Page 134 - DEDU504_EDUCATIONAL_MEASUREMENT_AND_EVALUATION_ENGLISH

Educational Measurement and Evaluation

Notes For example, if one hundred students were administered the test, then the first and third
part will have 33 students and the second part will have 34 students. The second part will be
covered with a white sheet and stapled. Now, the total of the right responses in the first part
and third part will be written in the white sheet. Now, the number of students in the third
part who have attempted a particular item correctly will be deducted from the number of
students in the first part who have attempted that particular item correctly. The difference
thus obtained will be divided by N/3. The coefficient thus obtained will be called its
discrimination index. Thus, discrimination index will be calculated for each item, and it will
be written in the last row of the chart.
Following is the formula for calculating discrimination index :
−
TB
V. I. =
N/3
Where, T = Top group
B = Bottom group
N = sample size
6. Now, each item of the test in the chart is studied to see whether its difficulty index is
between 20-80 or 10-50 as per the objective of the test and whether its discrimination index
is above .3 or not. The items which pass these two criteria are selected. As many items are
selected and written separately as are needed for the test. Thus, the final form of the test is
readied, which is further verified for its reliability, validity and norms in order to make it
standardized.
Some scholars do not take one-third of the samples as above, but they take only top 27%
and bottom 27% of the sample. It is tabulated which tells the above requirements easily.
According to the opinion of Kelly, the formula of top-bottom 27% is considered valid and
suitable because on its basis it can be decisively said whether the top group is superior and
the bottom group is inferior. If a smaller top-bottom group is taken, then these groups are
too small to decide something. Still, there is no definite rule in this regard. The higher is the
number of candidates in the top-bottom group, higher will be the discrimination index.
External Criterion for Item-analysis : In the above discussion, the basis for validity of each
item was internal, that is, it was accepted as the criterion measure of the whole test, and its
correlation was calculated in respect of the whole test. This type of assumption and logic is
rational. On this basis, we can say that each item evaluates that trait which is evaluated by
the test as a whole, but we cannot say how far it evaluates that trait which it ought to.
Generally we take an external criterion in order to ascertain that each item in the test is
evaluating the trait for which it has been included in the test, and thus each item is validated
on the basis of that external criterion. In the view of Ryens, no specific assistance is offered
by the external criterion.
Steps In Item Analysis
Let us take an example of teacher who has administered a 30 item multiple-choice objective
test in geography to 45 students in a secondary school classroom.
Step 1 : Obviously, upon receiving the answer sheet, the first step would to mark each of
the answer sheets.
Step 2 : Arrange the 45 answer sheets from the highest score obtained until the lowest
score obtained. The paper with the highest score is on top and the paper with the
lowest score is at the bottom.

128 LOVELY PROFESSIONAL UNIVERSITY