I've seen analyses by many people with impressive credentials that seem very impressive but amount to, a quote I remember, pseudoscientific sheep dip. tdracer cites the most relevant statistics - fatal accident rates. Even these are skewed but that is nit-picking. Pilot suicides are a hit on the airframe. Airframes lost to bombs are in the statistics. Pilot training - several hull losses have been followed by shutting down an airline due to safety deficiencies.
I've lived the life of engineering through manufacturing to failure analysis and accident investigations. Do out of tolerance parts get used in production? Yes. In fact at one company we engineers needed to analyze parts and determine if they had Critical or Major characteristics, mark those characteristics, and indicate on the part number if it had a Critical or Major characteristic. Referring to W Edward Deming's quality philosophy, I remember him stating in the first of his video series that quality started to go downhill with the hiring of the first inspector. Relying on an inspector to check whether or not you did your job correctly was an insult to workmanship.
In 1974 I had the opportunity to take a course Applied Fracture Analysis by Dr David Hoeppner who had worked for the Skunk Works but wanted to move on to a place where his research could be taught outside the confines of a small part of Lockheed. Our first class would have made those with weak stomachs leave as he brought in failed components that had resulted in fatal crashes. Our class following the crash of TK901 found Dr. Hoeppner bringing in a cargo door latch from a DC-10 and telling us we stupid engineers created this disaster - that was before knowing which part of the design caused the failure. One class, unnerved by his thoughts of engineers, I asked him if he were flying a transcontinental flight and could chose any plane, he selected a Convair 990 (remember, this is 1974). I asked him why not a Boeing? He said in his opinion they got burned early in the company's life and overdesigned everything. Boeing's carry a lot of extra weight due to overdesign. That was back in the days when the 737 was an infant. Dr. Heoppner was upset because he could not get TWA to send their engineers to his class as this was near their overhaul base. One day he entered the classroom particularly elated jumping on top of his desk and dancing. Seems the day before an axle had failed on one of TWA's L1011s. Being a Lockheed product, Dr. Heoppner was asked to investigate and diagnosed stress corrosion cracking. "Those idiots don't know how to identify a simple failure mode and they won't spend the money to send a group of engineers to my class." Unfortunately we engineers learn the most from failures - oh crap, well, I won't do that again." When we reach our peak we retire and let the new people repeat our errors.