Content uploaded by Sherry H Stewart
Author content
All content in this area was uploaded by Sherry H Stewart
Content may be subject to copyright.
P1: GDX
Journal of Psychopathology and Behavioral Assessment (JOBA) pp835-joba-464660 April 22, 2003 12:54 Style file version June 25th, 2002
Journal of Psychopathology and Behavioral Assessment, Vol. 25, No. 3, September 2003 (
C
°
2003)
Dyscontrol Experiences Questionnaire: Development,
Reliability and Validity
Margo C. Watt
1,3
and Sherry H. Stewart
2
Accepted December 2, 2002
This paper describes development of a retrospective measure of childhood exposure to parental
dyscontrol related to specific conditions such as a parent’s drinking, anger, or other negative
emotional states (e.g., stress, anxiety, depression). A 45-item questionnaire, the Dyscontrol
Experiences Questionnaire (DEQ), was designed to assess individuals’ childhood experiences
with parental loss of control. Two student samples (n = 465 and n = 536) completed a battery
of self-report measures. Responses were used to examine the psychometric properties of the
DEQ. Factor analysis revealed three lower-order factors accounting for 54.4% of the variance
in DEQ item scores. Cronbach’s coefficient alphas indicated acceptable internal consistency.
Moderate intercorrelations with two other retrospective measures of childhood experiencessuggested
construct validity. Moderate intercorrelations with two measures of psychological distress suggested
predictive validity. In summary, initial psychometric testing of the DEQ suggests it is a reliable
and valid tool for investigating an important developmental antecedent of adult psychological
distress.
KEY WORDS: anxiety; sensitivity; parental dyscontrol; measurement; psychometrics.
The concept of control is considered to be an im-
portant variable in determining individual susceptibil-
ity to anxiety-related distress (see Barlow, Chorpita, &
Turovsky, 1996; Chorpita & Barlow, 1998, for reviews).
Forexample, the experienceofanunexpected panic attack
mayleadtothedevelopment of panic disorder in individu-
als who perceive that they cannot control emotional expe-
riences or bodily reactions (Barlow, 1988). More specif-
ically, a perceived lack of control over panic sensations
may elevate an individual’s somatic activity and negative
thought processes, thereby increasing the probability of a
future panic attack.
Based largely on the results of studies with animals,
it has been hypothesized that people acquire “control
deficits” on the basis of experiences with uncontrollable
1
Department of Psychology, St. Francis Xavier University, Antigonish,
Nova Scotia, Canada.
2
Departments of Psychology and Psychiatry, Dalhousie University,
Halifax, Nova Scotia, Canada.
3
To whom correspondence should be addressed at Department of Psy-
chology, St. Francis Xavier University, P.O. Box 5000, Antigonish,
Nova Scotia, Canada B2G 2W5; e-mail: mwatt@stfx.ca.
aversive events (Mineka & Kelly, 1989). Chorpita and
colleagues (Chorpita, 2001; Chorpita & Barlow, 1998)
have suggested that early experience with uncontrol-
lable or unpredictable stimuli may influence low per-
ceptions of control (cf., Carton & Nowicki, 1994). For
example, a parenting style characterized by high protec-
tion and discouragement of autonomy has been shown
to foster an external locus of control (i.e., belief that
events are beyond one’s personal control) in children and
has been associated with later anxiety and depressive
disorders (see Parker, 1983; Schneewind, 1995). Other
ways in which certain childhood learning experiences
might engender concerns about control and subsequently
risk for anxiety symptoms/disorders remain relatively
underinvestigated.
It has been suggested that sufficient early experi-
ences with uncontrollable events may eventually lead to
an increased generalized tendency to perceive or process
events as not within one’s control (i.e., a psychological
vulnerability; Schneewind, 1995). In their analysis of
environmental influences on the development of anxiety,
Chorpita and Barlow (1998) suggested that experiencing
155
0882-2689/03/0900-0155/0
C
°
2003Plenum Publishing Corporation
P1: GDX
Journal of Psychopathology and Behavioral Assessment (JOBA) pp835-joba-464660 April 22, 2003 12:54 Style file version June 25th, 2002
156 Watt and Stewart
diminished control over events during development may
establish stored memories related to behavior–outcome
relations that would lead an individual to be more apt
to predict an aversive outcome in the face of stress.
In other words, stored information related to a history
of low control could result in heightened behavioral
inhibition system activity or increased anxiety (Gray &
McNaughton, 2000). It could be theorized that repeated
childhood exposure to a parent acting in an uncontrolled
fashioncoulddiminishachild’s sense ofcontrolover their
environment, resulting in psychological vulnerability to
distress. Consistent with social learning theory (Bandura,
1986), the more salient the parental dyscontrol behavior,
the more potent such a learning experience would be for a
child.
Ehlers (1993) investigated the relationship between
retrospectively reported childhood and adolescent learn-
ing experiences and panic attack development. Using a
modified version of a Learning History Questionnaire
(LHQ) originally designed by Whitehead, Busch, Heller,
and Costa (1986), Ehlers (1993) found that anxiety disor-
dered groups reported more frequent exposure to parental
uncontrolled behavior (e.g., associated with drunkenness
or anger) as compared to controls. This category of learn-
ing experiences comprised one of five LHQ subscales and
had been included in the study in response to anecdo-
tal clinician impressions that a relationship existed be-
tween anxiety patients’ concerns about loss of control and
parental alcoholism (Ehlers, 1993). Further support for
the relationship between exposure to parental dyscontrol
and anxiety outcomes was provided by Watt, Stewart, and
Cox (1998). Using the same measure as Ehlers (1993),
they found that individuals high in anxiety sensitivity
(fear of anxiety), reported significantly more episodes
of parental uncontrolled behavior due to drunkenness
and/or anger than individuals with lower levels of anxiety
sensitivity.
AsubsequentretrospectivestudyconductedbyScher
and Stein (in press) investigated the role of childhood ex-
posure to parental threatening, hostile, and rejecting be-
havior in the development of anxiety sensitivity. The de-
gree of childhood exposure to these threetypes of parental
behaviors predicted a significant amount of variance in
anxiety sensitivity levels among a sample of nonclinical
young adults.
