Well. Assuming that previous infections don't impact the future odds of infection survival. The equation of the odds of surviving the first scratch is 100%x0.75.
you got your first scratch your chances of surviving it is 75%. This is true for obtaining scratches in the future. Your chances of surviving a scratch in the future are 75% since we are assuming that previous scratches dont impact future odds.
But things get interesting when you start accumulating scratches.
Lets have a second assumption that the person has no traits that impact infection odds (thick skinned,thin skinned etc.)
imagine you got scratched twice escaping from being mauled. the odds of surviving one scratch mathematically is 75% BUT the odds of surviving both scratches is not 75%. We take the first term (75%) and reapply it back into our original equation. This gives the odds of surviving both scratches are 75%.0.75=56.25%
This means that the odds of surviving two scratches (regardless if they were back to back or over the course of weeks-months) is 56.25% instead of 50%. This can be reapplied as many times as you want
surviving total 3 scratches 56.25x0.75=42.19% (to two decimal places)
surviving total 4 scratches 42.19x0.75=31.64%
surviving total 5 scratches 31.64x0.75=23.73%
this can keep going forever
If somebody was actually bothered (or simply wants a challenge) somebody could create a simulation using the data and running 100,000 times in order to calculate on average how many scratches it actually takes for somebody to be zombified from a scratch. I am pretty confident the average would be 2.5 per person.