See discussions, stats, and author profiles for this publication at: [Link]

net/publication/225140110

Development and Validation of the Parental Authority Questionnaire – Revised

Article  in  Journal of Psychopathology and Behavioral Assessment · January 2002

CITATIONS READS

207 19,224

4 authors, including:

David Reitman Stephen D. A. Hupp

SEE PROFILE SEE PROFILE

All content following this page was uploaded by David Reitman on 26 August 2014.

The user has requested enhancement of the downloaded file.

Journal of Psychopathology and Behavioral Assessment, Vol. 24, No. 2, June 2002 (°
C 2002)

Development and Validation of the Parental Authority

David Reitman,1,3 Paula C. Rhode,2 Stephen D. A. Hupp,2 and Cherie Altobello2

Baumrind’s conceptualization of parenting style has influenced research and theorizing about child-
rearing for more than 30 years, yet few clinical applications of this work have been realized. The
purpose of this study was to examine the psychometric properties of a parenting style measure, the
Parental Authority Questionnaire – Revised (PAQ-R). This project furthered the development and val-
idation of the instrument by examining the factor structure, internal consistency, test-retest reliability,
and convergent validity of the PAQ-R in a large, ethnically and socioeconomically diverse sample. The
factor structure and reliability of the PAQ-R was modest, but generally within the acceptable range.
Specifically, the Authoritarian and Permissive subscales of the PAQ-R had modest reliability and
convergent validity; these PAQ-R subscales correlated with relevant subscales of the Parenting Scale
and the Parent–Child Relationship Inventory. By contrast, the Authoritativeness subscale had poorer
psychometric properties when used in lower SES, primarily African American populations. While
African American mothers in the sample reported being more Permissive than did Caucasian moth-
ers, these differences disappeared after controlling for income and maternal education. By contrast,
African American mothers continued to report more Authoritarian attitudes even after controlling for
income and maternal education.

KEY WORDS: parenting style; tests and measurement; child-rearing assessment.

Researchers have assumed parents to have critical in- demandingness” appears superior to either Permissive
fluences on children’s behavioral, emotional, personality, or Authoritarian parenting styles in terms of fostering
and cognitive development for over a century (Holden & academic achievement and social adjustment (Baumrind,
Edwards, 1989; Maccoby, 1984). Consequently, the re- 1991; Dornbusch, Ritter, Leiderman, Roberts, & Fraleigh,
lation of parenting practices to specific child outcomes 1987; Reitman & Gross, 1997). Baumrind’s three parent-
has been a prime research focus in child clinical and de- ing styles appear in Table I.
velopmental psychology (Maccoby & Martin, 1983). In a Despite the wealth of research concerning
comprehensive review, Maccoby and Martin (1983) con- Baumrind’s conceptualization of parenting style
cluded that the construct of “parenting style” had proven (Baumrind, 1971), little is known about the psychometric
exceptionally robust because it combined a variety of qualities of most parenting style measures (Foster &
heavily researched and empirically supported parenting Cone, 1995; Holden & Edwards, 1989; Shelton, Frick,
constructs (e.g., punitiveness, affection, warmth, involve- & Wootton, 1996). Robinson, Mandelco, Olsen, and Hart
ment, and control; see Baumrind, 1971; Baumrind & (1995) noted many problems with efforts to quantify
Black, 1967; Maccoby, 1984). In general, an Authoritative Baumrind’s parenting typology. Holden and Edwards
parenting style emphasizing both “responsiveness and (1989) noted that many researchers simply chose to
develop idiosyncratic measures rather than attempt to val-
1 Center for Psychological Studies, Nova Southeastern University, Fort idate others. Where they have been evaluated, measures
Lauderdale, Florida. such as the Child-Rearing Practices Report (CRPR; Block,
2 Louisiana State University, Baton Rouge, Louisiana.
3 To whom correspondence should be addressed at Center for Psycho- 1965), a 91-item Q-sort, and the modified Likert-type
logical Studies, Nova Southeastern University, 3301 College Avenue, CRPR (Rickel & Biasatti, 1982) have been shown to pos-
Fort Lauderdale, Florida 33314. sess poor psychometric qualities (Reitman & Gross, 1997;

119
0882-2689/02/0600-0119/0 °
C 2002 Plenum Publishing Corporation
P1: GVG/GCP
Journal of Psychopathology and Behavioral Assessment (JOBA) pp459-joba-372163 April 11, 2002 8:35 Style file version Nov. 07, 2000

