Latent Class Modelling (LCM) comprises a set of techniques aimed to model situations where different subgroups (or, more generally, classes) of entities (for example organizations or individuals) are present in data and group membership is not directly observable, but has an impact on phenomena of interest.
Latent Class Modelling permits to handle with unobservable heterogeneity, which represents a crucial problem for data collections such as RISIS. Heterogeneity mainly relies on differences between observation units (HEIs, people, organization) in terms of individual characteristics, organization, country factors or time. At the same time, heterogeneity is also an interesting source of information about the population investigated, and it is thus of a certain interest to model it.
With Latent Class Modelling individuals can be classified into mutually exclusive types (classes) using uncovers hidden patterns of association than can exist between observations (latent and not).
For inquiries and support on Latent Class Modelling in RISIS you can contact Barbara Antonioli Mantegazzini at the Università della Svizzera italiana in Lugano.