1.1.2 How does Magnitude Estimation work in principle?

Experiments have shown that “human beings are not really good at making absolute judgements, e.g. saying whether a line is 10 or 15 cm long” (Hoffmann, Chapter 5). Therefore, Magnitude Estimation makes use of a very simple principle: informants are asked to give relative judgements, e.g. they have to judge “whether a given line is longer or shorter than a reference line and try to express this difference in numerical terms (saying e.g. that a stimulus is half as short as the reference line)” (Hoffmann, Chapter 5). Linguists apply this relatively simple method to language (cf. Bard et al. 1996; Featherston 2004, 2005; Keller 2000; Keller and Alexopoulou 2005). Informants are “asked […] to give numerical judgements on sentences proportional to a constant reference sentence” (Hoffmann, Chapter 5). In more detail, this means informants first encounter a reference stimulus, i.e. a sentence to which they have to assign a number (cf. line length: reference value). In a next step, subjects are exposed to an experimental stimulus (another sentence) which they have to rate in proportion to the reference stimulus. Informants then rate the remaining experimental sentences one by one. Since they need to express all their judgements in relation to the reference stimulus, they must have access to the reference sentence and the reference value they assigned to it throughout the experiment.

Created with the Personal Edition of HelpNDoc: Full-featured Documentation generator