonstrate
adequate achievement on all objecti. deling with usage
parts of speech tested with the exce.pti,,
of:
a)
verb redundancy
b)
the use of a participle
St; Crci- school district students 3hOtvd adequate
of
achiment on all these object -ives except the use
of a participle.
Spelling
USVI examinee F; Showed adequate achieVeMAitj_n spelling
With
p value of .70 compared to .72 for.the Standard-
izat
On t-he lc itrst8 testing spelling sxills
;
-Agnificane.y nig,;er proportions scored correctly than
the Standardization g :oup on three ite
and :;ignificantly lower
prop-dr:
correcly on only three
it(uris;
Sonta_ce
Studeti
the
sample achieVed well on
items
designed to teat the
ob3ectives.
Their prOjortions scored
correctly were as high or higher than the Staidardization
groul Gn all
itoros Wi,:h a mean p of
;77 co:r. a '-cd to .76
for
the st;.indardization group.
Paragraph Arrangemel::-
USVI examineeL: achiel;ed adaquatr'ly on c'hie-;;;tives designed
k
to test their abillties to arrange paragra hs with a mean p
value of .61 compared to
.67 for the stan4 ar-ILzatiOn group.
CCU
Mathemat
Item Analysis
All items discriMinated Leguately with the following
exceptions:
Item # 1 and 2 = ThS6 were
particuIarl,J
sy items for the
USVI sample
and .88 respectively).
As a result of
this phenomenon it was apparent tha
IOW achieving students
had almost the same probability of scoring correctly on
these items as did high adhiVing students
This would
account for the low discriMination power which was obserVed.
Item 4 44 =- This was a particularly difficult item for student8
in the USVI sample (p=;12);
ThiS itt. is notable becauSe
Of the moderate negative
rimination in the St.
Croix
group
However, it was nc
tat 45-6 of the 2t; Thomas/
St. John examinees and 27% or
the St. Croix examinees
emit.Led this itetti (it was near the end of the 48 item test)
1.-t
is believed that any
t.:) interpret these
z °esults, ma,
be reading more validity into the item than is
warranted.
on the reading
test; there may be a timing
difficulty involved with this test; although it appears
less apparent witn this
teSt.
For instance; thti. rise in
the proportion of omits
L; net nearl
as smooth as the
rise on the reading test (See Table 5).
It seems as if
studeLits ilave had a chance to attcTupt most
items;
bu+-
left those they had diffic
iinSWei[ng blank,
This
Table 5
Proportions of the Twelfth Grade Sample
Omitting Mathematics Test Items
(Percents)
89
ITEM #
USYI
STT/STJ
S TX
1
0
0
0
2
0
0
0
0
0
0
4
7
5
9
0
0
6
0
0
7
4
9
8
0
0
0
9
5
5
4
IC
3
7
11
0
0
0
12
0
13
0
14
15
2
4
0
1
17
2
4
18
5
19
1
0
'7
21.
22
2
90
Table_5Colitinuez1
23
0
24
1
2
25
1
0
2
26
5
27
28
29
7
30
4
5
31
15
32
33
5
5
5
34
4
5
2
35
5
2
36
5
2
37
8
5
38
11
12
9
39
15
40
23
28
16
41
19
24
42
13
27
5
43
20
31
5
44
37
45
27
45
32
41
20
46
29
43
11
47
'6
39
7
48
33
46
15
91
seems apparent,
fOr instance; in the St.
CroiX sample
Where 27% of 4:L=:aminees omitted item number 44, but
only 7% c!r:7;_
:Htem # 47.
Again; a larger number of
omits was Observed for the St. Thomas/St; John sample
than fer the St. Croix sample.
Since the mean number
correct fOr these two samples differed by only 1.7 points
(See Bliss, 1982)
it is unlikely that the
differen,fes in Omits were due to differences in math
ability between groups.
Instead, as in the reading test,
it is more likely the differences observed were due
tb administration probleMS in the St.Thomas/St. John
sample.
SummarT__of_Student Skills
Nur-lb_S
and Sets
Examinees in the St;
ThomaS/St. John school distrt.
snowed adeqthte achievement on the following objectives:
a)
identification of place ValUe of a digit in a
decimal
b)
identifidation of the reIationShp between two Set,
They showed weakneSS5?s on tie fallowing Objective
)
rounding decimals-
intarpreting Venn diagram.,
c)
it tetp.L.E:ting abol.ttO values of e number.
St. Croix district examinees showed precisely
he opposite
strengths and weaknesses. They appeared to show adequate
achieveMent on the three objectVeS on which the St. Thmas/
St; 3.61in examinees
ere weak and to show weak achievement on
92
the two objectives on which the former group Df examinees
displayed adequate achievemnt.
