Internal Consistency Reliability is a subdivision of the test into two or more equivalent parts. In internal consistency reliability, we judge the reliability of the instrument by estimating how well the items that reflect the same construct yield similar results.
One method we can use is split-half. In split-half reliability, we randomly divide all items that purport to measure the same construct into two sets. How?
First, divide the test into half usually using the odd-even technique. As shown in the illustration below, the items were divided into odd-even numbered items, where odd-numbered items (1, 3, 5) forms the "x" group while even-numbered items (2, 4, 6) forms the "y" group.
Second, find the correlation of scores using Pearson r formula. To do this, make a table as shown in the example below.
Example: Four pupils took a 50-item test. Below are the results
Example: Four pupils took a 50-item test. Below are the results
The table above shows that odd-numbered items are denoted by (x) while even-numbered items are denoted by (y). The rows in the two columns show the scores of the four (4) pupils who took the 50-item test. For example, pupil 1 scored a total of 30 out of 50 in the test. When divided into odd-even numbered items, it shows odd=10 and even=20. Likewise, pupil 2 scored only 13 out of 50 in the test where odd-numbered items=5 and even-numbered items=8.
Moreover, the first two columns are given based on the result of the test while the last three columns need to be solved. For the third and fourth columns, all you have to do is to square the x and y variables. The last column is the product of x and y.
Finally, the last row provides the sum of all the columns. Once you have all this data you can now substitute the Pearson r equation. Just follow the example below.
Third, adjust & re-evaluate correlation using Spearman-Brown formula
The purpose of re-evaluating the correlation is to determine the reliability of the test as a whole. Since we divide the test into half, the result of the correlation in step 2 using Pearson r is only half of the test (r½). To determine the reliability of our test as a whole we use Spearman-Brown formula. Sample computation is shown below.
Analysis: the result above shows that 0.98 is closer to +1 hence the test is of excellent reliability. As a guide to the to interpret the coefficients of stability that are between 1 and 0, refer it below:
- 0.9 and greater: excellent reliability
- Between 0.9 and 0.8: good reliability
- Between 0.8 and 0.7: acceptable reliability
- Between 0.7 and 0.6: questionable reliability
- Between 0.6 and 0.5: poor reliability
