©2003-2013 Carolyn Kottmeyer
Testing and Assessmentof Gifted ChildrenApples and grapefruit?
New Jersey Association for Gifted Children (NJAGC)
This workshop is brought to you by the New Jersey Association for Gifted Children, An Affiliate Group of NJEAVisit us on the Convention Floor in the Affiliates Section of Main Street…
Latest state requirements for gifted education Membership information… Enter a drawing for a free
membership Professional development opportunities in gifted
education NJAGC conference in March 2014 Awards Discover what NJAGC can do for you. www.njagc.org Phone: 856-273-7530
©2003-2013 Carolyn Kottmeyer
©2003-2013 Carolyn Kottmeyer
Hoagies’ Gifted Education Page
The “all things gifted” page 1150+ pages = Overwhelming Site-wide search on every page!Not supported by advertising or grants Affiliates programs Click on Shop Hoagies’ Pagewww.hoagiesgifted.org/shop.htm
Hoagies’ Gifted Education Page
Facebook / LinkedIn / Twitter feed Research summaries Quick events and contests News Gifted Parent, Teacher & Professional
Q&A
©2003-2013 Carolyn Kottmeyer
©2003-2013 Carolyn Kottmeyer
1, 2, 3, Testing…Grade Level Achievement Test Out-of-level achievement testIndividual Achievement TestGroup Ability TestIndividual Ability / IQ Test
©2003-2013 Carolyn Kottmeyer
Ability Test Terms…Criterion-referenced (CRT) Measure how well a person has learned a specific
body of knowledge and skills Content based on its significance in the curriculum
Norm-referenced (NRT) Compare an individual's score against the scores
of a group who took the same exam, the “norm group” or “standardization sample”
Content is chosen by how well it discriminates among students
Some tests allow both comparisons…
©2003-2013 Carolyn Kottmeyer
Scoring terms…Mean Average score on test IQ test expected to have mean of 100Standard deviation (SD) Statistical term: difference from norm Two standard deviations above
average is considered gifted 130+ (for standard deviation of 15)
©2003-2013 Carolyn Kottmeyer
Normal Distribution
Distribution of IQ Scoreshttp://encarta.msn.com/media_461540296/Distribution_of_IQ_Scores.html
©2003-2013 Carolyn Kottmeyer
Scoring terms…Standard score Looks like IQ score, but may be very limited Mean = 100 standard deviation may or may not = 15
Age EquivalentGrade Equivalent Age / grade of average students with this
score
(http://alpha.fdu.edu/psychology/oat_cereal.htm)
©2003-2013 Carolyn Kottmeyer
Scoring terms…Percent Correct 90% means 90/100 (or 45/50 or 9/10)
correctPercentile If 10% of norming group got 100% correct,
perfect score is 90th percentile 50th percentile means average
Raw score Items correct of items attempted
©2003-2013 Carolyn Kottmeyer
Grade Level Achievement Test
Standardized for one grade level only Or less… early elementary grades have different
norms / tests available for fall and spring testingGrade equivalents compare how child of grade x would do on this grade level’s test NOT how this child would do on grade level x test!
Group test Testing in the classroom has built-in distractions…
©2003-2013 Carolyn Kottmeyer
Grade Level Achievement Test
Nationally normed ITBS, CAT-5, SAT 9, Terra Nova, CTBS, MAT Norm referenced (ITBS scores as both norm
and criterion referenced)State mandated tests ESPA, GEPA, PSSA (Pennsylvania), TAAS,
TAKS (Texas) Criterion referenced
MCAS (Massachusetts) Norm referenced, sort of…
©2003-2013 Carolyn Kottmeyer
Out-of-Level Achievement Test
Comprehensive achievement tests Developed for Talent Search or private schools More comprehensive content than typical
group achievement testsGiven to students 2-5 years younger than norms“Combs out” upper level achievement 95th-99th percentiles are spread across entire
percentile spectrum on higher level test
©2003-2013 Carolyn Kottmeyer
Out-of-Level TestingUsed by Talent Search programs Johns Hopkins CTY / Northwestern CTD / Duke
TIP / Rocky Mountain Talent Search Carnegie Mellon C-MITES
Identify gifted students vs. high achievers Those still scoring in higher percentiles on 2-4
grade level higher test Usually 50th percentile or above
Examples: SCAT, Plus, Explore, SAT
©2003-2013 Carolyn Kottmeyer
Individual Achievement Test
Less questions per grade level, but more grade levels coveredNorm-referenced scoring Compares child to average child of age/grade
level Does NOT evaluate child on specific curriculum
contentExamples WIAT (Wechsler) K-TEA (Kaufman) PIAT (Peabody) WJ-III (Woodcock Johnson)