NUMber
Properties and Operations_-
Whole NUmbers
St. Thomas/St. John district examinees displayed
adequate achievement on ;fill objectives in this grouping;
St. Croix district students showed adequate achievement
the following objectives:
a)
adding a broken column of three numbers
b)
subtracting with renaming
C)
multiplying involving zeroes
d)
multiplication of prime numbers
e)
application Of the Multiplicative property over addition.
They showed weakness in the area of
identifying
of prime numbers.
mPmber Properties and Operations
Common Inc; Decimal Fractions
St. Thomas/St; John district examinees showed adequate
achievement on all objectives
this area except:
a)
adding fractions with reduction
estimating the quotient of decimal diVisien
c)
identifying the least common denominator Of
liter,t1
fractions
St. Croik district examinees showed adequate achievement
on all 5bjectives in this area except:
a)
adding fractions with reduction
1-0
estimating the quotient in decimal
division
95
93
c)
subtracting common fractions with unlike denominators
after renaming.
Number Properties and Geratioi
-
Integers and Exponents
USVI examinees showed adequate achievement on all objectives
tested in this area;
Mathematical_Sentemoes_
St. ThoMas/St. John school district examinees showed
adequate achievement on all objectives tented in this area
except the ability to translate verbal prot:lems into
mathematicaI equaions.
Croix district students showed adequate achievement
on all objectivc3 tested in this item grouping
Geometry and Mr,asurement
There wore subr':,,etiaI differences
in achievement
this area between exam...nees in the two school
diSttiCtS.
St. Thomas/St.
John. district students di:;played adequate
achievement on the ability to approximate the inglish equivalent
of a lihbar tOttl-c unit.
They showed weak
evement on the
fcIl-
Hectl.ves:
itifyng characteristics of commoh polygons
,rm:',ning the relatiV
areas of potygohs
given
their dimensions
c)
naming the ang.:.c
that result from e. setting a right
angle .
d)
solving one step word problems that .nvolve
convering:
between English linear units.
94
St; Croix district examinees showed particularly strong
achievement on the following objectives:
a)
identifying characteristics of common polygons
b)
naming the angles that result from bisecting a
right angle
They showed adequate achievement on these objectives:
determining the relative areas of polygons; given
their dimensions
b)
approximating the English equivalent of a linear
metric unit
These students showed weakness in solving one step woid
problems that involve converting between engiisn
linear units.
Ratio and PPr_rent
St. Thomas/St. John district examinees displayed
adequate achievement on 411 Objectives tested in this area.
St. Croix diStriet Students showed adequate achievement
on the objectives tested in this item grouping except the
ability' to solve rate problems.
Graphs ;Probability;
and Statistics
St. Thomas/St; John district examinees
demonstrated
adequate achievement in the ability to use tables
to make
comparisons.
They showed weaknesses in the following areas:
a)
solving word problems by finding the average
two numbers
b)
finding the
average of positive and negative
numbers
solving word problems by interpreting a
table to find
percent success
9C5
d)
identifyinr the prObability of a particular eVt-4,I't
occurring;
TWO Other items had omits aboVe 40% and no attenpt will
be made to derive information from theM.
St. Croix diStriet examinees showed adequate
achieveme:n't
on the following objectives:
-a)
finds the average of a positive and negative numbcr
b)
using tables to make comparisons
C) choosing a set of ordered pairs that would be on
the same line as two given points;
They showed weaknesses On the follbing objectives:
a)
solving word probleMs by finding the average of two
nuMbers
b)
solVing word- problems by interpreting a table to
find percent of success
c)
identifying the probability of a particular event
occurring
d)
identifying coordinates on a
graph off a linear
equation
e)
choosing a set of ordered pairs that would be on
the same line as two given
Mathematical Reasoning
USVI examinees displayed adequate achievement on all
objectives tested in this
item group.
In the St. Themes/
8t. John sample,
one item had 46% omitted reSb6nSeS (item # 48)
and was het US-0d in this analysis.
96
Discussion
A detailed summary
tudorit basic skills across grades
and recommendations based on the results of the achievement
testing
will_ of necesSitY, have to be put off
Until the results of z,1:7.3ntary school testing are analyzed and
reported.
Nevertheless; there are some facts brought out by
thf..analytiS of the secondary school- examinee scores which are
worth, at leaSt tentatively;
touching upon.
Most striking is the time situation on the reading
comprehension tests.
On all three levels tested it was obvious
that many studentS did not have sufficient time to attempt all
the items on these tests. -Teachers and counselors were
instri :tea to Strictly adhere to the time limits SP-edified by
the test publishers.
