An Internet Journal devoted to the Psychology of Personal Constructs

PDF version
Journal Main
Contents Vol 9





Nigel J. Hopkins

Formerly Consultant Clinical Psychologist, Sheffield Health Authority, U.K.



Purpose. A heuristic PCT (Personal Construct Theory) model of the development of panic and subsequent phobic avoidance identified the PCT notion of ‘threat’ as a major causal variable (Hopkins, 1995). The formulation of this model arose out of the clinical observation that suggested many women diagnosed with agoraphobia had previously been focused on helping others, but life events had disrupted these activities. This report describes an initial assessment of the reliability of this clinical impression.
Method. The ‘self characterisations’ of twenty agoraphobic women were compared to those of twenty non-agoraphobic women. The two groups were found to display largely equivalent levels of anxiety and depression in the absence of agoraphobic avoidance.
Results.The self-descriptions of the agoraphobic women contained significantly more construct poles coded as ‘tenderness’ (Landfield) than were found in the self-descriptions of the non-agoraphobic women.
Discussion. Further investigation is suggested that would improve the reliability of the coding procedure and incorporate controls to aid distinction between ‘state’ and ‘trait’ effects.
Key words: Agoraphobic avoidance, self-characterisation, post-coding, reliability, core structure, attachment, and narrative.


