Test of Reliability Application and Appropriateness strengths weaknesses

TEST OF RELIABILITY APPLICATION AND APPROPRIATENESS STRENGTHS WEAKNESSES

Internal Consistency Internal consistency is best used when wanting to conclude the difference in reliability from lengthening or shortening a test With this type of measurement there are many intermediary variables that are brought about during the testing of reliability (Cohen & Swerdlik, 2010). The Spearman-Brown Formula is best used to help determine internal consistency (2010).

The more test items that are present causes the reliability of the test to increase (Cohen & Swerdlik, 2010). The Spearman-Brown Formula can help determine the level of reliability by shortening or lengthening the test (2010). This formula can also be used to help the researcher indicate how many items they should insert to reach the ideal reliability coefficient (2010). The Spearman - Brown Formula has issues with internal consistency because it is only operational with similar test items (Cohen & Swerdlik, 2010). Reliability are also higher for whole-test compared to half-test applications of the Spearman-Brown formula, the longer tests are more valid with this approach (2010).

Split-half Split-half is used when two scores are computed from comparable halves of one test given (Cohen & Swerdlik, 2010). This test is best used to measure reliability when there are restraints on time and/or expenses (2010). The test is best administered by assigning odd or evens to one of the items and the other item have the other items (2010). It would be inappropriate to use this test if a person only wanted to measure one item and no more than that.

If the test is split and the three steps are followed then the test will accurately be able to measure the two items with two different scores (Cohen & Swerdlik, 2010). This measure is a lot less burdensome and less time--consuming for the test taker than the corresponding form and it also takes a good measurement of internal consistency (Cohen & Swerdlik, 2010). Since both corresponding forms are taken at once, it helps keep intermediary variables that might bring about error variance in check (2010). If the test is not split appropriately it could cause a lowering in the reliability of the test (Cohen & Swerdlik, 2010). With this type of measurement there are many intermediary variables that are brought about during the testing of reliability (Cohen & Swerdlik, 2010). Fatigue is a possible variable that might be brought about for the second part of the test that could have an effect on the test.

Test/retest Test/retest is used to measure reliability estimate by gathering relating pairs of scores from the same person on two different test administrations of the same test (Cohen & Swerdlik, 2010). This would best be used when wanting to evaluate the reliability of a test that is measuring something that is moderately stable throughout time (2010). It would be inappropriate to measure something that tends to fluctuate (2010).

The strength of test/retest is that it helps engage measures of outcome such as perceptual judgment (Cohen & Swerdlik, 2010). The strength is because these types of traits (psychometric) do not diverge very much over time and are not sensitive to many outside variables that could sway the traits (2010). The major weakness of this approach of measurement is that the main constructs that are being tested can and may eventually alter with time which will cause the reliability of this measurement to lower (Cohen & Swerdlik, 2010). The true variance is a result of the aforementioned instead of error variance like other type of measurements (2010).

Parallel and alternate forms Parallel and alternate forms of reliability use numerous occurrences of the same test items at two separate times with the unchanged test takers (Cohen & Swerdlik, 2010). These approaches would be appropriate with measurements of traits that are unchanging throughout time. These approaches would not be appropriate when using anxiety levels or anxiety levels (2010).

This measurement is best to help measure the core construct through numerous modifications of the same test item(s) (Cohen & Swerdlik, 2010). If comparable results are located on several forms of the test item, then the reliability of the test will increase (2010). To help save time and money the researcher could use internal consistency to estimate reliability (2010). The test -taker would not have to take multiple tests if the aforementioned method is used. The major weakness for this measurement is that test-takers can become tired from having to