Assessment of Students with Special Education Needs

1
Assessment of Students with Special Education
Needs

• SPE 3374
• Margaret Gessler Werts
• 124D Edwin Duncan Hall
• 262 6365
• pm.appstate.edu/wertsmg

2

• What is a child with special needs?
• Why do we spend so much time trying to identify
  one when most people can spot one at a distance?

3
Assessment

• Why does your teacher assess you and your work?
• What does the assessment have to do with what you
  know?
• Which assessments are useful and which are not?

4

• Of what use are tests in any class?
• Of what use is the Praxis?
• What purpose does the Praxis and other
  standardized tests have for the test
  interpreters?

5
Overall Purpose of assessment Gathers
information to assist in decision making 

• 1.     Screening
• To determine whether there is a reason or
  suspicion that there is a reason for continued
  testing.
•  
• (Why wouldnt we just test anyway?)

6

• 2.      Monitoring students' progress. Assessing
  students formatively allows the teacher the
  opportunity to determine whether or not a student
  is making adequate progress toward desired
  instructional outcomes.
• 3.      Assigning grades or ranks.
• Assessing students provides the teacher with at
  least one type of evidence to assist in assigning
  grades.

7

• 4.      Determining one's own instructional
  effectiveness.
• Assessing students can provide teachers with
  information related to teaching effectiveness. If
  students are achieving the desired instructional
  outcomes, then teaching is likely, although not
  definitely, effective.

8

• 5.      Influencing public perceptions of
  educational effectiveness. Students' scores on
  standardized achievement tests and statewide
  assessments are often used to inform the public
  as to the effectiveness of schools and teachers.

9

• 6. Clarifying teachers' instructional intentions.
  
• What a teacher intends on assessing influences
  what is taught, and vice versa. Also, national
  and statewide assessments often feedback to
  influence a state's, school's, or teacher's
  curriculum.

10

• Definition
• Educational assessment is the systematic process
  of gathering educationally relevant information
  to make legal and instructional decisions about
  the provision of services.

11
Assessment should

• 1.     Relate to regular performance
• 2.     Gather from interdisciplinary sources
• 3.     Consider all aspects of the student's life

12
Basic terms

•   Assessment A general term used to denote the
  systematic collection and interpretation of data
  that is to be used in the making of educational
  decisions, including enhancing instruction.

13
More terms

•     Measurement The explicit quantification of
  the results obtained through testing that is
  designed to determine the degree of presence or
  absence of a particular characteristic, quality,
  or feature.
• Test An instrument or formal process that
  presents tasks that yields a specific type of
  measurement.

14
More terms

• Evaluation The placing of value or
  interpretation of a measurement. Evaluation
  encompasses the inferences that are drawn from
  measurements

15
Types of Assessments

• Formal
• Informal
• Observations
• Authentic

16
INFORMAL STRATEGIES.

• 1.     Observations.
• 2.     Curriculum based assessments.
• 3.     Work sample analysis.
• 4.     Portfolios.
• 5.     Task analysis.
• 6.     Inventories.

17
INFORMAL STRATEGIES.

• 7.     Criterion referenced tests.
• 8.     Diagnostic teaching.
• 9.     Checklists and rating scales.
• 10.            Interviews.
• 11.            Rating scales.

18

•                   FORMAL STRATEGIES
• Norm referenced tests- compare a students
  performance to that of a comparable group
•    Use limited to groups who are much like the
  norm group
• Very explicit instructions

19

• Two types
• Achievement ?
• Aptitude ?
• Uses a variety of score reporting types
• Raw scores, standard scores, percentiles,
  stanines, t scores, z scores, etc.

20
Formal tests

• Norm referenced tests- compare a students
  performance to that of a comparable group.
• 1.     group.
• 2.     individual.
• Use limited to groups who are much like the norm
  group.

21
Characteristics of formal tests

• Very explicit instructions.
• Uses a variety of score reporting types.
• Scores are comparable to other scores
• Some scores are better for this use than others

22
How do we know if a test is good for our purposes?

• Look at the norming sample.
• Look at the validity.
• Look at the reliability.
• Look at the standard error of measure.
• Where do we find this information?

23
The Examiners Manual

• The manual is your best friend!
• It contains
• Information about the concepts to be tested.
• Administration procedures (who should give it,
  basic test procedures, times, basals and
  ceilings).

24
Manual (cont.)

• Specific subtest administration instructions
• Script to read
• How to score (usually with examples)
• Explanations of reasons for the scoring
• Examiner qualifications
• How much training
• How much practice
• How to encourage and praise
• Time needed

25
Manual (cont).

• How to calculate the age of a subject
• How to record the subtest raw scores
• How to compute composite scores
• Tables for computing scores
• How to complete the graphics for interpretation

26
Administering items

• Some tests require that you administer all items
• Some have ranges of items that you administer
• The appropriate point to start is usually given
  in the test manual.

27
Basals

• A basal score is the point at which it may be
  assumed that all lower items would be scored as
  correct
• The student establishes a basal by responding
  correctly to a predetermined number of items

28
Ceilings

• The point at which it may be assumed that all
  harder items will be scored as incorrect
• Student establishes a ceiling by responding
  incorrectly to a predetermined number of items.

29
Scores

• The assignment of a number or level to an
  attribute or characteristic of a person.
• Raw scores.
• Number right and number wrong
• Obtained Score The score (measurement) that a
  student receives as a result of completing a
  particular test.