A study by MacPherson, Stewart, and McWilliams
(2001)examinedmoredirectlywhetherhighlevelsofanx-
iety sensitivity and its lower order components were as-
sociated with childhood exposure to parental drunken be-
havior. In this study, a sample of 213 university students
was administered a measure designed to detect degree
of childhood exposure to distressing behaviors associated
with parental problem drinking—namely the Children of
Alcoholics Screening Test (CAST; Jones, 1983). CAST
scores predicted anxiety sensitivity over and above par-
ticipant gender and parental alcoholism per se. The mea-
sureemployedby MacPherson et al. (2001), however,was
specific to parental uncontrolled behavior due to alcohol
intake (i.e., the CAST) and did not allow for the assess-
ment of general parental uncontrolled behaviors due to
other reasons (e.g., anger), leaving open the possibility
that exposure to uncontrolled behavior from any source
(not just parental drinking) would cause elevated anxi-
ety sensitivity levels in the child (cf. Scher & Stein, in
press).
Givenevidencefromtheempiricalliteraturesuggest-
ing that repeated exposure to parental uncontrolledbehav-
ior may predispose the child to anxiety (cf. Ehlers, 1993)
and the absence of a retrospective measure in the litera-
ture that allows for the separate quantification of degree
of childhood exposure to parental uncontrolled behavior
from different sources (not just parental drinking), we de-
signed a novel measure of exposure to parental dyscon-
trol. The Dyscontrol Experiences Questionnaire (DEQ)
was designed as a retrospective measure of the degree
of childhood exposure to parental dyscontrol, or loss of
control, associated with specific conditions such as a par-
ent’s drinking, anger, or other negative emotional states
(e.g.,stress,anxiety,depression).ModeledontheLHQde-
signed by Ehlers (1993), the DEQ permits more in-depth
probing of specific types of parental uncontrolled behav-
ior. For example, on the LHQ respondents havethe option
of indicating more than one source of parental dyscon-
trol (i.e., anger, drunkenness, other), as well as more than
one parental model of the behavior. This makes it im-
possible to discern to which source, or model, the fre-
quency of exposure rating refers. On the DEQ, however,
respondents are asked to indicate the frequency of a spe-
cific behavior and the parent who displayed the behavior.
The DEQ was designed to allow for a more precise de-
lineation of source and model of parental dyscontrol and
includes the two specific sources of parental dyscontrol
employed by Ehlers (1993) and Watt et al. (1998)—anger
and drunkenness—as well as allowing for respondents to
cite other sources (e.g., depression, anxiety, etc.). Indeed,
parental dyscontrol can derive from a variety of sources
(e.g., marital conflict) and respondents have the option to
cite any source that is relevant to their own experience.
This paper reports on the first two studies conducted with
the DEQ. Study 1 describes the development of the DEQ
and Study 2 describes the results of initial psychometric
testing.
P1: GDX
Journal of Psychopathology and Behavioral Assessment (JOBA) pp835-joba-464660 April 22, 2003 12:54 Style file version June 25th, 2002
Dyscontrol Experiences Questionnaire 157
STUDY 1: DEVELOPMENT OF DEQ
In Study 1, the DEQ was administered to a student
sample for exploratory factor analysis.
Method
Participants
Four hundred and sixty five (354 female; 111 male)
undergraduate students at two eastern Canada universi-
ties served as research participants in the present study.
Students were enrolled in first- and second-year psychol-
ogy courses and received course credit for their participa-
tion.Randomly selectedclasseswere invitedtoparticipate
and no students refused participation. Respondents’ mean
(SD) age was 20.4 (3.6) years with a range of 17–47 years.
The majority of respondents (67%) reported family-of-
origin incomes in excess of $35,000 per year and 92%
described themselves as single (never married). Further
information on the research participants is provided in
Table I and Fig. 1(a).
Measures
Dyscontrol Experiences Questionnaire (DEQ)
The DEQ invites respondents to answer a series of
questionsabouttheirexperiencesasachild(under18years
of age) with parental (or other’s) uncontrolled behavior.
There is a set of questions for each behavior—Anger,
Drunk, and Other and, for each set, respondents are asked
to indicate the individual to whom it applies. For “Other”
behavior, respondents are asked to specify the behavior
as well as the individual. Research participants in Study
1 responded to all 45 DEQ questions on a relative fre-
quency scale ranging from 0 to 3, where 0 = never;1=
seldom (once or twice a year); 2 = occasionally (three to
six times a year); and 3 = often (more than six times per
year). There were three DEQ composite scales defined as
“ExperienceAnger”(parentaldyscontrolrelatedtoanger),
“Experience Drunk” (parentaldyscontrol related to drink-
ing), and “Experience Other” (parental dyscontrol related
to sources other than anger or drunkenness). For this last
scale, participants were given the example of negative
emotions such as stress, anxiety, and/or dysphoria as a
source for the parental dyscontrol and were provided a
space in which to identify the source on which they were
reporting (e.g., stress due to marital discord). For each
Table I. Descriptive Data on Research Participants in Study Samples 1
and 2
Study Sample 1 Study Sample 2
(N = 465) (N = 533)
Sample size
Males 111 145
Females 354 388
Age (SD) [Range] 20.4 (3.6)[ 17–47] 19.6 (3.5) [17–60]
Annual family income
<$15,000 7% 5%
$15–25,000 10% 11%
$25–35,000 13% 11%
$35–50,000 25% 25%
>$50,000 44% 40%
Marital status
Single 92% 95%
Cohabitating 6% 4%
Other 1% 1%
Exposure to Parental Dyscontrol
DEQ DEQ
Anger
Number reporting 243 [52%] 301 [56%]
Mother 33% 32%
Father 56% 53%
Other 11% 15%
Drunk
Number reporting 96 [21%] 110 [21%]
Mother 15% 71%
Father 74% 71%
Other 11% 16%
Other
Number reporting 86 [18%] 125 [24%]
Mother 62% 68%
Father 22% 27%
Other 16% 5%
scale, respondents were asked to specify the individual
to whom the scale applied and to rate their frequency of
exposure (i.e., how often the parent behaved in an uncon-
trolled manner) for each indicated reason. Then respon-
dents were asked to rate a series of 10 questions pertain-
ing to the behavior of parents in response to each type of
dyscontrolexperience(i.e.,“neglectedhousehold chores,”
“tended to avoid others,” “engaged in risky behaviors”).
