You are on page 1of 18

543851

research-article2014
JFIXXX10.1177/0192513X14543851Journal of Family IssuesArtemis and Touloumakos

Original Articles
Journal of Family Issues
1­–18
“They Accept Me, © The Author(s) 2014
Reprints and permissions:
They Accept Me sagepub.com/journalsPermissions.nav
DOI: 10.1177/0192513X14543851
Not”: Psychometric jfi.sagepub.com

Properties of the
Greek Version of
the Child Parental
Acceptance–Rejection
Questionnaire–Short
Form

Giotsa Artemis1* and Anna K. Touloumakos2*

Abstract
This article examines the psychometric properties of the Greek adaptation
of Child Parental Acceptance–Rejection Questionnaire (PARQ)–Short
Form, Mother, and Father versions. Data on the 24-item long measure
were collected from 534 preadolescents and adolescents of the municipality
of Ioannina. Participants were asked to report their agreement or
disagreement on a 4-point Likert-type scale, developed to tap in to four
factors (subscales), namely, (a) perceived warmth/affection, (b) perceived
hostility/aggression, (c) perceived indifference/neglect, and (d) perceived
undifferentiated rejection. A confirmatory factor analysis was conducted
first, to explore if the structure of the four-factor model of Child PARQ is
corroborated with the Greek sample. Additionally, the internal consistency

*Both the authors have contributed equally to this paper.


1University of Ioannina, Ioannina, Greece
2University of Oxford, Oxford, UK

Corresponding Author:
Giotsa Artemis, University of Ioannina, Panepistimioupolis Dourouti, Ioannina 45110, Greece.
Email: agiotsa@uoi.gr

Downloaded from jfi.sagepub.com at NORTHERN ARIZONA UNIVERSITY on June 8, 2015


2 Journal of Family Issues 

of the proposed four factors (subscales) and the global scale were examined
through obtaining Cronbach’s alphas and Gutman’s split-half coefficients.
The analyses provide preliminary, yet strong, evidence for the structural
invariance of the Greek adaptation of Child PARQ–Short Form.

Keywords
parental acceptance–rejection theory, Child Parental Acceptance–Rejection
Questionnaire–Short Form, PARQ, psychometric properties, confirmatory
factor analysis, CFA, internal consistency

Family systems theory construes family as a whole that is different from the
sum of its parts (Bronfenbrenner, 1979; Carr, 2006; Fishman, Andes, &
Knolton, 2001; Minuchin & Lappin, 2011; Wetchler, 2003)—namely, the
individual members that comprise it. Among the key factors that make a fam-
ily an integrated whole are the unique emotional bonds among the family
members (McGoldrick, Giordano, & Pearce, 1996). Emotional bonds as
expressed feelings by family members—with emphasis on parental love—
are critical for the development of emotionally healthy children (Bengtson,
2001; Giotsa, 2007; Mylonas, Gari, Giotsa, Pavlopoulos, & Panagiotopoulou,
2006; Rohner, 1986; Rohner, Khaleque, & Cournoyer, 2012). In particular,
whether children perceive themselves as accepted or rejected by their parents
or other caregivers strongly predicts children’s psychological adjustment—
be it smooth or problematic (Carrasco & Rohner, 2013; Khaleque & Rohner,
2012; Rohner, 1975; Rohner & Khaleque, 2005).
Parental acceptance–rejection theory (PARTheory; Rohner, 1986, 2004;
Rohner et al., 2012; Rohner & Rohner, 1980) has emerged from this strand of
research investigating thoroughly the parent–child relations and their effect on
child development and psychological adjustment. “PARTheory is an evidence-
based theory of socialization and lifespan development that aims to predict
and explain major causes, consequences, and other correlates of interpersonal
acceptance and rejection worldwide” (Rohner & Khaleque, 2005, p. 4).
According to PARTheory, the psychological adjustment of children around the
world is likely to be affected depending whether they perceive themselves to
be accepted or rejected by their parents. Collectively, parental acceptance and
rejection form the warmth dimension of parenting, which captures the quality
of emotional bonds between family members. In particular, in the spectrum of
warmth, parental acceptance lies at one extreme, and refers to the expression
of positive feelings, love, and support toward the children; at the other extreme,
rejection incorporates indifference, absence of positive feelings, judgment,