30
Derived scores Scores that are weighted or
otherwise standardized.

• Standard scores.
• Those that have been transformed by statistics
• Standard deviation.
• How far a randomly drawn score will be from the
  mean

31
More Scores

• Z scores
• Mean 0
• T scores
• Mean 50
• Percentiles
• Rank of people
• Stanines
• Standard nines

32
Error Score

• The portion of a student's obtained score that is
  due to factors not associated with the student's
  actual level of achievement, such as guessing,
  poorly written questions, tiredness, and
  motivation.

33
Systematic Errors

• Systematic errors consistently raise or lower the
  obtained score for all (or most) of those tested,
  (e.g., faulty testing procedures, questions that
  give away the answer, misunderstanding the
  directions

34
Random Errors

• Random errors are the unpredictable sources of
  error that cause fluctuations in students' scores
  that are generally beyond the assessor's control,
  such as student fatigue, guessing, and
  motivation. We refer to these errors as "bad
  spaghetti days in the cafeteria."

35
True Score

• A true score is a score that is free from error
  and accurately conveys a student's true level of
  achievement. In practice, a true score is the
  hypothetical average resulting from taking the
  same test repeatedly. There is a minimal chance
  of ever achieving a true score and is minimal
  chance that you will know which score is a true
  score.

36
Validity

• The accuracy and appropriateness of the
  interpretations and inferences (evaluation) drawn
  from the results of a test. "Does this
  instrument measure what it says it does?"

37
Absence-of-Bias

• The absence of any characteristic associated with
  an assessment that might offend or unfairly
  penalize those being assessed and thus distort
  (depress or inflate) a student's score. "Does
  this instrument provide this examinee with a fair
  chance?"

38
Reliability

• The consistency of results (measurement) obtained
  from an assessment, based on the control,
  reduction, and/or elimination of measurement
  error. "Does this instrument measure the
  construct resulting similarly over time, parts of
  the test, administrators, etc.?"

39
Measuring Reliability (Reliability Coefficients).

• Measuring Stability.
•         Test-Retest a correlation between two
  successive measurements using the same test.

40

• Measuring Alternate-Form.
•       Alternate Form a correlation between two
  successive measurements using two parallel forms
  of the same test.       Inter-rater a
  correlation between two persons rating the same
  performance using the same scoring criteria.

41

• Measuring Internal Consistency.
•         Split Half a correlation between two
  halves of a test, obtained by dividing the test
  into two equal halves (e.g., odd questions vs.
  even questions).

42
Types of Reliability.

• Stability The consistency of test scores across
  different testing instances, using the same test.
  (i.e., consistency over time).
• Alternate-Form The consistency of test scores
  when comparing two supposedly equal forms of a
  test (i.e., consistency between tests).
• Internal Consistency The extent to which test
  items function consistently, assessing the same
  construct (i.e., consistency within a test).

43
Standard Error of Measurement.

• An estimate of the amount of error in a test
  score. The SEM is the standard deviation of the
  errors obtained in a series of measurements and
  is derived from the reliability coefficient. That
  is, how much can one expect a score to vary if
  the test were given again.
• where s standard deviation of the test scores
  and r reliability coefficient of the test.

44

• Since obtained scores contain error, the SEM can
  be used to construct confidence intervals or set
  "reasonable limits" on score interpretation
  regarding a student's true score. These
  intervals, or bands, are created by adding and
  subtracting the SEM to and from the individual's
  obtained score. That is, if a student has an
  obtained score of 80 on a test that has an SEM of
  3.0, then the teacher can be confident that the
  student's true score is somewhere between 77 and
  83.

45

• F. Estimating the Standard Error of Measurement
  (see Hopkins, 1998)
• Classroom teachers, most likely, will not know
  the SEM of the tests that they make. The SEM can
  be estimated given the number of questions that
  are asked, if each question on the quiz or test
  is worth 1 point. Remember, this is an estimate
  only.

46

• Using this method of adding and subtracting the
  SEM to and from the individual's score to
  determine "reasonable limits" is an effective way
  of determining whether two scores differ from
  each other in a meaningful way.
• For example, if Mica has an obtained score of 81
  and Liza has an obtained score of 86, is that a
  real difference?

47

• But if we know that the SEM for the test that
  they took was 4.0 then Mica has a confidence band
  of 77 to 85, while Liza has a confidence band of
  82 to 90. Since these bands represent the areas
  in which their true scores are likely to fall,
  and these bands overlap, the conclusion would be
  that their obtained scores of 81 and 86 are not
  meaningfully different.

48

• Obtained scores contain error. Combining the SEM
  and the normal curve can yield confidence
  intervals related to the true score.
• Approximately 68 of measurements fall between
  1 standard deviations from the mean.
• Approximately 95 of measurements fall between
  2 standard deviations from the mean.
• Approximately 99.7 of measurements fall between
  3 standard deviations from the mean.

49

• Thus, if Charles has an obtained score of 77 on a
  test with an SEM of 3.5 we can be
• 68 sure that his true score is between 73.5 and
  80.5 ( 1 SD),
• 95 sure that his true score is between 70.0 and
  84.0 ( 2 SD)
• 99.7 sure that his true score is between 66.5
  and 87.5 ( 3 SD).

50
Differences between error assumptions