Vegas 

et al. Psicologia: Reflexão e Crítica

https://doi.org/10.1186/s41155-022-00222-2
(2022) 35:18 Psicologia: Reflexão e Crítica

RESEARCH Open Access

Psychometric properties of the FACES IV

package for Spanish adolescents
María I. Vegas1*   , Manuel Mateos‑Agut2   , Pedro J. Pineda‑Otaola3    and Carlota Sebastián‑Vega4    

Abstract 
The family plays an essential role in the life of an adolescent. Hence, an acceptable understanding and an evaluation
of family functioning is fundamental for effective interventions with adolescents in the psychological, social, and
educational fields. The main purpose of this study is to examine the psychometric properties of the Family Adaptabil‑
ity and Cohesion Evaluation Scale (FACES IV), the Family Communication Scale (FCS), and the Family Satisfaction Scale
(FSS), for assessing the family functioning of Spanish adolescents. The sample was comprised of 1187 adolescents
between 14 -18 years old (49.96% boys and 50.04% girls; M = 16.17; SD = 1.31) from Castile and Leon (Spain), selected
from 23 educational centers, 10 university degree courses, and 18 specific juvenile centers for adolescents with either
family or behavioral problems. The scales of Balanced Cohesion, Balanced Flexibility and Disengaged showed good
convergent validity, while Enmeshed, Rigid, and Chaotic did not. For this reason some items were removed, obtain‑
ing a shortened version of FACES IV, that demonstrated acceptable reliability, and good convergent and predictive
validity. The FCS and FSS scales yielded excellent psychometric properties. The results confirmed the factorial struc‑
ture of the FACES IV, its transcultural applicability, and its validity for different ages. The hypotheses of the circumplex
model were confirmed, except for the dysfunctionality of two scales, Enmeshed and Rigid, that contrary to what was
expected, showed positive correlations with Family Communication, Family Satisfaction, Balanced Cohesion, and
Balanced Flexibility. In brief, our results present the FACES IV package as a useful instrument for the assessment of
family functioning of Spanish adolescents. Future studies will be necessary to confirm the trend observed for the two
aforementioned scales among adolescents.
Keywords:  FACES IV, Circumplex model, Adolescence, Family Communication, Family Satisfaction, Cohesion,
Flexibility, Family functioning, Validation, Family assessment

Introduction the conduct of adolescents. Hence, an acceptable under-

The family plays an essential role in the life of an ado- standing and an evaluation of family functioning is fun-
lescent. It is the principal source of protection, security, damental for effective interventions with adolescents in
emotional support, and socialization, and the model the psychological, social, and educational fields.
for the transfer of norms, values and beliefs, and for the Over 200 concepts have been used in family develop-
moral development of the adolescent (Bhugra & Fiorillo, ment theory, in order to try to assess the principal aspects
2012; Minuchin, 2009; Samper et  al., 2006; UN, 2016). of couple and family functioning. In the late 1970s,
Families are changing nowadays and many new problems these concepts were clustered into three main dimen-
have emerged in the family scene over recent decades, sions (Olson et  al., 2019), giving rise to the Circumplex
with direct impacts on the perceptions, the feelings, and Model of Family and Marital Systems (Olson et al., 1979).
Based on this model, the Family Adaptability and Cohe-
sion Evaluation Scale (FACES) was designed. This is one
*Correspondence: mivegas2000@hotmail.com
of the most important questionnaires at an international
1
Universidad de Burgos, C/ Alfonso XI, s/n, 09007 Burgos, Spain level for the assessment of family functioning, because of
Full list of author information is available at the end of the article

© The Author(s) 2022, corrected publication 2022. Open Access This article is licensed under a Creative Commons Attribution 4.0
International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you
give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes
were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated
otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not
permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To
view a copy of this licence, visit http://​creat​iveco​mmons.​org/​licen​ses/​by/4.​0/.
Vegas et al. Psicologia: Reflexão e Crítica (2022) 35:18 Page 2 of 12

its theoretical foundations, and its applicability to clini- Chaotic. The variables Balanced Cohesion and Balanced
cal scenarios and investigations with families (Hamilton Flexibility function as balanced scales and cover the
& Carr, 2016; Jiménez et  al., 2017; Olson, 2008; White intermediate parts of each dimension, while the vari-
& Klein, 2008). This model incorporates the systemic ables Disengaged, Enmeshed, Rigid, and Chaotic are the
theory and the theory of family development (Olson & extreme or unbalanced scales (Olson, 2011) and cover
Gorall, 2003) and has been used in over 1,200 studies the extremely high or extremely low levels of both
(Olson et  al., 2019). Sanderson et  al. (2009) pointed to dimensions.
the FACES questionnaire as the third measure of family These scales that cover the three basic constructs of
functioning that is most frequently used in investigations the circumplex model were later on supplemented with
on family and couple therapy. the Family Satisfaction Scale (FSS) by Olson (Olson,
According to the circumplex model, the three dimen- 2000; Olson et al., 2019), a form of measuring satisfaction
sions that are considered crucial to understand family through the three principal dimensions of the circumplex
functioning are Cohesion, Flexibility, and Communica- model.
tion (Barnes & Olson, 1985; Olson, 2011; Olson & Gorall, The FACES IV version has been validated for the adult
2003; Olson et  al., 1979). Cohesion is defined as “the population in various European countries (Baiocco et al.,
emotional bonding that family members have toward 2013; Gomes et al., 2017; Koutra et al., 2012; Margasiński,
one another” (Olson et al., 2019, p. 201). Flexibility is “the 2015; Mirnics et  al., 2010; Pereira & Teixeira, 2013;
amount of change in its leadership, roles relationships, Sequeira et al., 2021) and, likewise, in Spain by Martínez-
and relationship rules” (Olson et al., 2019, p. 202). Com- Pampliega et  al. (2017). The FCS and FSS scales have
munication is considered as a facilitating dimension of also been validated in Spain (Sanz et al., 2002) and other
the other two and is defined as “the positive communica- countries (Costa-Ball & Cracco, 2021; Cracco & Costa-
tion skills utilized in the couple or family system” (Olson, Ball, 2019; Gomes et al., 2017; Koutra et al., 2012; Pereira
2011, p.65). & Teixeira, 2013; Sequeira et al., 2021).
The principal hypotheses on family functioning arising Everri et al. (2020) insisted on the utility of the FACES
from the circumplex model are that partners and families IV scales for adolescents, because it provides more infor-
with balanced levels of Cohesion and Flexibility will func- mation on understanding the parent–offspring relations.
tion more effectively than those with extreme values in However, none of the FACES IV Package scales has been
both dimensions and that positive communication skills validated for the population of Spanish adolescents and
will permit the balanced systems to modify their levels a few validations have been conducted for adolescents
of Cohesion and Flexibility and will, therefore, give rise worldwide. We consider that an ex profeso validation
to systems with healthier family functioning (Olson & of the FACES IV Package for adolescents is necessary,
Gorall, 2003; Olson et al., 1979, 2019). because adolescents have very specific family needs and
It may be inferred from the above definitions that the present peculiar features, which requires a differenti-
dimensions of Cohesion and Flexibility have a curvilin- ated study of family functioning. In fact, the authors of
ear relation with family functioning, because either very international studies completed with adolescents using
high or very low values are problematic or dysfunctional. FACES IV (Baiocco et  al., 2013; Desautels et  al., 2016;
The same is not so with Communication, considered Everri et  al., 2015, 2016, 2020; Gouveia-Pereira et  al.,
as a linear dimension, as the higher the communicative 2020; Sebokova et al., 2016) all agree that the tool has to
value that is achieved, the better the system will function be validated with adolescents, due to the specific char-
(Olson et al., 1979, 2019). acteristics of this population. “When non-adult samples
The level of positive communication within the fam- are concerned, additional aspects need to be taken in
ily is measured through the Family Communication consideration” (Everri et al., 2020, p. 2508). In the Cana-
Scale (FCS) (Barnes & Olson, 1985), whereas Cohesion dian study, the results suggested a degree of fine-tuning
and Flexibility are measured through the questionnaire in the use of the version with adolescents, despite Olson
Family Adaptability and Cohesion Evaluation Scale and Gorall’s recommendations that the FACES IV may be
(FACES). Over the past four decades, there have been used with both adults and adolescents (Desautels et  al.,
numerous versions of the circumplex model, the last of 2016, p. 110).
which, FACES IV, validated by Olson in 2011, presented On the other hand, extremely low levels of Flexibility
the possibility of covering the central and the extreme (Rigid) or extremely high levels of Cohesion (Enmeshed)
parts of Cohesion and Flexibility through six scales. The are hypothesized as problematic for individuals and
dimension Cohesion includes three variables, Balanced relational development in general, according to the
Cohesion, Disengaged, and Enmeshed, while the dimen- predictions of Olson’s circumplex model (Olson et  al.,
sion Flexibility includes Balanced Flexibility, Rigid, and 2019). Nevertheless, these standards are not necessarily
Vegas et al. Psicologia: Reflexão e Crítica (2022) 35:18 Page 3 of 12

