Transcript Assessment - SpecialEducationOntario

Assessment
Assessment is the collection, recording and
analysis of data about students as they
work over a period of time. This should
include, teacher, self and peer assessment
Assessment Can Be:
• Diagnostic
• Formative
• Summative
Types of Assessments
•
•
•
•
•
•
•
•
Individual
Group
Formal
Informal
Norm- Referenced
Criterion-Reference
Teacher made
Commercial
Other Types of Assessment
• Dynamic
• Assisted
• Trial
• Authentic
Evaluation
Evaluation brings meaning to information
by making judgements about achievements
and progress in relation to expectations for
children of a particular age and grade.
Validity
Validity means that the test actually
measures the skills and knowledge
that it is designed to measure.
Reliability
Reliability in terms of standardized tests
means that if the same student were to take
the same test again within a short space of
time, they would get the same results. In
other words the timing or place of the test
has minimal impact on the results.
Measures of Reliability
Measures of reliability are generally reported as a
coefficient. 0 means the test has no reliability and
1 means a perfectly reliable test. A coefficient of
0.8 or higher is acceptable for most normreferenced tests.
Standard error of measurement
This is the statistical estimate of how
the results would vary if the test were
given again.
Normal Distribution
Norm referenced tests are created on normal distributions.
This means that about 68% of students will achieve a score
within one standard deviation of the mean. About 96% will
fall within two standard deviations, at either end and almost
100% will fall within three standard deviations.
Criterion Referenced Tests
Students are compared to a standard
or criterion.
E.G. Students will be able to add double digit
numbers with regrouping
Norm Referenced Tests
In a norm-referenced test students
performance is ranked according to other
students in the norm-group. Norm groups
are usually by grade or age.
Percentile
A student scoring at the 50th percentile would
mean that the student did better than 50% of
students his or her age. It gives no indication of
how many test items the student answered
correctly or incorrectly.
Grade Equivalents
These scores are highly misleading and must be used with
extreme caution. A score of 4.5 on a math indicates that the
student has done as well as the average student in mid-grade
four, this doesn’t mean that the student solved a mid-grade
four problem.
Normal Curve Distribution
While %ile scores are not evenly distributed,
normal curve equivalents are. They are calculated
on a 1-99 scale on how well students in a
particular category will do. So unlike %iles they
can be combined to get a total test score.
Stanines
Stanines rank students on a 1-9 scale, with
1,2 and 3 being below average, and 7, 8
and 9 being above average. 4,5 and 6 are
average.
Scaled Scores
Scaled Scores rank the student from the lowest
possible level of achievement on a given test to the
highest possible level of achievement. They are
calculated on a 0-999 scale and are used for long
term tracking.