УДК: 159.9.072.59:616.

89-07
Психологија: наука и практика, Vol. IV (7-8), 2020 Изворен научен труд

PSYCHOMETRIC PROPERTIES OF THE BRIEF SYMPTOM

INVENTORY IN A CLINICAL AND NONCLINICAL SAMPLE
Biljana Blazhevska Stoilkovska1*, Katerina Naumova*
Institute of Psychology, Faculty of Philosophy,
Ss. Cyril and Methodius University in Skopje

Abstract
The aim of this study is to investigate the proposed factor structure, reliability, and dis-
criminant validity of the Macedonian translation of one of the most widely used screening
and outcome measures. Both samples are drawn from two separate data sets. The clinical
sample (N = 149, 57% female) is composed of outpatients / participants, currently in use or
in need of mental health services, formally diagnosed by psychiatrists and/or psychologists
through a structured diagnostic clinical interview. All diagnoses were made in adherence
to ICD-10 criteria. The nonclinical sample (N = 180, 55% female) is composed of participants
not meeting diagnostic criteria for any mental disorder or not needing/using mental health
services in the previous six months. Confirmatory factor analysis was applied to examine
the construct validity of the BSI. The results supported the original nine-factor structure
in both samples, demonstrating acceptable model fit. Internal consistency of the overall
BSI was high. Discriminant validity was explored by comparing the clinical and nonclini-
cal sample on nine symptom dimensions and three global psychological distress indices.
As found, the BSI differentiates between the two groups with respect to all dimensions
and global indices. The study results indicate good psychometric properties of the BSI in
Macedonian context.

Keywords: Brief Symptom Inventory, factorial structure, discriminant validity

*  Both authors contributed equally to the manuscript and share first authorship.
1
  e-mail: [email protected]

-9-
B. BLAZHEVSKA STOILKOVSKA, K. NAUMOVA: PSYCHOMETRIC PROPERTIES OF THE BRIEF SYMPTOM…

The Brief Symptom Inventory (BSI) is one of the most widely used multidimen-
sional self-report instruments for the assessment of psychological symptoms. It has
been utilised both for screening, as well as for evaluation of treatment outcomes
with various clinical populations in different cultures (Derogatis, 2017; Derogatis
& Fitzpatrick, 2004). However, since the initial investigation of its factor structure
(Derogatis & Melisaratos, 1983), the evidence is inconclusive, as to whether the BSI
can accurately measure nine dimensions of psychopathology or if it is a unidimen-
sional measure of general psychological distress. In the past three decades, over
two dozen factor analytic studies of the BSI have demonstrated factor variance
across samples. It is worth noting that in most of these studies the results were
obtained from different types of exploratory factor analytic procedures (EFA) rather
than from confirmatory factor analysis (CFA).
When Derogatis and Melisaratos (1983) originally tested the hypothetical
structure of the instrument they used principal components analysis (PCA) with
varimax rotation to analyse data from a large sample of psychiatric outpatients.
Seven of the nine conceived symptom constructs were reproduced (psychoticism,
somatization, depression, hostility, paranoid ideation, obsessive-compulsive,
phobic anxiety). The eight factor (anxiety) was split into two well-defined clinical
components (panic, anxiety, and nervous tension), while the ninth (interpersonal
sensitivity) did not replicate, probably due to the small number of items (only four).
The nine factors accounted for 44% of the explained variance.
Subsequent studies with clinical samples did not find support for this factorial
solution. For example, one factor structure was derived from PCA in a relatively
homogeneous sample of forensic psychiatric inpatients and outpatients (Boulet
& Boss, 1991), as well as in adolescent and adult psychiatric inpatients, the lat-
ter mostly diagnosed with affective disorders (Piersma et al., 1994). Benishek et
al. (1998) found one and two-factor solutions in a sample of substance abusers
entering a treatment program using principal axis factoring (PAF) with oblimin
rotation, while the unidimensional model was supported by CFA with fit indices
below acceptable.
Among delinquent adolescents in detention, Whitt & Howard (2012) derived a
six-factor structure by PAF with oblimin rotation and confirmed it by maximum
likelihood (ML) CFA. A modified version of this model was supported by CFA in a
sample of adult men under criminal justice involvement (Valera et al., 2014).
In patients with spinal cord injuries, a six-factor solution was obtained from
PCA with oblique rotation and ML estimation (Heinrich & Tate, 1996), revealing
specific components of distress related to traumatic injury and physical illness.
Similarly, in a sample of primary care attenders, Schwannauer & Chetwynd (2007)

