You are on page 1of 23

AGGRESSIVE BEHAVIOR

Volume 27, pages 79101 (2001)

Measuring Aggression: Self-Reports, Partner


Reports, and Responses to Provoking Scenarios
Daryl B. OConnor,1* John Archer,2 and Frederick W.C. Wu1
1

Department of Endocrinology, Manchester Royal Infirmary, University of Manchester,


Manchester, England
2
Department of Psychology, University of Central Lancashire, Preston, Lancashire, England

.. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. ..
This paper describes three studies. The first is concerned with the relationship between the Aggression Questionnaire (AQ) and a peer-report version of the AQ (AQ-P). The results (from a sample of
77 men aged 19 to 55 years) showed agreement between the two versions, thus providing evidence for
the congruent validity of the AQ-P, and good internal consistency was shown for the subscales of the
AQ and AQ-P. The second study involved the development, piloting, and validation of a vignettebased assessment of aggressive behavior, the Aggressive Provocation Questionnaire. The third concerned the use of the finalised Aggressive Provocation Questionnaire in a sample of 130 men (aged 17
to 54 years). The results show that prior self-reported aggression (measured by the AQ) predicted
responses to provoking scenarios. Younger men (< 28 years) showed an association between physical
aggression and the aggressive action scale (of the Aggressive Provocation Questionnaire) in response
to provoking scenarios. Older men (> 28 years) showed an association between physical aggression
and the assertive action scale. The findings are discussed in relation to aggression measurement and
changes in aggression with age. Aggr. Behav. 27:79101, 2001. 2001 Wiley-Liss, Inc.

.. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. ..
Key words: Aggression Questionnaire; aggression; assertiveness; age; anger; hostility

INTRODUCTION

Real-life incidents of human aggression are difficult to record and to measure as they happen.
Behavioral scientists therefore fall back on a number of indirect methods, which can broadly be
divided into (1) laboratory procedures that simulate the sort of provoking situations that generate aggressive actions in the outside world [Berkowitz, 1989; Berkowitz and Donnerstein, 1982;
Anderson and Bushman, 1997], and (2) reports of real-life aggression by the protagonists. The
second of these can be subdivided into accounts of specific events or acts (e.g., the Conflict
Tactics Scale) [Straus, 1979; Straus et al., 1996] and measures of peoples propensity to act
aggressively (e.g., the Aggression Questionnaire) [Buss and Perry, 1992]. To this list we can add
a further, less commonly used measure, involving responses to hypothetical scenarios involving
provocationin other words, what people say they would do under situations which are likely
to evoke anger [e.g, van Goozen et al., 1994; Nisbett and Cohen, 1996].

Grant sponsor: World Health Organization; Grant number 96374.


Correspondence to: Dr Daryl B. OConnor, School of Psychology, University of Leeds, Leeds LS2 9JT.
E-mail: daryloc@psychology.leeds.ac.uk
Received 1 November 1999; amended version accepted 25 January 2000

2001 Wiley-Liss, Inc.

80

OConnor et al.

Of these various measures, the Aggression Questionnaire (AQ) [Buss and Perry, 1992] has
been widely used in recent studies involving trait aggressiveness [Bushman, 1995; Harris, 1995;
Harris et al., 1996; Meesters et al., 1996; Williams et al., 1996]. It is a relatively brief Likert-type
response measure that involves four subscales measuring physical and verbal aggression, anger,
and hostility. Confirmatory factor analyses have shown a reasonable fit with the original four
(correlated) factor model in most studies [Archer et al., 1995; Buss and Perry, 1992; Harris, 1995;
Meesters et al., 1996], Williams et al. [1996] being an exception. The scales generally show
acceptable internal consistencies [Archer and Haigh, 1997a, 1997b; Archer et al., 1995; Buss and
Perry, 1992]. Scores on the subscales show consistent intercorrelations [Archer et al., 1995; Archer and Haigh, 1997a, 1997b; Buss and Perry, 1992; Harris, 1996], and they show consistent
differences between men and women, with men showing significantly higher levels of both physical
and verbal aggression [Buss and Perry, 1992; Archer et al., 1995; Archer and Haigh, 1997a,
1997b; Meesters et al., 1996; Williams et al., 1996]. Physical aggression scores were higher among
prisoners convicted for violent offenses than among those convicted for nonviolent offenses in
one study [Archer and Haigh, 1997b] but not in a second [Williams et al., 1996].
Buss and Perry [1992] also used a form of peer nomination, based on the methods found in
studies of childrens aggression [e.g., Boulton and Smith, 1994; Huesman et al., 1984] to investigate the properties of their scale. This involved students from the same college fraternities
being asked to nominate aggressive individuals in response to four questions, each representing
an AQ subscale. Correlations with AQ scores ranged from .20 for verbal aggression to .45 for
physical. Although rather low, the correlation for physical aggression was similar to that for
extraversion obtained in the same way.
Other studies have found higher correlations, of around .60, between self-ratings of personality and those supplied by spouses [McCrae and Costa, 1990]. Similarly, acts of physical aggression to partners show correlations between self- and partner reports of around .50 [Moffitt et al.,
1997; Archer, 1999]. Undertaking a similar comparison between self- and partner reports for the
AQ is the aim of the first study. These data will also allow investigation of the internal reliability
of the subscales for both measures and explore the congruent validity for the Aggression Questionnaire-Partner Version (AQ-P) with the AQ.
The aim of the second study is to develop a scenario measure of aggression for men, the
Aggressive Provocation Questionnaire, and to compare responses on this to self-reports on the
AQ, in this way allowing the construct validity of the new measure to be investigated. Scenario
measures involve assessing a persons self-reported disposition to act aggressively in the face of
a standard set of provoking situations. Peoples responses are hypothetical ones, and therefore
may differ from reality for several reasons, such as the relative absence of provoking situations
from their lives, and the inhibition of risky aggressive responses when the cost to benefit ratio is
high [Bjrkqvist, 1994]. Talk is cheap, and imagination is even cheaper. Previously, the scenario
method has only been used patchily for assessing adult aggression, although there are a number
of studies involving children [e.g., Leifer and Roberts, 1972; Deluty, 1983]. van Goozen et al.
[1994] developed a scenario measure to assess womens responses to provocation, but this was
mainly concerned with the emotions that the situation generated. Nisbett and Cohen [1996] used
a single scenario to assess aggressive dispositions among men from the U.S. North or South,
who were either insulted or not. The present paper involves a series of scenarios designed for
use with a cross-section of young to early middle-aged men. It describes the derivation of this
measure, its subsequent modification, and comparison with the AQ.
The populations involved in the present studies were men, aged between 17 and 55 years, and
were drawn from several sources, rather than relying only on samples of undergraduates. Several studies, the age ranges of which extended beyond those of the typical undergraduate sample,

