Article

Evaluation Review
2015, Vol. 39(4) 428-446
ª The Author(s) 2015
Validity of the Reprints and permission:
sagepub.com/journalsPermissions.nav
Parenting Stress DOI: 10.1177/0193841X15600859
erx.sagepub.com

Index Short Form

in a Sample of
At-Risk Mothers

Javier Pérez-Padilla1, Susana Menéndez1,

and Oscar Lozano2

Abstract
Background: The Parenting Stress Index Short Form (PSI-SF) is a widely
used instrument in scientific literature to evaluate the levels of stress a par-
ent feels when facing parenting-related tasks. Despite the potential useful-
ness of the PSI-SF with at-risk families, no validation studies have been
carried out on this population in Spain. Objectives: The main objective
of this study is to report evidences of the reliability and validity of PSI-SF
with a sample of at-risk mothers. Specifically, (1) to examine the discrimina-
tive capacity of PSI-SF to differentiate between a community sample and
another sample composed of families with various levels of risk and (2)
to analyze the relationships with general health indicators and parental
sense of competence. Results: Analyses reported in this article show

1
Department of Developmental and Educational Psychology, University of Huelva, Huelva,
Spain
2
Department of Clinical, Experimental and Social Psychology, University of Huelva, Huelva,
Spain

Corresponding Author:
Javier Pérez-Padilla, Departamento de Psicologı́a Evolutiva y de la Educación, University of
Huelva, Campus del Carmen. Avda. Tres de Marzo s/n., Huelva, 21071, Spain.
Email: [email protected]

Downloaded from erx.sagepub.com at FLORIDA INTERNATIONAL UNIV on November 14, 2015

Pérez-Padilla et al. 429

satisfactory results regarding appropriate internal consistency coefficients

for the scale. With respect to the evidences of validity, results reported
here suggest that the total PSI-SF score, but not the two subscales, could
be useful to differentiate between different groups of mothers with different
levels of risk. The Childrearing stress subscale was associated with a poorer
perception as a mother as well as with an external locus of control, and the
Personal distress subscale was related to all the general health indicators,
explaining a high percentage of variance. Conclusion: The results reported
show that the total score or the subscales should be used in a differentiated
way according to the professional’s objectives. Hence, PSI-SF may be a use-
ful instrument for researchers and practitioners who work with at-risk
families.

Keywords
PSI-SF, validity, reliability, assessment, at-risk families

Parental stress is a complex process in which adults feel overwhelmed by

the tasks and responsibilities associated with their role as parents (Deater-
Deckard, 1998). The stress associated with this developmental task has
proven to be a relevant dimension in studies that examine the dynamics
taking place within family context. According to Abidin (1992), stress
may work as a motivational arousal and moderate levels are desirable, but
negative consequences are related to high parenting stress. Thus, clinical
stress levels have been associated with lower indicators for parent–child
relationship values (Crnic & Low, 2002) with a higher frequency of dys-
functional parenting practices (Bonds, Gondoli, Sturge-Apple, & Salem,
2002) and abuse (Stith et al., 2009) as well as with child adjustment
(Costa, Weems, Pellerin, & Dalton, 2006; Haapsamoa et al., 2013; Öst-
berg & Hagekull, 2013). With regard to the parents, parenting stress have
been associated with the quality of life of mothers (Cho & Hong, 2013)
and negative symptoms in the general health of the parents, particularly
those linked to anxiety and depression (Gerdes et al., 2007; Lyon,
2000; Ponnet & Wouters, 2014; Ponnet et al., 2013). Therefore, it is
important to appropriately detect parenting stress levels to prevent their
negative effects.
Social stressors have been studied through three perspectives (Sandı́n,
2003): life events, daily hassles, and role-related stress. The third one has
been discussed by Abidin (1992) in his theory regarding the determinants
of parenting. According to the Transactional Model of Stress proposed by

Downloaded from erx.sagepub.com at FLORIDA INTERNATIONAL UNIV on November 14, 2015

430 Evaluation Review 39(4)

Lazarus and Folkman (1984), Abidin points out that the cognitive appraisals
associated with the parenting role influence the degree of stress experi-
enced. Thus, the sense of competence and the perception of control as par-
ent have been commonly examined in studies about stress (Roddenberry &
Renk, 2010; Sevigny & Loutzenhiser, 2009).
Based on this theoretical model, Abidin designed the Parenting Stress
Index (PSI, 1983, 1997), a scale that assesses the negative feelings and
stress related to child-rearing. The extended version of this instrument
includes 120 items and, using two subscales, provides information about
the stress associated with parenthood (parent domain) and the stress
derived from the child (child domain). Based on the results obtained with
a sample of 530 mothers (Caucasian, married, and with problem-free chil-
dren around 4 years of age), Abidin (1995) developed an abbreviated ver-
sion of this scale, the Parenting Stress Index–Short Form (PSI-SF). It
includes 36 items, but the results of the factorial analyses reported by this
author did not replicate the originals, and PSI-SF yields three subscales:
Parental distress, Parent–child dysfunctional interaction, and Difficult
child. Psychometric properties of this instrument have been analyzed with
several populations obtaining satisfactory reliability coefficients. How-
ever, different factorial solutions have been reported, ranging from two
to five subscales with acceptable to excellent reliability (Deater-
Deckard & Scarr, 1996; Haskett, Ahern, Ward, & Allaire, 2006; McKel-
vey et al., 2009; Reitman, Currier, & Stickle, 2002; Whiteside-Mansell
et al., 2007; Zaidman-Zait et al., 2010). A number of studies have also
examined evidences of validity for this tool. Thus, Haskett, Ahern, Ward,
and Allaire (2006) showed that PSI-SF allows differentiating between par-
ents with a documented history of abuse and those without a known his-
tory of maltreatment and that the scores are related to general health
dimensions. According to Costa, Weems, Pellerin, and Dalton (2006),
Parental distress subscale shows high sensibility to depressive and anxiety
symptoms without collinearity problems. Thus, parenting stress is a robust
predictor of general health (Ponnet et al., 2013). Additionally, PSI-SF has
proved to be a useful assessment tool to determine the effectiveness of
psychoeducational programs (Bloomfield & Kendall, 2012; Marcynys-
zyn, Maher, & Corwin, 2011; Reitman et al., 2002).
To our knowledge, there is just one version of the PSI-SF in Spanish, and
it is developed by Dı́az-Herrero, Brito, López-Pina, Pérez-López, and Mar-
tı́nez-Fuentes (2010) in Spain with 129 mothers. Similar to Abidin’s initial
proposal (1983), Dı́az-Herrero and colleagues (2010) report two factors
related to the stress as a parent (Personal distress) and with respect to the

