Brief Symptoms Inventory psychometric properties supports the

hypothesis of a general psychopathological factor.

Authors

Leandro Fernandes Malloy-Diniz

Mental Health Department - Medical School. Universidade Federal de Minas Gerais.

Universidade FUMEC

Alexandre Luiz de Oliveira Serpa

Social and Cognitive Neuroscience Lab, Mackenzie Presbyterian University

Meta Cognitiv

Danielle de Souza Costa

Neuroscience Lab, Faculdade de Medicina da Universidade Federal de Minas Gerais

Mayra Isabel C. Pinheiro

Ministério da Saúde. SEGETS.

Alexandre Paim Diaz

Department of Psychiatry and Behavioral Sciences, The University of Texas Health

Science Center at Houston, Houston, TX, USA.

Jonas Jardim de Paula

Neuroscience Lab, Faculdade de Medicina da Universidade Federal de Minas Gerais

Department of Psychology, Faculdade de Ciências Médicas de Minas Gerais

Debora Marques de Miranda

Pediatrics Department - Neuroscience Lab, Faculdade de Medicina da Universidade

Federal de Minas Gerais

Antônio Geraldo Silva

Brazilian Psychiatry Association,

Porto University

Abstract

The existence of a general factor related to psychiatric symptoms is supported by studies

using a variety of methods and in both clinical and nonclinical samples. This study aims to

evaluate the psychometric properties of Brief Symptom Inventory assessing the hypothesis

that its scores fits better in a bifactor model with both general factor and specific cluster of

symptoms. Participants were 6,427 brazilian subjects (81% female). The mean age was 42.1

years old (SD = 13.6, Min = 13, Max = 80). All participants fulfill the online version of the

Brief Symptom Inventory. This scale presents a general score (GSI) and nine specific cluster

of symptoms (depression, anxiety, phobic anxiety, interpersonal sensibility, psychoticism,

paranoid ideation, obsessive-compulsive, hostility and somatization symptoms)

Confirmatory factorial analysis was performed to assess the factorial BSI structure. Results

supports the bifactorial solution for BSI. Nonetheless, the general factor was the main

dimension to explain the variability of examinees. The use of the nine specific clusters of

symptoms remains useful to assess specific clinical and research questions.

Introduction

The frontiers between psychiatric illnesses are much less established than those conceived in

the diagnostic manuals of mental disorders. In fact, the lack of precise boundaries among

mental illnesses has led the psychiatric diagnosis process to a gradual change from a

categorical perspective to a dimensional one1. Current categories of mental disorders are

highly comorbid among them, and such phenotypic covariance should not be neglected in

the clinical practice. Some authors even argue that the symptomatic similarity between

patients with specific disorders suggests a shared common core between mental disorders2-3.

Recently, Caspi and Moffitt4 pointed out that several psychopathologies often have the same

biomarkers and risk factors, and that therapeutic strategies seem to work for a broad range of

disorders. Hereupon, the assessment of psychopathological symptoms, in spite of their

nosological classification, has significant clinical applicability, so much for identifying

therapeutic targets as to assess interventions’ clinical relevance. Carragher and colleagues5

even reason that symptom-level analyses allow us to “unpack disorders”, uncovering the

empirically [nonarbitrary] based structure of psychopathology. In their study, a modified

bifactor model, with three correlated specific factors (internalizing, externalizing, thought

disorder) and one general psychopathology factor, the “p” factor, was found. The p factor

has since been corroborated by studies in childhood and adolescence 6 and with adults2, a

pattern likely to be stable throughout time7-8. In effect, if the p factor is not a spurious finding,

it will consistently appear in the psychometric modeling of instruments measuring different

types of psychopathological dimensions. Therefore, as in studies concerning constructs like

the general intelligence factor, also called “g”, the psychometric analyses of instruments
evaluating psychiatric disorders or psychological distress are a useful way to assess the

hypothesis of the “p” factor.

The Brief Symptoms Inventory (BSI)9 is an instrument of self-reported psychological distress

and psychopathological symptoms including nine dimensions: depression, anxiety,

somatization, obsession-compulsion, interpersonal sensitivity, phobic anxiety, hostility,

paranoid ideation, and psychoticism. The inventory also has a Global Severity Index (GSI),

which includes all symptoms assessed by the scale. The scale was developed before the third

edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-III)10 and

remains in use for both clinical and research purposes with psychiatric patients and in non-

clinical samples11-12. Throughout the decades of its existence the BSI was translated to

several languages13-20.

