There have been a lot of unfortunate cases of writers, painters, who have been melancholic, depressed, taken their own lives. I don’t think it goes with the territory. I think those people would have been depressed, or alcoholic, suicidal, whatever, even if they weren’t writing. I just think it’s their characterological makeup. Whatever that characterological makeup drove them to write or to paint, as well as to alcohol or to suicide, I don’t know. I know there are an awful lot of healthy writers and painters who have no thoughts of suicide...But the burden of consciousness is great for people who don’t—you know—want to kill themselves.

—Mark Strand, quoted in Csikszentmihalyi, 1996, pp. 73–74.

Quoting poet Mark Strand in his qualitative work interviewing prominent creative individuals, Csikszentmihalyi (1996) notes that a trait of sensitivity entails a lower threshold of pain that propels certain individuals to respond to what ails them in life with creativity. Insomuch as any trait can predict psycho-social behavioral patterns of a person only to some extent, let alone the complications of developmental, environmental, and situational influences, it is safe to assume that not every artistically inclined individual is weighed down equally by this burden of consciousness identified by Strand. However, this unnamed characterological makeup may have a scientific term: sensory processing sensitivity (SPS, Aron & Aron, 1997). The role of SPS in resilience is substantiated by its scientific relevance in the literature on individual differences in creativity and stress coping. This trait has been associated with the serotonin transporter linked polymorphic region by way of differential susceptibility to environmental stimuli, indirectly connected to stress vulnerability through depression and neuroticism (Homberg et al., 2016). The same trait also has been associated with responsiveness to positive resources through openness and plasticity (Belsky & Pluess, 2009), which are two hierarchical latent factors strongly associated with creativity (Feist, 2019). Jointly, the bivalent associations of SPS may function as a double-edged sword affecting individuals’ ability to bounce back from stress and their will to thrive in challenging circumstances.

Focusing on SPS, this paper sets out to identify factors associated with resilience, with follow-up analyses to explore the moderation effects of SPS in the relationship between creative self-concept and depression among artistically inclined individuals during the COVID-19 restriction (i.e., a period when the global outbreak of COVID-19 invoked a range of preventative measures through social restriction and physical confinement). This aim is based on the assumptions that both SPS and SCS are traits associated with artistic creativity and that the COVID-19 restriction period could have induced or exacerbated the pain of subjective loneliness among those without means of coping. The COVID-19 restriction period provided an adverse environmental context for identifying risk and protective factors among artistically inclined individuals in situ.

Literature review

Individuals scoring high in SPS, phenotyped as Highly Sensitive Persons (HSP, Aron, 2013), are influenced by a genetically based temperamental trait that correlates with sensitivity to subtle cues. People demonstrating the core features of HSP observed by Elaine Aron (2013) are described as individuals with greater awareness of subtleties and deeper information processing, both of which lead to stronger emotional reactivity. Consequently, many, but not all, highly sensitive individuals often cope with this reactivity through behavioral inhibition, manifested as a secondary behavioral style (e.g., a tendency to pause and check as well as a preference for solitude) and a personality characterized by introversion.

Since Aron and Aron’s publication in 1997, scholars have proposed theories that underpin SPS (e.g., biological sensitivity to context, Boyce & Eillis, 2005; environmental sensitivity, Pluess, 2015; for a comprehensive review, see Greven et al., 2019). The common theme shared by these theories, termed the “bivalent endowment hypothesis” in this paper, is that individuals who score high in SPS demonstrate a greater variability in health-related and performance outcomes; they fare better in a supportive environment and worse in an adverse environment as compared to those with lower SPS. Thus, this bivalent endowment of protective and risk effects on individuals is contingent upon contextual influences.

Sensory processing sensitivity and artistic creativity: occupational advantage or hazard?

Highly sensitive individuals are influenced by external stimuli (e.g., other people’s moods), sensory stimuli (e.g., visual, auditory, tactile, olfactory), and internal events (e.g., thoughts and feelings). They experience more heightened stress-reactivity than those low in SPS (e.g., Greven et al., 2019). Such features have also been associated with the visual processing ability to detect subtleties (Jagiellowicz et al., 2011), emotional awareness and empathy (Acevedo et al., 2014), as well as the personality traits of openness and neuroticism (e.g., Bröhl et al., 2020). Unsurprisingly, this sensitivity is well documented in creativity personality literature (e.g., Bridges & Schendan, 2019), characterizing individuals with a penchant for the arts and creativity (Aron, 2013; Csikszentmihalyi, 1996). These artistic and creative profiles typically resemble HSP features in a greater awareness of subtle cues, ease of overstimulation, deeper information processing, and stronger emotional reactivity (Aron, 2013; Aron et al., 2012; Greven et al., 2019; Homberg et al., 2016). Because this sensitivity entails responsiveness to positive (i.e., vantage sensitivity model, Greven et al., 2019; Pluess et al., 2022) and susceptibility to negative (i.e., diathesis-stress model, Pluess, 2015) environmental influences, in a taxing environment, SPS is associated with an increased risk for stress dysregulation (e.g., Golonka & Gulla, 2021; Pérez-Chacón et al., 2021) through neuroticism (Kotov et al., 2010; Langelaan et al., 2006). However, it is also this exact trait that enables individuals to reap greater benefits from positive and supportive experiences (Ellis & Boyce, 2011; Greven et al., 2019; Pluess, 2015). Considering this susceptibility to environmental influences and given the COVID-19 socio-environmental and temporal context as a stressor, based on the bivalent endowment hypothesis, it is possible that artistically inclined individuals who score high in SPS demonstrated a higher level of health risk due to stress reactivity during the pandemic. It is also possible that, to some extent, this health risk could be attributed to neuroticism, which correlates with SPS to a substantial degree.

