Studies on the role of disgust in the acquisition and maintenance of specific phobias

BEHAVIOUR RESEARCH AND THERAPY

PERGAMON

Behaviour Research and Therapy 36 (1998) 877±893

Studies on the role of disgust in the acquisition and maintenance of speci®c phobias Susan J. Thorpe *, 1, Paul M. Salkovskis 2 University of Oxford, Department of Psychiatry, Warneford Hospital, Oxford OX3 7JX, UK Accepted 3 February 1998

Abstract Disgust has been proposed as a possible factor in phobic acquisition and maintenance, particularly in spider phobia. Cognitions and processes concerning disgust were examined in a series of studies with spider phobics, other speci®c phobics and nonphobic controls. Beliefs about the disgusting nature of their phobic objects were present in phobics but did not contribute to an attentional bias. Measures of global disgust sensitivity were not closely linked to the phobic fear response. The disgust associated with phobic objects appears to have di€erent constituents to the disgust associated with objects that do not evoke the phobic response. In the light of evidence presented here, it seems unlikely that disgust plays a central role in the aetiology or maintenance of spider phobia in particular and speci®c phobias in general. It is proposed that when stimuli normally associated with disgust become the focus of phobic anxiety the disgust response may be ampli®ed. # 1998 Elsevier Science Ltd. All rights reserved.

1. Introduction The traditional account of speci®c phobias has concentrated on the fear evoked by phobic objects or situations, rather than other emotional responses. Watts (1986) suggested that disgust was important in phobias, particularly spider phobia. He hypothesised that it was disgust rather than anxiety which disrupted spider phobics' memory for spider information (and which therefore interfered with habituation), as indicated by experiments examining biases in their memory for spider stimuli. In a series of papers, Graham Davey and colleagues proposed that the connection between phobias and disgust may have had adaptive value (Matchett and Davey, 1991; Davey, 1992a,b,c, 1994; Davey et al., 1993). He postulated that * Corresponding author. 1 2

Susan Thorpe is supported by the Wellcome Trust. Paul Salkovskis is a Wellcome Trust Senior Research Fellow.

0005-7967/98/$19.00 # 1998 Elsevier Science Ltd. All rights reserved. PII: S 0 0 0 5 - 7 9 6 7 ( 9 8 ) 0 0 0 6 6 - 7

878

S.J. Thorpe, P.M. Salkovskis / Behaviour Research and Therapy 36 (1998) 877±893

humans are more likely to acquire disgust-based avoidance of stimuli which have been evolutionarily associated with disease or contamination. This is an extension of Seligman (1971) preparedness theory which states that particular classes of stimuli are more likely to be associated with conditioned fear responses due to evolutionary pressures. Seligman particularly cited the conditioned nausea response as one which was persistent and very easily acquired, and suggested that this had evolved in mammals because of the need to protect against poisoning. Support for this was to be found in the work of Garcia and colleagues (Garcia and Koelling, 1966; Garcia et al., 1977) who had discovered that taste aversions can be conditioned in a single trial, and that once acquired, this conditioning is particularly long-lasting and dicult to undo. This appears to be true only when there is an evolutionary based `belongingness' relationship: in the conditioning of taste aversion this means that the CS has to be novel taste (as opposed to auditory and visual stimuli) and the UCS has to be nausea (as opposed to pain). If disgust is involved it is not clear why this type of process should be relevant to phobias, in which the stimuli are usually visual and auditory, and reported conditioning experiences seldom involve nausea. Furthermore, the person who develops a lifelong aversion to shell®sh after an episode of food poisoning does not become phobic of mussels. Rozin and Fallon (1987) were amongst the ®rst to study disgust experimentally and to put it in the context of contamination sensitivity. They describe the disgust reaction as primarily a food-rejection response consisting of nausea, a characteristic facial expression, a distancing of the self from the o€ensive object, a sensitivity to both contamination from the object and the prospect of eating it, and a characteristic feeling-state (revulsion). They were also interested in the question of contagion, where an object could subjectively acquire the characteristics of another object by being physically close to it (Rozin et al., 1986) and similarity, where this happens when the object looks similar to another. So perhaps the disgust response in phobias is acquired through association in some way as yet unspeci®ed, and this may in part account for the development of the phobic response. In order to examine the question of disgust/ contamination sensitivity, Rozin et al. (1984) devised a scale based on how much those participating in the experiment would like to eat a bowl of soup from a dog bowl, stirred by a ¯y swatter, containing a dead grasshopper and so on (20 items in all). Additionally, they asked them how much they would like to eat a biscuit after a bite had been taken by an acquaintance, a good friend, a waiter in a restaurant or after it had fallen, unbitten, on the lawn while they were picnicking. They demonstrated an association between the perception of contamination, and avoidance of the object. Evidence for a speci®c association between the measure of disgust sensitivity and the degree of fear provoked by di€erent classes of animals has come from Matchett and Davey (1991), who found that measures of disgust and contamination sensitivity were correlated with scores on the animal subscale of the Fear Survey Schedule (Wolpe and Lang, 1964). In addition, disgust and contamination sensitivity correlated with fear of animals which were characterised as fear-relevant and not physically harmful (camel, rat, eagle, spider, cockroach, maggot, snake) and also correlated with fear of animals characterised as having high revulsion (maggot, cockroach, slug, rat, snake, spider, snail). This contrasted with the lack of correlation with animals they considered to be physically harmful, attacking or predatory (shark, tiger, lion, bear, snake, jelly®sh, wolf). It is not clear why snakes should appear in all categories and rat,

