It seems, the more people read, hear and watch about dogs and training, the more confused some get and pepper my inbox with questions. On one hand we have old school, alpha-wolf type people who enforce strict rules based on misconstrued hierarchical rituals largely taken out of context. Every behavior offered by the dog that falls into the illusional dominant category is corrected or punished.

On the opposite end there are the purely positive reinforcement clicker type trainers that perceive dogs as mere stimulus-response machines, have a scientific explanation for every behavior and an equally science based solution. They apply the teachings of Pavlov and Skinner, forgetting that both worked with laboratory animals without taking into consideration that humans, or any other mammal, have a mind. Not really the socially evolved urban companion most of us live with.

With either method the dog is coerced to behave or perform - one uses a stick - the other a carrot. Then there are trainers that use both, correctly stating that positive and negative reinforcement, and positive and negative punishment are all scientific learning methods.

So, which one works best? To answer that, I always explain how each method affects the dog, because the fact is that if a dog feels socially safe and belonging, he won't be fearful and/or stressed, which makes the owner happy as well because almost all behavior problems are based on fear and/or stress.

Alpha training, corrections and punishment makes for a fearful insecure dog lead by an erratic human bully who rules physically, not mentally. It makes a fearful dog more fearful and stressed, and a bossy dog rebellious and aggressive. Increases behavior problems.

Mechanical clicker type training, although not intimidating, doesn't necessarily mean that the dog is not under performance pressure and stressed. Purely mechanical training shapes one behavior at a time, relies heavily on reinforcers - often food, and doesn't mean that the dog necessarily learns to behave well in all circumstances, and doesn't automatically instill social safety and social belonging. Still, the gentle and humane method of positive reinforcement is a much better and more successful method to direct a dog into desired behavior. As a general rule decreases behavior problems, but shouldn't be used to the exclusion of leadership and should always take the dog's past, genetics and mind into consideration - and Bandura's Social Learning Theory.

Both methods in combination means that sometimes there is a correction and punishment and sometimes praise and reward. Kind of like the spouse who comes home with flowers for ten days and punches you in the face on day eleven. The worst type of training that leads to a confused and stressed out dog, who never knows what comes next. Increases behavior problems.

The best method is shaping behavior with positive reinforcement, with attention to reinforce with what the dog wants at the time, with isn't necessarily food, or not always food.
That should be combined with free shaping and free learning.
Attention to leadership includes resource control, enforcing set rules through redirecting and management (not give the dog opportunity to fail), and protection. That instills social safety and social belonging and the relationship is one of trust, mutual respect and opportunities to learn and have fun together.