©2003-2013 Carolyn Kottmeyer
Individual Achievement Test
Watch for test ceilings Average 12th grade math = Algebra I Average 12th grade reading level…Use WJ-III for older gifted kids 12th grade ceilings on WIAT / PIAT / K-
TEA 18th grade ceiling on WJ-III
18th grade is NOT post-graduate
©2003-2013 Carolyn Kottmeyer
Individual Achievement Test Scores
Scores include Percentile Standard score Age and grade equivalent
If school doesn’t share all scores, insist! FERPA
Can also report score as if in higher grade Child is rising 3rd grader, report scores also as
a rising 4th grader to see how child would compare to potential classmates
©2003-2013 Carolyn Kottmeyer
Curriculum Based Assessment (CBA)
Most valid for comparing child’s achievement to school’s curriculumUses actual curriculum materials Standard midterm or final exam Compiled from end of chapter tests
NOT chapter test after chapter test after chapter test after…
©2003-2013 Carolyn Kottmeyer
Curriculum Based Assessment (CBA)
Conducted thoughtfully… Review may be allowed Answer terminology questions during
testingScored reasonably… 85% - 90% correct Perfection is NOT the goal
If the child doesn’t get required 85-90%, parts of subject already mastered need to be skipped
©2003-2013 Carolyn Kottmeyer
Test Scores vs. Classroom Placement
1. Curriculum Based Assessment CBA level = grade level
2. Out of level Achievement Test Comprehensive and above level
3. Individual Achievement Test Not as comprehensive for each grade level Compared to average students (norm-referenced) Useful for comparing child to heterogeneous
elementary school classroom4. Group Achievement Test
Tested grade level only Shows child is at or below grade level
©2003-2013 Carolyn Kottmeyer
Ability / IQ TestAbility vs. Intelligence? Terms are interchangeableSome tests are more widely accepted, and have a more stable test/retest and test to test correlation Buros Mental Measurements Yearbook
Library reference section or www.unl.edu/buros/
©2003-2013 Carolyn Kottmeyer
Group Screening MeasuresBe aware… Normalization sample size Mean, deviation, and standard measure of
error Hard ceiling score
GATES and others Survey – only as good as teacher-training
SAGES and others Group test of both ability and achievement
©2003-2013 Carolyn Kottmeyer
Group Ability TestSame problems as group achievement tests Written for a single grade level Administered in classroom environment
Group ability tests often underestimate scores later received on individual tests Examples: OLSAT (Otis Lennon Scholastic Abilities Test) CogAT (Cognitive Abilities Test) Co-normed with
ITBS
©2003-2013 Carolyn Kottmeyer
Group Ability TestGifted children often over-think the questions No chance for tester to say, “What did you mean
by that?” “What’s a more common answer?”Study shows Reverse correlation between group ability test scores and gifted children The more gifted, the LOWER the test score!
Study shows LD/2e children underestimated These children are more accurately identified by
individual IQ test
©2003-2013 Carolyn Kottmeyer
Individual Ability TestConsidered best estimate of ability Still not perfect! Test style vs. child’s strengths Tester vs. child “fit”Examples: Wechsler series (WPPSI-III, WISC-IV,
WAIS-IV) Stanford-Binet (SB-5) Differential Ability Scales (DAS-II)
©2003-2013 Carolyn Kottmeyer
Intelligence (IQ) TestLess commonly used individual tests WJ-III cognitive (Woodcock Johnson)
Different definition of intelligence Excellent for locating LDs in 2e child
K-ABC (Kaufman Assessment Battery for Children)
Lots more… Inventory of Tests www.hoagiesgifted.org/tests.htm
©2003-2013 Carolyn Kottmeyer
Intelligence (IQ) TestNonverbal tests Naglieri Nonverbal Ability Test (NNAT) UNIT (Universal Nonverbal…) Ravens (group)
Research shows NO cultural “fairness” in non-verbal tests
David Lohman, University of Iowa
©2003-2013 Carolyn Kottmeyer
IQ Test CeilingsWISC-IV, WPPSI-III and SB-5 Ceiling of Full Scale 160 (SD=15)WJ-III Cognitive Ceiling of GIA depends on student age
(SD=15)DAS (Differential Abilities Scales) Ceiling of General 175 (SD=15)Other tests usually ceiling=160, SD=15
©2003-2013 Carolyn Kottmeyer
IQ Subtest CeilingsWISC-IV / SB-5 Subtests at scaled scores 17-19 (99th
percentile) Penalties for careful and thoughtful
completion (speed bonus) By some age, a child can get every WISC-III
question right, and still score under 130 due to time penalties!
