Reliability & Validity

Reliability & Validity

www.keydifferences.com

Activity 1:

1. What is meant by ”test reliability”, and why is test reliability important?

Answer: The reliability of a test is defined as the extent to which the results can be considered consistent or stable. It is important to be variable because an unreliable test might produce wildly different scores if it was taken again, and you could not be sure that you had placed the students in the right level.

2. Mention the three ways of estimating test reliability, and describe each!

Answer: Three ways of estimating test reliability:

a. Test-retest reliability: is usually estimated by administering the same test twice to one group of subjects and then calculating a correlation coefficient between the pairs of scores from the two administrations.

b. Equivalent forms reliability: is estimated by administering two equivalent tests ( say forms A and B of a test) to one group of subjects and calculating the correlation  coefficient between the scores and on the form.

c. Internal consistency reliabilities: it can be estimated in a number of ways, but the easiest method to understand conceptually is the split-half method.

3. What is meant by “test validity”, and why is it important?

Answer: it is defined as the degree to which a test measures what it claims to be measuring. It is important because it should test that should be test.

4. Mention the three types of test validity, and how is each one established?

Answer: three types of test validity:

a. Content validity: the test developer must answer the question, is the test a representative sample of the content of whatever the test is claiming to test?

b. Construct validity: it is related to psychological construct. It is a theoretical label that is given to some human attribute or ability that cannot be seen or touched because it goes on in the brain.

c. Criterion-related validity: it is basically just a subset, or variation, of the notions discussed under construct validity.

Activity 2:

Read Tuckman (1978:160-165), and answer the first four questions of the competency test exercise on pg. 193 (attached)

1. Match up the items on the left with those on the right:

a. tets-retest reliability	2. scores at time A versus scores at time B
b. alternate forms reliability	4. scores on form A versus scores on form B
c. split-half reliability	1. odd versus even items across test takers
d. Kuder-Richardson reliability	3. direct comparison of item scores by formula

2. Describe procedures for determining test-retest reliability as compared to split-half reliability.

Procedures for determining test-retest reliabilityis to give the same people the same test on more than one occasion and then compare each person’s performance on both testings. In this procedure, known as test-retest reliability, the scores obtained by each person on the first administration of the test are related to his or her score on the second administration to provide a reliability coefficient. While procedures for determining split-half reliability is splitting a test into two halves, usually the odd-numbered items and the even-numbered items, and then correlate the scores obtained by each person on one half with those obtained by each person on the other. This procedure, which yields an estimate called the split-half reliability, enables a researcher to determine whether the halves of a test are measuring the same quality or characteristic.

3. Match up the items on the left with those on the right.

a. predictive validity	3. test relates to behavior it is presumed to predict
b. concurrent validity	4. test relates to another test of the same thing
c. construct validity	2. test of concept relates to hypothetically related behavior
d. content validity	1. test adequacy samples from total range of relevant behaviors

4. Describe procedures for determining the concurrent validity of an I.Q. test as contrasted to its construct validity.

One of the concurrent validity procedure of a test is to relate on the test with performance on another, well-reputed test (if such exist), for instance, Intelligence test which are often validated concurrently by comparing performance on a newer, more experimental one with performance on an older, more established one. However, construct validity is established by relating a presumed measure of a construct or hypothetical quantity with some behavior or manifestation that it is hypothesized to underlie (the construct measured by the test attempts to explain the behavior).

References:

1. Brown, J.D. 1988. Understanding Research in Language Learning. Cambridge: CUP.

2. Nunan, D. 1989. Research Methods in  Language Learning. Cambridge: CUP.

3. Saleh, M. 2001. Pengantar Praktik Penelitian Pengajaran Bahasa. Semarang: IKIP Semarang  Press.

4. Tuckman, B.W. 1978. Conducting Educational Research. London: Harcourt Brace  Jacobovitz.

 Baca Juga: 

Research Instruments   Data   Variables   Research problems & hypotheses  Introduction (What is & why do research?) Sampling Techniques