Downloaded from jfi.sagepub.com at NORTHERN ARIZONA UNIVERSITY on June 8, 2015


Artemis and Touloumakos 3

and psychological or corporal aggression toward children. Cross-cultural


research findings during the past 50 years show that the same behaviors world-
wide signify parental love (Rohner, 2004; Rohner et al., 2012).
PARTheory has received extensive investigation through quantitative
methods such as self-report questionnaires, behavior observations, and meta-
analyses as well as through qualitative ethnographic research (Rohner et al.,
2012). The Parental Acceptance–Rejection Questionnaire (PARQ) tests
quantitatively important aspects of this theory. The tool has been now used in
approximately 500 studies internationally, has been translated into 35 lan-
guages to date, and measures individuals’ perceptions along the warmth
dimension of parenting (acceptance–rejection). Mother and Father versions
of the PARQ measure are available for adults (Adult PARQ), for children
(Child PARQ: Mother and Father), and for parents (Parent PARQ). Each of
these is also available in two forms: long (standard 60-item) and short
(24-item). The Adult PARQ measures how adults remember their parents’
accepting or rejecting behaviors in childhood; the Child PARQ measures
children’s perceptions of their parents’ current accepting and rejecting behav-
iors; the Parent PARQ measures the way parents reflect on their current
behavior toward their children (Rohner et al., 2012; Rohner & Khaleque,
2005).
All three instruments consist of the same four scales: (a) warmth/affection,
(b) hostility/aggression, (c) indifference/neglect, (d) undifferentiated rejec-
tion. Τhe warmth/affection scale measures participants’ perceptions of their
parents’ physical, verbal, and symbolic expressions of love, affection, caring,
and nurturance. Reverse scored, the warmth/affection scale is a measure of
parental coldness and lack of affection, which are expressions of parental
rejection. The hostility/aggression scale refers to parent–child relationships
where parents are perceived to express physical or verbal aggression, hostil-
ity, criticism, anger, and disapproval toward their children. The indifference/
neglect scale assesses participants’ perceptions about their parents’ lack of
interest in them and about their parents’ failure to pay attention to their needs
and wishes. Undifferentiated rejection refers to relationships where parents
are perceived to be rejecting although not clearly aggressive, unaffectionate,
or indifferent.
Unsurprisingly, to date the research on and with Child PARQ has been
rampant, establishing the theory underpinning the instrument as well as the
instrument itself. The advantages of such research, however, expand beyond
the theoretical level. In particular, research providing evidence for the reli-
abilities and validities of Child PARQ contributes substantially toward con-
ducting sound and valid research addressing a long-standing contemplation
of children about their parents: “They accept me, they accept me not.” In turn

Downloaded from jfi.sagepub.com at NORTHERN ARIZONA UNIVERSITY on June 8, 2015


4 Journal of Family Issues 

parental acceptance forms an important research area in the broader study of


family processes and the relations underpinning or existing alongside these
processes. Reviewing the literature on such evidence, therefore, is key to
identifying the psychometric soundness of Child PARQ, potential discrepan-
cies, and actions necessary to undertake to ensure its soundness. Gleaning
through the international literature, first we present evidence on the reliabili-
ties of Child PARQ, followed by evidence on its validity.
Rohner himself conducted the first study on the reliabilities of Child
PARQ Mother version standard form back in 1975 in the United States. This
study reported αs between .72 and .90 (Warmth/Affection: .90, Hostility/
Aggression: .87, Indifference/Neglect: .77, and Undifferentiated Rejection:
.72). In a later study, Rohner (1995) provided Cronbach’s α coefficients for
both Mother and Father versions of the global Child PARQ: .76 for the
Mother version and .77 for the Father version for the global scale. Alphas for
the four subscales in Mother and Father versions ranged between .78 and .94
and .83 and .94, respectively. In a similar vein, analysis of Child PARQ in a
study in Turkey (Erkman & Rohner, 2006) yielded global reliabilities equal
to .78 and .83 for mothers and fathers, respectively. Last, results from meta-
analytic studies provide further evidence for the measure’s internal consis-
tency reliabilities. In particular, in the sample of the three most recent studies
conducted in the United States and internationally on Mother versions, the
reported range of αs for the four subscales (Rohner & Khaleque, 2005) were
.75 to .91 (Bluestone & Tamis-LeMonda, 1999), .74 to .93 (Comunian &
Gielen, 1999), and .63 to .89 (Demetrious & Christodoulides, 2000). Taking
the 51 studies examined by the authors collectively, the mean weighted α
coefficient for the global Child PARQ Scale was .89.
Turning to the construct validity of the Child PARQ, evidence is available
only from studies exclusively using exploratory factor analysis (EFA). Note,
however, that recent developments on the Adult PARQ have moved toward
the use of the confirmatory factor analysis (CFA) to explore the construct
validity of the measure (see Gomez & Rohner, 2011). With respect to the
Child PARQ Mother version, the 1975 study by Rohner yielded a two-factor
solution accounting collectively for 58% of the variance (Rohner & Khaleque,
2005). The first factor had high loadings from the items designed to measure
each of the Hostility/Aggression, Indifference/Neglect, and Undifferentiated
Rejection, therefore being labeled as the “Rejection” factor; the second factor
had high loadings from items designed to tap into Warmth/Affection, thereby
labeled as the “Acceptance” factor. These results on the proposed higher
order structure of the Child PARQ were corroborated internationally, through
replicating studies in Egypt (Salama, 1990), India (Rohner & Chaki-Sircar,
1988), Italy (Comunian & Gielen, 2001), and Turkey (Varan, 2003).

