COMMENTS ON A.A.’S TRIENNIAL SURVEYS

INTRODUCTION

This paper is an analysis of the 1989 survey plus a review of comparable findings of previous surveys from 1977 forward (five in all). It contains the following sections:

Introduction The 1989 survey Limitations of the survey Appendices A: History B: Comparison with earlier surveys C: The first year D: Technical notes E: Regional Comparisons F: Population Values G: Progress in the Fellowship

In preparing this analysis the opportunity has been taken to reevaluate past surveys and to apply further analysis to them. In some cases the reduction of data has been done differently, and in a few cases errors have been corrected. Both in the interest of accuracy and consistency the data presented here, when different, should supersede earlier reports.

Purpose of the Survey. The original purposes of the survey are noted in Appendix A. The need for current and reliable information on A.A. for the public and for the professional community, and the need for a periodic inventory for the membership, seem to be as pressing today as in the past.

In connection with the function of inventory, certain findings of recent surveys seem to call for concern on the part of the Fellowship. For example, attention to our singleness of purpose might be suggested by the increase in the number of dually addicted members and the increase in members coming from rehabs. These trends are cited in the chapter on comparison of past surveys.

It is also noted that about half those coming to A.A. for the first time remain less than three months. Appendix C of this report, “The First Year,” is devoted to this subject. This observation may inspire groups to give more attention to the newcomer through such measures as greeting newcomers, beginners meetings, and interim sponsorship. We often say that no one gets to A.A. by accident, so the fact that we lose within three months half of those that begin our program may mean that we lose a great number of alcoholics who desperately need sobriety.

-2-

The crux of the recovery program of A.A. is to face up to difficult realities through inventory and to take action where possible to respond. The increasing number of members with drug problems is simply a thing we cannot change; a response through more effective teaching of the A.A. program is something we may achieve. Similarly the way in which we present A.A. to members coming from rehabs, directly or through literature, may present opportunities for improvement.

THE 1989 SURVEY

The 1989 survey was done through a random sampling process that produced a sample of 9.394 completed questionnaires from A.A. members of the population of groups registered at G.S.O. Except in a few cases that are noted in what follows, the sample can be taken to represent fairly well the sampled population.

Composition of the Sample. Thirty-five percent of the sample are women, 22% are under 31 years of age, and 3% are under 21 years of age. Fifty-six percent are from 31 to 50, and 23% are older than 50.

Help along the Way. Respondents found their way to A.A. through help from a number of sources (two cited per respondent); 34% from an A.A. member, 30% from a rehab, and 19% from their families, to cite the three most important factors. However, 27% cited “on my own,” and other less prevalent influences were counseling, 10%, and doctors, 7 percent. Other factors, such as the media, correctional facilities, Al-Anon member, friend, employer, and clergy, were noted by smaller percentages of the respondents.

Although these percentages cannot be added to give an overall indication of assistance from outside sources, a separate question determined that 68% of the sample had outside professional help of some kind prior to coming to A.A.

Sobriety. Thirty-five percent of the sample had been sober less than a year, 36% had between one and five years sobriety, and the remaining 29% more than five years. With regard to their practice of the A.A. program, 85% had a sponsor and 88% a home group. It appears that more detailed questions in this area might illuminate the way, on the average, that the A.A. program is practiced today.

The first year. Response to a question determining how many months before the survey date the respondents first came to A.A. can be interpreted to show that approximately 50% of those coming to A.A. leave within three months. The details of this determination are given in the appendix entitled “The First Year.” This is undoubtedly one of the most significant observations of the survey.

Role of physicians. Although few credited their physicians as an important factor in urging them in A.A., 28% believe that their physicians now present A.A. as a program of recovery to their patients, and 70% have told their physicians of their membership in A.A.

–3-

Frequency of A.A. meetings. Members of the population attended meetings an average of about 2.9 times per week. In Appendix C the way in which this number is derived from the sample result is presented. Needless to say, the sample result is heavily biased by those who attend meetings often.

Addiction to Drugs. About 46% of the sample indicated a problem with drugs in addition to their alcoholism. The age dependence of this characteristic is very strong, creating a significant cross-correlation between drug problems and the tendency to attend meetings more often. Thus, the kind of correction described in detail in Appendix C does make a slight difference, and the prevalence of past drug addiction in the population is about 42 percent.

