key: cord-0858164-gp520qen
authors: Somma, Antonella; Gialdi, Giulia; Krueger, Robert F.; Markon, Kristian E.; Frau, Claudia; Lovallo, Silvia; Fossati, Andrea
title: Dysfunctional personality features, non-scientifically supported causal beliefs, and emotional problems during the first month of the COVID-19 pandemic in Italy
date: 2020-05-26
journal: Pers Individ Dif
DOI: 10.1016/j.paid.2020.110139
sha: 6a7d503b3fea5d0358706527e5233c189d4b69f7
doc_id: 858164
cord_uid: gp520qen

The present study aimed at assessing the impact of demographic characteristics, maladaptive personality traits and causal beliefs about COVID-19 on perceived emotional problems in a sample of Italian community-dwelling adults (N = 1043) in the first month of the social distancing period due to the COVID-19 pandemic in Italy. Hierarchical logistic regression analysis results showed that dysfunctional personality domains and non-scientifically supported causal beliefs explained all the variance that was originally explained by demographic variables (i.e., age and gender). In particular, negative affectivity and detachment represented relevant risk factors for reduced emotional well-being in our sample. A significant positive association was observed also between emotional problems and supernatural causal beliefs on the COVID-19 infection. Our data supported the importance of considering the impact of quarantine measures on psychological well-being, while suggesting possible risk factors related to individual differences in personality and causal beliefs.

At the start of 2020, the 2019 coronavirus disease , originating from Wuhan in Hubei province, started to spread throughout China (Liu et al., 2020) and is now causing a pandemic (WHO, 2020) . There have been 4,735,622 confirmed cases of COVID-19, and 316,289 deaths (updated 19 May 2020; WHO, 2020) . In response to the COVID-19 outbreak, the Italian Government has ordered a nationwide school closure as an emergency measure to prevent spreading of the infection, strongly discouraging public activities (Tuite et al., 2020) .

In line with other countries, the Italian government decided to introduce norms to reduce the risk of spreading the infection, involving the separation and restriction of movement of all people meeting the standard for quarantine (Centers for Disease Control and Prevention, 2017) . A recent meta-analysis (Brooks et al., 2020) suggested that the psychological impact of quarantine may be wide-ranging, substantial, and long lasting, including a wide array of symptoms which may range from post-traumatic stress disorder features to anger outbursts. Notably, Brooks and colleagues (2020) stressed that data on the psychological impact of quarantine measures are sparse, suggesting the need for further studies (Greenbaum, 2020) , particularly as to the effect of participant characteristics, demographics, and personality features (Brooks et al, 2020) .

Further increasing the complexity of the emotional response to the current COVID-19 pandemic, social media platforms remain saturated with nonscientific and conspiratorial claims (Garrett, 2020) . Among them, conspiracy theories suggesting that COVID-19 does not have a natural origin have emerged (Calisher et al., 2020) , as well as other supernatural conspiracy theories (e.g., COVID-19 as a divine punishment). Unfortunately, misinformation may be related to fear and prejudice, which in turn may undermine the subject's willingness to implement the correct practices to prevent the COVID-19, thus putting at risk his/her own lives, as well as others' lives (Calisher et al., 2020) .

J o u r n a l P r e -p r o o f Against this background, we designed the present study in order to evaluate the impact of demographic characteristics, maladaptive personality traits and causal beliefs about COVID-19 on perceived emotional problems among Italian community-dwelling adults in the first month of the social distancing period due to the COVID-19 pandemic in Italy. To address this issue, we administered a measure of emotional problems to large online sample of adult participants. In particular, we administered the Strengths and Difficulties

Questionnaire Emotional Problems scale because it assesses anxiety, somatic complains, fear, and depressive features. Interestingly, the Emotional Problems Scale total score yields a wellcharacterized cut off score to identify people with clinically relevant emotional problems (Goodman, 1997; Findon et al., 2016) . Identifying people with clinically relevant distress may be useful in treatment planning. We relied on the emotional problem construct because reactions to quarantine may include a range of negative emotions rather than have a single, specific form. Moreover, participants were administered also self-report measures of DSM-5

Alternative Model of Personality Disorders (American Psychiatric Association, 2013) domains and COVID-19 related causal beliefs, respectively.

Starting from these considerations, we expected that participants with clinically relevant emotional problems may be efficiently discriminated by negative affectivity and detachment domains, as well as by non-scientifically supported causal beliefs. Indeed, previous findings showed that Negative Affectivity and Detachment were negatively associated with emotional well-being (e.g., Góngora, & Castro Solano, 2017) . Recently, Lai and colleagues' (2020) showed in a sample of 1,257 health care workers that women experienced more psychological burden than male participants. Thus, we also expected that female gender may act as a possible risk factor for clinically relevant emotional problems.