Measuring Aggression

81

have reported an inverse association between self-reported aggression and age. The correlations
for physical aggression ranged between -.12 and -.39, for verbal aggression between -.17 and
-.29, for anger between -.11 and -.31, and for hostility between -.21 and -.29 [Archer, 2000;
Archer and Haigh, 1997a, 1997b; Broadbent, 1999; Harris, 1996]. In view of these associations,
age was also included as a variable to be investigated in the present study.
To summarize, the present paper describes three studies: The first is concerned with exploring
the relationship between the AQ and a modified peer report version of the AQ. The second
outlines the development, piloting, and validation of a vignette-based assessment of aggressive
behavior, the Aggressive Provocation Questionnaire. The third concerns the use of the finalised
Aggressive Provocation Questionnaire, and tests whether prior self-reported aggressive behavior (measured by the AQ) predicts responses to provoking scenarios. All three studies address
the association of aggression measures with age.
STUDY 1: AN ASSESSMENT OF SELF- AND PARTNER RATINGS ON THE
AGGRESSION QUESTIONNAIRE

As outlined above, the first study involved an assessment of the agreement between self- and
partner assessments on the AQ. For the purpose of this study, the original AQ, which was general in its referents and can be characterised as a trait measure, was modified slightly to refer to
a specific time period in the recent past (i.e., last 4 weeks). This was undertaken to make it more
useful for measuring within-subject variation, which forms an important consideration for studies involving situational or biological influences on peoples aggressive dispositions and actions. It is interesting to note, however, that the unmodified AQ did show within-participant
increases in physical and overall aggression following 6 weeks of treatment with testosterone
cypionate [Kouri et al., 1995].
It was predicted that the strongest relationship between self- and partner reports would be
found for the physical subscale, because items of physical aggression are expected to have more
salience than verbal aggression and anger or hostility. In addition, the internal reliability of both
scales will be evaluated and the congruent validity assessed for the AQ-P.
Method

Questionnaires. The AQ [Buss and Perry, 1992] consists of 29 items concerning self-reports
of behavior and feelings, which are completed along a five-point scale (5: very often applies to
me to 1: never or hardly applies to me). Two items, 4 (I can think of no good reason for ever
hitting a person) and 19 (I am an even tempered person) are reverse-scored.
There are four subscales, physical (9 items), verbal (5 items), anger (7 items), and hostility (8
items). The first two are concerned with behavior (e.g., I have threatened people I know, and
I often find myself disagreeing with people), and the other two with feelings (e.g., anger: I
have trouble controlling my temper; hostility: I am sometimes eaten up with jealousy).
The AQ-P is an adapted version of the AQ [Buss and Perry, 1992] where the male respondents
partner is asked to rate him in relation to each of the 29 items. For example, I get into fights a
little more than the average person becomes My partner gets into fights a little more than the
average person.
Procedure. Two hundred and fifty questionnaires were distributed to a heterogeneous sample
of men and their partners derived from three sources: (1) a large family planning clinic in the
locality; (2) non-academic staff employed at the University of Manchester; (3) students from the
University of Central Lancashire (Preston) and Strathclyde University (Glasgow). Each couple

82

OConnor et al.

received a pack containing a male envelope and a partner envelope, with a covering letter
outlining the nature of the study and providing precise instructions on how to answer the questionnaires. This included an exclusion criterion for couples engaged in their present relationship
for less than 1 year. The importance of no collusion was also emphasized. The only demographic details collected were the age of the male respondent and length of present relationship.
In the family planning clinic, the questionnaires were distributed by the practice nurse, to university staff at designated meeting areas, and to the students at the end of lectures. Therefore,
nonrespondents could not be followed up. Seventy-seven men returned their questionnaires, 66
of whom had a partner, giving a response rate of 31% and 26%, respectively. The response rate
is low relative to other frequently cited American undergraduate studies. However, it should be
borne in mind that respondents in such studies are rewarded with course credits, which would
have a significant impact on improving the response rate. It is also more difficult to obtain
questionnaires from two people than from one, and male respondents are typically less willing
to fill in questionnaires than females are (see, in relation to partner aggression, Archer, 1999). In
addition, because nonrespondents could not be followed up, reminder letters were not possible.
Ages ranged from 19 to 55 years (mean = 33.12; SD = 9.88).
Statistical methods. Descriptive statistics were calculated for each of the variables and
Pearsons product moment correlation coefficients were used to investigate the relationship between self- and partner reports. Cronbachs alphas were calculated to assess the internal reliability of all the subscales. Congruent validity was evaluated by correlating scores from the AQ-P
with that of the previously validated AQ. All data were analysed with SPSS for Windows.
Results

Properties of the AQ and the AQ-P. Overall, the 29-item AQ showed a Cronbachs alpha of
.92; for the physical scale the value was .86; the verbal .77; anger .89; and hostility .86. Cronbachs
alpha for the total AQ-P was .93; for the physical scale .80; verbal .83; anger .86; and hostility
.84. All values are within acceptable boundaries [Cortina, 1993].
Descriptive statistics. Table 1 shows the descriptive statistics for the AQ and the AQ-P. Scores
for the present respondents were notably lower than those reported elsewhere for men [Archer
and Haigh, 1997b; Buss and Perry, 1992], particularly in the case of physical aggression and
hostility, resulting in a considerably reduced total aggression score. The partner scores were
lower than the self-reports, albeit not significantly.
Correlations. Table 2 shows that the values for the intercorrelations between the subscales of
the AQ were comparable to those from previous studies reviewed in the Introduction [e.g., ArTABLE I. Means, Standard Deviations (in Parentheses) for the Subscales of the Aggression Questionnaire
and the Aggression QuestionnairePartner for Study 1

Physical aggression
Verbal aggression
Anger
Hostility
Total score

Male

Partner

18.43
(7.16)
14.05
(3.78)
15.57
(5.88)
16.81
(6.18)
64.39
(17.32)

16.33
(6.02)
13.51
(4.37)
15.56
(5.87)
15.60
(5.93)
60.92
(18.35)

Measuring Aggression

83

TABLE II. Intercorrelations of the Aggression Questionnaire, the Aggression Questionnaire - Partner, and
Their Subscales, for Study 1 (N = 77)

Age
P
V
A
H
PP
PV
PA
PH

Age

PP

PV

PA

PH

.24*
.20
.10
.08
.38**
.01
.14
.06

.51**
.55**
.39**
.78**
.37**
.43**
.33**

.49**
.44**
.57**
.45**
.40**
.46**

.62**
.56**
.29*
.64**
.38**

.36**
.10
.36**
.55**

.53**
.62**
.52**

.62**
.48**

.61**

*P < .05.
**P < 0.01.

