ABSTRACT
This chapter helps people to understand: the utility of multiple-item scales; bipolar semantic differential scales; applications of nonverbal pictorial scales. The chapter includes the rank ordering of ordinal scales. The general purpose of most scales is to examine a variable across some range of possible values. Scales involve multiple questions or items designed to measure the same underlying concept or construct. To illustrate the concept of scale reliability, let's turn to some actual data collected by advertising agency Doyle Dane Bernbach (DDB) Needham. These data are part of an annual survey known as the Life Style survey. Most people are familiar with Likert-type scales, which present a statement and then ask respondents to rate their agreement with ratings such as "Strongly agree". As with Likert scales, semantic differentials are measured at least at the ordinal level of measurement. Pictorial scales use images as rating points rather than words.
