close× Call Us
close×
Title How Group Identity Shapes Opioid Treatment Policy Opinion
Post date 05/15/2019
C1 Background and Explanation of Rationale

Each day, nearly 200 Americans die from a drug overdose, making overdose the leading cause of death for Americans under age 50 (Katz, 2017; Sanger-Katz, 2018). The rise in overdoses has been driven by the use of opioids: two-thirds of overdoses in 2016 involved an opioid (SAMHSA, 2016). Though opioid use and overdoses are not a new phenomenon, the current crisis has permeated the media with stories of substance use and addiction.

What has caused this increased attention by both the media and the public to drug addiction? Popular narratives have focused on how the geography of the current crisis reaches more rural, whiter, conservative, and less wealthy parts of the United States (Jalal et al., 2018; Keyes et al., 2014). In turn, substance users depicted in media stories on the current opioid overdose epidemic have been whiter and from more rural areas than those of the crack scare (Harbin, 2018).

This difference in the identities of substance abusers may lead members of the public to support these policies because of their shared group membership. In other words, sharing racial, political, or location-based identities with policy beneficiaries may make people more sympathetic to those beneficiaries. This shared identity with people affected by drug addiction may even cross-pressure partisans who traditionally oppose drug treatment (Meier, 1994) and spending on social policy to instead support these type of policies (e.g., Klar, 2013, 2014; Lazarsfeld, Berelson, and Gaudet, 1944; Mutz, 2002).

In this project, we propose to empirically assess these questions using a factorial survey experiment that varies attributes of a news article on a substance user seeking help through treatment. By altering features of the recovering substance user/policy beneficiary, we will assess how both the identity and pathway to addiction of those beneficiaries can affect perceptions of deservingness and, consequently, policy support.

C2 What are the hypotheses to be tested?

Group Identity

H1: We expect that the race, gender, and location in a rural or non-rural location of policy beneficiaries depicted in a media story will affect support for treatment and punitive policies, operationalized as support for increasing funding for opioid treatment policy and punitive policy. Specifically, for the full sample, we expect a decrease in support for funding after reading about a black policy beneficiary compared to a white policy beneficiary.

H2: We expect that respondents will be more sympathetic to policy beneficiaries who share identities with the respondent -- e.g., black respondents will be more sympathetic to black policy beneficiaries depicted in the media, while white respondents will be more sympathetic to white policy beneficiaries. Viewing a profile with a shared identity will increase respondent support for funding treatment policy.

H3: We expect that any experimental treatment which increases respondent support for funding treatment policy will also decrease respondent support for funding law enforcement to arrest and prosecute drug users.

H4: We expect the ‘blame' outcome variable to negatively correlate with support for funding treatment policy and positively correlate with support for funding punitive policy.

Pathway to Addiction

H5: We expect that respondents will be more sympathetic to policy beneficiaries who become addicted to opioids via legally obtained prescription drugs linked to a legitimate medical need, e.g. knee surgery. This will make them more supportive of increased funding for treatment. Therefore, as discussed above, respondents will also be less supportive of increased punitive measures and less likely to believe that individuals are to blame for their own addiction.

Insurance Coverage

H6: We expect that when people receive treatment for addiction from private insurance, respondents will be more supportive of treatment funding than when it is provided by ACA-subsidized insurance or Medicaid.

H7: We expect that support for the ACA will be higher when the news article features a recovering addict who is able to obtain treatment via ACA-subsidized insurance.

Political Divisions

H8: We expect that when a black policy recipient is depicted in the story Republicans will be relatively less supportive of treatment program funding than when a white person is depicted in comparison to Democrats (i.e. a more negative treatment effect among Republicans).

H9: We expect that when a rural policy recipient is depicted in the story Republicans will be relatively more supportive of treatment program funding than when a non-rural person is depicted in comparison to Democrats (i.e. a more positive treatment effect among Republicans).

H10: We expect that when people began their addiction following surgery and a legal prescription for painkillers, Republicans will be relatively more supportive of treatment program funding than when addiction began with drugs offered at a party in comparison to Democrats (i.e. a more negative treatment effect among Republicans). This hypothesis stems from the Republican emphasis on personal responsibility when evaluating addiction.

H11: We expect that when people began their addiction with heroin, Republicans will be relatively less supportive of treatment program funding than when it began with non-heroin painkillers in comparison to Democrats (i.e. a more negative treatment effect among Republicans). This hypothesis stems from the historically racialized nature of heroin and other `street drugs'.

H12: We expect that when people receive treatment for addiction from private insurance, Republicans (Democrats) will be more (less) supportive of treatment program funding than when it is provided by ACA-subsidized insurance or Medicaid.