P = physical aggression; V = verbal aggression; A = anger; H = hostility; PP = partner physical aggression; PV =


partner verbal aggression; PA = partner anger; PH = partner hostility.

cher and Haigh, 1997a, 1997b; Buss and Perry, 1992]: the physical subscale was highly correlated with verbal and anger, and somewhat less with hostility; verbal was highly correlated with
anger, but only slightly less with hostility; anger and hostility were highly correlated. As in
previous studies (see Introduction), age showed a moderate negative relationship with the physical
scale, although it was not correlated with any other scale.
The subscales of the AQ-P generally followed a similar pattern of intercorrelations to those
for the AQ: partner physical was highly correlated with partner verbal and partner anger; however, physical was also highly related to hostility; verbal was highly correlated with anger but
notably less with hostility; anger and hostility were highly correlated.
As predicted, the physical subscale of the AQ was highly correlated (r = .78) with the same
scale on the AQ-P. Anger exhibited a strong relationship with partners ratings of anger, and
hostility showed a slightly lower correlation with partners ratings of hostility. The verbal subscale
showed the lowest correlation, although this value was still substantial (r = .45). These moderate-to-high correlation coefficients indicate congruent validity for the AQ-P.
Discussion

These findings provide further evidence for the internal consistency of the AQ subscales and
offer preliminary data indicating good internal consistency and congruent validity for the AQ-P.
Buss and Perry [1992] used peer nomination to assess the agreement between the AQ and observations from knowledgeable informants. They found a correlation between self reported physical aggression and peer nomination of .45, for verbal aggression the value was .20, for anger .29,
and hostility .24. The correlations presented in the present study were substantially higher, ranging from .45 to .78, suggesting that their method of peer nomination was limited.
As predicted, the strongest relationship between self- and partner reports was for the physical
aggression subscale. It is likely that this occurred because physical aggression is more salient
than all other forms of aggression and consequently is foremost in peoples memory. Surprisingly, verbal aggression, not hostility, showed the lowest correlation. The verbal scale is the
shortest of the subscales, and showed the lowest Cronbachs alpha for self-reports.
It is interesting to note that the scores for each of the AQ subscales were substantial lower
than reported elsewhere [Archer and Haigh, 1997b; Buss and Perry, 1992]. It is likely that this
difference was a result of using a more extended age-range than is typically the case for studies

84

OConnor et al.

relying wholly on undergraduates. In support of this, age showed a significant negative relationship with physical aggression, suggesting that older men tend to report lower levels of physical
aggression. This is consistent with the other studies reviewed in the Introduction.
STUDY 2: THE DEVELOPMENT OF THE AGGRESSIVE PROVOCATION
QUESTIONNAIRE

As indicated in the Introduction, hypothetical scenarios involving provocation have received


little use as measures of peoples aggressive tendencies. One exception is the research by van
Goozen et al. [1994], who developed a 33-item scenario measure of anger proneness, the
Anger Situation Questionnaire (ASQ). This measure assessed anger disposition in terms of experienced emotion, felt intensity, and action readiness in line with emotion theory [Frijda,
1986]. However, this measure was developed specifically for women, in view of empirical evidence showing that causes for anger may have sex-specific effects [van Goozen et al., 1994]. It
is the aim of Study 2 to develop, pilot, and validate a sex-specific scenario measure of aggression using culturally relevant vignettes for men. Sex specificity is very important as many of the
scenarios and action alternatives employed by van Goozen et al. [1994] are not suitable for a
British male population. For example, one deals with baking an apple pie and what the respondent would do upon finding the apples recently purchased from the green-grocers were rotten.
Another is concerned with realising that the pattern and colour of your new expensive blouse
does not flatter you. Yet another deals with how the respondent would react, whilst in the post
office, to overhearing two men commenting that they felt a womens rightful place was in the
kitchen. However, some suitable vignettes were modified and included in the Aggressive Provocation Questionnaire.
A vignette or scenario is a stand-in for a real-life situation which can be interpreted as a valid
representation of a real provocative situation [van Goozen et al., 1994]. Each vignette in the
present study is intended to provide suitable situational information from which the participant
can reasonably determine how he would respond. In this way, it provides a measure of aggression that involves a series of specific provoking situations, rather than the more general statements involved in questionnaire measures.
It was predicted that scores on the Aggressive Provocation Questionnaire would be correlated
with the AQ, and in particular the physical aggression subscale would be correlated with the
aggressive actions scale, thus providing evidence for its construct validity (see Statistical Methods section for more detail).
Method

Scenario generation. Scenarios to be included in the questionnaire were generated with three
different methods, outlined below. It was intended to include a range of scenarios which were valid
representations of real-life provocative situations encountered by young- to middle-aged men.
1. Twenty-five males (aged 18 to 45 years) were recruited from an opportunity sample of men
who lived near or worked within the University of Manchester, or who were contactable
via email. These men were not included in Study 3. They were asked to complete a short
questionnaire which involved outlining any recent provocative or aggressive situation they
had encountered and how they had responded.
2. Consultation of the literature relating to the measurement of aggressive responding.
3. Consultation between the authors.

Measuring Aggression

85

Twenty-two scenarios were included in the pilot version of the questionnaire. The respondent
was asked to try to imagine at this moment: (1) how he would feel in each situation (angry,
frustrated, and irritated), measured on a 5-point Likert scale extending from not at all to extremely and (2) how he would react to each situation by choosing one of five action responses.
Similar to van Goozen et al. [1994] the alternative action responses were randomly ordered and
followed the general categories outlined below:
1. Avoiding the situation, denying that something is wrong, or transforming it into something
positive
2. Doing nothing, although feel angry
3. Distant anger, indirect or delayed angry behavior
4. Assertive behavior, confronting the provoking person but without overt verbal or physical
aggression
5. Aggressive behavior, direct verbal or physical aggression.
It is important to note that van Goozen et al. [1994] considered the action responses to be part
of an ordinal anger-readiness scale. We did not, but rather regarded the categories as part of a list
of mutually exclusive alternative action responses to provocative situations. Figure 1 provides
an example of one of the items.
The Aggressive Provocation Questionnaire consists of three subscales (the angry scale, the
frustrated scale, and the irritated scale) which are scored by calculating the sum for all items,
and the five action responses, which are each scored by the number of times that particular
response was chosen across the twenty-two scenarios. The analysis concentrated on the values
for assertive and for aggressive actions.

Fig. 1.

Example item from the Aggressive Provocation Questionnaire

86

OConnor et al.

Procedure. The Aggressive Provocation Questionnaire and the AQ [Buss and Perry, 1992]
were distributed to the same sample as that used in Study 1. The covering letter encouraged
respondents to comment upon the content and format of the Aggressive Provocation Questionnaire. As outlined earlier, 77 men returned their questionnaires.
Statistical methods. Pearsons product moment correlation coefficient was used to investigate
the relationships between the measures and to assess the construct validity of the Aggressive Provocation Questionnaire. Construct validity was evaluated by testing hypotheses about how the new
instrument will perform. For example, it is likely that the tendency to respond aggressively to provoking scenarios would have a high association with prior self-reported physical aggression (on the
AQ). Therefore, an indication of the validity of the instrument would be provided by a significant
positive correlation between the physical aggression subscale from the AQ and the aggressive action scale. For the assertive action scale, one would predict a small, nonsignificant correlation with
prior self-reported physical aggression as assertive actions generally do not involve physical aggression. Significant positive correlations are hypothesised between the total angry, frustrated, and irritated subscales of the Aggressive Provocation Questionnaire and the physical aggression subscale
from the AQ. Cronbachs alphas were calculated to assess the internal reliability of all the subscales.
All data were analysed with SPSS for Windows.
Results