- 10 -
 ПСИХОЛОГИЈА: НАУКА И ПРАКТИКА, VOL. IV

applied ML with oblique rotation and detected six-factors dominated by a single

distress factor, while a series of CFAs in a sample of clinical psychology patients
in the same study supported a three-factor model. Finally, eight factors, highly
overlapping the original factor solution, were extracted from PCA with varimax
rotation in a sample of intellectually disabled participants (Kellet et al, 2004).
Equivocal results have also been reported in nonclinical community samples.
Hayes (1997) derived a six-factor model using PAF with oblimin rotation in college
and university counselling centre population. Likewise, in a study by Holden et al.
(2000)supported by CFA, three dimensions were extracted in university undergrad-
uates using PCA with varimax rotation. When comparing Indian and Canadian
university samples with multidimensional scaling, Watson and Sinha (1999) not
only found rather similar two dimensional solutions in both cultures, they also
found distinctive expressions of certain symptom dimensions, emphasizing the
necessity of developing national norms for psychopathology measures.
Since then, factor-analytic investigations of cultural adaptations of the BSI have
multiplied. For example, Ruipérez et al. (2001) extracted a six-factor solution from
PCA with oblimin rotation in a nonclinical adult sample in Spain, while Pereda et
al. (2007) found the original nine-factor solution in Spanish college students better
than a unidimensional model. However, they used a different translation and CFA
with elliptic robust least squares. Furthermore, a unidimensional solution was
derived from PCA with oblimin rotation in a nonclinical adult sample in Greece
(Loutsiou-Ladd et al., 2008).
In two separate analyses with large clinical samples, Printz et al. (2008, 2013)
report insufficient model fit of the original structure when investigating the
German translation. Contrary to these findings, in a large community sample
representative of the population of Hungary, Urban et al. (2014), using a series of
CFA’s found that the original nine-factor model had an acceptable fit. Similarly,
CFA on data from three separate representative Ukrainian general population
surveys (Sereda & Dembitskyi, 2016) also support the original internal structure
of the Ukrainian and Russian translations used in the studies. Kabát et al. (2018)
also report a sufficient fit of the original model using data from an adult sample
representative of the Czech population. However, taking into consideration the
ordinal response scale of the BSI and the expected non-normal distribution of
data when measuring psychopathology, these last three studies have appropri-
ately used the diagonally weighted least squares method (DWLS) to estimate the
parameters of the CFAs.
Findings from Asian and African cultures are sparse yet mixed. Daoud and Abo-
jedi (2010) report on a unidimensional invariate model extracted from PCA with

- 11 -
B. BLAZHEVSKA STOILKOVSKA, K. NAUMOVA: PSYCHOMETRIC PROPERTIES OF THE BRIEF SYMPTOM…

oblique rotation in nonclinical and clinical samples from Jordan, while Raghavan
et al. (2015) report on configural invariance of a two-factor model extracted from
PCA with oblimin rotation, across three groups of torture survivors from Tibet,
West Africa, and India.
Recently, the second order and bifactor model have received increasing attention.
Cross-ethnic measurement invariance using multiple-group CFA in individuals
treated for severe and persistent mental illness (Hoe & Brekke, 2008) supported a
secondary factor model, where the nine BSI subscales were indicators of a common
factor. Utilising multidimensional item response theory with archival clinical data,
Thomas (2012) found that the bifactor model slightly outperformed the original
model, while both outperformed the unidimensional model. The bifactor model
has also shown superior fit in CFA studies with community samples (Malloy-Diniz
et al., 2020; Urban, 2014).
In light of the existing variations in factor structure across samples, the aims
of this study are to (a) use confirmatory factor analysis to determine the factor
structure of the Macedonian translation of the BSI in a clinical and nonclinical
sample, and (b) examine the reliability and discriminant validity of the BSI.

Method
Participants and procedure

The participants are drawn from two separate data sets (Naumova, 2008; Nau-
mova & Naumov, 2019). The clinical sample (N = 149, 57% female, Mage = 34.3 years
± 13.8) is composed of outpatients formally diagnosed by psychiatrists (70%) and
participants using or in need of professional mental health care (30%) diagnosed
by psychologists through a structured diagnostic clinical interview (MINI, Sheehan
et al., 1998). All diagnoses were made in adherence to ICD-10 criteria: anxiety dis-
orders 48.3%, mood disorders 18.1%, stress-related 16.1% and other disorders 17.4%
(psychotic, conduct, personality and psychoactive substance use). The nonclinical
sample (N = 180, 55% female, Mage = 35.1 years ± 12.9) is composed of participants not
meeting diagnostic criteria for any mental disorder (69.4%) or not needing/using
professional mental health care in the previous six months (30.6%). The samples do
not differ significantly regarding age or gender. All participants provided informed
consent and completed the BSI as part of an extensive psychological assessment.