evident among adolescents: in their studies conducted reconstituted family. 1.4% lived in other family arrange-
with adolescents, Baiocco et  al. (2013) and Everri et  al., ments different from the above.
(2015, 2016), found positive correlations between Rigid All participants were Spanish.
and the balanced variables, Cohesion and Flexibility,
and between Rigidity and Enmeshed. In the studies of Procedure
Everri et  al., Rigidity consistently emerged as an indica- The sampling at the non-university educational centers
tor of adaptive family functioning, and an association was was completed in two stages: stratified sampling (center
found between Rigidity and parental monitoring, as well and educational level) and sampling by blocks (each
as between Rigidity and family responsibility. The par- group/class). All centers from the province of Burgos
ticipants ‘‘might have interpreted rigidity as a protective were classified into 12 levels, considering the educational
emotional bond related to more general parental engage- level of the center (a: pre-University; b: Obligatory Sec-
ment, e.g., awareness of their children’s activities, friends ondary; c: Intermediate Level Vocational Training; d:
and interests’’ (Everri et al., 2015, p. 3064). In the work of Basic Professional Training) and the location and type of
Everri et al. (2016), Rigidity was also positively associated center (a-urban public; b-rural public; c-urban private).
with family satisfaction. At each level, at least 5% of educational centers were
These works on adolescents, where consideration is selected and when a center refused to participate in the
given to adolescents’ points of view, raise a controversy study, it was replaced by another with similar character-
over the dysfunctionality of the variables Rigidity and istics. The centers with fewer than 100 students were not
Enmeshed, pointing to the need for a specific study of included in the sample because of their scant represent-
both variables among adolescents. ativeness. One group/class was selected for each center
The first objective of the present study is, therefore, and educational level, in which we had the collaboration
to test the validity and the reliability of the Family Com- of a teacher, and the voluntary participation of all stu-
munication Scale, the Family Satisfaction Scale, and the dents from the class was requested. The questionnaire
FACES IV scales among Spanish adolescents. was administered in person by the teachers with instruc-
A second objective is to test whether the variables tions from the researchers.
Rigidity and Enmeshed correlate negatively with fam- Convenience sampling was used at the university cent-
ily communication, family satisfaction, and the balanced ers. In this group, the questionnaire was completed on an
variables of FACES IV, in our sample of adolescents. individual basis outside the classroom.
All the juvenile centers in the province of Burgos were
visited for the sampling at non-educational centers and
Method participation was massive. At these centers, all the ado-
Participants lescents were invited to participate and the question-
A total of 1,187 adolescents (49.96% boys and 50.04% naires were individually administered with the assistance
girls), aged between 14 and 18  years old (M = 16.17; of an educator.
SD = 1.31), participated in the study. The sample had Informed consent from either the youth or their legal
the following distribution by age: 14  years old = 14.8%; guardians was obtained, ensuring the anonymity and the
15  years old  = 16.3%, 16  years old  = 22.8%; 17  years confidentiality of the data. The participation of all the
old = 28.8%; 18  years old = 17.3%. The adolescents were adolescents in the study was voluntary. All procedures
selected from 51 centers, grouped into three large clus- for this study were performed in accordance with the
ters: a) 23 educational centers within the province of Bur- ethical standards of the University of Burgos research
gos –Spain–, including Obligatory Secondary Education committee and with the Helsinki Declaration and its later
(35.5%), pre-university studies –Baccalaureate– (28.8%), amendments.
Intermediate Level Vocational Training (3.6%), and Basic
Professional Training (15.7%); b) 10 university degree Research instruments
courses at Burgos University –Spain– (6.9%), and c) 18 The adolescents were asked to complete a self-adminis-
juvenile facilities of Castile and Leon –Spain–, including tered questionnaire, composed of various socio-demo-
child protection centers and centers for adolescents with graphic questions and the FACES IV Package, which
family problems (4.1%), juvenile centers for youth with consisted of three questionnaires:
behavioral and drug dependency problems (1.7%), and
juvenile offenders (3.8%). a) the Spanish version of the FACES IV scale (Olson,
In all, 77.9% lived in a nuclear family, 9.9% lived only 2011), validated for adults (Martínez-Pampliega
with their mother, 2.8% lived only with their father, et al., 2017). This questionnaire has 42 items, 7 items
2.9% had shared parenting, and 5.1% belonged to a for each of their six variables: two balanced (Cohesion
Vegas et al. Psicologia: Reflexão e Crítica (2022) 35:18 Page 4 of 12