Properties of Aggressive Provocation Questionnaire. Cronbachs alphas for the three


subscales were: angry, .94; frustrated, .93; and irritated, .89. All three values were within acceptable boundaries [Cortina, 1993]. Reliability analysis could not be performed on the aggression
action or assertive action scales as these are categorical scales.
Correlations. Table 3 shows the intercorrelations between the Aggressive Provocation Questionnaire subscales and the AQ. These were generally high, and in the predicted direction. The
aggressive action scale was strongly correlated with the physical subscale of the AQ (r = .65),
but less so with verbal aggression (r = .41), and anger (r = .47). Hostility showed no significant
association with aggressive action. Verbal aggression showed the strongest relationship with the
assertive action scale (r = .42). Physical aggression, anger and hostility were all moderately
correlated with assertive action (r = .27 to .29). The aggressive action and the assertive action
scales were not significantly correlated with one another.
TABLE III. Intercorrelations of the Aggressive Provocation Questionnaire (APQ) and the Aggression
Questionnaire and Their Subscales for Study 2 (N = 77)

Age
AGA
ASA
TA
TF
TI
P
V
A
H

Age

AGA

ASA

TA

TF

TI

.12
.04
.16
.03
.05
.24*
.20
.10
.08

.18
.43**
.33*
.38**
.65**
.41**
.47**
.25

.33*
.19
.40**
.27*
.42**
.29*
.28*

.74**
.85**
.38*
.40**
.60**
.55**

.81**
.23
.25
.56**
.35*

.28*
.34*
.58**
.36**

.51**
.55**
.39**

.49**
.44**

.62**

*P < .05.
**P < 0.01.

AGA = aggressive action; ASA = assertive action; TA = total APQ angry; TF = total APQ frustrated; TI = total APQ
irritated; P = physical aggression; V = verbal aggression; A = anger; H = hostility.

Measuring Aggression

87

The total of the angry subscale items from the Aggressive Provocation Questionnaire was
highly correlated with the AQ anger scale (r = .60) and with hostility (r = .55), but less so with
physical (r = .38) and verbal aggression (r = .40). The total of the frustrated subscale items
showed a reasonably high correlation with AQ anger (r = .56) and a lower correlation with
hostility (r = .35). However, no significant correlation was found with physical or verbal aggression. The total of the irritated subscale items was moderately correlated with all the AQ subscales
(r = .28 to .58). The total of the angry, frustrated, and irritated subscale items from the Aggressive Provocation Questionnaire all showed high intercorrelations with one another (r = .74 to
.85), indicating a substantial amount of overlap between these variables. Age was negatively
correlated with aggressive action and the total Aggressive Provocation Questionnaire angry
scale, although not significantly so.
Discussion

The results demonstrate the construct validity of the Aggressive Provocation Questionnaire.
The significant positive correlation (r = .65) found between the aggressive action scale and the
physical aggression subscale support the hypothesis that these subscales would show a strong
association. Unexpectedly, the assertive action subscale showed a significant correlation with
the physical aggression subscale. However, the magnitude of this relationship was substantially
lower than that found for the aggressive action scale. As predicted, the total angry, frustrated,
and irritated subscales were all significantly correlated with the physical aggression subscales,
with the total angry scale exhibiting the strongest association (r = .43). This latter finding is not
surprising as feeling quite to extremely angry in response to a provoking scenario is likely
to be highly associated with higher levels of prior self-reported physical aggression. The internal consistencies of all the subscales were found to be highly satisfactory and well within acceptable boundaries [Cortina, 1993].
The aggressive and assertive action scales were found to be unrelated, suggesting that these
are alternative response tendencies. This finding casts doubt on one of the measures used by van
Goozen et al. [1994], who added the numbers of aggressive and assertive actions to form an
angry readiness score.
Although the correlation between age and the aggressive action scale was, as predicted, negative, it was non-significant (unlike the correlation between age and the physical subscale of the
AQ). This is contrary to a number of previous findings (reviewed in the Introduction) showing
that levels of aggression, in particular physical, are lower at older ages. It is possible that this
null finding is a consequence of the relatively small sample size involved. This will be investigated further in Study 3.
During the course of this study, respondents voiced a number of criticisms regarding the
format and content of the Aggressive Provocation Questionnaire. In particular, they highlighted
several vignettes as involving a range of action response alternatives that were unrealistic, and
in others action responses appropriate to the particular scenario were not included. It was also
clear that the existing response categories were not always readily distinguishable from one
another, particularly the two involving anger but without an overt aggressive response.
STUDY 3

The final study involved the validation of a revised version of the Aggressive Provocation
Questionnaire, and again used the physical aggression subscale from the AQ to evaluate its
construct validity. Based on the results of Study 2, it was predicted that the physical subscale of

88

OConnor et al.

the AQ would have a stronger relationship with the aggressive action scale, the total angry,
frustrated, and irritated subscales than with the assertive action scale. All scales would also
exhibit satisfactory internal consistencies.
Method

Questionnaires. The 29 item AQ [Buss and Perry, 1992] was slightly modified to refer to a
specific time in the recent past (last 4 weeks) as in the sample used for Studies 1 and 2.
The Aggressive Provocation Questionnaire was first revised to take account of the criticisms
raised by respondents in Study 2. Then each of the revised items and the response alternatives
were presented by the first author to the two second authors, who then classified each of the
responses according to the following five categories: (1) avoid, (2) no response, (3) anger, (4)
assertive behavior, and (5) direct aggression. In most cases, there was agreement, and disagreements were resolved by making the response categories more clear-cut. The revised version of
the Aggressive Provocation Questionnaire consisted of 21 scenarios with refined action responses
tailored to the modified general categories. See Fig. 2 for the modified item that was used as an
illustration in Study 2.
As in Study 2, the Aggressive Provocation Questionnaire was scored in terms of three subscales
(angry scale, frustrated scale, and irritated scale), and a total score for numbers of aggressive
and assertive options chosen across the 21 scenarios.
Procedure. Two hundred and fifty questionnaires were distributed to a heterogeneous sample
of men who were employed locally (in a blue-collar factory), to staff and students at the University of Manchester (non-academic staff), Strathclyde University, and University of Liverpool.
The questionnaires were distributed to staff at designated meeting areas and to the students at

Fig. 2.

Modified item used as illustration in Study 2.

Measuring Aggression

89

the end of a lecture. Therefore, non-respondents could not be followed up. Each pack contained
a covering letter outlining the nature of the study and precise instructions on how to answer the
questionnaires. The only demographic details collected were age of the respondent and length of
present relationship. One hundred and thirty men returned their questionnaires, giving a response rate of 52%. This response rate is comparable to other postal surveys and is likely to be
higher than reported in Study 1 and 2 reported because only self-report data was collected [e.g.,
OConnor et al., 2000; Rout and Rout, 1997]. Ages ranged from 17 to 54 years (mean = 28.80;
SD = 7.88).
Statistical methods. Descriptive statistics were calculated for each of the variables and
Pearsons product moment correlation coefficients were used to investigate the relationship between each of the subscales and to evaluate the construct validity of the revised Aggressive
Provocation Questionnaire. Construct validity was assessed using an identical strategy to that
employed in Study 2. Cronbachs alphas were calculated to assess the internal reliability of all
the scales. All data were analysed with SPSS for Windows.
Results