The role of neuroticism in sensory processing sensitivity and among artistically inclined individuals

According to the five-factor structure of personality traits (e.g., McCrae & Costa, 1997), neuroticism refers to the combination of emotional instability and negative emotionality (John & Srivastava, 1999). Individuals who score high in neuroticism have been found to be more worrisome, anxious, irritable, easily upset under pressure (Costa & McCrae, 1992), prone to boredom (Mercer-Lynn et al., 2013), and prone to developing anxiety disorders (i.e., anxiety sensitivity, Barlow et al., 2014a, b; Clark et al., 1994) if they have a tendency to experiential avoidance (Barlow et al., 2014b; Mohammadkhani et al., 2016) or of biased information processing toward ambiguous stimuli (Lommen et al., 2010). Although neuroticism overlaps with SPS to a considerable degree across studies (Greven et al., 2019), and both share a vulnerability to stress through sensitivity to negative stimuli, neuroticism by itself does not completely explain SPS (Aron & Aron, 1997) as SPS covers the additional sensitivity to positive stimuli.

Neuroticism is positively associated with the behavioral inhibition system (BIS) that captures negative emotional sensitivity (e.g., Smits & Boeck, 2006) whereas SPS is positively linked to both this BIS and the behavioral approach system (BAS), which captures sensitivity to positive stimuli (e.g., Pluess et al., 2018). Notably, the BIS described in neuroticism is coupled with the fight-or-flight system (Barlow et al., 2014b; DeYoung et al., 2021) in a situation in which one’s only motivation is to eliminate or avoid a threat whereas the BIS in HSP is thought to be motivated by the need to take more time to process information more deeply in novel or uncertain situations (e.g., Aron, 2013; Aron et al., 2010; Greven et al., 2019; Smolewska et al., 2006). This motivation is deemed self-agentic, advantageous to survival (Aron et al., 2010). In a study demonstrating that individuals who score high in SPS are less influenced by culture, Aron et al. (2010) suggest that the feature of pausing to process information more deeply before acting in HSP is an innate preference, akin to Gray and McNaughton’s (2000) conceptualization of BIS, which is not solely defined by passive avoidance but also by the crux of “active risk analysis behavior” (p. 10). In comparison, the BIS of neuroticism, driven by passive avoidance in responding to conflict or uncertainty (DeYoung et al., 2021), is construed as overly reactive and exaggerated (Barlow et al., 2014b).

The role of neuroticism among artistically inclined individuals is equivocally complicated. Although the positive association between neuroticism and artistic creativity has been documented (e.g., Götz & Götz, 1979; Haller & Courvoisier, 2010; Feist, 1998; Strong et al., 2007), an overview of literature indicates that there is no single phenotype of artistic personality. This could be supported by a sociocultural account theorizing that the identification of artistic creativity is resultant of the interactions among characteristics of individuals, domains, and fields (Csikszentmihalyi, 2015). Therefore, the identification of personality traits associated with artists is subject to variations in taste concerning the arts, bound to selection processes as a function of larger structural and systematic influences.

Notwithstanding that the sociocultural theory of artistic creativity may discount the idea of phenotyping artistic personality with neuroticism, artists are often examined through a psychopathological lens. Several studies have documented the associations between artistic creativity and neuroticism along with mood disorders. Although scholars in the psychology of creativity literature (e.g., Rothenberg, 1995; Silvia & Kaufman, 2010; Simonton, 2019) have noted limitations of the generalizability of these studies due to selection and confirmation biases,Footnote 1 from the lens of existential–humanistic psychology, neuroticism has an intricate relationship with artistic creativity through the notion of neurosis.

Neurosis is not a synonym for neuroticism as the former term has become antiquated with regard to discussions of empirically based classification systems for mental disorders (Barlow et al., 2014b); however, the two terms share the same characteristics of anxiety proneness and experiential avoidance. In examining what distinguishes a neurotic from an artist, May (1969/2007), following Otto Rank’s idea that all human beings have the urge to create and that an artiste manqué is someone with the inability to transmute internal conflicts into art, theorizes that both artists and neurotics undergo and express the implicit experiences stemming from their “sensitized consciousness” (p. 24) living in a culture full of contradictions, yet the inability to turn the experiences into meanings understandable to oneself and others prevents a neurotic from becoming an artist. Simply put, neurotics are potential artists, but not quite.

Parallel to the existential–humanistic view of neurosis, neuroticism has been viewed as a personality trait with implications for creativity. Bridges and Schedan (2019) posit that a profile of higher orienting sensitivity, lower ease of excitation, and lower sensory threshold in sensitivity temperament combined with lower neuroticism is advantageous to “resilience to novelty,” meaning “an openness and sensitivity to experience consciousness but a resilience so as not to suffer” (Bridges & Schedan, 2019, p. 187). But specifying the indicators or expressions of having the “resilience to novelty” or having “a resilience so as not to suffer” calls for another line of inquiry. In contrast, Strong et al. (2007) deem neuroticism as an affective advantage for creativity because it increases access to a broader range of emotions with unusually greater intensity and changeability. Thus, “resilience to novelty” could otherwise be interpreted as the continued courage to remain open, despite the vulnerability that comes with this sensitivity.

All things considered, what is clear is that neuroticism by itself is not a personality trait that singularly explains artistic personality. In examining creativity using the Huge Two Model of personality (DeYoung, 2006), Feist (2019) posits that creativity’s association with plasticity is stronger than its association with stability. Connecting the Huge Two Model with the Big-Five structure in order to relate the idea of “resilience to novelty” with creativity, it could be extrapolated that resilience among artistically inclined individuals is more strongly associated with openness than with neuroticism. Notably, the results of a longitudinal study on the changes in personality traits over a period of four years (Lüdtke et al., 2011) indicate that increasing openness leads to experiencing more positive and negative life events, inadvertently resulting in higher levels of neuroticism.