Participants responded to advertisements requesting potential volunteers for psychological research on the web (e.g., newsgroups, universities web) from March 16, 2020

to March 21, 2020. As a whole, 1,049 Italian community-dwelling adult participants volunteered to take part in the study receiving no economic incentive or academic credit for their participation. However, 6 (0.57%) participants were removed from the final sample because they completed the measures in less than 10 minutes (n=1) and because they provided questionable responses (n=5; e.g., repeating the same answer numerous times). Institutional Review Board approval was obtained. To be included in the sample, participants had to document that they were of adult age (i.e., 18 years of age or older), and to agree to online written informed consent in which the study was extensively described. All questionnaires were scored by an independent group of trainee psychologists who were blind to the aim of the study. Goodman et al., 1998) . The SDQ is a brief, 25-item self-report questionnaire which has been translated and validated in Italy (e.g., Tobia & Marzocchi, 2018) . In line with previous studies (e.g., Findon et al., 2016) , it has been adapted for use in adults with minor alteration of some items. For the purpose of the present study, we administered the Emotional Problems scale of the SDQ in order to assess common areas of emotional and behavioral difficulties. Maples et al., 2015) . The PID-5-SF is a 100-item questionnaire with a 4-point response scale, which was explicitly designed to measure the proposed DSM-5 traits and domains. It has been developed by Maples and colleagues (2015) through item response theory methods, and validated in its Italian translation (Somma et al., 2019 

Causal Belief Questionnaire (CBQ; Somma et al., 2020) . The COVID-19

CBS is a 16-item, Likert-type self-report questionnaire that was explicitly designed to assess the degree of agreement with some possible theories about the origin of the virus and its associated disease (Somma et al., 2020) . It was developed relying on independent online searches; firstly, possible origin theories for the virus were identified; then, these theories were assigned to one of the following theme: (a) supernatural beliefs; (b) conspiracy beliefs;

and (c) scientifically supported beliefs. Accordingly, 16 items listing causal beliefs for the COVID-19 were selected. Each COVID-19 CBQ item was measured on a 0 (Never) to 10 (Always) scale. COVID-19 CBQ item scores are summed to yield three scales, namely, the Supernatural Belief Scale (5 items), Conspiracy Belief Scale (9 items), and the Scientifically Supported Belief Scale (2 items). In a previous study on Italian community-dwelling adults (Somma et al., 2020) , full-information confirmatory factor analysis findings suggested the adequacy of a three-factor model for the COVID-19 CBQ items (comparative fit index = 

Hierarchical logistic regression analysis was used to evaluate the individual contribution of the independent variables in the multivariate context (i.e., taking into account their mutual overlap). Model selection was based on Akaike information criterion (AIC) minimization and the significance of the difference between omnibus likelihood ratio  2 test.

McFadden pseudo-R 2 was used as effect size measure in hierarchical logistic regression analysis. Based on our sample size, Cohen's d values of 0.20 and odds ratio values of 1.68

(both indicating small effects) could be detected with power greater than .95 even for p <.001.

In our full sample, the Emotional Problems Scale total score ranged from 0. Table 3 .

To the best of our knowledge, this study represents the first attempt at assessing the perceived emotional problems among Italian community-dwelling adults in the first month of the social distancing period due to the COVID-19 pandemic in Italy. Although we had no pre-COVID-19 assessment of our participants' psychological health, the average Emotional Problems Scale total score in our sample suggested that emotional problems were not spreading among Italian community-dwelling adult, and that the vast majority of our participants were endorsing relative emotional wellbeing. In our opinion, this finding may point to populationlevel resilience. Indeed, it should be observed that a substantial minority of our participants (13.2%) reported emotional difficulties of potential clinical relevance, at least according to the Emotional Problem Scale cut-off scores that were reported in the literature (i.e., SDQ EP7; Goodman, 1997; Findon et al., 2016) . In our opinion, this finding was somewhat consistent with the available literature (e.g., Brooks et al., 2020) on the impact of the SARS/COVID-19 epidemic on mental health, and highlights the importance of identifying the risk factors for perceiving clinically relevant emotional difficulties among subjects exposed to the COVID-19 epidemic. The cross-sectional design of our study did not allow to evaluate if the proportion of Italian community-dwelling adults experiencing clinically relevant emotional problems was likely to represent a stable phenomenon, or if it should be expected to increaseor to varywith the progression of the epidemic, as well as of the COVID-19 related social distancing interventions.

As a whole, our data seemed to suggest that several individual characteristics may represent significant risk factors for perceiving clinically relevant emotional distress among Italian community-dwelling adults, at least when they were assessed using self-reports based on a web survey. For instance, bivariate analyses showed that several demographic variables namely, female gender, being unmarried, being a university student, being on average lower than 30 years of age, and having on average less than 15 years of education -were significantly, albeit modestly effective in differentiating participants who reported clinically relevant emotional problems from participants who experienced no clinically relevant emotional problem. However, multivariate analysis results documented that only participant's (female) gender and age showed a significant unique contribution among demographic variables in differentiating participants who reported clinically relevant emotional problems from participants who were not experiencing emotional problems.