Instrument

The Brief Symptom Inventory (Derogatis & Melisaratos, 1983) consists of 53 items
measuring nine symptom dimensions: somatization (distress from perceptions of

- 12 -
 ПСИХОЛОГИЈА: НАУКА И ПРАКТИКА, VOL. IV

bodily dysfunction), obsessive-compulsive (unwanted, unremitting and irresistible

thoughts and impulses), interpersonal sensitivity (feelings of personal inadequacy
and inferiority), depression (dysphoric mood and affect, lack of motivation and loss
of interest in life), anxiety (nervousness, tension and panic), hostility (anger revealing
thoughts, feelings or actions), phobic anxiety (irrational and persistent fear response
to places, objects or situations), paranoid ideation (projective thoughts, suspicious-
ness, grandiosity and delusions), and psychoticism (social alienation to symptoms
of schizophrenia). Four items reflecting vegetative and other symptoms are not
included in the subscales, since they are related to several, but are not unique to
any specific dimension. Responses are given on a 5-point scale indicating degrees
of distress “over the past 7 days including today”, ranging from 0 (not at all) to 4
(extremely). Three global indices can also be calculated: (a) the General Severity
Index (GSI), measuring overall distress; (b) the Positive Symptom Total (PST) is the
count of reported symptoms; and (c) the Positive Symptom Distress Index (PSDI),
measuring symptom severity, adjusted for the number of experienced symptoms.
The BSI was translated into Macedonian and adapted by the second author and
a senior colleague as part of an international research project on mental health
outcomes in people exposed to war stressors (Priebe et al., 2004).

Results

Descriptive statistics for somatization, obsessive-compulsive, interpersonal sensi-

tivity, depression, anxiety, hostility, phobic anxiety, paranoid ideation, and psychot-
icism, as well as GSI, PST, and PSDI in both samples, are summarized in Table 1.
As can be seen, psychoticism and phobic anxiety symptoms were rated as least
distressing in both samples. In line with the prevalence of diagnosed disorders in
the clinical sample, anxiety symptoms are most pronounced in this group, followed
by obsessive-compulsive and depression symptoms, while the participants in the
nonclinical sample reported paranoid ideation and anxiety as most frequently
experienced symptoms.

Factorial validity
In order to test the nine first-order factors of the BSI Confirmatory factor
analysis (CFA) was performed. Due to sample sizes, the ordinal response scale,
and multivariate non-normality, the weighted least squares mean, and variance
adjusted (WLSMV) method was used for parameter estimation. This estimator was
chosen since it is more accurate than the well-known maximum likelihood (ML)
when the assumption of multivariate normality is not met (Li, 2016). As suggested
by Kline (2016), model fit was assessed by the following goodness of fit indices - χ2

- 13 -
B. BLAZHEVSKA STOILKOVSKA, K. NAUMOVA: PSYCHOMETRIC PROPERTIES OF THE BRIEF SYMPTOM…

test statistic, comparative fit index (CFI), root mean square error of approximation
(RMSEA), and standardized root mean square residual (SRMR). Criteria for good
model fit are CFA ≥ 0.95, RMSEA ≤ 0.06, and SRMR ≤ 0.08 (Hu & Bentler, 1999). If
the older guidelines for the model fit, i.e. CFA ≥ 0.90, RMSEA ≤ 0.10, and SRMR
≤ 0.10 are not met, then the model is not acceptable (Weston & Gore, Jr., 2006).
Accordingly, CFA between 0.90 and 0.95, RMSEA between 0.06 and 0.10, and SRMR
between 0.08 and 0.10 indicate acceptable fit of the estimated model.