S.J. Thorpe, P.M. Salkovskis / Behaviour Research and Therapy 36 (1998) 877±893

879

cockroach, spider and maggot in two categories. It should also be noted that these results were obtained from nonphobics. Matchett and Davey's suggestion that disgust sensitivity may be a vulnerability factor for the development of phobias is based on these ®ndings. If, as suggested, disgust is implicated in the development of phobias, it is to be expected that phobics who ful®ll clinical diagnostic criteria for speci®c phobia should have signi®cantly higher levels of disgust sensitivity than do fearful, but nonphobic individuals. Mulkens et al. (1996) have shown that spider phobic women are indeed more sensitive to disgust than are non spider phobic women, as measured by the Rozin and Fallon disgust sensitivity questionnaire, and by one of their own behavioural tests (reluctance to eat a biscuit over which a spider had crawled). However, in another behavioural test of disgust sensitivity, which did not involve a spider (drinking tea out of a scale-encrusted cup) there was no di€erence between the groups. One possible explanation for this and for the other reported di€erences is that the disgust sensitivity questionnaire, on which the majority of research in this area is based, consists of 24 items, 9 of which are concerned with the amount of disgust generated by insects: a dead grasshopper (6) and a ¯y swatter (3). Spider phobics may be particularly sensitised to these kinds of items purely because of the nature of the stimuli, i.e. `leggy' insects. Another possible explanation is that Ð in the test using the biscuit and the spider Ð the spider phobics were responding to a residual phobic reaction about the spider itself rather than to the biscuit as contaminated by the spider. In support of this, Thorpe and Salkovskis (1995) found that while phobics do have a high of level of endorsement of beliefs about a variety of disgusting qualities which their phobic objects possess, this does not appear to be related to the intensity of their fear. It is possible that, when phobic fear becomes associated with the stimuli which are commonly also associated with disgust, the fear may amplify the `normal' disgust response as part of the general negative evaluation of the phobic object. This would suggest that disgust may not be causally related to phobias. The present paper describes three studies which were designed to explore the links between disgust and phobias. In the ®rst, people ful®lling criteria for simple phobia completed fear and disgust ratings of the type used by Matchett and Davey with the possible confound of categories of animals examined in detail. This study assesses the generalisability of their earlier work with nonphobic respondents, to a phobic population. The second study looks at the way in which spider phobics, other speci®c phobics and nonphobics react to both spider stimuli and to stimuli which are generally regarded as `disgusting'. This allows a comparison of disgust reactions to phobic stimuli and to disgustevoking nonphobic stimuli of comparable intensity. The study thus aims to clarify the relative importance of (i) increased sensitivity to disgust-evoking stimuli in phobic individuals and (ii) the disgust response speci®c to spiders in particular. The ®nal part of the second study explores the possibility that spider phobics have an attentional bias to disgusting stimuli, at both the conscious and preconscious level, which would support the hypothesis, described above, that spider phobics have a greater disgust sensitivity than do nonphobics. The third study seeks to provide a further analysis of the relative contribution of fear and disgust to the phobic response. Spider phobics, other phobics and nonphobics chose the two items which disgusted them most then rated them on a variety of measures of disgust, fear, avoidance and physical characteristics.

