Types of Validity and Reliability
by Ramah Myers
0.0 stars - 0 reviews
range from 0 to 5
Types of Validity and Reliability
This is just a demo map that you can delete right away, if you feel like it...
Construct validity means that a test should
correspond to the subject matter it is covering,
and the scores on the test should increase
after teaching the students that subject. The
example Kubiszyn & Borich (2010) give is that
a student would do well on a math test, and
the score would increase after intensive
coaching in math, but not in foreign language.
However, a teacher should expect a student
who has a higher aptitude in whatever subject
is being covered to receive a higher score than
a student who may not be as talented in that
area. This is important because a teacher can
be better prepared for what grade the student
might get, and possibly help the student if they
start to slip. So, if a student who typically gets
a high mark on math tests gets a low mark, the
teacher can investigate further and possibly
help the student.
Criterion-Related Validity Evidence
The first type of criterion-related validity
evidence is Concurrent-Related validity
evidence. This measures the similarity between
a possible new test that might be given and a
well established test. If a new test is a possible
choice for test-makers, the test will be given
with a well established test and the scores will
be compared. If the scores are similar for the
respondents. then the tests are similar enough
to be given as equals. This is important
because this gives educators and test
distributors options, instead of just one test.
(Kubiszyn & Borich 2010)
The second type of
criterion-related validity evidence
is Predictive validity evidence.
Predictive validity evidence is the
ability of the test to predict how the
examinee will behave in the future.
So, this could be applicable on
aptitude tests or psychological
exams. This is important because
if a student is predicted to do well
in the future, changes can be
made to their curriculum (Kubiszyn
& Borich 2010).
Reliability is the consistency at which a test
yields results. A test should also yield the
same or very similar results if the same
student takes the test more than once
(Kubiszyn & Borich 2010).
There are three major types of reliability.
Test-Retest or Stability: What this means is
that the test must be able to have similar
scores if the test is taken twice. This shows
that the scores correlate.
Alternate forms or equivalence: Basically what
this means is that if there are two or more
types of a test that are similar, the different
types will be given and then the scores from
the two tests are compared. If the scores
correlate, it shows that the tests are both
relabile enough to use interchangably.
The third type is Internal consistency. With this,
the examinee is expected to answer
quesetions consistently. If a person is
answering questions, they should be
answering similar questions consistently within
a test. This shows they are reliable.
Content validity tells us whether or not
the test is valid in the content or not,
based on what the instructor wanted to
cover in the test. This means that looking
at the test, we should be able to match
the questions up to the material that was
supposed to be in the test. A test on the
Civil War should not have questions
about World War II on it. This is
important because it allows the student
to study properly and have the
opportunity to do well on the
assessment. Furthermore, it shows that
the eduator is staying in line with the
syllabus. (Kubiszyn & Borich 2010)