Table 1. Descriptive statistics of BSI symptom dimensions and global indices

Sample N M SD Mdn Min. Max. Skewness Kurtosis
Somatization clinical 149 1.16 .91 1.00 0.00 3.86 .49 -.63
nonclinical 180 .48 .55 .28 0.00 2.29 1.25 .81
Obsessive- clinical 149 1.54 1.04 1.33 0.00 3.83 .37 -.95
compulsive nonclinical 180 .58 .64 .33 0.00 3.17 1.34 1.46
Interpersonal clinical 149 1.30 .96 1.25 0.00 3.75 .50 -.74
sensitivity nonclinical 180 .35 .59 0.00 0.00 3.00 2.20 5.16
Depression clinical 149 1.41 1.00 1.33 0.00 3.83 .53 -.62
nonclinical 180 .40 .54 0.17 0.00 3.33 2.34 7.31
Anxiety clinical 149 1.75 .99 1.50 0.00 4.00 .47 -.61
nonclinical 180 .67 .64 0.50 0.00 3.33 1.37 2.25
Hostility clinical 149 1.11 .93 1.00 0.00 3.80 .88 -.01
nonclinical 180 .48 .60 0.40 0.00 4.00 2.56 9.29
Phobic clinical 149 1.00 .91 .80 0.00 3.80 1.02 .37
anxiety nonclinical 180 .23 .37 0.00 0.00 2.60 2.68 10.85
Paranoid clinical 149 1.40 .89 1.20 0.00 4.00 .54 -.25
ideation nonclinical 180 .69 .62 0.60 0.00 2.60 0.99 .61
Psychoticism clinical 149 .95 .82 0.80 0.00 3.00 .67 -.60
nonclinical 180 .20 .40 0.00 0.00 2.40 2.54 7.31
GSI clinical 149 1.30 .73 1.13 0.15 3.19 .60 -.50
nonclinical 180 .46 .42 .33 0.00 2.30 1.68 3.35
PST clinical 149 31.07 11.82 32.00 4.00 51.00 -.20 -.76
nonclinical 180 14.34 11.08 11.00 0.00 46 .97 .04
PSDI clinical 149 2.12 .55 2.06 1.16 3.64 .38 -.38
nonclinical 180 1.60 .47 1.57 0.00 4.00 .95 3.62

CFA results obtained on the clinical sample (Table 2) demonstrated that CFI
was higher than 0.90 (slightly below 0.95) and SRMR was below 0.08, indicating
acceptable model fit. The value below 0.05 of the RMSEA (90% CI[0.022, 0.040] de-
noted good fit of the model. WLSMV χ2 was statistically significant, revealing that
the observed and predicted covariance matrices differ significantly. It should be
noted that this fit statistic is liberal when variables are not normally distributed
(http://davidakenny.net/cm/fit.htm) and when strong correlation exists among
the observed variables (Kline, 2016). Hence, there is a possibility for many Type I

- 14 -
 ПСИХОЛОГИЈА: НАУКА И ПРАКТИКА, VOL. IV

errors. In addition, it is based on the proposition of exact fit, while reporting on

reasonable fit is adequate as well (Brown, 2006). Factor loadings ranged from 0.89
to 0.38 (all statistically significant at p < .001). Considering all four fit indices, the
BSI model observed here could be considered acceptable.
As shown in Table 2, CFA revealed adequate fit of the BSI model in the non-
clinical sample. Specifically, WLSMV χ2 was not statistically significant indicating
that there is no difference among the observed data and the predicted model.
The CFI was 0.90 and SRMR was slightly above 0.08 revealing that the model fit is
acceptable. RMSEA was lower than 0.05 with 90% CI[0.000, 0.028] denoting good
fit of the model. Factor loadings ranged from 0.84 to 0.14 showing that two items'
loadings were not statistically significant.

Table 2. Goodnes-of-fit indices for the nine-factor model of the BSI

Model WLSMVχ2 df CFI SRMR RMSEA (90% CI)
Clinical sample 1254.92*** 1091 0.94 0.07 0.03 [0.02, 0.04]
Nonclinical sample 1159.91 1091 0.90 0.09 0.02 [0.00, 0.03]
p < .001

Discriminant validity
Considering that the BSI dimensions were not normally distributed, the
Mann-Whitney U test was employed to explore discriminant validity of BSI. Dif-
ferences in GSI and all nine symptom dimensions between the clinical and non-
clinical sample were analysed.
Results showed that participants in the clinical group had significantly higher
scores on all symptom dimensions, as well as on all global indices (Table 3). These
findings confirm the discriminant validity of the BSI.

Reliability

Reliability of the scale and subscales was estimated using Cronbach alpha
internal consistency coefficient. Overall reliability of the BSI in the clinical and
nonclinical sample was high (α = .96 and α = .95, respectively). Cronbach’s alpha
coefficients for separate BSI symptom dimensions in the clinical sample ranged
from 0.86 for anxiety and depression to 0.70 for interpersonal sensitivity. Reliabil-
ity of BSI symptom dimensions in the nonclinical sample ranged from α = .82 for
obsessive-compulsive and interpersonal sensitivity to α = .56 for phobic anxiety.
All reliability coefficients are presented in Table 4.

