Self-Report Bias in Substance Use Disorder Treatment Outcomes

We typically determine effectiveness of psychosocial treatments for substance use disorders (SUD) using self-report measures.

Although in some cases, self-report can be validated through behavioral measures, like toxicology screens, researchers do not or cannot always use some corroborating measure of functioning.

In cognitive-behavioral SUD treatment for example, several gold standards for acquisition of cognitive-behavioral skills, like self-efficacy, coping, and core beliefs and attitudes regarding substance use, do not lend themselves easily to behavioral assessment and are often measured simply by asking participants.


If, however, patients have external motivation to bias their self-report in favor of improvement – say if they believe legal status depended on treatment progress – these self-reports may not be a valid indicator of treatment response.

In the current study, Davis, Doherty, and Moser examined this very idea with a sample of 1747 patients in a correctional facility, referred if alcohol and/or drug misuse was implicated in their offense and they self-reported problematic substance use. Then, level of misuse severity – mild, moderate, or severe — was determined by screening questionnaire.


Patients with severe substance misuse were referred to a 5-month high-intensity program (n = 316) and those with mild and moderate substance misuse were referred to a 6-week moderate-intensity program (n = 1431).

Authors investigated the role of social desirability, comprised here of “impression management” (denying behavior that might be perceived as “bad” or “wrong”, e.g., drinking to intoxication), and “self-deception enhancement” (exaggerating personal skills, qualities, or attributes) among participants’ reported response to cognitive-behavioral treatment, measured by beliefs about substance use, recovery coping, abstinence self-efficacy, and substance-related locus of control. An internal locus of control, or belief that one has control and responsibility for the outcomes they experience, was viewed as positive response to treatment.

Across all measurements of response, participants reported improvement in functioning from pre- to post-treatment with large effects (Cohen’s d = .8 or higher) or greater in the high-intensity program and moderate effects (Cohen’s d ~ .5) in the moderate-intensity program.

Interestingly, social desirability also increased during treatment, suggesting that patients were more likely to present themselves in a favorable light and deny common flaws.

In line with the authors’ predictions, when controlling for pretreatment levels of the treatment response indicator (e.g., abstinence self-efficacy), as well as pretreatment levels of the social desirability component, both impression management and self-deception enhancement were significant predictors of each indicator of treatment response.

Social desirability accounted for 5% of beliefs about substance use in the moderate-intensity and 7% in the high-intensity group, for 7% of beliefs about substance use cravings in the moderate intensity and 6% in the high-intensity group, for 8% of abstinence self-efficacy in the moderate intensity and 15% in the high-intensity group, for 2% of locus of control (i.e., greater internal) in the moderate intensity (only non-significant result) and 8% in the high-intensity group, and 8% of recovery coping in the moderate intensity and 3% in the high-intensity group.


The current study highlights the value of measuring and including personality traits in susbtance use disorder (SUD) recovery research.

The authors showed that part of what appears to be treatment response is better accounted for by patient’s desire to be perceived in a positive light.

To be sure, the proportions of response accounted for by social desirability are meaningful (and statistically reliable in almost all cases), but do not account for all of the observed response. That said, without considering desirability, the results are not entirely accurate and may in fact be misleading. The study has implications not only in forensic settings where external motivation is salient, but also in a variety of other SUD treatment settings where patients’ return to work or school environments as well as ones where family support may be contingent on successful treatment completion.

Note that presenting oneself in a favorable light is adaptive in many circumstances – for example, in job interviews. The point here is not that the patients were being “dishonest”, but rather that the natural human phenomenon of enhanced self-presentation when one’s well-being is at stake calls into question the use only of self-report measures in these cases.

In addition to social desirability, others have shown that neuroticism, the vulnerability to react with unpleasant emotion to stressful situations, hinders the acquisition of recovery skills in SUD treatment.


Davis, C. G., Doherty, S., & Moser, A. E. (2014). Social desirability and change following substance abuse treatment in male offenders. Psychology of Addictive Behaviors, 28(3), 872.