We reasoned that the more that a parent responded to
their dyscontrol experiences, the more salient the dyscon-
trol experience would be for the child. We then calculated
composite scores for each of the three scales by multiply-
ing the frequency of exposure score by the mean score
across the 10 “salience” items pertaining to the behavior
of parents in response to dyscontrol in the given context
(e.g., anger; cf. Watt et al., 1998). Although not
P1: GDX
Journal of Psychopathology and Behavioral Assessment (JOBA) pp835-joba-464660 April 22, 2003 12:54 Style file version June 25th, 2002
158 Watt and Stewart
Fig. 1. (a) Frequency of Reported Parental Dyscontrol on DEQ as a function of Source of Dyscontrol in
Study 1. (b) Frequency of Reported Parental Dyscontrol on DEQ as a function of Source of Dyscontrol
in Study 2.
used in scoring, participants also provided informa-
tion on whom exhibited the uncontrolled behavior in
each context (e.g., father, mother), and the predictabil-
ity of the behavior and how fearful the respondent
felt at the time of the exposure (rated on 1–
10 scales).
Results
Factor Analytic Investigation of DEQ Structure
An investigation of DEQ factor structure was con-
ductedbyexploratoryfactoranalysistodeterminewhether
P1: GDX
Journal of Psychopathology and Behavioral Assessment (JOBA) pp835-joba-464660 April 22, 2003 12:54 Style file version June 25th, 2002
Dyscontrol Experiences Questionnaire 159
there was empirical support for separate factors pertain-
ing to the source of the parental dyscontrol experience
(e.g., drunkenness, anger, etc.), as well as to assess the
relative proportion of variance accounted for by each fac-
tor.Principalcomponentsanalyses(PCA)wereemployed,
consistent with Nunnally’s (1978) recommendation that,
in an exploratory analysis, PCA is a reasonable analytic
strategy. PCA uses sums of the observed variables to opti-
mally weight the maximal variability and reliability of the
resultant factor solution (Floyd & Widaman, 1995). The
Kaiser–Meyer–Olkinindexofsamplingadequacywas.86,
indicating that the partial correlations among variables
were small and that the DEQ was appropriate for PCA.
As DEQ dimensions were expected to be intercorrelated,
oblique(Oblimin)rotations wereused,whichallowforin-
tercorrelation among factors (Tabachnick & Fidell,1989).
PCA was performed on 30 of the 45 DEQ items.
Items included were the“salience” items pertaining to the
behavior of parents in response to dyscontrol in a given
context (e.g., anger). Items referring to frequency of ex-
posure to each type of dyscontrol (Items 1, 16, and 31),
who performed each behavior (Items 2, 17, and 32), pre-
dictability of behavior (Items 3, 4, 18, 19, 33, and 34),
and reported magnitude of fear of behavior (Items 15, 30,
and 45) were not included in analysis. An initial estimate
of the number of factors was provided by the commonly
employed eigenvalue greater than one criterion, which in-
dicated support for a six-factor lower-order solution (see
eigenvalues, Table II). A second, more conservative crite-
rion, the scree test of eigenvalues plotted against factors
(Cattell, 1966), suggested that only three dimensions un-
derlietheDEQ.Finally,parallelanalyses(Longman,Cota,
Holden, & Fecken, 1989) were conducted. According to
Zwick and Velicer (1986), as compared to other methods
such as Kaiser’s eigenvalue greater than one and Cattell’s
scree test, parallel analysis is the most accurate rule for
determining the number of components to retain. Paral-
lel analysis is used to identify, statistically, the break in
the scree plot. Consistent with the recommendations of
Longman et al. (1989), the parallel analyses were con-
ducted twice, using mean and 95th percentile eigenval-
ues, respectively. Comparison of the obtained eigenvalues
with both mean and 95th percentile eigenvalues provided
significant support for a three-factor solution.
Table II displays the obliquely-rotated factor load-
ings for the three-factor lower-order solution. Variables
are ordered and grouped by size of loading to facilitate in-
terpretation.Thissolutionaccountedfor54.4%ofthevari-
ance in DEQ item scores. Prior to rotation, the first factor
accounted for 27.3% of the variance,the second factor ac-
countedfor 18.2%,andthethirdfactoraccountedfor9.0%
ofthevariance.Salient loadings wereevaluated asthosein
excess of 0.55 (30% overlapping variance)—a cutoff con-
sidered “good” by Comrey and Lee (1992). The retained
solution had good simple structure (Thurstone, 1947) as
indicated by (1) the relatively large number of items with
salient loadings per factor (i.e., 7–10 salient loadings per
factor); (2) the absence of items with complex loadings;
and (3) the presence of only three hyperplane items (i.e.,
items that failed to load on any of the factors identified;
see Table II). However, because these three items showed
reasonably high loadings on their expected factors (i.e.,
loadings > 0.22), a decision was made to retain these
items in subsequent analyses. Examination of the pattern
of loadings in Table II suggests that Factor I pertains to
dyscontrol due to parental drinking (“Drunk”), Factor II
to childhood exposure to dyscontrol due to parental neg-
ative emotions other than anger (“Other”), and Factor III
to childhood exposure to dyscontrol due to parental anger
(“Anger”). Correlations between the three factors ranged
from0.09(Drunkwith Other) to 0.34 (Drunk withAnger).
STUDY 2: PSYCHOMETRIC TESTING
OF THE DEQ
Method
Research Participants
Fivehundredandthirty-three(388female;145male)
undergraduate students at an eastern Canada university
served as research participants in the second study. Stu-
dents were enrolled in first-year psychology courses and
received course credit for their participation. Randomly
selected classes were invited to participate and no stu-
dents refused participation. Respondents’ mean (SD) age
was19.6(3.5)yearswitharangeof17–60.Themajorityof
respondents (65%) reported family-of-origin incomes in
excess of $35,000 per year and 95% described themselves
as single (never married). Further information on the re-
search participants is provided in Table I and Fig. 1(b).
