Ipsative Scales

Characteristics of Ipsative Scales

Formally, a scale is considered fully or purely ipsative when the sum of the scores of all the assessed attributes equals a constant for all the [Page 881]respondents. As a consequence, all respondents have the same mean score across all the attributes, and the scores on each measured attribute can be understood as deviations from that mean.

Ipsative measures are contrasted in the literature with normative measures that attempt to assess each of the attributes of interest separately, allowing in principle the possibility that a respondent would score high or low in all of them. The person-centered focus of Ipsative Scales makes them suitable only for intraindividual comparisons as opposed to interindividual comparisons.

Ipsative Scales can be constructed through the use of specific item formats, such as forced-response items or ranking items. Ipsative forced-response items can assess the relative preference of the respondents by, for instance, presenting them with multiple options—each associated with one of the attributes being assessed—and asking the respondents to pick among them the one that represents them the most. Ranking items can produce ipsative scores by asking respondents to assign a ranking to all the options presented to them in a question, prompting them to weigh the relative strengths of the attributes represented by each option. Additionally, ipsative scores can be obtained through the use of traditional item formats, such as Likert-type scales, by centering a set of Likert-type items that assess multiple attributes on the mean score for each person (i.e., subtracting the person mean from each item), generating ipsatized data from items originally conceived to produce normative measurements.

Advantages and Disadvantages of Using Ipsative Scales

Ipsative Scales are most often used for the purpose of reducing or controlling self-report bias due to social desirability or halo effects, among others. The idea is that Ipsative Scales can be more robust to these kinds of biases by, for instance, forcing respondents to choose among options with similar levels of social desirability or by asking respondents to rank among multiple options, therefore eliminating the possibility of assigning high scores to all options. In sum, the dependency among the scores of multiple attributes is considered an advantage from this perspective, as it can be used to counter or eliminate certain response biases.

...