Downloaded from jfi.sagepub.com at NORTHERN ARIZONA UNIVERSITY on June 8, 2015


Artemis and Touloumakos 5

Despite the abundance of studies on the psychometrics (reliability and


validity) of Child PARQ internationally this evidence is restricted in the use
of EFA and stops with Greece. In particular, evidence about the psychometric
characteristics of the Child PARQ indicates that (a) in Greece there is only
one other study researching the psychometric properties of Child PARQ
(Tsaousis, Giovazolias, & Mascha, 2010) not currently published, (b) the
focus of prior research on the factorial validity of the Child PARQ is exclu-
sively on exploratory factor analytic approaches and the higher order struc-
ture of the tool, and (c) in the great majority of cases conclusions deal with
the Mother version (standard form) of the measure. Seeking to address these
gaps, and provide answer to the question about the structural invariance of
Child PARQ Greek version, this article presents findings from a CFA of the
proposed four-factor structure of both the mother and father Child PARQ
(Short Form), and the corresponding reliability coefficients. In investigating
the construct validity of the Child PARQ, CFA is deemed the most appropri-
ate method (compared with EFA) because it tested the hypothesis about the
four-factor structure fitting the data—and, with it, the structural invariance of
Child PARQ Greek version (Fabrigar, Wegener, MacCallum, & Strahan,
1999; Gomez & Rohner, 2011). Addressing this hypothesis, in turn, was
deemed rather important before moving forward the ongoing research about
family relations in Greece.

Method
Participants
A sample of 534 Greek preadolescents and adolescents from the municipality
of Ioannina was drawn to assess the structure of the Child PARQ. Ioannina is
the eighth largest city in Greece (Hellenic Statistical Authority; http://www.
statistics.gr/portal/page/portal/ESYE) and as such embodies characteristics
of both large and smaller cities of Greece, and provides a representative
example of Greece as whole. This became a key criterion for selecting
Ioannina as the place from which the sample of participants was drawn.
Participants’ age was between 9 and 18 years. The mean age was 12.22
years (SD = 2.49). Gender was balanced, with 276 males (51.7%). Mean ages
of males and females were approximately equal (for males M = 12.11, SD =
2.48; and for females M = 12.34, SD = 2.52). Four hundred forty-four partici-
pants (83.1%) were originally from Greece; the remaining 90 (16.9%) were
of other ethnic origins such as Albania, Romania, and other countries.
Reflecting the levels of immigration, economic or political, in Greece in the
past decades, these 90 participants were construed and treated as integrated

Downloaded from jfi.sagepub.com at NORTHERN ARIZONA UNIVERSITY on June 8, 2015


6 Journal of Family Issues 

members of the Greek society and Greek schools. They typically were born
in their country of origin, moved to Greece when their families migrated, and
became integrated members of the Greek society. These students are either
bilingual, or native speakers (depending on the dominant language in their
family); in any case they are fluent users of the Greek language. Among this
cohort of participants, 85% of parents were married and 15% were divorced.

Instrumentation
The study used the short form of the Parental Acceptance–Rejection
Questionnaire: Child version (Child PARQ: Mother version, Child PARQ:
Father version; Rohner & Khaleque, 2005). The Child PARQ short version
encompasses 24 items tapping in to four dimensions/factors: Warmth/
Affection, Hostility/Aggression, Indifference/Neglect, and Undifferentiated
Rejection. The 24 items are statements about how participants feel their par-
ents treat them. Participants are asked to evaluate each statement on a Likert-
type scale from 1 (almost never true) to 4 (almost always true). The scale is
summed and keyed in the direction of perceived rejection. Scores from each
scale are added together to produce the overall test score, which indicates the
level of perceived acceptance–rejection from the parent.
The Warmth/Affection Scale is composed of eight items including state-
ments such as my father/mother “says nice things about me,” “makes it easy
for me to tell her things that are important to me,” “is really interested in what
I do,” “makes me feel wanted and needed,” and “makes me feel what I do is
important.” The Hostility/Aggression Scale is composed of six statements,
including statements such as my father/mother “hits me, even when I do not
deserve it,” “punishes me severely when she is angry,” and “says many
unkind things to me.” The Indifference/Neglect Scale has six items, including
statements such as “pays no attention to me,” “is too busy to answer my ques-
tions,” and “pays no attention when I ask for help.” Last, the Undifferentiated
Rejection Scale incorporates four statements such as my father/mother “sees
me as a big nuisance,” “seems to dislike me,” and “makes me feel unloved if
I misbehave.”

Translation and Back-Translation


The Greek version of the Child PARQ (short version) was obtained through
translation, back-translation, and pretesting (piloting). Translation of the
Child PARQ was completed by two Greek teachers of English. Two bilingual
family psychologists then back-translated the two versions (Father and
Mother). Finally, a pilot study (N = 20) using the Greek version was carried

Downloaded from jfi.sagepub.com at NORTHERN ARIZONA UNIVERSITY on June 8, 2015


Artemis and Touloumakos 7

out to determine how items and rating scales worked with preadolescents and
to identify potential practical problems in following the research procedure
(Lindquist, 1991; Van Teijlingen, Rennie, Hundley, & Graham, 2001).
Administration was carried out in exactly the same way as in the main study
and participants were asked to give us feedback in order to identify ambigui-
ties and difficulties in items. Time taken to complete the measure and reac-
tions about the process and the measure were recorded too. Following
feedback from this exercise, refinements and amendments were made in
rephrasing items that were ambiguous, leading to the final Greek versions of
the measures. For example, Item 5 (sees me as a big nuisance) and Item 23
(pays no attention to me as long as I do nothing to bother her) had to be
rephrased and simplified because they seemed to be challenging for the
younger participants (9-11 years old).