- 15 -
B. BLAZHEVSKA STOILKOVSKA, K. NAUMOVA: PSYCHOMETRIC PROPERTIES OF THE BRIEF SYMPTOM…

Table 3. Results of the Mann-Whitney U test: differences in BSI symptom

dimensions and global indices between the clinical and nonclinical sample
Sample N Mean Rank Sum of Ranks U z
Somatization clinical 149 206,34 30745,00 7250,000 -7,22***
nonclinical 180 130,78 23540,00
Obsessive- clinical 149 216,33 32233,50 5761,500 -8.95***
compulsive nonclinical 180 122,51 22051,50
Interpersonal clinical 149 222,06 33087,50 4907,500 -10,14***
sensitivity nonclinical 180 117,76 21197,50
Depression clinical 149 223,29 33269,50 4725,500 -10,20***
nonclinical 180 116,75 21015,50
Anxiety clinical 149 223,93 33366,00 4629,000 -10,25***
nonclinical 180 116,22 20919,00
Hostility clinical 149 205,39 30603,50 7391,500 -7,07***
nonclinical 180 131,56 23681,50
Phobic anxiety clinical 149 218,68 32583,00 5412,000 -9,57***
nonclinical 180 120,57 21702,00
Paranoid ideation clinical 149 208,67 31092,00 6903,000 -7,61***
nonclinical 180 128,85 23193,00
Psychoticism clinical 149 221,42 32991,00 5004,000 -10,26***
nonclinical 180 118,30 21294,00
GSI clinical 149 230,49 34343,00 3652,000 -11,36***
nonclinical 180 110,79 19942,00
PST clinical 149 227,09 33837,00 4158,000 -10,78***
nonclinical 180 113,60 20448,00
PSDI clinical 149 213,32 31784,00 6211,000 -8,38***
nonclinical 180 125,01 22501,00
*** p < .001

Table 4. Internal consistency coefficients of the BSI symptom dimensions

and GSI
Clinical sample Nonclinical sample
BSI dimensions N of items Cronbach’s alpha Cronbach’s alpha

Somatization 7 .84 .77

Obsessive-compulsive 6 .86 .82
Interpersonal sensitivity 4 .70 .82
Depression 6 .86 .78
Anxiety 6 .86 .80
Hostility 5 .82 .76
Phobic anxiety 5 .74 .56
Paranoid ideation 5 .74 .64
Psychoticism 5 .74 .67
GSI 53 .96 .95

- 16 -
 ПСИХОЛОГИЈА: НАУКА И ПРАКТИКА, VOL. IV

Descriptive, reliability, and nonparametric tests were performed in SPSS 24.0.

The CFA analyses were conducted with the Lavaan package (Rosseel, 2012) in R
environment (R Core Team, 2020).

Discussion

Our analysis confirmed the Macedonian BSI factor structure, as proposed by

Derogatis and Melisaratos (1983), both in clinically diagnosed adults as well as in
a community sample. More precisely, the original model fit the data adequately
in the clinical sample. The majority of items had satisfactory factor loadings >.60,
while only a few items had loadings between .45 and .55 and only one had a load-
ing slightly below .40. Accordingly, most of the items shared at least 36% of the
variance with the factors, thus providing support for the conclusion regarding
the model fit, i.e. classification of the measured psychopathology symptoms into
nine groups as predicted.
In addition, the reliability coefficients of all subscales, i.e. somatization, ob-
sessive-compulsive, interpersonal sensitivity, depression, anxiety, hostility, phobic
anxiety, paranoid ideation, and psychoticism varied from acceptable to very good,
representing satisfactory internal consistency. Taken together, both analyses
proved the psychometric characteristics of the BSI, namely its construct validity
and reliability in the clinical sample.
Findings obtained from the nonclinical sample suggested that the BSI model
could be considered acceptable, although several constraints were registered in
our analyses, such as small factor loadings on several items, and lower Cronbach’s
alpha coefficients for three subscales (phobic anxiety, paranoid ideation and psy-
choticism). However, due to the varying manifestation of psychological distress
in the general population as opposed to in defined clinical populations, prior to
item or subscale modification, additional confirmatory factor analysis is advised
with larger urban and educationally more diverse community samples.
The Mann-Whitney U test revealed that BSI subscales and global indices distin-
guish participants in the clinical sample from those in the nonclinical sample. That
is, all nine symptom dimensions, overall distress, number of reported symptoms,
and level of their severity were higher among clinically diagnosed individuals,
thus confirming the discriminant validity of the BSI.
Our findings are in line with recent studies that have also used robust CFA es-
timators (Kabát et al., 2018; Sereda & Dembitskyi, 2016; Urban et al., 2014) lending
further evidence to the notion that previously reported variations in the factorial
structure of the BSI could result to a greater degree from the use of less powerful

- 17 -
B. BLAZHEVSKA STOILKOVSKA, K. NAUMOVA: PSYCHOMETRIC PROPERTIES OF THE BRIEF SYMPTOM…

and methodologically diverse factor analytic procedures, rather than from inherent
instability of the original BSI model (Loutsiou-Ladd et al., 2008).
An additional strength of our study is both the use of a clinically diagnosed
sample and the diagnostic screening of the nonclinical group, thus eliminating
potential threats to the discriminant validity of the measure resulting from of-
ten undetected yet increased psychological distress or treatment involvement in
community samples (Thurston et al., 2008).

