Scribd
Upload
485 views1 page

Nunnally On Reliability: R R R R K

Nunnally (1978) recommends reliability of about .70 or better for instruments used in basic research, and that increasing reliability much beyond .80 is unnecessary. For applied settings where important individual decisions are made based on test scores, reliability should be at least .90, preferably .95 or higher. The formula provided allows calculating additional items needed to raise an instrument's reliability to a desired level, assuming additional items are of similar quality.

Uploaded by

Shanthakumari S S
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOC, PDF, TXT or read online on Scribd
485 views1 page

Nunnally On Reliability: R R R R K

Nunnally (1978) recommends reliability of about .70 or better for instruments used in basic research, and that increasing reliability much beyond .80 is unnecessary. For applied settings where important individual decisions are made based on test scores, reliability should be at least .90, preferably .95 or higher. The formula provided allows calculating additional items needed to raise an instrument's reliability to a desired level, assuming additional items are of similar quality.

Uploaded by

Shanthakumari S S
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOC, PDF, TXT or read online on Scribd
You are on page 1/ 1

Nunnally on Reliability Nunnally (1978, p. 245) recommends that instruments used in basic research have reliability of about .

70 or better. He adds that increasing reliabilities much beyond .80 is a waste of time with instruments used for basic research. On the other hand, with instruments used in applied settings, a reliability of .80 may not be high enough. Where important decisions about the fate of individuals is made on the basis of test scores, reliability should be at least .90, preferably .95 or better. Nunnally (1978, p. 244), shows how to calculate how many additional items one would need to raise the reliability of an instrument to the desired value (assuming that the additional items are as good as the items already on hand).
rd (1 re ) , where rd is the desired reliability , re is the reliability of the existing re (1 rd ) instrument, and k is the number of times the test would have to be lengthened to obtained the desired reliability. For example, suppose you have a 5 item test whose k=

reliability is .66. To raise the reliability to .70, k =

.7(1 66) = 1.2. Thus, you would .66(1 .7)

need a test with 1.2(5) = 6 items. To raise the reliability to .75, k = 1.5, you would need 7 or 8 items. To raise the reliability to .80, k = 2.06, you would need 10 or 11 items.

Nunnally, J. C. (1978). Psychometric theory (2nd ed.). New York: McGraw-Hill.

Reliab-Nunnally.doc

You might also like