One hundred and forty-four siblings (89 sisters; 55
brothers) of the student respondents (i.e., 27%) also par-
ticipated in the study by completing a sibling version of
the DEQ as a validity check. PCA performed on DEQ
items with sample two confirmed the same three-factor
lower-order solution obtained with sample one. Follow-
ing rotation with sample two, the first factor (parental
dyscontrol related to drinking) accounted for 35.8% of
the variance, the second factor (parental dyscontrol re-
lated to other negative emotions) accounted for 13.5%,
and the third factor (parental dyscontrol related to anger)
accountedfor10.5%ofthevariance.Correlationsbetween
P1: GDX
Journal of Psychopathology and Behavioral Assessment (JOBA) pp835-joba-464660 April 22, 2003 12:54 Style file version June 25th, 2002
160 Watt and Stewart
TableII. FactorLoadingsforPrincipalComponentsAnalysisandDirectObliminRotationonDyscontrol
Experiences Questionnaire (DEQ) Items
Item number and description F1 Drunk F2 Other F3 Anger
DEQ Subscale Drink
21. Was she/he more apt to neglect household chores? .83
∗
−.05 .09
29. Was she/he apt to engage in risky behaviors? .79
∗
.03 .04
24. Did she/he tend to avoid others? .78
∗
.02 .02
27. Was she/he more likely to ignore you? .78
∗
.02 .03
22. Did she/he cancel social activities? .77
∗
−.02 −.00
23. Did she/he expect more attention? .76
∗
−.04 .12
25. Was she/he more apt to be indulgent? .71
∗
.03 −.17
26. Was she/he more apt to punish you? .71
∗
.02 .09
20. Was she/he more likely to stay home from work? .70
∗
−.06 −.07
28. Was she/he more likely to seek out your company? .64
∗
−.02 .00
DEQ Subscale Other
39. Did she/he tend to avoid others? .00 .89
∗
−.03
36. Was she/he more apt to neglect household chores? −.07 .89
∗
−.00
37. Did she/he cancel social activities? −.02 .88
∗
.02
42. Was she/he more likely to ignore you? −.06 .80
∗
.00
38. Did she/he expect more attention? −.01 .79
∗
.03
41. Was she/he more apt to punish you? −.08 .72
∗
.00
35. Was she/he more likely to stay home from work? −.02 .70
∗
−.00
43. Was she/he more likely to seek out your company? .05 .66
∗
−.03
40. Was she/he more apt to be indulgent? .05 .64
∗
−.09
44. Was she/he apt to engage in risky behaviors? .05 .62
∗
.05
DEQ Subscale Anger
12. Was she/he more likely to ignore you? −.09 −.02 .83
∗
6. Was she/he more apt to neglect household chores? .07 −.04 .80
∗
11. Was she/he more apt to punish you? −.04 −.01 .78
∗
9. Did she/he tend to avoid others? −.02 −.01 .77
∗
7. Did she/he cancel social activities? .01 .08 .76
∗
8. Did she/he expect more attention? .03 .02 .74
∗
14. Was she/he apt to engage in risky behaviors? .12 .04 .68
∗
5. Was she/he more likely to stay home from work? −.01 −.05 .42
10. Was she/he more apt to be indulgent? .06 .23 .29
13. Was she/he more likely to seek out your company? .16 .11 .22
Eigenvalues: 8.19, 5.45, 2.69, 1.32, 1.21, 1.03.
Note. Loadings >.55 (30% of variance) are considered salient and are indicated by asterisks (
∗
). Items
2, 3, 4, 15, 17, 18,19, 31, 32, 33, 34, and 45 are not included because these items arenot used in scoring.
thethreefactorsrangedfrom0.24(Drunk withOtherNeg-
ative Emotions) to 0.44 (Drunk with Anger).
Measures
Parent Threat Inventory (PTI; Scher, Stein, Ingram,
Malcarne, & McQuaid, 2002)
The PTI is a 17-item self-report retrospective mea-
sure of parental threatening behavior in three domains:
Threat of Rejection, Threat of Abandonment, and Threat
of Punishment/Neglect. Respondents rate the degree to
which each item applies to each parent on a 5-point scale
from “Never” to “Very True.” The PTI has demonstrated
reliability and validity and has been found to predict both
anxious and depressive symptomatology, as well as life-
time symptoms of anxiety and depressive disorder (Scher
& Stein, in press). DEQ subscales were predicted to cor-
relate positively with each of the three PTI dimensions.
Parental Bonding Instrument (PBI; Parker, Tupling, &
Brown, 1979)
The PBI is a 25-item self-report retrospective mea-
sure of two dimensions of parenting: (1) Care dimension,
characterized by affection, warmth, and closeness; and
(2) Overprotection dimension, characterized by control,
intrusiveness, and discouragement of autonomy. Respon-
dents rate the degree to which each item applies to each
parent on a 4-point scale from “Very like” my parent to
“Very unlike” my parent. Scores range from 12 to 48 on
P1: GDX
Journal of Psychopathology and Behavioral Assessment (JOBA) pp835-joba-464660 April 22, 2003 12:54 Style file version June 25th, 2002
Dyscontrol Experiences Questionnaire 161
the Care dimension and from 13 to 52 on the Overpro-
tection dimension. The PBI is perhaps the most widely
used retrospective measure of parenting and has demon-
strated reliability and validity (Parker et al., 1979). Low
scoresontheCare dimension and high scores on the Over-
protection dimension have been linked to many forms of
psychological distress, including anxiety (Rapee, 1997).
The DEQ subscales were expected to correlate negatively
with PBI Care scores and correlate positively with PBI
Overprotection scores.
State–Trait Anxiety Inventory—Trait Scale (STAI-T)
The STAI-T (Spielberger, Gorsuch, Lushene, Vagg,
&Jacobs,1983) is a 20-item self-report measure on which
participants rate how anxious they “generally” feel on 4-
point Likert-type scales of relative frequency that range
from almost never (1) to almost always (4). The STAI-T
contains nine reverse-scored, anxiety-absent items. Total
STAI-T scores can range from 20 to 80. Considerable ev-
idence supports the good psychometric properties of the
STAI-T including high reliability, and good construct va-
lidity in predicting anxiety in response to stressors (see
review by Spielberger et al., 1983). The STAI-T was in-
cludedinthepresentstudyasameasureofthefrequencyof
general anxiety symptoms (cf. MacPherson et al., 2001).
