Donate to the Baldwin Research Institute

Addiction & Treatment

1992 Gathering Research to Increase Program Effectiveness

Observations and Methodology

In 1992, the researchers purchased a 125 year old broken-down Victorian mansion for the purpose of conducting a study on the efficacy of a social/education program (later to develop into a social/learning program) to help people deal with alcohol and drug problems. The researchers settled on the social/educational model for two specific reasons. First, these were the two components observed in the previous studies that had produced the 75% success rate. While at the time of the Baldwin Research Study of 1990, the social aspect of the study was, for the most part, ignored, as time went on it became apparent that, indeed, social influences played a significant role in the study subjects' commitment to accepting the idea that they were "recovered." Nevertheless, the educational aspect of the still undefined program was far more obvious to the researchers. The educational aspect utilized the methodology of presenting to the subjects, that contrary to Alcoholics Anonymous' dogma, the study subjects could, in fact, recover and return to their normal selves. Further, throughout the 1980's references to the beneficial effects of socialization and educational approaches with respect to helping people deal with substance use problems had been made in some obscure research.

By 1993, the researchers were unable to continue their observational studies because the number of subjects had grown and many lived remotely, thus making it impossible to observe them on a daily basis. To cope with these new circumstances the researchers adopted a more conventional method of study: polling, self-reporting and a corroborating observer. During this time, until 2002, the researchers still polled 100% of the program attendees. The researchers were conducting one poll a year. Each subject signed an agreement authorizing Baldwin Research Institute to contact them each time Baldwin Research Institute conducted a poll. Additionally, each subject provided the name and contact information for Baldwin Research Institute to contact a third party to corroborate the subjects' self-reporting. [Note: This agreement was, and is, voluntary and is still used today at all Saint Jude Retreat facilities.] By signing this agreement, the subjects knew that they were being studied, including the original subjects of the Baldwin Research Study of 1990.

Even though the researchers were aware of the self-reporting study method alleged shortcomings, their concerns were lessened by two important considerations. First, because of the subjects' participation in the study, the researchers were known to the subjects, so talking with each other on the telephone was a familiar activity. Second, the study included confirmation of the subjects' reports by third parties, their corroborators. The corroborator was an extension of the observational study technique, i.e. the corroborator was an observer of the subject's behavior daily or at some regular intervals. Corroborators, however, served a secondary function beyond simply corroborating or refuting the report of the subject. The study subjects most certainly took into consideration, prior to answering the poll question(s), whatever answer they gave would be confirmed or contradicted by each subject's corroborator. So, what were the results?

From 1992 to 2004 the results ranged from a low of 53% to a high 87%. The mean was 71.923% for the first 13 years of the study.

The first 13 years of research was not conducted for publication; it was being conducted for BRI internal purposes only. For example, from time to time new and differing information was introduced into the educational aspect of the program and into the social aspect of the program, and the research we conducted was used to measure trends that may occur related to these changes. That is to say that the researchers were tracking variances from year to year as opposed to conducting studies to determine abstinence rates. The reality is that any drug and alcohol program that conducts its own abstinence rate studies and then actually publishes it as their abstinence rate is delusional. By way of explanation, efficacy studies of treatment programs, educational programs, or any kind of program that is designed to help people with drug and/or alcohol problems are subject to unavoidable flaws.  Furthermore, when these studies are conducted by the same drug and alcohol program that provides the treatment, education, etc, additional critical flaws are introduced. The additional flaws are (1) interviewer effect, creating (2) response expectancy.

In the case of Baldwin Research Institute's polling, the attendees of the then Tri-Key Program and later the Jude Thaddeus Program, interviewers were known to the respondents as providers of the program services that were being asked about (polled). To wit, the Baldwin Research studies conducted from 1993 through and including 2004 were influenced by the phenomena of interviewer effect and response expectancy. This is a natural occurrence among respondents when the interviewer is known to the respondent. Response expectancy occurs when a respondent provides a response that is pleasing to the interviewer. This interviewer effect does not affect the total population, but it is not known to what extent it increases the frequency of positive responses. Response expectancy does influence, to some extent, all respondents. This particular phenomenon actually comes into play with virtually all polling studies. It is the result of respondents anticipating an expected response to the question and then giving that expected response, without regard for accuracy.

Taking into account all the previously discussed limitations, the absolute values of the study results were:

Taking into account all the previously discussed limitations, the absolute values of the study results were:



























For the record, it is acknowledged that these results are inflated due to interviewer effect and response expectancy. However, there is no exact way of determining to what extent these two variables, interviewer effect and response expectancy, affected the results. Nevertheless, it is reasonable to trust that by limiting the study to the same single interviewer, using the same questions, with the same population characteristics, year after year, the results comparing year to year were predominately influenced by program performance. This 10-year study, 1995 through and including 2004, yielded meaningful results insofar as determining the trendline for the data. Over that 10-year period the trend was +8% for the entire period during which the Baldwin researchers migrated the program from an Alcoholics Anonymous based program to a non-Alcoholics Anonymous, positive-reinforcement based program.


learn about our researched programs