The negativity bias, revisited: Evidence from neuroscience measures and an individual differences approach

ABSTRACT

Past research has provided support for the existence of a negativity bias, the tendency for negativity to have a stronger impact than positivity. Theoretically, the negativity bias provides an evolutionary advantage, as it is more critical for survival to avoid a harmful stimulus than to pursue a potentially helpful one. The current paper reviews the theoretical grounding of the negativity bias in the Evaluative Space Model, and presents recent findings using a multilevel approach that further elucidate the mechanisms underlying the negativity bias and underscore the importance of the negativity bias for human functioning.

KEYWORDS:

• ERPs
• fMRI
• neuroticism
• personality
• gender
• age

Acknowledgments

I would like to acknowledge my collaborators on the research reviewed in this paper, including (alphabetically): Kristin (Wood) Flanary, Zachary Ingbretsen, JS Irick, and George Monteleone. In addition, I thank Sky Deswert for her help with conducting multiple literature searches on the negativity bias, threat bias, and other related topics for this review.

Disclosure statement

No potential conflict of interest was reported by the author.

Notes

¹ The ESM predicts a second asymmetry in affective processing, the positivity offset, which is dominant at low levels of input to the systems and results in a slightly greater output for positivity than for negativity; this second affective asymmetry is beyond the scope of the current paper and will not be discussed further. See Norris et al. (2010) for a theoretical perspective on the positivity offset and Norris et al. (2011) for empirical evidence.

² Note that this is only one (arguably the strongest) prediction of the ESM regarding the negativity bias. Based on other aspects of the model, we would also predict that (a) the slopes of the activation functions relating input to output should differ, such that the negativity slope is steeper than the positivity slope, and (b) the degree of curvature (i.e., negative deceleration) of the activation functions should differ, such that the positivity slope decelerates at a higher rate than the negativity slope. All three predictions are made by the ESM; but comparing output of the two systems at higher levels of (equivalent) input is the most straightforward, empirically valid hypothesis.

³ Note that other approaches to studying differences in emotional responses to pleasant and unpleasant stimuli have provided evidence both in support of and at odds with the functioning of a negativity bias. For example, research using facial expressions (e.g., happy vs. sad/angry/fearful faces) often has shown that negative/unpleasant expressions elicit stronger behavioral, neural, and physiological responses to the latter than the former (e.g., Adolphs, 2008; Leppänen et al., 2007; Yang et al., 2012; but cf. Kauschke, Bahn, Vesker, & Schwartzer, 2019 for a more nuanced view across development). To the contrary, studies using categories of IAPS images have sometimes shown equivalent or even stronger responses to pleasant (e.g., erotica) than to unpleasant (e.g., mutilation, snakes) images (e.g., Weinberg & Hajcak, 2010; for sex differences in this approach cf. Bradley, Codispoti, Sabatinelli, & Lang, 2001; Sabatinelli et al., 2004). Both types of evidence, however, fail to meet the criterion of equivalency and are therefore beyond the scope of the current paper.

⁴ It is important to note here the difference between “ambivalent” and “ambiguous” stimuli. “Ambivalent” stimuli are those that are “ambi-“ (both) “valent”; or both positive and negative at the same time. “Ambiguous” stimuli are those that do not inherently carry one specific valence but may be interpreted as either positive or negative (e.g., surprise faces). The line of research discussed here concerns the latter.

⁵ It is also worth noting that participants in Bradley et al. (2003) were all male; we return to gender differences in the negativity bias and to neural and behavioral responses to erotica and mutilation pictures later.

⁶ Note that we argue that both process-based measures, which do not rely on self-reports of how one “typically” feels or responds, and more traditional reflective measures (i.e., surveys; such as the Behavioral Inhibition and Activation Scales [BIS/BAS; Carver & White, 1994] and the Positive and Negative Affect Schedules [PANAS; Watson, Clark, & Tellegen, 1988]) both are important for understanding individual differences in affective processing. Interested readers are referred to Norris and colleagues (2011) for more discussion of these issues.