Downloaded from erx.sagepub.com at FLORIDA INTERNATIONAL UNIV on November 14, 2015

Pérez-Padilla et al. 431

child (Childrearing stress), with good levels of internal consistency (a ¼

.90 and .87, respectively).
Parents from families at psychosocial risk are a specific population that
usually experience high levels of parental stress (Anderson, 2008; Raikes
& Thompson, 2005). Several studies focusing in this topic indicate that
many parents of at-risk families tend to experience clinical levels of par-
enting stress (Bloomfield & Kendall, 2012; Hurley et al., 2012; Lanier,
Kohl, Benz, Swinger, & Drake, 2014). For a broad range of reasons, the
parents have difficulties to adequately provide for their children and there-
fore hinder their well-being but not severely enough to require the chil-
dren’s placement in foster care (Rodrigo, Byrne, & Álvarez, 2012;
Sanders & Cann, 2002). According to transnational recommendations
(i.e., Committee of Ministers of the Council of Europe, 2006), at-risk fam-
ilies should be attended with preventive and supporting interventions
designed to preserve the family unit. In Spain, to date, these interventions
are delivered in state agencies by interdisciplinary professional teams
(psychologists, social workers, and educators), and throughout a wide
range of diversified services. At-risk families cover approximately 80%
of the cases attended by Child and Family Protection Services (CFPS).
Therefore, correctly identifying and evaluating parental stress is a key
topic to better understand the process that characterizes the families and
also to design and implement effective preservation and support interven-
tions. Hence, reliable and cost- and time-efficient instruments to assess
parental stress should be available to researchers and practitioners who
work with at-risk families (Harnett & Dawe, 2008; Hutchings &
Webster-Stratton, 2004; Rodrı́guez, Camacho, Rodrigo, Martı́n, & Mái-
quez, 2006). Despite the potential usefulness of the PSI-SF, there are no
validation studies for this tool available with Spanish families at psycho-
social risk. In this regard, the recommendations of the American Educa-
tional Research Association, the American Psychological Association,
and the National Council on Measurement in Education (AERA, APA,
& NCME, 2002) concerning the need to collect empiric evidences using
assessment instruments with noncommunity populations should be high-
lighted. According to Ramayah, Yeap, and Ignatius (2014), the reliability
and validity of standardized scales must be studied through repeated appli-
cation of an instrument in diverse contexts and among different population
groups (e. g., Conrad, Riley, Conrad, Chan, & Dennis, 2010).
The main objective is to evaluate and report evidences about the reliabil-
ity and validity of the PSI-SF in a sample of at-risk mothers. Specifically, a
series of psychometric analysis will be computed:

Downloaded from erx.sagepub.com at FLORIDA INTERNATIONAL UNIV on November 14, 2015

432 Evaluation Review 39(4)

1. to analyze the properties of the items and the reliability coefficient

of the PSI-SF and
2. to report evidences of validity. Specifically:

2.1. To explore the utility of PSI-SF to predict community,

moderate-risk, or high-risk status.
2.2. To analyze the relationships between PSI-SF scores and data
about other individual indicators: general health and parental
sense of competence.

Method
Participants
The target population were at-risk mothers receiving supporting and
strengthening interventions at the CFPS. The sample framework
included 9 of the 11 social work areas in the province of Huelva
(Spain), which covers 62.34% of the population. The at-risk sample was
composed of 109 mothers, which represents 16.03% of the recipients of
CFPS. Inclusion criteria were the following: (1) to be formally sup-
ported by the aforementioned agencies for family preservation and (2)
to have at least one child below 12 years. At the time of data collection,
the families had been receiving some support intervention over a period
of approximately 3 years (M ¼ 2.85, SD ¼ 3.55). The mean age of the
women were around 35 years (M ¼ 35.35, SD ¼ 7.25). Only 36.1%
were employed, and their educational level was quite low: 42.6% were
illiterate, 21.8% completed primary school, and only 31.7% and 4% had
initiated or finished high school or university studies, respectively. The
families had two or three children (M ¼ 2.41, SD ¼ 1.12) with an aver-
age age of 8 (M ¼ 8.07, SD ¼ 3.33). The weighted of monthly family
incomes per consumption unit and their contrast with population and
official data in Spain (Observatorio de la Infancia en Andalucı́a,
2013) showed that most (56%) of the families lived below the poverty
threshold.
A sample of community families was contacted in 10 schools from the
same areas where the at-risk families resided. Mothers from the commu-
nity sample were not supported by agencies for family preservation. This
sample was composed of 40 mothers with an average age of 40.15 (SD ¼
6.52), with most of them being employed (75.70%) at the time of the
study. The families had two or three children (M ¼ 2.44, SD ¼ 1.10), with
an average age of 11 (M ¼ 11.60, SD ¼ 1.16).