Procedure
The research was conducted through the help of the school parents’ associa-
tions. Moreover, the research was conducted following the Code of Ethics of
the University of Ioannina. First, we informed parents about the aim of the
study and we asked them to complete an informed consent form concerning
their children’s participation in the study. After getting consent from parents,
permission to enter classrooms was further sought by the school headmaster.
The next step was then to ask the children participating in the study to com-
plete an informed consent form containing details about the aim, the proce-
dure, the data protection issues, the right to withdraw from the research at any
time, the person responsible for the research, and the contact details of the
principal investigator. Data collection went through group administration at
the end of class and lasted approximately 20 minutes. Participants were asked
to complete first the demographic form and then the two versions of the Child
PARQ (Father and Mother). Questionnaires were administered between
December 2011 and May 2012.

Results
CFA was pursued using AMOS Version 18 (Arbuckle, 2005) in order to
determine the extent to which the four-factor model proposed by Rohner and
Khaleque (2005) was confirmed with the present Greek sample. Initial checks
for univariate and multivariate normality and kurtosis (DeCarlo, 1997) sug-
gested that, as expected, the data violated the assumptions of normality. In
particular Kolomogorov–Smirnov (Massey, 1951) and Shapiro–Wilk
(Shapiro & Wilk, 1965) tests on each individual item of the two forms were

Downloaded from jfi.sagepub.com at NORTHERN ARIZONA UNIVERSITY on June 8, 2015


8 Journal of Family Issues 

statistically significant with no exception (p = .000). Furthermore Mardia’s


coefficient for the Mother version was 244.948 (critical ratio [c.r.] 80.114 >
1.96) and for the Father version was 249.341 (c.r. 81.550 > 1.96), indicating
severe deviations from multivariate normality. In particular, our data were
positively skewed, suggesting that the reported levels of parental rejection
among our participants were very low. These results were unsurprising when
considering previous research in Greece that corroborate them (Giotsa &
Touloumakos, 2011; Giotsa & Zergiotis, 2010); they were surprising, how-
ever, when compared with other international research on Child PARQ
(Rohner, 1975).
As a consequence a reciprocal (1/x) transformation was applied to the
data. The results from Mardia’s test of normality, however—while correcting
considerably for the problems with normality—did not lead to an acceptable
c.r. for any of the Mother and Father versions (c.r. > 1.96). Given these
results, initial raw data were used, but the normal theory estimation method
of maximum likelihood (ML) was deemed inappropriate for use with this
data set. Instead, the use of an alternative estimation, namely, the unweighted
least squares (ULS), was selected and used for the assessment of the pro-
posed hypothesis, the model fit, albeit at the expense of not reporting a χ2 and
its associated significance value.
The preferred estimation method and use of the statistical software AMOS
combined led to the decision of reporting four indices: three were absolute
(i.e., not using an alternative model as a basis for comparison), namely, the γ
or goodness-of-fit index (GFI), the adjusted goodness-of-fit index (AGFI),
and the root mean square residual (SRMR); and one was relative (i.e., using
the null model as a basis for comparison), namely, the normed fit index (NFI).

CFA of Child PARQ–Short Form for Mother and Father Versions


The Child PARQ model proposed by Rohner and Khaleque (2005) consists
of 24 items clustering around four latent factors: Warmth/Affection, Hostility/
Aggression, Indifference/Neglect, and Undifferentiated Rejection. Based on
the data from the present sample there was evidence of model fit as shown in
Figures 1 and 2. (For guidelines on cutoff scores for model fit see Hu &
Bentler, 1999; Schumacker & Lomax, 2004.)
In particular, for the Mother version results included: GFI = .97, AGFI =
.96, NFI = .95, and SRMR = .056. Similarly, for the Father version goodness-
of-fit indices results were GFI = .97, AGFI = .97, NFI = .96 and SRMR =
.057. For GFI, AGFI, and NFI, typically values over .90 indicate adequate
model fit whereas values over .95 indicate good model fit (Hu & Bentler,
1999). All values for both versions were over .95. Similarly, SRMR values

Downloaded from jfi.sagepub.com at NORTHERN ARIZONA UNIVERSITY on June 8, 2015


Artemis and Touloumakos 9

Figure 1.  Confirmatory factor analysis of Child Parental Acceptance–Rejection


Questionnaire Mother version: Four-factor structure.

below .08 are indicative of adequate model fit whereas SRMR values below
.05 indicate really good fit. SRMR values achieved in this study for both the
PARQ Mother and PARQ Father revealed good fit. Furthermore, three alter-
native models were juxtaposed to this primary one: (a) a single-factor model
corresponding to the “warmth” dimension in Child PARQ; (b) a two-factor

