Page 283 - DEDU504_EDUCATIONAL_MEASUREMENT_AND_EVALUATION_ENGLISH

Unit 24 : Question Bank

these educational objectives into test items. This is essential because the educational objectives Notes
may differ from program to program. We have intentionally left the discussion of the third
domain i.e. psychomotor. Psychomotor domains deals with behaviours like gross bodily
movements, finely co-ordinated bodily movements, non-verbal communications and speech
behaviours which are more suited for the practical components of the exam.
Computerised question banks
One essential activity for the “On demand” examination system is the preparation of question
banks. For this, let us try to relate the type of questions with the educational objectives. The table
below shows the domains of learning outcomes and the question type, which will measure them
effectively.
Table : Taxonomy of educational objectives vs. question types

Domains Question Type
A : Cognitive
1. Knowledge True-False type, Multiple Choice,
2. Comprehension Matching, Short Answer type
3. Application
4. Analysis Essay type
5. Evaluation
B : Affective
1. Interests
2. Values Essay type
3. Attitudes

So, one can easily classify the question paper in three sections i.e. very short answer types, short
answer types and long answer types. These sections will comprise of the following.

Very Short answer type ⎯⎯→ True — False type Multiple choice, Matching

Short answer type ⎯⎯→ Short questions

Long answer type ⎯⎯→ Essay type questions

So, while preparing question banks one has to take care of these three sections. An ideal question
bank will have all three of questions and when one attempts to draw a set of question paper from
it, the only thing he has to decide is the number of questions to be taken from each section.
Using simulation and random number generators
Parallel sets can be made very easily with the help of random number generators. Arithmetic
methods, particularly suited for computer operation are available for generating random numbers.
The most common method is with the help of Linear Congruence Generator.

Z = ( Za i – 1 + ) c mod m, a,c,m,Z are + ve integers and a,c,Z ≤ xm
1
0
0
The choice of numbers, depends upon the seed, which initialise the recursive computation. To
make the calculations simple and to maintain a uniformity we have a set of 20 random numbers
taken from Uniform (0,1) distribution on six columns.

LOVELY PROFESSIONAL UNIVERSITY 277