Downloaded from erx.sagepub.com at FLORIDA INTERNATIONAL UNIV on November 14, 2015

Pérez-Padilla et al. 433

Instruments
PSI-SF. This instrument assesses the feelings of stress a person experiences
regarding his or her role as a parent (Abidin, 1995). It is composed of
36 items with a Likert-type answer format of five options. According to the
original author, total stress scores of 90 or above may indicate a clinical
level of stress. The Spanish adaptation of this scale (Dı́az-Herrero, Brito,
López-Pina, Pérez-López, & Martı́nez-Fuentes, 2010) revealed a bifactorial
structure: stress generally associated with parenthood (Personal distress,
12 items; e.g., I feel that I cannot handle things) and specifically to
child-rearing (Childrearing stress, 24 items; e.g., My child doesn’t giggle
or laugh much when playing). The levels of internal consistency in this
study were a ¼ .79 for Personal distress and a ¼ .85 for Childrearing stress.

Parental sense of competence (PSOC). This instrument explores the perception

the parent has of his or her abilities regarding parental role (Johnston &
Mash, 1989). It is a self-report tool with 16 items rated on a 6-point
Likert-type scale. Two scores are computed: Effectiveness as a parent
(a ¼ .73, 7 items; e.g., I honestly believe I have all the skills necessary
to be a good mother to my child) and Satisfaction with the parental role
(a ¼ .50, 9 items; e.g., Being a parent makes me tense and anxious).

Parental Locus of Control. Parental locus of control (PLOC) is a 47-item (a ¼

.71) scale to assess whether a parent views his or her child’s behavior as a
direct consequence of their parenting efforts (internal locus of control) or as
outside the reach of his or her parenting efforts (external locus of control;
e.g., When something goes wrong between me and my child, there is little
I can do to correct it; Campis, Lyman, & Prentice-Dunn, 1986). Each item
is rated on a 5-point Likert-type scale. The highest scores indicate a more
external PLOC.

General Health Questionnaire. This tool is a self-report instrument designed to

assess general state of health by the presence of different symptomatology
(Goldberg & Williams, 1996). General Health Questionnaire 28 (GHQ-28)
is composed of 28 items rated on a 5-point scale scoring four subscales:
Somatic symptoms (a ¼ .94, 7 items; e.g., Have you recently been feeling
perfectly well and in good health?), Anxiety and insomnia (a ¼ .92, 7 items;
e.g., Have you recently lost much sleep over worry?), Social dysfunction
(a ¼ .73, 7 items; e.g., Have you recently been able to enjoy your normal
day-to-day activities?), and Depression (a ¼ .91, 7 items; e.g., Have you

Downloaded from erx.sagepub.com at FLORIDA INTERNATIONAL UNIV on November 14, 2015

434 Evaluation Review 39(4)

recently felt that life isn’t worth living?). The highest scores indicate a high
frequency of symptoms and, hence, a worst level of general health.

Family risk and case history indicators. Practitioners completed a semistruc-

tured protocol designed ad hoc to collect data about the trajectory and the
current services received from CFPS (number of years and number of dif-
ferent concrete services received per family), and the level of family risk
characterizing the global situation (by means of a single score ranging from
0 to 10, the higher the score, the higher the level of risk).

Procedure
A series of meetings were conducted with the psychologists of CFPS to
describe the objectives of the research, the target sample of the study, and
the cooperation required from each professional. These practitioners were
requested to (a) select from the group of parents with whom they were
working a sample of mothers from declared at-risk families (the children are
at risk for being removed for their homes), (b) arrange an appointment of a
trained researcher with each mother to complete the aforementioned tools,
and (c) complete the semistructured protocol described earlier. The mothers
signed an informed consent form, and confidentiality was guaranteed.
Directors of the 10 schools were requested to send an evaluation protocol
to the students’ parents. It included an informative document (describing
the objectives of the project and the confidentiality rules) and the PSI-SF.
All this information was collected anonymously and in a sealed envelope.
The questionnaires were collected by school administration and delivered
to the research team. At each school, families were asked to participate
in the study voluntarily, although only families having a child aged 12 years
or below and who did not have an active CFPS file were considered as part
of the community sample for this study.

Data Analyses
The skewness and kurtosis of Personal distress and Childrearing stress
scores were calculated. Item-total correlations were computed, and test
reliability was established by Cronbach’s a for internal consistency. To pro-
vide evidence of validity, the at-risk sample was subdivided into two groups
with respect to the level of family risk reported by practitioners. These two
groups (moderate and high risk) were established considering +1 standard
deviation of the mean of family risk as limiting criteria. Analysis of

Downloaded from erx.sagepub.com at FLORIDA INTERNATIONAL UNIV on November 14, 2015

Pérez-Padilla et al. 435

variance and post hoc contrasts (minimum significant difference test

[DMS]) were computed to analyze the differences between the groups, tak-
ing into account the estimated effect size (Cohen, 1988). A multinomial
logistic regression analysis was performed with each PSI-SF subscale to
explore the differences between groups (Tabachnick & Fidell, 2007).
The explanatory capacity of the PSI-SF subscales with regard to the gen-
eral health and the PSOC was explored using multiple regression analysis.
For each of the aforementioned analysis, the parametric test assumptions
were validated. Analyses were conducted by using SPSS 18 software.