When it comes to corporeal reality, it is unlikely neuroticism can be teased apart from a flesh and blood person, be it the HSP phenotype or an artistically inclined individual with adverse life experiences. Additionally, neuroticism is considered functionally adaptive in late adulthood as it may increase awareness of age-related vulnerability, prompting individuals to develop behaviors in compensation for declined physical resources (Wagner & Muller, 2020). Following this line of reasoning, one auxiliary assumption of this study is that it is of little practical value to draw implications from study results that identify predictors of resilience among artistically inclined individuals independent of the influence of neuroticism. This assumption is built upon the consideration that resilience among artistically inclined individuals can be interpreted as the ability to handle the burdens of experiencing fine-tuned consciousness due to neuroticism as well as the ability to process sensitivity in-depth, tapping into the idea that a creative consciousness requires the courage to meet anxiety (May, 1953/2009) and that it demands inner strength to regulate both pain and passion (Csikszentmihalyi, 1996).

The role of creative self-concept in resilience

Creative self-concept is defined as a self-belief about one’s ability in the global domain of creativity based on one’s own affective and cognitive judgment (Beghetto & Karwowski, 2017). Informed by the dissonance theory, which posits people are motivated to behave in a way consistent with attitude (e.g., Glasman & Albarracín, 2006; Harmon-Jones et al., 2018), this study hypothesized that people who score high in creative self-concept would be more resilient during the COVID-19 restriction period. The importance of having a creative self-concept is not about the production of art; the significance is placed on the mental effort invested in the process of creating that generates self-belief in one’s ability to solve problems creatively. Thus, creative self-concept may help generate inner strength to sustain emotional and psychological resources when handling challenges, especially when the challenges appear to be novel. It is from this perspective that creative self-concept is conceptualized as a potential strength to be cultivated through psychosocial education, akin to self-competency in problem-solving with optimism.

Understanding how creative self-concept may or may not be associated with resilience among older adults can also help formulate ideas for designing psychosocial educational approaches in the area of aging well. Toward this objective, exploring psychological constructs as resources for secondary control can be useful. Because the second half of the human lifespan comes with challenges accompanying decreases in biological resources, these decreases may motivate aging individuals to rely more on secondary control (i.e., altering one’s psychology in one’s relations to the world, as opposed to primary control, which aims to directly alter the objective circumstances, Rothbaum et al., 1982) in order to address problems that cannot be solved directly. Further, because aging has been associated with decreases in openness to experience (Mickler & Staudinger, 2008; Reitz & Staudinger, 2017; Wagner & Mueller, 2020) and curiosity (e.g., Sakaki et al., 2018), framing CSC as a psychological resource for secondary control may promote adaptive aging by encouraging openness to new experience and curiosity. A study on how SPS interacts with CSC among older adults with a penchant for the arts may provide insight into understanding how the effects of CSC are complicated by other factors commonly associated with SPS. This could yield future research directions and applied implications for researchers and practitioners interested in adapting CSC to formulate approaches that promote human thriving, adaptive stress-coping, or slowing down aging-related decline. To the author’s best knowledge, research on the role of CSC in stress-coping among artistically inclined adults from middle to later life is scarce.

Overview of the study

The overarching purpose of this study was to examine the role of SPS in resilience during the COVID-19 restriction. Because the pandemic’s impact was global and massive, it was reasonable to assume that during this period, this stressor had disrupted people’s sense of normalcy for more than a year since its declaration as a global pandemic by the World Health Organization (WHO) on March 11, 2020 (Cucinotta & Vanelli, 2020). The sense of uncertainty in coping with this stressor was likely further perpetuated by the disruption of critical mental health services worldwide (WHO, 2020) and the political climate, which had impeded public health implementations. All of these socio-environmental tensions may have posed adverse risks to the general public, particularly for those high in SPS.

This study consisted of two stages, the Relevant Factor Identification Stage (Stage 1 Analyses) and the Influence of SPS Stage (Stage 2 Analyses). Stage 1 sought to identify risk and protective factors among artistically inclined individuals from middle to later life during the COVID-19 restriction period with particular attention paid to SPS. Based on the bivalent endowment hypothesis and given the specific temporal context of global exposure to the COVID stressor, the first prediction was that SPS would be negatively associated with resilience. Additionally, since individuals who scored high in creative self-concept may have invested a higher level of mental effort to solve problems, the second prediction was that creative self-concept would be positively associated with resilience.

Of note, perceived barriers to the arts were conceptualized as risk-transmission factors in Stage 1, assuming that lack of resources central to artistically inclined individuals can discourage their endeavor to thrive. This assumption was informed by the idea that individual differences and domain-specific resources need to be taken into account when studying resilience (Windle, 2011). No other specific prediction was made, but three corresponding open-ended research questions were pursued, outlined as follows:

  • Research Question 1: What were the levels of general distress, in terms of depression, anxiety, and stress in the sample of artistically inclined individuals?

  • Research Question 2: Controlling for demographic factors (i.e., age, gender, education, and social-economic status), which risk transmission factors (i.e., perceived barriers to the arts) and endogenous factors (i.e., general distress, SPS, and CSC) were associated with resilience among artistically inclined individuals?

  • Research Question 3: Were there different SPS sensitivity patterns between the relatively-high and relatively-low resilience groups?

Stage 2 analyses were triggered by Stage 1 results. Following up on the Stage 1 results, which indicated that creative self-concept was a protective factor and depression a risk factor for resilience, Stage 2 analyses sought to explore whether creative self-concept could serve as a strength-based psychosocial resource for preventing the development of, or mitigating the severity of, depression for people with varying degrees of SPS. As creativity can be broadly defined as “the interplay between imagination, improvisation, and innovation” (Karwowski, 2015, p. 165), a self-belief in creativity may encourage the willingness to imagine broader possibilities for problem solving, thus serving as a psychological resource for turning imagined possibilities into action in challenging situations, indirectly leading to lower levels of depression.

Informed by Stage 1 results, the first goal of Stage 2 was to explore whether CSC and SPS were independently associated with depression, controlling for neuroticism, using the same cross-sectional dataset analyzed in Stage 1. Additionally, because it may be informative to explore how the effectiveness of creative self-concept differs among people with varying levels of SPS, the second goal of Stage 2 was to explore whether SPS had a moderating effect on the relationship between creative self-concept and depression.

