Page 78 - DEDU504_EDUCATIONAL_MEASUREMENT_AND_EVALUATION_ENGLISH

Educational Measurement and Evaluation

Notes these variables can be divided into two categories. For example, if want to calculate correlation
between school attendance and current employment. Here, we shall divide the individuals
into two categories. In the first category will be the individuals who have passed high school
and the second category will have individuals who have not passed high school; on then we
shall divide these individuals into employed and unemployed categories, and then go on to
calculate correlation. In the same way, if we want to see correlation between intelligence and
social maturity, then we will have to make such categories, as above average, below average,
socially mature and socially immature etc. in order to calculate r . t
INTELLIGENCE

Above Average Below Average

MATURITY
Poorly a b
Adjusted
SOCIAL
Well c d
Adjusted

Assumptions : The assumptions of the tetrochoric correlation are as follows :
(a) Both variables X and Y should be continuous.
(b) Both variables should be normally distributed.
(c) Both variables should be lineally related.
Supposing, the following two questions were given to students :
(a) Do you find pleasure in meeting others ?
(b) Do you find pleasure in working with others ?
Here, the assumptions related to continuity and normality will be as follows :
The students who have written Yes in these questions, but have not done it with the same
degree of consistency, and those who have written NO and have not done so with the same
degree of consistency. So, each question presents a continuous variable, and this continuous
scale has consistence on one end and lack of consistency on the other end, or approval on one
end and disapproval on the other end. When a variable is taken as continuous, then NPC can
be taken for a psychological trait.
5. Multiple Correlation Method : This method can be used when more than one scales are used
for some hypothesis. In this, scores of different scales are used jointly in the forms of statistics,
and it is correlated with the third scale in order to know its multiple correlation. Multi coefficient
expresses correlation between one distribution of scores and two or more co-distributions. It
indicates how the co-distributions are correlated with some definite hypotheses.
If we want to select trainees for some training, such as engineering, medicine or teaching etc.,
then we have to predict their success in the given profession; for example, we can use the
intelligence test, study habits test and vocabulary test for admission to college. In the same
way, before buying a car, we shall contact such people who already possess a car or some
mechanic or will test-drive the car, and shall decide about its purchase only when satisfied.
More than one predictor make our prediction more accurate.
Supposing we want to predict success of students in a college. For this, the vocabulary test was
used, and supposing its criterion (test scores) is .5, then it would mean that 25% variance of the
evaluation has been described by the vocabulary test.

72 LOVELY PROFESSIONAL UNIVERSITY