Results
Estimating Test Reliability
The item-total correlations for each of 36 items were calculated (see Table
1). The analysis showed that most of these items had adequate item-total
correlations (>.20), except Items 11 (Personal distress), 14, 22, and 32
(Childrearing stress), which decreased the overall a coefficient. PSI-SF was
highly internally consistent (a ¼ .89), and each subscale had an internal
consistency of a ¼ .79 for Personal distress and a ¼ .85 for Childrearing
stress.

Evidence of Validity
PSI-SF scores according to the level of family risk. According to the practitioners’
point of view, at-risk families were characterized by a global level of risk of
around 5 in a 0–10 scale (M ¼ 5.31, SD ¼ 2.08, Zskewness ¼ 0.17, Zkurtosis
¼ 1.05). The at-risk sample was divided into two groups considering +1
standard deviation of the mean: moderate risk (M ¼ 3.66, SD ¼ 1.16) and
high risk (M ¼ 7.13, SD ¼ 0.82).
Total stress scores were significantly different among the three groups of
families (community, moderate risk, and high risk; see Table 2). The com-
munity sample obtained the lowest mean and the high-risk group the higher
total score. Moderate- and high-risk groups showed clinical levels of par-
enting stress. Post hoc analysis indicated that the PSI-SF total stress score
was statistically different for each of the three groups. The mothers from the
moderate-risk group differed from those in the community sample as well
as those from the high-risk group (medium size). However, both subscales
only differentiated the high-risk group from the other two.
The three groups were included in a series of multinomial logistic regres-
sions to explore the relation of PSI-SF subscales to each group. Table 3

Downloaded from erx.sagepub.com at FLORIDA INTERNATIONAL UNIV on November 14, 2015

436 Evaluation Review 39(4)

Table 1. Item-Total Correlations and Test Reliability.

Item-Total Correlation a If Item Deleted

Items—Parental distressa
1 .225 .793
2 .489 .768
3 .403 .776
4 .526 .765
5 .594 .758
6 .350 .782
7 .416 .775
8 .339 .783
9 .573 .759
10 .557 .761
11 .167 .802
12 .565 .760
a ¼ .795
Items—Childrearing stressb
13 .537 .844
14 .026 .863
15 .582 .841
16 .575 .841
17 .360 .849
18 .315 .851
19 .535 .844
20 .518 .844
21 .314 .851
22 .133 .855
23 .555 .842
24 .443 .846
25 .373 .849
26 .290 .852
27 .532 .843
28 .440 .846
29 .488 .844
30 .457 .846
31 .357 .849
32 -.071 .861
33 .538 .843
34 .512 .844
35 .586 .843
36 .516 .843
a ¼ .853
Note. a ¼ Cronbach’s a.
a
Zskewness ¼ 2.08; Zkurtosis ¼ .075. bZskewness ¼ 0.01; Zkurtosis ¼ .091.

Downloaded from erx.sagepub.com at FLORIDA INTERNATIONAL UNIV on November 14, 2015

Pérez-Padilla et al. 437

Table 2. Variability in the Level of Parenting Stress Between Community,

Moderate-, and High-Risk Groups.

Moderate
Community Risk High Risk
n ¼ 40 n ¼ 46 n ¼ 44 ANOVA

M (SD) M (SD) M (SD) F DMSa db

Parental 79.23 (16.83) 90.45 (20.93) 102.87 (26.38) 9.957*** 1–2* 0.59
stress— 1–3*** 1.07
Total 2–3* 0.52
stress
Personal 27.93 (9.03) 30.73 (9.71) 35.32 (10.25) 6.062** 1–3*** 0.76
distress 2–3* 0.50
Childrearing 53.08 (12.19) 58.57 (15.38) 68.57 (17.69) 10.212*** 1–3*** 1.12
stress 2–3** 0.65

Note. ANOVA ¼ analysis of variance.

a
Post hoc analysis: DMS test. bCohen’s d effect size.
y
p < .1. *p < .05. **p < .01. ***p < .001.

Table 3. Multinomial Logistic Regression Model Parameters Using the Community

Sample as a Reference.

w2 OR inf. OR sup.
B Wald OR p 95% 95%

Parenting Stress Index–SF: Subscalesa

Moderate risk
Intercept .318 1.676
Personal distress .287 0.936 0.745 1.33 0.745 2.380
Childrearing stress .274 0.818 0.726 1.32 0.726 2.386
High risk
Intercept .064 0.058 0.809
Personal distress .463 2.170 1.590 0.141 0.858 2.945
Childrearing stress .833 6.568 2.301 0.010 1.217 4.351
Parenting Stress Index–SF: Totalb
Moderate risk
Intersection .563 4.344 0.037
Parental stress—total stress .664 4.732 1.943 0.030 1.068 3.535
High risk
Intersection .339 1.412 0.235
Parental stress—total stress 1.222 13.963 3.393 0.000 1.788 6.439
Note. SF ¼ Short Form; OR ¼ odds ratio; OR inf. ¼ OR lower; OR sup. ¼ OR upper.
a
Goodness of fit: w2 ¼ 234.03, p ¼ .343. 2LL: w2 ¼ 238.19, p < .001. Pseudo R2 of Nagelkerke
¼ .177. bGoodness of fit: w2 ¼ 127.39, p ¼ .351. 2LL: w2 ¼ 172.38, p < .001. Pseudo R2 of
Nagelkerke ¼ .168.

