Semantic Differential

A special type of graphic rating scale, the semantic differential, has come to be used widely in marketing research. Perhaps its main use has been in connection with brand and company image studies; it permits the development of descriptive profiles that facilitate comparison of competitive items.

For example, Figure shows the profiles of national brands, private brands, and generic products as they were developed with this technique.

The unique characteristic of the semantic differential is the use of a number of bipolar scales, such as those shown above, to rate any product, company, or concept of interest. Respondents are given a group of these scales and asked to check on each one the point that indicates their opinion of the subject in question. As originally developed, each scale consisted of two opposing adjectives, such as good/bad and clean/dirty, which were separated by a continuum divided into seven segments. Each segment was separated by a continuum divided into seven segments. Each segment was assumed to represent one step in moving from the meaning of the adjective at one end to that at the other. Respondents checked the segment that represented the degree of the characteristic involved that mostly closely coincided with their opinion of the product or other item being rated. There is no reason why the space between the opposing adjectives must be divided into seven segments, but that has become the most common. In one project, the entire scale was omitted and respondents were asked to pick one or the other of the polar adjectives that most clearly coincided with their opinion of the subject. Many researchers now use an even number of segments to force respondents to take a position. The neutral midpoint in a semantic differential seems to attract many respondents, but the results from scales of odd and even numbers differ only slightly.

Although the original semantic differential scales could presumably be used for rating any items, most researchers develop their own scales for specific projects. For example, in a study to determine how people viewed a picture that was to be used in an advertisement, the following pairs of adjectives were used:
Rich – Poor
Intelligent – Unintelligent
Worried – Unworried
Honest – Dishonest
Healthy – Unhealthy
Likable – Unlikable
Wholesome – Unwholesome
Cheerful – Sad

The opposing adjectives used in a specific project are determined by the objectives of that project. In the above example, the advertisement was for an insurance company particularly concerned that the man pictured should not be thought to be worried or sad. Opposing adjectives must be devised to be readily understood as opposites by respondents.

The basic concept of the semantic differential has been modified in various ways. Figure shows three variations all reported to have produced highly reliable data.

Semantic differential scales can be used to obtain total attitude scores. Pairs of adjectives must be selected that are relevant to the attitude to be measured. In a study to determine the relative favorability of consumer attitudes toward two brands of the same product, pairs such as fair tasting / excellent tasting, average value for money / excellent value, and very healthful / not very healthful were used. The adjectives were separated by a six segment scale; the least favorable location on each scale was given a value of 1, the next least favorable a value of 2, and so on. After all the scales had been completed for each brand, these values were summed to obtain an attitude score for each brand. It should be noted that the paired adjectives in this project did not represent opposing extremes in each cases. It was apparently reasoned that few people would consider any food product on the market as less than fair tasting. The use of a total score obtained in this manner as an indication of the overall attitude toward a brand assumes that all the important factors that determine consumer attitude toward a brand have been included and they are all of relatively equal weight. A positive correlation was found between attitude scores so determined and brand purchases. The student will note that this methodology assumes the scale is an interval scale rather than the ordinal scale it really is.

When the semantic differential is used to develop an image profile, it provides a good basis for comparing images of two or more items. It is best when used for image descriptive purposes and is not recommended for overall attitude measurement. The one big advantage of the semantic differential is its simplicity, while producing results comparable with those of the more complex scaling methods. The method is easy and fast to administer, but is also sensitive to small differences in attitude, highly versatile, reliable, and generally valid.