Hello darkness my old friend: preferences for darkness vary by neuroticism and co-occur with negative affect^*

* Supplementary materials for this project are available in Open Science Framework under the project name “Preferences for darkness vary by neuroticism and co-occur with negative affect”.

ABSTRACT

Metaphors frequently link negative affect with darkness and associations of this type have been established in several experimental paradigms. Given the ubiquity and strength of these associations, people who prefer dark to light may be more prone to negative emotional experiences and symptoms. A five study investigation (total N = 605) couches these ideas in a new theoretical framework and then examines them. Across studies, 1 in 4 people preferred the perceptual concept of dark over the perceptual concept of light. These dark-preferring people scored higher in neuroticism (Studies 1 and 2) and experienced greater depressive feelings in daily life (Study 3). Moreover, dark preferences shared a robust relationship with depressive symptoms (Study 4) as well as generalised anxiety symptoms (Study 5). The results provide novel insights into negative affectivity and extend conceptual metaphor theory in a way that is capable of making individual difference predictions.

KEYWORDS:

• Neuroticism
• negative affect
• conceptual metaphor
• darkness
• preferences

Disclosure statement

No potential conflict of interest was reported by the authors.

ORCID

Michelle R. Persich http://orcid.org/0000-0001-5322-6685

Notes

* Supplementary materials for this project are available in Open Science Framework under the project name “Preferences for darkness vary by neuroticism and co-occur with negative affect”.

1 For all studies, we sought adequate power (.80) to detect medium effect sizes, which were posited (Robinson & Fetterman, 2014). With the two-group designs of Studies 1–4, 128 participants would give us this degree of power and we obtained sample sizes in this range. This was done by running the studies for a fixed period of time, one that had resulted in similar sample sizes in previous research within the lab. We also sought approximately 120 participants in Study 5 and we ran this study for the same length of time. However, the participant pool was particularly depleted during the relevant semester and we ended up with a sample size of 72. Post-hoc power analyses revealed that Study 5 had .75 power to detect a medium effect size.

2 We report results concerning all of the dark preference measures that were collected in the five studies. That is, preferences for darkness were measured with a single item in Studies 1–4 and with three scales (e.g. luminance preferences) in Study 5. Similarly, we report results concerning all measures of negative affectivity that were collected. Within each study, there were also assessments that were unrelated to the present hypotheses and these included measures of personality (e.g. extraversion), interpersonal behaviour (e.g. arrogance), cognitive style (e.g. the Cognitive Reflection Test), and social orientation (e.g. empathy). The latter measures were included for routine assessment purposes or for other projects.

3 There was a brief manipulation at the beginning of this study. Some participants completed an initial word evaluation task (1 = very unpleasant; 6 = very pleasant) with the lights turned off while others completed the same task with the lights at their highest setting. Under dark conditions, participants made more negative evaluations (M = 3.51) than under light conditions (M = 3.87), F(1, 70) = 6.36, p = .014, ${\eta }_p^2$ = .08, a hypothesised effect. After the word evaluation task, though, the lights were placed on a medium setting for all participants, who then completed a number of filler tasks prior to the main study. Accordingly, the initial lighting conditions (−1 = light; +1 = dark) did not affect scores for any of the key variables from the study, all |rs| < .15, all ps > .30.

4 We also asked people which colours – white (scored 1), grey (scored 2), or black (scored 3) – they would choose for various products: a car, a jacket, shoes, stereo speakers, a laptop, a t-shirt, and a refrigerator. This is not a very good measure of dark preferences because people will choose product colours for many reasons such as ability to hide dirt, scuff-resistance, etc., and these reasons will vary by product (α = .32). Stated another way, the sorts of self-related processes emphasised in Figure 1 would not necessarily be operative in these product choices. Despite these comments, there was a positive correlation between the linguistic measure of dark preferences and the product-based measure, r = .25, p = .030.