Better on WISC-IV, but still strong time bias
©2003-2013 Carolyn Kottmeyer
IQ TestingWISC (Wechsler)"He [Dr. Wechsler] rejected most attempts that I made to add easy or hard items to the WISC-R saying firmly, 'My scales are meant for people with average or near-average intelligence, clinical patients who score between 70 and 130.'" "They are clinical tests." When I reminded him that psychologists commonly use his scales for the extremes, and want to make distinctions with the 'below 70' and 'above 130' groups, he answered, "Then that is their misfortune. It's not what I tell them to do, and it's not what a good clinician ought to do. They should know better." (Kaufman, Intelligent testing with the WISC-III ,1994)
©2003-2013 Carolyn Kottmeyer
IQ Subtest CeilingsWISC-IV and WPPSI-III The WISC is not a good measure for children
scoring outside of 3 deviations from the mean. An average subtest score of 14 or 15 is 2 standard deviations outside the mean, an average subtest score of 16 or 17 is 3 standard deviations outside the mean. Sattler and Dumont do not discuss the use of the test above those levels at all… Summary of Assessment of Children WISC-IV and
WPPSI-III Supplement, by Sattler and Dumont
©2003-2013 Carolyn Kottmeyer
The View from the Ceiling…
Was completion criteria reached? Completion criteria:
<3 questions correct of 5 consecutive asked
Did student run out of test questions, or score possibilities? If student answered more questions
correctly, could score increase?
©2003-2013 Carolyn Kottmeyer
Flynn EffectAccording to Flynn…
IQ increases by 3 points every 10 yearsDoes Flynn effect apply at the extremes? From SB-4 to SB-5, ~4 points at
mean=100 Gifted sample SB-4 to SB-5 scores vary by
7+
©2003-2013 Carolyn Kottmeyer
Flynn Flam…?Smart as We Can Get?: Gains on certain tests of intelligence are ending in some places, American Scientisthttp://www.americanscientist.org/ Some countries are experiencing a Flynn effect with a reversed sign…
©2003-2013 Carolyn Kottmeyer
Levels of GiftednessLevel IQ
range OccurrenceBright / High Achiever
115-130 1 in 6
Moderately Gifted 130-145 1 in 44
Highly Gifted 145-160 1 in 1,000
Exceptionally Gifted 160-180 1 in 10,000
Profoundly Gifted 180 + <1 in 1,000,000
Excerpt from “Exceptionally and Profoundly Gifted Students: An Underserved Population” by Miraca U.M. Gross
©2003-2013 Carolyn Kottmeyer
Levels of Giftedness (SB-5)Level IQ range
2 grades ahead by start of K (MG) 125-135
Push the limits of gifted program (HG) 130-140
Need big accommodations,
multiple grade skips (EG)
135-141+
Probably never fit… (PG) 141+
“Ruf Estimates of Levels of Giftedness,” http://www.talentigniter.com/ruf-estimates
©2003-2013 Carolyn Kottmeyer
Levels of Giftedness (WISC-IV)
Suggestion of new range above 130, would create these levels:
Level IQ rangeModerately gifted 130-138
Highly gifted 138-145Exceptionally gifted 145-152
Profoundly gifted 152-160Assessment of Children WISC-IV and WPPSI-III Supplement, by Sattler and Dumont
©2003-2013 Carolyn Kottmeyer
ID the Gifted with WISC-IVWISC-IV subscales are lowering full scale IQ Identifying far fewer gifted students
For gifted sample, full scale IQ = 123.5 Verbal Comprehension average 124.7 Perceptual Reasoning score average 120.4 Working Memory average 112.5 Processing Speed average 110.6
(source: History of the WISC IV http://www.psychpage.com/learning/library/intell/wisciv_hx.html)
©2003-2013 Carolyn Kottmeyer
ID the Gifted with WISC-IVFull Scale IQ (FSIQ) should not be reported if the variance from the highest to lowest composite score is 23 points or greater Essentials of WISC-IV Assessment,
Flanagan and Kaufman (2004)
©2003-2013 Carolyn Kottmeyer
WISC-IV Global Ability Index
Publisher’s bulletin on Global Ability Index (GAI) GAI removes Working Memory and
Processing Speed Index from calculation Do not assess what’s commonly included in gifted
programs Do not show the same 15 point standard deviation
Published AFTER the WISC-IV Psychologists NOT notified after they purchase /
train
©2003-2013 Carolyn Kottmeyer
WISC-IV Extended Norms
Publisher’s bulletin on Extended Norms Gives credit for correct answers above
subtest ceilings Used when child has 2+ 99th percentile (18 or
19) subtest scores Published LONG after the WISC-IV – February
2008 Psychologists NOT notified after they purchase /
train
©2003-2013 Carolyn Kottmeyer
Stanford Binet Extended Norms
Publisher’s bulletin on Extended Norms Available only for those who achieve FSIQ
scores above 150 Theoretical ceiling EX=225 ~166 people at this level in the entire U.S.
