DIRECTIONS: READ THE FOLLOWING LEARNER POST AND RESPOND. PLEASE CITE REFERENCES APA STYLE.
Discussion : Meaning of Reliability Scores
According to Cohen, Swerdlik, and Sturman (2012), "The reliability coefficient helps the test developer build an adequate measuring instrument, and it helps the test user select a suitable test." In this discussion example, three scores are given that calculate the reliability of test THING.
The first score given is an internal consistency reliability coefficient, which estimates the reliability of a test's internal consistency without having to make an alternate test or have someone take the same test twice by evaluating the extent that test items relate to each other (Cohen, Swerdlik, &Sturman, 2015).
The second score is the alternate forms reliability coefficient, which evaluates the reliability of different forms and their consistency with the original test. The test-retest reliability estimates the reliability of a single users score on the same test taken at different times.
Individual Results of Error Variance
The internal consistency reliability coefficient score of .92 is about an "A" grade in terms of the test's reliability and tells the test user that the test is internally consistent (Cohen, Swerdlik, &Sturman, 2015).
The alternate forms reliability coefficient score of .82 is a low "B" grade of coefficient. This grade tell us that the alternate test forms are mostly reliable in terms of the extent to which they differ from the original and have been affected by sampling errors or other possible errors (Cohen, Swerdlik, &Sturman, 2015). The test-retest reliability coefficient score of .50 tells us that a test takers scores largely varied on the same test taken at different times.
Synthesis
In most cases the internal consistency reliability coefficient of .92 would be acceptable, however we do not know the content of what the test is measuring. If this test were measuring a life or death situation, then perhaps, even a very high internal consistency reliability coefficient of .92 would not be high enough, and would require a coefficient of .95 or higher (Cohen, Swerdlik, &Sturman, 2015).
The alternate forms reliability coefficient is a low "B" grade, or a "B2" grade coefficient and tells us that alternate test forms THING 1 and THING 2 are mostly consistent with the original test. Again, the criteria and purpose of the specific test would determine whether this coefficient would be acceptable or not.
The test-retest reliability coefficient is very low, however the importance of this score depends on the content of the test and the purpose of the test measurement. For example, if this test were measuring stability over time, then a higher score would be necessary.
However, if the test is measuring something that is assumed to fluctuate over time, the test-retest reliability coefficient is not as important, because the test taker's score would be expected to be variable (Cohen, Swerdlik, &Sturman, 2015).
Reference
Cohen, Ronald J., Mark Swerdlik, Edward Sturman. Psychological Testing and Assessment: An Introduction to Tests and Measurement, 8th Edition. McGraw-Hill Learning Solutions, 07/2012. VitalBook file.