The dispositional need for cognitive closure indirectly predicts mock jurors’ sentencing decisions through right-wing authoritarianism

Abstract

One reason people are motivated to hold right-wing authoritarian beliefs is the need to manage uncertainty. Right-wing authoritarianism provides a stable source of black-and-white ‘answers’ about the social world – obey established authorities and norms and show hostility to deviants. Right-wing authoritarianism, in turn, is positively associated with more punitive attitudes and judgements. The purpose of this study was to examine the relationship between mock capital jurors’ need for cognitive closure and sentencing decisions through right-wing authoritarianism. Four-hundred and fifty-one jury-eligible adults read a hypothetical capital case, weighed aggravating and mitigating evidence and chose a sentence. They also responded to items measuring right-wing authoritarianism and the need for cognitive closure. The need for cognitive closure was indirectly related to choosing a death sentence through right-wing authoritarianism and the weighing of aggravators and mitigators. Theoretical and practical implications, as well as directions for future studies, are discussed.

Keywords:

• death penalty
• motivated social cognition
• need for cognitive closure
• right-wing authoritarianism
• sentencing

Notes

1 MTurk is a platform by which researchers (and others) can recruit people to complete ‘Human Intelligence Tasks’ (HITs), such as a survey (see Buhrmester et al., 2018; Paolacci & Chandler, 2014, for review). MTurk samples are more advantageous than other convenience samples in that they tend to be more representative (e.g. Heen et al., 2014; Paolacci et al., 2010) and more attentive (e.g. Hauser & Schwarz, 2016; Peer et al., 2014). MTurk samples are widely used for social scientific studies in general (e.g. Pogarsky et al., 2017) and for capital jury decision-making studies in particular (e.g. Vaughan et al., 2019).

2 We used one criterion for jury eligibility: participants must have voted in the most recent presidential election at the time of the survey.

3 Racial/ethnic identity and gender identity were open-ended questions. Approximately 10% of participants did not respond to these questions.

4 Beyond this, six participants had missing data on at least one item. These participants were not excluded because the estimation routines we used allow for the information the participants did provide to contribute to model estimation (see Asparouhov & Muthén, 2010; Schafer & Graham, 2002).

5 Maximum likelihood estimation with robust standard errors and geomin rotation were used (see Browne, 2001, for review of geomin rotation). This estimation routine is robust to departures from normality and produces a scaled chi-square test statistic that is equivalent to Yuan and Bentler’s (2000) T₂* statistic. The standard errors are computed using a sandwich estimator (see Freedman, 2006). Item responses were treated as continuous because they were on 7-point scales (see Rhemtulla et al., 2012).

6 The root mean squared error of approximation (RMSEA) value was .103, which surpassed thresholds for what is generally considered a good, a mediocre and even a poor fit (Kenny et al., 2015; MacCallum et al., 1996). The values of the comparative fit index (CFI) and the Tucker–Lewis Index (TLI) were .844 and .818, respectively. These values fell below thresholds for good and reasonable fit (Bentler, 1990; Hu & Bentler, 1999). In contrast to these criteria, the value of the standardized root mean squared residual (SRMR) was .064, which did suggest a good fit (Hu & Bentler, 1999). However, considered holistically, we did not deem the model an acceptable fit to the data.

7 This routine uses probit regression for the categorical variables. A full weight matrix is used to calculate the chi-square statistic, while a diagonal weight matrix is used to calculate standard errors. The analysis was conducted using Mplus (Muthén & Muthén, 1998–2017).

8 Specifically, this method involved random resampling of the data 5000 times, multiplying path coefficients, and using that distribution to construct confidence intervals.

9 There is a notable weakness in trying to use death qualification standards to exclude prospective jurors, however. Namely, trial judges have discretion in determining whether a particular juror is excludable according to death qualification standards (Uttecht v. Brown, 2007). This could be, in part, why excludable jurors are not always, in fact, excluded (Dillehay & Sandys, 1996; Sandys & Trahan, 2008), as well as why there is variation across jurisdictions in the success of challenges for cause (Nietzel et al., 1987).

10 This has been explicitly noted by courts. For example, ‘Counsel is not . . . required to present all mitigation evidence. . . . Considering the realities of the courtroom, more is not always better. Stacking defenses can hurt a case. Good advocacy requires ‘winnowing out’ some arguments, witnesses, evidence and so on, to stress others (Chandler v. United States, 2000). ‘Good advocacy’ might be tied to the type of defense attorney. Some studies indicate that defendants with a privately-retained attorney are less likely to receive a death sentence than defendants with a court-appointed attorney or public defender (Kramer et al., 2017).

11 Although not focused on capital cases specifically, Bornstein and McCabe (2005) review studies on the issue of ‘consequentiality’.

12 A recent meta-analysis does suggest, however, that the trial medium presentation moderates the relationship between sample type and sentencing decisions (Bornstein et al., 2017). In short, the combination of a student sample and written materials might be problematic. See also Bornstein (1999).