Drug addiction is most prevalent among young members. In the older age groups women are significantly more prone to this problem than men.

Occupations. Respondents were asked to classify themselves according to standard employment categories with the following results:

OCCUPATION OF RESPONDENTS

Manager or administrator (except farm) 12.8% Operative (except transportation) 1.4% Service Worker (except private households) 5.3% Transportation equipment operative 2.2% Professional or technical worker 22.9% Housewife or homemaker 5.1% Private household worker 0.7% Farm laborer or foreman 0.5% Farmer or farm manager 0.5% Sales worker 5.9% Craft worker 5.0% Student 4.0% Disabled 3.1% Military 0.7% Retired 7.7% Unemployed 6.7% Clerical worker 5.5% Laborer (except farm) 6.6% Other 4.8%

Changes from Previous Surveys. The proportions relating to a few of these factors have changed substantially over the 12-year period reviewed here. This includes large increases in the proportion of young people, of people coming from rehabs, and people citing past drug dependence as well as alcoholism. Graphical representations of these trends are among the graphs given in Appendix B. Other factors which have remained constant or changed slightly are the average sobriety, the proportion of women, and the frequency of attendance at meetings.

–4-

Other information. There is other information available in this and past surveys. Much was collected according to the hope that a consistent data base would be built over the years so that future members might be able to study historical trends, even if there were not time or effort available to analyze or report on all of it at present. The opportunity to study the incidence of the various attributed and their interrelationship is certainly great. For example, an entire report could be constructed on the women of the survey; another on the young people; still others on young men or women. Correlations of sex, age and occupation with factors like sobriety, drug addiction, origin in a rehab, etc., could occupy analysts for some time, and might help us to understand what is happening in our Fellowship.

Limitations. There are limitations to the data of the surveys. Perhaps most tantalizing is that the rate at which members abandon active participation in A.A. can be estimated, but there isn’t a clue as to the causes. In surveys of this kind in the outside world, provision is often made for follow-up to determine such vital information, but that has not seemed possible for us.

There are limitations in the survey coverage. The population surveyed is the 48,000 groups registered at G.S.O. (roughly 980,000 members.) The many groups that are not registered at G.S.O. are not included. It is conceivable that the members attending such groups have different characteristics. It is also possible that some members in groups that were surveyed failed to or refused to complete the questionnaires, and that such members have different characteristics.

In other respects the way in which questionnaires are administered to groups has undoubtedly not been uniform, even though a good random selection of groups is made to start with.

Nevertheless, the consistency of results from survey to survey (Appendix B) suggests that each presents a good overall picture of our Fellowship.

–5-

APPENDIX A: HISTORY

The first survey was conducted in 1968, to respond to a request of Dr. Jack Norris. He found that some of our communications to the professional community had very little credibility because of a lack of objective data about the Fellowship. However, the survey was also seen as useful to the Fellowship itself as a kind of inventory. In the foreword to a report on the 1968 survey Dr. Jack said:

“There were two major reasons for undertaking the survey:

1. “To enable A.A. to furnish more accurate data about the Fellowship and its effectiveness to the growing number of professionals – doctors, psychiatrists, social workers, law enforcement officials and others – who are working today in the field of alcoholism.

2. “To provide A.A. with more information about itself so that members can work more effectively in helping the many millions of alcoholics who still suffer throughout the world.”

Dr. Jack’s desire was implemented by the Public Information Committees of the board and the Conference (there was no separate C.P.C. function at that time). A copy of the first questionnaire is attached.

In 1970 the board formed a separate C.P.C. function. It is my impression that, by that time, a third use for survey material had been identified in the efforts of the P.I. committee to inform the public about A.A. Thus the continuance of the survey was assigned to the surviving P.I. committee rather than the new C.P.C. committee when they were separated. For example, in a proposal for the 1971 survey it was stated that:

As a result of these findings (the 1968 survey)…(1) we have prepared and directed specific messages to the family of the alcoholic, (2) we are conducting a continuing evaluation of what the public thinks about A.A., (3) we are exploring ways and means to overcome the negative features of A.A.’s reputation, and (4) the Professional Relations Committee has been established to work with doctors and interested professionals.