880

S.J. Thorpe, P.M. Salkovskis / Behaviour Research and Therapy 36 (1998) 877±893

2. Methods and results The three studies are described below. For clarity, assessment instruments used in the ®rst two studies will be described ®rst, followed by the methods and results for each of the studies in turn. The questionnaire used in study 3 will be described along with the study. 2.1. Measures 2.1.1. Descriptive questionnaire This was a series of 14 questions which dealt with demographics (age of onset of phobia, familial factors in phobia) descriptions of phobic incidents and coping strategies. It also asked subjects to list no less than 5 words describing their phobic object. 2.1.2. Depression Beck depression inventory (Beck et al., 1961). 2.1.3. Anxiety Spielberger STAIy (state-trait anxiety index) self evaluation questionnaire (Spielberger, 1983). 2.1.4. Phobic intensity Participants were asked to rate on a scale of 0 to 8: (i) how much they would avoid a spider because of fear or other unpleasant feelings; (ii) how much fear or other unpleasant feelings they would experience if in the proximity of a spider; (iii) how much the problem upsets them and/or interferes with normal activities. They were also asked to ®ll in a visual analogue containing seven items rated on a scale of 0 to 100, in steps of 10, as follows: (a) I do not feel at all anxious/I feel extremely anxious, (b) I do not want to escape at all/I want to escape very much, (c) I do not ®nd the spider at all disgusting/I ®nd the spider extremely disgusting, (d) I can easily cope with my anxiety about spiders/I cannot cope with my anxiety about spiders at all, (e) I feel I am coping with this situation very well/I do not feel I am coping with this situation well, (f) I do not feel at all anxious about spiders/I feel extremely anxious about spiders, (g) I could easily approach and deal with a spider/I could not approach and deal with a spider. 2.1.5. Spider and other speci®c phobia The Watts and Sharrock (1984) spider phobia questionnaire (WQS) was used in its original form of 43 forced choice questions to do exclusively with spiders, as well as in an amended form of 30 forced choice questions which could be applied to other phobias. 2.1.6. Disgust Matchett and Davey (1991) Fear of Animals Questionnaire. This is a list of animals which fall into three categories. (i) `High fear/high predatory': animals which were rated as highly likely to attack and harm human beings and were also highly fear-evoking, (shark, tiger, lion, bear, snake). (ii) `High fear/low predatory': animals rated relatively low on tendency to attack

S.J. Thorpe, P.M. Salkovskis / Behaviour Research and Therapy 36 (1998) 877±893

881

human beings but which were still rated high in fear evocation (rat, eagle, spider, cockroach, maggot, snake). (iii) `High revulsion': animals that were rated as evoking high revulsion responses (maggot, cockroach, slug, rat, snake, spider, snail). When these were reconstituted with exclusive categories the lists of animals were as follows (i) High fear/high predatory: shark, tiger, lion, bear, snake. (ii) High fear/low predatory: rat, eagle, spider. (iii) High revulsion: maggot, cockroach, slug, snail). The fear ratings of animals in each category were combined to give a single fear score for each of the categories of animals. 2.1.7. Disgust sensitivity Disgust and contamination sensitivity. This variable was measured by the only previously validated measure of disgust sensitivity, that of Rozin et al. (1984) as constructed for their study of family resemblances in attitudes to food. This was amended by Matchett and Davey (1991) in order to remove Americanisms and consists of a series of 24 questions all involving some sort of contamination of normal food. Participants rated on a 9 point scale how much they would like to eat each of the items. An overall measure of disgust sensitivity was obtained by summing the scores. 2.1.8. The fear survey schedule (FSS) This was used to measure fear responses to various categories of objects (Wolpe and Lang, 1964). This is a 73 item scale divided into 6 categories of anxiety provoking stimuli: (i) animals, (ii) social and interpersonal, (iii) tissue damage, illness, death and their associations, (iv) noises, (v) other classical phobias and (vi) miscellaneous.

