Methods
Participants
We collected data from 126 participants recruited via Amazon Mechanical Turk. The sample size was determined by a power analysis based on the effect size for the effect of article type on interrace contact motivation scores reported in Williams & Eberhardt (2008). A power analysis was conducted on GPower 3.1 for a one-way ANOVA testing differences between 3 groups. The power analysis determined that a sample size of 126 was required to achieve 80% power. The procedures reported below were approved by the McMaster University Research Ethics Board (project ID: 5593).
Materials
Participants read one of three news articles (see Appendix A). In the pro-essentialism condition, participants read a news report about a scientific study that had discovered the genetic basis of opioid addiction. In the anti-essentialism condition, participants read a news report about a scientific study that had determined there was definitively no genetic basis of opioid addiction. These two articles were modified articles from Williams & Eberhardt (2008). Additionally, there was a control article, about the discovery of new dinosaur fossils and completely unrelated to opioid addiction or essentialism.
After reading the articles, participants completed three questionnaires. The first questionnaire (see Table 1) was an essentialism questionnaire based on Bastian & Haslam (2008) that we adapted to be about addiction rather than a general essentialism questionnaire. The second questionnaire was an addiction stigma questionnaire, which combined items from Kennedy-Hendricks et al. (2017) and Barry et al. (2014). Items adapted from Kennedy-Hendricks et al. (2017) primarily asked about perception of people with addiction (e.g., “People with an addiction are more dangerous than the general population”). Items adapted from Barry et al. (2014) asked about interacting with people with addictions (e.g., “Would you be willing to have a person with drug addiction work closely with you on a job?”) as well as statements about the perception of social support for people with addiction (e.g., “I am in favour of increasing government spending on the treatment of addiction.”) See Table 2 for the full addiction stigma questionnaire. Lastly, there was a personality questionnaire that included items based on the Big 5 personality structure. These items were filler items to obscure the true nature of the study.
Table 1. Essentialism questionnaire adapted from Bastian & Haslam (2008). Reverse scored items are denoted by (R). Questions related to biological essentialism are denoted by an asterisk (*). Participants responded to each item with a 1-7 Likert scale.
|
The boundaries that define the differences between addicts and non-addicts are clear-cut.
|
A person either has addictive tendencies, or they do not.
|
There are different types of people (i.e., addicts or non-addicts) and those types can be easily defined and are relatively clear-cut.
|
The kind of person someone is, is clearly defined, they either are an addict or they are not.
|
People fall into distinct personality ‘types’.
|
Generally speaking, once you know someone in one or two contexts, it is possible to predict how they will behave in most other contexts.
|
It is possible to know about many aspects of a person once you learn they are an addict.
|
When getting to know a person, it is possible to determine if they are an addict or not very quickly.
|
Knowing that someone is an addict can lead to accurate predictions of their future behaviour.
|
Although addicts may have some basic identifiable traits, it is never easy to make accurate judgments about how they will behave in different situations (R).
|
With enough scientific knowledge, addiction can be traced back to genetic causes. *
|
Whether someone is an addict or not can be determined by their biological make-up. *
|
With enough scientific knowledge, the basic qualities of addicts can be traced back to, and explained by, their biological make-up. *
|
A person being an addict can largely be attributed to their genetic inheritance. *
|
Table 2. The addiction stigma questionnaire adapted from Kennedy-Hendricks et al. (2017) and Barry et al. (2014). Reverse scored items are denoted by (R). Participants responded to each item with a 1-7 Likert scale.
|
Individuals with an addiction are to blame for the problem.
|
Some people lack the self-discipline to use drugs without becoming addicted.
|
I am unwilling to have a person with an addiction marry into the family.
|
I am unwilling to work closely with a person with an addiction.
|
Discrimination against people with drug addiction is a serious problem. (R)
|
Employers should be allowed to deny employment to a person with drug addiction.
|
People with an addiction are more dangerous than the general population.
|
Landlords should be able to deny housing to a person with drug addiction.