To serve these several objectives of the survey many questions have been added to the survey questionnaire and reworked by the trustees’ P.I. Committee as time has gone on, resulting in the present questionnaire (copy attached). In particular, Dr. Jack and Milton Maxwell made many suggestions, and both repeatedly urged the importance of the survey effort. Suggestions also came from C.P.C. committee members.

Interest has also been expressed in many topics that were ruled out for various reasons; these included ethnic and religious background, income level, follow-up procedures, details of drug use, legal involvements, etc.

–6-

In the past decade the desire to leave the questionnaire essentially the same, in the interest of having comparable data from one survey to the next, has inhibited the introduction of new topics.

The processing of the information on computer was undertaken by an A.A. member from Michigan. From the beginning through 1980 he, his family, and his students did an enormous amount of work in writing the computer programs, reviewing and entering the data from the many questionnaires, and running the programs.

In 1983 the processing was done by A.A. volunteers, and a data processing firm under contract handled the computer work. In 1986 all was done by G.S.O. employees, and in 1989 we reverted to the 1983 arrangement.

At first, the raw data were treated with extreme caution. Dr. Jack said in 1968, “Survey results should not be used to project figures for…the Fellowship of A.A….A.A. members who filled out the questionnaire represent only those who attended the particular meetings at which the survey was conducted. This represented 5% of the groups in each of the 50 states and Canada.” Elsewhere it was cautioned that the sample represented those members who attend meetings most frequently.

As successive surveys showed consistency of results and trends, confidence began to grow that the results were, within limits, representative of the Fellowship as a whole. For example, data within the survey made it possible to estimate and compensate for the effect of frequent attendance at meetings. More important, the adoption of a scientific sampling process in 1983 made it possible to decrease the number of questionnaires processed by a factor of three and yet increase the confidence that the results are representative of the Fellowship as a whole, although not of any specific group or area.

It was also realized that a good deal of information about the A.A. process was implicit in the raw data, but so far only limited use of that fact has been made.

A consultant, Professor of Statistics Joseph Sedransk, then of the University of New York in Albany, was engaged to help with and review the characteristics of the sampling system that was devised for the 1983 survey. At the same time a member with expertise in the physical design of questionnaires greatly improved the format.

The number of completed questionnaires for the various surveys was:

1968 11,355 1971 7,194 1974 13,467 1977 15,163 1980 24,950 1983 7,611 1986 6,977 1989 9,394

–7-

The reduction in 1983 illustrates savings made possible by the stratified sampling scheme. The size of the sample was established to be large enough to give good estimates of general characteristics of the entire population, but it was not practical to make it so large as to give a reliable picture of a single area.

In 1980 the survey form was used by about six other countries and a report prepared on the composition of the Fellowship in those countries. In 1986 the U.K. and Australia used the questionnaire again and G.S.O. processed their returns.

–8-

APPENDIX B: THE SURVEYS FROM 1977 TO 1989

The importance of the 1989 survey is enhanced by the ability to compare the findings to those of previous surveys.

Surveys have been conducted every three years since 1968. In this appendix the last five, from 1977 to 1989, are examined to indicate trends in the nature of the Fellowship and the characteristics of its members.

A number of characteristics of members are shown in the form of graphs in Figures B-1 and B-2. A few show strong trends of the past 12 years; namely: factors that led to A.A. (the A.A. member declining in importance, rehabs and counseling increasing), addiction to drugs (increasing), professional help before and after coming to A.A. (increasing), young people in A.A. (increasing), and older members (decreasing).

A second group of factors changing less strongly would include the proportion of women (increasing), and the prevalence of long-term sobriety (increasing).

Factors that show little significant change include the frequency of meetings attended and average sobriety.