The nine factors structure of the BSI has been replicated through confirmatory and

exploratory factor analysis in samples from several countries like Italy 21 and Azerbaijan22,

but most studies fails to support this model. Applying the factor analysis framework, usually

a strong unidimensional factor emerges23-24. Evidence of unidimensionality of the BSI may

be expected also due to the GSI score, but most important because psychometric studies have

not been designed to forecast a general factor and its possible relations to other specific

dimensions25.

Thomas26 evaluated the BSI demonstrating that the bifactor model fits better the data in

comparison with unidimensional and multidimensional data and also shown additional

evidence of criterion-related validity by predicting more accurately DSM-IV-TR depressive,

generalized anxiety, phobic and somatization disorders than those models. This result brings

attention to the clinical relevance of the bifactorial structure to explain the variability in the
respondents' responses. The bifactorial structure was replicated by Urbán et al25, in a

hungarian sample, and Urbán et al.,27, in hungarian and dutch samples, demonstrating better

fit than other multidimensional and hierarchical structures. Bifactorial structure was

replicated for both men and women and the global scores were understood as a global distress

factor and the general factor seems adequate to represent a global psychological distress

index. Nonetheless, the bifactor solution was not always found in other populations, like, for

example in a Greek nonclinical sample24. In this case, the authors argue that the BSI-

properties does not justify the use of specific factors to assess psychopathology. But even in

that situation, the authors suggest that the BIS global score can be used as a proxy of

psychological distress being in nonclinical samples. Should the psychological distress index

be a proxy or a measure of the “p” factor? Is the bifactor structure specific to a sample and/or

country and not replicable abroad?

The present study aims to analyze BSI scores in a large Brazilian nonclinical sample to assess

if: a) a bifactor structure could be observed in Brazilian sample; b) several

psychopathological symptoms could be explained by a bifactor model with a broad general

psychopathology factor and narrower specific clusters of not correlated symptoms or b)

multiple correlated factors without a general factor of psychopathology. The assessment of

these alternative possibilities is important since a bifactorial structure of symptoms

presentation, with a general score, independent from specific clusters of symptoms, supports

the idea of hierarchical psychopathological organization.

Method

Participants
We included 6,427 (81% female) Brazilian subjects from all states of the country. The mean

age was 42.1 years old (SD = 13.6, Min = 13, Max = 90). In regard of educational attainment,

68% were classified from 6 to 8 on the International Standard Classification of Education -

2011 (UNESCO, 2012), 51% are married or have a stable relationship and 2,468 (85.7%

female) participants self-reported previous lifetime psychiatric diagnosis for at least one

condition.

Procedure

Participants were recruited from the internet by a social media campaign in May-June 2020.

All the questionnaires were delivered by the SurveyMonkey platform. The participants

agreed with the informed consent before starting to answer the tests and questionnaires. The

procedures have been approved by the National Commission on Ethics in Research under the

process number CAAE: 30823620.6.0000.5149 in accordance with the Declaration of

Helsinki.

Measure

The Brief Symptom Inventory (BSI) is a 53-item instrument designed to identify relevant

psychological symptoms9. The inventory covers nine symptom dimensions (Somatization,

Obsession-Compulsion, Interpersonal Sensitivity, Depression, Anxiety, Hostility, Phobic

Anxiety, Paranoid Ideation, and Psychoticism) plus three global indices of distress (Global

Severity Index – GSI, Positive Symptom Distress Index – PSDI, and Positive Symptom Total

– PST). The items were answered on a 5-points Likert scale from 0 (not at all) to 4

(extremely), could be self or interviewer administered and has norms for adolescents until 13
years old and adults for both clinical and non-clinical groups. The reliability reported on the

original manual ranged from .71 on Psychoticism to .85 on Depression dimension. Test-retest

reliability was provided by global indices, ranging from .87 (PSDI) to .90 (GSI) and for all

dimensions that ranges from .68 (Somatization) to .91 (Phobic Anxiety).