Conclusion

In this study, the nine-factor structure of the BSI examined using CFA was repro-
duced in both clinical and non-clinical samples thus demonstrating its construct
validity. In addition, the study results confirmed its discriminant validity and
internal consistency. The exception was found to be the phobic anxiety subscale
in the nonclinical sample.
Second-order and bifactor models of the BSI could be investigated in future
studies. Confirmed psychometric characteristics of the Macedonian translation of
the BSI justify its application both in research and clinical settings. The results are
particularly important since this is an easily administered self-report measure of
psychological status that can be utilized for screening, assessment, and treatment
outcomes in various clinical populations.

- 18 -
 ПСИХОЛОГИЈА: НАУКА И ПРАКТИКА, VOL. IV

References

Benishek, L. A., Hayes, C. M., Bieschke, K. J., & Stöffelmayr, B. E. (1998).

Exploratory and confirmatory factor analyses of the Brief Symptom
Inventory among substance abusers. Journal of Substance Abuse, 10(2),
103–114. https://doi.org/10.1016/S0899-3289(99)80127-8
Boulet, J., & Boss, M. W. (1991). Reliability and validity of the Brief Symptom
Inventory. Psychological Assessment: A Journal of Consulting and Clinical
Psychology, 3(3), 433–437. https://doi.org/10.1037/1040-3590.3.3.433
Brown, T. A. (2006). Confirmatory Factor Analysis for applied research. The
Guilford Press.
Daoud, F. S., & Abojedi, A. A. (2010). Equivalent factorial structure of the
Brief Symptom Inventory (BSI) in clinical and nonclinical Jordanian
populations. European Journal of Psychological Assessment, 26(2), 116–121.
https://doi.org/10.1027/1015-5759/a000016
Derogatis, L. R. (2017). Symptom Checklist-90-Revised, Brief Symptom Inventory,
and BSI-18. In M. E. Maruish (Ed.), Handbook of psychological assessment in
primary care settings (p. 599–629). Routledge/Taylor & Francis Group.
Derogatis, L. R., & Fitzpatrick, M. (2004). The SCL-90-R, the Brief Symptom
Inventory (BSI), and the BSI-18. In M. E. Maruish (Ed.), The use of
psychological testing for treatment planning and outcomes assessment:
Instruments for adults (p. 1–41). Lawrence Erlbaum Associates Publishers.
Derogatis, L. R., & Melisaratos, N. (1983). The Brief Symptom Inventory: An
introductory report. Psychological Medicine, 13(3), 595–605. https://doi.
org/10.1017/S0033291700048017
Hayes, J. A. (1997). What does the Brief Symptom Inventory measure in college
and university counseling center clients? Journal of Counseling Psychology,
44(4), 360-367. https://doi.org/10.1037/0022-0167.44.4.360
Heinrich, R. K., & Tate, D. G. (1996). Latent variable structure of the Brief
Symptom Inventory in a sample of persons with spinal cord injuries.
Rehabilitation Psychology, 41(2), 131–147. https://doi.org/10.1037/0090-
5550.41.2.131
Hoe, M., & Brekke, J. S. (2008). Cross-ethnic measurement invariance of the
Brief Symptom Inventory for individuals with severe and persistent
mental illness. Social Work Research, 32(2), 71-78. https://doi.org/10.1093/
swr/32.2.71
Holden, R. R., Starzyk, K. B., McLeod, L. D., & Edwards, M. J. (2000).
Comparisons among the Holden Psychological Screening Inventory
(HPSI), the Brief Symptom Inventory (BSI), and the Balanced Inventory

- 19 -
B. BLAZHEVSKA STOILKOVSKA, K. NAUMOVA: PSYCHOMETRIC PROPERTIES OF THE BRIEF SYMPTOM…

of Desirable Responding (BIDR). Assessment, 7(2), 163–175. https://doi.

