So, requirement of IB high school program is Extended Essay, 4000 words research on some topic. And, as I'm really into dogs, I wanted to do one, on dogs. Here is my brainstorming on possible topics:
- Dynamics of a pack – social hierarchy
- can do something like David Mech
- http://www.npwrc.usgs.gov/resource/mammals/alstat/index.htm
- PRO's – quite biological
- CON's – will be hard to find pack of dogs, don't know the origin of the dogs (family or random dogs that got together)
- Communication – calming signals
- can look at dog's reactions to calming signals of another dog or situations in which dogs are giving off calming signals and with which frequency
- PRO's – interesting, people don't know about it much, communication involved, biological
- CON's – what are calming signals, are they just submissive behavior, not enough written on it, highly observational, collecting data??? (Preferably, data should be quantitative)
- Will social behavior put on cue have the same effect as the one naturally performed
- putting calming signals on cue and looking if they'll have the same effect
- not much to investigate cause it's capturing of the behavior and will probably give the same results
- PRO's – interesting and good for dog training
- CON's – very hard data collection, how will I analyze my data?
- Affects on learning:
a) Breed – can relate it to selective breeding, things already done on it (Stanley Coren's Intelligence of dogs)
- taking ten FCI groups (1. Sheepdogs and Cattledogs (except Swiss cattledogs; 2. Pinscher and Schnauzer type, Molossian and Swiss mountain & cattledogs; 3.Terriers; 4.Dachshunde; 5. Spitz and primitive types; 6. Scent hounds and related breeds; 7. Pointing Dog; 8. Retrievers, Flushing Dogs, Water Dogs;9.Companions and Toys: 10. Sighthounds) or seven AKC groups (1. Sporting; 2. Terrier; 3. Non-Sporting; 4. Hound; 5. Toy; 6. Working; 7. Herding) and maybe mongrels (not as good options since I don't know their parents… Maybe I can, though have a question like: are mongrels smarter/do they learn faster than pure breed dogs?), probably young adults
- PRO's – quite straight forward, have some list to compare my results to, probably will go with AKC list since it means less work.
- CON's – something already done (in a different way, I think Coren looked at solving problems, I'll do how fast do they learn), maybe problems with finding dogs – contact breed clubs, go to Hong Kong dog show in April and ask owners themselves.
b) Age - makes sense to look at it, would argue that the youngest dog's learn the fastest. Can split them into 4 groups: puppies (2-6 months), young dogs (6 months – 1.5/2 years), adult dogs (1.5/2 years – 7 years) and old dogs (>7 years), inside the same breed, maybe two or three dogs of the same age.
- Problem with training experience. If the dog has been trained it can affect the final result. Can argue that it doesn't matter if dog hasn't been clicker trained. Then again, if dog has been trained using P+ it maybe won't be as easy to shape.
c) Treats used - can use it, but more of a math studies thing (they do statistics and preference of treats would be a good thing for that)
d) Sessions: few short or one long – very psychological.. And I know the outcome and I don't think much can be said.
So, I can actually go and see how breed and age affect the rate of learning.
I can then take 4 dogs of different age from breeds in seven different groups. I would try and shape a simple behavior as ‘touch', maybe even without charging the clicker. I would look then at how many times I had to click to get the dog to touch my hand when he sees it, how many errors did the dog make. Probably would work in very short session+ 1-2 minutes, maybe 20 clicks. Would have to take care of following:
a) dog's didn't have much previous experience with clicker training, neither they were trained with a lot of P+
b) dogs are willing to work for food
c) dogs haven't met me before
d) I would work with dogs right before the meal to get them to be hungry
e) Would use high value treats, best something dog's didn't have before
f) Very low distraction environment, owner not present
RESEARCH QUESTION – how do breed and age of a dog effect it's learning rate.
HYPOTHESIS – the younger the dog, it'll learn faster. Dogs from breeds that are/were required to do complicated task would learn faster than those that didn't
VARIABLES
- independent variables: breed and age of a dog
- dependent variables: number of clicks/sessions for dog to learn the behavior
- controlled variables: bunch of stuff I have to control but don't know how.. probably how hungry is the dog, distractions, does he know the trainer. all those things I wrote above as things I have to look at.
The only problem I'm facing is finding the dogs… How to solve it:
a) Finishing the experiment back home?
b) Post flyers around Sai Kung (part of Hong Kong with huge amounts of dogs) and wherever to find people willing to do that
c) Talk to owners on the dog show
I've come up to here with my thoughts... Probably it's quite hard to understand... I did it for myselfand am just posting it here, too for future reference since, if i decide to do breed and age thing, I will probably keep the track of where I'm going here. Any suggestions on whatever are welcomed.
