Page 85 - DEDU504_EDUCATIONAL_MEASUREMENT_AND_EVALUATION_ENGLISH

Unit 7 : Reliability – Types , Methods and Usability

Suppose, a cloth length has been measured at 3 m by a scale. Now the question arises, is this scale Notes
reliable ? If we measure that cloth length by any other scale and its length comes to 3 m, then we
can easily say that it is reliable. However, if the cloth measures 3 m at one instance and 3.5 m at
another instance, then this scale cannot be called reliable. The same thing applies to mental tests
too. If a mental test measures a child’s IQ at 110, and the same scale measures it at 110 again, then
we shall call the first test as reliable. It is clear that by reliability of a test is meant is constancy
of the score of measurement. It refers to the consistency of the measurement. That is, if the test is
repeated under the same set of known conditions and the obtained scores should remain the
same, then the test is said to be reliable.
Supposing, an individual is given a test today, and then the same test is administered to him after
an interval of time, and his scores are equal in both cases, then the test will be said to be reliable.
In the same manner, if a test is administered on several individuals repeatedly, and its scores are
equal like the first instance, then the test will be considered reliable.
According to Anastasi, “The reliability of a test refers to the consistency or scores obtained by
the same individuals on different occasions or with different set of equivalent items.”
Thus, reliability is that attribute of a test which is responsible for making its results equal. On the
contrary, if the scores of a student or a group of them change sufficiently, then the reliability of
this test will be considered to be low. The more reliable is a test, the more consistent will be the
scores of a student or a group of them. It is not necessary that these scores will be exactly equal
under all equal circumstances, there can be minor variation between them due to some other
causes, but this variation should not be excessive. In theory, reliability is considered the measurer
between observed scores and true scores. According to Guilford, an obtained score is the sum of
true score and error score. That is :
X = T + E
Where, X = Obtained score
T = True score
E = Error score
This thing has been explained by Helmesteder using the following formula :

t × e
X = i i
i
Where, X = Score by a certain student
i
T = True score by a certain student
i
E = Error score by a certain student
i
7.1.2 Reliability Defined in Terms of Variance
The formula used for reliability X = T + E tells about the relation of obtained score, true score and
error score; however, because reliability is an attribute of testing, not of the scores of a certain
individual; so the following formula can prove to be more meaningful :
SX = ST + SE 2
2
2
Where, SX = Total variance
2
ST = True variance
2
SE = Error variance
2
That is, on the basis of the above formula, total variance is found in two types of variances : true
variance and error variance. True variance is found due to systematic sources, that is, personal
distinction of individuals; and error variance occurs due to unsystematic sources, such as guess,
recording error or scoring error etc.

LOVELY PROFESSIONAL UNIVERSITY 79