org/10.1177/107319110000700208
Hu, L.-t., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance
structure analysis: Conventional criteria versus new alternatives.
Structural Equation Modeling, 6(1), 1-55.
Kabát, J., Kaščáková, N., Fürstová, J., Bartůš-Ková, L., Glogar, P., Šolcová, I. P.,
& Tavel, P. (2018). Psychometrické charakteristiky české verze stručného
inventáře příznaků (BSI-53) [Psychometric characteristics of the Czech
version of the Brief Symptom Inventory (BSI -53)]. Československá
Psychologie: Časopis Pro Psychologickou Teorii a Praxi, 62, 19–39.
Kellett, S., Beail, N., Newman, D. W., & Hawes, A. (2004). The factor structure
of the Brief Symptom Inventory: Intellectual disability evidence. Clinical
Psychology & Psychotherapy, 11(4), 275–281. https://doi.org/10.1002/cpp.410
Kenny, D. A. (2020, June 5). Measuring Model Fit. http://davidakenny.net/cm/fit.
htm
Kline, R. B. (2016). Principles and Practice of Structural Equation Modeling (4th
ed.). The Guilford Press
Li, C. (2016). Confirmatory factor analysis with ordinal data: Comparing robust
maximum likelihood and diagonally weighted least squares. Behavior
Research Methods, 48, 936– 949. https://doi.org/10.3758/s13428-015-0619-7
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(1), 90–110. https://doi.
org/10.1080/15305050701808680
Malloy-Diniz, L., de Oliveira Serpa, A. L., de Souza Costa, D., Pinheiro, M. I.
C., Diaz, A. P., de Paula, J. J., Miranda, D. M., & Silva, A. G. (2020). Brief
Symptoms Inventory psychometric properties supports the hypothesis
of a general psychopathological factor. [Preprint]. PsyArXiv. https://doi.
org/10.31234/osf.io/7jk2r
Naumova, K. (2008). Traumatskite iskustva, mentalnoto zdravje i kvalitetot na
zhivot kaj licata koi dozhiveale voeni stresovi [Traumatic experiences,
mental health, and quality of life in individuals exposed to war stressors]
(Unpublished master’s thesis). Ss. Cyril and Methodius University in
Skopje.
Naumova, K., & Naumov, F. (2019). [Unpublished raw data on cognitive,
emotional and personality factors in a clinical and nonclinical sample].
Pereda, N., Forns, M., & Peró, M. (2007). Dimensional structure of the Brief
Symptom Inventory with Spanish college students. Psicothema, 19(4),
634-639.

- 20 -
 ПСИХОЛОГИЈА: НАУКА И ПРАКТИКА, VOL. IV

Piersma, H. L., Boes, J. L., & Reaume, W. M. (1994). Unidimensionality of the

Brief Symptom Inventory (BSI) in adult and adolescent inpatients.
Journal of Personality Assessment, 63(2), 338–344. https://doi.org/10.1207/
s15327752jpa6302_12
Priebe, S., Gavrilović-Janković, J., Schützwohl, M., Galeazzi, M. G., Lečić-Toševski,
D., Ajduković, D., Francikovic, T., Kucukalic, A. & Popovski, M. (2004).
Study of long-term clinical and social outcomes after war experiences in
ex-Yugoslavia: Methods of the ‘CONNECT’project. Psihijatrija danas, 36(1),
101-110.
Prinz, U., Nutzinger, D.O., Schulz, H., Petermann, F., Braukhaus, C., &
Andreas, S.(2008). Die Symptom Checkliste 90-R und ihre kurzversionen:
Psychometrische analysen bei patienten mit psychischen erkrankungen
[The Symptom Check List 90-R and their short versions in patients
with mental disorders]. Physikalische Medizin, Rehabilitationsmedizin,
Kurortmedizin, 18(6), 337–343. https://doi.org/10.1055 / s-0028-1093323
Prinz, U., Nutzinger, D. O., Schulz, H., Petermann, F., Braukhaus, C., & Andreas,
S. (2013). Comparative psychometric analyses of the SCL-90-R and its short
versions in patients with affective disorders. BMC Psychiatry, 13(1), 104.
https://doi.org/10.1186/1471-244X-13-104
R Core Team (2020). R: A language and environment for statistical computing. R
Foundation for Statistical Computing.
Raghavan, S. S., Rosenfeld, B., & Rasmussen, A. (2015). Measurement
invariance of the Brief Symptom Inventory in survivors of torture and
trauma. Journal of Interpersonal Violence, 32(11), 1708–1729. https://doi.
org/10.1177/0886260515619750
Rosseel, Y. (2012). lavaan: An R Package for Structural Equation Modeling.
Journal of Statistical Software, 48(2), 1–36.
Ruipérez, M. Á., Ibáñez, M. I., Lorente, E., Moro, M., & Ortet, G. (2001).
Psychometric properties of the Spanish version of the BSI. European
Journal of Psychological Assessment, 17(3), 241–250. https://doi.
org/10.1027//1015-5759.17.3.241
Schwannauer, M., & Chetwynd, P. (2007). The Brief Symptom Inventory: A
validity study in two independent Scottish samples. Clinical Psychology &
Psychotherapy, 14(3), 221–228. https://doi.org/10.1002/cpp.539
Sereda, Y., & Dembitskyi, S. (2016). Validity assessment of the Symptom
Checklist SCL-90-R and shortened versions for the general population in
Ukraine. BMC Psychiatry, 16(1), 300. https://doi.org/10.1186/s12888-016-
1014-3

