Introduction to Test Construction in the Social and Behavioral Sciences by Fishman Joshua A. Galguera Tomás. & Tomás Galguera

Author:Fishman, Joshua A.,Galguera, Tomás. & Tomás Galguera [Fishman, Joshua A.]
Language: eng
Format: epub
ISBN: 9781461665588
Publisher: Rowman & Littlefield Publishing Group, Inc.
Published: 2013-06-26T04:00:00+00:00

Guttman Scaling

Another method of providing a metric for items is the method perfected by Louis Guttman (1916-1987). Thurstone’s metric derives from the category assignments of judges, whereas Guttman’s metric derives from what the implications are for the subjects’ responses. Guttman scales have the advantage that a single score allows researchers to predict patterns of response for all the items on the scale. Moreover, when the Guttman scale represents a single dimension, scores can be thought of as moving from general to specific. Though the items on the scale may or may not appear in order, a positive response (i.e., acceptance, agreement) toward a specific item presupposes acceptance of more general items.

Guttman scales are characterized by the “implicational” or “scalable” nature of their items. Correspondingly, tasks that can be successfully completed only when their component subtasks or preconditions are completed in a given order are said to be “implicational” or “scalable” in nature. If one is informed that the student reached step 6, one can be sure that he or she had reached steps 5, 4, 3, 2, and 1 as well. Some fields of knowledge are implicational (e.g., calculus assumes trigonometry, which assumes algebra, which assumes arithmetic). The beauty of Guttman scaling is that its final scores are equivalent to the highest (i.e., most extreme, most advanced) item that subjects have agreed with or answered correctly, and from that score we immediately know all the other items that the subject has agreed with or answered correctly. That score itself implies or presupposes all the others. When this condition exists, the scale is said to be perfectly or fully scalable or implicational.

In the attitudinal domain, some attitudes are scalable as well. In table A6.1, items 2 through 6 scale perfectly, but item 7 is only 83 percent scalable, since from the responses to it one can predict the responses to only five of the six other items. In other words, not everyone who answered item 7 “right” also answered the other items “right.” Therefore, the scale is not perfectly implicational.

Download

Copyright Disclaimer:
This site does not store any files on its server. We only index and link to content provided by other sites. Please contact the content providers to delete copyright contents if any and email us, we'll remove relevant links or contents immediately.