Downloaded from jfi.sagepub.com at NORTHERN ARIZONA UNIVERSITY on June 8, 2015


10 Journal of Family Issues 

Figure 2.  Confirmatory factor analysis of Child Parental Acceptance–Rejection


Questionnaire Father version: Four-factor structure.

model corresponding to the two ends of the warmth spectrum, namely,


“acceptance” and “rejection,” suggested in PARTheory; and (c) a higher
order model, with the four lower order factors loading on a single higher
order factor (see Gomez & Rohner, 2011). Table 1 presents the results from

Downloaded from jfi.sagepub.com at NORTHERN ARIZONA UNIVERSITY on June 8, 2015


Artemis and Touloumakos 11

Table 1.  Fit Indices for Three Alternative Models: One-Factor Model, Two-Factor
Model, and Higher Order Single-Factor Model.

Mother version Father version

Model GFI AGFI NFI SRMR GFI AGFI NFI SRMR


One-factor .95 .95 .92 .07 .95 .94 .93 .08
Two-factor .96 .96 .94 .06 .96 .96 .95 .07
Higher order single-factor .97 .96 .94 .06 .97 .96 .96 .06

Note. GFI =goodness-of-fit index; AGFI = adjusted goodness-of-fit index; NFI = normed fit
index; SRMR = standardized root mean square residual.

these analyses for Mother and Father versions separately. Based on this evi-
dence, it is suggested that the four-factor model fits the data slightly, yet bet-
ter than the one- and three-factor models we compared it against (although in
the case of the higher order model this supremacy of the four-factor model is
very minimal). Furthermore, it seems that evidence from the third model cor-
roborates the higher order structure of Child PARQ suggested by Rohner
(Rohner et al., 2012; Rohner & Khaleque, 2005).

Intercorrelations Among Factors


Moving away from the fit indices, and checking the standardized coefficients
reported in the initial four-factor analysis, the picture becomes clearer. In
particular, the scale “Undifferentiated Rejection” especially in the Mother
version seems to be alarmingly highly correlated with the factor “Hostility/
Aggression” (see Figure 1), thus implying that the relationship between items
and factor are worth investigating further. Otherwise, the correlations between
factors in both Mother and Father versions are strong. Correlations between
all four latent variables with the items that comprise them for Mother version
(see Figure 1) range between .43 through .7, with the exception of Item 15,
which correlates .35 with the “Indifference Neglect” factor. Focusing on the
Father version (see Figure 2), all correlations are even better, ranging between
.46 and .77. Comparing the results yielded from the CFAs in Mother and
Father versions, it can be argued that results are relatively similar, with the
lowest correlation between factors between “Warmth/Affection” and
“Hostility/Aggression” in both cases and with the highest correlations
between “Undifferentiated Rejection” and “Hostility/Aggression” in both
cases; furthermore, “Undifferentiated Rejection” reports high correlations
with all other factors, justifying even further the need to explore this factor.

Downloaded from jfi.sagepub.com at NORTHERN ARIZONA UNIVERSITY on June 8, 2015


12 Journal of Family Issues 

Table 2.  Descriptive Statistics and Reliability Coefficients for the Child PARQ–
Short Form, Mother and Father Versions.

Subscales M SD α Split-half
Mother version
 Warmth/Affection 11.67 3.69 .83 .81
 Hostility/Aggression 8.63 2.85 .72 .69
 Indifference/Neglect 9.85 2.98 .65 .69
  Undifferentiated Rejection 5.70 1.93 .53 .56
  Total PARQ score 35.86 9.06 .88 .86
Father version
 Warmth/Affection 13.02 4.60 .86 .87
 Hostility/Aggression 8.50 2.99 .76 .76
 Indifference/Neglect 11.06 3.50 .73 .74
  Undifferentiated Rejection 5.71 2.07 .61 .60
  Total PARQ score 38.29 10.75 .91 .91

Note. PARQ = Parental Acceptance–Rejection Questionnaire.

Reliabilities of Child PARQ–Short Form for Mother and Father


Two key measures were used to provide evidence for the internal consistency
and coherence of each of the factors of the Child PARQ: Mother and Father–
Short Form versions: (a) Cronbach’s alpha and (b) Guttman’s split-half. As
shown in Table 2, Cronbach’s α coefficient for the global Mother version and
Father version was excellent (.88 and .91, respectively). When seen indepen-
dently “Warmth/Affection” and “Hostility/Aggression” scales in the Mother
version provided Cronbach’s α well above .7, whereas the scales
“Undifferentiated Rejection” and “Indifference/Neglect” failed to meet this
threshold. In the Father version, Cronbach’s α was above the .7 threshold in
all scales but the “Undifferentiated Rejection.” Guttman’s split-half coeffi-
cients were similar for both Mother and Father versions. In particular, for the
Mother version Guttman’s split-half was excellent for the global scale, just
below .7 (.69) for the scales “Hostility/Aggression” and “Indifference/
Neglect” and lower for “Undifferentiated Rejection”; for the Father version
Guttman coefficients for the global scale was excellent and only
“Undifferentiated Rejection” provided a coefficient below .7. Note here that
the lower coefficients (Cronbach’s alpha and Guttman split-half) reported in
both Mother and Father versions are those referring to Factor 4
(Undifferentiated Rejection). Also note that the reported correlation between
this and other factors as yielded in the CFA presented before raised some

Downloaded from jfi.sagepub.com at NORTHERN ARIZONA UNIVERSITY on June 8, 2015


Artemis and Touloumakos 13

questions about this factor too. Table 2 also presents other descriptive statis-
tics for subscales and the global Child PARQ.