Statistical procedures

Structural models for the BSI should take in consideration its multidimensional nature. From

the global indices, only the GSI was taken on the models. It happens due its calculation,

which takes in account all the nine dimensions and the four items which do not load on any

of the dimensions (items 11, 25, 39 and 52). For that reason, the GSI could be interpreted as

a latent dimension of the BSI. The other two indices demand calculations to be produced that

do not make them able to be considered a latent factor of the instrument. In addition, all the

nine dimensions are also included in the models.

<<Table 1 around here>>

Three competitive structural models were built due to possible nature of GSI dimension, the

first and the second considering the GSI as a high-order factor that emerges from the nine

symptom dimensions and the third model a bifactor structure. The difference between the

first and second factors was the inclusion or not of the four items of GSI on the structure.

A confirmatory factor analysis was conducted on lavaan package 28 in R software29 using

weighted least squares mean and variance adjusted estimation with Satorra-Bentler

correction, to correct the standard errors and qui-squared estimates30. The full sample was
randomly split by two subsamples and each model was verified for each of them. After both

models were compared in regard of its fit indices to verify the stability of estimates. The

global model fit was evaluated using the comparative fit index (CFI), Tucker-Lewis index

(TLI), the root mean square error of approximation (RMSEA) and the standardized root mean

square residual (SRMR). To interpret model fit, values higher than .95 for CFI and TLI, and

lesser than .05 for RMSEA and .08 for SRMR are indications of good fit for the models31.

The quality of the models has been verified by several indices. The H index was developed

to evaluate the construct replicability, measuring in which degree the latent variables are

appropriately represented by the indicators. The threshold of .70 is generally accepted as a

criterion for that index32. Coefficients omega (ω) and omega hierarchical (ωH) have been

shown. The omega hierarchical is useful for bifactor models for assessing the percent of the

common variance attributable to the general factor and Reise, Bonifay and Haviland 33 argues

that higher the omega hierarchical, higher is the relevance of the general factor to explain the

variance of the data. In that case, the general factor could reflect an essentially

unidimensional structure that explains the variance in the examines’ scores.

To help the evaluation of the essential unidimensionality of the factors, the explained

common variances for general (ECV), specific (ECV_SG and ECV_GS) and item levels (I-

ECV) were calculated. The ECV index evaluates the proportion of the common variance

explained by the general factor, whereas the ECV_SG and ECV_GS explains the common

variances which are due to the specific factors and the variance in each factor due to the

general factor, respectively34. In the item level, it represents the proportion of the items

variance that could be explained by the general factors32. The percent of uncontaminated

correlations (PUC) specified the possible bias in data of interpreting multidimensional into
unidimensional data. The semPlot (Epskamp, 2019) and BifactorIndicesCalculator (Dueber,

2020) in R and the Jamovi software (The jamovi project, 2020) were also used in the analysis.

Results

The confirmatory factor model does not converge for models one and two and a solution was

not found for both subsamples and for the whole sample. For the bifactor structure a solution

has been found and the model fits properly for both subsamples (CFIsample1 = .991, TLIsample1

= .990, RMSEAsample1 = .058 [.057-.059], SRMRsample1 = .051; CFIsample2 = .991, TLIsample2 =

.991, RMSEAsample2 = .058 [.057-.059], SRMRsample2 = .050). Due the model fit are equal, a

final model was estimated for the whole sample and the indices indicates a good fit for the

bifactor model (CFI = .991, TLI = .990, RMSEA = .057 [.057 - .058], SRMR = .049).

<<Figure 1 around here>>

The general factor emerges as the main dimension to explain the variability of examinees'

answers. The H index indicates that the best construct replicability was achieved by the

general factor whereas the other dimensions have low estimates and suggested that they are

not adequately defined, except the Phobic Anxiety dimension (Table 2). For GSI, the omega

is .98 and omegaH is .95 which suggests that around 97% of the reliable variance is due

general factor, 3% are due the specific factors and 2% squarely to random error 32. For the

specific factors, the omega range between .83 and .93 and are higher compared to their omega

hierarchical values. Those results suggested an essentially unidimensional structure due a

strong general factor that explains most of the reliable variance and is less affected by the

multidimensionality induced by specific factors.

<<Table 2 around here>>

The ECV of GSI explains around 77% of variance and in conjunction with the PUC of .918

indicates that the common variance might be interpreted as essentially unidimensional. Yet,

the comparison of ECV_SG and ECV_GS implies that most of the explained variance on the

specific factors are due the general factor and not to the items’ composition of the dimensions

itself. In addition, most of the BSI items showed high communality in virtue of the GSI

dimension, suggesting they are practical markers of measures of the general factor. Thus,

evidence suggests that one general factor is sufficient to explain the score variability of the

BSI.

