Words Matter: How we describe individuals with substance problems affects perceptions

Language matters and language around substance use has evolved to decrease the stigma of substance use disorders.  The terms ‘addict’ and ‘addiction’ share negative associations that can be measured implicitly and explicitly and may pose a barrier to pursuing behavioral health treatment.


Terminology in the addiction field is evolving in recognition of the impact caused by the implications of certain labels.  Highly stigmatized views of labels such as ‘alcoholic’ has been associated with reduced rates of pursuing treatment. It has been shown also that there is explicit bias even amongst mental health professionals against individuals described as ‘substance abusers’ (as opposed to ‘having a substance use disorder’).  However, how we perceive terms implicitly, automatically, often is the last to evolve.  Despite many studies examining explicit conscious bias, relatively little has been done to explore implicit bias, whether terms are perceived unconsciously as neutral, negative or positive, in the substance use disorder field.  Studying implicit bias has the additional advantage that it can be objectively measured, and may be present even after explicit biases have been corrected.


The Go/No-Go Association Taks (GNAT) is an established measure of implicit bias and was administered remotely after two practice blocks.


DESIGN: Using the Go/No-Go Association Taks (GNAT), four test blocks were administered, focused on measuring response times when the terms addict/ person with a substance use disorder were associated with the terms good/bad.


SETTING: The study was performed remotely, with participants using their own personal computers.


PARTICIPANTS: 25 adult participants recruited from around the United States using digital media platforms for individuals interested in or affected by substance use disorders with a final inclusion of 23 participants.


MEASUREMENT: The sensitivity measure (d’) was derived from the proportion of correct ‘go’ items for target words as compared with incorrect ‘go’ responses for distractor words as well as the latency of those correct and incorrect responses.  An example of a test block would include the target words “addict” or “bad”, with the distractors phrases/words being “substance use disorder” or “good”.  A different test block might have the target phrases/words be “substance use disorder” or “bad” with the distractors “addict” or “good”.  All four combinations of these terms were tested.


Both the words ‘addict’ and ‘substance use disorder’ are more strongly associated with the word bad than good (as measured by the proportion of correct responses when different combinations of these words were tested).


The association of the word ‘addict’ with ‘bad” was significantly stronger than the association of ‘person with substance use disorder’ with ‘bad.’  This suggests that person with substance use disorder is less stigmatizing than addict although the implicit associations are still mildly negative.


The associations did not differ significantly when participants were in recovery or whether they worked in the behavioral health field.  However, the data trended towards stronger negative associations for both the terms tested for those working in the mental health field as well as for those in recovery.  This was an unexpected finding but was not statistically significant.


Certain terms in the substance use field have strong negative associations (addict; substance abuser). Even the more neutral term -“person with substance use disorder” – has some negative associations (as measured by a quantitative test of subconscious word associations), although less so than the aforementioned terms.  Tests of implicit bias (objective, quantitative analyses) together with explicit self-report, could be used to continue to modify our language to further reduce the stigma of substance use disorders.

  1. This was a small pilot study with n=23. When the conclusions are no difference (e.g. recovery status or working in the behavioral health field), it could be that the sample was too small to detect a statistically reliable (i.e., significant) difference, rather than the “true” absence of an effect. Larger samples in future studies would help clarify this.
  2. Almost 90% of the sample was white and thus not representative of minority groups. Of note though is that 56% of the participants were female, in contrast to the majority of substance use disorder research.
  3. Per the authors themselves, implicit bias has not been found to accurately predict behavior. In this study however, they point out that the results are in line with studies of explicit bias.


  • For individuals & families seeking recovery: Words matter and working to change the language around substance use disorders may help individuals identify less with negative terms and more with neutral terms around substance use.  The language used by family and friends around substance use may influence how individuals feel they are perceived and supported.
  • For scientists: There is a need for future studies on the effects of bias or discriminatory behavior in the substance use disorder field and whether it affects treatment engagement or effectiveness.  For example, treaters with no implicit bias against those with substance use disorders may be more effective in their treatment interactions than those with biases, even if they are not explicitly aware of such biases.
  • For policy makers: The broad use of stigmatizing language throughout the general population can generate biases that impact the effectiveness of public policies designed to address the public health crisis around substance use disorders.  There is a pressing need to reach consensus on appropriate terminology (for example, see the RRI Addictionary™) to reduce the continued reinforcement of negative associations attributed to persons with substance use disorders. If we want addiction destigmatized we need a language that is unified and that more accurately reflects of the true nature of these conditions.
  • For treatment professionals and treatment systems: Implicit bias around substance use disorders exists in the mental health field (and the data presented here trends towards suggesting that bias is greater than in the general population, albeit with a small sample size).  Treatment professionals should prioritize consistently changing the terminology they use, but should also work on addressing their implicit associations, or at least monitor them, as it likely interferes with the effectiveness of treatment.  For example, clinicians often use the term ‘they’ to refer to persons with substance use disorders, as opposed to the less biased term ‘my patients.’


Ashford, R. D., Brown, A. M., & Curtis, B. (2018). The Language of Substance Use and Recovery: Novel Use of the Go/No–Go Association Task to Measure Implicit BiasHealth communication, 1-7.