120 Reitman, Rhode, Hupp, and Altobello

Table I. Baumrind’s Parenting Styles style prototypes described in Table I. Positive correlations
were expected between the Authoritativeness scale of the
Authoritarian Parent is high in control and maturity demands
PAQ-R and the Communication and Limit Setting scales
and low in responsiveness and communication
Authoritative Parent is high in control, responsiveness, of the PCRI. Negative correlations were anticipated
communication, and maturity demands between Authoritativeness and the Laxness and Over-
Permissive Parent is low in control and maturity demands reactivity scales of the PS. The Authoritarian scale of the
and high in communication and responsiveness PAQ-R was expected to positively correlate with the Limit
Setting scale of the PCRI and the Overreactivity scale of
the PS, but negatively correlate with the Communication
Robinson et al., 1995). Other efforts to develop measure- scale of the PCRI and PS Laxness. The Permissive scale
ment tools have relied solely on rational, as opposed to em- of the PAQ-R was expected to positively correlate with
pirical approaches, rendering them of questionable gener- the Communication scale of the PCRI and the Laxness
ality and utility (Kochanska, Kuzynski, & Radke-Yarrow, scale of the PS, but negatively correlate with the Limit
1989; Trickett & Susman, 1989). Robinson et al.’s own Setting scale of the PCRI and the Overreactivity scale of
62-item measure (Robinson et al., 1995) was empirically the PS. Scores on the Authoritative, Authoritarian, and
derived, yet developed almost exclusively with middle- Permissive scales were not expected to correlate with the
class Caucasian parents from intact families in Utah. An- PCRI Social Desirability scale.
other promising instrument, the Parental Authority Ques- A second purpose of this study was to evaluate the
tionnaire (PAQ), was developed by Buri (1988, 1991). performance of the PAQ-R across a diverse group of par-
The 30-item instrument consisted of three 10-item scales ents. Much of Baumrind’s research has been conducted
representing authoritative, authoritarian, and permissive with Caucasian, upper middle class samples, and she has
parenting styles. In validation studies (Buri, 1988, 1991), frequently cautioned that generalization to families of
items were constructed based on Baumrind’s descriptions lower socioeconomic status (SES), or to more ethnically
of the parenting style prototypes and then subjected to diverse samples, may be inappropriate (Baumrind, 1972,
multidisciplinary expert review. The PAQ appeared to 1991). Still, relatively few studies have addressed gen-
have good internal consistency (range = .74–.87) and eralization issues directly (Okazaki & Sue, 1995). When
test-retest reliability ranged from .77 to .92. Furthermore, generalizability issues have been addressed, results have
PAQ scores did not appear vulnerable to social desirability been mixed regarding the influence of socioeconomic
response bias. Although the PAQ was originally designed factors and ethnicity on parenting style. Early studies
to assess parenting style based on retrospective adolescent found disciplinary permissiveness increased with family
ratings, it was chosen for further development because of SES (Hoffman, 1960; Sears, Maccoby, & Levin, 1957).
its brevity and clear relationship to Baumrind’s original More recently, studies suggest that ethnicity may play an
parenting style taxonomy (Buri, 1988). important role in shaping parenting style. For example,
The purpose of this study was to assess the reliability African American parents may be more strict and less
and validity of a parent self-report version of the PAQ; permissive than are Caucasian parents (Dornbusch et al.,
the Parental Authority Questionnaire – Revised (PAQ-R). 1987; Portes, Dunham, & Williams, 1986; Randolph,
In clinical settings, such an instrument could be useful 1995; Reis, Barbara-Stein, & Bennett, 1986; Steinberg,
for assessing response to parent training (see Kazdin & Lamborn, Dornbusch, & Darling, 1992). Darling and
Weisz, 1998). First, the PAQ was modified for parent Steinberg (1993) have suggested that differences in per-
report and altered to improve readability (see Appendix ceptions of “normal parenting practices” may vary across
for a copy of the revised instrument). Second, factor ana- cultures. In contrast, Gutman and Eccles (1999) have ar-
lytic and reliability data were obtained from three diverse gued that economic, rather than ethnic, factors are most
samples. It was expected that the three-factor structure influential in explaining differences in parenting style. In
and favorable psychometric properties of the original this study, the influence of SES on the relation between
PAQ would be preserved in the PAQ-R. Third, to establish ethnicity and PAQ-R scores was evaluated by controlling
convergent validity, PAQ-R subscales were compared for SES in some analyses.
with scales from two other parenting measures; the Par-
enting Scale (PS; Arnold, O’Leary, Wolff, & Acker, 1993)
and the Parent–Child Relationship Inventory (PCRI; INITIAL SCALE DEVELOPMENT
Gerard, 1994). The PS measures disciplinary laxness
and overreativity, whereas the PCRI samples domains Participants in the initial item revision of the PAQ
such as communication and limit setting. Together, these were two doctoral-level psychologists and two graduate
measures sample behavioral concomitants of the parenting students. The doctoral-level researchers, including the
P1: GVG/GCP
Journal of Psychopathology and Behavioral Assessment (JOBA) pp459-joba-372163 April 11, 2002 8:35 Style file version Nov. 07, 2000