Downloaded from erx.sagepub.com at FLORIDA INTERNATIONAL UNIV on November 14, 2015

438 Evaluation Review 39(4)

displays the results obtained using the community sample as a reference. As

indicated, the Personal distress subscale did not differ between the three
groups of mothers. Nevertheless, the Childrearing stress revealed differ-
ences between the community sample group and the high-risk group:
One-unit change in Childrearing stress was associated with a 2.31 times
higher odds of being in the high-risk group compared to being in the com-
munity group. Contrary to the subscales, the highest scores in total parental
stress correlated with a greater probability that a mother from the commu-
nity sample group would become part of the moderate-risk group (one-unit
change in parenting stress was associated with a 94% increase in odds of
being in the moderate-risk group compared to the community group) or a
member of the high-risk group (one-unit change in parenting stress was
associated with a 239% increase in odds of being in the moderate-risk group
compared to the community group). Both models explained 17% and 18%
of the variance in the group factor, respectively.

Relationships between the PSI-SF scores and scores on criterion measures. The
results obtained indicated that the total PSI-SF score was related to all mea-
sures, except for the effectiveness as a parent. As shown in Table 4, greater
levels of parental stress were associated with an external locus of control,
poorer satisfaction as a mother, and higher indices of depressive symptoms,
anxiety, social dysfunction, and somatic symptoms. With regard to the sub-
scales, Personal distress followed the same pattern as the total score, while
greater Childrearing stress was negatively related to the satisfaction and
effectiveness as a parent and with a more external PLOC, but this subscale
failed to correlate with any general health indicator.
Finally, a multiple regression analysis was computed to calculate the
proportion of total variance of the general health indicators explained by the
subscales. To control any possible influences of the PSOC and the PLOC, a
hierarchical regression was used. PLOC was introduced in the first block
(since this shows the highest correlation coefficient with the stress scores),
the second block was completed with the subscales related to parental sense
of competence, and in the third block, Personal distress and Childrearing
stress were entered simultaneously.
The model explained approximately 29% of the variance in general
health (see Table 5). Both the first and the second block obtained no signif-
icant change in F, on the contrary, block three showed a significant change
increasing the model explanation by 32%. Attending to the regression coef-
ficients, the scores for PLOC, parental satisfaction, and parental effective-
ness did not obtain any significant b. Therefore, they did not contribute to

Downloaded from erx.sagepub.com at FLORIDA INTERNATIONAL UNIV on November 14, 2015

 Table 4. Pearson Correlations Coefficient Between Parental Stress and the Locus of Control, Parental Efficacy and Parental Satisfaction, and
General Health.

M (SD) 2 3 4 5 6 7 8 9 10 11

1. Parental stress— 86.14 (20.70) .736*** .893*** .484*** .154 .343** .393** .284** .369** .269* .359**
total stress
2. Parental stress— 29.16 (9.33) — .492*** .339** .137 .327** .570*** .410*** .513*** .425*** .558***
Personal distress
3. Parental stress— 56.57 (14.63) — .474*** .242* .352** .240y .178 .205 .239y .213
Childrearing stress
4. Parental locus of 111.93 (18.23) — .182 .217y .068 .040 .003 .155 .146
control
5. Parental efficacy 29.66 (7.05) — .114 .050 .007 .014 .147 .005
6. Parental satisfaction 30.62 (6.75) — .155 .149 .142 .037 .072
7. General malaise 54.38 (17.49) — .886*** .916*** .648*** .883***
8. General malaise— 14.76 (5.36) — .811*** .428*** .626***
Somatic
9. General malaise— 15.89 (6.68) — .389** .697***
Anxiety
10. General malaise— 13.19 (3.02) — .652***
Social dysfunction
11. General malaise— 10.91 (5.18) —

Downloaded from erx.sagepub.com at FLORIDA INTERNATIONAL UNIV on November 14, 2015

Depression
y
p < .1. *p < .05. **p < .01. ***p < .001.

439
440 Evaluation Review 39(4)

Table 5. Summary for the Multiple Regression Analysis Model and Coefficients for
Personal Distress and Childrearing Stress as Explanations for Health.

Model R2 corrected Change in R2 Change in F Durbin-Watson

General health 1 .014 .005 .622 2.004

2 .031 .022 .568
3 .286 .325 <.001

Regression Coefficients

B t p

Block 1
Locus of parental control .068 0.496 .622
Block 2
Parental effectiveness .033 0.234 .816
Parental satisfaction .147 1.039 .304
Block 3
Personal distress .611 4.623 .000
Childrearing stress .104 0.722 .444

the explanation of the mothers’ general health. In the third block, the main
effects for the subscales were observed but only the Personal distress sub-
scale showed a significant b.