Hopkins (1995) suggests that the panic experience, described by individuals who go on to become phobic of those circumstances they perceive as being associated with the panic’s occurrence, is a consequence of the PCT (Personal Construct Theory) notion of ‘threat’. ‘Threat’, is defined as “the awareness of an imminent comprehensive change in core structure” (Kelly, 1955). In this study it is hypothesised to be created by the removal of the opportunities to validate core structure, and especially ‘core role’, and this could take place through the effects of life events, for example the loss of a loved one or other major changes in the demands made upon the individual that had given them a sense of purpose and a sense of significance.
Stefan (1977, p. 283) notes that core structure, which posits an identity or meaning to behaviour, is superordinate in the PCT system. He goes on to point out that core structure is not directly influenced by experience, but responds to the success of its subordinate or first order constructs in their prediction of events. As such core structure can be vulnerable to deviations from the limits set by, for example, having a core structure of being a Christian. They are now constrained to behave in a Christian way and not in some other way. Their core structure will be largely validated by Christian acts and not to such an extent by other acts. Individuals at greater risk of experiencing ‘threat’ may have such a superordinate core structure that limits their freedom to be otherwise.
The feelings associated with ‘threat’ may lead to panic especially away from the familiar and more validating surroundings of home. Travel away from home may involve decreases in validation and a loss of a sense of purpose, threat is experienced, panic occurs and becomes anticipated in unfamiliar and un-validating settings that are now avoided.
A ‘self characterisation’ (Kelly, 1955, p. 321) is a very open-ended approach to the problem of obtaining a person’s own view of their life, that is their ‘core structure’. It can identify individual differences in self description that may have greater significance in relation to the development of an understanding of an individual’s purposes than can be obtained using a questionnaire. The latter asks each person the same questions, thereby imposing the researcher’s own model of people onto the data. For example, if asked, most people would claim to care about other people, but how many would volunteer such information about themselves? The general research question was “Do people displaying agoraphobic avoidance characterise themselves differently to other patients who experience similar levels of anxiety and depression in the absence of such agoraphobic avoidance?”
In a sense a person’s self-characterisation is a form of invitation to another person to interact with them in a certain way. A written self-characterisation of the type described above is a particularly formal type of self-presentation. Usually our public self-presentations are delivered verbally and behaviourally. The constructs described in a self-characterisation, and on show, are not merely descriptions of past events. Instead, like all constructs, they serve to enable a person to anticipate the future. In anticipating the future a person acts in particular ways and these actions prompt a response from those around the person, and in this sense a person’s predictions have a certain causal quality or effect not just on others, but also on the actor themselves. The results of those predictions can lead to a validation of construct systems or to an absence of validation, invalidation and a loss of a sense of purpose, and even the loss of the blueprint they had been following and which had structured their life.
This line of thinking lies behind Theodore Sarbin’s foreword to Hallam’s 1985 book on anxiety. Sarbin advocates anxiety as a construction or metaphor, suggesting that this and other metaphors need not be restricted to those drawn from science and medicine but can include “game playing, narrative, drama and rhetoric” (p. x). Sarbin uses the term ‘contextualism’ to describe this level of discourse and says that its “root metaphor is the historical act”. “Actors who participate in this historical act are agents . . . who engage in intentional actions not only to solve problems of a practical nature, but also to maintain or enhance their identity. To this end they construct their worlds” (Sarbin, in Hallam, 1985, p. xi). Hallam and Sarbin’s view is paralleled by a more recent emphasis on personal narratives that relate our life story (McAdams, 2008).
Sarbin believes that Hallam’s critical review supports the conclusion that people should not be regarded as passive, but as active agents. Causality should perhaps be sought in relation to the intentions of the individual rather than in relation to processes within the organism. These parallels with Kelly’s position are very strong. Sarbin and Hallam put forward positions that can be seen in some respects as elaborations of Kelly’s theory.
Put in another way, by construing the world well we learn about causality, how things and other people, work. We can use this information to bring about desired effects. By construing ourselves in a certain way and by communicating this self to others in the dramaturgical way that Kelly, Sarbin, and Harre (1979) describe we create a context within which others are to some extent constrained to act themselves. An agoraphobic person’s self-characterisation may capture features of self-construction that distinguish them from other people experiencing other types of painful emotional state.
Were such differences found to exist between these clinical groups in this study it was recognised that it would not be possible to determine whether such measures were a function of the subject’s clinical state or a reflection of more enduring pre-morbid characteristics or traits.
Winter (1983) and Winter and Gournay (1987) utilised Landfield’s (1971) system of construct pole classification in carrying out a content analysis of the (bi-polar) constructs used in their study, and they suggest that agoraphobics “submerge” the ‘low tenderness’ pole of Landfield’s ‘high-low tenderness’ construct (that is, tend not to use this potentially available classification). They further suggest that the agoraphobic person constricts their perceptual field to avoid dealing with interpersonal conflict. This is in line with other theorists, most notably Goldstein and Chambless (1978) and Chambless and Goldstein (1980b, 1981, 1982). More recent work by Winter, Gournay, and Metcalfe (1996), Winter et al. (2006), and Winter and Metcalfe (2005) applies PCT therapeutic techniques that encourage experimentation aimed in particular to the dilation of the perceptual field and on to subsequent elaboration of the persons construct system. In addition to the submergence of the low tenderness pole, another perhaps complementary tendency may also exist, namely that agoraphobics elaborate the ‘high tenderness’ pole of this same dimension.
Dorothy Rowe drawing on clinical observation suggested to the author that agoraphobic people are group oriented (Rowe, 1985). This hypothesised gregarious nature may have significance in a possible link with a need not to alienate themselves from others on whom they may depend for validation. Hopkins (1995) suggested that one reason for this may lay in a particular reliance the pre-agoraphobic person may have on the recognition or validation of their self image by others, and through interaction with others, lacking perhaps more than most people an ability to sustain their core structure by more independent means. Perhaps this is because of the nature of this other directed core structure, and perhaps for reasons that also relate to the very development of such other directed core structure; in this respect Guidano and Liotti (1983) and Liotti (1991) implicate the attachment process (e.g. how should one behave in order to be valued by one’s primary carer?). Perhaps, too, they have some block to seeing themselves as an assertive self directed person as such assertion may create interpersonal conflict. That is, such a construction of themselves may prove to be incompatible with others’ superordinate core structure.
Strodl and Noller (2003) examined the attachment styles of agoraphobic participants using the five dimensions of attachment measured by the Attachment Style Questionnaire (Feeney, Noller and Hanrahan, 1994). They report that ‘preoccupation with relationships’ was associated with agoraphobic behaviour, and catastrophic cognitions about bodily sensations partly mediated this association. They also speculate on the possibility that interpersonal conflict increases fears of abandonment and thus separation anxiety. Being unable to construe interpersonal conflict due to their submergence of construct poles and constriction of perceptual field, which Winter and Gournay (1987) note is referred to by Kelly (1955, p.1139) as a ‘retreat to safety’, they are rendered unable to deal with the ‘real’ issues that require changes to take place and instead have catastrophic cognitions, and for example, misattribute their feelings as signs of impending death and they can imagine a variety of disorders that they fear are about to bring that event about. That they do sense that they are about to come to an end gives further impetus to the idea that they are experiencing Kellian ‘threat’.
Hopkins (1995) also hypothesised that when experiencing agoraphobic avoidance, or immediately prior to having such an experience, a person may have a proneness to rigidity or be ‘closed to alternatives’, and to have a generally more judgmental stance when it comes to surveying the world of others around them, reflecting perhaps a further aspect of ‘submergence’ and the constriction of perceptual field identified by Winter and Gournay (1987).
More positively, clinical experience suggests the hypothesis that agoraphobics tend to describe themselves formerly as being energetic people who were used to getting things organised and done. This may relate to an aspect of assertiveness that does not involve interpersonal conflict.
The hypotheses are:
1. Agoraphobics elaborate the ‘high tenderness’ pole, and will therefore include more frequent statements to this effect in their self-characterisations than will the contrast group.
2. Agoraphobics will portray themselves as being of an especially sociable nature by using the ‘active social interaction’ category frequently.
3. They will reveal signs of an underlying selectivity to knowledge about life through their use of the ‘closed-minded’ construction category.
4. They will describe themselves as energetic and organised people who get things done (‘forcefulness’ and ‘organisation’ categories.)

