AA Success Rate Statistics: What Every Widely Cited Number Actually Shows

Search for the AA success rate and you will find pages confidently telling you it is 5 percent, 8 to 12 percent, 40 percent, 50 percent, 60 percent, or 75 percent. Those numbers cannot all be right, and none of them means what the pages quoting them imply. Some describe different decades, some describe different outcomes, and several cannot be traced to any study at all.

This guide maps every widely cited Alcoholics Anonymous statistic to its actual primary source, leading with the best modern evidence, the 2020 Cochrane review. For background on how the 12-step program works, see our separate primer. This page is strictly about outcomes.

Key numbers at a glance

• Manualized 12-step facilitation produced continuous abstinence at 12 months in roughly 42 percent of participants, versus roughly 35 percent for established treatments such as CBT (418 versus 345 per 1,000; Cochrane review, 2020).
• The 12-month continuous-abstinence comparison pools two randomized trials with 1,936 participants and was rated high-certainty evidence; the full Cochrane review behind it covers 27 studies with 10,565 participants (Cochrane, 2020).
• The abstinence advantage persisted at 24 and 36 months of follow-up (Cochrane, 2020).
• Twelve-step facilitation also produced substantial healthcare cost savings versus outpatient treatment and CBT in three economic studies (Cochrane, 2020).
• 44 percent of surveyed AA members reported more than 10 years of continuous sobriety; 23 percent reported less than one year (AA 2022 Membership Survey, published 2023).
• 75 percent of surveyed AA members have attended a meeting online or by phone (AA 2022 Membership Survey).
• Among 2,778 male veterans followed after substance use treatment, 40 percent of those who had engaged with 12-step groups had disengaged one year later; this is the closest real source for the famous "40 percent dropout" claim (Kelly and Moos, 2003).
• 67 percent of people who participated in AA for 27 or more weeks in their first year of recovery were abstinent 16 years later, versus 34 percent of non-participants (Moos and Moos, 2006).
• A longitudinal comparison of 12-step groups, SMART Recovery, LifeRing, and Women for Sobriety found no significant differences in efficacy after accounting for members' abstinence goals (Zemore et al., 2018).
• 27.9 million Americans aged 12 and older, 9.7 percent, had past-year alcohol use disorder in 2024 (2024 NSDUH, published 2025); only 7.6 percent of them received alcohol use treatment (NIAAA, 2025).
• New Hampshire's statewide AA directory listed 834 weekly meetings and eastern Massachusetts' directory listed 2,662, as counted July 3, 2026 (NH Area 43 and AA Eastern Massachusetts directories; methodology below).

Why the AA success rate is so hard to pin down

Before any number makes sense, it helps to understand why this question resists a single answer.

AA is anonymous by design. There is no intake paperwork, no roster, and no follow-up registry, so nobody can count everyone who ever walked into a meeting and check on them five years later. Every AA statistic comes from voluntary surveys of current members or from studies that recruit and track their own participants.

People select themselves in and out. Those who keep attending differ from those who leave, in motivation, illness severity, and life circumstances. Surveys of current members describe the people who stayed, not everyone who tried.

"Success" has competing definitions. Continuous abstinence, percentage of days abstinent, reduced drinking intensity, and remission are all defensible outcomes, and they produce very different percentages. A person sober eleven months after one lapse in month two is a failure on one definition and a substantial success on another. It is the same definitional trap we unpack in our guide to relapse statistics.

Most pages ranking for this topic skip that framing and quote incompatible numbers side by side. The next section untangles them.

Every widely cited AA success rate, reconciled

This table maps each circulating figure to its actual origin, what it measures, and whether the 2020 Cochrane review supersedes it. To our knowledge, no other page on this topic reconciles these numbers to their primary sources.

Two patterns are worth naming. The highest numbers all trace to a single 1955 paragraph, and the lowest all trace to retention arithmetic rather than measured outcomes. And the pages quoting these figures rarely link to any source in this table; they link to each other.

