Semantic Differential Scaling

Initial Research

As psychology developed as a science, researchers began to use rating scales to measure people’s attitudes and beliefs. Thus, to place Osgood’s development of the semantic differential in [Page 1504]historical context, it is instructive to briefly consider other prominent measurement techniques that were proposed in the 1920s to 1940s. Most notably, the Thurstone, Guttman, and Likert methods assess people’s level of agreement with a series of unique questions about a particular topic. For example, on a Likert-type scale, research participants could be required to answer numerous, differently worded questions about various aspects of the Democratic Party on a strongly disagree to strongly agree scale.

Figure 1 Example of a series of semantic differential items used to evaluate attitudes toward the Democratic Party

In the early 1950s, Osgood and several other scholars proposed the semantic differential as a technique for measuring the meaning that people place on particular concepts. In direct contrast to the contemporary measurement techniques mentioned earlier, Osgood’s semantic differential relied on a standardized set of items, on which participants can rate nearly anything. Thus, a key benefit of semantic differentials is that they are less taxing on research participants and researchers than other methods.

Several early investigations found that semantic differentials could efficiently capture the changing nature of social stereotypes. These initial studies also revealed that people’s judgments fall into the three dimensions noted earlier—evaluative, potency, and activity. Moreover, the findings from Osgood’s work provided initial evidence that semantic differential scales were relatively objective, reliable, and valid ways of measuring a wide range of concepts.

Updates and Current Uses

A key result of early research on semantic differential scaling was that compared to the potency and activity dimensions, evaluative questions revealed the most about people’s overall assessment of an object or person. A great deal of research confirms this result, as evidenced by the publication of over 500 studies, on a range of topics, from all over the globe. Thus, for researchers interested in people’s attitudes, of which evaluation is a major component, using semantic differentials is a key measurement tool.

...