and Flexibility) and four unbalanced (Disengaged, the PCA (Hair et  al., 2010; Hattie, 1985; Lloret-Segura
Enmeshed, Rigid, and Chaotic). Each item could be et  al., 2014; Malhotra, 2010). A sufficiently high KMO
answered on a 5-point Likert-type scale: 1 (Strongly value (higher values than 0.7 ensure sufficient data ade-
Disagree) to 5 (Strongly Agree). quacy, higher than 0.8 a satisfactory fit, and below 0.5,
b) the Family Communication Scale (FCS) (Barnes an inadequate fit), and a p-value < 0.05 in Bartlett’s test of
& Olson, 1985), with 10 items and responses on a sphericity indicate the appropriateness of factor analysis.
5-point Likert-type scale: 1 (Strongly Disagree) to 5 Descriptive statistics for each scale were calculated:
(Strongly Agree). mean and standard deviation. The reliability analysis was
c) the Family Satisfaction Scale (FSS) (Olson, 2000), performed using Cronbach’s alpha coefficients.
with 10 items, with responses on a 5-point Likert- Finally, the inter-scale correlations were assessed with
type scale: 1 (Very Dissatisfied) to 5 (Very Satisfied). Pearson’s correlation coefficient and predictive valid-
ity through discriminant analysis, following the specific
Both the FCS and the FSS have been validated in Spain guidelines from Olson (2011).
for the adult population (Sanz et al., 2002). The Spanish For the statistical analyses, CFA was performed with
versions were used for our study. the statistical software package IBM SPSS AMOS 16.0
and AMOS 23, and PCA, reliability, descriptive statistics
Data analysis
and the Pearson correlations were calculated with IBM
First, a Confirmatory Factor Analysis (CFA) for all 42 SPSS Statistics 22 software.
items was performed, in order to validate the structure
of the FACES IV constructs and to test their convergent Results
Factor analysis of the FACES IV scales
validity. The following quality indices of the model were
considered (Chen et al., 2008; Ferrando & Anguiano-Car- A Confirmatory Factor Analysis (CFA) of the FACES IV
rasco, 2010; Hair et  al., 2010; Lloret-Segura et  al., 2014; model with the 42 items originally proposed by Olson
Malhotra, 2010; Montero & León, 2007; Xia & Yang, (2011) was performed (see Table  1). The CFA showed 7
2019): the minimum discrepancy or likelihood ratio Chi- items with factor loadings below 0.3, specifically items
square test divided by its degrees of freedom (CMIN/
df ), the Goodness of Fit Index (GFI), the Adjusted Good-
ness of Fit Index (AGFI), the Normed Fit Index (NFI), the Table 1 Confirmatory Factor Analysis (CFA) results for the 42
Incremental Fit Index (IFI), the Comparative Fit Index FACES IV items
(CFI), and the Root Mean Squared Error of Approxi- Factor loadings for each item Factor loadings for each item
mation (RMSEA). The Composite Reliability (CR), and
Item 1 Balanced Cohesion .549 Item 38 Balanced Flexibility .722
the Average Variance Extracted (AVE) were used to
Item 7 Balanced Cohesion .754 Item 32 Balanced Flexibility .381
complete the convergent validity. The values consid-
ered to guarantee the quality of goodness of fit indexes Item 13 Balanced Cohesion .720 Item 26 Balanced Flexibility .423
were: CMIN/df < 5; GFI > 0.95; AGFI > 0.80; NFI > 0.80; Item 19 Balanced Cohesion .561 Item 20 Balanced Flexibility .546
IFI > 0.85; CFI > 0.90; RMSEA < 0.05 (acceptable between Item 25 Balanced Cohesion .573 Item 14 Balanced Flexibility .602
0.05 and 0.08), and a p-value > 0.05. CR should be greater Item 31 Balanced Cohesion .655 Item 8 Balanced Flexibility .553
than 0.7 and AVE greater than 0.5. A minimum value of Item 37 Balanced Cohesion .630 Item 2 Balanced Flexibility .540
0.3 was required to maintain the factor loadings within Item 40 Enmeshed .323 Item 5 Rigid .625
each dimension (Hair et al., 2010). The t-student test was Item 34 Enmeshed .096a Item 11 Rigid .556
applied, to check that the factor loadings were signifi- Item 28 Enmeshed .723 Item 17 Rigid .644
cantly non-zero. Item 22 Enmeshed .147a Item 23 Rigid .525
As an additional tool to establish the structure and the Item 16 Enmeshed .519 Item 29 Rigid .211a
functioning of the 6 variables of FACES IV and both the Item 10 Enmeshed .050a Item 35 Rigid .614
FCS and the FSS scales, a Principal Components Analysis Item 4 Enmeshed .631 Item 41 Rigid .170a
(PCA) with the Varimax rotation method was separately Item 3 Disengaged .614 Item 42 Chaotic .697
performed for each variable, in order to determine the Item 9 Disengaged .688 Item 36 Chaotic .511
unidimensionality of each construct (Hattie, 1985). The Item 15 Disengaged .436 Item 30 Chaotic .113a
Kaiser–Meyer–Olkin (KMO) measure, the Bartlett’s test Item 21 Disengaged .638 Item 24 Chaotic .342
of sphericity (χ2), the Carmines and Zeller criterion (vari- Item 27 Disengaged .646 Item 18 Chaotic .674
ance explained by each factor must be at least 40%), and Item 33 Disengaged .334 Item 12 Chaotic .050a
the Kaiser criterion (eigenvalues should be greater than Item 39 Disengaged .392 Item 6 Chaotic .584
one) were used to determine the sampling adequacy of a
items with factor loadings below 0.3
Vegas et al. Psicologia: Reflexão e Crítica (2022) 35:18 Page 5 of 12

10, 22, 34, 29, 41, 12, and 30, corresponding to the scales Chi-squared test in Bartlett’s test of sphericity were sig-
Enmeshed, Rigid, and Chaotic. nificant, guaranteeing the adequacy of the data set for the
On the other hand, the CFA fit indices yielded no factor analysis that was performed (PCA). In the case of
acceptable fit, as the criteria of some of the indices were Cohesion, Flexibility, and Disengaged, the percentage of
not at an optimum level of quality (Hair et  al., 2010): variance explained by the single factor with an eigenvalue
GFI = 0.87 (yet should be > 0.95); NFI = 0.769 (yet should greater than 1 was over 40%. A result that confirmed the
be > 0.80); IFI = 0.798 (yet should be > 0.85); CFI = 0.811 unidimensionality of these three variables. However, two
(yet should be > 0.90); p-value of the RMSEA = 0.006 (yet eigenvalues greater than one were identified by the PCA
should be > 0.05). on the scales of Enmeshed, Chaotic, and Rigid, and the
As a complementary measure to establish the struc- variance explained by the first factor was less than 40%. In
ture and the functioning of the six variables of FACES addition, for Enmeshed and Chaotic, the quotients of the
IV in the Spanish sample of adolescents, we carried out differences between the first and second eigenvalues and
a Principal Component Analysis (PCA) on each of the between the second and third eigenvalues were less than
six theoretical constructs proposed by Olson, with the 3 in both cases (Carmines & Zeller, 1979). It may there-
aim of confirming the unidimensionality of each con- fore be affirmed that the constructs of both Enmeshed
struct. A global Exploratory Factor Analysis (EFA) on the and Chaotic as they are defined are not consistent. In
42 items was not conducted, in search of a new under- the case of the Rigid scale, it was not unidimensional,
lying structure of the data, because the six-factor struc- although it was very close to being so. Likewise, the items
ture has been confirmed by all the validation studies on with factor loadings below 0.3 in the CFA were almost
FACES IV over the past 15 years, both among the adult exactly the same items that were not saturating the first
and adolescent population, with either the original ver- PCA factor, in its respective scales: Enmeshed, Rigid, and
sion or the abridged versions of FACES IV (Baiocco et al., Chaotic.
2013; Costa et  al., 2013; Desautels et  al., 2016; Everri At this point, we decided to remove eight items from
et  al., 2020; Gomes et  al., 2017; Gouveia-Pereira et  al., the three scales (specifically, items 10, 22, 30, and 40
2020; Koutra et al., 2012; Martínez-Pampliega et al., 2017; from Enmeshed; 12 and 30 from Chaotic; 29 and 41
Pereira & Teixeira, 2013; Rivero et  al., 2010; Sebokova from Rigid), in order to improve the quality of the model.
et al., 2016). This decision was based on the following criteria: a) the
We therefore performed a PCA with Varimax rotation excessively low factor loadings of the items in the CFA
method on each of the six constructs (variables), using (all below 0.33 and six below 0.17); b) the incompliance
the Kaiser criterion (the eigenvalues-greater-than-1 rule). of unidimensionality in these three constructs (the eight
The results are presented in Table 2. items removed were not saturating in the first compo-
Previously, we had calculated the KMO values and the nent of their scale in the PCA, as shown in Table 2); c) the
Bartlett’s test of sphericity. All the KMO test values for lack of convergent validity (measured with both CR and
the six scales were greater than 0.67 and the results of the AVE); d) the deleted items were the ones that produced