Discussion
The article sought to examine the psychometric properties of the Child
PARQ–Short Form, Mother and Father versions with a sample of 534 chil-
dren (ages 8-19) from the municipality of Ioannina, Greece. The central aim
was to address the question about the measure’s structural invariance (i.e., its
construct validity with a Greek sample), and ultimately enhance the research
on family and family relations. In doing so, a CFA was conducted, as opposed
to the widely and internationally used EFA, to assess the structure of the
measure. Three important findings resulted through this endeavor:

1. Data were skewed, thus indicating the low levels of parental rejection
reported by the children from our Greek sample compared with inter-
national evidence (for details, see Rohner, 1975).
2. Findings from CFA provide preliminary, yet adequate evidence for
the structural validity of the Child PARQ–Short Form Father and
Mother versions with the Greek sample as proposed by Rohner and
Khaleque (2005) and Rohner et al. (2012). Furthermore, these results
along with evidence from the research of other colleagues in Greece
(Tsaousis et al., 2010) and abroad (Comunian & Gielen, 2001; Ripoll-
Nunez, Comunian, & Brown, 2012; Rohner & Chaki-Sircar, 1988;
Varan, 2003) provide support for the argument about the structural
invariance of that Child PARQ with ethnically and culturally diverse
populations.
3. Findings on reliabilities are relatively consistent with the literature on
all four factors (for details, see Rohner & Khaleque, 2005). However,
reliabilities for the “Undifferentiated Rejection” in both Mother and
Father versions were lower in this study than reported in international
research, but were considered and reported as acceptable on the basis
that the factor comprised four items only and alpha is a function of the
number of items comprising a scale (see Cortina, 1993; Green, Lissitz,
& Mulaik, 1977; Nunnally & Bernstein, 1994).

Considering these points together allows us the confidence that it is worth


the time and effort to investigate further the Child PARQ–Short Form Greek
version. However, limitations of this work should be considered and addressed
first. Limitations identified in this study are in three areas: (a) the factor
“Undifferentiated Rejection,” which alarmed us not only in terms of

Downloaded from jfi.sagepub.com at NORTHERN ARIZONA UNIVERSITY on June 8, 2015


14 Journal of Family Issues 

factor-to-factor correlations but also in terms of the low internal consistencies


it reported; (b) issues of generalizability, considering that the sample came
exclusively from Ioannina; and (c) issues relating to the analysis. Remedies
for these limitations, in turn, lie in the next steps we propose to take to move
this research forward.
First, a good place to start would be the revision of the factor
“Undifferentiated Rejection.” Although international research on Child
PARQ indicates that the “Undifferentiated Rejection” scale has the lowest
alpha coefficient among all factors (see, e.g., Anjel, 1993; Demetrious &
Christodoulides, 2000; Myers et al., 1992), it is only in very few of the 51
studies included in Rohner and Khaleque’s (2005) meta-analysis that the fac-
tor falls below the threshold of .7. Moreover, the factor does not seem to be
problematic in the EFAs conducted in the aforementioned studies, or in the
CFA used with the Adult version of PARQ (Gomez & Rohner, 2011).
Accordingly, items need to be reviewed, refined, and amended as needed
alongside potentially other items from this tool where there are high correla-
tions, with a clear aim to achieve a fair translation yet ensuring the distinc-
tiveness and cohesiveness of such factors.
Second action to be taken would be to ensure the funds to turn this study
in to a standardization of the Child PARQ–Short Form in Greece. This would
involve the development of a sophisticated design and the considerable
expansion of the sampling through the use of advanced sampling procedures,
potentially with different samples for different age ranges (e.g., distinguish-
ing between adolescents and preadolescents). As part of this process it would
be worth investigating whether the levels of perceived rejection by Greek
kids is consistent throughout Greece, therefore corroborating this research,
and, furthermore, identifying potential variations (that could lead to a differ-
ent study altogether). Such a step would enhance considerably the generaliz-
ability of the results from this study as well claims made on the basis of
results from this analysis.
Moreover, the analyses conducted with the existing data set would benefit
from the use of additional statistical software (e.g., R or Lisrel)—which pro-
vide alternative robust methods of dealing with nonnormal data—allowing a
deeper understanding of the factors and items they encompass. In addition,
the current data set could be used to explore the psychometric properties of
Child PARQ: convergent/discriminant validity, criterion validity, and,
through retesting a subsample of the participants, test–retest reliability.
Although these are important steps for moving this work forward and
would effectively address the limitations to this study, evidence presented
here gives clear indications that the Mother and Father versions of the Greek
Child PARQ–Short Form offer valuable tools for conducting research in the

