Like many departments our highest amount of usages are building searches. Due to the high number of false alarms, the find ratio is realy low. In order to keep the dog motivated for building searches it was suggested (on the false ones) that once the building has been serched an officer hides inside with a tug and the dog is sent in to find him with the tug being the reward. It was suggested that this be done on more than 50% of the false alarms so it keeps the dog's motivation high, Opinions?

I agree that on occasion, you need to allow the dog to find someone. This can be accomplished, depending on the number of searches that no one is found, during training or at the conclusion of a negative search. Behavioralists will tell you, the optimum behavior control is achieved when a dog is trained using a variable reward system. Variable reward is nothing more than the dog being trained where his primary reward is not given each time a find is made, but using a random interval. It sounds complicated, but really isn't. During intial training, the dog is given it's primary reward each time is successfully completes the task, as training progress, the reward is given, once for every two successful completions, then for every third successful completion and so on. Then the reward is given on a random basis, somewhere within the variable you have created. In this manner, the dog learns that it will be reward at some point during the training/work session. That reward could come at anytime from the first to say the sixth, as an example.

DFrost

I realize that this is a bit of a "Catch 22" situation for a small agency trying to justify the costs of a program, but, frankly, the use of a PSD to search residences, especially, as a routine thing is not usually a good policy for many reasons...note, please, that I said to actually "search"...not "respond to". Business alarms too, do not always call for the dog to be put on the ground just because an alarm went off.

While doing some finds in the variable manner David describes will probably help, a better fix for the problem might be a review of policy. JMO

Mike... I think you took the question out of context. We don't search residences as a matter of policy with the dog unless there is a sign of forced entry. We only search buisnesses after alarms where unsecured doors are located and there is a possability that someone might be inside. I don't see in the post where department policy was ever mentioned.

Matthew;
I took that info directly from your Department website's narrative on the Canine Unit:

"The K9 Unit responds to most burglar alarms within Ross Township and routinely searches the interiors of residences and businesses to search for possible burglars that may be hiding inside during break-ins."

Sorry if it is incorrect.

I am involved with a large number of departments in which this very concern arises.
Simply, the solution is this. If you deploy in an environment resulting in a no find situation you must, as quickly as time permits, repeat this deployment or an approximation of the deployment in which the dog is successful.
In a building search, a bite is not necessary, but a find is. In fact I discourage much biting and encourage civil work under those circumstances you discribe.
In addition, if possible, you need to arrange searches where you arrive and deploy just like a call. But, have prearranged auccesses. A few dogs figure out that the good search is second.
Variable reward in the context of dog training is pretty theoretical. There is some good evidence to show that in the way we apply dogs all a variable reward system does is lower the average intensity level of the response. But, in regards to street work we can't hardly do any different once we get into deployment vs. training.
Variable length (time in this case)prior to success is more important in respect to maintaining an intense search that any other factor. Lots of smaller quick searches interspersed with a few vaariable length longer searches are more effective than trying to staistep your way into longer searches.

Actually, all behavior is theoretical. However there is an inordinate amount of research that supports the variable reward as being the strongest behavior control. It applies to not only animals, but humans as well. Find a quarter in a soda machine and you check soda machines for months without finding another one. That, in itself is variable reward. None the less, performance should tell the story. Document, document, document. Documentation tells you where you are, standards tell you where you need to be.

DFrost

 Quote:

Originally posted by David C.Frost:
Actually, all behavior is theoretical. However there is an inordinate amount of research that supports the variable reward as being the strongest behavior control. It applies to not only animals, but humans as well. Find a quarter in a soda machine and you check soda machines for months without finding another one. That, in itself is variable reward. None the less, performance should tell the story. Document, document, document. Documentation tells you where you are, standards tell you where you need to be.

DFrost

I completely understand variable reward theory David. What I am saying is that there is some recent research involving dogs and dog training (not humans and quarters) that supports the idea that the variable reward system may not be as important in the context of what we do with dogs as the general theory that we have applied over the years might lead us to believe. Because I happen to believe that much of what people spout as truth is simply convention, it is important to raise the question as to whether or not what we think to be truth really is or not. Whether what we have done is really supported by anything more than the fact that that is the way it has been done. Anything less is a dead end for the critical thinker.

"subvert the dominant paradigm"

Kevin, I'd like to read it; can you tell me where to find this research?

Originally posted by Lee Baragona - Sch3FH2:
Kevin, I'd like to read it; can you tell me where to find this research? Let me hunt for it Lee. I think I can find a url