Discussion

Our results strongly support the idea of bifactorial structure in the assessment of psychiatric

symptoms by BSI. They are in accordance with those reported by Thomas26, Urbán et al25,

Urbant et al.,27 reinforcing the bifactorial nature of BSI regardless of cultural influences and

mental health condition. As found in these studies, the present study found that the nine BSI’s

original factors present adequate psychometric properties supporting bifactor structure for

BSI. We did not find evidence supporting the hypothesis from Loutsiou-Ladd et al.,24 which
suggests that BSI is unidimensional at least in nonclinical samples since the general

population presents those symptoms assessed by BSI in a less pronounced way.

Nonetheless, our results suggest that the general symptom index presents strongest

psychometric properties in comparison to specific factors. The idea of a “p” factor is

supported by previous psychometric studies which argue that a bifactorial structure of

symptoms explains most of the variability of presentation of psychopathological traits in the

population. For example, Gluschkoff et al.,8 analyzed the results of interviews based on DSM

criteria in a large nonclinical sample. They found that the clusters of symptoms related to

specific diagnostics are explained by the bifactorial structure with specific symptom clusters

of mania, generalized anxiety disorder, major depressive disorder, dysthymia, posttraumatic

stress disorder, agoraphobia, panic disorder, social phobia, specific phobia, antisocial

personality disorder, distress, externalizing disorder and internalizing disorder. Besides these

specific factors, a general factor was also founded. It is interesting to note that this bifactor

structure remains relatively stable throughout the longitudinal follow up of the sample

suggesting that despite changes in symptoms presentation, a general “p” factor continues to

influence clinical presentation in psychopathology. The “p” factor is also supported by its

ecological relationship. Recently Pettersson et al. 35 assessed a large population sample and

also found that the general psychopathological factor (assessed by general scores derived

from self and other-report psychiatric scales) was associated to some adverse outcomes the

use of both prescribed and illegal drugs, criminality, and both low income and educational

level.

Our results present a bifactor structure for BSI and even both general and the nine specific

scores presented appropriated in terms of psychometric properties. The GSI emerges as the
main factor for the screening of mental health in general population, while the specific scores

have none or low discriminative power. Nonetheless, the use of specific factors seems to be

useful to address specific questions in both clinical and research. For example, in recent
36
studies concerning mental health related to the pandemics, Wang et al., and Ellis et al., 37

used hostility and depression subscales to address specific questions related to psychological

distress in cancer patients and adolescents in isolation, respectively. Thus, our results support

a similar score interpretation approach to address specific issues in mental health for the

Brazilian BSI version.

Our study presents some limitations which have been addressed by future studies. First of

all, the gender imbalance in our sample can represent a bias. Future studies are needed to

assess the gender invariance in BSI scores. The second limitation is that our sample was

nonclinical and even 38% has reported suffering from a psychiatric illness, we do not have

an in-depth assessment to verify this information. Nonetheless, our data is similar with those

reported by Viana et al.,38 in an epidemiological Brazilian study, and, therefore, our sample

can be considered similar with those previously studied in Brazil.

As pointed by Loutsiou-Ladd et al.,24 symptoms expression in clinical samples can be

different from nonclinical and therefore, can affect the factorial structure of symptom

presentation. The third limitation is related to the period of data collection (from May to June

2020). Since COVID-19 pandemic can impact population mental health increasing distress

symptoms in the population39 the non-clinical characteristic of our sample is, in a certain

way, questionable. Therefore, future studies should assess if this factorial structure will

remain in a similar community sample in a post pandemic scenario.

Despite its limitation, our study reinforces the bifactorial structure of psychiatric symptoms,

as assessed by BSI corroborating the accumulated evidence which suggests the existence of

a general psychiatric factor independent from specific clusters of psychiatric symptoms.

