AA has a higher success rate than 5%

There is a frequently quoted 5% success rate figure for AA's success. This is a popular figure which anti-steppers love to bring up in the comments sections of articles about AA, almost always without citation. This number comes from no less than three different sources, all of which use incorrect methods to derive the 5% number:

• The 1990 Triennial Survey which the anti-steppers love to bring out had, of the population of people in their first year, 5% in their 12th month. To say this shows a 5% success rate shows a profound ignorance of statistics, as a HindsFoot article clearly shows.
• Agent Orange started a rumor that Vallant in The Natural History of Alcoholism Revisited showed a 5% success rate. The old green-papers site refuted it quite nicely; I summarize those findings in the Wikipedia talk page for Valliant's book. The actual AA success figures in that book, for people who attended 300 or more meetings over a ten-year period, is on page 197. The "this many people who attended 300 or more meetings" numbers are not in that table, but easily enough determined with some simple math: 74% of the people who went to 300+ meetings had stable remission, 21% of the 300+ meetings attenders had intermittent alcoholism, and only 5% of people who went to 300 or more meetings were still chronic alcoholics -- numbers, that, interestingly enough, agree with the figures in the preface to the second edition of AA's own Big Book (50% got sober right away, 25% got sober after relapsing, and the rest showed improvement).
• Lance Dodes used some questionable math in The Sober Truth to come up with this same figure; Gabrielle Glaser, in her 2015 hit piece, parroted Dodes' figures. Thomas Beresford, writing for the National Council on Alcoholism and Drug Dependence, directly refutes Dodes' bogus 5% figure in one article and shows some actual success figures for AA in another article.