In split half method, two scores are obtained for each person by dividing the test into equivalent halves. To find split-half reliability a procedure that is adequate for most test purpose is to find the scores on the odd and even items of the test.

There are several formulas that are used widely to find reliability in split-half method. For example,

Mosier formula:

Mosier offers a short-cut computing formula . it requires the scoring of only one of the parts. The formula is

Here,

Roe=reliability.

Correlation between odds score and total score

=SD of odds scores

=SD of total scores

Spearman-brown formula:

Spearman brown prophecy formula is used to measure split half reliability. Split-half reliability is used in single test, consisting of two parallel forms-odd items and even items. Each of which measure the same things. We may administrate a test and assign separate scores to every participant on two arbitrarily selected half of the test. For example, a participant may be given score of odd items, second score on the even items. Then the correlation between two score is a parallel form reliability coefficient. We can assume that two halves are equivalent. We can use the spearman brown formula

Here,

r (1/2 1/2)= correlation between odd and even items.

The Kuder-Richardson Formulas:

The Kuder-Richardson Formula 20 (K-R 20):

Here,

n = number of items

p= proportion of correct response to each item in turn

q= 1-p

sigma suare t= total variance of test.

Again,

Here,

n = number of items

Pi =proportion of examinees getting item I correct

S suare= variance of the total test score.

The Kuder-Richardson Formula 21(K-R 21)

Here,

n = number of items

sigma suare t= total variance of test.

p= proportion of correct response to each item in turn

q= 1-p.

and are average of p and q for n items

Again,

### Like this:

Like Loading...