70 operationally deployed dog–handler teams, 166 randomised double-blind field trials with standardised four-way starts (single-trail and scent-discrimination scenarios). Directional decisions at the start point were consistent with chance across all conditions, and in discrimination trials teams reached a wrong person more than twice as often as the target. Scope is limited to initial trail selection — not full operational trailing — but it is the most rigorous negative evidence to date.

Dogs proficient in directional tracking were tested on trail segments where available footsteps were experimentally restricted. Dogs required odour from about five sequential footsteps to determine trail direction, supporting a mechanism of comparing odour intensity/age across successive steps within seconds of sniffing.

22 police dogs each worked 10 human-laid trails to test direction determination. Only 8 dogs (36%) were consistently correct; males and younger dogs performed better. The 8 proficient dogs still identified true direction when trails were laid walking backwards, showing they used odour cues (not footprint shape) to read direction.

Experimental study of trained tracking dogs’ directionality on human-laid trails. Tracking dogs were tested on straight tracks 200 paces long to determine to what extent they would go in the direction a track-layer had walked when brought in at 90° angle at the mid-point of a track. None of the dogs tested had any previous experience with this task. A total of 66 tracks were run. In 60.6% of these, the dogs tracked in the direction the track-layer had walked. These results are not significant enough to say that the dogs tested were going in the direction of the track-layer any more frequently than one would expect from random chance (P> 0.05). Thirty-two percent of the dogs exhibited a position preference with regard to the direction they first turned at the decision point. However, the preferred direction was not the same for all these dogs. As a group, the dogs showed no overall position preference (P > 0.05).

Six dogs trained on Odor print training aids for TATP had a 51% detection rate on real TATP, compared to 87% for 5 dogs trained on real TATP.

The study tests if working dogs trained to indicate on pseudo training aids for explosives will also indicate on real explosives, and vice versa. Two brands of pseudo training aids are tested. The study was conducted on 18 dogs with no previous experience of the odours in question.

Training the dogs on pseudo odours did not result in detection of actual explosives, nor did dogs trained on actual explosives show interest in the pseudo odours.

The use of electronic shock collars (“e-collars”) is one of the most controversial topics in dog training. In this study, we compared e-collars to methods relying entirely on food rewards in order to stop dogs running after a lure. We found that dogs receiving shocks from e-collars stopped chasing a lure within two sessions of ten minutes of lure running per session. These dogs also refrained from chasing the lure in three out of four test sessions. Two groups of dogs trained with food reward did not refrain from chasing the lure across five training sessions and failed all four test sessions. Aside from presumably pain-induced yelps in the dogs with e-collars when they received shocks, none of the dogs in any groups showed any signs of stress or distress. E-collars may be an appropriate tool in the hands of expert trainers training behaviors that have important welfare impacts, such as running after cars or other animals. Future studies should investigate the levels of expertise needed to use e-collars effectively, the kinds of behavioral problems for which they are best suited, and the longer-term implications of their use.