What the 2020 Cochrane review actually found

The 2020 Cochrane review by John Kelly, Keith Humphreys, and Marica Ferri (PubMed record) is the most rigorous evaluation of AA ever conducted and the reference point every other number should be checked against. It analyzed 27 studies with 10,565 participants, including 21 randomized or quasi-randomized controlled trials, comparing AA and Twelve-Step Facilitation (TSF, clinician-delivered therapy designed to connect patients with 12-step groups) against treatments such as cognitive behavioral therapy.

The headline findings, stated precisely:

• Continuous abstinence at 12 months: manualized TSF beat other established treatments, with a risk ratio of 1.21 (95 percent confidence interval 1.03 to 1.42). In absolute terms, roughly 418 per 1,000 participants remained continuously abstinent with AA/TSF versus 345 per 1,000 with comparison treatments, which is where "about 42 percent versus 35 percent" comes from. Cochrane rated this high-certainty evidence.
• Durability: the advantage held at 24 months (risk ratio 1.37) and 36 months (risk ratio 1.42).
• Other drinking outcomes: on measures like percentage of days abstinent and drinking intensity, AA/TSF performed about as well as the comparison treatments.
• Costs: in three economic studies, AA/TSF produced higher healthcare cost savings than outpatient treatment, CBT, and no treatment; the review concluded it "probably produces substantial healthcare cost savings." It published no per-person dollar figure, so treat any specific dollar amount you see quoted as unsourced.

Two corrections matter because the errors ranked highly in search results as of our July 2026 review. A widely shared statistics page (SoberSpeak, dated November 2025) claims the review found a "60 percent higher success rate"; the actual relative advantage at 12 months was about 21 percent, and no 60 percent figure appears anywhere in the review. And the largest treatment-industry guide to this question (American Addiction Centers', updated December 2025) omits the Cochrane review entirely, quotes an unreconciled 5 to 75 percent range, and still leans on AA's superseded 2014 survey.

It is also fair to note what critics used to lean on: Cochrane's own 2006 review (Ferri et al.) concluded that no experimental study at that time had unequivocally demonstrated AA's effectiveness. That was accurate in 2006. Better randomized trials of manualized TSF arrived in the years since, which is why the 2020 review reached a different conclusion. Anyone still quoting the 2006 review as the current state of the science is fourteen years out of date.

"The honest clinical read of the Cochrane findings is that a free, widely available peer fellowship earned its place beside professionally delivered therapy. That does not make meetings a substitute for medical care. In my practice the two do different jobs, and the patients who tend to do best are the ones we help engage with both."

Dr. Richard Marasa, Medical Director, Clear Steps Recovery

What AA's 2022 membership survey shows

Every few years, AA's General Service Office surveys a random sample of members in the United States and Canada. The most recent edition, the 2022 Membership Survey (published 2023 as pamphlet P-48), gathered responses from more than 6,000 members. Most pages on this topic still quote the 2014 edition.

Length of sobriety among surveyed members (AA, 2022):

Read correctly, this distribution is encouraging and cautionary at once. Among people in the rooms today, long-term sobriety is the norm: 44 percent report more than a decade. At the same time, nearly a quarter are in their first year, and the survey cannot see anyone who left. AA itself states plainly that the results describe respondents, not all alcoholics or the general population.

Other current figures from the same survey (AA, 2022):

• Members attended an average of about 2 meetings per week.
• 89 percent belong to a home group.
• 81 percent have a sponsor, and 73 percent got that sponsor within their first 90 days.
• 75 percent have attended a meeting virtually, online or by phone.
• The average member age is 52; membership skews male (about 64 percent) and white (about 88 percent).
• Treatment remains a major doorway into AA: 29 percent were introduced through a treatment facility and 12 percent through a counselor or mental health professional.

