Both simple frequencies
and relative frequencies appear as numerical and pictorial representations.
These statistical formats appear in the context of several
informational environments, namely sample size, number/percentage
of buyers in the sample and supply costs in the different versions of
the managerial problem. Further assumptions are that two sample
sizes exist (two different numbers/percentages of buyers) and that the
market has asymmetric costs (either over-supplying costs or under supplying
costs).More over, the choices present three levels of difficulty.
The impact of each statistical format in decision making undergoes
quantification in terms of both the time required to make decisions
and the accuracy of these decisions. Third, in this experiment, the number
of factors to consider, namely the statistical formats and the different
informational environments, is quite large. Specifically, five factors
receive attention (statistical formats, sample size, number/percentage
of buyers, costs and responses). Three of these factors have two levels,
one factor has three levels and the remaining factor has four levels.
The possible combinations of these factors define 96 different choices
for the subjects. Such a high number of decisions might overwhelmsubjects
or require several sessions with different subjects. An orthogonal
design whereby the number of alternatives reduces to 25 (a quarter of
the total number) helps resolve this problem, thus allowing the experiment
to run in just one session and with the same pool of subjects.
The methodology of orthogonal design allows for a reduction in the
number of decisions without substantially lowering the capacity for
analysis (mainly related to the interaction order between factors).
Thus, this approach also offers a contribution to the experimental design
methodology.