Remember kids... always wear a helmet. (The almighty bikeradar helmet thread)

Jeez, you have completely ignored a very important point that MDCC has made:

this is why we ignore anecdote and study populations; if you measure the passing distance enough times, you get a statistically significant view of whether some factors relating to the cyclist influence the behaviour of a population of drivers.

Much of what you have said, especially in point 1, is based on well thought out anecdotes and assumption.