Over a period of two years forty of the general practitioner referrals to a department of clinical psychology were of women who were identified as belonging to one of two groups. One group of twenty was confirmed by the author as meeting the DSM III criteria for ‘agoraphobia’. The other group of twenty was not agoraphobic, but in other respects it was thought to be likely, based on initial interview, and later confirmed by assessment, that their symptoms would be of a similar type and level of intensity. Both groups were free of any psychotic characteristics on clinical examination.
The mean ages of the two groups were: agoraphobics mean = 38.96, S.D. = 12.73 (range 18 to 71); non-agoraphobics mean = 35.28 years, S.D. = 11.51 (range 24 to 59). [t (38) < 1.00, NS.].
(Note: The term ‘agoraphobia’ is used here to describe the type of avoidance being displayed and ‘agoraphobic’ participant refers to people in the study who are handicapped by this avoidance. The descriptions are used to identify their main complaint, not to infer an underlying medical condition.)
The assumption being made had the implication that differences between the groups on non-symptom variables could be taken to indicate the likelihood of an association with agoraphobia as opposed to being a feature shared by all participants sharing equivalent symptoms of anxiety and depression. Assessment provides reasonable confirmation that the above assumption of symptom equivalence was correct:
Comparison of the scores of eighteen of the twenty agoraphobic participants with fifteen of the twenty non-agoraphobic participants scores on the Beck Depression Inventory found no significant difference, means = 17.46 and 17.68 respectively t (31) NS.
On scales of Social Avoidance and Distress there again was no significant difference, mean Ag = 13.5 and mean NonAg = 14.75 t (31) NS.
 On the Fear of Negative Evaluation no significant difference was found with Agoraphobic means = 17.60, Non-Agoraphobic means = 21.30 t (31) NS.
In the case of eighteen of these twenty agoraphobic participants and fifteen of the Non-agoraphobic participant group the Crown Crisp Experiential Index (CCEI) measures indicated differences on just two of the scales, as expected on the ‘Phobia’ scale, Ag means =11.68, Non-Ag means 8.00 t (31) p<0.01 1 tail, and (unexpectedly) on the ‘Hysteria’ scale Ag = 5, Non-Ag = 8 t (31) p<0.05 2 tail.
On the CCEI ‘Anxiety’ scale both groups demonstrated high levels mean Ags =12.13, mean Non-Ags =12.36. t (31) NS.
Summing of the six CCEI subscales gives an Agoraphobic group mean = 53.56 and the Non-Agoraphobic group mean = 56.3 (t (31)=0.59 NS).
To assess and monitor levels of avoidance Johnston et al’s (1984) Guttman Scale measure of agoraphobic avoidance was adapted for use. Initial scores were Agoraphobics mean = 32.35, Non Agoraphobics mean = 17.40 t (38) p<0.0005.
The self characterisation was elicited by asking each participant to respond to the following task:
“Write a brief character sketch of yourself as if you were the principal character in a play.  Write sympathetically and intimately as would a close friend, but perhaps no friend could know you as well as this. Write in the third person thus: “He/she is the sort of person who. . .” “(name) often wonders if he/she . . . etc.” It may help if you imagine that you are the director of a play and are trying to describe the character to a leading actor or actress, and your aim is to communicate the essential qualities and characteristics of that character.”
Each Self Characterisation (SC) was started in the presence of the author and some were completed on the day, others were completed as homework and brought along the following week. Participants were asked not to confer with friends or relatives prior to completing their SC as it was important that this was their own view of themselves.
Landfield (1965) has produced a method of construct ‘content postcoding’ in which the two poles of each construct can be rated independently. Sets of constructs from different people can be compared using this technique and this allows freely elicited constructs to be used in studies with all the advantages this brings and also allows for within group and between group comparisons to be made in relation to construct content. Landfield (1971) reports good interjudge reliability.
It was decided to use Landfield’s (1965) version of his rating manual which involves twenty two rating categories. Applying Landfield’s categories to a self characterisation script necessitates an additional step in the process. This involved the identification of statements that could be taken to identify a construct pole. Two graduate psychologists employed as assistant psychologists in the clinical psychology department carried out this initial task dividing the scripts from each participant group between them equally and working independently. They then went on to apply the method of categorisation set out in Landfield’s 1965 manual. These coders were unaware of both the experimental hypotheses and the nature of the investigator’s line of theorising. Their instructions were as follows: “Read the self-characterisation through to gain an overview. Write out the self-characterisation again, putting each adjectival statement onto a separate line. A single sentence may contain several descriptions and each one needs to be rated separately. Below is an example”. The directions in Landfield’s manual were discussed and the raters asked to adhere to these instructions.
As the identification of construct poles within self-characterisations was not part of Landfield’s study a check on inter-rater reliability was incorporated into this study. Another psychology assistant was asked to randomly select ten self-characterisations from each of the two participant groups’ total of twenty. Using the same instructions she repeated the analysis and her results were compared to those of the first raters. The findings of this reliability check are set out in the results section below.
As the nature of the data provides reasonable grounds for being cautious in making the assumption that it is normally distributed or that the level of measurement achieved reached the criteria required for interval or ratio scales, both requirements needing to be met if the unambiguous use of parametric tests of significance is to be realised, the distribution free method of chi-square was applied, along with the non-parametric Mann-Whitney test.
Reliability check of the post-coding procedure
Stage one of the Self-Characterisation post-coding analyses involved the identification of individual construct poles from within each self-characterisation, and a measure of the reliability of this procedure would be the degree of agreement between raters in making these identifications. The Cohen’s Kappa was used as an index of the extent of such agreement:
The overall level of agreement between the two sets of classifications was found to be Kappa = 0.59, which is only modest. In fact, agreement was found to be Kappa = 0.60 for the Agoraphobic group and Kappa = 0.57 for the Contrast group, a two tailed test demonstrated that these levels were not significantly different, ( t (18) =1.65 ,p>0.05,N.S.).
These figures as they stand suggest that the results of this study require cautious interpretation. An analysis of the disagreements show that they are largely a result of the rater carrying out the reliability exercise tending to create more subdivisions within statements than did the two earlier raters. This practice can lead to a duplication of construct poles and Landfield’s manual instructions advise vigilance on this point (Landfield, 1965).
Stage two involved the post coding of these construct poles in terms of Landfield’s categories and produced a correlation of 0.66 for the comparison of all the pairs of rater’s scores for the total number of categories in both groups of patients combined. For the agoraphobic group and the contrast group considered separately the reliability coefficient was 0.74 and 0.60 respectively.
There was a fair degree of variation in the reliability of rating the individual categories within and between the two sets of patients. This variation may in part be due to the level of disagreement in pole identification discussed above. The reliability coefficients for the five categories of interest to this study are examined below:
The category of greatest importance (‘tenderness’ [17a]), as it relates to the study’s main hypothesis, has quite an acceptable level of inter-rater reliability with a coefficient of 0.78 for the groups combined, 0.84 for the agoraphobic group and 0.77 for the contrast group.
Similarly inter-rater reliability for ‘organisation’ and ‘forcefulness’ constructs was good (0.80 overall, 0.95 agoraphobics, 0.75 contrast, and 0.80 overall, 0.89 agoraphobics, 0.84 contrast respectively).
‘Active social interaction’ and ‘closed to alternatives’ were less reliably identified however (0.53 overall, 0.57 for agoraphobics, 0.91 for contrast, and 0.57 overall, 0.74 for agoraphobics, 0.48 for contrast group respectively).
The above results suggest that in general the identification of construct poles within self-characterisations was moderately reliable and their post-coding reliability, particularly in relation to the main variables of interest, lay in the moderate to good range.
Hypothesis 1
As predicted, the agoraphobic patients described themselves as being tender minded on the ‘tenderness (high)’ scale 17a, (‘any statement denoting susceptibility to softer feelings towards others such as love, compassion, gentleness, kindness, considerateness, or the opposite’- (low is 17b).’ [Landfield, 1965, p. 12]), with a greater frequency than did the non-agoraphobic patient contrast group. Medians = 2.85 and 1.65 respectively, chi-square = 3.61, p<0.03, 1 tail (phi = .3, a medium effect), calculated using Yates correction for small expected values and dichotomising the data according to the median method described by McNemar (1962). [Median for both groups combined =1.11]. A Mann-Whitney test, adjusted for ties, gives an exact p = 0.0024 1 tail. There was no difference between the two groups in their expression of the low-tenderness pole 17b. Agoraphobic median = 0.70, non-agoraphobics = 0.76, chi-square = 0.45, p N.S. (1 tail with reference to Winter and Gournay (1987) finding).
Hypothesis 2
In retrospect not surprisingly, given that they were currently agoraphobic, and current active social interaction was greatly reduced, the agoraphobic group did not use statements in relation to ‘active social interaction’ at a level that distinguished them from non-agoraphobic patients. Chi-square = 1.83, N.S 1 tail, but these statements were expressed with a relatively high frequency by both groups (medians = 3.99 and 4.99 respectively). Even so, hypothesis 2 remains unconfirmed, due to the absence of non-patient control data.
Hypothesis 3
Contrary to the hypothesis the agoraphobic group did not produce significantly more statements classifiable as the construct pole category ‘closed to alternatives’, [Category 10d, closed to alternatives: ‘Fit: always realistic, avoidant, bigoted, conservative, dogmatic, inhibited, narrow-minded, never angry, no emotions, one-track mind, rigid, status quo,’ Landfield, 1965, p 9.] Chi-square =2.01, p N.S 1 tail. Agoraphobic median = 1.59 and Contrast median = 0.99.
Hypothesis 4
The agoraphobic group did use many self descriptions that involved ‘forcefulness (high)2a: ‘Any statement denoting energy, overt expressiveness, persistence, intensity, or the opposite’- (low 2b [Landfield,1965, p 4]), but so did the contrast group (medians = 4.59 and 3.39 respectively, chi-square = 1.5, p N.S. 1 tail). The ‘organisation (high)’ self statements were used with low frequency by both groups (medians = 1.16 and 1.28 respectively, chi-square = 0.85, p NS 1 tail). Hypothesis 4 was therefore only partly confirmed.
The ‘self now’ quality of a self characterisation makes this analysis difficult, in that ‘trait’ qualities may be affected by ‘state’ processes, from both a behavioural and attitudinal viewpoint. Even so, the confirmation of the agoraphobics’ greater tendency to perceive themselves as tender minded, relative to the non-agoraphobic group, might reasonably be taken to suggest that this self-description does reflect a pre-onset core construction and possibly the most important core construct. If as Hopkins (1995) suggests, the agoraphobic is experiencing ‘threat’ then PCT predicts that the effect of ‘threat’ is to compel the client to “claw frantically for (their) basic construct” (Kelly, 1963, p 167). The strong expression of this construct here may then reflect an elaboration of this quality. Again cautious interpretation is required due to lack of non-patient control data and the modest level of inter-rater reliability.
The agoraphobics’ freely expressed descriptions of themselves as being ‘forceful’ combine usefully with their ‘tender-mindedness’. ‘Forcefulness’ may not be a contradiction of ‘tenderness’ when applied in the service of others for this is consistent with this study’s view of the pre-agoraphobic person as helper and carer (Hopkins, 1995), but as has just been argued, as these people are currently agoraphobic the attribution of particular levels of these qualities to them pre-morbidly, on the basis of this data, needs to be received with caution.
Although the two groups were broadly equivalent on type and level of symptomatology and therefore a reasonable control for this source of possible causal linkage, a non-patient group could have been utilised as an additional control. It would then have been possible to indicate whether the agoraphobia group and the comparison group were more or less ‘tender minded’ or ‘forceful’ and so on, than a non-patient population.
In an effort to distinguish between ‘state’ and ‘trait’ characteristics the subjects could have been asked to produce two self-characterisations one ‘as I am now’ and one ‘as I was formerly’, that is, prior to the onset of agoraphobia. This may have led to a better demonstration of the hypothesised differences between the two clinical groups especially in the area of their social interaction and general effectiveness as individuals, prior to the onset of their complaints.
The reliability of the construct pole identification stage could be improved. In this study the reliability procedure took place at too late a stage for the raters to meet and confer and arrive at a consensus where there were differences in identification. In addition the training, of the main study’s raters and the reliability rater themselves, in both unitisation and subsequent coding was insufficient to achieve higher levels of agreement. Reliance on the Landfield procedure alone does omit material of interest within the context of the narrative, the identification of themes for example, and the balance of meaning provided by the contrast pole is lost with the separation of emergent poles from contrast poles in this coding process.
Further exploration of their use of these apparently differentiating constructs and other constructs is required whilst the patients are still agoraphobic and on their recovery. Such studies, reported on in Hopkins (1995), were given direction and impetus by the encouraging result of this first test of the PCT model of agoraphobia. A further analysis of constructs using the repertory grid method and the sophisticated statistical methods now available to describe the results will enable hypotheses to be tested that relate not only to content but also to the structure of the person’s construct system in its application both to themselves and to other people in their lives.
The analysis of each person’s self-characterisation utilised an adaption of Landfield’s (1965) method of construct ‘content postcoding’, thus converting the idiosyncratic results of the self-characterisations into a form that allows for comparisons to be made between individuals and groups. The major finding was that, as predicted, participants displaying agoraphobic avoidance describe themselves as being of a ‘nurturant’ or ‘tender-minded’ nature more frequently than equally depressed and anxious participants not displaying agoraphobic avoidance. The limitations of the results confirmed the need for additional control groups and reliability procedures, but it was judged that the data did provide reason to undertake a more detailed examination of the personal construction of agoraphobic people that would look not only at construct content, but would explore the possibility that differences exist at a structural level that may increase their vulnerability.


