British Jownal of ClinicaL P g d o l o u (1993), 32, 327-331 Printed in Great Britain 327

0 The British Psychological Society

The Herald of Free Enterprise disaster:

Measuring post-traumatic symptoms
30 months on

Steve Joseph*
Department of Pydology, Universio of Uhter at Coleraine, Co. Londonderry, BT52 I S A , Northern Ireland

William Yule and Ruth Williams

Department of Psychology, Institate of Psychiatry, London

Peter Hodgkinson
Centre for Crisis Psychology, Skipton, N . Yorks

The aim of the present study was to explore the psychometric properties of the
Impact of Events Scale. Evidence is presented that the scale is composed of two
major factors tapping intrusion and avoidance which are associated with greater
psychological distress.

The introduction of Post-Traumatic Stress Disorder (PTSD) acknowledged a causal

role for traumatic events in the onset of psychiatric disorder, and fuelled interest into
the systematic quantification of the psychological phenomena associated with
disaster. PTSD is characterized primarily by symptoms associated with persistent re-
experiencing of the event, and persistent avoidance of related stimuli (American
Psychiatric Association, 1987).
One self-report instrument that has been widely used for the assessment of
intrusive and avoidant symptoms is the Impact of Events Scale (IES). This yields a
seven-item measure of intrusion (items 1, 4, 5, 6, 10, 11 and 14) and an eight-item
measure of avoidance (items 2, 3, 7, 8, 9, 12, 13 and 15) (Horowitz, Wilner &
Alvarez, 1979; Zilberg, Weiss 8t Horowitz, 1982).
However, although researchers and clinicians have tended to follow this scaling
procedure, there is other evidence that several of the items do not load consistently
on either factor (items 2 and 12) or emerge as a third factor (items 8 and 15)
(Schwarzwald, Solomon, Weisenberg & Mikulincer, 1987 ; Joseph, Williams, Yule
& Walker, 1992).
The aim of the present study, therefore, was to analyse the responses of adult

* Requests for reprints.

13-2
328 Steve Joseph and others
survivors of the Herald of Free Enterprise disaster at 30 months in order to confirm
the use of the intrusion and avoidance scales, and to examine their relationship to
other measures of psychological distress.

Method
Background
In March 1987, the Herald of Free Enterprise ferry capsized shortly after leaving Zeebrugge harbour
in Belgium. Out of a total of 600 on board, 193 passengers and crew died.

Subjects
IES measures were obtained from 7 3 adult survivors of the Herald of Free Enterprise disaster
(approximately 18 per cent of the total surviving population) comprising 46 men and 27 women ranging
in age from 17 to 55 (M = 35 years), of whom 23 were bereaved in the disaster, 25 were coded as not
being in a current relationship (single, divorced or widowed) and 47 were in a current relationship
(married or co-habiting).

Measures
The Impact of Events Scale (IES) was devised by Horowitz e t al. (1979) to assess the emotional sequelae
of extreme stress. The questionnaire describes 15 emotional reactions to which the respondent is asked
to indicate on a four-point scale ranging from ‘not at all’ to ‘often’ how frequently each reaction has
been experienced in the last week. Scores are obtained by assigning the values 0, 1, 3 and 5 to each of
the frequency categories. Table 1 presents the IES Scale items. Horowitz e t a/. reported high internal
consistency of the intrusion and avoidance subscales (Cronbach‘s alpha for intrusion = .79, for
avoidance = .82).
The 28-item General Health Questionnaire (GHQ-28 : Goldberg & Hillier, 1979) was also included.
This is a measure of psychiatric well-being and yields four subscale scores: somatic symptoms, anxiety
and insomnia, social dysfunction and severe depression.

Procedure
Survivors of the Herald disaster completed the IES between two and three years following the capsize
as part of their participation in a postal survey carried out by the Herald Research Team (Yule,
Hodgkinson, Joseph, Parkes & Williams, 1990). The research was carried out in conjunction with Kent
Social Services who agreed to post questionnaires to all survivors on their database. Questionnaires
were returned directly by the survivors to the first author at the Institute of Psychiatry.