Literature indicates that SPS, influenced by a polymorphism in the serotonin-transporter gene (5-HTTLPR) linked to neuroticism through the two-way interaction between the 5-HTTLPR and negative life events (Pluess et al., 2010), is associated with an increased risk of depression (e.g., Homberg et al., 2016; Liss et al., 2005). Yet, the association between SPS and depression after controlling for neuroticism in this sample of artistically inclined individuals was unclear. Given the paucity of research aligning with the contextual and sample characteristics of the paper, the two research subquestions are presented as follows without a priori hypotheses:

  • Subquestion 1: Controlling for neuroticism, were creative self-concept and SPS associated with depression?

  • Subquestion 2: Did SPS moderate the relationship between creative self-concept and depression, controlling for neuroticism?

As previously discussed, the relationship between neuroticism and artistic creativity is intricate. Thus, for the purpose of drawing practical implications, neuroticism was not included in the regression model of the Stage 1 analyses. More importantly, the phenotype of HSP among artistically inclined individuals without the attributes of neuroticism is not a focus of the study, assuming that in real life for some individuals with a self-history of adversity, neuroticism could be a price to pay for continuing to be highly open to experience or novelty. Such justification was motivated by pragmatic consideration, apposite to those who use their sensitivity as an instrument for creative endeavors. Namely, the objective of the Stage 1 analyses was to identify factors associated with resilience among artistically inclined individuals without controlling for the effect of neuroticism. However, to satisfy scientific curiosity, neuroticism was included in the discussion of the Stage 1 results through post-hoc analyses and was also included in the Stage 2 analyses.

Method

Procedures

Data used in this study were based on an existing dataset, deidentified and anonymized. The original data collection (October 2020–January 2021) was approved by a university institutional review board. Individuals who enjoy artistic creativity with an age over 40 years old and with a working comprehension of written English were invited to participate in a research study on creativity and human resilience from middle to later life, thereby earning the eligibility to enter a drawing for a gift card worth USD$ 50.00. Individuals aged at least 40 years with varying degrees of interest and involvement in visual arts were recruited using online postings through social media, newsletters of art organizational websites, and email invitations to individuals in the arts communities.

Sample characteristics

The sample (N = 224) consisted of a majority of participants residing in the United States of America (72%) and a minority scattered among Canada, Germany, the United Kingdom, Australia, Belgium, Brazil, France, Greece, Ireland, Norway, Singapore, and Taiwan. The sample consisted of men (n = 63) and women (n = 128), with an additional four individuals collectively coded as “other” (one individual choosing “non-binary,” two individuals choosing “prefer not to say,” and one individual choosing “self-described”). The average age of the sample was 54.08 years (SD = 10.08, range = 40–84). Participants’ educational backgrounds ranged from high school or equivalency (12.9%), associate degree (8%), bachelor’s degree (27.2%), master’s degree (26.8%), to doctoral and professional degree (12.1%). With regard to participants’ disciplinary backgrounds in visual arts, 41% reported having either their primary or secondary discipline in painting/drawing, 25% craft, 22% photography, 17% computer arts/digital media, 16% textile art, 16% graphic design/interactive design, 13% sculpture, 9% illustration, 6% printmaking, 5% filmmaking, and 12% others.

Measures

Demographics questionnaire

The demographic questionnaire contained variables of gender, age, educational attainment, subjective social-economic status (SSES, measured by the MacArthur scale of subjective social status, Adler & Stewart, 2007), country of residence, primary and secondary disciplines in artistic creativity, and an index of items of perceived barriers to the arts. Each checked item was coded as 1 and each unchecked item as 0.

General distress

During the COVID-19 pandemic period, general distress was assessed using the Depression, Anxiety, and Stress Scale (DASS-21; Lovibond & Lovibond, 1995). The DASS-21 has three 7-item subscales that assess level of depression (e.g., “I felt down-hearted and blue”), anxiety (e.g., “I found myself getting agitated”), and stress (e.g., “I found it difficult to relax”) among adults. The scale requires individuals to indicate how much each item applies to themselves over the course of the previous week on a scale ranging from 0 (did not apply to me at all) to 3 (applied to me much for most of the time). For the purpose of comparing results with those of the DASS-42, the scores were summed and multiplied by two to create three scores for the subscales, each ranging from 0–42. The DASS-21 had an internal consistency of ω = 0.92 in this current sample (depression ω = 0.91, anxiety ω = 0.78, and stress ω = . 85).

Creative self-concept

Self-concept in the global domain of creativity was assessed using the Short Creative Self-Concept Scale (SSCS; Karwowski 2012). The SSCS assesses individuals’ self-conviction about their personal identity and self-efficacy in the global domain of creativity. The SSCS is an 11-item instrument that requires individuals to self-rate on a 5-point scale ranging from 1 (Definitely not) to 5 (Definitely yes). Sample items include “Ingenuity is a characteristic which is important to me” and “I am sure I can deal with problems requiring creative thinking.” The internal consistency ω was 0.89 in this sample.

Resilience

During the COVID-19 period, resilience was assessed using the Brief Resilience Scale (BRS; Smith et al., 2008). The BRS is a 6-item scale that assesses people’s ability to bounce back from health-related stress, in which respondents are asked to indicate how much they agree with each statement (e.g., “It does not take me long to recover from a stressful event”) on a scale ranging from 1 (strongly disagree) to 5 (strongly agree). The omega coefficient ω was 0.89 in this current sample.