Beck Depression Inventory—II (BDI-II; Beck, Steer, &
Brown, 1996)
The BDI-II is a 21-item self-report instrument
for measuring the severity of depression in adults and
adolescents (aged 13 years and older) consistent with
the DSM-IV (American Psychiatric Association [APA],
1994). The BDI-II reflects both cognitive and affective, as
well as somatic and vegetative symptoms of depression.
Each item is rated on a 4-point scale ranging from 0 to
3. The BDI-II is one of the most widely accepted instru-
ments for assessing severity of depression in diagnosed
patients and for detecting possible depression in normal
populations. Its psychometric properties have been well-
demonstrated (Beck et al., 1996).
Results
Internal Consistency of DEQ scales
Internal consistencies (Cronbach’s alphas) for the
multi-item DEQ scales were calculated based on partic-
ipant responses (N = 400) to respective subscale items.
Internal reliability estimates ranged from 0.89 for the ten
items on the “Anger” scale (all but Items 1, 2, 3, 4, and
15) to 0.93 for the “Drunk” scale (all but Items 16, 17, 18,
19, and 30), and “Other” scale (all but Items 31, 32, 33,
and 34).
As a preliminary test of validity, correlations were
performed between the three DEQ subscales of the 140
student–sibling pairs. One-tailed tests indicated that sib-
lings’ and students’ responses were significantly posi-
tively correlated with one another in each case: r = .40,
p <.001 for Anger; r = .44, p <.001 for Drunk; and
r = .43, p <.001 for Other (see Table III). As a base-
line for comparing these validity coefficients,correlations
were also performed between the three DEQ subscales of
the 140 sibling respondents with a random sample of 140
students who had not had siblings respond (and who in-
deed had siblings themselves according to their responses
to a demographic measure). The correlations were not
significant for any of the three subscales: r =−.03, ns
for Anger; r = .07, ns for Drunk, and r =−.03, ns for
Other. The fact that the correlations were clearly larger
for the student–sibling pairs than for this random com-
parison sample provides further evidence of the DEQ’s
validity.
According to Meier (1994), construct validity
answers the question: “Does the test measure what it was
intended to measure?” and is the type of validity upon
which other types of validity rest. One way of demon-
strating construct validity is to examine intercorrelations
between the construct of interest (in this case, parental
dyscontrol behavior) and conceptually related but distinct
constructs. Such intercorrelations should be moderately-
sized and in direction(s) predicted by theory. To test the
construct validity of the DEQ, intercorrelations between
DEQ subscales and subscales of the Parent Threat Inven-
tory (PTI: Scher & Stein, in press) and Parental Bonding
Instrument (PBI; Parker et al., 1979) were examined. The
PTI and PBI were selected for the present study as two
retrospective measures of parental behavior (e.g., threat,
rejection, lack of care) associated with psychological
distress. The DEQ taps similar (e.g., rejection, punish-
ment) but also other behavior as a function of parental
dyscontrol.
As predicted, DEQ subscales correlated positively
witheachofthePTIsubscales. Correlationsbetween DEQ
subscales and PTI Rejection ranged from .20 to .48; cor-
relations between DEQ subscales and PTI Abandonment
rangedfrom.23to.45andcorrelationsbetweenDEQsub-
scales and PTI Punishment rangedfrom.20 to .54. As pre-
dicted, DEQ subscales correlated positively with the PBI
Overprotection subscale with correlations ranging from
.21 to .43 and correlated negatively with the PBI Care
P1: GDX
Journal of Psychopathology and Behavioral Assessment (JOBA) pp835-joba-464660 April 22, 2003 12:54 Style file version June 25th, 2002
162 Watt and Stewart
Table III. Pearson Product Moment Correlations Between Student and Sibling Reporting
of Parental Dyscontrol due to Anger, Drinking, or Other Negative Emotions
Student and Sibling DEQ Subscales
Student Sibling
Anger Drunk Other Anger Drunk Other
Student Anger 1.00
Student Drunk .50
∗∗∗
1.00
Student Other .35
∗∗∗
.24
∗∗∗
1.00
Sibling Anger .40
∗∗∗
.25
∗∗
.29
∗∗
1.00
Sibling Drunk .28
∗∗
.44
∗∗∗
.26
∗∗
.50
∗∗∗
1.00
Sibling Other .37
∗∗∗
.18
∗
.43
∗∗∗
.50
∗∗∗
.14
∗
1.00
∗
p <.05.
∗∗
p <.01.
∗∗∗
p <.001 one-tailed.
subscale with correlations ranging from −.15 to −.38.
The correlation matrix is presented in Table IV.
Predictive validity refers to the degree of correlation
between the results of a given measurement and an ex-
ternal criterion (Frankfort-Nachmias & Nachmias, 1996).
Given that the DEQ was developed in order to better un-
derstand the significance of parental dyscontrol behavior
inthedevelopmentofpsychologicaldistress,itwasimpor-
tant to assess its ability to predict psychological distress.
To test the predictive validity of the DEQ, intercorrela-
tions between DEQ subscales and the Beck Depression
Inventory-II (BDI-II: Beck et al., 1996) and State–Trait
AnxietyInventory–Trait(STAI-T;Spielbergeretal.,1983)
were examined. As predicted, DEQ subscales correlated
positively with theSTAI-T with correlations ranging from
.19 to .28 and correlated positively with the BDI-II with
correlations ranging from .17 to .36 (see Table IV).
