The behaviour implied by these results indicates an intriguing decision calculus, in which the proportional allocation of freight tasks across modal alternatives is aligned with proportional demands across the range of cost/level-of-service trade-offs available. That is, while the study involves a range of attribute levels, and hence the nature of the trade-offs is not identical in com- position across the choice sets faced by a given respondent, structural differ- ences in the qualities of each alternative may play a large role in the allocation of freight activity across modes in the study. The ability to allocate choices proportionally may be important here, as a strict (non-representative) binary choice could be expected to lead to divergent implications. Regardless of the specification of the choice variable, the appearance of a unique subset of influential independent variables within the estimated utility functions for each alternative highlights important behaviour in relative preferences for each mode. This is supported through the inability to identify a significant sensitivity to service frequency for any mode within the study. While one may reasonably expect higher frequencies of departure to be of value, the choices of respondents were not sensitive to the range of service frequencies on offer. Hence, this indicates the presence of a multi-tiered choice calculus in which general frequencies of service (for example, high-frequency road freight, mid-frequency rail freight and low-frequency short sea freight) are considered in the allocation of cargo across modes, but where, conditional upon this choice, variations within general levels of frequency do not impact demand patterns. That is, respondents have indicated a preference for modes that include some critically high level of service frequency, without further significant sensitivities to different frequencies on either side of their per- ceived threshold value.
The behaviour implied by these results indicates an intriguing decision calculus, in which the proportional allocation of freight tasks across modal alternatives is aligned with proportional demands across the range of cost/level-of-service trade-offs available. That is, while the study involves a range of attribute levels, and hence the nature of the trade-offs is not identical in com- position across the choice sets faced by a given respondent, structural differ- ences in the qualities of each alternative may play a large role in the allocation of freight activity across modes in the study. The ability to allocate choices proportionally may be important here, as a strict (non-representative) binary choice could be expected to lead to divergent implications. Regardless of the specification of the choice variable, the appearance of a unique subset of influential independent variables within the estimated utility functions for each alternative highlights important behaviour in relative preferences for each mode. This is supported through the inability to identify a significant sensitivity to service frequency for any mode within the study. While one may reasonably expect higher frequencies of departure to be of value, the choices of respondents were not sensitive to the range of service frequencies on offer. Hence, this indicates the presence of a multi-tiered choice calculus in which general frequencies of service (for example, high-frequency road freight, mid-frequency rail freight and low-frequency short sea freight) are considered in the allocation of cargo across modes, but where, conditional upon this choice, variations within general levels of frequency do not impact demand patterns. That is, respondents have indicated a preference for modes that include some critically high level of service frequency, without further significant sensitivities to different frequencies on either side of their per- ceived threshold value.
การแปล กรุณารอสักครู่..