Ba FIGURE B-1 less than 21 20f years of age 15 less than 31 qears of age 10 5 77 80 83 86 89 77 08 E8 86 b8 t08 31- 50 years of age 40 60 30 4o 20 more Hhan 51 20 years of aqe fo1 77 0g E8 8 的 77 9g 83 8 58 80 Aveng Mmths f Sobrielty 30 60 Months4o 20 percentage of women 10 10 77 83 98 b& 77 8 E8 b8

-10- F1GURE B-2 FACTORSTHAT FACDRSTHAT q.077 LEOTOAA 40 4 30 30 % 20 20 10 Ceunse lline 10 77 80 83 86 8 77 8 83 98 b8 DISTRIBUTION OFSOBRIETY 40 40 30 20 20 —/ess than I yr Also addicted 1-5yrs. to drugs 10 10 more than5yrs. i7 E8 98 b8 77 E8 98 68 g 60 metines perweek (population) heda sponsor Freqvency prot.heip bcton AA pri help doring Ni 77 g % 77 08 8 99 83 89

– – 11 –

APPENDIX C: THE FIRST YEAR

It is possible to calculate from completed questionnaires, by month, the number of members that have “been around” a given number of months. This relies on the question that determines the month and year that the respondent first came to A.A.

The calculation has been performed for the twelve months of the first year for the five surveys, and the results are plotted in Figure C-1. Such results can be interpreted to show the probability that a member will remain in the Fellowship a given number of months.

To be more explicit: if all the members who report they have been in the Fellowship for less than a month were still present a month later, then the number who report being in A.A. between one and two months should equal the number that report being in less than a month, subject, of course, to month-to-month fluctuations and to any possible seasonal effects. The same should apply to succeeding months.

However, it is observed that there is a steady decline, (subject to inevitable fluctuations).

This has been the case for each of the five surveys we are reporting on, and the remarkable similarity of results for the surveys is shown in Figure C-1, where all five are plotted on a single scale by taking into account the size of each survey. That figure also tabulates the average over the five surveys, and that average strongly suggests that about half those who come to A.A. are gone within three months. Unfortunately there seems to be no way in which the reasons for departure can be determined.

It seems impossible that such a systematic effect could be achieved by any mechanism other than a slow attrition of newcomers during the first year. It is little comfort to suggest that many who leave return later, because those who have done that are already counted in the numbers shown here.

After the first year, survey results show that attrition continues, but at a much slower rate. During such years it is likely that many whose activity in A.A. diminishes remain sober but are no longer adequately represented in the sample.

It is not part of the survey to attempt to determine the causes of this phenomenon beyond noting that suggestions would include individuals sent against their will, individuals who are not convinced of their alcoholism, individuals who are unable to accept one or another characteristic of the A.A. program, etc. But it does appear that this result and its implied challenge to A.A. should be widely understood in the Fellowship.

12- % of those Coming t。 AA within the fist yeer +hat hare remained the indicaked rumbr of marths 22 1999 ot 19.86 113 18 08b1 LLb1 16 14 Averogeof5Surveus 19% 137 10% 9% 8%777767676767 57 3 10 Momth FIGUREC-1

– – 13 –

Individuals may rebel against this result as contradicting our time-honored statement that “half get sober right away, another 25% eventually make it,” etc. That statement applies to observations made at an earlier time, and there is no reason to doubt that changes in society and in A.A. since that time could create a different circumstance today.

Like other findings of the survey, this may be a challenge to the membership to “change the things we can.”

– – 14 –

APPENDIX D: TECHNICAL NOTES

Sampling method: Questions repeatedly arise regarding the meaning of the numbers determined in the survey and their reliability, particularly the meaning of a “random sample.”

The ordinary connotation of the word random suggests imprecision rather than the opposite. If, however, the objective is a practical yet reliable determination of the average characteristics of a large population of people or objects, then random sampling is the way it has to be done.

If Time magazine wanted to predict the outcome of a presidential race, it would first discard the idea of polling all voters, simply on the basis of impracticality. Next it might consider getting the opinions of all subscribers to Time Apart from questions of practicality, such a project would be fatally biased; the subscribers are not typical, economically or politically. In fact, they have also been biased to some extent by reading the same magazine. (Used in the statistical sense the word biased is not a pejorative.) Thus a very large survey would still be an unreliable indicator of public sentiment.

How is it, then, that random surveys of a few thousand people can closely predict the outcome of national elections? It is because the surveyors succeed in their deliberate efforts to achieve randomness (absence of bias).

