IFCS 2009 conference location:Dresden date:13-18 March 2009
Summary. In this paper we will focus on the modeling of binary data regarding
individual differences in the emotions that people experience in specific situations.
Underlying such data, emotion psychologists typically assume a sequential process
with two links: situations activate specic appraisals in persons (link 1); subse-
quently, specific patterns of activated appraisals elicit the experience of particular
emotions (link 2). It is further hypothesized that these two sorts of links may differ
across persons. An important challenge then consists of retrieving the place and
the nature of the key individual dierences in the process under study. To meet
this challenge, Ceulemans & Van Mechelen (2008) recently introduced the CLASSI
model. However, the CLASSI model requires the persons and situations to be fully
crossed, implying that each person has to rate the same set of situations. This is
a major restriction since not all situations are equally relevant for every person.
To overcome this restraint we propose an extension of the CLASSI model which
permits the set of rated situations to dier across the persons, implying that the
situations are nested within persons rather than crossed. Like the original CLASSI
model for crossed data, the new CLASSI model for nested data (1) reduces the sit-
uations, appraisals, emotions, and persons to a few types, and (2) defines linking
structures between the situation types and the appraisal types on the one hand and
the appraisal types and emotion types on the other hand, which represent individual
differences in these two sorts of links.