Downloaded from jfi.sagepub.com at NORTHERN ARIZONA UNIVERSITY on June 8, 2015


Artemis and Touloumakos 15

area of child and family psychology. Parenting and its effect in child develop-
ment and psychosocial adjustment is undoubtedly one of the main pillars in
family studies research. Ongoing research conducted in this area, therefore,
progresses alongside the development and adaptation of psychometrically
sound instruments. Contributing to this research, therefore, is key to promot-
ing our understanding of family and family processes, can offer cross-cultural
comparisons, and can inform family psychology as practice -and the pro-
grams developed to support this practice. Against this background, this study
is an important first step toward contributing to the development of this
strand of research in Greece.

Acknowledgment
We would like to thank dearly Professor R. P. Rohner for his invaluable comments
and suggestions in writing this article.

Declaration of Conflicting Interests


The author(s) declared no potential conflicts of interest with respect to the research,
authorship, and/or publication of this article.

Funding
The author(s) received no financial support for the research, authorship, and/or publi-
cation of this article.

References
Anjel, M. (1993). The transliteral equivalence, reliability, and validity studies of
Parental Acceptance-Rejection Questionnaire Mother Form: A tool for assess-
ing child abuse (Unpublished master’s thesis). Istanbul, Turkey: Bogazici
University.
Arbuckle, J. L. (2005). AMOS 6.0. User’s guide. Spring House, PA: Amos
Development.
Bengtson, V. L. (2001). Beyond the nuclear family: The increasing importance of
multigenerational bonds. Journal of Marriage and Family, 63, 1-16.
Bluestone, C., & Tamis-LeMonda, C. S. (1999). Correlates of parenting styles in
predominantly working and middle class African American mothers. Journal of
Marriage and the Family, 61, 881-893.
Bronfenbrenner, U. (1979). The ecology of human development experiments.
Cambridge, MA: Harvard University Press.
Carr, A. (2006). Family therapy. concepts, process and practice. West Sussex,
England: John Wiley.
Carrasco, M., & Rohner, R. (2013). Parental acceptance and children’s psychological
adjustment in the context of power and prestige. Journal of Child and Family
Studies, 22, 1130-1137.

Downloaded from jfi.sagepub.com at NORTHERN ARIZONA UNIVERSITY on June 8, 2015


16 Journal of Family Issues 

Comunian, A. L., & Gielen, U. P. (1999). Parental acceptance and rejection in Italian
culture. Unpublished manuscript.
Comunian, A. L., & Gielen, U. P. (2001). An Italian study of parental acceptance-
rejection. In R. Roth & S. Neil (Eds.), A matter of life: Psychological theory,
research, and practice (pp. 166-174). Lengerich, Germany: Pabst Science.
Cortina, J. M. (1993). What is coefficient alpha? An examination of theory and appli-
cations. Journal of Applied Psychology, 73, 98-104.
DeCarlo, L. T. (1997). On the meaning and use of kurtosis. Psychological Methods,
2, 292-307.
Demetrious, L., & Christodoulides, P. (2000). Parental acceptance-rejection in the
Cypriot family: A social-psychological research on the Child PARQ. Unpublished
manuscript, Frederick Institute of Technology, Limassol, Cyprus, Greece.
Erkman, F., & Rohner, R. P. (2006). Youths’ perceptions of corporal punishment,
parental acceptance, and psychological adjustment in a Turkish metropolis.
Cross-Cultural Research, 40, 250-267.
Fabrigar, L. R., Wegener, D. T., MacCallum, R. C., & Strahan, E. J. (1999). Evaluating
the use of exploratory factor analysis in psychological research. Psychological
Methods, 4, 272-299. doi:10.1037/1082-989X.4.3.272
Fishman, H. C., Andes, F., & Knolton, R. (2001). Enhancing family therapy: The
addition of a community resource specialist. Journal of Marital & Family
Therapy, 27, 111-116.
Giotsa, A. (2007). Values and the family. A cross-cultural research. Psychology,
13(4), 111-128.
Giotsa, A., & Touloumakos, A. K. (2011, December 8-11). Children’s and ado-
lescents’ representations of parental behaviour. Paper presented at the Eighth
PanHellenic Conference of Social Psychology, Stylida, Greece.
Giotsa, A., & Zergiotis, A. (2010, July). Father love in Greece: Children’s perceptions
of their parents’ acceptance-rejection. Paper presented at the Third International
Congress on Interpersonal Acceptance and Rejection, Padua, Italy.
Gomez, R., & Rohner, R. P. (2011). Tests of factor structure and invariance in the
USA and Australia using the Adult version of the Parental Acceptance-Rejection
Questionnaire. Cross-Cultural Research, 45, 267-285.
Green, S. B., Lissitz, R. W., & Mulaik, S. A. (1977). Limitations of coefficient alpha as
an index of test unidimensionality. Educational and Psychological Measurement,
37, 827-836.
Hu, L., & Bentler, P. M. (1999). Cut off criteria for fit indexes in covariance struc-
ture analysis: Conventional criteria versus new alternatives. Structural Equation
Modeling, 6, 1-55.
Khaleque, A., & Rohner, R. (2012). Pancultural associations between perceived
parental acceptance and psychological adjustment of children and adults: A meta-
analytic review of worldwide research. Journal of Cross-Cultural Psychology,
43, 784-800.
Lindquist, R. (1991). Don’t forget the pilot work! Heart & Lung, 20, 91-92.
Massey, F. J. (1951).The Kolmogorov-Smirnov tests for goodness of fit. Journal of
American Statistical Association, 46, 68-78.

