The main objectives of this study are to identify the constraints on shifting freight in New Zealand (NZ) from road to rail and/or coastal shipping, and to quantify the trade-off between factors affecting shippers’ perceptions, to assist in increasing the share of freight moved by non-road transport modes. This was done by three logistic regression methods. The ranked logit results show that NZ shippers ranked transport time as the most significant constraint upon distributing goods by rail, while accessibility and load size were the most significant constraints upon using coastal shipping. The study also identifies how NZ shippers’ modal shift constraints vary according to the firm’s individual or logistical characteristics (e.g. their use of logistics facilities, lead time, and delivery distance). Mode choice models, consistent with econometric theory and based on transport cost, time, reliability and modal frequency, are developed. Multinomial logit (MNL) and mixed logit (ML) models are estimated, to identify the factors influencing the choice between road, rail and coastal shipping, for domestic inter-island freight flows. Finally, the models are used, with empirical data on transport cost, time and reliability, to estimate the effect (on mode split) of policies to alter the values of these variables.