Sensory processing sensitivity

Sensory processing sensitivity was assessed using the Highly Sensitive Person Scale (HSP-27, Aron & Aron, 1997). The HSP-27 assesses how individuals process information from the environment due to biological sensitivity to context temperament. Originally developed as a single factor construct, the scale was later found to yield three subscales: (a) ease of excitation (EOE), (b) aesthetic sensitivity (AES), and (c) low sensory threshold (LST) (Smolewska et al., 2006). The scale requires respondents to indicate how often each item applies to themselves on a Likert scale ranging from 1 (not at all) to 7 (extremely). Sample items include “Are you easily overwhelmed by strong sensory input?” “Do you seem to be aware of subtleties in your environment?” and “Do you find it unpleasant to have a lot going on at once?” The HSP-27 had an internal consistency of ω = 0.83 in this current sample (EOE ω = 0.82, AES ω = 0.76, and LST ω = 0.85).

Neuroticism

Neuroticism was assessed using the Big Five Inventory–Neuroticism (BFI-N; John & Srivastava, 1999). The BFI-N assesses the neurotic tendency of emotional instability and negative emotionality. The BFI-N is an 8-item subscale that requires individuals to indicate to what extent each statement applies to themselves on a Likert scale ranging from 1 (Disagree strongly) to 5 (Agree strongly). Sample items include “I am someone who gets nervous easily” and “I am someone who is emotionally stable, not easily upset (reverse coded).” The Omega coefficient ω was 0.86 in this current sample.

General quality of social relations

The general quality of social relations was assessed using the Positive Relations with Others Subscale of the Psychological Well-Being Scale (Ryff et al., 2007). This subscale is a 7-item instrument that requires individuals to self-rate on a 7-point scale ranging from 1 (Strongly disagree) to 6 (Strongly agree). Sample items include “I enjoy personal and mutual conversations with family members or friends,” and “I know that I can trust my friends, and they know they can trust me.” The internal consistency ω was 0.80 in this sample.

Creative personal identity

Defined as a self-belief about the importance of being a creative person, the creative personal identity was assessed using the Short Creative Personal Identity Scale (CPI; Karwowski 2012). CPI is a 6-item instrument that requires individuals to self-rate on a 5-point Likert scale ranging from 1 (Definitely not) to 5 (Definitely yes). Sample items include “I think I am a creative person” and “Being a creative person is important to me.” The internal consistency ω was 0.87 in this sample.

Creative self-efficacy

Defined as one’s self-conviction about one’s ability to solve problems requiring creative thinking in the global domain of creativity, the subjects’ creative self-efficacy was assessed using the Short Creative Self-Efficacy Scale (CSE; Karwowski 2012). CSE is a 6-item instrument that requires individuals to self-rate on a 5-point Likert scale ranging from 1 (Definitely not) to 5 (Definitely yes). Sample items include “I trust my creative abilities,” and “Many times I have proven that I can cope with difficult situations.” The internal consistency ω was 0.83 in this sample.

Analyses

Stage 1 analytic strategies

Descriptive statistics (i.e., mean, standard deviation) of the three DASS-21 (Lovibond & Lovibond, 1995) subscales were used to answer the first research question regarding the levels of general distress in terms of depression, anxiety, and stress in the sample. For the second research question, on identifying factors associated with resilience, a regression analysis was conducted. A normality probability plot of the regression standardized residuals, and a scatter plot as well as collinearity diagnostics were performed prior to testing the multiple linear regression model. To answer the third research question on SPS sensitivity patterns, a profile analysis (Tabachnick & Fidell, 2014) was used to see if relatively high and low resilience groups demonstrated different patterns among the SPS components (Smolewska et al., 2006), which include aesthetic sensitivity (i.e., a tendency to notice subtle beauty), ease of excitation (i.e., a tendency to be mentally affected by stimuli), and low sensory threshold (i.e., a tendency to be overwhelmed by unpleasant sensations).

Stage 2 analytical strategies

Regression analyses were used to answer the two research subquestions of Stage 2 analyses. The following statistical equation was devised to test the two subquestions exploring the effects of SPS and creative self-concept on depression while controlling for the effect of neuroticism,

$$Y={i}_{Y}+{b}_{1}\mathrm{X}+{b}_{2}\mathrm{W}+{b}_{3}\mathrm{XW}+\mathrm{C}+{e}_{Y}$$

where \(Y=\) depression, X = CSC, W = SPS, C = neuroticism, and XW = the interaction term of CSC by SPS; b1, b2, and b3 were the raw coefficients associated with X, W, and XW, respectively. A simple linear moderation model with the plot option (PROCESS macro, Hayes, 2018) was used to test the conditional effect of creative self-concept on depression moderated by SPS, controlling for the effect of neuroticism. The PROCESS macro for the simple moderation model by default produced conditional effects of creative self-concept on depression among individuals at the 16th, 50th, and 84th percentiles of the distribution of SPS in the study sample. The SPS scores from the individuals at the 16th, 50th, and 84th percentiles were also used for generating a visual depiction of the interaction effects.

Results

Stage 1: relevant factor identification analyses

Table 1 displays the descriptive information of and correlations between the endogenous variables. Bivariate correlational analyses indicated a moderate negative association (r =  − 0.40, p < 0.001) between SPS and resilience. SPS was positively associated with anxiety, depression, and stress. All correlations between the psychological variables were statistically significant at the level of p < 0.01, except for those of CSC with anxiety and stress.

Table 1 Mean, standard deviations, bivariate correlations, and p values of endogenous variables

Research question 1: general distress in terms of depression, anxiety, and stress

The means for anxiety and stress scores fell within the normal ranges, yet the mean score for depression was slightly above the normal range. Using the cutoff scores specified by Lovibond and Lovibond (i.e., normal range of depression = 0–9, anxiety = 0–7, and stress = 0–14, 1995), the scores of the general distress measure suggest some portions of the sample, excluding subjects with missing data (n = 19), were experiencing some degree of depression (n = 84, 41%), anxiety (n = 57, 27.8%), and stress (n = 55, 26.8%) beyond the normal ranges. These figures were lower in comparison to findings in a study using the same measure with a sample of 678 adults aged 18 years and beyond in the global population during the COVID-19 pandemic (depression = 58.6%, anxiety = 50.9%, stress = 57.4%. Shah et al., 2021). However, in comparison with another study by Alzueta et al. (2021) of a sample of 6882 individuals (Mage = 42.30, depression = 36.2%) from the global population using the same DASS-21 depression subscale, the rate of depression in this sample is higher.