Discussion
The objective of this study was to estimate the reliability and report evi-
dences of validity of the PSI-SF in a sample of at-risk mothers attended
by CFPS. Since sample size did not allow computing a confirmatory factor
analysis, the internal structure of PSI-SF was examined using other statisti-
cal procedures. Analyses performed and reported in this article show satis-
factory results regarding appropriate internal consistency coefficients for
the scale as a whole and for the two subscales that Dı́az-Herrero et al.
(2010) recommend for the Spanish female population. Hence, even the
higher a if item deleted indicates that the scales might be shortened without
loss of information, this approach was not considered as appropriate accord-
ing to the reliability results and because it would affect the validity of the
scores (it implicitly includes changing the operational definition of the
construct).
With respect to the evidences of validity, results reported here suggest
that the total PSI-SF score (but not the two subscales) could be useful to

Downloaded from erx.sagepub.com at FLORIDA INTERNATIONAL UNIV on November 14, 2015

Pérez-Padilla et al. 441

differentiate between different groups of mothers with different levels of

risk. These results are similar to the ones founded by Haskett and colleagues
(2006) using the PSI-SF with a sample of parents with or without a docu-
mented history of abuse, but in this work, the scores for the two subscales
also differentiated between the parents. Probably, the subscales of this tool
allow a more precise differentiation when there are greater differences
between the considered levels of risk. If the differences are less acute or
intense (as in the case of at-risk families), only the global score of PSI-
SF discriminate between groups. The obtained results also reveal that a
large percentage of the at-risk sample presents clinical levels of stress as has
also been reported in other studies (Anderson, 2008; Raikes & Thompson,
2005). Summing up, these results suggest that PSI-SF may be a useful
assessment tool for at-risk mothers.
The scores obtained with PSI-SF were related to the mother’s perception
of control, results that have also been reported in other studies (Rodden-
berry & Renk, 2010; Sevigny & Loutzenhiser, 2009). Specifically, the
Childrearing stress subscale was associated with a poorer perception as a
mother as well as with an external locus of control, in line with the theore-
tical framework (Abidin, 1992, 1995; Lazarus & Folkman, 1984). Contrary
to the previous one, the Personal distress subscale was related to all the gen-
eral health indicators, explaining a high percentage of variance (Lyon,
2000). These results are similar to those reported by McKelvey et al.
(2009), who founded a greater relationship of the Personal distress scores
with depressive symptoms, compared with Childrearing stress subscale.
Therefore, both subscales are related to diverse aspects of parental role in
this sample, obtaining different results depending on the domain they are
focusing on (Abidin, 1995; Haskett et al., 2006). These results not only pro-
vide evidences about the validity of this instrument but also suggest that the
total score or the subscales should be used in a differentiated way according
to the professional’s objectives. The Personal distress subscale is sensitive
to the mother’s general health, and therefore, it may be useful in therapeutic
and clinical interventions. The Childrearing stress subscale is related to the
self-perceptions about the parental role. Thus, it could provide professionals
working with mothers in a more preventive and strengthening way with use-
ful information. Similar conclusions have been observed in previous studies
(McKelvey et al., 2009; Whiteside-Mansell et al., 2007), which examine the
usefulness of decomposing the first two PSI-SF scales into more narrowly
defined and shorter subscales. Although the two dimensions proposed in the
present study is supported by both the Abidin’s parenting stress theory
(1983, 1995) and the Spanish adaptation of this scale (Dı́az-Herrero

Downloaded from erx.sagepub.com at FLORIDA INTERNATIONAL UNIV on November 14, 2015

442 Evaluation Review 39(4)

et al., 2010), further studies to capture specific sources of parenting stress in

at-risk families are planned.
Finally, some limitations of this study must be underlined, mainly the
fact that the sample includes only women, reflecting the reality in CFPS.
Furthermore, the level of internal consistency of satisfaction with the par-
ental role scale was weak. Nevertheless, the results reported in this article
show that short version for PSI may be a useful instrument for researchers
and practitioners who work with at-risk families. Moreover, the two-
dimensional structure is appropriate for at-risk mothers, consistently with
the theoretical framework of Abidin (1983, 1992).

Declaration of Conflicting Interests

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

Funding
The authors received no financial support for the research, authorship, and/or pub-
lication of this article.

References
Abidin, R. R. (1983). The parenting stress index. Charlottesville, VA: Pediatric
Psychology Press.
Abidin, R. R. (1992). The determinants of parenting behavior. Journal of Clinical
Child Psychology, 21, 407–412. doi:10.1207/s15374424jccp2104_12
Abidin, R. R. (1995). Parenting stress index (Short Form). Odessa, TX:
Psychological Assessment Resource.
Abidin, R. R. (1997). Parenting stress index: A measure of the parent-child system.
In C. P. Zalaquett & R. Wood (Eds.), Evaluating stress: A book of resources (pp.
277–291). Lanham, MD: Scarecrow Press.
American Educational Research Association, American Psychological Association
& National Council on Measurement in Education. (2002). Standards for educa-
tional and psychological testing. Washington DC: American Educational
Research Association.
Anderson, L. S. (2008). Predictors of parenting stress in a diverse sample of parents
of early adolescents in high-risk communities. Nursing Research, 57, 340–350.
doi:10.1097/01.NNR.0000313502.92227.87
Bloomfield, L., & Kendall, S. (2012). Parenting self-efficacy, parenting stress and
child behaviour before and after a parenting programme. Primary Health Care
Research and Development, 13, 364.