Table 2  Principal components analysis with Varimax rotation on each of the FACES IV constructs
Balanced Cohesion Balanced Flexibility Disengaged Enmeshed Rigid Chaotic

Item 1 Item 2 Item 3 Item 4 Item 10 Item 5 Item 29 Item 6 Item 12

Item 7 Item 8 Item 9 Item 16 Item 22 Item 11 Item 41 Item 18 Item 24
Item 13 Item 14 Item 15 Item 28 Item 34 Item 17 Item 36 Item 30
Item 19 Item 20 Item 21 Item 40 Item 23 Item 42
Item 25 Item 26 Item 27 Item 35
Item 31 Item 32 Item 33
Item 37 Item 38 Item 39
Pa 49.8% 41.11% 40.28% 25.9% 22.5% 33.7% 17.4% 31.9% 21.0%
Rb 44.5 22.75 10.70 1.87 (< 3) 10.43 1.75 (< 3)
KMO test .885 .828 .840 .676 .767 .749
Bartlett´s test of χ2 = 2435.72 χ2 = 1377.4 χ2 = 1478.39 χ2 = 853.73 χ2 = 1262.3 χ2 = 1152.6
sphericity (χ2) sig. .000 sig. .000 sig. .000 sig. .000 sig. .000 sig. .000
a
P  = Percentage of variance explained by each AFE factor (eigenvalues-greater-than-one rule)
Rb = ratio between the difference of the first and second eigenvalues and the difference between the second and third eigenvalues
Vegas et al. Psicologia: Reflexão e Crítica (2022) 35:18 Page 6 of 12

the highest levels of reliability on their scale –measured to extremely close emotional ties or dependency on par-
with Cronbach’s alpha– when they were removed from ents or siblings.
the model. Furthermore, in the three aforementioned
scales, the modification indexes proposed a high number The new FACES IV scales: convergent validity
of correlations between errors, to improve the quality of and correlation analysis
the model. Thus, before dropping the items from each After removing the eight items, we performed a new CFA
construct, the errors were correlated according to the of the FACES IV, with the 34 remaining items. The results
modification indexes. However, the results of the model are shown in Fig. 1.
were of no better quality following this procedure, since Without the deleted items, all the new factor loadings
the main problem of the model was the inadequate satu- were significantly non-zero coefficients and higher than
ration of the items in its construct and not the possible 0.3.
correlation of its errors. After having removed the eight items specified above,
At a theoretical level, we proposed an explanation for the improved fit indices of the new quality model
the inadequate fit and the low saturation of these eight were sufficient (CMIN/df  = 4.142; p-value = 0.000;
items on their respective scales. On the Chaotic scale, GFI = 0.903; AGFI = 0.886; NFI = 0.835; IFI = 0.870;
there was ambiguity over the significance of items 12 and CFI = 0.869; RMSEA = 0.051; p-value of the
30. Item 12 (“It is hard to know who the leader is in our RMSEA =  0.143), such that the adapted FACES IV
family”) is not necessarily an indicator of chaos within instrument may be considered a good model. Three of
the family. Not knowing who the family leader is may the five values that never reached the optimum levels in
mean that there is no leader or that there are two lead- the first CFA were in the new CFA within the recognized
ers with the same level of power. The sentence of item 30 levels for a good adjustment of the model. One of the
(“There is no leadership in our family”) can imply chaos parameters that showed significant improvement was
or disengagement within the family, as there is very the RMSEA p-value (from 0.006 to 0.143), a key element
low involvement or interaction (Olson & Gorall, 2003, to test the improved goodness of fit, because a p-value
Appendix 19.1). lower than 0.05 invalidates the quality of the model
On the Rigid scale, item 29 (“Our family becomes (Chen et al., 2008).
frustrated when there is a change in our plans or rou- The reliability of the new constructs also increased.
tines”) should be changed for the adolescent population, All the Cronbach’s alpha coefficients were above 0.7 (see
because when filling in the questionnaire, many of the Table  3), except for the Enmeshed scale, whose coeffi-
participants were unaware of the meaning of “frustrate”. cient was not far from this value.
Item 41 (“Once a decision is made, it is very difficult to The CR and the AVE are also necessary to analyze
modify that decision”) has no clear meaning. Who takes the convergent validity (see Table  3). The CR coeffi-
the decision? Who does not permit it to be changed? cients were all over 0.7 except for the Enmeshed scale,
Does the decision refer to rules, roles, personal issues, and the AVE values were all higher than 0.3. The two
family issues, discipline consequences, etc.? scales, Rigidity and Chaotic, improved their CR by 7
On the Enmeshed scale, items 4, 16, and 28 explicitly and 12%, respectively and their AVE by more than 35%
refer to extreme emotional closeness or fusion, depend- in both cases. Notably, both the CR and the AVE of
ency, and maximization of time spent together. How- the Enmeshed scale also improved by 26 and by 110%,
ever, the other four items are in a negative key and in our respectively.
understanding, do not necessarily express Enmeshed, A correlation analysis was also performed, in order to
as it was defined by the authors (Olson & Gorall, 2003, evaluate the relations between the six scales (see Table 3).
Appendix  19.1). Item 22 (“Family members have little The strongest correlation was between the two bal-
need for friends outside the family”) might also indi- anced scales, Cohesion and Flexibility, which according
cate lack of interest or personal apathy. Under item 10 to Olson (2011) demonstrates concordance in the bal-
(“Family members feel pressured to spend most free time anced family function within the zone with the healthi-
together”), item 34 (“We resent family members doing est function levels. There was a high negative correlation
things outside the family”), and item 40 (“Family mem- between Cohesion and Disengaged (-0.6) and a moder-
bers feel guilty if they want to spend time away from the ate one between Flexibility and Chaotic (–0.36). How-
family”), reference is made to negative feelings, suppos- ever, the sample presented moderate positive correlations
edly associated with engagement, in other words, coer- between Cohesion and Flexibility, with both Enmeshed
cion in relation to item 10, jealousy under item 34, and and Rigid.
guiltiness under item 40. However, these feelings at the The results were unequal for each of the two princi-
stage of adolescence do not necessarily have to be linked pal dimensions with regard to the non-balanced scales.
Vegas et al. Psicologia: Reflexão e Crítica (2022) 35:18 Page 7 of 12

FACES5 e29
,64 ,22
FACES11 e30
,57 -,23
,63 FACES17 e31 -,33
e14 FACES4 ,61
,46 RIGID FACES23 e32
e12 FACES16 ENMESHED ,52

e10 FACES28 ,63 FACES35 e34

,79

,63 ,45
FACES38 e22

FACES32 e23 -,17

,19
e7 FACES37 ,73
,65 ,41 FACES26 e24
e6 FACES31 ,42
e5 FACES25 ,62
,52 FACES20 e25
,55 ,61
e4 FACES19 ,58 BALANCED COHESION -,42 -,14 BALANCED FLEXIBILITY FACES14 e26
,76 e27
e3 FACES13 FACES8
,73 ,56
e2 FACES7 FACES2 e28
,55 ,52
e1 FACES1