AA dropout and retention: what the 40 percent figure really is

Nearly every article about AA repeats some version of "about 40 percent of members drop out within the first year." Trace the citations and you find treatment-center blogs citing other treatment-center blogs; none lands on a study of AA's general membership, because no such study exists. AA has no roster from which to compute a dropout rate.

The closest real number comes from Kelly and Moos (2003), who followed 2,778 male patients through the year after substance use treatment in the Veterans Affairs system. Of the 91 percent who had engaged with 12-step groups before or during treatment, 40 percent were no longer attending at one year. That is a genuine, peer-reviewed 40 percent, but note what it is: disengagement from meetings among male veterans leaving professional treatment, not a claim about everyone who tries AA. And stopping meeting attendance is not the same as returning to drinking; some people disengage and stay sober, others keep attending and struggle.

The same study carries the practical lesson: patients who started 12-step activities during treatment, rather than being handed a meeting list at discharge, were significantly less likely to drop out. Early, supported engagement predicts retention.

What predicts success in a 12-step program

If the honest answer to "what is the AA success rate" is "it depends on engagement," the next question is which forms of engagement matter. The research is unusually consistent.

Duration of participation in the first year. The landmark evidence is Moos and Moos (2006), a 16-year prospective study of 461 people with alcohol use disorder who were untreated at the start. Of those who participated in AA for 27 weeks or more during their first year, 67 percent were abstinent at the 16-year follow-up, versus 34 percent of those who did not participate that year. Continued AA involvement in later years was also associated with better 16-year outcomes.

Sponsorship. Having a sponsor is one of the strongest correlates of sustained participation: 81 percent of surveyed members have one, and most connected with a sponsor in their first 90 days (AA, 2022). We explain how that relationship works in our guide to what an AA sponsor does.

Meeting frequency and a home group. Surveyed members average about 2 meetings per week, and 89 percent anchor themselves to a home group (AA, 2022). Regularity appears to matter more than intensity: the Moos data favored sustained participation over brief bursts.

Starting during treatment. As the dropout data above shows, people who begin 12-step involvement while still in professional care are more likely to still be engaged a year later.

"The pattern I see in patients who do well is engagement you can point to: a sponsor they actually call, a home group where somebody notices if they miss a week, and step work they take seriously alongside their treatment plan. I offer that as a pattern from clinical experience, never as a promise, but it is a remarkably consistent one."

Dr. Richard Marasa, Medical Director, Clear Steps Recovery

The Big Book's 50 percent claim is history, not evidence

The oldest number in circulation comes from AA itself. The Foreword to the Second Edition of Alcoholics Anonymous, written in 1955 about the fellowship's early years, states: "Of alcoholics who came to A.A. and really tried, 50% got sober at once and remained that way; 25% sobered up after some relapses, and among the remainder, those who stayed on with A.A. showed improvement."

Added together, that passage is the source of every "75 percent success rate" claim you will encounter. It is a movement's own impression of its first members, written before modern research methods, with no control group, no defined follow-up period, and an unquantified qualifier ("really tried") doing enormous work. Some modern statistics pages still present it alongside current research as comparable evidence. It is not. The 1955 foreword is a historically interesting artifact; the 2020 Cochrane review is data.

Where the 5 to 10 percent figure comes from

At the other extreme, AA's best-known critic, psychiatrist Lance Dodes, argued in The Sober Truth (Beacon Press, 2014) that AA's real success rate is between 5 and 10 percent. That estimate was assembled from retention figures: what fraction of people who ever encounter AA go on to achieve long-term sobriety in it, under unfavorable assumptions about how many newcomers stay.

The estimate deserves a fair reading. It captures something true: most people who ever attend an AA meeting do not become long-term members. But retention is not effectiveness. By the same math, most gym memberships and most prescriptions "fail," because attrition from everything is high. The question a family actually needs answered is different: for people who engage, does it work? On that question, the randomized evidence Cochrane assembled in 2020, six years after The Sober Truth, points to a clear yes for abstinence outcomes. The 5 to 10 percent figure and the 42 percent figure are answers to different questions, one about attrition, one about efficacy.

