“Technique for the Measurement of Attitudes, A”

Often, researchers in the social sciences would like to quantitatively describe or explain people's attitudes or beliefs about an issue. The problem is that attitudes or beliefs—such as “hope,” “customer satisfaction,” or “conservatism”—are qualitative and cannot be directly measured the way weight, height, or speed can be measured. This is the problem that Rensis Likert and others were up against in the early 20th century, and which Likert addressed in his article “A Technique for the Measurement of Attitudes.”

Until 1932, the best procedures available for measuring attitudes were those developed by Louis Thurstone. In Thurstone's approach, as many as 200 experts, called judges, each rated a large number of statements for their favorableness toward a certain position or object. Through a process of elimination, 20 to 30 statements would be retained and ordered along a scale representing the attitude toward the position or object, from negative through neutral to positive. Based on its position on the continuum, each statement was then assigned a scale score. Thereafter, to measure a person's attitude toward that object or issue, he or she would be asked whether they agree with each statement. The person's attitude score was obtained by taking the average scale scores of all the statements that person agreed with.

Likert identified two shortcomings in Thurstone's procedure. First, it required many judges and a long, laborious process. Second, the use of the attitude scales made several statistical assumptions that had not been verified. In his 1932 article, “A Technique for the Measurement of Attitudes,” Likert introduced a streamlined approach to improve on Thurstone's methods. His main goal was to simplify the process by creating the scale from participants’ responses and thus eliminating the influence of the judging group. Whereas previous methods engaged separate processes for assigning scale scores to statements and attitude scores to individuals, in Likert's study, each statement became a scale in itself, and a person's responses to each statement were assigned a score. These response scores were then combined by using a median or mean to obtain an attitude score.

Although Likert's 1932 article is most noted for the scales that carry his name, the issue he tackled was not purely a matter of simplifying measurement but of determining whether in fact social attitudes could be measured reliably and validly. If so, it should be possible not only to measure a general attitude but also to justify its difference from other attitudes. Likert's 1932 study showed, to his own surprise, that general social attitudes could be measured validly and reliably.

This remainder of this entry describes the development of the Likert scales and examines their shortcomings.