3. Study one 3.1. Method: links between disgust and phobias This study sought to replicate the Matchett and Davey results Ð which found links between disgust sensitivity and fear of animals in nonphobics Ð in participants who were phobic. 3.1.1. Participants These were 25 nonstudent volunteers from a variety of backgrounds and occupations, ranging in age from 18 to 42 with a mean age of 23.6 years. 17 were female, 8 male. All ful®lled the criteria for phobia as de®ned by achieving a score of 14 points on the Watts and Sharrock (1984) spider phobia questionnaire and an equivalent cut o€ point on an amended general version for other phobias (Thorpe, 1990). The mean of scores for the spider phobics was 22, and for other phobics was 16. Types of phobia were as follows: 6 spiders, 4 heights, 2 patterns (leaf patterns and networks), 2 snakes, and one each of; blood/injury, vomit, dentists, dogs, cats, slugs, maggots, birds, rodents, buttons and claustrophobia. 3.1.2. Procedure Phobics were contacted by post. If they ful®lled the criteria for speci®c phobia they were sent the questionnaires which they ®lled in and returned for analysis.

882

S.J. Thorpe, P.M. Salkovskis / Behaviour Research and Therapy 36 (1998) 877±893

Table 1 Correlations between disgust sensitivity score, trait anxiety and factors of the fear survey schedule in nonphobics (Matchett and Davey, 1991) and phobics Factors of the FSS animal Nonphobics Disgust sensitivity Trait anxiety Phobics Disgust sensitivity Trait anxiety *

p < 0.05;

**

social

illness

noises

classical

miscellaneous

ÿ0.3171*** 0.1692

ÿ0.1966 0.4986***

ÿ0.3253*** 0.1634

ÿ0.207 0.1467

ÿ0.1686 0.2228*

ÿ0.0897 0.3931**

ÿ0.3117 ÿ0.0713

ÿ0.4926* 0.4852*

ÿ0.4224* 0.1743

ÿ0.3508 0.2263

ÿ0.1227 0.2318

ÿ0.5162* 0.3978*

p = 0.052;

***

p < 0.001.

3.2. Results Matchett and Davey previously noted correlations between disgust sensitivity and fear ratings in unselected student volunteers. The ®rst study used these same measures with a range of phobics: a correlational analysis was therefore carried out. Table 1 shows these results together with those reported by Matchett and Davey (1991). The results of this analysis are broadly consistent with the disgust theory, although the wider range of scores might have been expected to produce larger correlations. However, there was no evidence found of a correlation between the phobia speci®c avoidance rating and disgust sensitivity (r = 0.028, p = 0.9). Results may have been confounded by the fact that snakes appear in all three categories and rat, cockroach, spider and maggot in two categories. When the scores are recalculated with exclusive categories the correlations for the two categories in which disgust would be most likely to be important (i.e. those relatively low in threat and high in revulsion) disappear (as again shown in Table 2).

Table 2 Correlations between disgust sensitivity and the mean fear scores for groups of animals rated as high fear/high predatory, high fear/low predatory and high revulsion Disgust sensitivity HFHP Matchett and Davey (normals) All phobics All phobics (exclusive categories)

HFLP

HREV **

ÿ0.1124

ÿ0.2937

ÿ0.2824**

ÿ0.1391 ÿ0.1570

ÿ0.1512 ÿ0.0782

ÿ0.1407 ÿ0.0148

Note (DNI = disgust not involved, DI = disgust involved).

S.J. Thorpe, P.M. Salkovskis / Behaviour Research and Therapy 36 (1998) 877±893

883

4. Study two 4.1. Method: reaction to spider stimuli and to disgust stimuli 4.1.1. Overall procedure Participants ®lled in the questionnaires then went on to take part in the three experiments in this study in one session. After each part they ®lled in the state anxiety questionnaire and the phobic intensity questionnaire as described above. 4.1.2. Participants These were 75 volunteers from a variety of sources, of whom 21 were nonphobics, 20 were phobic controls and 34 were spider phobics. All the phobics ful®lled DSM-III R criteria for simple phobia and in addition scored above the cut-o€ point in the Watts and Sharrock (1984) spider phobia questionnaire or (in nonspider phobic phobics) the general phobia questionnaire, in which items from the Watts and Sharrock were amended to substitute a general term for the speci®c phobic object. The groups did not di€er in measures of trait anxiety or depression, nor were they di€erent in a variety of demographic features. Participant characteristics are shown in detail in Table 3, including a variety of measures concerned with spiders. There was no di€erence in sex ratio between the groups (w2[2 df ] = 1.743, p = 0.418).