Chambless, D. L. & Goldstein, A. J. (1982). (Eds.) Agoraphobia: Multiple perspectives on theory and treatment. New York: John Wiley and Sons.
Chambless, D. L. & Goldstein, A. J. (1981). The treatment of agoraphobia. In M. Mavissakalian & D. H. Barlow (Eds.). Phobia: Psychological and pharmacological and pharmacological treatment (pp. 105-127). New York: The Guilford Press.
Chambless, D. L. & Goldstein, A. J., (1980a). The treatment of agoraphobia. In A. Goldstein & E. Foa (Eds). Handbook of Behavioural Interventions (pp. 322-415). New York: Wiley.
Chambless, D. L. & Goldstein, A. J. (1980b). Anxieties: Agoraphobia and hysteria. In A. Brodsky & R. Hare-Mustin (Eds.) Women and psychotherapy (pp.113-134). New York: Guilford.
Cohen, J. (1960). A coefficient of agreement for nominal scales. Educational and Psychological measurement, 20, 37-46.
Feeney, J.A., Noller, P. & Hanrahan, M. (1994) Assessing adult attachment. In M.B. Sperling & W.H. Berman (Eds) Attachment in Adults: Clinical and developmental perspectives (pp.128-154). New York: Guilford Press.
Goldstein, A. J. & Chambless, D. L. (1978). A reanalysis of agoraphobia. Behaviour Therapy, 1978, 9, 47-59.
Guidano, V. F. & Liotti, G. (1983). Cognitive processes and emotional disorders. New York: Guilford Press.
Hallam, R. S. (1985). Anxiety: psychological perspective on panic and agoraphobia, foreword by Theodore Sarbin. Academic Press, London.
Harre’, R. (1979). Social being, Basil Blackwell, Oxford.
Hopkins, N.J. (1995). On the origin of panic in agoraphobic women, Ph.D. Thesis, University of Sheffield.