The Parental Authority Questionnaire – Revised 121

first author, had extensive experience in parent training, Parents drawn from Samples A and B were recruited
assessment, and child behavior therapy; two other re- through letters sent home with their children and asked
viewers were advanced graduate students in psychology to return completed questionnaires containing only the
with extensive knowledge of the parenting literature and PAQ-R and a demographic form. Children returned
Baumrind’s typologies. For the initial revision of the PAQ, signed, sealed envelopes within a week of the original
items were subjected to informal expert review to insure distribution of questionnaires. To enhance response rates,
content validity concomitant with Baumrind’s parenting a $50 gift certificate was promised to a randomly selected
typologies and improve item clarity. Item content from parent in each sample.
the PAQ was retained, and questions were edited to con- To assess convergent validity, the PAQ-R, a question-
vert the second-person grammatical format to first-person naire requesting demographic information (i.e., ethnicity,
pronominals. A more formal analysis of subscale item marital status, education, and income), and two other par-
content was undertaken by the two doctoral-level review- enting instruments (described below) were administered
ers who were asked to assign each item to its rationally to a third group (Sample C) of families at mandatory preen-
derived factor (e.g., Authoritative, Authoritarian, or Per- rollment orientation meetings for Head Start. All parents
missive). No errors were made in the assignment of items willing to participate signed a consent form, completed the
to factors. No items were deleted from the original PAQ, research instruments, and returned them at the conclusion
but 13 items were altered slightly for content, grammar, of the meeting. Parents were advised that their participa-
and/or readability. After revision, the reading level of the tion was voluntary and that nonparticipation would not
PAQ-R was approximately sixth grade (Reference Soft- adversely affect their status in Head Start. Teachers and
ware International, 1986). The revised PAQ, or PAQ-R, is research assistants were available to answer questions re-
a parental self-report version of the original PAQ, intended garding individual items and to orally administer the mea-
for use with parents of children 3–8-years-old. The PAQ-R sures to those parents who had difficulty reading them.
consists of 30-items, with three 10-item scales represent- Oral administration was required in only three instances,
ing authoritative, authoritarian, and permissive parenting precluding statistical analysis of response bias induced by
styles. Items are rated on a 5-point Likert-type Scale rang- oral administration. However, a review of subscale means
ing from 1 (strongly disagree) to 5 (strongly agree). Sub- from these protocols indicated their PAQ scores gener-
scale scores range from 10 to 50. Greater appraised levels ally fell within a standard deviation of other protocols. To
of the parental authority prototype yield higher scores. reduce missing data, questionnaires were reviewed imme-
diately upon completion and parents were asked to provide
missing information. Questionnaires missing data for one
METHOD
or more items were eliminated from subsequent analyses,
resulting in a final sample of 171 participants.
Participants
MEASURES
Parents and their children in grades pre-k through 5,
were recruited from one public and one private elemen-
The PCRI (Gerard, 1994) is a 78-item self-report
tary school and a Head Start program in the south-central
questionnaire designed to assess parents’ attitudes toward
United States. Sample A (private school) consisted of
parenting and their children. Ongoing research cited in
87 older, predominantly Caucasian, married, higher SES
the manual indicates that the scales have good internal
mothers who had attended some college. Sample B (pub-
consistency, content validity, and adequate construct and
lic school) consisted of 102 young, single, predominantly
criterion-related validities (Gerrard, 1994). However, in-
African American, lower SES mothers who had not at-
dependent validation of the PCRI is still needed (Heinze
tended college. Participants in Sample C were 171 chil-
& Grisso, 1996). The PCRI has seven content scales and
dren, aged 3–5 years, and their primary caregivers, re-
two validity indicators, four of the content scales appeared
cruited during a preenrollment orientation at a Head Start
relevant to Baumrind’s parenting typologies. The Com-
program. Response rates were 50, 35, and 83% for Sam-
munication scale assesses the parent’s perception of how
ples A, B, and C, respectively.
effectively he or she communicates with the child and the
Limit Setting scale assesses the parent’s disciplinary prac-
Procedures tices. The Involvement and Autonomy subscales showed
poor internal consistency in this sample (r = .36 and .55,
One private and one public elementary school were respectively) and were dropped from subsequent analy-
chosen at random from a list of local schools. School ses. Coefficient alphas for the Communication (r = .67)
principals were then contacted and agreed to participate. and Limit Setting (r = .75) subscales were minimally
P1: GVG/GCP
Journal of Psychopathology and Behavioral Assessment (JOBA) pp459-joba-372163 April 11, 2002 8:35 Style file version Nov. 07, 2000