References

1. Haslam, N., Holland, E., & Kuppens, P. (2012). Categories versus dimensions in
personality and psychopathology: a quantitative review of taxometric research.
Psychological medicine, 42(5), 903.
2. Caspi A, Houts RM, Belsky DW, Goldman-Mellor SJ, Harrington H, Israel S, et al.
The p factor. Clin Psychol Sci (2014) 2(2):119–37.
3. Haltigan JD. Editorial: Putting Practicality Into "p": Leveraging General Factor
Models of Psychopathology in Clinical Intervention. J Am Acad Child Adolesc
Psychiatry. 2019;58(8):751-753. doi:10.1016/j.jaac.2019.03.005
4. Caspi A, Moffitt TE. All for One and One for All: Mental Disorders in One
Dimension. Am J Psychiatry. 2018;175(9):831-844.
doi:10.1176/appi.ajp.2018.17121383.
5. Carragher, N., Teesson, M., Sunderland, M., Newton, N. C., Krueger, R. F.,
Conrod, P. J., … Slade, T. (2015). The structure of adolescent psychopathology: a
symptom-level analysis. Psychological Medicine, 46(05), 981–994.
6. Murray AL, Eisner M, Ribeaud D. The Development of the General Factor of
Psychopathology 'p Factor' Through Childhood and Adolescence. J Abnorm Child
Psychol. 2016;44(8):1573-1586. doi:10.1007/s10802-016-0132-1
7. Polek E, Neufeld SAS, Wilkinson P, Goodyer I, St Clair M, Prabhu G, Dolan R,
Bullmore ET, Fonagy P, Stochl J, Jones PB. BMJ Open. 2020 May 11;10(5):e032494.
doi: 10.1136/bmjopen-2019-032494. PMID: 32398331
8. Gluschkoff K, Jokela M, Rosenström T. The General Psychopathology Factor:
Structural Stability and Generalizability to Within-Individual Changes. Front
Psychiatry. 2019;10:594. Published 2019 Aug 30. doi:10.3389/fpsyt.2019.00594
9. Derogatis, L. R. (1993). BSI bibliography. Minneapolis: NCS Pearson, Inc.
10. American Psychiatric Association. (1987). Diagnostic and Statistic Manual of
Mental Disorders (DSM-Ill). Washington: American Psychiatric Association.
Revised version.
11. Gavazzi, S. M., Julian, T. W., & McKenry, P. C. (1996). Utilization of the Brief
Symptom Inventory to Discriminate between Violent and Nonviolent Male
Relationship Partners. Psychological Reports, 79(3), 1047–1056.
12. Yasak, Y., & Esiyok, B. (2009). Anger amongst Turkish drivers: Driving Anger
Scale and its adapted, long and short version. Safety Science, 47(1), 138–144.
doi:10.1016/j.ssci.2008.02.003
13. Francis, V. M., Rajan, P., & Turner, N. (1990). British community norms for the
Brief Symptom Inventory. British Journal of Clinical Psychology, 29(1), 115–116.
doi:10.1111/j.2044-8260.1990.tb00857.x
14. Leo, D., Frisoni, G. B., Rozzini, R., & Trabucchi, M. (1993). Italian community
norms for the Brief Symptom Inventory in the elderly. British Journal of Clinical
Psychology, 32(2), 209–213.
15. Gilbar, O., Ben-Zur, H. Adult Israeli Community Norms for the Brief Symptom
Inventory (BSI). International Journal of Stress Management 9, 1–10 (2002).
16. Sahin, N.H., Durak, A., 1994. Kısa Semptom Envanteri: Türk gencleri
icinuyarlanmasi (Adaptation of the brief symptoms inventory for Turkish
adolescents). Turk Psikoloji Dergisi (Turkish Journal of Psychology) 31, 44–56.
17. Urbán, R., Kun, B., Farkas, J., Paksi, B., Kökönyei, G., Unoka, Z., Felvinczi, K., ...
Demetrovics, Z. (January 01, 2014). Bifactor structural model of symptom
checklists: SCL-90-R and Brief Symptom Inventory (BSI) in a non-clinical
community sample. Psychiatry Research, 216, 1, 146-54.
18. Jahn DR, DeVylder JE, Drapalski AL, Medoff D, Dixon LB. Personal Recovery as
a Protective Factor Against Suicide Ideation in Individuals With Schizophrenia. J
Nerv Ment Dis. 2016;204(11):827-831. doi:10.1097/NMD.0000000000000521
19. Kwee, C. M. B., & van den Hout, M. A. (2018). Anxiety Sensitivity Does not
Predict Treatment Outcome or Treatment Length in Obsessive-Compulsive
Disorder and Related Anxiety Disorders. Journal of Obsessive-Compulsive and
Related Disorders.
20. Boer, S., Dekkers, O. M., Cessie, S. le, Carlier, I. V., & van Hemert, A. M. (2019).
Prediction of prolonged treatment course for depressive and anxiety disorders in an
outpatient setting: The Leiden routine outcome monitoring study. Journal of
Affective Disorders, 247, 81–87. doi:10.1016/j.jad.2018.12.035
21. Adawi, M., Zerbetto, R., Re, T.S., Bisharat, B., Mahamid, M., Amital, H., Del
Puente, G., Bragazzi, N.L., 2019. Psychometric properties of the brief symptom
inventory in nomophobic subjects: Insights from preliminary confirmatory factor,
exploratory factor, and clustering analyses in a sample of healthy italian volunteers.
Psychol. Res. Behav. Manag. 12, 145–154.
22. Kerimova, M., Osmanli, N., 2016. The Brief Symptom Inventory: A validity-
