Why is it by each manufacturer individually?
To show that this is a correlation seen across manufacturers. You provided an excellent example of why this is necessary when you singled out BMW drivers, when the data showed this correlation across all manufacturers.
Why do some manufacturers have a much higher percentage of accidents even though they don't make high power cars?
Potentially because they're less safe, or they draw riskier drivers?
Why are some manufacturers percentage of accidents in their high power cars lower than other manufacturers low power cars?
See above.
You'll probably need other slices of the data to reason about those questions. However what we see here is there is a positive correlation across all manufacturers between horsepower and accidents.
This looks like a chart that was specifically structured to prove a point the researcher started from.
I imagine this data was analyzed with answering "do higher horsepower cars have more accidents" in mind, so the presentation of the data will naturally have answering that question in a comprehensive manner in mind. Please remember that you're not seeing a table with all of their data - you're seeing a particular subset that they believe is relevant to the question at hand.
So....at the end of the day, do you agree that the data shows there's a positive correlation between increased horsepower and percentage of cars having reported accidents?