These time
had bean empirically determined
to be sufficient for all
studentS to complete the test during
the standardization stuies described in Technical Report
# I.
It appears
that those time limits may no'.
be appropriate for
USVI students who may 1.),: slower; more deiberate readers.
This
characteriStid of local students is not
necessarily undesiruble
provided students end up_understandina what theY read.
Nevertheless, it tray' account for low rea ling scores since most
standardized tests count'omits as
incorrect answers.
FrOm
a technical point of view, the result of this phenomenon may
be that students arc taking a
timed test rather than a power
test and this brings up numerous problem; in score interpretation.
97
It is notable that time seems to be sufficient on most other
_
7_ests.
No_attempt is made here to describe the causes of this
phenomenon;
Many alternative hypothesis present
themselves;
HoweVer, it is important that research be conducted to establish
the nature and causes of this difficulty.
Second; there appears to be a pattern that can; at least
tentatively. be observed in student achievement.
USing the
taxonomy of edUbatiOnal objectives in the cognitive deMain
(Bloom; 1956) as a theoretical framework; some sense can be
made out of this pattern.
-U.S. Virgin Islands students generally
appear to achieve at least adequately on- objectives that would
be classified in the lower three categories of the taxonomy
(i.e.-knOWledge;_comprehension; and appliCatioa).
They achieve
leS-fJ well on objectives that would be elaSSifi2d in the upper
three categories (analysis
synthesisi and eVa uation). Again
the reading comprehension tests exhibit this phenomenon quite
clearly.
USVI examinees achieve adequately on
items which
test their abilities to
understand facts that tte explicitly
stated in the text (comprehension
level object veS); bUt are
weak iA determining implicit, global; and cont!xtual Meanings
(analysis and synthesis objectives).
Those stAdents seem to
achieve well on objectives which can be achiev_:d by memorization
(knowledge objectives) or the application of a general procedure;
rule or algorithm (comprehension and application level);
They
scored extremely
wall
in spelling and Singl:.) step mathematical
computations.
They scored
lower on itemS reqiiring
98
them to organize data; choose information relevant to solving
problems, and to cheese a series of rules and/or algorithms
to use in 'solving problems (analysis; synthesis; and evaluation
objectives)
.
Note the weaker achievement in mathematics
4
applications problems; multistep mathematical computations)
and the use of dictionary entries; for example;
Again, there
is no attempt here to determine the causes of this pnenowenon;
but if analy8is of the elementary achievement tests substantiates
this tentative observation, further research into this phenomenon
is important.
Third, U.S. Virgin Islands students in the grade levels
tested displayed weakness in their knowledge of standard
English vocabulary.
Finally i
there is the issue of test taking Skill8.
one
particular phenomenon was very apparent.
Many examinees omitted
items which they had a chance to attempt to answer.
Although
t was not explicitly stated in the directions; th_ score
students received was the number of items they answered correctly.
There was no correction for guessing and an omit was counted
the Same as a wrong answer.
This should have been explained
in the directions and the lack of such an explanation shows
_a oeficit in the test publishers°
thorouginess in test con-
struction.
Nevertheless; most tests students take use this
"number right" scoring procedure (exceptions are the Preliminary
Scholastic Aptitude Test and the Scholastic Aptitude Test which
are not given earlier than grade eleven)
and the best strategy
99
for examinees wishing to maximize their scores when confronted
with an item they did not know the answer to would have been
to try to discard alternatives which they know were wrong and
to guess randomly from the remaining alternatives or; if they
could identify no incorrect answers, to simply take a random
guess.
Many examinees seemed to be unaware of this strategy
or unwilling or unable to use it.
100
References
Allen, M.J. & Yeti,
W.M.Introduction_toneasurement theory.
Monterey, CA: BrOOkS/Cole Publishing Co.,
1979.
Bliss, L.B. Virgin Islands of the United
States public school
ha'fc skills achievement survey;
Technical_Report #1:
Validation of the use of the Stanford Achievement Test
with_ U-S.V-T. students._ St. Thomas, USVI: Caribbean
-
Research Institute, College of the Virgin Islands; 1982.
Bloom; B.
(Ed.) Taxonomy of educational objectiv_Randbook I:
Cognitive_damain_._
New York: Longmans, Green
& Co., 1956.
Cooper, V. Personal communication, June
I,
1982.
Cureton, E.E. The upper and lower
twenty -seven per cent
rule;
Psychometrika, 1957,72_, 293-296.
Kelly, T.L. The selection of upper and lower groups for the
validation -or test items.
Journal of EduCational_Psychology,
1939, 30; 17- -24.