Beyond the firm's initial declaration: Are preannouncements of new product introductions and withdrawals alike?

Previous preannouncement research has primarily focused on product preannouncements regarding the firm's intention to introduce a new product and, for the most part, has ignored preannouncements that update the status of new product introductions (e.g., delays in launch dates and cancellation of new product programs). This study's goal is to examine if different factors influence preannouncements of new product introductions (NPIs) versus new product withdrawals or delays (NPWs). A model of six antecedents that could influence a firm's propensity to issue NPIs and NPWs is developed and tested using a sample of 265 CEOs and Presidents from manufacturers of new products. Three of the antecedents are organizational in nature; specifically, first mover predisposition, firm information interactivity, and reputation building. Also, the effects of two environmental constructs, industry innovativeness and competitive hostility, are examined. Finally, the model incorporates the effect of buyer involvement on a firm's propensity to issue NPIs and NPWs. The results indicate that NPIs and NPWs are very alike regarding their antecedent factors. Reputation building, defined as a firm's tendency to pursue a high profile leadership position within its industry, and buyer involvement are the primary motivators of a firm's propensity to issue both NPWs and NPIs. Future directions for research include the development of a normative preannouncement framework and the examination of NPIs and NPWs as nonadvertising forms of marketing communication targeted at numerous audiences such as buyers, employees, channel members, industry influencers (e.g., business and trade related press), and investors.     

Identification structure of linear structural models

In this paper identification problems in linear structural models are discussed. A special but practically important form of these problems, the structural identifiability, is defined. Thereafter it is shown that this problem can be studied — without empirical data — by a set system defined on the set of free parameters of a model. This set system is the identification structure, or the matroid, of the model. It gives insight into the dependence of unidentified parameters. The advantages of this approach in research planning and in correcting nonidentifiable models is demonstrated. Finally we introduce the FORTRAN program LISRAN, which is designed to analyze a given linear structural model and to provide its identification structure.