-,80 -,56
e21 FACES39 FACES42 e36
,20 ,40 ,72
e20 FACES33 FACES36 e37
,52
e19 FACES27
,34
,64
,63 ,33
e18 FACES21,43 DISENGAGED CHAOTIC FACES24 e39

e17 FACES15 FACES18 e40

,69 ,66
e16 FACES9 ,62 ,57
e15 FACES3 FACES6 e42

Fig. 1  Confirmatory Factor Analysis (CFA) for the new FACES IV scales

Table 3  Descriptive statistics (average scores), convergent validity, and correlation analysis for the new six FACES scales*
FACES IV scales Mean SD α CR AVE FACES IV Scales—Correlations
Balanced Balanced Disengaged Enmesheda Rigida
Cohesion Flexibility

Balanced Cohesion 3.632 .754 .830 .826 .408

Balanced Flexibility 3.493 .678 .730 .744 .301 .77**
Disengaged 2.463 .723 .747 .744 .305 –.60** –.49**
Enmesheda 2.675 .837 .655 .657 .400 .43** .42** –.30**
Rigida 2.827 .776 .713 .734 .357 .28** .37** –.07* .32**
Chaotica 2.219 .753 .697 .701 .332 –.32** –.36** .52** –.06 –.07*
(*) Cronbach’s alpha (α), CR, AVE for Convergent Validity. Pearson’s correlation matrix for Correlation Analysis. N = 1187
(.a) Adapted new scales, without deleted items
(*) p-value < .05
(**) p-value < .01

Enmeshed and Disengaged showed a weak negative cor- moderate positive correlation was also found between
relation, which makes sense, as if one increases, the other Rigid and Enmeshed (0.32).
should diminish; while no significant relation was found
between Rigid and Chaotic, showing the independence
of both variables and that a system could be chaotic and Factor analysis, validity, and reliability for the FCS
rigid at the same time. and the FSS scales
The important positive correlation (0.52) between Dis- A CFA was performed for each scale, FCS and FSS. The
engaged and Chaotic captures the attention, in so far as factor loadings of the items are shown in Table 4.
it might indicate that adolescents tend to perceive the The factor loadings were all over 0.48 in the FCS and
systems where no emotional links exist as chaotic. A over 0.55 in the FSS, and most of these factor loadings on
Vegas et al. Psicologia: Reflexão e Crítica (2022) 35:18 Page 8 of 12

Table 4  Factor loadings for FCS and FSS scales The Cronbach’s alpha coefficients for reliability were
FCS Factor loading FSS Factor loading
also excellent on both scales (0.894 for FCS and 0.913 for
FSS). To complete convergent validity, we calculated CR
FCS1 .713 FSS1 .735 and AVE (see Table 5), which were also within the desir-
FCS2 .693 FSS2 .725 able limits (CR greater than 0.8; AVE of 0.5 for FSS and
FCS3 .695 FSS3 .715 slightly lower for FCS).
FCS4 .719 FSS4 .742 All these results yielded very good psychometric prop-
FCS5 .685 FSS5 .727 erties for FCS and FSS scales, confirming the robustness
FCS6 .477 FSS6 .736 of their theoretical constructs.
FCS7 .587 FSS7 .549
FCS8 .721 FSS8 .725 Correlation analysis
FCS9 .698 FSS9 .666 The correlations between the FCS and FSS scales with the
FCS10 .694 FSS10 .687 other FACES scales are shown in Table 6.
The correlation between Family Communication and
Family Satisfaction was very high (0.79), meaning that
good communication in the family leads to wellbeing
both scales surpassed the threshold of 0.7 or were very within the family system. The relation between these
close to it. dimensions with the balanced variables was also highly
All values of the goodness of fit indices were within the positive (almost 0.7 in both cases) and followed the same
required limits (see Table  5). GFI, AGFI, NFI, IFI, and trend with the non-balanced variables as were found for
CFI were all over 0.95, which indicated an excellent fit of Balanced Cohesion and for Balanced Flexibility, which is
the model (Hair et al., 2010). to say, that they were negatively correlated with Disen-
As was done for the six variables of the FACES IV scale, gaged and with Chaotic, and positively correlated with
a PCA with Varimax rotation was performed for each of both Enmeshed and Rigid.
the FCS and the FSS scales, to test their unidimensional-
ity. The KMO test had previously been calculated –val- Discriminant analysis (Predictive Validity) for the FACES IV
ues of 0.918 for the FCS, and 0.933 for the FSS– and the scales
results of Bartlett’s test of sphericity (χ2) were statistically Following the methodology described by Olson (2011),
significant for both scales, which guaranteed satisfactory a discriminant analysis for the new six FACES IV scales
sampling adequacy and the validity of the component was performed, to determine the capacity of these scales
analysis. For each of the FCS and the FSS scales, there to distinguish between problematic and non-problem-
was only one factor with an eigenvalue greater than one atic families. There were no specific or external criteria
–values of 5.1 and 5.6, respectively– and the variance to classify problematic and non-problematic families, so
explained by the unique factor was over 40–51.8% for the these groups were defined on the basis of each person’s
FCS and 56.4% for the FSS–, which confirmed the unidi- score on two family assessment instruments, established
mensionality of both scales. as valid measures of family functioning, specifically the

Table 5  Descriptive statistics (total score), AVE, CR, and the main fit indices for the FCS and the FSS scales
Scale Mean SD CR AVE CMIN/df GFI AGFI NFI IFI CFI RMSEA p-value

FCS 34.68 7.92 .891 .452 4.605 .982 .959 .978 .983 .983 .055 .182
FSS 37.14 7.83 .907 .494 4.144 .981 .962 .982 .987 .987 .052 .372

Table 6  Correlation analysis of FSS/FCS and correlations of FACES IV with validation scales (FCS and FSS)
Scales FSS Cohesion Flexibility Disengaged Enmesheda Rigida Chaotica

FCS .788** .704** .684** –.571** .427** .199** –.368**

FSS – .696** .668** –.555** .464** .264** –.296**
(.a) Adapted new scales, without deleted items
(*) p-value < .05
(**) p-value < .01
Vegas et al. Psicologia: Reflexão e Crítica (2022) 35:18 Page 9 of 12