Johnstone, M., Johnstone, D.W., Wilkes, H., Burnes, L.E., & Thorpe, G.L. (1984). Cumulative scales for the measurement of agoraphobia. British Journal of Clinical Psychology, 23,133-143.

Kelly, G. A. (1955). The psychology of personal constructs, Vols. I and II., New York: W. W. Norton and Company.
Kelly, G. A. (1963). The autobiography of a theory. In Maher, B. (Ed.), 1969 Clinical psychology and personality (pp. 46-65). New York: John Wiley and Sons.
Kelly, G. A. (1964). The threat of aggression (prepared for a conference on humanistic psychology, Old Saybrook, Connecticut, November 27-29th, 1964).
Landfield, A. W., (1965) Personal construct systems in psychotherapy. Functionally Independent Construction, University of Missouri.
Landfield, A. W. (1971). Personal construct systems in psychotherapy Rand McNally & Co., U.S.A.
Liotti, G. (1991). Insecure attachment and agoraphobia. In C. M. Parkes, J. Stevenson-Hinde & P. Marris (Eds.) Attachment across the life cycle (pp. 216-233). London: Routledge.
McAdams, D.P. (2008). Personal narratives and the life story. In O.P. John, R.W.Robbins & L. A. Pervin (Eds), Handbook of personality:Theory and research (pp.241-261). New York: Guilford Press.
McNemar, Q. (1962). Psychological statistics. (Third Edition), New York and London: Wiley.
Rowe, D. (1985) Personal communication.
Stefan, C. (1977) Core structure theory and implications. In D. Bannister (Ed.), New perspectives in personal construct theory. (pp. 281-298). London, New York, San Francisco: Academic Press.
Strodl, E. & Noller, P. (2003). The relationship of adult attachment dimensions to depression and agoraphobia. Personal Relationships, 10, 171-185.
Sarbin, T. (1985). Foreword. In R. S. Hallam, Anxiety: psychological perspective on panic and agoraphobia, London: Academic Press.
Winter, D. A. (1983). Constriction and construction in agoraphobia. Paper presented at 5th International Congress of Personal Construct Psychology, Pine Manor College, Boston, USA.
Winter, D. & Gournay, K. (1987). Construction and constriction in agoraphobia. British Journal of Medical Psychology, 60, 233-244.

