CHANG, AI-HWA; PHD
UNIVERSITY OF ILLINOIS AT URBANA-CHAMPAIGN, 1992
BUSINESS ADMINISTRATION, MARKETING (0338)
This dissertation proposes a new diffusion model for an innovation in durable
goods markets and
compares the model with four existing models. The distinctive feature of the
new diffusion model is an
explicit consideration of the competition between an existing product and a
new product due to the
comparative advantage of the new product and also due to the closer introduction
times of the existing
product and the new product. Adopting an individual-level modeling approaching,
this diffusion model
incorporates the phenomena of: interaction of evaluations for successive generation
innovations,
risk-adjusted utility of the new product, increased new product utility over
time, and the Bayesian process
of uncertainty reduction. The proposed model has been compared with existing
models using both
empirical data from the color TV market, and computer-simulated market data.
The comparison criteria are
(1) goodness of fit in model estimation samples, and (2) forecast accuracy in
model validation samples.
The results of empirical comparison show that the proposed model performs better
than the compared
models on both criteria. The results of computer simulation comparison show
that the proposed model
and Lattin and Roberts' (1989) RAU model are the two best models with the latter
being better than the
former on goodness-of-fit in seven of twenty-eight market types, and the former
being better than the
latter on the accuracy of one-step-ahead sales forecasting in twenty-six markets.
In addition, in markets
characterized by strong consumer resistance, the proposed model is superior
to the RAU model on both
performance criteria. The proposed model is also superior to the Bass model
(1969), a
first-order-autoregressive model, and the Lattin and Roberts' (1989) model with
price consideration in all
the markets examined. Besides, the superiority of the proposed model to these
three models increases
more often in markets in which diffusion-deterring factors are present. Therefore,
the proposed model
has demonstrated its superior capabilities in modeling the diffusion of a new
product in an environment
characterized by stronger consumer resistance. Managerial implications of the
proposed model in new
product prelaunch forecasting and post-introduction adoption diagnosis are also
discussed.
Social
Systems Simulation Group
P.O. Box 6904 San Diego, CA 92166-0904 Roland Werner, Principal Phone/FAX (619) 660-1603 |