Family Satisfaction Scale (FSS) and the Family Commu- discriminant analyses with similar findings. Compared to
nication Scale (FCS). Discriminant analyses were carried the original American validation, the percentage of accu-
out for each of the six scales of FACES IV entered as indi- racy for our sample was lower for the Disengaged, the
vidual independent discriminators (or predictors) of the Chaotic, and the Cohesion scales, but higher for the Flex-
problematic/non-problematic groupings, as well as an ibility, the Enmeshed, and the Rigid Scales.
analysis where the six scales were entered together as the
independent variable. Discussion
We proceeded to define the groups, according to the This paper presents an important contribution to the
individual scores on both the FCS and the FSS: first, if the study of adolescent family functioning, as it is the first
score was in the top 50% of the FSS, the individual was validation of the FACES IV package for adolescents,
assigned to the non-problematic group or with healthier based on a large and heterogeneous sample, and using an
family functioning. Conversely, if the score was in the elaborate sampling strategy.
bottom 50% of the FSS, the individual was included in the The indicators of internal consistency, convergent
problematic group or with poorer family functioning. A validity, and predictive validity of the scales of the FACES
similar procedure was followed by using the individual IV package provided a good fit for the scales of Balanced
scores on the FCS. Analogous problematic and non- Cohesion, Balanced Flexibility, and Disengaged and an
problematic groups were created using both the upper excellent fit for Family Communication and Family Satis-
40% and the lower 40% of the FSS (or the FCS) scale as faction. However, the results for convergent validity were
their respective cut-off points. Table  7 illustrates the not good for Enmeshed, Rigid, and Chaotic, which in
results of the discriminant analysis, where the four col- addition never complied with the condition of unidimen-
umns represent the four different ways of grouping the sionality. These three scales also showed problems for the
families. validation of the FACES for Slovak adolescents (Sebok-
The grouping method that showed the highest percent- ova et al., 2016). In the Uruguayan validation, Enmeshed
age accuracy at discriminating between the problematic and Rigid showed no fit with a single factor model, which
and non-problematic groups was the upper 40% ver- could indicate that more than one concept is involved in
sus the lower 40% on the FSS validation scale. Correct these scales (Costa et al., 2013).
placement ranged from 58.5% (Rigid) to 80.7% (Balanced In our case, we decided to eliminate eight items,
Cohesion), reaching 83% when all six scales were consid- obtaining a reduced version of the FACES IV, but with
ered together. That means that using jointly the 6 FACES good psychometric properties. Some of the items
variables (and employing the FSS scale as the valid cri- removed in our study also showed insufficient saturation
terion), the scores can predict whether a family will in other European validations (Gouveia-Pereira et  al.,
or will not be problematic in 83% of cases. Both Olson 2020; Koutra et al., 2012; Martínez-Pampliega et al., 2017;
(2011) and Martínez-Pampliega et  al. (2017) reported Mirnics et  al., 2010) and were also almost all removed

Table 7  Discriminant analysis of functional/dysfunctional families (Percent Accuracy in Discriminating Groups) for FACES IV scales
Upper versus lower 50% Upper 40% versus lower 40% Upper versus lower 50% Upper 40% versus
on FSS on FSS on FCS lower 40% on FCS

N for each group Upper N = 575 Upper N = 505 Upper N = 599 Upper N = 471
Lower N = 603 Lower N = 474 Lower N = 579 Lower N = 513
Unbalanced Scales
 Disengaged 70.5% 74.8% 71.0% 73.9%
 ­Chaotica 62.1% 65.3% 61.0% 62.1%
 ­Enmesheda 65.5% 68.0% 68.6% 66.6%
 ­Rigida 57.6% 58.5% 60.1% 61.4%
Balanced Scales
  Balanced Cohesion 75.6% 80.7% 75.6% 79.4%
  Balanced Flexibility 72.4% 77.8% 74.7% 78.5%
  Six scales together 77.3% 82.9% 78.4% 82.6%
a
(. )Adapted new scales, without deleted items
GR 1 = Upper N = Non- problem Families
GR 2 = Lower N = Problem Families
Vegas et al. Psicologia: Reflexão e Crítica (2022) 35:18 Page 10 of 12

by Rivero et  al. (2010) in the abridged Spanish version. Everri et  al. (2020) also found positive correlations
In the Canadian study with adolescents (Desautels et al., between Enmeshed and Rigidity with Family Commu-
2016), saturation was very low in practically all the same nication, Cohesion, and Flexibility in his sample of Ital-
items as in our work, while the items from Enmeshed and ian adolescents. It therefore appears that Rigidity and
Chaotic showed acceptable or good saturation on their Enmeshment are not necessarily perceived as negative
respective scales when the authors conducted their vali- by the adolescents. It may even be precisely at this stage
dation among adults. These results indicate that adoles- of life where parental control and some strong emotional
cents understand some questions differently than adults. links are necessary for the proper development of the
The confirmatory factor analysis of our final shortened young person. Pereira and Teixeira (2013), Rivero et  al.
version produced an acceptable model for the six scales (2010), and Martínez-Pampliega et al. (2017) had already
theorized by Olson, with good reliability indices, except pointed in that direction, as they affirm that Rigid and
for the Enmeshed scale (0.66), a quite reasonable value, Enmeshed appear not to be as dysfunctional as Disen-
on the other hand, taking into account that the scale only gaged and Chaotic, and that the families with rules and
had three items. clear and strict consequences function better. Baiocco
As regards the correlation analysis, the signs of the cor- et  al. (2013) also found that the younger adolescents
relations found in our sample among the non-balanced scored higher for Enmeshed and Rigid and although the
scales of FACES IV mostly match the results from other authors related it with greater family dysfunctionality, it
validation studies (Baiocco et al., 2013; Costa et al., 2013; may have more to do with the necessary parental super-
Koutra et  al., 2012; Martínez-Pampliega et  al., 2017; vision at those ages. In turn, in the studies of Everri et al.,
Olson, 2011; Pereira & Teixeira, 2013). (2015, 2016), the adolescents linked Rigid to a protec-
The important positive correlation found between Dis- tive and emotional tie, related with greater interest and
engaged and Chaotic and the weak positive one between parental commitment and deduced from their results
Rigid and Enmeshed are in accordance with the findings that Rigid is not negative per se, but that it depends on
of Olson (2011) and other transcultural works. In the the positive or negative dimensions of family functioning
opinion of both Pereira and Teixeira (2013) and Mirnics with which it is associated.
et al. (2010), although the scales of Enmeshed and Rigid Olson (2008) had previously noted that the non-bal-
appear to be independent from other scales, they are anced variables were not necessarily always dysfunctional
inter-related. and that the time of the life cycle, along with other cul-
Likewise, the high correlations within our sample tural and religious aspects, should be also considered
between Family Satisfaction and Family Communication when interpreting these two variables. It is important to
with the two balanced scales, as well as the high nega- note that almost all the studies where the behavior of the
tive correlations of those four scales with Disengaged and variables Enmeshed and Rigidity was unexpected, have
Chaotic, are almost generalizable to all the validations been carried out in Mediterranean cultures (Portugal,
that have been performed and they confirm some of the Spain, Italy, and Greece). Enmeshed families and rigidity
hypotheses proposed by Olson. They demonstrate that are perhaps viewed within those family environments as
communication significantly improves family satisfaction part of the culture and are even considered desirable for
and that cohesion and flexibility lead to healthier family acceptable functioning of a family.
systems and with more satisfactory functioning. In turn, In conclusion, the results of our study corroborate the
disengaged and chaos are frequent in dysfunctional fam- factorial structure of the circumplex model proposed by
ily systems. Olson (2011) and demonstrate the transcultural applica-
However, unlike other studies, we found a negative cor- bility of FACES IV, FCS and FSS questionnaires. The ver-
relation between Enmeshed-Disengaged and moderate sion validated in this work (reduced for Enmeshed, Rigid,
positive correlations between the “healthy” dimensions and Chaotic) presents the FACES IV package as a valid
(Communication, Satisfaction, Cohesion, and Flexibility) and useful instrument for the assessment of family func-
with Rigid and Enmeshed, which calls into question the tioning of Spanish adolescents.
familiar dysfunctionality of these two unbalanced scales. Moreover, in the sample of Spanish adolescents, the
In the study by Sequeira et  al. (2021), Enmeshed and scale of Family Communication, the scale of Family
Rigid seemed to be positively connected to healthy family Satisfaction, and the scales of Balanced Cohesion, Bal-
functioning, probably due to “Portuguese cultural specifi- anced Flexibility, Disengaged and Chaotic of FACES
cities, namely its traditional values and manifest ideologi- IV function in the way that was anticipated through
cal familism, that emphasizes affective closeness, explicit the Circumplex Model, and that has been confirmed
solidarity norms, (…) marked respect for authority and in all the validations with the adult population. How-
hierarchies” (Sequeira et al., 2021, p. 1660). ever, the scales Enmeshed and Rigid work differently
Vegas et al. Psicologia: Reflexão e Crítica (2022) 35:18 Page 11 of 12