Another simple illustration from a writer on statistics: a manufacturer of fireworks obviously wants his product to detonate when ignited. But he can’t test all his products for this property or the only ones he will have left will be the defective ones. This fellow has a special need for sampling.

He might simply take one day’s production, test all of it and assume that his result applies to the entire year’s production. But now the possibility of bias creeps in. The conditions of the raw materials and processes on that day may not be representative of what happens the rest of the year.

The best he can do is to select samples to test, at random, all through the year. In this way he can avoid the various kinds of bias that might otherwise creep in.

The meaning of random selection in this case is that he will establish conditions under which any firecracker is as likely to be selected for testing as any other.

The establishment of such conditions, and particularly the fraction of firecrackers he must select for testing in order to achieve a satisfactory degree of assurance in the reliability of his product, is a subject for mathematical statistics. The result can be illustrated from one of the findings of our survey.

– – 15 –

As a final example, then, consider some calculations that were done in 1983 of the percentage of women in the Fellowship in the United States and Canada. The problem was to establish conditions under which any member in that population was equally likely to be chosen for determination of their sex. The approximate solution was to establish conditions under which each group listed at G.S.O. was equally likely to be selected for survey. In 1983 somewhat less than 2% of the groups listed at G.S.O. were selected by a process in which each group in the population was (approximately!) equally likely to turn up in the two percent. On the average, about 20 members per group are expected to be reached in that way, and in 1983 we had responses from about 7,000 individuals.

A biased result would have been obtained if all these individuals had been selected from one state, or one region, or by a number of people around the country selecting the groups they thought were the best groups, or the groups they thought were most typical, etc.

In the random sample actually used, the percentage of women in 1983 was determined to be 31 percent. Because we did not determine the sex of every member of the Fellowship, this result can only have approximate validity when applied to the population. However, statistics allows precise statements to be made about the extent of its validity, in terms of probabilities.

Confidence Limits: If we have succeeded in establishing random conditions, then mathematical statistics can give a table of the relationship between (1) the probability that the percentage of women in the entire population falls within certain limits around the percentage in the sample, and (2) the width of those limits; for example:

Statement of Confidence in the result the statement

31% +- 2% 99% 31% +- 1% 90% 31% +- 0.5% 75%

The meaning of the entries in the table is this: If I want to make a statement that is 99% likely to be correct, I can only say that the percentage of women in the population lies between 29 and 33 percent. If, however, I am content to make a statement that is only 90% likely to be correct, then I can say that the percentage of women lies between 30 and 32 percent. Statisticians usually attribute significance only to statements made with more than 90% probability of being correct.

This table approximates the actual confidence one can have in the statement of the 1983 survey about the percentage of women in the sample, and, when corrected for frequency of attendance at meetings, in the population sampled. However, the term “A.A. member” is very loosely defined and includes many not in the population sampled. Whether such members differ from the sampled population in other respects, and by how much, is not known.

– – 16 –

Other general results of the survey, such as the average sobriety, the average age, the percentage addicted to drugs, etc., have roughly this kind of reliability, subject always to ambiguities associated with the way in which sampling is done, questions are posed by the questionnaire, etc. However, any results which are derived from a small fraction of the population (for example, the sex of teenagers, or of a single area) have limited validity.

Correction for Attendance Rates: Obviously the sampling will pick up disproportionately those members that attend meetings often.

Fortunately the survey includes a question which determines the frequency with which the member attends meetings. The prevalence of such a member in the population sampled is inversely proportional to this frequency.

When this correction is applied to determine the average frequency of meeting attendance in the population sampled, a significantly different result from the sample average is obtained. For example, the sample average frequency was more than 4 meetings per week in 1989, but the population average turns out to be 2.9 meetings per week.

This population average is obtained by correcting, for frequency, meeting frequency data for each of 22 age groupings and 7 ranges of frequency.

In principle the same correction should be applied to all attributes determined by the survey. However, the detail to which the data is reported makes this easy only for the 22 age groups, not for the seven levels of frequency. As an example, data on the sex of 19- and 20-year-olds can be corrected for the average frequency of attendance at meetings, but if there were a crosscorrelation of meeting frequency with sex of the respondent for that age group, the correction would not take it into account.

