Direct Effects of City Logistics Measures and Urban Freight Demand Models - Sustainable Urban Logistics: Concepts, Methods and Information Systems - page 216

4.4 A New Shop Type Choice Model
This model gives us the probability (p[k/so]) of purchasing freight type s in a
outlet of type k (i.e. small, medium and large) departing from zone o. As there are
few examples of shop type choice models (Gonzalez-Benito 2004 ), below we
present some models for the simulation of this choice. Besides, although they can
refer to different categories of end-consumer i, for simplicity of notation, the class
index i will be understood unless otherwise stated.
The results are based on surveys carried out in Rome where more than
300 households were interviewed, considering both home-based and non-home-
based shopping trips.
From survey data analysis, it emerged that the choice of retail outlet mainly
depends on freight types. Different multinomial logit models for the choice of
retail outlet types were then calibrated according to the four main identified freight
types: foodstuffs, hygiene and household products, clothing and shoes, other
products.
In the following four tables, the multinomial logit models calibrated for the four
freight types are reported. Our analysis reveals that foodstuffs and hygiene/
household products are bought at all three different types of retail outlets, while
clothing and other products are purchased at small or large retail outlets. Hence,
for the latter freight types, only two alternatives are considered. All parameters are
correct in sign and are statistically significant as shown by t-st and p-values, while
values of q 2 are similar to those present in the literature for models of this type (i.e.
discrete choice models).
From the calibration reported in Table 2 for foodstuffs, it emerges that the
probability of making a purchase in small retail outlets increases if the purchase is
made in the early morning (i.e. before 11 am), on Saturday and the customer is a
woman. Increasing the money spent, the number of goods types to buy and the size
of the group, the probability of customers buying in larger retail outlets increases.
The probability of buying in a large retail outlet increases if the customer is
younger than 29, while it decreases if the trip starts from home. The results
confirm that many people travel to shop together or to buy many items and that
large retail outlets are preferred if time for shopping is available.
As regards hygiene and household products (Table 3 ), the probability of pur-
chasing in a small retail outlet increases if the customer is housewife, while the
probability of shopping in large retail outlets increases if the trip is made to buy
many products, if the customer is young, has much time available and/or travels
with other people. As the time allocated to purchasing increases, the probability of
choosing a medium retail outlet decreases. This result shows the inclination of
customers to choose this type of outlet for rapid shopping for already-chosen
products, such as washing-up liquid or soap powder. The results also confirm that
younger customers travel to larger retail outlets (e.g. to find special discounts and
because they have more free time).