To further examinetheincremental validityofthe ex-
posure to parental dyscontrol as assessed by DEQ scales
over and above exposure to negative parental behavior,
per se, DEQ scores were entered into a set of hierarchi-
cal multiple regressions predicting anxiety (STAI-T) and
depression (BDI-II), respectively. In each case, scores on
Table IV. Pearson Product Moment Correlations Between DEQ
Subscales and Subscales of the PTI, PBI, STAI-T, and BDI-II
DEQ Subscales
Anger Drunk Other
PTI—Rejection .48
∗∗∗
.34
∗∗∗
.20
∗∗∗
PTI—Abandonment .45
∗∗∗
.44
∗∗∗
.23
∗∗∗
PTI—Punishment/Neglect .54
∗∗∗
.54
∗∗∗
.23
∗∗∗
PBI—Care −.38
∗∗∗
−.17
∗∗∗
−.15
∗∗
PBI—Overprotection .43
∗∗∗
.21
∗∗∗
.21
∗∗∗
STAI-T .28
∗∗∗
.19
∗∗∗
.19
∗∗∗
BDI-II .36
∗∗∗
.25
∗∗∗
.17
∗∗∗
∗∗
p <.01.
∗∗∗
p <.001 one-tailed.
the PBI and PTI were entered into the regression equation
in the first block (as measures of negative parental behav-
iors) followed by the three DEQ composite scores in the
second block.
In the regression predicting STAI-T scores, the fi-
nal model including PTI, PBI, and DEQ scores explained
a significant 21.9% of the variance in anxiety scores,
F(8, 445) = 15.60, p <.001. In the first block, PTI and
PBI scores together predicted a significant 20.2% of the
variance in anxiety scores, F(5, 448) = 22.62, p <.001.
The addition of the three DEQ subscale scores in the sec-
ondblock significantly improvedthepredictionof anxiety
scores on the STAI-T, F
inc
(3, 445) = 3.32, p <.05, indi-
cating incremental validity for the DEQ. In the final equa-
tion, significant univariate predictors of increased anxiety
were lack of care on the PBI (β = .12, t = 1.97, p =
.05), overprotection on the PBI (β =−.25, t =−4.06,
p <.001), and experience other on the DEQ (β = .09,
t = 2.06, p <.05). There was also a marginal indepen-
dent effect of experience drunk on the DEQ (β = .09,
t = 1.71, p = .09) in predicting anxiety (see Table V).
Table V. Summary of Hierarchical Regression Analysis for
Measures of Parental Behavior (DEQ, PBI, PTI) in Predicting
Anxiety (STAI-T) scores
Variable BSEB β
PBI—Care 0.09 0.05 0.12
∗
PBI—Overprotection −0.22 0.05 −0.25
∗∗∗
PTI—Rejection 0.03 0.09 0.02
PTI—Abandonment −0.00 0.08 −0.01
PTI—Punishment/Neglect 0.16 0.13 0.08
DEQ—Anger 0.01 0.32 0.01
DEQ—Drunk 0.58 0.34 0.09
t
DEQ—Other 0.50 0.24 0.09
∗
Note.
t
= trend. The PBI is keyed in such a way that a high
score indicates lack of parental care and low overprotection.
∗
p <.05.
∗∗
p <.01.
∗∗∗
p <.001.
P1: GDX
Journal of Psychopathology and Behavioral Assessment (JOBA) pp835-joba-464660 April 22, 2003 12:54 Style file version June 25th, 2002
Dyscontrol Experiences Questionnaire 163
TableVI. SummaryofStepwiseHierarchicalRegressionAnal-
ysis for Measures of Parental Behavior (DEQ, PBI, PTI) in
Predicting Depression (BDI-II) scores
Variable BSEB β
PBI—Care −0.08 0.04 −0.11
t
PBI—Overprotection 0.14 0.05 0.18
∗∗
PTI—Rejection 0.07 0.08 0.05
PTI—Abandonment −0.02 0.07 −0.02
PTI—Punishment/Neglect 0.24 0.12 0.14
∗
DEQ—Anger 0.52 0.28 0.11
t
DEQ—Drunk 0.50 0.30 0.09
t
DEQ—Other 0.03 0.21 0.00
Note.
t
= trend. The PBI is keyed in such a way that a high
score indicates lack of parental care and low overprotection.
∗
p <.05.
∗∗
p <.01.
∗∗∗
p <.001.
In the regression predicting BDI-II scores, the final
model including PTI, PBI, and DEQ scores explained a
significant 25.4% of the variance in depression scores,
F(8, 445) = 18.95, p <.001. In the first block, PTI and
PBI scores together predicted a significant 23.7% of the
variance in depression scores, F(5, 448) = 27.86, p <
.001. The addition of the three DEQ subscale scores in
the second block significantly improved the prediction
of depression scores on the BDI-II, F
inc
(3, 445) = 3.38,
p <.05, indicating incremental validity for the DEQ. In
the final equation, significant univariate predictors of in-
creased depression were overprotection on the PBI (β =
−.18, t =−2.87, p = .01) and punishment on the PTI
(β = .14, t = 2.12, p <.05). There was also a marginal
independent effect of lack of care on the PBI (β = .11,
t = 1.88, p = .06), and both experience anger (β = .11,
t = 1.86, p = .06) and experience drunk on the DEQ
(β = .09,t = 1.67, p = .09)inpredictingdepression(see
Table VI).
Discussion
This paper reports on the development and initial
psychometric testing of the DEQ in two student samples.
The DEQ is a 45-item retrospective measure of childhood
exposure to parental dyscontrol, or loss of control, due
to specific conditions such as a parent’s drinking, anger,
or other negative emotional states (e.g., stress, anxiety,
depression). Given the findings of earlier studies (Ehlers,
1993; Watt et al., 1998), the authors developed the DEQ
in an attempt to better understand the role of loss of con-
trol behaviorinthe etiology of psychological distress.The
conceptofcontrolisconsideredtobeanimportantvariable
in determining individual susceptibility to psychological
distress (e.g., anxiety, depression). Early childhood expe-
rience with uncontrollable or unpredictable stimuli may
contribute to low perceptions of control and predispose
the individual to psychopathology in adulthood.
Certainparentingpracticeshavealsobeenimplicated
in the development of adult psychopathology. For exam-
ple, low levels of parental care and high levels of parental
overprotection have been linked to a number of adult psy-
chological dysfunctions including anxiety and depression
(see Scher, Stein, Ingram, Malcarne, & McQuaid, 2002),
and childhood exposure to parental threatening, hostile,
and rejecting behavior has been linked to the develop-
ment of anxiety sensitivity (Scher & Stein, in press). It
was theorized that repeated childhood exposure to a par-
ent acting in an uncontrolled fashion could diminish a
child’s sense of control over their environment, resulting
in psychological vulnerability to distress. According to
social learning theory (Bandura, 1986), the more salient
the parental dyscontrol behavior, the more potent such a
learning experience would be for a child.