Properties of the revised Aggressive Provocation Questionnaire and the AQ. The angry
subscale of the Aggressive Provocation Questionnaire showed a Cronbachs alpha of .92; for the
frustrated scale the value was .92, and for the irritated scale the value was .93. Reliability analysis could not be performed on the aggression action or assertive action scales as these are categorical scales. The AQ showed a Cronbachs alpha of .88 overall; for the physical scale, the
value was .81; the verbal .68; anger .82; and hostility .74. All values were within acceptable
boundaries [Cortina, 1993].
Descriptive statistics. Table 4 shows the descriptive statistics for the AQ and the Aggressive
Provocation Questionnaire. As in the first sample, scores for all scales were notably lower than
those reported elsewhere [Archer and Haigh, 1997b; Buss and Perry, 1992], particularly for
physical aggression and hostility.
Correlations. Table 5 shows the intercorrelations between the Aggressive Provocation Questionnaire subscales and the AQ. In this case, the correlations between the aggressive action scale
and the AQ subscales were substantially smaller than those found in Study 2, although all in the
predicted direction. Surprisingly, the anger subscale showed the strongest correlation (r = .34),
closely followed by physical (r = .27) and verbal aggression (r = .24). Similar to Study 2, hostility was not significantly related to the aggressive action scale. Therefore, the AQ physical aggression subscale was again more closely related to the aggressive action scale of Aggressive
TABLE IV. Means, Standard Deviations for the Aggression Questionnaire and the Aggression Provocation
Questionnaire and Their Subscales for Study 3 (N = 130)

Physical aggression
Verbal aggression
Anger
Hostility
Total AQ Score
Aggressive action
Assertive action
Total angry
Total frustration
Total irritated

Male

Standard Deviations

18.58
13.90
15.33
16.67
63.98
1.13
12.20
37.35
37.36
47.26

5.88
3.35
4.93
4.73
13.79
1.72
3.37
14.90
15.98
16.14

90

OConnor et al.

TABLE V. Intercorrelations of the Aggressive Provocation Questionnaire (APQ) and the Aggression
Questionnaire and Their Subscales for Study 3 (N = 130)

Age
AGA
ASA
TA
TF
TI
P
V
A
H

Age

AGA

ASA

TA

TF

TI

.05
.23*
.03
.08
.08
.16
.10
.16
.23*

.12
.48**
.30**
.41**
.27**
.24
.34**
.08

.03
.02
.14
.18
.04
.07
.16

.70**
.65**
.22*
.04
.31**
.27**

.83**
.04
.00
.28**
.28**

.14
.10
.34**
.22*

.23*
.46**
.29**

.41**
.35**

.47**

*P < .05.
**P < 0.01.

AGA = aggressive action; ASA = assertive action; TA = total APQ angry; TF = total APQ frustrated; TI = total APQ
irritated; P = physical aggression; V = verbal aggression; A = anger; H = hostility.

Provocation Questionnaire than it was to the assertive action scale. In fact, the assertive action
scale was not significantly correlated with the AQ subscales. The total angry, frustrated, and
irritated scales from the Aggressive Provocation Questionnaire showed high intercorrelations (r
= .65 to .83), indicating substantial overlap between these variables. However, as predicted, all
three subscales were significantly correlated with the aggressive action scale, but not with the
assertive action scales.
Age differences. Consistent with other studies [Archer, 2000; Harris, 1996], age was negatively correlated with hostility (r = -.23), and (although nonsignificantly) with physical aggression (r = -.16). A significant positive correlation was also found between age and the assertive
action scale (r = .23) suggesting that older men are more likely to respond in an assertive manner
to a provocative situation.
In order to explore these age differences further, the sample was split at the median age (28
years). Tables 6 and 7 show the intercorrelations between the Aggressive Provocation Questionnaire and AQ for the younger and older sub-samples respectively. Among the younger men (<
28 years), the physical subscale of the AQ is very strongly correlated with the aggressive action
TABLE VI. Intercorrelations of the Aggressive Provocation Questionnaire (APQ) and the Aggression
Questionnaire and Their Subscales for Males Aged Under 28 Years in Study 3

AGA
ASA
TA
TF
TI
P
V
A
H

AGA

ASA

TA

TF

TI

.11
.35*
.15
.30*
.56**
.37**
.39*
.23

.11
.08
.23
.19
.01
.11
.18

.81**
.75**
.32*
.16
.32*
.49**

.83**
.05
.08
.27*
.28*

.12
.17
.34**
.23

.27*
.52**
.50**

.54**
.42**

.50**

*P < .05.
**P < 0.01.

AGA = aggressive action; ASA = assertive action; TA = total APQ angry; TF = total APQ frustrated; TI = total APQ
irritated; P = physical aggression; V = verbal aggression; A = anger; H = hostility.

Measuring Aggression

91

TABLE VII. Intercorrelations of the Aggressive Provocation Questionnaire (APQ) and the Aggression
Questionnaire and Their Subscales for Males Aged Over 28 Years in Study 3

AGA
ASA
TA
TF
TI
P
V
A
H

AGA

ASA

TA

TF

TI

.15
.64**
.49**
.49**
.03
.11
.28*
.05

.12
.11
.04
.51**
.06
.01
.13

.56**
.54**
.09
.07
.31*
.04

.83**
.00
.06
.29*
.27

.18
.04
.36*
.21

.20
.35*
.00

.90*
.32*

.41**

*P < .05.
**P < 0.01.

AGA = aggressive action; ASA = assertive action; TA = total APQ angry; TF = total APQ frustrated; TI = total APQ
irritated; P = physical aggression; V = verbal aggression; A = anger; H = hostility.

scale (r = .56), but not at all with the assertive action scale. Conversely, in older men (> 28 years)
the physical subscale of the AQ is not correlated with the aggressive action scale, but is instead
correlated with the assertive action scale (r = .51). These findings reveal the reason for the low
overall correlation between physical aggression and the aggressive action scale in this sample.
Older men who rate themselves as physically aggressive on a general questionnaire are more
likely to choose an assertive rather than an aggressive response when asked about their responses to provocative situations. Among younger men, those endorsing physical aggression on
the AQ choose the overtly aggressive responses to the scenarios.
To investigate this finding further, the best working items from the Aggressive Provocation
Questionnaire were identified from the item-to-whole point-serial correlations; that is, items
with a low item-to-whole point-biserial correlation (rpb < .30) were removed. Although this
procedure reduced the number of items from 21 to 12 the reliabilities of the subscales were not
affected. The angry scale showed a Cronbachs alpha of .88, the frustrated scale .88, and the
irritated scale .89. Using this reduced scale, we found that, the AQ physical, verbal, and anger
subscales were only correlated with the aggressive action scale (not with the assertive action
scale) among younger men (rpb = .57, .30, .26, respectively). Among older men, physical aggression was correlated with the assertive action scale (rpb = .46), but other AQ subscales were
unrelated to both the assertive and aggressive action scales.
Discussion