Research question 2: factors associated with resilience

Collinearity diagnostics did not indicate the presence of multicollinearity; the values for the variance inflation factor ranged from 1.127 to 2.791. No violations of the assumptions of normality, linearity, multicollinearity, and homoscedasticity for the regression analysis were observed.

Table 2 displays the full results of the multivariate linear regression testing the strength of unique associations for age, gender, SSES, education, perceived barriers to the arts, general distress (depression, anxiety, and stress), and SPS on the outcome variable resilience. Statistically significant predictors at the level of 0.001 included lack of peer support of shared interests in the arts (β =  − 0.23, p < 0.001), creative self-concept (β = 0.26, p < 0.001), depression (β =  − 0.29, p < 0.001), and SPS (β =  − 0.30, p < 0.001). The two specific hypotheses predicting SPS to be a risk factor and CSC to be a protective factor were supported.

Table 2 Multivariate linear regression predicting resilience

Research question 3: patterns among the subscales of SPS

As SPS was found to be negatively associated with resilience, it was not certain if the high- and low-resilience groups produced the same pattern of means on the SPS subscales. Therefore, a profile analysis was performed on its three subscales: AES, EOE, and LST. The grouping variable was resilience, divided into relatively- high versus relatively-low resilience groups, based on the binned variable of a continuous score on BRS with 1 cut point (50th percentile above and below). No univariate and multivariate outliers were detected. Assumptions of homogeneity of variance–covariance matrices, linearity, and multicollinearity were met with no violations noted. Using Wilks’ criterion, the profiles, seen in Fig. 1, deviated significantly from parallelism, F(2, 206) = 17.62, p < 0.001, Wilks’ Lambda = 0.85, \(\eta_p^2=\;0.15\). For testing the differences in the subscales of SPS between the groups, a one-way between-groups MANOVA with Bonferroni adjusted alpha 0.017 was used. The results indicate that the differences in EOE (F(1, 207) = 37.64, p < 0.001, \(\eta_p^2=\;0.15\)) and LST (F(1, 207) = 25.79, p < 0.001, \(\eta_p^2=0.11\)) were statistically significant. The relatively low resilience group had a significantly higher mean for EOE (M = 4.86, SD = 0.94) and LST (M = 4.49, SD = 1.37) than that of the relatively high resilience group (EOE: M = 4.10, SD = 0.81; LST: M = 3.53, SD = 1.32). The mean scores for AES between the groups did not differ significantly (M = 5.70 and 5.66, SD = 0.87 and 0.84 for the relatively- low and high resilience groups, respectively).

Fig. 1
figure 1

Profiles of sensory processing sensitivity subscale scores for the relatively-high and relatively-low resilience groups

Stage 2: follow-up analyses on the influence of SPS

Excluding the interaction term of SPS and creative concept in the model, collinearity diagnostics did not indicate issues with multicollinearity. The values of variance inflation factor for the three regressors ranged from 0.638 to 1.567. As seen in Table 3, independent from the influence of neuroticism, both SPS and creative self-concept were associated with lower levels of depression. The results supported an answer of yes to the Stage 2 Subquestion 1 regarding the associations for creative self-concept and SPS with depression, controlling for neuroticism. Moreover, there was an interaction effect of creative self-concept by SPS (b3 = 2.330, p = 0.014), as shown in Table 3 and Fig. 2. The results showed that SPS attenuates the association between creative self-concept and depression; as the SPS score increases, the effect of creative self-concept on depression decreases. Further, the effects of creative self-concept on depression were only statistically significant among people with an SPS score lower than the 50th percentile of the sample. For people with an SPS score higher than the 84th percentile of the sample, the conditional effect of creative self-concept on depression was − 0.221 and statistically non-significant. Although the results supported an answer of yes to the Stage 2 Subquestion 2 regarding the moderating role of SPS, the direction of SPS’s moderating effect was unanticipated.

Table 3 Regression model estimating creative self-concept and sensory processing sensitivity on depression, controlling for neuroticism
Fig. 2
figure 2

The conditional effect of creative self-concept on depression by sensory processing sensitivity, controlling for neuroticism

Discussion

The results from the Relevant Factor Identification Stage (Stage 1 analyses) indicated that, depression, sensory processing sensitivity, and lack of peer support of shared interests in the arts emerged as risk factors, and creative self-concept was identified as a protective factor. The results from the Influence of SPS Stage (Stage 2 analyses) indicated that, controlling for neuroticism, SPS was associated with lower levels of depression and that it played a moderating role in the relationship between creative self-concept and depression; however, the direction of the moderation was unexpected. Each found factor is discussed below, and the results from the Stage 2 analyses are further considered with an eye toward plausible explanations for the unexpected findings regarding the moderation effect of SPS in the relationship between creative self-concept and depression.

The relevant factor identification

Depression

Holding all other factors constant, individuals with a higher level of depression were at the greatest risk for a reduced ability to bounce back from stress; however, the effects of anxiety and stress were not found to be associated with resilience. It should be noted that the three different types of general distress: anxiety, stress, and depression, were correlated to a moderate degree. Taken together, these findings suggest that depression could be the most proximate factor associated with a reduced ability to bounce back from stress, manifested as fatigued affective dysregulation.

Lack of peer support of shared interests in the arts

Similar to previous studies regarding the benefits of social support in resilience (e.g., Ozbay et al., 2007), lack of peer support of shared interests in the arts emerged as a risk factor, suggesting that artistically inclined individuals with a lack of social support through shared interests in the arts were at greater risk for a reduced ability to bounce back from stress. It is also possible that COVID-19 restrictions had accentuated the loneliness felt by individuals with lack of peer support of shared interests in general, leading them to lower levels of resilience in comparison with those with peer support of shared interests.