Fortunately there appears to be little variation of average frequency with age group except for very young members, with the result that the sample results for almost all attributes other than frequency itself and drug addiction appear to be representative of the sample population (see Appendix E). This has been verified for all the attributes cited in this report for the 1989 and 1986 surveys, and has been verified as well for many of the attributes of earlier surveys.

Therefore, comparison from one survey to another in Appendix B, for the purpose of identifying trends, is made using sample averages, which are easily accessible in earlier surveys.

– – 17 –

APPENDIX E: POPULATION VALUES

The numerical values of the attributes discussed in this paper have usually been those pertaining to the samples taken rather than the populations those samples represent. There are, of course, many reasons why these may differ by amounts greater than the sampling errors themselves. Most obvious is that inherent in the sampling at group meetings is a bias toward A.A. members that attend meetings often. Thus a simple sample average of the reported frequency of attendance at meetings will register a higher frequency than that of the sampled population.

Other attributes might also be biased if correlations existed between those attributes and the frequency of attendance.

Because the frequency of attendance is reported in the questionnaire, it is possible approximately to remove this source of bias, and this has been done over the years, with the finding that in most cases little bias exists in the sample attributes. To illustrate, the calculations done in 1986 and 1989 are presented below, and it will be seen that the only attribute strongly affected is the average frequency of attendance itself.

1989 1986 ATTRIBUTE POP SAMP POP SAMP frequency 2.87 4.56 3.22 4.50 % rehabs 0.29 0.29 0.29 0.29 % drugs 0.42 0.44 0.42 0.43 % AAmemb. 0.34 0.34 0.34 0.34 % counslg 0.10 0.11 0.11 0.11 % family 0.18 0.18 0.19 0.18 % self 0.27 0.27 0.27 0.27 mos.sobr. 51.61 49.35 52.07 50.45 treat bfr 0.67 0.67 0.67 0.67 treat aft 0.60 0.60 0.59 0.60 % doctors 0.07 0.06 0.06 0.06 HaveSpons 0.84 0.84 0.84 0.84 %under 31 0.18 0.22 0.20 0.21 % 31-50 0.57 0.54 0.54 0.55 %51-plus 0.25 0.23 0.26 0.24

There are, of course, many other ways in which the sample averages could differ from population averages in ways other than sampling errors. The population itself is that of groups registered at the General Service Office, and there are many other A.A. groups. A study done in 1987 suggested that about 20% of groups registered with intergroups and central offices are not registered at G.S.O. There are groups not registered with any of these.

There are undoubtedly members who, for whatever reason, do not complete the questionnaire; they may constitute a set with other significant differences. Moreover, only about 2/3 of the groups that were selected actually participated; perhaps they systematically differ in other respects.

– – 18 –

The method of distribution of the questionnaires was through the ninety-one area delegates, so that, ipso facto, there were differences in administration of the questionnaires. There are undoubtedly other factors at work.

Nevertheless, the consistency of results over the years, exhibited either as stationary values of some attributes or systematic trends of others, suggests that all sources of error are not sufficient to render the results without meaning.

– – 19 –

APPENDIX E: REGIONAL STATISTICS

This appendix presents the attributes of the individual regions that go to make up the total sample. In accordance with the requirements of the stratified sample the number of groups sampled in each region was intended to be in proportion to the total number of groups in the region. Consequently, the number of returns from the different regions is quite variable.

The percentages are those of the raw data except in the case of frequency of attendance, where the correction for frequency is applied as described in Appendix D. As noted in Appendix E, sample averages in that case alone can be widely different than population averages.

An analysis of variance has not been done for these regions; consequently, confidence limits for the attributes are not available. The attributes for the Northeast, Pacific, and East Central regions have much narrower confidence limits than the others and can be thought of as significant. At the other extreme the attributes for Western Canada have to be regarded as possessing wide uncertainty.