Results
The total mean score on the IES was 35 indicating a high level of intrusion ( M = 19)
and avoidance ( M = 16). This is substantially higher than the total mean score of 17
reported by McFarlane (1988) for his group of 32 firefighters diagnosed as suffering
from persistent chronic PTSD 29 months following the Australian bushfire. Also,
over two-thirds (49/73-67 per cent) of the sample scored higher than the cut-off
point of 4 on the total GHQ-28 taken to indicate high risk of psychiatric disturbance
( M = 10). However, we do not know how representative the present sample is of
the remaining surviving population and we are mindful of Thompson’s (1991)
observation that partial samples may underestimate disaster effects because those
with most severe post-traumatic responses do not reply.
 PTSD 30 months after the event 329

Principal component anaEysis

The solution obtained yielded three factors with eigenvalues greater than 1.OO, and
extracted 63.0 per cent of the total variance. Factor one with an eigenvalue of 6.18
accounted for 41.2 per cent of the extracted variance and had loadings ranging from
0.71 to 0.84 on five items (items 1, 5, 10, 11 and 14). This factor is best characterized
as intrusion. Factor two with an eigenvalue of 2.04 accounted for 13.6 per cent of the
extracted variance and had loadings ranging from 0.59 to 0.87 on five items (items
3, 7, 9, 12 and 13). This factor is best characterized as avoidance. Factor three with
an eigenvalue of 1.22 accounted for 8.1 per cent of the extracted variance and had
loadings ranging from 0.54 to 0.77 o n five items (items 4 , 6 , 7 , 8 and 15). This factor
contains items characterized by both intrusion and avoidance. Inspection of these
items suggests that this factor is primarily tapping items specific to sleep disturbance
(items 4 and 6) and emotional numbing (items 8 and 15).
However, only two factors were found above the point at which there was a
marked ‘elbow’ in the scree plot of eigenvalues. For this reason, a two-factor forced
solution was carried out (see Table 1). Factor one, with an eigenvalue of 6.73,
accounted for 45 per cent of the extracted variance and had loadings ranging from
.62 to .78 on seven items (items 3, 7, 8, 9, 12, 13 and 15). This factor seems best
characterized as avoidance and, except for the omission of item 2, is in accord with
that obtained by Zilberg e t al. (1982). Factor two, with an eigenvalue of 1.63,
accounted for 11 per cent of the extracted variance and had loadings ranging from

Table 1. Factor loadings for a forced two-factor solution

Item Factor 1 Factor 2

1. I thought about it when I didn’t mean to. - .12 .79*
2. I avoided letting myself get upset when I thought about it .40 .41
or was reminded of it.
3. I tried to remove it from memory. .75* .11
4. I had trouble falling asleep or staying asleep because of .45 .52*
pictures or thoughts about it that came into my mind.
5. I had waves of strong feelings about it. .17 .85*
6. I had dreams about it. .33 .58*
7. I stayed away from reminders of it. .78* .20
8. I felt as if it hadn’t happened or it wasn’t real. .62* .20
9. I tried not to talk about it. .69* .13
10. Pictures about it popped into my mind. .21 .76*
11. Other things kept making me think about it. .24 .74*
12. I was aware that I still had a lot of feelings about it, but I .65* .31
didn’t deal with them.
13. I tried not to think about it. .77* .04
14. Any reminder brought back feelings about it. .29 .70*
15. My feelings about it were kind of numb. .67* .33
~ ~ ~~

* indicates loading greater than or equal to .50 in the factor.

330 Steve Joseph and others
.52 to .85 on seven items (items 1, 4, 5, 6, 10, 11 and 14). This factor seems best
characterized as intrusion and is in accord with Zilberg e t af.
The factor structure is largely consistent with the currently accepted scaling
procedure. Scales were therefore computed according to Horowitz e t af. (1979) and
Zilberg e t af. (1982) and correlations computed between scores on the IES and the
GHQ-28. Higher scores on all four GHQ-28 scales were associated with greater
intrusion (lowest r = .28, p < .Ol) and greater avoidance (lowest r = .37, p < .OOl).
The IES would seem, therefore, to have concurrent validity as a measure of distress.
No significant difference was obtained on any of the symptom measures between men
and women (highest F = .71, n.s.) and no significant correlation was obtained
between age and any of the symptom measures (largest r = 21, n.s.).