in adolescents than they do in adults, because both Author details

1
 Universidad de Burgos, C/ Alfonso XI, s/n, 09007 Burgos, Spain. 2 Day Hos‑
scales are not always perceived as dysfunctional by ado- pital, Psychiatry Service, Burgos Universitary Hospital, Avda. Islas Baleares 3,
lescents. On the contrary, these two variables are fre- 09006 Burgos, Spain. 3 Technical Professor of Community Services, Junta de
quently considered positive among adolescents and are Castilla y León, Burgos, Spain. 4 Social Services of Aranda de Duero (Burgos),
Aranda de Duero, Burgos, Spain.
associated with greater family emotional attachment,
flexibility, communication and satisfaction. Received: 4 November 2021 Accepted: 26 May 2022

Limitations of the study
The strengths of the present study are the size of the References
Baiocco, R., Cacioppo, M., Laghi, F., & Tafà, M. (2013). Factorial and construct
sample and its variability. However, the conclusions
validity of FACES IV among Italian adolescents. Journal of Child and Family
were based on the perceptions of adolescents, without Studies, 22(7), 962–970. https://​doi.​org/​10.​1007/​s10826-​012-​9658-1
any information provided by other family members, Barnes, H. L., & Olson, D. H. (1985). Parent-adolescent communication and the
circumplex model. Child Development, 56(2), 438–447. https://​doi.​org/​10.​
which might give it a certain bias.
2307/​11297​32
Additionally, as the six FACES scales have a differ- Bhugra, D., & Fiorillo, A. (2012). Families, functioning and therapies. Interna-
ent number of items, it is not possible to calculate the tional Review of Psychiatry, 24(2), 79–80. https://​doi.​org/​10.​3109/​09540​
261.​2012.​657915
ratio scores. It would therefore be recommendable to
Carmines, E.G., Zeller, R. A. (1979). Reliability and validity assessment. Sage publi‑
substitute the items that have been removed or to adapt cations. https://​doi.​org/​10.​4135/​97814​12985​642
them to the adolescent vocabulary and to the specific Chen, F., Curran, P. J., Bollen, K. A., Kirby, J., & Paxton, P. (2008). An empirical
evaluation of the use of fixed cutoff points in RMSEA test statistic in struc‑
requirements of this stage of the life cycle.
tural equation models. Sociological Methods & Research, 36(4), 462–494.
In future investigations, further studies should be https://​doi.​org/​10.​1177/​00491​24108​314720
performed on the Enmeshed scale, in order to improve Costa, D., González, M. L., Arca, D., Olson, D. H. (2013). Propiedades Psicomé‑
tricas del Faces IV: Estudio de Validez en Población Uruguaya. Ciencias
its consistency and reliability. Likewise, additional stud-
Psicológicas, VII(2), 119–132. http://​www.​scielo.​edu.​uy/​scielo.​php?​pid=​
ies will be necessary to confirm the trend observed for S1688-​42212​01300​02000​02&​script=​sci_​artte​xt&​tlng=​pt
Enmeshed and Rigid scales among adolescents. Costa-Ball, C., & Cracco, C. (2021). Psychometric properties of the family satis‑
faction scale to Uruguayan families. Anales De Psicología, 37(1), 161–167.
Acknowledgements https://​doi.​org/​10.​6018/​anale​sps.​383381
Not applicable. Cracco, C., Costa-Ball, C. D. (2019). Psychometric Properties of the Family
Communication Scale in College Students. Revista Iberoamericana de
Authors’ contributions Diagnóstico y Evaluación Psicológica, 2, 77–86. Retrieved from https://​
Study conceptualization and design were performed by MM, PP, CS and MV. www.​aidep.​org/​sites/​defau​lt/​files/​2019-​04/​RIDEP​51-​Art6.​pdf
Material preparation and data collection were performed by MV. Literature Desautels, J., Lapalme, M., Touchette, L., & Pauzé, R. (2016). Validation de la
Search and Data analysis, by MV. The first draft of the manuscript was written version française du FACES IV auprès d’une population d’adolescents
by MV; review and editing by MM, PP, and CS. All authors commented on francophones issus d’un échantillon populationnel et clinique. Therapie
previous versions of the manuscript. All authors read and approved the final Familiale, 37(1), 95–113. https://​doi.​org/​10.​3917/​tf.​161.​0095
manuscript. Everri, M., Mancini, T., & Fruggeri, L. (2015). Family functioning, parental
monitoring and adolescent familiar responsibility in middle and late ado‑
Funding lescence. Journal of Child and Family Studies, 24(10), 3058–3066. https://​
Not applicable. doi.​org/​10.​1007/​s10826-​014-​0109-z
Everri, M., Mancini, T., & Fruggeri, L. (2016). The role of rigidity in adaptive and
Availability of data and materials maladaptive families assessed by FACES IV: the points of view of adoles‑
The datasets generated and analysed during the current study are available cents. Journal of Child and Family Studies, 25(10), 2987–2997. https://​doi.​
in the Harvard Dataverse repository, https://​datav​erse.​harva​rd.​edu/​datas​et.​ org/​10.​1007/​s10826-​016-​0460-3
xhtml?​persi​stent​Id=​doi:​10.​7910/​DVN/​UA5GTO. Everri, M., Caricati, L., Mancini, T., Messena, M., & Fruggeri, L. (2020). Italian
validation of family adaptability and cohesion scale (FACES IV) short ver‑
sion for adolescents: SAD_FACES. Journal of Child and Family Studies, 29(9),
Declarations 2507–2514. https://​doi.​org/​10.​1007/​s10826-​020-​01771-9
Ferrando, P. J., & Anguiano-Carrasco, C. (2010). El análisis factorial como técnica
Ethics approval and consent to participate de investigación en psicología. Papeles Del Psicólogo, 3(1), 18–33.
All procedures performed in this study were in accordance with the ethical Gomes, H., Peixoto, F., & Gouveia-Pereira, M. (2017). Portuguese validation of
standards of the University of Burgos research committee and with the the family adaptability and cohesion evaluation scale–FACES IV. Journal
Helsinki Declaration and its later amendments. The study was approved by the of Family Studies, 25(4), 477–494. https://​doi.​org/​10.​1080/​13229​400.​2017.​
Bioethics Committee of University of Burgos (IR 16/2020). 13861​21
Consent to participate: Informed consent was obtained from individual par‑ Gouveia-Pereira, M., Gomes, H., Miranda, M., & Candeias, M. de J. (2020). Family
ticipants included in the study or from their legal guardians. cohesion and flexibility: Validation of the FACES IV package with portu‑
guese adolescents. Analise Psicologica, 38(1), 111–126. https://​doi.​org/​10.​
Consent for publication 14417/​ap.​1651
The authors consent the journal Psicologia: Reflexao e crítica for publishing Hair, J.F., Black, W.C., Babin, B.J., & Anderson, R.E. (2010). Multivariate data analy-
this article. sis (7th ed.). Prentice-Hall, Inc.
Hamilton, E., & Carr, A. (2016). Systematic review of self-report family assess‑
Competing interests ment measures. Family Process, 55(1), 16–30. https://​doi.​org/​10.​1111/​
The authors declare that they have no competing interests. famp.​12200
Vegas et al. Psicologia: Reflexão e Crítica (2022) 35:18 Page 12 of 12

