The swim-with-dolphin industry was established in New Zealand in 1989. It is managed by the Department of Conservation under a permitting  system designed to control the level of commercial marine mammal based tourism. Swimming with dolphins is allowed, but it is illegal to  harass or disturb dolphins, or to swim with groups containing  juveniles.

In 1997, there were almost 60 commercial operations taking people out  to swim with primarily dusky, Hector's, bottlenose, and common  dolphins. Our research focused on bottlenose and common dolphins in  the Bay of Islands (RC), Hector's dolphins in Porpoise Bay (LB), and  dusky dolphins in Kaikoura (SY). We used both land- and boat-based  methods to study the effects of the swim-with-dolphin industry. Land-based research using a theodolite is a non-invasive technique  allowing for accurate measures of boat and dolphin speed and  direction. It is ideal for data collection on large groups (up to  several hundred dolphins) and on group behavior. However, theodolite studies do not allow for detailed behavioral observations on an  individual level. To date, boat-based studies on impacts of  swim-with-dolphins have concentrated on detailed group behavior, but  data on individual dolphin behavior can also be collected.  Unfortunately, boat-based studies can be invasive and do not allow for measurement of avoidance reactions to boats at distances. In New Zealand, boats are most often the platform from which  swim-with-dolphin operations take place. It is therefore important to  also consider boating impacts when evaluating the overall effects of  swim-with-dolphin operations.

R. Constantine began boat-based MSc research on the bottlenose and  common dolphins in the Bay of Islands in 1994, and is currently  completing a PhD on the bottlenose dolphins. Results from 1994-95  showed dolphin response to swimmers was affected by swimmer placement  from the commercial vessels. Common dolphins were less likely to  interact with, and more likely to avoid swimmers than bottlenose  dolphins (24% vs 48% interaction, and 38% vs 22% avoidance). Recent  results from research on bottlenose dolphins comparing data from  1994-95 to 1997-98, have shown a significant decrease in the success  of swims, and an overall increase in avoidance responses to swimmers.  Photo-identification has shown an 82% resight rate of bottlenose  dolphins in the Bay. These dolphins are not resident in the Bay but  have had a number of exposures to swim attempts over the past six  years and therefore it is likely that this population has become  sensitised to swim-with-dolphin tours.

S. Yin conducted theodolite and boat-based acoustic research on dusky dolphins in Kaikoura from 1994-1997 for her MSc. Results on the  effects of boat presence on small groups (<25) of dolphins peripheral  to the main group, suggested that the presence of the tourist vessels  had not changed the overall mean speeds since Cipriano and Wursig's pre-tourism research in the 1980's. Caution is necessary as there was a great deal of variation in mean leg speed recorded within and between groups of animals which suggests that this variable may not have been an adequate measure of disturbance. Linearity of groups was higher with boat and post boat conditions which suggests dolphins may meander more and move in a less directed manner when no boats are nearby. The linearity and reorientation rates both revealed slight differences between no boat and boat condition.

L. Bejder conducted land-based theodolite observations to quantify Hector's dolphin group reactions to boats and swimmers in Porpoise Bay from 1995-1997. 
The commercial dolphin-watching operation in Porpoise  Bay does not offer swim-with-dolphin encounters. Instead,  non-commercial swim-with-dolphin attempts occur off the beach.  Dolphins were accompanied by swimmers (within 200m) for 11.2% of the  total observation time (251 hours), whereas boats accounted for an  additional 12.4%. Analyses of relative orientation indicated that  dolphins tended to approach the dolphin-watching boat in the initial  stages of an encounter, but became less interested as the encounter  progressed. By 70 minutes into an encounter, dolphins were either  actively avoiding the boat or equivocal towards it. Dolphin groups  were significantly more tightly bunched when a boat was present in the bay. Forty two percent of swim-with-dolphin attempts were classified  as at least potentially disturbing (dolphins left the 200m vicinity of swimmer entry either immediately or within 5 minutes of swimmer  entry).

There are a number of practical limitations when conducting research  on the effects of swimming with dolphins (e.g. geographical  limitations of research site, temporal variations in habitat use of  study animals). Furthermore, the duration of data collection is often compromised by time constrains set forth by management committees.  Research designs should incorporate methods of data collection that  allow for rapid detection/indication of possible impacts. The short  term focus should be incorporated into a long term research plan that  emphasises long term cumulative impacts. Management recommendations  based on short term research should be highly precautionary.

Currently, only the dusky dolphin population of Kaikoura have received detailed study before being targeted by commercial operations.  Unfortunately for all other species and populations currently exposed  to tourism, "before and after" comparisons are impossible. Research  designs must account for the level of tourism established and the  feasibility of collecting control data, the latter is often difficult  to achieve. Future research designs should, where possible, strive to  include focal individual observations and multiple data collection  strategies to assess impacts. Simultaneous data collection of acoustic and behavioural reactions (on a group and individual level) will allow for a more complete assessment of impacts. Furthermore, a more  comprehensive approach will highlight how data collection methods can  affect conclusions about impacts on dolphins and thereby also  management recommendations.