AA vs. SMART Recovery, LifeRing, and Women for Sobriety

AA is not the only mutual-help option. The best comparative data comes from Zemore and colleagues (2018), who followed 647 adults with alcohol use disorder drawn from 12-step groups, Women for Sobriety, LifeRing, and SMART Recovery for 12 months.

The initial numbers appeared to favor 12-step groups: SMART-affiliated participants fared worse across outcomes and LifeRing members showed lower odds of total abstinence. But those differences disappeared once the researchers controlled for members' baseline recovery goals. Commitment to abstinence, not group branding, predicted outcomes. The authors' conclusion: the alternatives appear about as effective as 12-step groups, and people do best when they commit to abstinence in a group that fits them. We profile these options in our guide to secular organizations for sobriety.

One more corner of this debate deserves a direct answer. Some pages ranking for AA statistics are advocacy sites for naltrexone-first approaches (often called the Sinclair Method), framing AA's contested numbers as proof that mutual help fails. The evidence supports a less dramatic conclusion: anti-craving medication is genuinely effective and underused, and AA outperformed or matched established therapies in randomized trials. Those findings are not in tension, because medication and mutual help are not competitors.

Can you combine 12-step meetings with medication?

Yes, and many people should. NIAAA's clinician guidance on treatment options lists FDA-approved medications for alcohol use disorder (naltrexone, acamprosate, and disulfiram) alongside behavioral therapies and mutual-help groups as complementary, evidence-based tools, and explicitly supports combining them.

The gap between evidence and practice is stark: among the roughly 28 million people with past-year alcohol use disorder in 2024, only 2.5 percent received medication for it (NIAAA, 2025, analyzing 2024 NSDUH data). Nothing in AA's literature prohibits prescribed medication, though meetings vary in culture, and some members report pressure about psychiatric prescriptions. The clinical consensus is simple: medication decisions belong with a prescriber, not a meeting. A structured medication-assisted treatment program can run alongside 12-step participation, with each covering what the other cannot.

Online meetings and mutual help in the telehealth era

The pandemic permanently changed how mutual help is delivered. In AA's 2022 survey, 75 percent of members reported having attended a meeting virtually, online or by phone, a question that did not meaningfully exist in the 2014 edition most websites still quote.

Early outcome research is cautiously positive. Timko and colleagues (2022) examined online mutual-help attendance among 647 adults with alcohol use disorder and found that online attendees tended to be earlier in recovery, with lower abstinence at baseline, yet showed abstinence comparable to in-person-only attendees by 12 months. Their interpretation: online meetings appear to appeal to, and help, people earlier in recovery who might not otherwise connect with support.

The practical upshot: distance, transportation, childcare, and privacy are weaker barriers than they were a decade ago, on both the mutual-help side and the professional side, where structured clinical care now works remotely too. We cover that shift in our article on virtual intensive outpatient programs.

When meetings alone are not enough: co-occurring conditions

The statistics above describe populations. For an individual, the more important question is whether mutual help alone is the right level of care, and for a large group of people it is not.

Depression, anxiety disorders, PTSD, and other psychiatric conditions frequently co-occur with alcohol use disorder, and a peer meeting, whatever its strengths, cannot diagnose a mood disorder, manage medication, or deliver trauma-focused therapy. Meetings also vary in how they treat psychiatric medication; most are supportive, but a newcomer who hears that antidepressants are a "crutch" is hearing advice no clinician would endorse. Signs that professional, integrated care should accompany mutual help include drinking that self-medicates panic, trauma symptoms, or low mood; any history of psychosis or suicidal thinking; and repeated returns to drinking despite genuine engagement with meetings. In immediate crisis, call or text 988.

None of this is an argument against AA. It is an argument for matching the level of care to the person, with meetings as one layer of support rather than the entire plan when a second condition is in play.