reliability Study of a Sample from Azerbaijan. J. Educ. Train. Stud. 4, 153–159.
23. Elliott, R., Fox, C.M., Beltyukova, S.A., Stone, G.E., Gunderson, J., Zhang, X.,
2006.
Deconstructing therapy outcome measurement with Rasch analysis of a
measure of general clinical distress: the Symptom Checklist-90-Revised.
Psychological Assessment 18, 359–372.
24. Loutsiou-Ladd, A., Panayiotou, G., & Kokkinos, C. M. (2008). A review of the
factorial structure of the Brief Symptom Inventory (BSI): Greek evidence.
International Journal of Testing, 8, 90–110. doi:10.1080/15305050701808680.
25. Urbán, R., Kun, B., Farkas, J., Paksi, B., Kökönyei, G., Unoka, Z., ... &
Demetrovics, Z. (2014). Bifactor structural model of symptom checklists: SCL-90-
R and Brief Symptom Inventory (BSI) in a non-clinical community
sample. Psychiatry research, 216(1), 146-154.
26. Thomas, M. L. (March 01, 2012). Rewards of bridging the divide between
measurement and clinical theory: Demonstration of a bifactor model for the Brief
Symptom Inventory. Psychological Assessment, 24, 1, 101-113.
27. Urbán, R., Arrindell, W. A., Demetrovics, Z., Unoka, Z., & Timman, R. (2016).
Cross-cultural confirmation of bi-factor models of a symptom distress measure:
Symptom Checklist-90-Revised in clinical samples. Psychiatry Research, 239, 265–
274. doi:10.1016/j.psychres.2016.03.039
28. Yves Rosseel (2012). lavaan: An R Package for Structural Equation Modeling.
Journal of Statistical Software, 48(2), 1-36. URL http://www.jstatsoft.org/v48/i02/.
29. R Core Team (2020). R: A language and environment for statistical computing. R
Foundation for Statistical Computing, Vienna, Austria. URL https://www.R-
project.org/.
30. Finney, S. J., DiStefano, C., & Kopp, J. P. (2016). Overview of estimation methods
and preconditions for their application with structural equation modelling. In: K.
Schweizer & C. DiStefano. Principles and Methods of Test Construction (p. 218-
246). Hogrefe: Göttingen.
31. DiStefano, C. (2016). Examining fit with structural equation models. In: K.
Schweizer & C. DiStefano. Principles and Methods of Test Construction (p. 166-
193). Hogrefe: Göttingen
32. Rodriguez, A., Reise, S.P., Haviland, M.G., 2016. Evaluating bifactor models:
Calculating and interpreting statistical indices. Psychol. Methods 21, 137–150.
https://doi.org/10.1037/met0000045
33. Reise, S.P., Bonifay, W.E., Haviland, M.G., 2013. Scoring and Modeling
Psychological Measures in the Presence of Multidimensionality. J. Pers. Assess. 95,
129–140. https://doi.org/10.1080/00223891.2012.72543
34. David Dueber (2020). BifactorIndicesCalculator: Bifactor Indices Calculator. R
package version 0.2.0. https://CRAN.R-
project.org/package=BifactorIndicesCalculator
35. Pettersson, Erik, Henrik Larsson, Brian M. D'Onofrio, Sven Bölte, and Paul
Lichtenstein. "The General Factor of Psychopathology: a Comparison with the
General Factor of Intelligence with Respect to Magnitude and Predictive Validity."
World Psychiatry. 2020. 19.2: 206-213. Print.
36. Wang, Y., Duan, Z., Ma, Z. et al. Epidemiology of mental health problems among
patients with cancer during COVID-19 pandemic. Transl Psychiatry 10, 263 (2020).
https://doi.org/10.1038/s41398-020-00950-y
37. Ellis, W. E., Dumas, T. M., & Forbes, L. M. (2020). Physically isolated but socially
connected: Psychological adjustment and stress among adolescents during the initial
COVID-19 crisis. Canadian Journal of Behavioural Science / Revue canadienne des
sciences du comportement, 52(3), 177-187. http://dx.doi.org/10.1037/cbs0000215
38. VIANA, Maria Carmen and ANDRADE, Laura Helena. Lifetime Prevalence, age
and gender distribution and age-of-onset of psychiatric disorders in the São Paulo
Metropolitan Area, Brazil: results from the São Paulo Megacity Mental Health
Survey. Rev. Bras. Psiquiatr. [online]. 2012, vol.34, n.3 [cited 2020-08-09], pp.249-
260. Available from:
<http://www.scielo.br/scielo.php?script=sci_arttext&pid=S1516-
44462012000300004&lng=en&nrm=iso>. ISSN 1516-
4446. http://dx.doi.org/10.1016/j.rbp.2012.03.001.
39. DA SILVA, Antônio Geraldo et al. Mental health: why it still matters in the midst of
a pandemic. Braz. J. Psychiatry [online]. 2020, vol.42, n.3 [cited 2020-08-09],
pp.229-231. Available from:
<http://www.scielo.br/scielo.php?script=sci_arttext&pid=S1516-
44462020000300229&lng=en&nrm=iso>. Epub Apr 03, 2020. ISSN 1809-
452X. http://dx.doi.org/10.1590/1516-4446-2020-0009.
Table 1
Descriptive statistics of BSI score (sum of individual items answers).
BSI_S BSI_OC BSI_IS BSI_D BSI_A BSI_H BSI_PA BSI_PI BSI_P BSI_GSI