The DEQ was modeled on the Learning History
Questionnaire (LHQ) developed by Ehlers (1993) to ex-
plore the relationship of childhood learning experiences
and panic. Whereas the LHQ included one question re-
garding parents behaving in an uncontrolled way due to
being angry or drunk, the DEQ sought to examine this be-
haviorinmoredepthbyprobingspecific“salient”parental
behaviors exhibited when the parent was drinking, angry
or experiencing other negative emotions. The DEQ was
firstadministeredto a student sample for exploratory prin-
cipal components analysis. A three-factor solution was
confirmed and adopted for further psychometric investi-
gation with a second student sample. Cronbach’s alpha
coefficients ranging from .87 to .93 revealed acceptable
internal consistency reliability for the entire measure and
for each subscale.
The construct validity of the DEQ was assessed by
examining intercorrelations between the DEQ and two re-
lated retrospective measures of parental behavior (i.e., the
PTI and PBI). As predicted, scores on the DEQ subscales
correlated significantly and positively with the PTI sub-
scales of Threat of Rejection, Threat ofAbandonment and
Threat of Punishment/Neglect. In other words, students
who reported greater exposure to parental dyscontrol due
to either drinking, anger and/or other negative emotions
also reported more parental threatening behavior in child-
hood.Also, aspredicted,scoresonthe DEQsubscalescor-
relatedpositively with the parental Overprotection dimen-
sion and negatively with the parental Care dimension of
the PBI. In other words, students who reported experienc-
ing parenting characterized by control, intrusiveness and
discouragementof autonomyreportedmore childhoodex-
posure to parental dyscontrol, whereas students who re-
ported experiencing parenting characterized by affection,
P1: GDX
Journal of Psychopathology and Behavioral Assessment (JOBA) pp835-joba-464660 April 22, 2003 12:54 Style file version June 25th, 2002
164 Watt and Stewart
warmthand closeness reported less childhood exposureto
parental dyscontrol.
TheDEQwasdevelopedinordertobetterunderstand
the significance of parental behavior, specifically parental
dyscontrol related to drinking, anger or other negative
emotions, in the development of psychological distress. It
wasimportant,therefore,toassessitsabilitytopredictpsy-
chological distress as indicated by students’ scores on the
STAI-T and BDI-II. As expected, students who reported
greater exposure to parental dyscontrol in childhood re-
vealed greater psychological distress in young adulthood
as indicated by measures of their anxious and depressive
symptomatology.
As a further test of the construct validity of the DEQ,
hierarchical regression analyses were conducted to assess
whether the DEQ would add to the prediction of psycho-
logical distress over and above the contribution of the PTI
and PBI. Indeed, analyses revealed that the DEQ did add
to the prediction supporting the authors’ contention that
it measures something more than rejection, abandonment,
andotheraversiveexperiences.Whereasthisfindingis en-
couraging, further research is required to determine if it is
the uncontrollability of the parental behavior per se or the
aversive nature of the negative parental behavior that con-
tributesto the distress. Chorpita and Barlow(1998) would
suggest that it is the uncontrollability, having equated the
uncontrollability of both aversives and reinforcement on
the development of anxiety. The addition of a perceived
control or perceived predictability measure in future stud-
ies might serve to better elucidate this matter.
An important strength of the current psychometric
assessment of the DEQ was the availability of sibling val-
idation data in the second student sample. Data on 140
student–sibling pairs allowed for a comparison of report-
ing of parental dyscontrol behavior between siblings. In-
deed, data revealed moderate correlations between sib-
ling and student reporting of parental dyscontrol with the
strength of their association approaching 20%.
Overall, the DEQ shows promise as a reliable and
valid tool with which to evaluate the certain possible de-
velopmental antecedents of psychological distress among
adults. However, certain limitations need to be addressed.
First, the DEQ asks participants to specify one individual
who behaved in an uncontrolled manner due to drinking,
angerorothernegativeemotionandtorespondtoinquiries
with that person in mind. This need to specify one indi-
vidual does not allow for the possibility that more than
one caretaker (e.g., both parents) may have behaved in an
uncontrolledmannerforsimilarreasons.Itmightbeadvis-
able to query participants about the number of significant
attachment figures and allow for responding to more than
one if applicable.
AsecondlimitationoftheDEQisitsbroadretrospec-
tive nature. For example, it asks participants to respond to
inquiries about experiences prior to age 18. This does not
take into consideration how parental behavior might be
perceived differently at different developmental levels or
that parental behavior may change over the course of re-
spondents’ childhood and adolescence. What is perceived
to be uncontrolled behavior in early childhood may not
seem so uncontrolled in adolescence. To increase the sen-
sitivity of the DEQ, researchers and clinicians may wish
to query respondents regarding the time period(s) during
which particular behaviors occurred.
A third limitation of the DEQ relates to the criti-
cism of retrospective measures in general. Retrospective
reports are often criticized as being subject to biases and
distortions due to selective memory of past events; the in-
fluences of current attitudes, behaviors, and experiences;
and demand characteristics of the study itself. For exam-
ple,scoresontheDEQsubscalesmayrepresentanover-or
underestimate of parental dyscontrol as a result of current
relations with the parentrather than reflecting actual early
experiences. In the present study, an attempt was made to
validate students’ retrospective reports of their childhood
experiences by inviting their siblings to complete com-
parable versions of the DEQ. Indeed, brothers’ and sis-
ters’ responses were found to provide statistically signifi-
cant corroboration of the responses of their siblings. This
finding supports Brewin, Andrews, and Gotlib’s (1993)
conclusion that “provided that individuals are questioned
about the occurrence of specific events or facts that they
were sufficiently old and well placed to know about, the
central features of their accounts are likely to be reason-
ably accurate” (p. 94).
Finally, the usefulness of the DEQ with clinical sam-
ples remains to be examined. Given that a stronger re-
lationship between parental dyscontrol and psychologi-
cal dysfunction might be expected in clinical samples as
compared to student samples, normative data on clinical
samples of men and women, as well as tests of reliability
and validity in these samples should be collected.