122 Reitman, Rhode, Hupp, and Altobello

acceptable. The PCRI social desirability subscale had an to evaluate a three-factor structure. Items were retained if
alpha of .74. Because low scores on the social desirability they had a loading of at least .30 (Robinson et al., 1995).
scale indicate high social desirability, the sign or direction In the present analyses, all items loaded that highly on at
of relations (e.g., “+” or “−”) was reversed for interpretive least one subscale. Both eigenvalue and scree plot analysis
purposes. appeared to converge on a three-factor solution.
The PS (Arnold et al., 1993) is a 30-item self-report As shown in Tables III and IV, the PAQ-R demon-
questionnaire designed to measure dysfunctional disci- strates a fairly stable factor structure for two of the three
pline practices for parents with young children. The PS has PAQ-R scales. An average of 90% of the Authoritarian
three content scales relating to dysfunctional discipline items and 83% of the Permissive subscale items loaded
style. The Laxness subscale describes parents who give above .30 on their predicted subscale for all samples com-
in to their children, allow rules to go unenforced, or posi- bined. However, for the Authoritative subscale the re-
tively reinforce misbehavior. The Overreactivity subscale sults diverged across samples. In the primarily Caucasian,
surveys parent reports of anger or irritability in discipline higher SES sample, 90% of items loaded as expected,
encounters, whereas Verbosity reflects an overreliance on whereas only 70 and 60% loaded as expected for the pri-
explaining discipline practices to the child. The PS has marily African American, lower SES Samples B and C,
moderate construct and predictive validity and evidence respectively. Specifically, in Samples B and C, several
of convergent relations with measures of child behavior, items on the Authoritativeness scale either did not load
marital functioning, maternal depression, and behavioral on a factor or loaded on another factor (often Authoritar-
ratings (Arnold et al., 1993; Reitman et al., 2001). In ianism). Table IV presents eigenvalues and percentage of
the present analysis, only the Laxness and Overreactiv- variance accounted for in each sample. The proportion of
ity scales were used, because the Verbosity subscale was variance accounted for by the Authoritativeness scale is
not hypothesized to reflect a theoretical construct relevant much smaller for Samples B (8.3%) and C (11.6%) than
to this study. Psychometric data from previous research that for Sample A (16.9%).
has indicated subscale internal consistencies in the .80s or Table V presents PAQ-R subscale means and standard
higher. Coefficient alphas for the present sample were .71 deviations along with the scale internal consistencies for
for Laxness and .72 for Overreactivity. the three samples. According to Anastassi (1988), reliabil-
ity coefficients of .80 or higher are considered desirable;
RESULTS therefore, the internal consistency of the PAQ-R should
be considered only modest overall (see Table V). Coeffi-
Table II shows the demographic makeup of the three cient alphas for the Authoritarian and Permissive scales
samples. Although the gender distribution was even across ranged from .72 to .76 across all samples. By contrast,
samples, the samples were significantly different ( p < the Authoritativeness scale registered an alpha of .77 in
.001) in terms of mother’s education, mother’s age, fam- the predominantly Caucasian, higher SES Sample A, but
ily’s yearly gross income, child age, ethnicity, and marital yielded considerably lower estimates in Samples B and C,
status. .56 and .66, respectively. One-month test-retest reliability
Tables III and IV present the results of the preliminary for 22 mothers from Sample A was .61 for Authoritative,
factor analyses with samples A and B as well as the results .87 for Authoritarian, and .67 for Permissive subscales.
of a factor analysis conducted with sample C. A principal- One-month test-retest reliability data were also gathered
components analysis with varimax rotation was performed from Sample C (n = 40) and yielded values of .54, .88,
and .74, for Authoritative, Authoritarian, and Permissive
Table II. Participants and Setting subscales, respectively.
Across samples, significant mean differences emer-
Sample A Sample B Sample C ged on the Authoritarian, F(2, 348) = 18.01, p < .001,
and Permissive, F(2, 348) = 6.06, p < .001, subscales,
Child’s age (average years) 7.1a 7.4a 3.5b
Mother’s age (average years) 41a 30b 28b
whereas differences on the Authoritative scale were
College or higher (%) 69a 12b 2c nonsignificant, F(2, 348) = 2.73, p > .05. Specifically,
Black/White/other (%) 11/79/10a 78/21/1b 83/14/4b parents from Samples B and C (primarily lower SES,
Income above 30 K (%) 89a 25b 15c African American parents) scored higher on the Author-
Married (%) 79a 28b 17c itarian scale than did parents from Sample A (primarily
Sample size 87 102 171
Caucasian). Parents from Sample C also scored higher
Note. Numbers with different superscripts are significantly different at than parents from Sample A on the Permissive subscale.
p < .01. Further examination of these differences (accomplished
P1: GVG/GCP
Journal of Psychopathology and Behavioral Assessment (JOBA) pp459-joba-372163 April 11, 2002 8:35 Style file version Nov. 07, 2000

The Parental Authority Questionnaire – Revised 123

Table III. Factor Loadings for All PAQ-R Items Across Three Samples: Organized by Subscales

