AI@CBS

We propose and test new adaptive question design and estimation algorithms for partial-profile conjoint analysis. Polyhedral question design focuses questions to reduce a feasible set of parameters as rapidly as possible. Analytic center estimation uses a centrality criterion based on consistency with respondents' answers. Both algorithms run with no noticeable delay between questions. We evaluate the proposed methods relative to established benchmarks for question design (random selection, D-efficient designs, adaptive conjoint analysis) and estimation (hierarchical Bayes).

For many years, average bed occupancy level has been the primary measure that has guided hospital bed capacity decisions at both policy and managerial levels. Even now, the common wisdom that there is an excess of beds nationally has been based on a federal target of 85% occupancy that was developed about 25 years ago. This paper examines data from New York sate and uses queueing analysis to estimate bed unavailability in intensive care units (ICUs) and obstetrics units. Using various patient delay standards, units that appear to have insufficient capacity are identified.

This paper studies dynamic competition in markets characterized by the introduction of technologically advanced next-generation products. Firms invest in new product effort in an attempt to attain industry leadership, thus securing high profits and benefiting from advantages relevant for the success of future product generations. The analysis reveals that when the current leader possesses higher research and development (R&D) competence, it tends to invest more in R&D than rivals and to retain its lead position.

The purpose of this paper is to understand buyer/seller adoption dynamics in independent, buyer-side B2B exchanges. In a stylized model, we assume that the main role of the exchange is to reduce search costs for buyers. Buyers and sellers enter or exit the exchange based on the relative economic surplus (loss) they receive inside vs. outside the exchange. We contrast two situations: one where participants' switching cost to join the institution is negligible and another, in which it is significant.

The author studies the pricing of information with private value (e.g., management consulting, legal advice, medical diagnosis). Anecdotal evidence shows that in some of these markets, competing information sellers split the business to sell only first or second opinions to their customers. The author explains this pricing practice by showing that second-opinion markets are a result of temporal differentiation.

The authors investigate two competing hypotheses about how chronic vividness of imagery interacts with the vividness and salience of information in decision making. Results from four studies, covering a variety of decision domains, indicate that chronic imagery vividness rarely amplifies the effects of vivid and salient information. Imagery vividness may, in fact, attenuate the effects of vivid and salient information. This is because, relative to nonvivid imagers, vivid imagers rely less on information that appears obvious and rely more on information that seems less obvious.

The authors propose a new methodology called the "coupled-hazard approach" to study the global diffusion of technological innovations. Beyond its ability to describe discontinuous diffusion patterns, the method explicitly recognizes the conceptual difference between the timing of a country's introduction of the new technology (the so-called implementation stage; Rogers 1983) and the timing of the innovation's full adoption in the country (the confirmation stage).

The authors study global adoption processes where the units of observation are countries, which sequentially adopt a particular technology. The authors’ goal is to provide a better understanding of how exogenous and endogenous country characteristics affect this diffusion process. They develop a general model of global adoption processes, which allows researchers to test extant theories of cross-country adoption, and illustrate the approach using data from the cellular telephone industry for 184 countries.