The question provides a table and asks:
Number of Attempts Fraction persisting in fibrillation
"Assume that the probability p of defibrillation on one attempt is independent of other attempts. Obtain an equation for the probability that the patient remains in fibrillation after N attempts. Compare it to the data and estimate p."
The Attempt at a Solution
I used the binomial distribution for my equation to estimate the probability that the patient remains in fibrillation. I'm not concerned about the "number of successes" in each attempt, so I believe this problem is similar to asking a coin toss question. For example, the probability that a coin will return heads after 1 attempt is 0.50. After 2 attempts, 0.5*0.5, etc.
Likewise, there are two possibilities: fibrillation and defibrillation. Instead of the coin example, the probability that the patient remains in fibrillation is 0.37. After two attempts, 0.37*0.37. After 3 attempts, 0.37*0.37*0.37, etc. It models the data rather well.
So then to estimate "p", the probability of defibrillation in each, p+q = 1 ---> p= 1-q
Does this sound reasonable?