Winter, D., Gournay, K., & Metcalfe, C. (1996). An investigation of the effectiveness of a personal construct psychotherapy intervention. In J.M. Fisher & D.J. Savage (eds.).  Beyond Experimentation into Meaning (pp. 146-160). Farnborough: EPCA Publications.
Winter, D., Gournay, K., Metcalfe, C., & Rossotti, N. (2006). Expanding agoraphobics’ horizons: an investigation of the effectiveness of a personal construct psychotherapy intervention. Journal of Constructivist Psychology, 19, 1-29.
Winter, D.A. & Metcalfe, C. (2005). From constriction to experimentation: personal construct psychotherapy for agoraphobia. In D.A. Winter & L.L. Viney (eds.). Personal Construct Psychotherapy: Advances in Theory, Practice and Research. (pp. 148-164). London: Whurr.



This study was carried out whilst the author was working as a Consultant Clinical Psychologist at Sheffield’s Northern General Hospital. It is one aspect of a broader investigation into the problems of panic and agoraphobia using a Personal Construct Theory approach submitted to the Faculty of Medicine of the University of Sheffield in part fulfillment of the degree of Doctor of Philosophy.



Nigel Hopkins: After completing training in Clinical Psychology at Birmingham University England in 1971 I worked for the next twenty-five years mainly with Adults. In 1996 I took up posts based in Forensic settings. From 2005 I worked in Cornwall with adolescents in Residential Care, and then, in Plymouth, I practised as the team psychologist. Starting out in 1971 with behavioural skills, and like many others supplementing these, first with Rogerian approaches and then with Rational Emotive Therapies, I moved onto Beck, and Cognitive Behavioural inspired ways of working. Around 1982 I began to take a serious ‘look’ at Personal Construct Theory and, following substantial training, made PCT my default starting point when putting case formulations together.



Hopkins, N. J. (2012). Is there a typical agoraphobic core structure?
Personal Construct Theory & Practice, 9, 19-27, 2012
(Retrieved from http://www.pcp-net.org/journal/pctp12/hopkins12.html)

Correspondence address: nigel.hopkins4@btinternet.com


Received: 12 September 2012 – Accepted: 10 December 2012 Published: 28 December 2012


ISSN 1613-5091

Last update 28 December 2012