Downloaded from jfi.sagepub.com at NORTHERN ARIZONA UNIVERSITY on June 8, 2015


Artemis and Touloumakos 17

McGoldrick, M., Giordano, J., & Pearce, J. K. (1996). Ethnicity and family therapy
(2nd ed.). New York, NY: Guilford Press.
Minuchin, S., & Lappin, J. (2011). Salvador Minuchin on family therapy [DVD].
Retrieved from http://www.psychotherapy.net
Myers, H. F., Alvy, K. T., Arrington, A., Richardson, M. A., Marigna, M., Huff,
R., . . .Newcomb, M. D. (1992). The impact of a parent training program on
inner-city African-American families. Journal of Community Psychology, 20,
132-147.
Mylonas, K., Gari, Α., Giotsa, A., Pavlopoulos, V., & Panagiotopoulou, P. (2006).
Greece. In J. Georgas, J. W. Berry, F. van de Vivjer, C. Κagitçibasi, & Y. H.
Poortinga (Eds.), Families across cultures: A 30 nation psychological study (pp.
344-352). New York, NY: Cambridge University Press.
Nunnally, J. C., & Bernstein, I. H. (1994). Psychometric theory (3rd ed.). New York,
NY: McGraw-Hill.
Ripoll-Nunez, K. J., Comunian, A. L., & Brown, C. M. (2012). Expanding horizons.
Current research on interpersonal acceptance. Boca Raton, FL: Brown Walker
Press.
Rohner, R. P. (1975). They love me, they love me not: A worldwide study of the effects
of parental acceptance and rejection. New Haven, CT: HRAF Press.
Rohner, R. P. (1986). The warmth dimension: Foundations of parental acceptance-
rejection theory. Beverly Hills, CA: Sage.
Rohner, R. P. (1995). McPAC Project. Unpublished raw data. Retrieved from http://
csiar.uconn.edu/bibliography/
Rohner, R. P. (2004). Extended parental acceptance-rejection bibliography. Retrieved
from http://csiar.uconn.edu/bibliography/
Rohner, R. P., & Chaki-Sircar, M. (1988). Women and children in a Bengali village.
Hanover, NH: University Press of New England.
Rohner, R. P., & Khaleque, A. (2005). Handbook for the study of parental acceptance
and rejection. Storrs, CT: Rohner Research.
Rohner, R. P., Khaleque, A., & Cournoyer, D. E. (2012). Introduction to parental
acceptance-rejection theory. Retrieved from http://csiar.uconn.edu/bibliography/
Rohner, R. P., & Rohner, E. C. (1980). Parental acceptance-rejection and parental
control: Cross-cultural codes. Ethnology, 20, 245-260.
Salama, M. (1990). Perceived parental rejection and cognitive distortions: Risk fac-
tors for depression. Egyptian Journal of Mental Health, 30, 1-17.
Schumacker, R. E., & Lomax, R. G. (2004). A beginner’s guide to path analysis,
structural equations and causal inference. Cambridge, England: Cambridge
University Press.
Shapiro, S. S., & Wilk, M. B. (1965). An analysis of variance test for normality (com-
plete samples). Biometrika, 52, 591-611. doi:10.1093/biomet/52.3-4.591
Tsaousis, I., Giovazolias, T., & Mascha, K. (2010, July). Translation and psychometric
properties of the Child Parental Acceptance-Rejection Questionnaire (PARQ)-
Short Form in Greek. Paper presented at the Third International Congress on
Interpersonal Acceptance and Rejection, Padua, Italy.

Downloaded from jfi.sagepub.com at NORTHERN ARIZONA UNIVERSITY on June 8, 2015


18 Journal of Family Issues 

Van Teijlingen, E., Rennie, A. M., Hundley, V., & Graham, W. (2001). The impor-
tance of conducting and reporting pilot studies: The example of the Scottish
Births Survey. Journal of Advanced Nursing, 34, 289-295.
Varan, A. (2003). Reliability and validity of the Turkish Child PARQ/C (Mother and
Father forms) and the Turkish Child PAQ. Retrieved from http://www.azmi-
varan.com/arastirma/ekar-kurami-arastirmalari/
Wetchler, J. (2003). The history of marital and family therapy. In L. Hecker & J.
Wetchler (Eds.), An introduction to marital and family therapy (pp. 3-38). New
York, NY: Haworth.

Downloaded from jfi.sagepub.com at NORTHERN ARIZONA UNIVERSITY on June 8, 2015

You might also like