ACKNOWLEDGMENTS
This research was supported in part by a grant from
the Social Sciences and Humanities Research Council
(SSHRC) awarded to the first author.
REFERENCES
American Psychiatric Association. (1994) Diagnostic and statistical
manual of mental disorders (4th ed.). Washington, DC: Author.
P1: GDX
Journal of Psychopathology and Behavioral Assessment (JOBA) pp835-joba-464660 April 22, 2003 12:54 Style file version June 25th, 2002
Dyscontrol Experiences Questionnaire 165
Bandura, A. (1986). Social foundations of thought and action: A social
cognitive theory. Upper Saddle River, NJ: Prentice-Hall.
Barlow, D. H. (1988). Anxiety and its disorders. New York: Guilford
Press.
Barlow,D.H.,Chorpita,B.F.,&Turovsky,J.(1996).Fear,panic,anxiety,
anddisordersofemotion.InD.A.Hope(Ed.),NebraskaSymposium
on Motivation: Perspectives on anxiety, panic, and fear Vol. 43.
(pp. 251–328). Lincoln: University of Nebraska Press.
Beck, A. T., Steer, R. A., & Brown, G. K. (1996). Manual for the Beck
Depression Inventory—second edition (BDI-II). San Antonio: The
Psychological Corporation.
Brewin, C. R., Andrews, B., & Gotlib, I. H. (1993). Psychopathology
and early experience: A reappraisal of retrospective reports. Psy-
chological Bulletin, 113, 82–98.
Carton,J.S.,&Nowicki,S.(1994).Antecedentsofindividualdifferences
inlocusofcontrolreinforcement:Acriticalreview. Genetic, Social,
and General Psychology Monographs, 120, 31–81.
Cattell, R. B. (1966). Thescree test for the number of factors. Multivari-
ate Behavioral Research, 1, 245–276.
Chorpita, B. F. (2001). Control and the development of negative emo-
tion. In M. W. Vasey & M. R. Dadds (Eds.), The developmental
psychopathologyof anxiety (pp.112–142). New York: Oxford Uni-
versity Press.
Chorpita, B. F., & Barlow, D. H. (1998). The development of anxiety:
Theroleofcontrolintheearlyenvironment.PsychologicalBulletin,
124, 3–21.
Comrey, A. L., & Lee, H. B. (1992). A first course in factor analysis
(2nd ed.). Hillsdale, NJ: Erlbaum.
Ehlers, A. (1993). Somatic symptoms and panic attacks: A retrospective
study of learning experiences. Behaviour Research and Therapy,
31, 269–278.
Floyd, F. J., & Widaman, K. (1995). Factor analysis in the development
and refinement of clinical assessment instruments. Psychological
Assessment, 7, 286–299.
Frankfort-Nachmias, C., & Nachmias, D. (1996). Research methods in
the social sciences (5th ed.). New York: St. Martin’s Press.
Gray, J.A., & McNaughton, N. (2000).The neuropsychology of anxiety:
Anenquiryintothefunctionsofthesepto-hippocampalsystem.New
York: Oxford University Press.
Jones, J. W. (1983). The Children of Alcoholics Screening Test (CAST).
Chicago, IL: Camelot Unlimited.
Longman, R. S., Cota, A. A., Holden, R. R., & Fecken, G. C. (1989).
A regression equation for the parallel analysis criterion in princi-
pal components analysis: Mean and 95th percentile eigenvalues.
Multivariate Behavioral Research, 24, 59–69.
MacPherson, P. S. R., Stewart, S. H., & McWilliams, L. A. (2001).
Parental problem drinking and anxiety disorder symptoms in adult
offspring:Examiningthemediating role of anxietysensitivitycom-
ponents. Addictive Behaviors, 26, 917–934.
Meier, S. T. (1994). The chronic crisis in psychological measurement
and assessment: A historical survey. San Diego: Academic Press.
Mineka, S., & Kelly, K. (1989). The relationship between anxiety, lack
of control, and loss of control. In A. Steptoe & A. Appels (Eds.),
Stress, personal control, and worker health. Brussels: Wiley.
Nunnally, J. C. (1978). Psychometric theory (2nd ed.). New York:
McGraw-Hill.
Parker, G.(1983). Parentaloverprotection: A risk factor in psychosocial
development. New York: Grune & Stratton.
Parker, G., Tupling, H., & Brown, L. B. (1979). A Parental Bonding
Instrument. British Journal of Medical Psychology, 52, 1–10.
Rapee, R. M. (1997). Potential role of childrearing practices in the de-
velopment of anxiety and depression. Clinical Psychology Review,
17, 47–67.
Scher, C. D., Stein, M. B., Ingram, R. E., Malcarne, V. L., & McQuaid,
J. R. (2002). TheParentThreatInventory: Development,reliability,
and validity. Child Abuse and Neglect, 26, 207–225.
Schneewind,K. A. (1995). Impact offamilyprocessesoncontrolbeliefs.
In A. Bandura (Ed.), Self-efficacy in changing societies (pp. 114–
1148). New York: Cambridge University Press.
Spielberger, C.D.,Gorsuch, R. L., Lushene, R. E., Vagg, P. R., & Jacobs,
G. A. (1983). Manual for the State–Trait Anxiety Inventory. Palo
Alto, CA: Consulting Psychologists Press.
Tabachnick, B. G., & Fidell, L. S. (1989). Using multivariate statistics
(2nd ed.). New York: HarperCollins.
Thurstone, L. L. (1947). Multiple-factor analysis: A development and
expansion of the vectors of mind. Chicago: University of Chicago
Press.
Watt, M., Stewart, S. H., & Cox, B. J. (1998). A retrospective study of
the origins ofanxiety sensitivity. Behaviour Research and Therapy,
36, 505–525.
Whitehead, W. E., Busch, C. M., Heller, B. R., & Costa, P. T. (1986).
Social learning influences on menstrual symptoms and illness be-
haviour. Health Psychology, 5, 13–23.
Zwick, W. R., & Velicer, W. F. (1986). Comparison of five rules for
determining the number of components to retain. Psychological
Bulletin, 99, 432–442.