"Meetings can carry someone a long way, but a support group cannot diagnose depression, adjust a medication, or treat trauma. When a patient keeps struggling despite honest effort in the rooms, I do not read that as failing the program. I read it as a sign that something underneath needs professional attention alongside it."

Dr. Richard Marasa, Medical Director, Clear Steps Recovery

12-step meeting availability in New Hampshire and Massachusetts

National statistics say little about what recovery support looks like where you live, so we compiled the numbers for our two states directly from the official intergroup meeting directories on July 3, 2026.

Population figures are the U.S. Census Bureau's Vintage 2024 estimates. Two honest caveats: the eastern Massachusetts directory excludes the separate western Massachusetts intergroup, so the true statewide per-capita figure is somewhat higher than the floor shown, and directories list meetings, not attendance. Even so, New Hampshire's roughly 59 listed meetings per 100,000 residents each week is unusually dense coverage, and 164 of NH's listings are tagged as online meetings. In both states, a newcomer can realistically find a meeting any day of the week, in person or virtually.

For treatment-system context, SAMHSA's Treatment Episode Data Set recorded 59,728 admissions to state-reported substance use treatment in Massachusetts in 2022, a rate of 975 per 100,000 people, among the higher rates in the country (TEDS 2022 annual report, published 2024). New Hampshire's TEDS line shows only 500 admissions at a rate of 40 per 100,000, among the lowest reported in the nation; that reflects a gap in what New Hampshire's state agency reported to the federal dataset, not actual treatment volume, and it is a useful reminder to check how a statistic was collected before quoting it. Anyone seeking professional alcohol addiction treatment in either state has options across levels of care, with detox available by referral when medically needed.

How 12-step facilitation fits into outpatient treatment

One Cochrane finding deserves special emphasis because it changes how families should think about "AA versus treatment." The interventions that performed best in the randomized trials were not meetings alone; they were Twelve-Step Facilitation, professional therapy designed to get patients durably connected to 12-step fellowships. The review concluded that the abstinence advantage came largely from fostering AA participation that continued after the therapy ended. The strongest evidence for AA is simultaneously evidence for professional treatment that engages with it deliberately.

That is how mutual help is integrated in practice. In a structured outpatient setting, such as an intensive outpatient program, clinicians can treat the medical and psychiatric picture, introduce 12-step or secular mutual-help options during care rather than after it, and troubleshoot early obstacles like finding a sponsor or a compatible meeting. The data throughout this article, from the dropout research to the 16-year follow-up, consistently favors that combination: professional care and sustained peer support, started together. Clear Steps Recovery provides this model of outpatient care in Londonderry, New Hampshire and Needham, Massachusetts; our admissions team at (603) 769-8981 can answer questions about how mutual help fits into a treatment plan.

Methodology and sourcing notes

Every statistic on this page links to a primary source, and each figure is labeled with the year of the underlying data rather than the year a page repeating it was last updated. We used the peer-reviewed literature named above, AA's own 2022 membership survey pamphlet, and federal datasets (the 2024 NSDUH, NIAAA's treatment statistics, and the TEDS 2022 annual report). Where a circulating figure could not be traced to any primary source, such as the "8 to 12 percent" claim, we say so rather than citing a secondary page. The New Hampshire and Massachusetts meeting counts are our own compilation: we counted the weekly listings in the NH Area 43 directory (nhaa.net) and the AA Eastern Massachusetts directory (aaboston.org) on July 3, 2026, and computed per-capita rates against the Census Bureau's Vintage 2024 state population estimates; listings change continually, so treat those counts as a dated snapshot. Statements about competing articles reflect our review of the top-ranking pages in July 2026. This article reports population-level research; it does not predict any individual's outcome, and no program, including AA and including professional treatment, can promise sobriety for a given person. This page was medically reviewed by Dr. Richard Marasa and last reviewed on July 3, 2026.