That's better.

Your first source is good, no issues there.

The second one is a law review, which is not a relevant paper (and I think not a paper at all but I'm not a law student). So it'd only be worth anything if a witness had suffient credentials, and no alternate motive, and then it'd still lack peer review, academic oversight, ...
The paper needs to be on-topic, I can't expect a biology paper to get art history right. Usually papers constrain themselves to topics they know for that reason, but for a law review I can see why that isn't possible.
Papers also make their statements more directly, so you would find a more clear statement about breeds and attacks. The fact this is missing should give you some warning signs.

You did well then digging out the third source, which is fine again. You should have just skipped the second one entirely, since it's not primary anyway, just citing.
This third might be what wikipedia was using, numbers look familiar.

Fourth source looks fine.
It is about a different topic, agression not violence, but you use it relevantly.

While checking the first source I saw the full results section, containing "Prior to 1980, the majority of dog bites reported in peer-reviewed literature were attributed to the German Shepherd breed (68.4%). From 1981–2000 German Shepherds still accounted for the largest minority of breeds identified (20.1%), with mixed breeds (19.6%) and Pit Bull type breeds (14.1%) accounting for the 2nd and 3rd largest minorities. Since 2001, Pit Bull type breeds have accounted for the largest subset of dog bites reported in the medical literature (37.5%) [...]" continuing with your quote.
It is interesting to note that German Shepherds apparently used to take the same statistical position you are arguing pitbulls hold in the present.

Ok now. You mentioned "Notably you'll notice that a ban [...] was the result of an ~89.5% reduction in pitbull attacks (1-(5.7/54.4)).".
That is mostly meaningless for our purposes, i.e. determining what factor dog breed plays in dog-human harm.
Firstly, it's a ratio, so speaking of reduction is incorrect. Reduction refers to numbers, so you need to consider the total amount. If the total number of dog attacks remained stable, then this would be accurate.

If that were the case though, it would be precisely not the point you wanted, since it would mean banning pitbulls did nothing for the underlying problem: dog-human harm (not pitbull-human harm).

I saw the third article have a discussion on more useful metrics for judging pitbulls, but ima just roll my own rq:
As mentioned above the next step would be seeing if limiting pitbulls reduces dog attacks in total, not just pitbull attacks. That would cover cases like problematic owners shifting breeds.
An even more general one, one the article makes, is factoring in total dog ownership. That also checks if maybe reduczions are just people having less dogs in general.

Basically you'd want the same you have for for example large medical studies, factoring in many quantities and doing a multi-variable correlation analysis to hopefully determine all the various independent correlations and have a better chance at establishing causes.

After that you could see further. You may for example reach the conclusion that the only way is either adressing owners, or banning all large dogs, which if the case should change your strategy for action. If in that case you had already banned pitbulls you'd have shot yourself in the foot, since them people would be sick of it, the motivation would be used up, and you'd have to more than start over for the next breed or the actual wider methods.

On the other hand if it is indeed mostly a breed issue you would have far better arguments for getting breed-specific regulations.

The 4th source is a good start for establishing hypotheses if pitbulls are shown to be a problematic breed. You would need another metric like harm potential for the strenth, bit damage, etc. of dogs, since an agression rating led by chihuahuas is perhaps not the most useful here. Then you can multiply agression and harm potential for an estimate of natural inclination to endanger humans.
I haven't looked deeper into the paper as to how "aggression" was measured, that method would also have to be airtight against training, so probably measured for dogs trained in a controlled environment etc..

If you do find a controlled link between such a harm potential rating and actual human harm using the methods of the last section, then that would immediately show both problem and solution. You would have a method by which to rate any new breeds or vatiations and could write regulations to target that. So even if someone creates new breeds to target the aggressive dog market, the regulations would cover that without any new changes. And similarly you could breed pitbulls to be more passive and get that variation exempt from limitations. That should probably cover most peoples interests in this debate.