This study involved refinement of the Aggressive Provocation Questionnaire response categories, in order to eliminate the ambiguous categories apparent in Study 2 and to evaluate the
construct validity and internal reliability of the tool. Yet it produced a result that was at first sight
paradoxical. The associations between the scenario measures of direct aggression and the AQ
scales were lower than those found with the previousapparently crudercategories. The reason for this became apparent when the sample was divided according to age. Younger men
showed the association between physical aggression and the aggressive action scale found in
Study 2, whereas older men had switched to an association between the AQ physical subscale
and assertive responses on the Aggressive Provocation Questionnaire. Possible reasons for this
are covered in the general Discussion section.
Even after age was controlled, the association between the AQ physical subscale and aggres-

92

OConnor et al.

sive scenario responses was still slightly lower than in Study 2, which involved the unrefined
questionnaire. There are probably two considerations that contribute to this. First, the original
list of assertive actions used in Study 2 did not always seem to be sufficiently clearly distinguished from directly aggressive actions: this impression is supported by the higher correlations
between assertive actions and AQ subscales found in Study 2 than in Study 3. The second consideration is that the other way in which these categories were refined involved the three action
categories that were not scored in the present analysis. In effect they provided alternative
nonaggressive categories that enabled the nonassertive and nonaggressive respondent to chose
from a range of other alternatives. Refining these appears to have had little influence on whether
people chose an aggressive/assertive or another category.
DISCUSSION

The studies reported in this paper provide further evidence for the robustness of a widely used
self-report measure of aggression, the AQ [Buss and Perry, 1992]. In Study 1, ratings of the
target person obtained from his spouse or long-term partner show a good level of agreement
with self-reports, the correlations ranging from .45 to .78. These data provide preliminary congruent validity for a new measure, the AQ-P. Agreement between these scales was greater in the
case of the physical subscale than for the other subscales. This is probably a consequence of the
higher salience of acts of physical than verbal aggression, and because anger and hostility refer
to mental states rather than specific behavioral acts. Since there was this variation in interpartner
agreement between the different subscales, the levels of agreement found in Study 1 are unlikely to have arisen from collusion between the partners, because this would have produced
similar levels of agreement irrespective of the content of the subscale.
The results presented in these studies demonstrate the psychometric properties of the Aggressive Provocation Questionnaire. Among the first sample, construct validity was indicated by the
high correlation between the physical subscale of the AQ and endorsing aggressive reactions to
scenarios. This was further corroborated by the substantially lower correlation between the AQ
physical subscale and the assertive action scale.
Among the second sample, the findings were apparently different. Although this study involved a refined version of the Aggressive Provocation Questionnaire to clarify the response
categories for scoring, it produced lower correlations with AQ measures than in Study 2. Correlations between the Aggressive Provocation Questionnaire responses and the physical, verbal,
and anger subscales were all significant, but ranged from .24 to .34, much lower than in Study 2.
Despite the lower correlations, these findings provide clearer evidence for the construct validity
of the Aggressive Provocation Questionnaire: that is, as predicted, the physical aggression scale
from the AQ is significantly correlated only with the aggressive action scale (r = .27) and not
with the assertive action scale. The internal consistencies of all the scales were found to be
highly satisfactory.
Unfortunately the data collected in Studies 2 and 3 only allowed for the evaluation of the
construct validity of the Aggressive Provocation Questionnaire. It is important that future research investigate alternative construct validation strategies and also assess the convergent validity of the measure. It would also be interesting to extend the work into different populations,
for example comparing violent with nonviolent offenders. This paper has demonstrated that
scores on the AQ can predict aggressive responding to provocative scenarios. However, can
scores on the latter predict aggressive responding in vivo?
The main reason why the correlations between aggressive responses to scenarios and the AQ

Measuring Aggression

93

were lower in Study 3 was revealed when the sample was divided into younger and older subgroups, using the median age of 28 years as the dividing line. Among the younger men, correlations comparable with those in Study 2 were found. As in the total sample, there were no significant
associations between AQ measures and assertive responding. Among the older men, there was
only one significant correlation between aggressive responding to scenarios and AQ measures,
a low (.28) correlation with anger. The physical subscale showed almost no relation to aggressive responses on the Aggressive Provocation Questionnaire. However, there was instead a correlation of about .5 between AQ physical and assertive responses to scenarios in these men. It
seems that among older men who rate themselves as physically aggressive on the AQ, an assertive rather than an overtly aggressive response is chosen when specific provoking situations are
presented to them.
The Aggressive Provocation Questionnaire is a new measure of aggression, designed to assess mens tendency to act aggressively when presented with a hypothetical set of provoking
situations. It is therefore different in emphasis from the AQ which asks people to rate how they
typically behave in terms of general statements, and also different from scales that ask people
whether they have used specific acts of behavior [Gergen, 1990; Harris, 1992]. All three can
play a part in the measurement of aggressive intentions and actions. In the Introduction, we
suggested that people who might in real life avoid an overt aggressive confrontation because of
its high potential cost would be free to opt for the aggressive alternative on the Aggressive
Provocation Questionnaire. Hence, their scores might reflect what they would like to have done
rather than what they really would have done in a given situation. The findings for the older age
group that the assertive choice was most strongly related to the physical scale of the AQ leads us
to question this. Even those men who rated themselves as more physically aggressive on the AQ
did not opt for the aggressive alternative on the Aggressive Provocation Questionnaire. Instead
they endorsed the assertive one. This raises a question that can only be answered by future
research, the extent to which responses to these hypothetical scenarios do accurately reflect
what the person would do when provoked in the way described.
Age is a neglected individual-differences variable in aggression research. While many studies, reviews, and commentaries have concentrated on sex differences in aggression [e.g.,
Bettencourt and Miller, 1996; Eagly and Steffen, 1986], only a handful of psychological studies
have examined age differences. Those that have (see Introduction) find lower values for measures of aggression at older ages among adults. It has been known for over 150 years that crimes
by men, including crimes of violence, decrease with age [Quetelet, 1833/1984; Courtwright,
1996], and this occurs irrespective of the absolute level of violent acts in a particular place [Daly
and Wilson, 1990]. Explanations for the decline range from Quetelets original emphasis on
declining physical strength and passion with age, to Daly and Wilsons [1988, 1990] view that
young mens aggression represents reproductive competition arising from sexual selection.
There are a number of convincing accounts of the widespread propensity of young men
particularly those with nothing to loseto take risks and not to put events into a long term
perspective [e.g., Gilmore, 1990; Daly and Wilson, 1990; Courtwright, 1996]. As Daly and
Wilson have correctly identified, sexual selection is the ultimate reason for young men underestimating the risks when it comes to intermale competition. According to this view, a more cautious estimate of risk and benefit gradually develops with age. This is of course accompanied by
declining physical strength, the peak of which between 25 and 30 years of age coincides with
that for intermale homicides [Daly and Wilson, 1990; Quetelet, 1833/1984]. Future studies that
concentrate on age differences in aggressive tendencies throughout the adult male life span
might concentrate on measuring the perceived effectiveness of overtly aggressive actions, per-

94

OConnor et al.

ceived risks involved in confrontations, and the perceived benefits of aggressive actions. The
last of these should take into account the relative weightings given to short- and long-term
benefits at different ages.
ACKNOWLEDGMENTS

We would like to thank Dr R Kirkman (The Palatine Centre, Manchester), Nicola GrahamKevan (University of Central Lancashire), Dr Rory OConnor (University of Strathclyde), Dr
Simon Chu, and Shane Johnson (University of Liverpool) for their help with questionnaire distribution.