Downloaded from erx.sagepub.com at FLORIDA INTERNATIONAL UNIV on November 14, 2015

Pérez-Padilla et al. 443

Bonds, D. D., Gondoli, D. M., Sturge-Apple, M. L., & Salem, L. N. (2002). Parenting
stress as a mediator of the relation between parenting support and optimal parenting.
Parenting: Science and Practice, 2, 409–435. doi:10.1207/S15327922PAR0204_04
Campis, L. K., Lyman, R. D., & Prentice-Dunn, S. (1986). The parental locus of
control scale: Development and validation. Journal of Clinical Child and
Adolescent Psychology, 15, 260–267. doi:10.1207/s15374424jccp1503_10
Cho, K. S., & Hong, E. J. (2013). A path analysis of the variables related to the qual-
ity of life of mothers with disabled children in Korea. Stress and Health, 29,
229–239. doi:10.1002/smi.2457
Cohen, J. (1988). Statistical power analysis for the behavioral sciences. Hillsdale,
NJ: Lawrence Erlbaum.
Committee of Ministers of the Council of Europe. (2006). Recommendation 19 on
policy to support positive parenting. Retrieved August 2013, from wcd.coe.int/
wcd/ViewDoc.jsp?id¼1073507&Site¼CM
Conrad, K. J., Riley, B. B., Conrad, K. M., Chan, Y. F., & Dennis, M. L. (2010).
Validation of the Crime and Violence Scale (CVS) against the Rasch measure-
ment model including differences by gender, race, and age. Evaluation
Review, 34, 83–115. doi:10.1177/0193841X10362162
Costa, N. M., Weems, C. F., Pellerin, K., & Dalton, R. (2006). Parenting stress
and childhood psychopathology: An examination of specificity to internalizing
and externalizing symptoms. Journal of Psychopathology and Behavioral
Assessment, 28, 113–122. doi:10.1007/s10862-006-7489-3
Crnic, K., & Low, C. (2002). Everyday stresses and parenting. In M. H. Bornstein
(Ed.), Handbook of parenting: Vol. 5. Practical issues in parenting (pp.
243–267). Mahwah, NJ: Lawrence Erlbaum.
Deater-Deckard, K. (1998). Parenting stress and child adjustment: Some old hypoth-
eses and new questions. Clinical Psychology: Science and Practice, 5, 314–327.
doi:10.1111/j.1468-2850.1998.tb00152.x
Deater-Deckard, K., & Scarr, S. (1996). Parenting stress among dual-earner mothers
and fathers: Are there gender differences. Journal of Family Psychology, 10,
45–59. doi:10.1037//0893-3200.10.1.45
Dı́az-Herrero, A., Brito, A., López-Pina, J. A., Pérez-López, J., & Martı́nez-Fuentes,
M. T. (2010). Estructura factorial y consistencia interna de la versión española
del parenting stress index–short form. Psicothema, 22, 1033–1038.
Gerdes, A. C., Hoza, B., Arnold, L. E., Pelham, W. E., Swanson, J. M., & Wigal, T.
(2007). Maternal depressive symptomatology and parenting behavior: Exploration
of possible mediators. Journal of Abnormal Child Psychology, 35, 705–714.
Goldberg, D., & Williams, P. (1996). Cuestionario de salud general GHQ. (General
Health Questionnaire). Guı´a para el usuario de las distintas versiones.
Barcelona, Spain: Masson.

Downloaded from erx.sagepub.com at FLORIDA INTERNATIONAL UNIV on November 14, 2015

444 Evaluation Review 39(4)

Haapsamoa, H., Pollock-Wurmanb, R. A., Kuusikko-Gauffina, S., Ebelinga, H.,

Larinena, K., Soinic, H., & Moilanena, I. (2013). Maternal stress and young chil-
dren’s behavioural development: A prospective pilot study from 8 to 36 months
in a Finnish sample. Early Child Development and Care, 183, 1841–1852.
Harnett, P. H., & Dawe, S. (2008). Reducing child abuse potential in families identi-
fied by social services: Implications for assessment and treatment. Brief Treatment
and Crisis Intervention, 8, 226–235. doi:10.1093/brief-treatment/mhn010
Haskett, M. E., Ahern, L. S., Ward, C. S., & Allaire, J. C. (2006). Factor structure
and validity of the parenting stress index-short form. Journal of Clinical Child
and Adolescent Psychology, 35, 302–312. doi:10.1207/s15374424jccp3502_14
Hurley, K. D., Griffith, A., Ingram, S., Bolivar, C., Mason, W. A., & Trout, A.
(2012). An approach to examining the proximal and intermediate outcomes of
an intensive family preservation program. Journal of Child and Family
Studies, 21, 1003–1017.
Hutchings, J., & Webster-Stratton, C. (2004). In M. Houghugi & N. Long (Eds.)
Community based support for parents (pp. 334–351). Sage Publications.
Johnston, C., & Mash, E. J. (1989). A measure of parenting satisfaction and efficacy.
Journal of Clinical and Child Psychology, 18, 167–175. doi:10.1207/
s15374424jccp1802_8
Lanier, P., Kohl, P. L., Benz, J., Swinger, D., & Drake, B. (2014). Preventing mal-
treatment with a community-based implementation of parent–child interaction
therapy. Journal of Child and Family Studies, 23, 449–460.
Lazarus, R., & Folkman, S. (1984). Stress, appraisal and coping. New York, NY:
Springer.
Lyon, B. (2000). Stress, coping, and health. In V. Hill-Rice (Ed.), Handbook of
stress, coping, and health (pp. 3–23). Washington, DC: Sage.
Marcynyszyn, L. A., Maher, E. J., & Corwin, T. W. (2011). Getting with the (evi-
dence-based) program: An evaluation of the incredible years parenting training
program in child welfare. Children and Youth Services Review, 33, 747–757. doi:
10.1016/j.childyouth.2010.11.021
McKelvey, L. M., Whiteside-Mansell, L., Faldowski, R. A., Shears, J., Ayoub, C., &
Hart, A. D. (2009). Validity of the short form of the parenting stress index for
fathers of toddlers. Journal of Child and Family Studies, 18, 102–111. doi:10.
1007/s10826-008-9211-4
Observatorio de la Infancia en Andalucı́a. (2013). Pobreza y desigualdad.
Cifras y Datos n 10. Junta de Andalucı́a, Spain: Consejerı́a de Salud y Bienestar
Social.
Östberg, M., & Hagekull, B. (2013). Parenting stress and external stressors as pre-
dictors of maternal ratings of child adjustment. Scandinavian Journal of
Psychology, 54, 213–221.