Noticeably, in our study individual differences in dysfunctional personality domains, at least as they were operationalized in the PID-5, and non-scientifically supported causal pandemic may have a substantial role in shaping preventive interventions on perceiving emotional problems, at least among Italian community-dwelling adults. In our study, the

McFadden pseudo-R 2 value for the final model was .43; beside suggesting excellent model fit (Domencich & McFadden, 1975) , this finding seemed to indicate that taking into account these predictors may produce a 43% improvement over chance of the possibility to correctly identify subjects experiencing clinically relevant emotional problems. Indeed, simulation data suggest that McFadden pseudo-R 2 values in the .40s are likely to correspond to R 2 values in the .70-.80 range (Domencich & McFadden, 1975) .

Of course, our findings should be considered in the light of several limitations. Our sample was composed of highly educated participants (years of education M = 15.37, SD = 3.37), mainly of female gender; these characteristics inherently limit the generalizability of our findings. However, the high rate of female participants that was observed in our study was consistent with extant research documenting a relative overabundance of women subjects participating in online studies (e.g., Paolacci et al., 2010) . Of course, this limitation suggests that care should be used in generalizing our findings to the Italian general population.

We had no opportunity to assess the pre-COVID-19 mental health of our participants.

Thus, we cannot rule out that the 13.2% of participants who reported having clinically relevant emotional problems had these struggles also before the COVID-19 pandemic. The cross-sectional nature of our study prevented us from making any causal inference; future longitudinal studies on this topic are needed. We relied exclusively on self-report questionnaire. Further studies based on different methods of assessment are badly need before accepting our findings; however, during Italy lockdown, web-based survey based on self-report questionnaires seemed to represent the most effective strategy to get preliminary, albeit potentially useful data on emotional problems, dysfunctional personality domains and causal beliefs among Italian community-dwelling adults.

In the attempt to facilitate subjects' participation in our on-line survey, we had to rely on short measures and to limit the number of demographic variables that were assessed in the present study. For instance, we assessed emotional problems relying on the Emotional Problem Scale of the SDQ which is a short, albeit well validated (e.g., Findon et al., 2016) , self-report measure of emotional difficulties. Although we relied on a validated empirical cutoff (e.g., Findon et al., 2016) to assign participants to the emotional problem group, this method choice was not meant to suggest that emotional problems have a latent categorical structure. Actually, the results of taxometric analyses clearly showed the dimensional nature of emotional problems. Our decision was indeed related to the fact that in applied settings (e.g., clinical assessment), it may be important to identify participants with clinically relevant problems for treatment planning. Moreover, we decided to rely on the Emotional Problem Scale because it has no item content overlap with the PID-5-SF. Of course, it should be observed that relying on different measures to assess the same constructs may lead to different findings. Further studies including an extended set of sources of information/demographic variable, and a fine-grained assessment of their use should be carried out before accepting our conclusions.

Even keeping the limitations of our study in mind, we feel that our study provided useful information on the relationships between emotional problems, and dysfunctional personality domains and COVID-19 causal beliefs, at least as they were assessed among Italian community-dwelling adults during the COVID-19 pandemic. In line with Brooks and colleagues' (2020) meta-analysis results, our findings suggested that depriving people of their liberty for the wider public good, while representing an effective life-saving measure, is often fraught and needs to be handled carefully. Note. #: One (0.1%) participant refused to disclose his/her civil status (N = 1042); §: Five (0.5%) participants refused to disclose their occupation; df: Degrees of freedom. Percentages with different superscripts were Bonferroni-significant in post-hoc multiple comparisons. 1: Phi coefficient; 2: Cramer's V coefficient; 3: Cohen's d coefficient; 4: Common language effect size. * p <.05; ** p <.01; *** p <.001. 

Diagnostic and statistical manual of mental disorders

The psychological impact of quarantine and how to reduce it: rapid review of the evidence

Statement in support of the scientists, public health professionals, and medical professionals of China combatting COVID-19

Economic costs of neuroticism: a population-based study

Urban travel demand-a behavioral analysis (No. Monograph)

Screening for co-occurring conditions in adults with autism spectrum disorder using the strengths and difficulties questionnaire: A pilot study

COVID-19: The medium is the message. The Lancet

Pathological personality traits (DSM-5), risk factors, and mental health

The strengths and difficulties questionnaire: A research note

The Strengths and Difficulties Questionnaire: A pilot study on the validity of the self-report version

Psychologist leads innovative approach to tackle psychological toll of COVID-19

Factors Associated with mental health outcomes among health care workers exposed to Coronavirus Disease

Online mental health services in China during the COVID-19 outbreak

Testing whether the DSM-5 personality disorder trait model can be measured with a reduced set of items: An item response theory investigation of the Personality Inventory for DSM-5

Psychological intervention measures during the COVID-19 pandemic

Running experiments on Amazon Mechanical Turk

Conceptualization, Methodology, Writing -Original draft preparation

Markon: Writing -Review and & Editing

Investigation, Data curation

Data curation