Sample A Sample B Sample C

Authoritarian scale
2 — — −.36 — — — — .49 —
3 .69 — — .65 — — .69 — —
7 .63 — — .65 — — .58 — —
9 .36 — — .58 — — .51 — —
12 .66 — −.32 .41 — — .50 — —
16 .59 −.36 — .52 — — .56 — —
18 .61 — — .49 — — .42 .42 —
25 .79 — — .45 — — .35 — —
26 .47 — — .43 — — .57 — —
29 .38 — — .66 — — .59 — —
Authoritative scale
4 — .58 — — .51 — .51 .39 —
5 — .33 .52 — — −.36 .32 .39 —
8 — .48 — .42 .41 — — .50 —
11 — .67 — — .53 — — — —
15 — .73 — .30 .56 — .35 .41 —
20 — .61 — — .48 — — .42 —
22 — .38 — −.46 .38 .36 — .38 .46
23 — .76 — — .55 — — 44 —
27 — .48 .38 — .39 — — .57 .35
30 — .56 — — .48 — — .63 —
Permissive scale
1 — — .67 — — — — — .58
6 — — .57 — — .47 — — .42
10 — — .49 — — — — — .59
13 — — .45 — — .58 — — —
14 — — .49 — — .54 — — .55
17 — — .57 — — .58 — — .66
19 — — .37 — — .77 — — .36
21 −.48 — .48 — — .35 — — .49
24 — — — — — .41 — — .64
28 — — — — — .51 — — .58

Note. Sample A, N = 87; Sample B, N = 102; Sample C, N = 171. ATR = Authoritarian, ATT = Authoritative, PER = Permissive. The dash (—)
indicates item loaded less than .30.

by combining Samples A, B, and C; N = 360) was different when contrasts for family income, maternal ed-
revealing. Although Permissive and Authoritarian sub- ucational level, and ethnicity were performed. Because
scale scores did not differ as a function of marital status, ethnicity is frequently cited as a factor in parenting style
mother’s or child’s age, or child gender, scores on the without consideration of SES (Gutman & Eccles, 1999),
Permissive and Authoritarian subscales were significantly the relation of ethnicity and SES to parenting style was
evaluated with a series of ANCOVAs. For Permissiveness,
Table IV. Summary of Factor Analyses for PAQ-R Subscales Across entering maternal education, F(1, 342) = 11.31, p <
Three Samples .001, or family income, F(1, 349) = 15.94, p < .001,
as covariates resulted in a significant adjustment to the
Sample A Sample B Sample C means and rendered differences between African Amer-
Factor EV % var EV % var EV % var ican and Caucasian mothers nonsignificant. However,
African American parents continued to have more Au-
Authoritative 5.06 16.9 2.49 8.3 3.47 11.6 thoritarian attitudes than did Caucasian parents. Entering
Authoritarian 4.07 13.6 4.54 15.1 3.40 11.3
Permissive 2.37 7.9 3.23 10.9 2.91 9.7
maternal education, F(1, 342) = 2.18, p > .10, and fam-
ily income, F(1, 349) = 2.46, p > .10, as covariates re-
Note. EV = Eigenvalue; % var = variance accounted for. sulted in nonsignificant adjustments to the overall model.
P1: GVG/GCP
Journal of Psychopathology and Behavioral Assessment (JOBA) pp459-joba-372163 April 11, 2002 8:35 Style file version Nov. 07, 2000