Table 3 Means of participant characteristics Group

Nonphobic

n Sex ratio (f:m) Age (years) Age of onset WQS Spielberger trait anxiety Rozin and Fallon BDI Avoidance Fear Interference Spielberger state anxiety General anxiety Desire to escape Find spiders disgusting Coping: anxiety re spiders Coping: situation Anxiety about spiders Approach and deal with spiders *

p < 0.01,

***

p < 0.0001.

21 18:3 25.4
5 35.86 131.26 4.00 1.00 1.20 0.33 28.47 6.00 3.33 12.00 6.67 5.33 12.67 18.00

Phobic controls 20 15:5 24.1 9 6.05 36.70 130.15 5.65 1.22 1.00 0.61 33.00 15.56 8.33 12.22 11.11 7.78 12.22 16.67

Spider phobics 34 30:4 26.3 3.8 22.50*** 40.85 116.29 6.97 6.00*** 5.18*** 3.65*** 41.74* 50.00*** 24.71* 66.47** 60.88*** 33.82*** 67.35*** 78.53***

884

S.J. Thorpe, P.M. Salkovskis / Behaviour Research and Therapy 36 (1998) 877±893

4.2. Part 1: spider attributes 4.2.1. Procedure 40 spider phobics (6 from the ®rst study and 34 from the second set of studies) were asked to list no less than 5 words describing a spider. The words were collated and mentions of each word totalled. 4.3. Results Spider phobics chose a total of 98 physical appearance words, 56 movement words, 14 disgust words (if `horrible' is categorised as disgust), and 22 evaluative words (ugly, menacing and evil). This is comparable to Davey's method and results. This suggests that movement is a more important characteristic of spiders for phobics than is disgust. In a further exploration of this, disgust sensitivity scores from all three groups were compared. There was no signi®cant di€erence between disgust sensitivity scores obtained from the spider phobic, other phobic and nonclinical controls (F = 1.36): spider phobics scored a mean of 116, nonphobics 131 and other phobics 130 (low score indicates high disgust sensitivity). Unfortunately, mean scores for Matchett and Davey's nonphobic group are not available. 4.4. Part 2: disgust sensitivity and its relation to fear 4.4.1. Method This experiment is divided into two parts. In the ®rst, participants were shown 40 pictures from 7 categories of pictures Ð including disgust relevant and irrelevant Ð and were asked to rate them for disgust. In the second part, 20 pictures of spiders were ®rst rated for disgust and then for fear. This part of the study was designed (i) to evaluate the relative intensity of the disgust reaction to spiders (compared with other potentially disgustevoking stimuli) in spider phobics and (ii) to examine the relationship between disgust and fear elicited by spiders alone. 4.4.2. Materials For the ®rst set, spider phobics, other speci®c phobics and nonphobic controls were shown 40 pictures divided into: 10 carnivorous plants and 5 each of the following: `aesthetically pleasing' spiders, revolting spiders, owls, slugs, centipedes and human disease. Spiders were divided into the two types, based on prior ratings given by nonphobic evaluators. Pictures were taken from a variety of sources, were all in colour and were mounted on pastel A4 card and placed in the transparent plastic leaves of a presentation folder. All were photographs in natural settings, except for the human pictures which were obtained from a book of diagnostic picture tests for use in clinical medicine, and were consequently con®ned to particular anatomical areas. For the second set, 20 colour pictures of spiders cut from `SPIDERS an illustrated guide' by Rod Preston-Mafham (1991) were used. Pictures were mounted on A4 paper and put in