Although this finding taps into the established literature consensus on the benefits of social support on mental and physical health (Taylor, 2011), it may suggest an additional layer of quality enriched by the kindred connection among artistically inclined individuals. Described as a more expansive mode of knowing (Rosch, 2001), the experience of unique commonality shared among artistically inclined individuals could be recognized by peers with similar levels of artistic sensibility, which may have reduced the existential interpersonal isolation experienced during the COVID-19 restriction period.

Creative self-concept

Independent of the influence of SPS, creative self-concept was found to be a protective factor associated with higher levels of resilience in Stage 1 analyses. The finding can be explained by two possibilities. First, for those who were novices to a creative field or those who subscribed to the myths of creativity (e.g., viewing creativity a synonym for art or having a naïve view equating creativity with complete freedom; for details, see Benedek et al., 2021), the health theory of positive illusions (Taylor, 1989) may explain the effect of creative self-concept as such individuals may use it as an additional socioemotional resource to generate an illusion of self-control in challenging situations. Second, for those with a realistic understanding that creativity is an “extraordinary result of ordinary processes” (Sternberg & Lubart, 1996, p. 681), creative self-concept could be a bona fide indicator of creative ability that led to a higher level of resilience during the COVID-19 restriction period. The two plausible explanations may not be mutually exclusive; regardless, translated into practical implications, creative self-concept appeared to have a utility in promoting resilience without the influence of SPS.

Sensory processing sensitivity

Independent of the influences of general distress (i.e., anxiety, stress, and depression), SPS was found to have a negative association with resilience. Further, the subsequent profile analysis results suggest that people who scored higher in comparison to those who scored lower in resilience have different profiles in SPS. As a group, the relatively-low resilient individuals experienced more unpleasant sensory arousal and negative reactions to stimuli, meaning that they were more vulnerable to the input of negative sensory information. Literature elsewhere indicates that EOE and LST together are often found to be associated with anxiety and depression, and LST alone is an indicator of sensory discomfort (Greven et al., 2019). These findings give rise to the question: to what extent are AES, EOE, and LST intertwined? As SPS is well-known for its association with neuroticism, it is possible that the findings merely reflect the influence of neuroticism. Thus, two post-hoc tests controlling for neuroticism were performed to examine 1) the partial correlations among the three SPS components, and 2) if different patterns, indicated by the subscales found in SPS between the relatively-high and relatively-low resilience groups would remain the same after controlling for neuroticism. The results showed that the three components of SPS are correlated (raes-eoe = 0.27. p < 0.001; reoe-lst = 0.55, p < 0.001; raes-lst = 0.43, p < 0.001), with the effect of neuroticism being partialed out. After controlling for neuroticism, the profiles between the relatively- high and low resilience groups remained deviated from parallelism.Footnote 2 Taken together, the results indicate that neuroticism did not completely explain the intercorrelations among AES, EOE, and LST in this sample of artistically inclined individuals.Footnote 3

The influence of SPS

The results of Stage 2 analyses indicated that SPS was a protective factor associated with a lower level of depression, independent of the influence of neuroticism. This finding contradicted the Stage 1 results, which identified SPS as a risk factor associated with a lower level of resilience. The fact that SPS played as a risk factor in Stage 1 and a protective factor in Stage 2 may be explained by neuroticism’s overlap with sensitivity as a negative facet of SPS. When the effect of neuroticism was controlled, SPS was seen in a more positive light and associated with a lower level of depression.

The finding that individuals scoring higher in SPS benefited less from having a high level of creative self-concept was unexpected. The beneficial effect of creative self-concept was found only for those with relatively-low SPS scores, controlling for neuroticism. Why was it that creative self-concept appeared to be less beneficial against depression for people scoring high in SPS, independent of the influence of neuroticism? First, it is possible that participants scoring high in SPS in this sample were less likely to endorse a positive illusion; in this case, creative self-concept. The very idea of defining oneself as a creative person can work as a double bind, working against the trait of openness commonly observed among highly sensitive people (Bröhl et al., 2020; Lionetti et al., 2019) and among creative individuals (Feist, 2019). Additionally, a psychologically sound, artistically inclined adult who has experienced the ironies and burdens of creativity might hesitate to prioritize self-importance as a creative person, considering the following chain of reasoning: 1) Within the marketing culture that commercializes individual uniqueness, being a creative person can mean many different things to a point of meaning nothing; and 2) People scoring high in SPS are less likely to be influenced by culture due to their deliberative thinking style (Aron et al., 2010). Furthermore, serious endeavors in creativity can be psychologically costly as they bear the burden of a fine-tuned consciousness (Csikszentmihalyi, 1996) and in the eminent case of Big-C creativity, social responsibility (Lebuda & Csikszentmihalyi, 2020). Taken together, this self-concept may appear undesirable for a highly sensitive adult with a deliberative mindset (Stenmark & Redfearn, 2022) and few life illusions.

Alternatively, the analysis may have overlooked the complexity of creative self-concept as this self-concept contains two dimensions (Karwowski, 2012): creative personal identity (CPI) and creative self-efficacy (CSE). Ideally, a stable sense of CPI should be a byproduct of CSE over time, assuming the relationship between CSE and CPI is reciprocal (Karwowski, 2016). Because CPI has a focus on an ostensive self-image for being creative (e.g., “I think I am a creative person”), a creative self-concept disproportionally based on a self-image without the bolstering of creative self-efficacy may suggest a lack of sense of personal mastery in the global domain of creativity. From this perspective, a creative self-concept chiefly derived from CSE measured at a fixed-point in time would be more indicative of having an authentic creative self-concept than the one predominantly derived from CPI measured at a fixed-point in time. A combination of a high CPI score and low CSE score could suggest an unstable or inflated sense of creative self-concept. Additionally, people who score high in SPS could be more vulnerable to having a discrepancy between CSE and CPI due to their sensitivity and deliberative mindset. To test this plausible explanation, an exploratory post-hoc analysisFootnote 4 was performed comparing the moderation effects of CPI versus CSE in the relationship between SPS and depression. The results partially supported the plausible explanation (seen in Table 4); independent of the influences of CSE and neuroticism, a higher level of CPI was associated with a higher level of depression among those who scored higher in SPS (the interaction effect of CPI × SPS: b = 4.0719, p = 0.002). Independent of the influence of CPI, there was not enough evidence to indicate an interaction effect between CSE and SPS on depression (b =  − 1.0845; p = 0.344).