(source: Riverside: Use of the SB5 in the Assessmentof High Abilities http://www.assess.nelson.com/pdf/sb5-asb3.pdf)
©2003-2013 Carolyn Kottmeyer
Measures of General Intelligence
Good Measures of “g” Arithmetic .768 optional subtest on WISC-IV Vocabulary .751 Information .748 Similarities .733
Fair Measures of “g” Matrix Reasoning .687 Block Design .672 Word Reasoning .648 Comprehension .646 Letter-Number Seq. .621 Picture Completion .616 Picture Concepts .582 Symbol Search .568 Digit Span .525
Poor Measure of “g” Coding .454
Poorest Measure of “g” Cancellation .209
(Keith, Fine, Taub, Reynolds, & Kranzler, 2004)
©2003-2013 Carolyn Kottmeyer
WISC-IV scoring… Harcourt Assessment WISC-IV Technical
Report #4 General Ability Indexhttp://www.pearsonassessments.com/NR/rdonlyres/1439CDFE-6980-435F-93DA-05888C7CC082/0/80720_WISCIV_Hr_r4.pdf
Harcourt Assessment WISC-IV Technical Report #7 WISC–IV Extended Normshttp://www.pearsonassessments.com/NR/rdonlyres/C1C19227-BC79-46D9-B43C-8E4A114F7E1F/0/WISCIV_TechReport_7.pdf
NAGC Position Paper Use of the WISC-IV for Gifted Identificationhttp://www.nagc.org/index.aspx?id=2455
©2003-2013 Carolyn Kottmeyer
How to ID the Highly Gifted?
Not easy!Research shows parents often best identifiersHighly gifted intellectually doesn’t necessarily correlate to high achiever academicallySocially optimal IQ is about 125-140 (new test versions)
©2003-2013 Carolyn Kottmeyer
Gifted AssessmentSchool or PrivateTesting
IQ test (group or individual)
Achievement test (group or individual)
Score report
PrivateTesting & Assessment
Individual IQ test Individual
achievement test Test / subtests for LD
identification Other measures as
neededRecommendations“Assessing Gifted Children” by Julia B. Osborn
www.hoagiesgifted.org/assessing_gifted.htm
©2003-2013 Carolyn Kottmeyer
Gifted AND Learning Disabled
Can this be? Yes!Do these kids need to be identified? YES! They need appropriate work in their
areas of strength, AND appropriate remediation or support in their areas of weakness
GT/LD kids are NOT average!
©2003-2013 Carolyn Kottmeyer
Gifted AND Learning Disabled
How can you identify these students? Combination of achievement and ability
testing, find subtest scatter and discrepancies
Look at subtest scores, not just full scale Verbal score is considered best estimate
of IQ Recommended: WJ-III cognitive and
achievement, in the hands of a gifted/LD-knowledgeable tester
©2003-2013 Carolyn Kottmeyer
ResourcesTesting and Assessmentwww.hoagiesgifted.org/testing.htmAn Inventory of Testswww.hoagiesgifted.org/tests.htmA Parent’s Guide to IQ Testing and Gifted Educationamazon.com/exec/obidos/ASIN/0977109852/
thehoagiesgifted
©2003-2013 Carolyn Kottmeyer
Resource ArticlesWhy Test?www.hoagiesgifted.org/why_test.htmWhat Do Tests Tell Us?www.hoagiesgifted.org/tests_tell_us.htmWhy Do My Child’s Test Scores Vary From Test to Test?www.hoagiesgifted.org/iq_varies.htm
©2003-2013 Carolyn Kottmeyer
Testing and Assessmentof Gifted ChildrenQuestions?
Thanks!
©2003-2013 Carolyn Kottmeyer
Stanford Binet version 5Change sensitive scores (age related) can be used to calculate…Rasch ratio scores Verbally gifted children: Rasch ratio
scores are similar to SB L-M scores Twice exceptional children: Rasch ratio
scores are higher than SB L-M scores, perhaps indicating verbal AND non-verbal strengths (SB L-M was entirely verbal)