S.J. Thorpe, P.M. Salkovskis / Behaviour Research and Therapy 36 (1998) 877±893

885

individual transparent leaves in a plastic wallet. All were taken in the wild in a natural setting and were of a variety of spiders. 4.4.3. Procedure Subjects were seated at a small table. The experimenter sat on the opposite side of the table and gave the following instructions for the ®rst picture set (mixed pictures of spiders, plants, owls, human disease, centipedes and slugs): I am going to show you some pictures. As we go through them I want you to tell me how disgusting you judge them to be on a scale of 0 to 10, where 0 is not at all disgusting and 10 is extremely disgusting. Do not think too hard about your answer: it is your immediate response which I want. Please tell me when you are ready to begin. The experimenter opened the folder and showed each picture for about 10 s to the participant who gave a response in the form of a number from zero to ten. For the second series (spider pictures only) the instructions above were repeated and the set of 20 spider pictures was shown. The experimenter noted the disgust rating for each. Immediately after this, a fear rating to the same set of pictures was obtained using the following instruction: You are on a country walk when you come across a spider. It is about a metre away from you. Look at the pictures again and as we go through them tell me how frightened you would be if you saw the spider in that situation. Use a scale of 0 to 10, where 0 is not at all frightened and 10 is the most frightened you have ever been. The ratings for the ®rst, mixed, set of pictures allows a comparison of the disgust response to spiders and to other disgust-evoking stimuli in the three participant groups. The ratings for the second set (spider pictures only) allows a comparison of fear and disgust responses in the three groups to spider stimuli alone. 4.5. Results 4.5.1. Disgust ratings for ®rst set of pictures Mean scores were obtained for fear ratings of the forty mixed pictures. These were subjected to a two way analysis of variance with one between group factor (control, other phobic or spider phobic) and one within group factor (type of picture). There was a signi®cant main e€ect of group (F(2,66) = 9.86, p < 0.0001) and of picture type (F(6,396) = 48.35, p < 0.0001) There was a signi®cant interaction between group and picture type (F(12,396) = 7.06, p < 0.0001). Analysis of simple main e€ects indicated that spider phobics found both sets of spider pictures signi®cantly more disgusting than did the other two groups: revolting spiders (F(2,66) = 19.5, p < 0.0001) and aesthetically pleasing spiders (F(2,66) = 30.5, p < 0.0001). In addition, there was a main e€ect of

886

S.J. Thorpe, P.M. Salkovskis / Behaviour Research and Therapy 36 (1998) 877±893

group for centipedes (F(2,66) = 3.91, p < 0.025). Multiple comparisons con®rmed that both the spider phobics and the other phobics found centipedes signi®cantly more disgusting than controls. There were no other di€erences between the groups. Means of each of the categories can be found in graph form in Fig. 1. Spider phobics found the aesthetically pleasing spiders less disgusting than centipedes and human disease, and found slugs only marginally less disgusting than the aesthetically pleasing spiders. 4.5.2. Disgust and fear ratings for the second set of pictures Mean scores were obtained for fear and disgust ratings of twenty spider pictures. One way analyses of variance were carried out on fear ratings and disgust ratings for each of the groups. For fear ratings, there was a signi®cant main e€ect of group (F(2,66) = 40.7, p < 0.0001) Multiple comparisons con®rmed that the spider phobics were signi®cantly more fearful of the spider pictures than were the other two groups. A similar result was obtained for disgust ratings (F(2,66) = 37.47, p < 0.0001). Again, multiple comparisons con®rmed that the spider phobics were signi®cantly more disgusted by the spider pictures than were the other two groups. Scores of fear and disgust were correlated in all three groups: spider phobics r = 0.73, p < 0.0001; other phobics r = 0.64, p < 0.01; controls r = 0.64, p < 0.025. These results suggest that there is an association between self-rated disgust and fear of spiders.

Fig. 1. Mean disgust ratings for all categories of pictures: aps = aesthetically pleasing spider; rs = revolting spider; plant = carnivorous plant; cent = centipede; slug = slug; owl = owl; hum = human disease.

S.J. Thorpe, P.M. Salkovskis / Behaviour Research and Therapy 36 (1998) 877±893

887

4.6. Part 3: the Stroop test 4.6.1. Method An information processing task was chosen as the next measure in an attempt to identify the extent to which disgust concepts are an important element of the phobic experience as evidenced by possible conscious and preconscious processing biases. This experiment used the Stroop colour naming task using both masked and unmasked words (that is, unavailable and available to consciousness) with the groups previously de®ned (spider phobics, other speci®c phobics and nonphobic controls). The Stroop procedure used here was identical to that reported in Thorpe and Salkovskis (1997a). In that study, a threat-speci®c e€ect was identi®ed for spider words only in the unmasked (`strategic processing') presentation. Disgust words were included in that experiment, and it is those results which are reported here. 4.6.2. Stimulus items The words used as experimental stimuli were drawn from four classes of words: spider, disgust, emotional and neutral. All groups of words were matched on average length and word frequency with all other groups. Stimuli were as follows: 1. 2. 3. 4.