124 Reitman, Rhode, Hupp, and Altobello

Table V. PAQ-R Scale Means, Standard Deviations, and Coefficients Desirability scale of the PCRI. Nonsignificant correlations
Alpha, for Three Samples were found between the Authoritative (r = −.03, p <
.66) and Permissive (r = −.13, p < .09) subscales of the
Sample
PAQ-R and the Social Desirability scale of the PCRI, sug-
A B C gesting that responses on these scales were not adversely
Authoritarian affected by social desirability bias. However, a signifi-
Mean 34.5a 38.8b 38.1b cant correlation was found between the Authoritarian scale
Standard deviation 4.8 5.6 5.3 (r = .26, p < .01) of the PAQ-R and the Social Desirabil-
Alpha .76 .75 .72 ity scale.
Permissive
Mean 21.6a 22.8ab 24.1b
Standard deviation 4.8 6.2 6.3 DISCUSSION
Alpha .74 .74 .73
Authoritative This study examined the psychometric properties
Mean 40.5 41.4 40.2 of the PAQ-R in a diverse sample of parents of pre-
Standard deviation 4.2 3.7 4.3
Alpha .77 .56 .66
school- and elementary school-aged children. This study
did not support the original three-factor structure of the
Note. Numbers with different superscripts are significantly different at PAQ-R across all populations sampled, as suggested by
p < .01. Buri (1991) and Reitman, Altobello, and Hupp (1997).
Instead, the data suggest that the factor structure of the
Foster and Cone (1995) have argued that demonstra- PAQ-R is influenced significantly by sample character-
tions of reliability and factor integrity do not speak to an istics such as SES or ethnicity or both. Several items
instrument’s broader representativeness or validity. Thus, loaded on unpredicted or multiple factors, particularly
we also sought to establish the convergent validity of the those items intended to discriminate between authorita-
PAQ-R, along with evidence of its susceptibility to social tive and authoritarian disciplinary practices. For example,
desirability bias. in Sample C, two of the items predicted to load on the
As reflected in Table VI, bivariate correlations Authoritative scale loaded on both this scale and the Au-
were calculated between the three PAQ-R subscales, thoritarian scale. One item predicted to load on the Author-
the Laxness and Overreactivity scales of the PS, and itarian scale loaded on the Authoritative scale instead, and
the Communication and Limit Setting scales of the three other items loaded comparably on both the Author-
PCRI. As predicted, Permissiveness was associated with itarian and Authoritative scales. If replicable, these data
greater Laxness and less Limit Setting. However, con- suggest that the distinction between authoritative and au-
trary to predictions, Permissiveness was associated with thoritarian parenting styles may be less meaningful than in
greater Overreactivity, and less, rather than more Com- previously studied populations (e.g., Caucasians of higher
munication. Authoritative parenting was associated with SES; cf. Baumrind, 1971, 1978; Baumrind & Black, 1967;
greater Communication but was nonsignificantly associ- Dornbusch et al., 1987).
ated with other factors. Authoritarian parenting was asso- Reliability data were generally consistent with the
ciated with greater Overreactivity, but interestingly, more results of the factor analyses. Two of the three scales of
Communication. the PAQ-R have modest reliability, with Authoritativeness
Finally, bivariate correlations were calculated be- demonstrating unacceptably low reliability in lower SES,
tween the three subscales of the PAQ-R and the Social primarily African American samples. Coefficient alpha is
considered to be a reasonably good indicator of the degree
to which the items comprising a particular construct are
Table VI. PAQ-R Subscales and Selected PS, PCRI Scale Correlations
homogenous in content (Anastassi, 1988). As reflected in
PCRI subscales Table V, the internal consistency of the PAQ-R subscales
PS subscales ranged from .56 to .77. Although coefficient alphas of .80
Limit
Laxness Overreactivity Communication setting and above are generally considered to be desirable, Clark
and Watson (1995) note that “it is not uncommon for con-
Authoritarian −.08 .24∗∗ .25∗∗ −.12 temporary researchers to characterize reliabilities in the
Authoritative −.03 −.02 .34∗∗ −.13 .60s and .70s as good or adequate” (p. 315). Neverthe-
Permissive .26∗∗ .27∗∗ −.18 −.30∗∗
less, reliabilities (both test-retest and internal consistency)
Note. Sample C, N = 171. for the Authoritarian and Permissive subscales were mod-
∗∗ p < .001. est. The internal consistency of the Authoritative subscale
P1: GVG/GCP
Journal of Psychopathology and Behavioral Assessment (JOBA) pp459-joba-372163 April 11, 2002 8:35 Style file version Nov. 07, 2000

