Public Support for Solitary Confinement: A Randomized Experiment of Belief Updating and Confirmation Bias

ABSTRACT

This randomized control experiment investigates whether providing educational information about solitary confinement (SC) cause respondents to update their beliefs about the practice. It also examines whether prior views about punishment moderate the influence of the message content on changes in level of support for SC. Study participants were recruited from Amazon’s MTurk workplace to complete an on-line survey about their level of punitive orientation and support for SC. Participants were randomly assigned to one of two treatment conditions that provided differing messages about SC through brief, educational videos. Study results suggest participants who received information that SC is necessary for institutional order increased their support for the practice at post-test, while those who viewed information that SC is harmful to its inhabitants decreased their support for its use. This study did not find evidence that punitive orientation influenced the impact of the message received on changes in attitude toward SC.

KEYWORDS:

• solitary confinement
• administrative segregation
• restrictive housing
• prisons
• public support

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1. MTurk is an online crowdsourcing space that allows for the recruitment of workers to complete various tasks, such as online surveys and document transcription (Mullinix et al., 2015). Research indicates that participants recruited using crowdsourcing sites are more representative of the U.S. population than university samples (Berinsky et al., 2012), and further that MTurk participants pay more attention to survey instructions than college student participants (Hauser & Schwarz, 2015).

2. There were an additional 63 people who responded to the MTurk task, but failed the attention check questions. These individuals were not surveyed or included in this study.

3. Participants were asked to rate the conditions in three prison photos as “too harsh,” “too comfortable,” or “about right.” This information was not used in this study.

4. Prior to the current study, a separate sample of 41 participants were recruited from MTurk to assess if the two treatment conditions conveyed the intended message. After watching either the SC is necessary (n = 20) or SC is harmful video (n = 21), participants were asked a series of questions to identify its primary message. The results of this pilot test indicated that participants appropriately understood the content in both videos. Those in former group correctly identified 93.6% of the responses that aligned with the position that SC is necessary for institutional order, while those in the latter group correctly identified 98.8% of the responses that aligned with the position that SC is harmful to its inhabitants. Given these findings, we elected to use the same two videos in the current study without modification. The SC is necessary video can be viewed at https://youtu.be/HupNOJejBLg and the SC is harmful video can be viewed at https://youtu.be/MrO94XwBhwk.

5. This study applies Cohen’s (1988) guidelines for interpreting the magnitude of the effect sizes, where Cohen’s d =.2 represents a small difference, Cohen’s d =.5 represents a medium difference, and Cohen’s d =.8 represents a large difference.