Disgust words: vomit, disgusting, pus, diseased, decay, faeces. Emotional words: harmony, ecstatic, fatal, wounded, painful, joyous. Neutral words: sponge, rectangular, wrap, woven, windy, seating. Spider words: spider, tarantula, web, scuttling, cobweb, lurking.

4.6.3. Treatment of data Means were trimmed to 2 standard deviations in accordance with convention (Watts et al., 1986). 4.7. Results The data are summarized in Table 4. An analysis of variance was carried out using one between group factor (control, other phobic or spider phobic) and two within group factors: masking mode (masked vs. unmasked) and wordtype (disgust, emotional, neutral or spider). There was no overall main e€ect of group (F(2,66)=1.94, p>0.1). There was a signi®cant main e€ect of masking (F(1,66) = 161.03, p < 0.0001). There was no signi®cant group  masking interaction (F < 1). There was no signi®cant main e€ect of wordtype (F < 1). There was a signi®cant group  wordtype interaction (F(6,198) = 2.34, p < 0.05) which was accounted for by the spider phobics' signi®cantly longer reaction time latencies to spider words only (cf. Thorpe and Salkovskis, 1997a for detailed analysis of spider word results). There was no signi®cant masking  wordtype (F < 1) interaction, or group  masking  wordtype (F < 1) interaction.

888

S.J. Thorpe, P.M. Salkovskis / Behaviour Research and Therapy 36 (1998) 877±893 Table 4 Trimmed mean colour naming latencies in ms. (Standard deviations in parentheses) Nonphobics

Phobic controls

Spider Phobics

Masked Disgust Emotional Neutral Spider

631 620 626 620

(63) (61) (64) (66)

572 591 576 577

(68) (87) (74) (75)

596 593 595 598

(61) (59) (54) (53)

Unmasked Disgust Emotional Neutral Spider

696 685 687 693

(82) (84) (93) (79)

645 651 647 636

(107) (97) (99) (92)

674 664 676 694

(84) (93) (79) (104)

In order to ensure that disgust word e€ects were not being lost as a result of this relatively strong e€ect of spider words, an additional analysis was carried out on the disgust, neutral and emotional words alone, without the spider words There was no overall main e€ect of group (F(2,72) = 1.82, p>0.1) As expected the masked  unmasked main e€ect was signi®cant (F(1,66) = 150.73, p < 0.0001, but the group  masking interaction was not (F < 1). There was no signi®cant wordtype main e€ect (F < 1), wordtype  group interaction (F(4,132) = 1.59, p>0.1), masking  wordtype (F < 1) interaction, or group  masking  wordtype (F < 1) interaction. These results indicate that the groups did not di€er in the way they processed the varying wordtypes (disgust, emotional and neutral). Neither spider phobics nor a mixed group of phobics appear to process disgust words di€erently from nonphobics. Given that our previous studies (Thorpe and Salkovskis, 1997a,b) detected spider word interference only in the unmasked presentation (consistent with Watts et al., 1986), that analysis was carried out on disgust words in an attempt to ensure that an e€ect was not being overlooked. An analysis of variance was carried out using one between group factor (control, other phobic or spider phobic) and one within group factor (wordtype: disgust, emotional or neutral). There was no overall main e€ect of group (F < 1). There was no signi®cant main e€ect of wordtype (F < 1). There was no signi®cant group  wordtype interaction (F < 1). Thus, when the analysis was conducted in the way which previously identi®ed spider speci®c interference most clearly, there was no evidence of any disgust speci®c interference.

5. Study three 5.1. Method: disgust questionnaire As previous work on the disgust response has been based on a limited set of questions, this study sought to explore disgust reactions in more detail.

