Development of a Therapeutic Alliance Measure for 12-Step Attendees & Their Sponsors

During the past 20 years a growing number of studies with increasing levels of sophistication support the role of 12-step mutual-help organizations (MHO) such as Alcoholics Anonymous (AA) and interventions that facilitate engagement with 12-step MHOs in promoting abstinence and other health outcomes among individuals with substance use disorders (SUDs).

Individuals that attend 12-step mutual-help organizations (MHO) groups are often encouraged to get a “sponsor” who is a more experienced person in recovery that can provide support, guidance, and help him/her work the 12 steps.

Among the many behaviors and activities in which mutual-help organizations (MHO) members can participate, research shows that having a sponsor, specifically, promotes better abstinence rates. Zemore and colleagues, for example, found that among over 500 individuals with SUD participating in a trial of a 12-step facilitation intervention, meeting attendance and currently having a sponsor were the only variables that predicted abstinence across the entire year of follow-up assessments.


Although often measured with a simple yes/no question, the relationship between sponsor and sponsee may vary substantially from individual to individual and contain several relational nuances. However, very little is known about how to measure the quality of this key recovery relationship. In professional treatment, the “therapeutic alliance” is deemed crucial to successful treatment.


his is all about the quality of the relationship and whether the clinician and the patient are on the same page regarding the treatment expectations and clinical course. Despite the prominence of this construct in the professional realm, there is nothing known about this for mental health organizations (MHOs) despite MHOs being the most commonly sought source of help for substance use disorders (SUDs) in the U.S.

In the current study, Kelly and colleagues sought to develop a measure of the “therapeutic alliance” in the MHO realm. Specifically, the measure assessed the quality of the relationship an individual 12-step member has with his/her sponsor, and tested its psychometric properties; that is, how well does it measure the construct in question, and how well it relates to key recovery behaviors such as 12-step MHO attendance and abstinence.




The study used a sample of 157 young adults (ages 18-24) that attended residential treatment, had a 12-step mutual-help organizations (MHO) sponsor, and completed the new measure called the “Sponsor Alliance Inventory”(SAI) at one or more of the follow-up assessments (3 months, 6 months, or 12 months after being discharged from treatment). Initially, the SAI was a 12-item questionnaire that asked participants to rate their level of agreement from 1 (never) to 7 (always) on the nature and quality of the sponsor/sponsee relationship, particularly as related to the sponsee’s 12-step MHO participation. It was adapted from the widely used Working Alliance Inventory (WAI), assessing the relationship between therapist and patient in studies of psychotherapy.


Authors used several strategies to develop the measure and to determine its ability to assess an individual’s alliance with his/her sponsor:


  1. First, principal factor analysis was used to fine tune the measure so that only items that generally belong together (i.e., measuring sponsor alliance) were included
  2. Next, it tested whether this fine tuned version of the SAI differed by several demographic and clinical factors
  3. Finally, it examined whether SAI was statistically associated with relevant recovery behaviors including 12-step attendance and abstinence over time



Participant scores ranged from 7 to 70, with median (50th percentile) scores of about 60 at each of the follow-up assessments. Items were highly correlated with each other and with the overall sponsor alliance construct. Sponsor alliance scores generally did not vary much across gender and substance of choice signifying that it may be a good measure across different subgroups and populations.


Authors found that a 10 item questionnaire was the best way to assess sponsor alliance.

Controlling statistically for possible biases, “Sponsor Alliance Inventory” (SAI) was generally an excellent predictor of relevant recovery related constructs at:


  • SAI scores at 3 months and 12 months were associated with 12-step mutual-help organizations (MHO) attendance, 12-step MHO active involvement (a six item index of participation that excluded sponsor-related variables), and abstinence at 3-month and 12-month follow-ups, respectively
  • SAI at 6 months was associated with 12-step MHO attendance and active involvement at 6-month follow-up.


SAI was not as good at predicting these behaviors in the subsequent follow-up period (e.g., the effects of 3-month SAI on 6-month 12-step MHO attendance).

This study suggests that an individual’s alliance with his/her sponsor is a construct that can be reliably measured and is associated with related and important recovery behaviors such as 12-step MHO participation (i.e., has good criterion validity).


This study is among the first to develop a measure of the therapeutic alliance construct specific to the sponsor/sponsee relationship with attention to the nature and quality of the connection. In a majority of prior studies, the construct was measured by presence/absence of a sponsor, and at times, whether or not one has contact with his/her sponsor outside meetings. Such a measure could be an important piece of what explains benefit from 12-step MHO participation that is seldom included in clinical research.


The nature and quality of the alliance between young adults with SUD and their sponsors (among those who have sponsors) may be an important consideration in fully understanding and being able to predict their recovery experience and outcomes.

  1. The study sample included only young adults 18-24 who were mostly Caucasian; it is unclear if the SAI is a valid measure of the sponsor/sponsee relationship in other age and racial/ethnic groups.


Future work should determine whether “Sponsor Alliance Inventory”(SAI) scores predict 12-step mutual-help organizations (MHO) participation and abstinence over and above simply having a sponsor and/or having contact with one’s sponsor outside of meetings, as other studies have examined.


  • For individuals & families seeking recovery: Studies related to the one reviewed here suggest having a sponsor can have a substantial and positive influence on your recovery. More research is needed, however, to fully understand the relative importance of how well you work and get along with your sponsor.
  • For scientists: Although it should be tested in other settings and with more heterogeneous samples, the Sponsor Alliance Inventory holds promise as a measure of the quality and strength of the sponsor/sponsee relationship.
  • For policy makers: MHOs are freely available, widespread, and are the most commonly sought sources of help for substance use disorders. Consider greater funding dedicated to researching nuanced dynamics of the 12-step mutual-help organizations (MHO) experience to fully maximize its potential utility including research on the sponsor/sponsee relationship.
  • For treatment professionals and treatment systems: Studies related to the one reviewed here suggest having a 12-step MHO sponsor can have a substantial and positive influence on the recoveries of your patients. Encouraging patients to attend MHOs and obtain a sponsor is likely to enhance patients’ chances of recovery. More research is needed, however, to fully understand the relative importance of how well he/she works and gets along with a sponsor in terms of recovery outcomes.


Kelly, J. F., Greene, M. C., Bergman, B., Hoeppner, B. B., & Slaymaker, V. (2015). The Sponsor Alliance Inventory: Assessing the Therapeutic Bond Between 12-Step Attendees and Their Sponsors. Alcohol Alcohol. doi:10.1093/alcalc/agv071




**Note: One or more authors of this study were Recovery Research Institute Staff, including the director and/or other research scientists. As with all summaries, staff made the greatest possible effort to recognize and account for any potential biases in the review of this article.