Social-psychological variables are typically measured using either cumulative or unfolding response processes. In the former, the greater the

No abstract

No abstract

Joint maximum likelihood estimation (JMLE) for the hyperbolic cosine model (Andrich & Luo, 1993) is extended to the situation in which the units of items are allowed to vary. Four issues were considered in developing this JMLE procedure: (1) the indeterminacy caused by the unit parameter of each item, (2) a necessary constraint to eliminate this indeterminacy, (3) the initial estimates of the unit and location parameters of the items, and (4) a practical correction procedure to account for inconsistency in parameter estimates. Four estimation cycles were designed to address these issues. Results from two sets of simulation studies (ideal and non-ideal situations) indicated that the algorithm yielded reasonably accurate recovery of the parameters.