Discussion
The data extend previous work by applying the IES to survivors of mass civilian
disaster. A three-factor solution was obtained for the IES. The first two factors were
best characterized as intrusion and avoidance. The third factor seemed to be
specifically tapping those items concerned with sleep disturbance and emotional
numbing. The emergence of this factor is interesting as previous work has also
identified a third factor containing emotional numbing items (Joseph e t af., 1992;
Schwarzwald e t a/., 1987). It is suggested, therefore, that future psychometric work
should develop the item pool of the IES to include these aspects of response.
However, the percentage of variance extracted by the third factor in the present
study was small and examination of the scree plot suggested carrying out a forced
two-factor solution. This yielded a primary and secondary factor almost identical to
that reported by Horowitz e t a/. (1979) and Zilberg e t af. (1982). Only one
discrepancy was noted: item 2 (‘I avoided letting myself get upset when I thought
about it or was reminded of it’) loaded equally well on both avoidance and intrusion.
A possible criticism of this work is that the number of subjects, although close to
meeting the criteria of five subjects per variable proposed by Gorsuch (1983), falls
short of the minimum of 200 subjects suggested by Comrey (1988) when carrying out
structural analysis of a psychological test. However, it is noted that the present
sample size compares similarly with that reported by Zilberg e t af. (1982) and
Horowitz e t af. (1979).
In conclusion, the present study confirms the use of the IES with survivors of
civilian trauma. However, it is also suggested that future psychometric work should
develop the item pool further to include aspects of sleep disturbance and emotional
numbing.

Acknowledgements
Thanks are due to the survivors who took part in this survey, and to Kent Social Services for
administration.
References
American Psychiatric Association (1987). The Diagnostic and Statistical Manualof Mental Disorders, 3rd rev.
ed. Washington, DC: APA.
 PTSD 30 months after the event 331
Comrey, A. L. (1988). Factor-analytic methods of scale development in personality and clinical
psychology. Journal of Consulting and Clinical Psychology, 54, 754-161.
Goldberg, D. P. & Hillier, V. F. (1979). A scaled version of the General Health Questionnaire.
Psychological Medicine, 9, 139-1 45.
Gorsuch, R. L. (1983). Factor analysis. Hillsdale, N J : Erlbaum.
Horowitz, M., Wilner, N. & Alvarez, W. (1979). Impact of Event Scale: A measure of subjective stress.
Psychosomatic Medicine, 41, 209-218.
Joseph, S., Williams, R., Yule, W. & Walker, A. (1992). Factor analysis of the Impact of Events Scale
with survivors of two disasters at sea. Personality and Individual Differences, 13, 693-697.
McFarlane, A. C. (1988). The longitudinal course of posttraumatic morbidity: The range of outcomes
and their predictors. Journal of Nervous and Mental Disease, 176, 30-39.
Schwarzwald, J., Solomon, Z., Weisenberg, M. & Mikulincer, M. (1987). Validation of the Impact of
Event Scale for psychological sequelae of combat. Journal of Consulting and Clinical Psychology, 55,
251-256.
Thompson, J. (1991). Kuwait airways hijack : Psychological consequences for survivors. Stress Medicine,
7, 3-9.
Yule, W., Hodgkinson, P., Joseph, S., Parkes, C. M. & Williams, R. (1990). The Herald of Free
Enterprise : 30-month follow up. Paper presented at the second European Conference on Traumatic
Stress, The Netherlands, 23-27 September 1990.
Zilberg, N. J., Weiss, D. S. & Horowitz, M. J . (1982). Impact of Events Scale: A cross-validation study
and some empirical evidence supporting a conceptual model of stress response syndromes. Journal of
Consulting and Clinical Psychology, 50, 407414.

Received 28 November 1991; revised version received 1 1 May 1992