[Limdep Nlogit List] Bass-type diffusion of innovation models

[Fred Feinberg]

Another way to do this, if your probabilities are on a 'grid' (i.e., they
have been expressed, for example, as 0%, 10%, 20%, ..., 90%, 100%) is to
treat them as ordinal, and use an ordered probit or logit model. This isn't
strictly correct from a modeling perspective, but at least it accounts for
more shades of gray between 0% and 100%.

A better way is to use Rost's (1985) rank-ordered binomial model. I'm not
sure if LIMDEP natively supports this (Prof. Greene would know). I've
estimated the model from scratch using MATLAB, and also natively using MLWin;
the likelihood isn't difficult to write down, and the outcomes are
probabilities, so the dependent variable type is correct. It also
accommodates 0% and 100% responses.

Fred

=====

Prof. Fred Feinberg
Ross School of Business
University of Michigan
feinf at umich.edu

Erik Ferguson wrote:

> Upananda--
>
> Since the '0' and '1' options represent infinity in terms of relative
> utility, tobit would work, assuming you have intermediate values between
> 0 and 1 for partial adoption, but this is not particularly elegant from
> a mathematical perspective.
>
> You could also define three (or more) 'choices' representing no adoption
> (F=0), full adoption (F=1), and partial adoption (0<F<1), and use an
> ordered probit or logit model to estimate the equation. If none of your
> values were 0 or 1, you could estimate choice probabilities directly a
> la Theil (1970).
>
> Is your data for the degree of adoption at one moment in time, or do you
> have multiple data points? If you have time series data, you could
> estimate one of several Bass-type diffusion of innovation models.
>
> I was wondering if LIMDEP can be used to estimate continuous Bass models
> using MLE procedures?
>
> For the discrete Bass model, deltaAt = pm + (q-p)At-1 - q/mAt-1^2. I
> have estimated this model in LIMDEP by manually creating the three
> primary variables (Y,X1, X2), all of which are simple transformations of
> a single time series.
>
> Is there a simple way to enter a set of say 85 independent time series
> variables, and estimate the Bass model without manually creating 3*85
> separate transformed variables?
>
> Thanks!
>
> Erik Ferguson
> Master of Urban Planning Program
> School of Architecture and Design
> American University of Sharjah
> PO Box 26666
> Sharjah, UNITED ARAB EMIRATES
> eferguson at aus.edu
> 971 6 515 2878 office
> 971 6 515 2800 fax
> 971 50 369 8355 mobile
>
> Paragahawewa, Upananda wrote:
> > Hi
> > Could you please tell me whether Tobit model would applicable in the
> > following case or some other options will better suit?
> >
> > A given technology has been adopted fully by some individuals; to some
> > extent by some individuals, and not at all by some. (example, a new
> > clone of a crop grown in entire land area available in some case (1) ;
> > part of the available land in some cases (0.7) and not at all (0) but
> > still practice the old crop in some cases). Thanking you in advance
> >
> >
> > Kind Regards
> >
> > Upananda
> >
> >
> >
> > =======================================================================
> > Attention: The information contained in this message and/or attachments
> > from AgResearch Limited is intended only for the persons or entities
> > to which it is addressed and may contain confidential and/or privileged
> > material. Any review, retransmission, dissemination or other use of, or
> > taking of any action in reliance upon, this information by persons or
> > entities other than the intended recipients is prohibited by AgResearch
> > Limited. If you have received this message in error, please notify the
> > sender immediately.
> > =======================================================================
> > _______________________________________________
> > Limdep site list
> > Limdep at limdep.itls.usyd.edu.au
> > http://limdep.itls.usyd.edu.au
> >
> >
> >
>
> _______________________________________________
> Limdep site list
> Limdep at limdep.itls.usyd.edu.au
> http://limdep.itls.usyd.edu.au