OK, first of all, I understand, know, and have experience with everything you say.
Side point: within my extended family (uncles, cousins, nieces, etc.) I have 2 neck injuries due to seatbelts (one permanent and near-fatal), and 2 cases of severe stomach damage due to aspirin. I almost agree with your point, but you lucked on two bad examples.
Anyway, about those tests.
1. The 32% number varies dramatically with the type of thing being tested (pain relief vs life-saving drugs, drugs vs devices, etc.) Also, the 32% number is a statistically reduced value. The raw data has a surprising variance across studies.
2. Given the near-violent nature of a lot of the posts around here attacking or defending specific Zapper designs, vendors, and theories, I think the response pool is, uh, biased? Stepping back and looking at the traditional medicine crowd and the alternative medicine crowd, it is blatently obvious that the majority of people in both crowds are biased and not reliable reporters of fact. That's not a bad thing at all, and is an inevitable consequence of advocacy, but it also is a thing to be recognized and taken into account.
Take, for example, "offset". (Oh, god, here we go...) Read the books, read the postings, read your stuff. Can't find anywhere that someone tested offset: Is offset really contributing to effectiveness? If so, how much is enough? Since the nervous system is asymmetrical in most people (right-left-handedness), does the polarity of the offset with respect to body geometry matter? What about more offset, but a smaller AC component - a square wave varying between 8V and 12V, for example (4V wave on 10V offset)? Considering how I'm going to be blasted for even asking such a question, imagine trying to get clean data from a group of Zapper enthusiasts, all swearing by their favorite Zapper's effectiveness, without DB testing.
As for the more basic issue of testing Zapper effectiveness with placebo devices.... Not today. Got bounced from the forum once for asking that question. Once is enough.