- 21 -
B. BLAZHEVSKA STOILKOVSKA, K. NAUMOVA: PSYCHOMETRIC PROPERTIES OF THE BRIEF SYMPTOM…

Sheehan, D. V., Lecrubier, Y., Sheehan, K. H., Amorim, P., Janavs, J., Weiller,
E., Hergueta, T., Baker, R., & Dunbar, G. C. (1998). The Mini-International
Neuropsychiatric Interview (M.I.N.I.): The development and validation of
a structured diagnostic psychiatric interview for DSM-IV and ICD-10. The
Journal of Clinical Psychiatry, 59 (Suppl 20), 22–57.
Thomas, M. L. (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. https://doi.org/10.1037/
a0024712
Thurston, I.B., Curley, J., Fields, S., Kamboukos, D., Rojas, A., & Phares, V.
(2008). How nonclinical are community samples? Journal of Community
Psychology, 36(4), 411-420. https://doi.org/10.1002/jcop.20223
Urbán, R., Kun, B., Farkas, J., Paksi, B., Kökönyei, G., Unoka, Z., Felvinczi, K.,
Oláh, A., & 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. https://doi.
org/10.1016/j.psychres.2014.01.027
Valera, P., Fullilove, R., Cali, S., Nunes, E., Chiongbian, V., Clark, W., & Covey,
L. (2015). The psychometric properties of the Brief Symptom Inventory in
men under criminal justice involvement: Implications for forensic social
workers in practice settings. British Journal of Social Work, 45(7), 2210–
2223. https://doi.org/10.1093/bjsw/bcu074
Watson, D. C., & Sinha, B. K. (1999). A cross-cultural comparison of the Brief
Symptom Inventory. International Journal of Stress Management, 6(4), 255-
264. https://doi.org/10.1023/A:1021940321129
Weston, R., & Gore, P. (2006). A brief guide to Structural Equation Modeling. The
Counseling Psychologist, 34(5), 719-751.
Whitt, A., & Howard, M. O. (2012). Brief Symptom Inventory factor structure in
antisocial adolescents: Implications for juvenile justice. Research on Social
Work Practice, 22(2), 166–173. https://doi.org/10.1177/1049731511418786

- 22 -
 ПСИХОЛОГИЈА: НАУКА И ПРАКТИКА, VOL. IV

ПСИХОМЕТРИСКИ КАРАКТЕРИСТИКИ НА КРАТКИОТ

ИНВЕНТАР НА СИМПТОМИ КАЈ ИСПИТАНИЦИ ОД
КЛИНИЧКА И НЕКЛИНИЧКА ПОПУЛАЦИЈА
Билјана Блажевска Стоилковска
Катерина Наумова

Кратка содржина
Целта на ова истражување е да се испита предложената факторска структура, релија-
билноста и дискриминантната валидност на македонската верзија на една од најчесто
користените мерки за тријажа и евалуација на третман. Испитаниците од двете групи
се извлечени од две постоечки бази на податоци. Клиничкиот примерок (N = 149, 57%
жени) се состои од испитаници со формалнo дијагностицирани ментални растројства
од страна на психијатри, кои се тековно на вонхоспитален третман и од испитаници
кои користат или имаат потреба од услуги за третман на нарушено ментално здравје,
дијагностицирани од страна на психолози со примена на структурирано клиничко
интервју. Сите дијагнози се поставени во согласност со критериумите на МКБ-10.
Неклиничкиот примерок (N = 180, 55% жени) се состои од лица кои не задоволуваат
дијагностички критериуми за ниту едно ментално растројство или во последните
шест месеци не користат/немаат потреба од услуги за третман на нарушено ментално
здравје. За проверка на валидноста на конструктот беше применета конфирматорна
факторска анализа. Резултатите ја поддржаа оригиналната структура од девет фак-
тори во двата примероци, со прифатливи вредности на индексите на согласување на
моделот. Покрај тоа, утврдена беше и висока внатрешна конзистентност на мерката.
Дискриминантната валидност беше проверена преку споредба на двете групи испита-
ници во однос на деветте димензии на психопатолошки симптоми и трите глобални
индекси на психичка вознемиреност. Инвентарот успешно ги диференцира двете групи
во однос на сите поединечни супскали и трите општи индекси. Резултатите во целост
упатуваат на задоволителни психометриски карактеристики на македонската верзија
на Краткиот инвентар на симптоми.

Клучни зборови: Краток инвентар на симптоми, факторска структура,

дискриминантна валидност

- 23 -