# Reliability and Validity Matrix

Essay by   •  May 26, 2013  •  Essay  •  1,447 Words (6 Pages)  •  1,724 Views

## Essay Preview: Reliability and Validity Matrix

Report this essay
Page 1 of 6

Validity and Reliability Matrix

For each of the tests of reliability and validity listed on the matrix, prepare a 50-100-word description of the type of reliability/validity, its purpose and under what conditions these types of reliability and validity would be used as well as when they would be inappropriate. Then, prepare a 50-100-word description of each test's strengths and a 50-100-word description of each test's weaknesses.

TEST of Reliability Description, Purpose, Application and Appropriateness Strengths Weaknesses

Inter-item Consistency Inter-item consistency is the parallel of all items on a scale calculated from one trial of a test. It is used in assessing how consistent various raters and observers are of the same phenomenon. When asking questions to research an idea this test can assess the response of the test-taker against the idea. Different questions that test the same idea give consistent results. This is appropriate for example, in testing employee performance at different times over a period of time. Employers can use this to determine if an employee is eligible for a raise or promotion. Test score calculations are more accurate and clear when there is a lot of consistency. Inter item consistency is great at measuring if a test is reliable and consistent based on the length or shortness of a test. The inter-item consistency test can show reliability over a period of time. On the flip side, errors among items can be broken down and new ones can be added to reach a reliability measurement. Inter-item consistency cannot measure intelligence or personality. If the items are not homogenous with the same difficulty and length, it would be ineffective determining internal consistency. Even the Spearman-Brown formula would fail. Inter-item consistency works best on tests that are whole-test and long in length rather than half-test uses or short test.

Split-half Split-half reliability randomly divides all items that mean to measure the same idea into two sets. When it is difficult to measure reliability with two test or perform a test two times, split half reliability is suitable. It is appropriate with uneven random assignment splits need to be measured. It also can be used to create a small parallel form of the same test.

Split-half reliability has its strength in being efficient and less tedious for test-takers than the parallel form. It measures internal consistency well. It also can check middle variables that may cause an error in the analysis since the both portions of the test are taken at one time. It is not wise to divide a test in half straight down the middle because the content and difficulty of questions will not be distributed evenly. Many intermediary variables are created such as fatigue during the second half of the test. Deviations in difficulty and subjects of the items on the first part of the test compared to the second part.

Test/retest Test-Retest reliability is about taking the same test with the same people and two different times to measure how stable an idea is over time. If an idea being measured is supposed to change over a period then the scores would vary. It is inappropriate when measuring for example, computer skills of college students. A series of lessons about computers would be on the first and second test, then the test would show variance because of the education provided to all testtakers.

Test-retest is strong in reliability because the results measure an individuals reaction time and perceived judgment. Such traits are stagnant and do not change a lot over time and are not sensitive to many intervening variables. Test-retest reliability is weak in that the roots of an idea being tested can alter over time. It would produce sensitive results that make the score of reliability appear lower than the actual measurement. For example, a college student may have excellent skills when assessed on using a HP computer but when assessed on a MAC they could fail or when assessed on a computer from 15 years ago, they could falter.

Parallel and alternate forms Parallel and alternate forms that test reliability use many occurrences of the same test items at two separate times with the same test-takers. It is appropriate in measuring traits that are stagnant over a long period of time and not effective when measuring limited emotions or anxiety levels. Parallel forms can be done with another form such as split-half.

...

...