Downloaded from erx.sagepub.com at FLORIDA INTERNATIONAL UNIV on November 14, 2015

Pérez-Padilla et al. 445

Ponnet, K., & Wouters, E. (2014). Stress and mental health in families with different
income levels: A strategy to collect multi-actor data. Journal of Medical Internet
Research, 16, 1–10.
Ponnet, K., Wouters, E., Mortelmans, D., Pasteels, I., De Backer, C., Van Leeuwen,
K., & Van Hiel, A. (2013). The influence of mothers’ and fathers’ parenting
stress and depressive symptoms on own and partner’s parent-child communica-
tion. Family Process, 52, 312–324.
Raikes, H. A., & Thompson, R. A. (2005). Efficacy and social support as predic-
tors of parenting stress among families in poverty. Infant Mental Health
Journal, 26, 177–190. doi:10.1002/imhj.20044
Ramayah, T., Yeap, J. A., & Ignatius, J. (2014). Assessing knowledge sharing
among academics a validation of the Knowledge Sharing Behavior Scale
(KSBS). Evaluation Review, 38, 160–187. doi:10.1177/0193841X14539685
Reitman, D., Currier, R., & Stickle, T. (2002). A critical evaluation of the
Parenting Stress Index-Short Form (PSI-SF) in a head start population.
Journal of Clinical Child and Adolescent Psychology, 31, 384–392. doi:10.
1207/153744202760082649
Roddenberry, A., & Renk, K. (2010). Locus of control and self-efficacy: Potential med-
iators of stress, illness, and utilization of health services in college students. Child
Psychiatry and Human Development, 41, 353–370. doi:10.1007/s10578-010-0173-6
Rodrigo, M. J., Byrne, S., & Álvarez, M. (2012). Preventing child maltreatment
through parenting programmes implemented at the local social services level.
European Journal of Developmental Psychology, 9, 89–103. doi:10.1080/
17405629.2011.607340.
Rodrı́guez, G., Camacho, J., Rodrigo, M. J., Martı́n, J. C., & Máiquez, M. L. (2006).
Evaluación del riesgo psicosocial en familias usuarias de servicios sociales
municipales. Psicothema, 18, 200–206.
Sanders, M. R., & Cann, W. (2002). Promoting positive parenting as an abuse pre-
vention strategy. In K. D. Browne, H. Hanks, P. Stratton, & C. Hamilton (Eds.),
Early prediction and prevention of child abuse: A handbook (pp. 145–163).
West Sussex, England: John Wiley.
Sandı́n, B. (2003). El estrés: Un análisis basado en el papel de los factores sociales.
International Journal of Clinical and Health Psychology, 3, 141–157.
Sevigny, P. R., & Loutzenhiser, L. (2009). Predictors of parenting self-efficacy in
mothers and fathers of toddlers. Child: Care, Health and Development, 36,
179–189. doi:10.1111/j.1365-2214.2009.00980.x
Stith, S. M., Liu, T., Davies, L. C., Boykin, E. L., Alder, M. C., Harris, J. M., . . . Dees,
J. (2009). Risk factors in child maltreatment: A meta-analytic review of the
literature. Aggression and Violent Behavior, 14, 13–29. doi:10.1016/j.avb.
2006.03.006

Downloaded from erx.sagepub.com at FLORIDA INTERNATIONAL UNIV on November 14, 2015

446 Evaluation Review 39(4)

Tabachnick, B. G., & Fidell, L. S. (2007). Using multivariate statistics (5th ed.).
Boston, MA: Pearson Education.
Whiteside-Mansell, L., Ayoub, C., McKelvey, L., Faldowski, R. A., Hart, A., &
Shears, J. (2007). Parenting stress of low-income parents of toddlers and pre-
schoolers: Psychometric properties of a short form of the parenting stress index.
Parenting: Science and Practice, 7, 27–56. doi:10.1207/s15327922par0701_2
Zaidman-Zait, A., Mirenda, P., Zumbo, B. D., Wellington, S., Dua, V., &
Kalynchuk, K. (2010). An item response theory analysis of the parenting stress
index-short form with parents of children with autism spectrum disorders.
Journal of Child Psychology and Psychiatry, 51, 1269–1277. doi:10.1111/j.
1469-7610.2010.02266.x

Author Biographies
Javier Pérez-Padilla is a Professor of University of Huelva (Spain) and his research
is centered in parenting and family dynamics in at-risk families, specifically in par-
enting stress and coping.
Susana Menéndez is a Professor of University of Huelva (Spain). Her research
interests center on studying at-risk families and explore the psychometric properties
of instruments with these families.
Oscar Lozano is a Professor of University of Huelva (Spain) and his research is
centered in opiate-dependent patients and in psychometric evidences evaluation.

Downloaded from erx.sagepub.com at FLORIDA INTERNATIONAL UNIV on November 14, 2015