N 4039 4764 4764 4476 4764 4696 3878 4228 4764 2918
Missing 2388 1663 1663 1951 1663 1731 2549 2199 1663 3509
M 4.70 7.55 3.98 7.41 7.47 4.71 5.90 4.28 4.03 53.6
Md 3 6.00 3.00 6.00 6.00 4.00 5.00 3.00 3.00 45.0
SD 4.74 5.71 4.08 5.83 5.62 4.25 4.58 3.86 4.12 38.2
Min 0 0 0 0 0 0 0 0 0 0
Max 27 24 16 24 24 20 20 20 20 189

Note. BSI_S = Somatization; BSI_OC = Obsession-Compulsion; BSI_IS = Interpersonal Sensitivity; BSI_D =

Depression; BSI_A = Anxiety; BSI_H = Hostility; BSI_PA = Phobic Anxiety; BSI_PI = Paranoid Ideation;
BSI_P = Psychoticism; BSI_GSI = Global Severity Index.
Table 2
Reliability, sources of variances and replicability of the Bifactor Model for BSI
GSI S O IS D A H PA PI P
ECV_SG 0,77 0,04 0,03 0,01 0,01 0,03 0,04 0,04 0,02 0,01
ECV_GS 0,77 0,62 0,77 0,85 0,92 0,79 0,66 0,5 0,74 0,91
Omega 0,98 0,87 0,92 0,91 0,93 0,94 0,93 0,84 0,83 0,87
OmegaH 0,96 0,33 0,19 0,13 0,07 0,18 0,3 0,38 0,21 0,00
H 0,98 0,62 0,56 0,34 0,26 0,54 0,65 0,71 0,45 0,23
PUC 0,92
Note. S = Somatization; OC = Obsession-Compulsion; IS = Interpersonal Sensitivity; D = Depression; A =
Anxiety; H = Hostility; PA = Phobic Anxiety; PI = Paranoid Ideation; P = Psychoticism; GSI = Global Severity
Index; ECV_SG = Explained common variance – specific to group; ECV_GS = Explained common variance –
group to specific; Omega = Omega; OmegaH = Omega hierarchical; H = Construct replicability coefficient.
Figure 1 – See supplemental materials