PERCENT OF SAMPLE, EXCEPT WHERE NOTED

WCa ECa NE SE EC1 WC1 SW PA age < 21 0.8% 1.1% 3.1% 2.6% 2.4% 3.2% 2.9% 3.7% age < 31 16.2% 14.3% 27.9% 20.6% 18.6% 18.4% 18.0% 24.2% age 31-50 62.1% 54.1% 51.7% 57.2% 54.8% 61.0% 56.9% 54.2% age 51+ 22.6% 31.7% 20.3% 21.9% 26.6% 20.5% 25.2% 21.4% women 21.6% 30.1% 35.9% 36.9% 32.4% 24.9% 42.6% 36.6% sober <1 29.6% 27.8% 35.5% 38.0% 35.5% 26.2% 32.9% 35.5% sober 1-5 34.4% 34.6% 38.2% 36.6% 34.1% 33.8% 39.5% 37.0% sober 5+ 36.0% 38.0% 26.2% 25.5% 30.4% 39.9% 27.7% 27.5% mtgs/wk 2.6 2.6 3.8 3.4 2.9 2.3 2.6 3.3 AA member 49.6% 43.0% 34.5% 30.5% 30.0% 37.0% 33.9% 34.5% self 31.2% 22.2% 26.0% 29.4% 29.1% 22.3% 26.7% 27.9% family 20.8% 19.3% 20.6% 18.6% 18.6% 17.4% 16.4% 17.0% rehab 18.4% 17.7% 27.6% 33.2% 33.7% 41.7% 35.6% 28.1% cnselling 4.8% 3.9% 12.1% 9.7% 12.2% 11.6% 8.5% 9.2% doctor 5.6% 8.7% 7.8% 7.1% 7.8% 5.2% 6.1% 5.5% group 93.5% 92.1% 88.0% 87.3% 88.3% 93.1% 92.3% 85.6% sponsor 85.6% 84.7% 87.0% 86.7% 80.9% 85.0% 87.5% 84.2% hlp bfore 58.1% 59.0% 68.4% 69.6% 69.2% 74.3% 73.9% 65.3% hlp aftr 64.2% 61.2% 63.6% 59.1% 63.3% 56.1% 62.2% 58.9% drugs 32.5% 36.9% 49.9% 45.2% 39.6% 40.6% 45.0% 52.4% mths sber 60 64 45 44 52 60 46 48 #in smpl 125 717 2607 915 1668 466 587 2309

– – 20 –

APPENDIX G

PROGRESS IN THE FELLOWSHIP

Introduction

As remarked elsewhere in this document, there is a great deal of information about the A.A. Fellowship implicit in the raw data. An earlier example is the findings regarding survival in the first year, in Appendix C.

In this Appendix some further results regarding progress of members in the Fellowship are described. The data allows some rough inferences to be made, illustrated by statements such as:

About 40% of the members sober less than a year will remain sober and active in the Fellowship another year.

Similarly, of the members sober five years, about 90% will remain sober and active in the Fellowship another year.

There is one extremely important qualification of such statements. They do not predict the number that will remain sober. Those that remain sober but do not remain active in the Fellowship cannot be estimated from our data. The statements only give lower limits to the number that will remain sober.

We use the term survival rate to describe this joint characteristic of sobriety and activity in the Fellowship.

Progress in the Fellowship

To describe how such approximations can be arrived at, Figure 1 is presented. The raw data in the Figure are the percentages of the 1989 sample that have been sober the indicated number of years, indicated by open circles connected by dotted lines. The corrected data (as will be described) are shown in solid circles and connected by solid lines.

– – 21 –

Figure 1

These plots are on semilog scales, which means that a straight line connects data that has a constant percentage change each year. The raw data can be fitted rather well by the three straight lines (although surely by other ways as well) indicating (1) that the number in the sample sober between 1 and 2 years is 38% of the number sober less than a year; (2) that the number sober 2, 3, or 4 years is about 76% of the number sober one less year, and that the number sober 5, 6, 7 or more years is about 84% of the number sober one less year.

