12.1 Reflective and formative indicators

Psychologists mainly work with unobservable entities (e.g., depression, impulsiveness, intelligence), not with the items or scales (i.e., measures). The items of a given test are the indirect way to measure the underlying psychological construct (i.e., a latent variable). The domain covered by any psychological construct is sampled through items. Consequently, a scale (e.g., 9 items measuring job burnout) intends to reflect the underlying psychological construct (i.e., job burnout). Latent variables are regarded as causes of item scores. The magnitude (i.e., strength) of the latent variable is assumed to cause an item to take on a certain value. Sometimes, some subdomains emerge (i.e., factors). These subdomains or factors are identified using iterative conceptual and empirical analyses (e.g., exploratory factor analysis). In sum, observed measures (e.g., items) reflect the latent variable (Figure 12.1).

Figure 12.1: Reflective and formative indicator models.