REFERENCES
Anderson CA, Bushman BJ. 1997. External validity of
trivial experiments: the case of laboratory aggression. Rev Gen Psychol 1:1941.
Archer J. 1999. Assessment of the reliability of the Conflict Tactics Scale: a meta-analytic review. J Interp V
14:12631289.
Archer J. 2000. Which attitudinal measures predict selfreported aggression? Unpublished manuscript, Department of Psychology, University of Central Lancashire,
Preston, UK.
Archer J, Haigh A. 1997a. Do beliefs about aggression
predict self-reported levels of aggression? Br J Soc P
36:83105.
Archer J, Haigh A. 1997b. Beliefs about aggression among
male and female prisoners. Aggr Behav 23:405415.
Archer J, Kilpatrick G, Bramwell R. 1995. A comparison
of two aggression inventories. Aggr Behav 21:371
380.
Berkowitz L. 1989. Laboratory experiments in the study
of aggression. In: Archer J, Browne K, editors. Human aggression: naturalistic approaches. New York:
Routledge. p 4261.
Berkowitz L, Donnerstein E .1982. External validity is
more than skin deep: some answers to criticism of laboratory experiments. Amer Psychol 37:245257.
Bettencourt AA, Miller N. 1996. Gender differences in
aggression as a function of provocation: a meta-analysis. Psych Bull 119:422447.
Bjrkqvist K. 1994. Sex differences in physical, verbal
and indirect aggression: a review of recent research.
Sex Roles 30: 177188
Boulton MJ, Smith PK. 1994. Bully/victim problems in
middle-school children: stability, self-perceived competence, peer perceptions and peer acceptance. Br J
Dev Psychol 12:315329.
Broadbent J. 1999. Age differences in aggression. Unpublished BSc Dissertation, Department of Psychology,
University of Central Lancashire, Preston, UK.
Bushman BJ. 1995. Moderating role of trait aggressive-

ness in the effects of violent media on aggression. J


Pers Soc Psychol 69:950960.
Buss AH, Perry M. 1992. The aggression questionnaire. J
Pers Soc Psychol 63:452459
Cortina JM. 1993. What is coefficient alpha?: an examination of theory and applications. J Appl Soc Psychol
78:98104.
Courtwright DT. 1996. Violent land: single men and social disorder from the frontier to the inner city. Cambridge, MA: Harvard University Press.
Daly M, Wilson M. 1988. Homicide. New York: Aldine
de Gruyter.
Daly M, Wilson M .1990. Killing the competition: female/
female and male/male homicide. Hum Nat 1:81107.
Deluty RH. 1983. Childrens evaluations of aggressive,
assertive, and submissive responses. J Clin Child
Psychol 12:124129.
Eagly A, Steffen VJ. 1986. Gender and aggressive behavior: a meta-analytic review of the social psychological literature. Psych Bull 100:309330.
Frijda, NH. 1986. The emotions. New York: Cambridge
University Press.
Gergen M .1990. Beyond the evil empire: horseplay and
aggression. Aggr Behav 16:381398.
Gilmore DD. 1990. Manhood in the making: cultural concepts of masculinity. New Haven, CT: Yale University Press.
Harris JA. 1995. Confirmatory factor analysis of the Aggression Questionnaire. Behav Res Ther 8:991993.
Harris JA, Rushton JP, Hampson E, Jackson DN. 1996.
Salivary testosterone and self-report aggressive and
pro-social personality characteristics in men and
women. Aggr Behav 22: 321331.
Harris MB. 1992. Sex and ethnic differences in past aggressive behaviors. J Fam Viol 7:85102.
Harris MB. 1996. Aggressive experiences and aggressiveness: relationship to ethnicity, gender, and age. J Appl
Soc Psychol 26:843870.
Huesmann LR, Lagerspetz K, Eron LD. 1984. Intervening variables in the TV violence-aggression re-

Measuring Aggression
lation: evidence from two countries. Devl Psychol
20:746775.
Kouri, EM, Lukas SE, Pope HG Jr, Oliva PS. 1995. Increased aggressive responding in male volunteers following the administration of gradually increasing
doses of testosterone cypionate. Drug Alcohol Depend
St 40:7379.
Leifer AD, Roberts DF. 1972. Childrens responses to television violence. In: Murray JP, Rubinstein EA,
Comstock GA, editors. Television and social behavior, Vol. 2:Television and social learning. Rockville,
MD: National Institute of Mental Health. p 43180.
McCrae RR, Costa PT 1990. Personality in adulthood.
New York: Guilford Press.
Meesters C, Muris P, Bosma H, Schouten E, Beuving S.
1996. Psychometric evaluation of the Dutch version
of the Aggression Questionnaire. Behav Res Ther
34:839843.
Moffitt TE, Caspi A, Krueger R, Magdol L, Margolin G,
Silva PA, Sydney R. 1997. Do partners agree about abuse
in their relationship?: a psychometric evaluation of
interpartner agreement. Psychol Assess 9:4756.
Nisbett RE, Cohen D. 1996. Culture of honor: The psychology of violence in the South. Boulder, CO:
Westview Press.
OConnor DB, OConnor RC, White BL, Bundred PE.

95

2000. The effect of job strain on British general


practitioners mental health. J Men Hth 9:637654.
Quetelet A. 1833/1984. Recherches sur le Penchant au
Crime aux Differens Ages. Bruxelles: M. Hayez. Trans
Sylvester SF. Research on the propensity for crime at
different ages. Cincinnati, OH: Anderson.
Rout U, Rout JK. 1997. A comparative study on occupational stress, job satisfaction and mental health in British general practitioners and Canadian family
physicians. Psychol Health Med 2:181190.
Straus MA. 1979. Measuring intrafamily conflict and violence: The Conflict Tactics (CT) Scales. J Mar Fam
41:7588.
Straus MA, Hamby SL, Boney-McCoy S, Sugarman DB.
1996. The revised Conflict Tactics Scales (CTS2): Development and preliminary psychometric data. J Fam
Issues 17:283316.
van Goozen SHM, Frijda NH, Kindt M, van de Poll NE.
1994. Anger proneness in women: development and
validation of the anger situation questionnaire. Aggr
Behav 20:79100.
Williams TY, Boyd JC, Cascardi MA, Poythress N. 1996.
Factor structure and convergent validity of the Aggression Questionnaire in an offender population.
Psychol Assess 8: 398403.

APPENDIX A: AGGRESSIVE PROVOCATION QUESTIONNAIRE


1. Imagine yourself in the following situation:

It is Saturday evening and you are queuing to buy a lottery ticket. Its very busy and the shop
is soon to close. You have already been waiting for 10 minutes. Just when its your turn, someone else pushes in front of you.
How would you feel in this situation?

Angry
Frustrated
Irritated

Not at all

A little bit

Moderately

Quite a bit

0
0
0

1
1
1

2
2
2

3
3
3

What do you think you would do in this situation?


Feel angry but do nothing.
Push him and shout wait your turn.
Wait patiently until he had been served.
Say Im sorry but it was my turn.
Walk out of the shop.

