Thurstone Scaling

Two-Step Scaling and the Thurstone Approach

As in any scaling exercise, the beginning step is to operationalize the construct of interest and several statements that help to define that attitude such as beliefs, feelings, and/or behaviors are written. These must be reviewed by judges (in this case, the participants in a scale development study) in the first part of the two-step approach to scaling. The goal of this step is not to obtain the actual opinions of the participants toward the construct of interest. Rather, what are needed here are their judgments about the favorableness of those statements on that construct, to identify how items differentially reflect specific levels of an attitude or trait. Typically, many more statements than are strictly necessary are written, as some winnowing of statements might occur in the scale development process to create the final measurement instrument from the assortment of scaled stimuli.

Next, taking the statement ratings across judges, the stimuli can then be sequenced on a subjective continuum from least to most favorable. Positions on this continuum should span the full range from strongly negative through neutral to strongly positive. Then, for the last piece of scale assembly, the statements to be included are deliberately sequenced in no particular order; that is to say, they are not sorted by their scale values.

Once the scale is assembled in this manner, the second step of scaling occurs through administration of the scale itself, as people's endorsements (or not) of the statements on the scale allow for positioning the respondents to the survey on the attitude continuum. The attitude score for a person is the mean (or median) scale score of the stimuli that the individual agreed with. An important point to be made here is that the actual scaling portion of Thurstone scaling in fact references three distinct approaches to the judges’ task of rating the stimuli, as described below.

...