It is natural to think of these percentages as survival rates of members in the Fellowship, but a little thought suggests that they must be modified in at least two ways. The most important way can be illustrated by the following example: those in the sample that are sober ten years must have come into the fellowship at least ten years previously. (Neglecting differences from any that might have been sober before they came to A.A.) Thus they were part of a much smaller membership than the 1989 sample represents (the membership in 1979 was 45% as large as the membership in 1989), and they represent a larger percentage of that 1979 group than of the 1989 sample which the raw data give. The former will therefore be the meaningful survival percentage for that group. And, of course, similarly for those sober any other number of years.

Among the approximations that are involved in this modification of the data of the 1989 survey are that many of those sober ten years may have been in the Fellowship a longer period of time, and some may have been sober before coming to the Fellowship. These two effects will, to some extent, cancel each other. Another approximation is that the group of newcomers in any year is proportional to the total membership that year.

– – 22 –

One other key approximation: the interpretation of the decline from the corrected percentage of 9-year people to the corrected percentage of 10-year people (say) as a survival rate depends also on the observation that the distribution in sobriety in the survey samples has shown reasonable similarity from 1968 forward; that is to say, there is a kind of constancy in operation here. Figure 2 illustrates this.

Survey Year Sober < 1 Sober 1-5 Sober 5+ 1977 37.3 38.0 24.7 1980 36.4 37.2 26.4 1983 37.7 36.9 24.9 1986 32.8 38.4 29.0 1989 34.6 36.4 28.9

Figure 2

The second, much less drastic, correction is to apply the frequency of attendance at meetings, as discussed in Appendix D (the section on attendance rates.) The frequency of those sober less than a year is about 10% greater than that of members with 5 years and more, so that they are somewhat overrepresented.

Applying these two corrections to the raw data gives the corrected data of Figure 1, or in summary:

(1) About 41% of those sober less than a year will remain sober and active in the Fellowship another year.

(2) About 83% of those sober less than five years will remain sober and active in the Fellowship another year.

(3) About 91% of those sober five years or more will remain sober and active in the Fellowship another year.

The calculations described has been repeated for surveys from 1974 forward and have always given results within a few percent of these values. Because of the approximation involved, we have generally cited 40%, 80% and 90% as the three numbers in question.

There seems to be no acceptable way to follow up individuals who leave the Fellowship to determine the reasons, although we are all familiar with many examples of members with many years of sobriety who no longer attend meetings. Therefore, it is important to re-emphasize that the survival rates refer to members still active in the Fellowship, and the numbers, particularly for significant lengths of sobriety, are lower limits to the percentages of those remaining sober in or out of the Fellowship; that is, they are less than what might be called “success rates.”

– – 23 –

It should be apparent that such calculations are the development of an approximate description of average properties of the A.A. Fellowship, and by no means rigorous statistical findings. The methods are more akin to those of operations research or systems analysis. They are useful descriptions for the use of the Fellowship, but averages so broad that professionals may wish for more detailed research that is beyond the purposes and Traditions of A.A.

Other Data

Three sets of data are required for the construction of Figure 1: 1989 survey data on length of sobriety, membership numbers compiled by the General Service Office of A.A. from 1970 forward, and average frequencies of attendance at meetings for members with different lengths of sobriety.

Length of Sobriety (Data of 1989 survey)

Sobriety Range % of sample 0-1 years 34.5% 1-2 13.3% 2-3 9.8% 3-4 7.4% 4-5 5.8% 5-10 17.2% 10-15 6.8% 15-20 2.8% 20-25 1.0% 25-30 0.5% 30-35 0.3% 35-40 0.1% 40-45 0.1% 45 plus 0.0% no response 0.4%

– – 24 –

Membership, US & Canada

Year Membership Year Membership (thousands) (thousands) 1970 193 1980 476 1971 211 1981 520 1972 244 1982 585 1973 261 1983 656 1974 331 1984 702 1975 337 1985 751 1976 369 1986 804 1977 404 1987 853 1978 410 1988 917 1979 445 1989 979

Average Frequency of Attendance (Meetings per week)

Sobriety Frequency Sobriety Frequency (years) (years) 0-1 3.04 15-20 2.74 1-2 3.03 20-25 2.70 2-3 2.95 25-30 2.93 3-4 2.91 30-35 2.87 4-5 2.88 35-40 2.65 5-10 2.82 40-45 3.08 10-15 2.73 45+ 2.98

5M/12-90/TC