7
7
7
7
7

Extremely
4
4
4

96

OConnor et al.

2. Imagine yourself in the following situation:

You have gone out to have a couple of drinks with your partner. Whilst you are at the bar, a
stranger approaches your partner and grabs her/his backside. On your return, your partner
tells you.
How would you feel in this situation?

Angry
Frustrated
Irritated

Not at all

A little bit

Moderately

Quite a bit

0
0
0

1
1
1

2
2
2

3
3
3

Extremely
4
4
4

What do you think you would do in this situation?


Leave and go somewhere else.
Do nothing.
Threaten the stranger and swear at him.
Tell him that such behavior is unacceptable and out of order.
Feel angry but do nothing at the time.

7
7
7
7
7

3. Imagine yourself in the following situation:

You are in a great hurry and right in front of you a car stops. A man gets out but he carries on
talking to the driver, blatantly ignoring your calls for him to move. You cannot get past the car.
How would you feel in this situation?

Angry
Frustrated
Irritated

Not at all

A little bit

Moderately

Quite a bit

0
0
0

1
1
1

2
2
2

3
3
3

What do you think you would do in this situation?


Get out of your car, walk over to the man and threaten him.
Reverse the car and take another route.
Sit in the car and fume with anger, but do nothing.
Calmly wait until he moved.
Go over to him, tell him that he is being unreasonable and ask him to move.

7
7
7
7
7

Extremely
4
4
4

Measuring Aggression

97

4. Imagine yourself in the following situation:

Your boss believes you have made a minor mistake at work. In the presence of all your work
mates, he embarrasses you by calling you an incompetent imbecile.
How would you feel in this situation?

Angry
Frustrated
Irritated

Not at all

A little bit

Moderately

Quite a bit

0
0
0

1
1
1

2
2
2

3
3
3

Extremely
4
4
4

What do you think you would do in this situation?


Shout back to him that it wasnt your fault.
Tell him that this is not the right way to talk to his employees.
Fell angry, but do not do anything.
Shrug it off, and go back to work.
Walk away from him.

7
7
7
7
7

5. Imagine yourself in the following situation:

You are in the cinema watching a movie. Behind you two lads are talking, laughing loudly
and kicking the back of your seat all the time.
How would you feel in this situation?

Angry
Frustrated
Irritated

Not at all

A little bit

Moderately

Quite a bit

0
0
0

1
1
1

2
2
2

3
3
3

What do you think you would do in this situation?


Turn around and ask them to be quiet or to leave.
Feel angry, and do nothing.
Move to another seat.
Try to ignore them.
Turn around and threaten to hit them if they do not keep quiet.

7
7
7
7
7

Extremely
4
4
4

98

OConnor et al.

6. Imagine yourself in the following situation:

You are driving down the motorway. As you are in the process of changing to a slower lane, a
reckless driver speeds out from the inside lane, cutting you off, causing you to slam on your
brakes, swerve, and nearly lose control of your car.
How would you feel in this situation?

Angry
Frustrated
Irritated

Not at all

A little bit

Moderately

Quite a bit

0
0
0

1
1
1

2
2
2

3
3
3

Extremely
4
4
4

What do you think you would do in this situation?


Blast your horn several times at them.
Feel angry but do nothing.
Try to move away from that driver
Chase after the other car and try to do the same to them.
Just carry on driving.

7
7
7
7
7

7. Imagine yourself in the following situation:

You are out with a group of friends and there is one guy who is continually taking the piss
out of you and generally insulting your family.
How would you feel in this situation?

Angry
Frustrated
Irritated

Not at all

A little bit

Moderately

Quite a bit

0
0
0

1
1
1

2
2
2

3
3
3

What do you think you would do in this situation?


Tell him to shut his mouth and threaten him if he doesnt.
Leave and go home.
Feel angry but do nothing.
Tell him that he is not funny and should stop.
Laugh it off and try not to let it get to you.

7
7
7
7
7

Extremely
4
4
4

Measuring Aggression

99

8. Imagine yourself in the following situation:

You find out from a friend that your partner has been unfaithful to you on one occasion, after
a works Christmas party.
How would you feel in this situation?

Angry
Frustrated
Irritated

Not at all

A little bit

Moderately

Quite a bit

0
0
0

1
1
1

2
2
2

3
3
3

Extremely
4
4
4

What do you think you would do in this situation?


Confront your partner about it next time you see her/him.
Get angry creating a big scene when you next see her/him.
Be inclined not to believe what I had heard.
Just not bother about it.
Feel very angry but do not do anything.

7
7
7
7
7

9. Imagine yourself in the following situation:

You are walking down the street on your way to an interview for a new job. As you turn the
corner, a window clearner nearby, accidentally spills soapy, hot water on your newly drycleaned suit.
How would you feel in this situation?

Angry
Frustrated
Irritated

Not at all

A little bit

Moderately

Quite a bit

0
0
0

1
1
1

2
2
2

3
3
3

What do you think you would do in this situation?


Move away from the scene as quickly as possible.
Feel angry but dont do anything.
Attract his attention, shout and swear at him.
Attract his attention and point out what he had done.
Just walk on and think that you were unlucky today.

7
7
7
7
7

Extremely
4
4
4

100

OConnor et al.

10. Imagine yourself in the following situation:

Youre sat on a train quietly reading the newspaper. A couple of football supporters are sitting
a few seats in front shouting, swearing and generally being obnoxious. Suddenly, one of them
throws an empty beer can in the air and it accidentially hits you.
How would you feel in this situation?

Angry
Frustrated
Irritated

Not at all

A little bit

Moderately

Quite a bit

0
0
0

1
1
1

2
2
2

3
3
3

Extremely
4
4
4

What do you think you would do in this situation?


Sit there feeling angry.
Try to ignore them.
Find somewhere else to sit.
Attract their attention and ask them to be more careful.
Go over to them and threaten them.

7
7
7
7
7

11. Imagine yourself in the following situation:

It is Saturday afternoon and you are looking for a parking space in the centre of town. You
drive into a car park and just as you are about to reverse into one of the few remaining spaces
another car speeds into your space.
How would you feel in this situation?

Angry
Frustrated
Irritated

Not at all

A little bit

Moderately

Quite a bit

0
0
0

1
1
1

2
2
2

3
3
3

What do you think you would do in this situation?


Drive away to look for another space.
Get out of the car, go over to the other driver and shout and swear at him.
Do nothing.
Go over to the other driver and tell him that this was your space.
Feel angry but do nothing.

7
7
7
7
7

Extremely
4
4
4

Measuring Aggression

101

12. Imagine yourself in the following situation:

You arrive home from work, it has been a long day. The kids are screaming and running
around the living room whilst you are trying to relax and watch the television.
How would you feel in this situation?

Angry
Frustrated
Irritated

Not at all

A little bit

Moderately

Quite a bit

0
0
0

1
1
1

2
2
2

3
3
3

What do you think you would do in this situation?


Feel angry but do nothing at the time.
Say Children, please sit down and be quiet.
Shout at the kids to be quiet.
Sit patiently and ignore the children.
Get up and go into another room.

7
7
7
7
7

Extremely
4
4
4

You might also like