Table 4 Regression model estimating the interaction effects of creative personal identity and creative self-efficacy by sensory processing sensitivity on depression, controlling for neuroticism

Final discussion

This study augments existing literature by providing empirical evidence for supporting hypotheses built on the bivalent endowment of SPS and scholarly discussions on how the mechanisms between creative self-concept and depression vary as a function of SPS, complicated by the consideration of whether neuroticism is or is not integral to SPS.

Several questions arise from the study. First, although Stage 1 analyses mathematically demonstrated that the relatively-high and relatively-low resilience groups shared the same level of aesthetic sensitivity and differed in ease of excitation and low sensory threshold, it is uncertain what conditions could result in a personality profile high in aesthetic sensitivity and low in the other two SPS components. As discussed earlier, it appears that the three components are intercorrelated in this sample, even when controlling for neuroticism. Hence the following questions are suggested as possible directions for future research: is the tendency to be moved by subtle beauty inseparable from the tendency to be mentally affected by stimuli and overwhelmed by unpleasant sensations among artistically inclined individuals? Is the inter-correlational pattern found in the study an idiosyncrasy that uniquely characterizes artistically inclined individuals from middle to later life? Although mathematically any factor can be controlled for or partialed out, any statistical analysis alone cannot address the real-life questions: How common are such people, high in SPS and low in neuroticism, in the real world? Under what conditions can such a personality profile be fostered?

While this study was not designed to answer these questions, it is noteworthy that its results may carry a higher external validity because of the sample’s real-world temporal context. The findings may provide unique insights more applicable to artistically inclined adults from middle to later life, in comparison to studies conducted among relatively young college students in confined experimental labs.Footnote 5

Limitations and future directions

The study findings should be interpreted in the context of caveats and a cautionary note that no two individuals are completely alike or utterly different, artistically inclined or not. First, the cross-sectional design of the study limits the ability to draw causal conclusions. For instance, it cannot exclude the possibility that depression could lead to a lower self-reported creative self-concept. Future research may employ longitudinal designs to test this possibility. Additionally, a fixed-point cross-sectional dataset is limited when it comes to demonstrating affective dysregulation. Specifically, although stress and anxiety were not found to play a statistically significant role as risk factors for lower levels of resilience in this cross-sectional study, it should be noted that anxiety and stress can induce depression (e.g., Weger & Sandi, 2018), and the delayed effects may not be captured at a fixed point in time. Second, as the analyses were based on a convenient sample involving self-selection biases, generalizability is limited. The fact that all of the study participants were recruited using digital media (i.e., email invitations and online postings) implies that all participants in this sample had some sort of access to the Internet and the sample is neither globally nor culturally representative. However, that statistical significance was found indicates that findings related to the identified risk factors are worth examining in future studies with assessments of different social groups at multiple time points. Third, while depression was identified as a risk factor for resilience, the author acknowledges that there is heterogeneity in depression complicated by issues pertaining to the measurement of depression (Fried et al., 2022). Further, depression can be viewed as a manifestation, not merely as an isolated clinical diagnosis, of broader issues of developmental influences and socio-environmental disparities (Slavich, 2020). Fourth, most of the study participants might have wanted to believe that they were creative as the majority of the study participants had a background in visual arts. This may have conflated the idea of art with creativity. However, although the field of visual arts might value creativity in some way, art and creativity are not the same thing. Future research may want to include measures that can clarify this potential conflation as well as constructs related to self-impression management, inflated self-esteem, and adherence to cultural orientations in the survey. In sum, many other important factors associated with resilience among individuals from middle to later life were not included in this study.

Finally, special attention is paid to the use of the self-report measure of sensory processing sensitivity in this paper. Because measuring sensitivity may also involve capturing emotional reactivity to stimuli, it is possible that the act of simply responding to the HSP-27 items could have elicited emotional reactivity intrinsic to respondents who scored high in sensory processing sensitivity and/or neuroticism (Aron et al., 2012). Future research could employ advanced techniques to extract psychophysiological reactions using biological measures (e.g., fMRI scanning of brain activations, Aron et al., 2010, Jagiellowicz et al., 2011; Schaefer et al., 2022; physiological stress responses, Weyn et al., 2022) and behavioral observations (Lionetti, 2020).

Conclusion

Based on the bivalent endowment hypothesis of SPS, the primary objective of this study was to examine the role of SPS and identify factors associated with resilience among artistically inclined individuals from middle to later life during the COVID-19 restriction. The paper uniquely homes in on factors that carry theoretical and practical implications for researchers of sensory processing sensitivity. The vulnerability to lower levels of resilience seems to be greatest among individuals with higher levels of depression, ease of excitation, low sensory threshold, and lack of peer support of shared interests in the arts. This last finding may have broader implications for the significance of having peer support of shared interests in general, not limited to interests in the arts.

The findings of post-hoc analyses provide future research directions to investigate the relationships among aesthetic sensitivity, ease of excitation, low sensory threshold, and neuroticism in different samples and contexts. Additionally, sensory processing sensitivity, without the influence of neuroticism, was found to be a moderator in the relationship between creative self-concept and depression. Interpretations of the study results should be placed into the temporal context of the data collection, yet the implications of the study may not be uniquely limited within the scope of the COVID-19 restriction. Specifically, this paper may provide insights into understanding resilience in the broader context of stress coping. The moderation effects of SPS found in the study may also provide directions for future research and applied work focusing on highly sensitive and artistically inclined individuals from middle to later life.