Hattie, J. (1985). Methodology review: Assessing unidimensionality of tests and Sequeira, J., Vicente, H. T., Daniel, F., Cerveira, C., Silva, M. I., Neves, S., Espírito
ltenls. Applied Psychological Measurement, 9(2), 139–164. https://​doi.​org/​ Santo, H., Guadalupe, S. (2021). Family Adaptability and Cohesion Evalu‑
10.​1177/​01466​21685​00900​204 ation Scale–Version IV (FACES IV): Validation Study in the Portuguese
Jiménez, L., Lorence, B., Hidalgo, V., & Menéndez, S. (2017). Análisis factorial de Population. Journal of Child and Family Studies, 30 1650-1663. https://​doi.​
las escalas FACES (Family Adaptability and Cohesion Evaluation Scales) org/​10.​1007/​S10826-​021-​01941-3
con familias en situación de riesgo psicosocial. Universitas Psychologica, UN (2016). Resolution 32/23 adopted by the Human Rights Council “Protec‑
16(2), 1–12. https://​doi.​org/​10.​11144/​Javer​iana.​upsy16-​2.​afef tion of the family: role of the family in supporting the protection and
Koutra, K., Triliva, S., Roumeliotaki, T., Lionis, C., & Vgontzas, A. N. (2012). Cross- promotion of human rights of persons with disabilities.” https://​docum​
cultural adaptation and validation of the greek version of the family ents-​dds-​ny.​un.​org/​doc/​UNDOC/​GEN/​G16/​156/​45/​PDF/​G1615​645.​pdf?​
adaptability and cohesion evalution scale IV package (FACES IV Package). OpenE​lement
Journal of Family Issues, 34(12), 1647–1672. https://​doi.​org/​10.​1177/​01925​ White, J., & Klein, D. (2008). Family Theories ­(3rd ed.). Sage Publications.
13X12​462818 Xia, Y., & Yang, Y. (2019). RMSEA, CFI, and TLI in structural equation modeling
Lloret-Segura, S., Ferreres-Traver, A., Hernández-Baeza, A., & Tomás-Marco, I. with ordered categorical data: the story they tell depends on the estima‑
(2014). El análisis factorial exploratorio de los ítems: una guía práctica, tion methods. Behavior Research Methods, 51, 409–428. https://​doi.​org/​10.​
revisada y actualizada. Anales De Psicología, 30(3), 1151–1169. https://​doi.​ 3758/​s13428-​018-​1055-2
org/​10.​6018/​anale​sps.​30.3.​199361
Malhotra, N.K. (2010). Marketing research: an applied orientation (6th ed., global
ed.). Pearson. Publisher’s Note
Margasiński, A. (2015). The polish adaptation of FACES IV-SOR. Polish Journal of Springer Nature remains neutral with regard to jurisdictional claims in pub‑
Applied Psychology, 13(1), 43–66. https://​doi.​org/​10.​1515/​pjap-​2015-​0025 lished maps and institutional affiliations.
Martínez-Pampliega, A., Merino, L., Iriarte, L., & Olson, D. H. (2017). Propiedades
psicométricas de la versión española de la escala de evaluación de la
adaptabilidad y la cohesión familiar. Psicothema, 29(3), 414–420. https://​
doi.​org/​10.​7334/​psico​thema​2016.​21
Minuchin, S. (2009). Familias y terapia familiar. Gedisa.
Mirnics, Z., Vargha, A., Tóth, M., & Bagdy, E. (2010). Cross-cultural applicability of
FACES IV. Journal of Family Psychotherapy, 21(1), 17–33. https://​doi.​org/​10.​
1080/​08975​35100​36185​77
Montero, I., & León, O. G. (2007). A guide for naming research studies in
Psychology 1. International Journal of Clinical and Health Psychology, 7(3),
847–862.
Olson, D. H. (2008). Circumplex model of marital and family systems. Journal of
Family Therapy, 22(2), 144–167. https://​doi.​org/​10.​1111/​1467-​6427.​00144
Olson, D. H. (2011). FACES IV and the circumplex model: validation study. Jour-
nal of Marital and Family Therapy, 37(1), 64–80. https://​doi.​org/​10.​1111/j.​
1752-​0606.​2009.​00175.x
Olson, D. H., Sprenkle, D. H., & Russell, C. S. (1979). Circumplex model of marital
and family systems i cohesion and adaptability dimensions, family types
and clinical applications. Family Process., 18(1), 3–28. https://​doi.​org/​10.​
1111/j.​1545-​5300.​1979.​00003.x
Olson, D. H., Waldvogel, L., & Schlieff, M. (2019). Circumplex model of marital
and family systems: an update. Journal of Family Theory and Review, 11(2),
199–211. https://​doi.​org/​10.​1111/​jftr.​12331
Olson, D. H., Gorall, D. M. (2003). Circumplex Model of Marital and Family
Systems. In F. Walsh (Ed.), Normal Family Processes. ­(3rd ed.) (pp. 514–547).
Guilford. https://​doi.​org/​10.​4324/​97802​03428​436_​chapt​er_​19
Olson, D. H. (2000). Family Satisfaction Scale. Administration Manual. In FACES-
IV Package. Prepare/Enrich Insights. www.​faces​iv.​com.
Pereira, M. G., & Teixeira, R. (2013). Portuguese validation of FACES-IV in adult
children caregivers facing parental cancer. Contemporary Family Therapy,
35(3), 478–490. https://​doi.​org/​10.​1007/​s10591-​012-​9216-4
Rivero, N., Martínez-Pampliega, A., & Olson, D. H. (2010). Spanish adaptation
of the FACES IV questionnaire: psychometric characteristics. The Family
Journal: Counseling and Therapy for Couples and Families, 18(3), 288–296.
https://​doi.​org/​10.​1177/​10664​80710​372084
Samper, P., Cortés, M. T., Mestre, V., Nácher, M. J., & Tur, A. M. (2006). Adaptación
del child’s report of parent behavior inventory a población española.
Psicothema, 18(2), 263–271.
Sanderson, J., Kosutic, I., Garcia, M., Melendez, T., Donoghue, J., Perumbilly, S.,
Franzen, C., & Anderson, S. A. (2009). The measurement of outcome vari‑
ables in couple and family therapy research. American Journal of Family
Therapy, 37(3), 239–257. https://​doi.​org/​10.​1080/​01926​18080​24059​35
Sanz, M., Iraurgi, I., & Martínez-Pampliega, A. (2002). Evaluación del funcion‑
amiento familiar en toxicomanías: Adaptación española y características
de adecuación métrica del FAP-FACES IV. In I. Iraurgi & F. González (Eds.),
Instrumentos de Evaluación en drogodependencias (pp. 403–434). Aula
Médica.
Sebokova, G., Jurisova, E., Popelkova, M., Uhlarikova, J., & Zatkova, M. (2016).
Factorial validity and reliability of the Slovak version of the FACES IV in
adolescents. Ceskoslovenska Psychologie, 60(3), 278–289.