Personal Exposure

H13: We expect that respondents who have personally known someone who has struggled with addiction will express greater support for addiction treatment funding.

Outcome Variables
1. If you were making up the budget for the federal government this year, would you increase, decrease, or keep spending the same for treatment for those addicted to opioids?
a. Increase a lot
b. Increase a little
c. Keep the same
d. Decrease a little
e. Decrease a lot

2. If you were making up the budget for the federal government this year, would you increase, decrease, or keep spending the same for law enforcement to arrest and prosecute those addicted to opioids?
a. Increase a lot
b. Increase a little
c. Keep the same
d. Decrease a little
e. Decrease a lot

3. Given what you know about the Affordable Care Act (also known as “Obamacare”), do you have a generally favorable or unfavorable opinion of it?
a. Strongly favorable
b. Somewhat favorable
c. Neither favorable nor unfavorable
d. Somewhat unfavorable
e. Strongly unfavorable

4. Would you agree or disagree that individuals addicted to opioids are to blame for their own addiction?
a. Strongly agree
b. Somewhat agree
c. Neither agree nor disagree
d. Somewhat disagree
e. Strongly disagree

C3 How will these hypotheses be tested? *

Shared Identity

The following approaches will be used to test each hypothesis:

H1: Two-tailed t-tests of difference in means of treatment funding support between each vignette treatment group (e.g., ‘rural') vs. all the others in that category (e.g., ‘suburban' and ‘urban'), using one treatment group in each category as the baseline category.

H2: Two-tailed t-tests of support for treatment funding on each identity treatment within respondent subgroups detailed below. Second, interact identity treatment effect with an indicator for the respondent's identity subgroup. Third, measure effect of shared identity via omnibus model using new indicator for a shared identity between each respondent to the identity treatment they received.
• For the effect of the race experimental manipulation, subgroups by survey respondent race/ethnicity (black vs. non-Hispanic white).
• For the effect of the gender experimental manipulation, subgroups by respondent gender (male vs. female).
• For the effect of the location experimental manipulation, subgroups by respondent location (rural vs. non-rural, as well as three subgroups matching the manipulated levels of rural, urban, and suburban).

H3: Conduct tests for H1 and H2 using punitive outcome. Repeat additional tests below using punitive outcome.

H4: Conduct tests for H1 and H2 using blame outcome. Repeat additional tests below using blame outcome.

Pathway to Addiction

H5: Two-tailed t-tests for the difference in means of treatment funding support between surgery as pathway to addiction vs. all other addiction pathways.

Insurance Coverage

H6: Two-tailed t-tests for the difference in means of treatment funding support between the private insurance treatment group vs. all forms of insurance.

H7: Two-tailed t-tests for the difference in means of ACA support between the private insurance treatment group vs. all other forms of insurance.

Political Divisions

H8: Two-tailed t-tests of support for treatment funding on racial treatment. Interact treatment effect with indicator for respondent partisanship (Republican vs. Democrat, coding leaners with nearest party).

H9: Two-tailed t-tests of support for treatment funding on location treatment, comparing rural to non-rural treatments. Interact treatment effect with indicator for respondent partisanship (Republican vs. Democrat, coding leaners with nearest party).

H10: Two-tailed t-tests of support for treatment funding on pathway to addiction treatment, comparing legal prescription from knee surgery to all other pathway to addiction treatments. Interact treatment effect with indicator for respondent partisanship (Republican vs. Democrat, coding leaners with nearest party).

H11: Two-tailed t-tests of support for treatment funding on pathway to addiction treatment, comparing heroin at a party to all other pathway to addiction treatments. Interact treatment effect with indicator for respondent partisanship (Republican vs. Democrat, coding leaners with nearest party).

H12: Two-tailed t-tests of support for treatment funding on insurance coverage treatment, comparing private insurance to all other insurance coverage treatments. Interact treatment effect with indicator for respondent partisanship (Republican vs. Democrat, coding leaners with nearest party).

Personal Exposure

H13: Two-tailed t-tests for the difference in means of treatment funding support between respondents based on personal exposure to those with opioid addiction.

C4 Country United States
C5 Scale (# of Units) The experiment will involve 3,100 survey respondents
C6 Was a power analysis conducted prior to data collection? Yes
C7 Has this research received Insitutional Review Board (IRB) or ethics committee approval? Yes
C8 IRB Number 18-07-12
C9 Date of IRB Approval 02/06/19
C10 Will the intervention be implemented by the researcher or a third party? NORC using their AmeriSpeak panel
C11 Did any of the research team receive remuneration from the implementing agency for taking part in this research? No
C12 If relevant, is there an advance agreement with the implementation group that all results can be published? Implementation group is contracted by TESS to provide survey respondents and run the survey only.
C13 JEL Classification(s) not provided by authors