|
The treatment options for persons with drug addiction are effective at controlling symptoms. (R)
|
Most people with drug addiction can, with treatment, get well and return to productive lives. (R)
|
I am in favour of requiring insurance companies to offer benefits for the treatment of drug addiction that are equivalent to benefits for other medical services. (R)
|
I am in favour of increasing government spending on the treatment of drug addiction. (R)
|
I am in favour of increasing government spending on programs that help people with drug addiction find jobs and provide on-the-job support as needed. (R)
|
I am in favour of increasing government spending on programs to subsidize housing costs for people with drug addiction. (R)
|
Procedure
Participants completed the study online via Amazon MTurk. They clicked a link in the study description and were sent to a page that hosted a consent form. Participants were informed in the study description that the study was a memory experiment, where they would read a news article, then complete a questionnaire, before completing a memory check. Upon clicking “Continue” on the consent form, participants were randomly assigned to either the pro-essentialism, anti-essentialism, or control news article. Participants could spend as much time as they needed to read the article and could read it as many times as they wished. Once they were done reading the article, participants began the questionnaires. Questions from all three questionnaires were intermixed and presented in a random order. Each question was presented with a 7-point Likert scale where participants clicked the option they agreed with (there was no starting point). Participants had the option to skip questions they were uncomfortable with. (This occurred on less than 2% of trials.) After completing the questionnaire, participants were presented with two memory check questions asking about the general topic of the article. Participants were then presented with a reconsent form where they were informed of the true purpose of the study and had the option to reconsent or withdraw. Participants were paid $7.50 CAD for their participation.
Data Availability Statement
The data that support the findings are openly available on the Open Science Framework (OSF): https://osf.io/ny7dq/?view_only=6a5e3d86239246a8bb93b6e2811f7557.
Results
Prior to analysis, 5 participants’ data were removed due to missing responses. As a manipulation check, we assessed the effect of article type on participants’ biological essentialism scores. We found that article type had a marginal effect on biological essentialism (F(2, 118) = 2.94, p = 0.05)[1]. Mean biological essentialism scores were highest for the pro-essentialism condition (M = 4.05, SD = 0.93) relative to the anti-essentialism (M = 3.57, SD = 1.45) or control (M = 3.41, SD = 1.27) conditions. To assess the effect of article type on addiction stigma, we conducted a one-way ANOVA. There was no significant effect of article type on stigma scores (F(2, 118) = 0.692, p = 0.50). As there was no significant effect of article type on either variable, the rest of the analyses collapses the data across conditions.
We broke down participants’ essentialism scores into biological essentialism and non-biological essentialism. Participants’ biological essentialism significantly correlated with their stigma toward individuals with opioid addiction (r = 0.31, p < 0.001, 95% CI [0.14, 0.46], see Figure 1). However, participants’ non-biological essentialism and addiction stigma was more strongly correlated (r = 0.53, p < 0.001, 95% CI [0.39, 0.65], see Figure 2). To test if there was a significant difference between the strength of the two correlations, we conducted a Fisher’s z-test for dependent correlations. The correlation between non-biological essentialism and stigma was significantly stronger than the correlation between biological essentialism and stigma (z = 3.029, p = 0.001).
Discussion
We found that priming participants with articles that promoted either biologically essentialist or anti-essentialist views about addiction did not affect levels of stigma around opioid addiction. Participants’ biological essentialism scores were significantly associated with their addiction stigma scores. However, participants’ non-biological essentialism scores were even more strongly correlated with stigma.
There has been a debate in the literature about whether genetic explanations will increase or decrease stigma surrounding mental illness (Phelan, 2002). Those arguing that genetic explanations will decrease stigma surrounding mental illness and addiction argue that it will decrease culpability (Phelan, 2002). Additionally, they argue that having information about the biological nature of addiction will suggest that addiction is treatable, and will lead people to advocate for more health-based treatment (Richter et al., 2019). Others argue that spreading information about the biological nature of addiction will lead to increased stigma because it makes addiction seem unchangeable (Phelan, 2002). Our data supports the latter argument: Participants’ biological essentialism was strongly associated with stigma against people with an opioid addiction. This suggests that believing that addiction is traceable to an individual’s biological make-up does promote stigma. However, it appears that non-biological essentialism may be an even stronger contributor to stigma.
Essentialism is associated with the perception that category membership is discrete and immutable (Haslam et al., 2000). The perception of addiction as immutable is a hallmark belief of addiction stigma (Ben-Zeev et al., 2010; Richter et al., 2019). This view may motivate punishment rather than health-based treatment (Richter et al., 2019). Additionally, the belief that addiction is a discrete category highlights the difference between individuals with addiction disorders and individuals without. The use of generic language, like calling someone an “opioid addict” rather than “a person with an opioid use disorder”, highlights the discreteness of the category boundary and is associated with increased stigma (Goodyear et al., 2018). The use of generic language also increases essentialist beliefs about categories (Rhodes et al., 2018). These findings suggest that a general essentialist framework accounts for many of the key aspects of stigma against individuals with opioid addiction. Additionally, it may be more important to focus on essentialist views of addiction generally rather than focusing specifically on the effect of biological explanations of addiction when trying to understand sources of prejudicial beliefs.
[1] Some argue that ANOVAs are not ideal for Likert scale data, and that mixed-effects linear models are superior (Kizach, 2014). We replicated the same analyses with mixed-effects linear models and found the same results.