ABSTRACT
The models that have been considered so far imply that each person has a particular level of a trait which influences the way in which they respond to items in a scale (and how they behave in other ways). Network analysis is different. Rather than looking for traits, it explores how individual items interact – whether the way a person responds to one item influences their response to other items in a questionnaire or test. It displays the interrelationships between these items graphically, and has been used to explore whether items measuring two or more traits (according to factor analysis) form separate clusters, or whether any pairs of items from different traits interact.
