RESEARCH METHOD

75 Semantic Differential

Semantic differentials can help reveal “felt” meanings that are a direct product of one’s experiences, culture, and dearly held beliefs.¹

The semantic differential scale is a linguistic tool designed to measure people’s attitudes toward a topic, event, object, or activity, so that its deeper connotative meaning can be ascertained. Although used in marketing surveys to evaluate products and services, its original intention was to measure social attitudes by exposing the outer limits of a semantic space. Its recent popularity is probably due to its straightforward format: the respondent is asked to indicate where on a continuum a concept is best described. For instance, given the concept of “Art:”

Much care must go into the design of an effective semantic differential in order to yield useful results. Each of these components should be considered before conducting a semantic differential:²

Concepts The stimuli of the semantic differential, concepts can be a topic, event, object, or activity. Concepts should be carefully chosen based on research objectives, and should be meaningful to respondents.

Bipolar Word Pairs Usually, pairs of antonyms are selected as the polar ends of a semantic differential scale. They can be complementary antonyms (e.g., pleasant—unpleasant) or more nuanced, gradable antonyms (e.g., the opposite of friendly isn’t necessarily unfriendly; shy or guarded could be a more meaningful opposite). Poles should be randomized so that negative and positive connotations don’t consistently fall on the same side.

Survey Scale It is common to see six-and seven-point scales, but the seven-point scale is preferred because it provides a neutral midpoint. A neutral answer could indicate apathy, indecisiveness, or that the concept is socially irrelevant, all of which are meaningful judgments. The distance the rating is from the midpoint reflects the intensity of the judgment.

Dimensions for Classification All bipolar word pairs belong to a dimension of classification. Osgood et al. recommend three dimensions to classify concepts: evaluation (e.g., valuable—worthless), potency (e.g., strong—weak, heavy—light), and activity (e.g., active—passive, excitable—calm).³

After multiple concepts are assessed against the same dimensions, the semantic differential between concepts can be mapped. The differences in where concepts are mapped in a semantic space reflect their differences in connotative meaning.

1. The Semantic Differential Scale (SDS) was pioneered in 1957 by Charles Osgood, George Suci, and Percy Tannenbaum. The methodology and theory were documented in their book The Measurement of Meaning, and has since been used extensively in language attitude studies. See:

Osgood, Charles, George Suci, and Percy Tannenbaum. The Measurement of Meaning. Urbana, IL: University of Illinois Press, 1957.

2. Al-Hindawe, Jayne. “Considerations when Constructing a Semantic Differential Scale.” Dissertation: Linguistics Program at La Trobe University, 1996.

3. See note 1 above.

4. Bartneck, C. “Who Like Androids More: Japanese or US Americans?” Proceedings of the 17th IEEE International Symposium on Robot and Human Interactive Communication, 2008.

Further Reading

Williams, Frederick. “The Identification of Linguistic Attitudes.” International Journal of the Sociology of Language 3, no. 1 (1974): 21–32.

Semantic differential scales are particularly powerful when eliciting cross-cultural attitudes and perceptions to the same stimuli. In Christoph Bartneck’s research study “Who like Androids More: Americans or Japanese,” eight semantic differential scales were used to investigate the degree to which a person’s cultural background influences one’s perception of a robot’s anthropomorphism and likeability.⁴ The experiment used static pictures of 18 different robots (like the iCat and Geminoid HI-1, above) as the stimuli for the study.

Courtesy of Christoph Bartneck