The Parental Authority Questionnaire – Revised 125

averaged only .66, with values of .56 and .66 in Samples B between parents and children. Indeed, clear communica-
and C, respectively. The lower internal consistency of this tion, albeit with an authoritarian “flavor,” may be an impor-
scale is also reflected in the factor loadings of individual tant feature of disciplinary practices for parents of children
PAQ-R items. Given that alpha estimates are known to rise living in lower SES communities (see Peters, 1988).
as more items are added to a scale (Cohen, 1996), it may As noted by Randolph (1995), Authoritarian child-
desirable to add or significantly revise items on this scale rearing practices may be valued in the African American
to better capture the Authoritativeness construct. community because “strictness is seen by African
Modest convergent validity was demonstrated for American mothers as protection from the streets.” More-
the PAQ-R subscales via their predicted correlations with over, these parents “emphasize obedience, but this is not
conceptually similar scales on the PS (Arnold et al., 1993) seen by them as negative, instead it is viewed as necessary
and the PCRI (Gerard, 1994). Authoritativeness was because it ‘would make life easier for my child,’ ‘means
associated with effective parent–child communication, respect,’ ‘is equated with my love,’ or ‘is necessary if
Authoritarianism was associated with overreactivity my child is to achieve in school’ ” (Randolph, 1995;
in parent–child interactions, and Permissiveness was p. 137). However, although some argue that Authoritarian
correlated with lax and nonrestrictive disciplinary prac- parenting may be moderated by SES, with mothers
tices. Contrary to predictions, the authoritarian style was becoming less authoritarian as income rises (Gutman &
positively correlated with communication as measured Eccles, 1999; Zegiob & Forehand, 1975), that did not
by the PCRI, and permissiveness was associated with appear to be the case in this study. After controlling for
overreactivity. It was also surprising that neither the maternal education and family income, ethnic differences
authoritative nor authoritarian pattern was associated in Authoritarian subscale scores remained significant,
with greater limit setting as would be predicted by although not very large in absolute terms.
Baumrind’s taxonomy (Baumrind, 1971). There are several limitations to this study. For ex-
A positive correlation was found between the Au- ample, as with all studies of parenting, this study lacked
thoritarian scale of the PAQ-R and the Social Desirabil- a “gold standard” against which to validate the measure.
ity scale. Although not strong, evidence of social desir- Additionally, with the exception of the Limit Setting scale
ability bias may complicate the interpretation of relations of the PCRI, the scales and behavioral dimensions used
between authoritarian parenting and other psychosocial for comparative purposes in this study were not explicitly
variables. Indeed, although the relationship of authoritar- based on Baumrind’s taxonomy. Instead, they were chosen
ian parenting to social desirability could be related to eth- based on their perceived conceptual similarly to the com-
nicity or SES (i.e., in the African American community ponents of the parenting styles identified by Baumrind.
parents may regard Authoritarian values as normative; see As such, it may be unrealistic to expect high correlations
Peters, 1988; Randolph, 1995), the ethnic homogeneity of for many of the comparisons. Finally, although the sizes
the Head Start sample precluded an analysis of its relation- of the reliability and validity coefficients are modest, it
ship to social desirability. Interestingly, social desirabil- should be noted that the very low SES of some participants
ity was most strongly correlated with the following items almost certainly increased the extent of measurement er-
from the Authoritarian subscale of the PAQ-R: Item 7, ror, and hence may have attenuated relationships between
“I do not allow my children to question the decisions that variables.
I make” (r = .33) and Item 9, “Other parents should use The above limitations notwithstanding, the meth-
more force to get their children to behave” (r = .26). ods used in this study represent a more rigorous test
Psychometric characteristics of the PAQ-R and its re- of psychometric properties than has been employed in
lation to other measures across these diverse populations most previous studies of parenting instruments (Holden
differed from those hypothesized in several ways. For ex- & Edwards, 1989). This study also appears to be unusual
ample, Baumrind’s conceptualization of authoritarian par- in that it attempted to validate a new parenting measure
enting describes parents who are high on the dimensions based on comparisons to more established measures of
of control and maturity demands, but low in responsive- conceptually similar constructs. Future work in minority
ness and communication. However, in the current sam- and clinical populations may prove especially valuable in
ple, although authoritarian parenting was associated with elaborating its utility, especially where sample character-
overreactivity, it was also correlated with effective parent– istics (e.g., marital status, parent and child age, etc.) and
child communication. This surprising finding suggests SES are carefully controlled. Studies holding ethnicity
that the negative outcomes associated with authoritarian constant while contrasting SES, and observational studies
parenting practices in other populations could be attenu- involving mothers of clinically referred children may be
ated in this population by high levels of communication particularly illuminating.
P1: GVG/GCP
Journal of Psychopathology and Behavioral Assessment (JOBA) pp459-joba-372163 April 11, 2002 8:35 Style file version Nov. 07, 2000

126 Reitman, Rhode, Hupp, and Altobello

APPENDIX
P1: GVG/GCP
Journal of Psychopathology and Behavioral Assessment (JOBA) pp459-joba-372163 April 11, 2002 8:35 Style file version Nov. 07, 2000

The Parental Authority Questionnaire – Revised 127

REFERENCES Maccoby, E. E., & Martin, J. (1983). Socialization in the context of fam-
ily: Parent–child interaction. In P. H. Mussen (Series Ed.) & E. M.
Hetherington (Vol. Ed.), Handbook of child psychology: Vol. 4.
Anastassi, A. (1988). Psychological testing. New York: Macmillan. Socialization, personality, and social development (pp. 1–101).
Arnold, D. S., O’Leary, S. G., Wolff, L. S., & Acker, M. M. (1993). The New York: Wiley.
parenting scale: A measure of dysfunctional parenting in discipline Okazaki, S., & Sue, S. (1995). Methodological issues in assessment
situations. Psychological Assessment, 5, 137–144. research with minorities. Psychological Assessment, 7, 367–375.
Baumrind, D. (1971). Current patterns of parental authority. Develop- Peters, M. F. (1988). Parenting in black families with young children: A
mental Psychology Monographs, 4(1, Pt. 2), 1–103. historical perspective. In H. P. McAdoo (Ed.), Black families (2nd
Baumrind, D. (1972). An exploratory study of socialization effects on ed.). Newbury Park, CA: Sage.
black children: Some black and white comparisons. Child Devel- Portes, P., Dunham, R. M., & Williams, S. (1986). Assessing child-
opment, 43, 261–267. rearing style in ecological settings: Its relation to culture, social
Baumrind, D. (1978). Parental disciplinary patterns and social compe- class, early age intervention and scholastic achievement. Adoles-
tence in children. Youth and Society, 9, 239–275. cence, 21, 723–735.
Baumrind, D. (1991). The influence of parenting style on adolescent Randolph, S. M. (1995). African American children in single-mother
competence and substance abuse. Journal of Early Adolescence, families. In B. J. Dickerson (Ed.), African American single moth-
11, 56–95. ers: Understanding their lives and families. Thousand Oaks, CA:
Baumrind, D., & Black, A. E. (1967). Socialization practices associated Sage.
with dimensions of competence in preschool boys and girls. Child Reis, J., Barbera-Stein, L., & Bennett, S. (1986). Ecological determinants
Development, 38, 291–327. of parenting. Family Relations, 35, 547–554.
Block, J. (1965). The child-rearing practices report. Berkeley: Institute Reitman, D., Altobello, C., & Hupp, S. D. A. (1997, November). Devel-
of Human Development, University of California. opment of a self-report parenting measure: The Parental Authority
Buri, J. R. (1988). The nature of humankind, authoritarianism, and self- Questionnaire – Revised (PAQ-R). Poster presented at the 31st an-
esteem. Journal of Psychology and Christianity, 7, 32–38. nual meeting of the Association for the Advancement of Behavior
Buri, J. R. (1991). Parental authority questionnaire. Journal of Person- Therapy, Miami, FL.
ality Assessment, 57, 110–119. Reitman, D., Currier, R. O., Hupp, S. D. A., Rhode, P. C., Murphy,
Clark, L. A., & Watson, D. (1995). Constructing validity: Basic issues M., & O’Callaghan, P. M. (2001). Psychometric characteristics of
in objective scale development. Psychological Assessment, 7, 309– the Parenting Scale in a Head Start population. Journal of Clinical
319. Child Psychology, 30, 514–524.
Cohen, R. J. (1996). Psychological testing and assessment: An introduc- Reitman, D., & Gross, A. M. (1997). The relation of maternal child-
tion to tests and measurement. Mountain View, CA: Mayfield. rearing attitudes to delay of gratification among boys. Child Study
Darling, N., & Steinberg, L. (1993). Parenting style as context: An inte- Journal, 27, 279–300.
grative model. Psychological Bulletin, 113, 487–496. Reference Software International. (1986). Grammatika II [Computer
Dornbusch, S. M., Ritter, P. L., Leiderman, H., Roberts, D. F., & Fraleigh, program]. Salt Lake City, UT: Author.
M. J. (1987). The relation of parenting style to adolescent school Rickel, A. U., & Biasatti, L. L. (1982). Modification of the Block child-
performance. Child Development, 58, 1244–1257. rearing practices report. Journal of Clinical Psychology, 38, 129–
Foster, S. L., & Cone, J. D. (1995). Validity issues in clinical assessment. 134.
Psychological Assessment, 7, 248–260. Robinson, C. C., Mandelco, B., Olsen, S. F., & Hart, C. H. (1995).
Gerard, A. B. (1994). The Parent–Child Relationship Inventory. Los Authoritative, authoritarian and permissive parenting practices:
Angeles: Western Psychological Services. Development of a new measure. Psychological Reports, 77,
Gutman, L. M., & Eccles, J. S. (1999). Financial strain, parenting be- 819–830.
haviors, and adolescents’ achievement: Testing model equivalence Sears, R. R., Maccoby, E. E., & Levin, H. (1957). Patterns of child
between African American and European American single- and two rearing. Evanston, IL: Row & Peterson.
parent families. Child Development, 70, 1464–1476. Shelton, K. K., Frick, P. J., & Wootton, J. (1996). Assessment of parenting
Heinze, M. C., & Grisso, T. (1996). Review of instruments assessing par- practices in families of elementary school-age children. Journal of
enting competencies used in child custody evaluations. Behavioral Clinical Child Psychology, 25, 317–329.
Sciences and the Law, 14, 293–313. Steinberg, L., Lamborn, S., Dornbusch, S., & Darling, N. (1992). Im-
Hoffman, M. L. (1960). Power assertion by the parent and its impact on pact of parenting practices on adolescent achievement: Authorita-
the child. Child Development, 31, 129–143. tive parenting, school involvement, and encouragement to succeed.
Holden, G. W., & Edwards, L. A. (1989). Parental attitudes toward Child Development, 63, 1266–1281.
childrearing: Instruments, issues and implications. Psychological Trickett, P. K., & Susman, E. J. (1989). Perceived similarities
Bulletin, 106, 29–58. and disagreements about childrearing practices in abusive and
Kazdin, A. E., & Weisz, J. R. (1998). Identifying and developing em- nonabusive families: Intergenerational and concurrent family pro-
pirically supported child and adolescent treatments. Journal of cesses. In D. Cicchetti & V. Carlson (Eds.), Child maltreat-
Consulting and Clinical Psychology, 66, 19–36. ment: Theory and research on the causes and consequences
Kochanska, G., Kuczynski, L., & Radke-Yarrow, M. (1989). Correspon- of child abuse and neglect. New York: Cambridge University
dence between mother’s self-report and observed child rearing prac- Press.
tice. Child Development, 60, 56–69. Zegiob, L., & Forehand, R. (1975). Maternal interactive behavior as
Maccoby, E. E. (1984). Socialization and developmental change. Child a function of race, socioeconomic status, and sex of child. Child
Development, 55, 317–328. Development, 46, 564–568.

View publication stats