S.J. Thorpe, P.M. Salkovskis / Behaviour Research and Therapy 36 (1998) 877±893

889

5.1.1. Participants 154 participants were drawn from a wide range of sources. They each completed two questionnaires about disgust. The questionnaires were ®lled in anonymously. Of those ®lling in the questionnaires, 69 were nonphobic and 85 were phobic, dividing into the following categories of phobia: 1 faeces, 5 blood, 9 small animals, 6 invertebrates, 37 spiders, 7 insects, 1 rotting meat, 1 vomit, 2 injections, 13 heights and 3 miscellaneous. 5.1.2. Measures The two questionnaires were identical and consisted of 36 items concerning the characteristics of disgust, including attributes of the object's physical appearance as well as the responses (physiological, emotional, cognitive) it provoked. Participants were asked to note down the object that most disgusted them and to keep it in mind while they answered the questions: they then endorsed one of ®ve possible responses ranging from `never' to `always' for a series of statements about the object. They then competed an identical questionnaire but this time were asked to note down and think about their second most disgusting object. It was anticipated that when disgust was relevant to the phobic object, phobics would specify this as their most disgusting object, allowing comparison with nonphobics' most disgusting object. The second questionnaire (which was a nonphobic disgust object in all instances) would allow a comparison of the scores of both phobic and nonphobic respondents for a disgusting, nonphobic, object. This comparison would ascertain whether any di€erences found for the ®rst questionnaire were likely to be due to it being phobic, or due to di€erences between groups in terms of their disgust reaction to all objects. 5.2. Results The scores of a mixed group of phobics on the ®rst questionnaire was compared with nonphobics on a variety of disgust, avoidance or fear-related items. Only those phobics whose phobic object matched their disgust object were used in the analysis, so 31 phobics were excluded from the analysis, leaving 54 phobics whose phobic and disgust object matched. If disgust is important in the aetiology of disgust relevant phobias then phobics should rate disgust associated with their phobic object more highly than nonphobics do about an object disgusting to them but about which they are not phobic. The ratings of 54 phobics and 69 nonphobics on a variety of statements concerning the object they considered most disgusting were collected and compared. The results relating to the `most disgusting object' are shown in Table 5. Items have been divided into categories of fear-relevant, disgust-relevant and avoidance-relevant for ease of comprehension. Phobics are more likely than controls to rate fear or avoidance statements higher in relation to their phobic/disgusting object (12/14 instances). Phobics were less likely to rate their phobic object as associated with disgust items than controls (10/22 instances). For the second most disgust-evoking item (which did no involve the phobic object for phobics), there were few di€erences. Phobics rated 2 items signi®cantly higher than the controls

890

S.J. Thorpe, P.M. Salkovskis / Behaviour Research and Therapy 36 (1998) 877±893

Table 5 Signi®cant di€erences between phobics and nonphobics with regard to disgusting objects. P = phobic, C = control Fear I am afraid I become paralysed I start to sweat I tremble It can move fast My skin crawls

Disgust/contamination P>C P>C P>C P>C P>C P>C

My nose wrinkles I close o€ my nostrils I feel nauseous I think it is slimy I think it smells My stomach heaves I think it is dirty I think it could I could swallow it It is biological

Avoidance P < C** P < C** P < C* P
No di€erences were found between the groups on the following statements I think it is harmful I grimace I feel contaminated I hold my breath I dislike the thought of I cover my mouth I feel dirty I could eat something it I want to wash my hands I feel unclean in its I wash my hands I think it looks diseased I feel unclean in its presence p < 0.001 except where

I leave quickly I would get help to I will avoid places I I try to avoid it It sounds repellant I would not touch it

P>C P>C P>C P > C* P > C* P > C*

I look away

**

p < 0.01 and *p < 0.05.

(``I am afraid'': p < 0.001; ``I would get help to remove it'': p < 0.01) and two items signi®cantly lower than the controls (``I think it smells unpleasant'': p < 0.05; ``It is biological'': p < 0.05). 6. Discussion The results of the questionnaires and experimental studies reported here provide little evidence to support the view that disgust plays a major role in phobias. In study 1 the association between disgust sensitivity and animal fears was similar to that previously found by Matchett and Davey (1991) when the same type of analysis was used. However, in a subsequent analysis using exclusive categories the association appeared to be greatly attenuated in the low fear categories where the association should be strongest. In study 2, spider phobics were found not to di€er from other phobics and nonphobic controls in their disgust sensitivity scores. When spider phobics and the control groups were asked to rates pictures of di€erent classes of objects including `revolting' and `aesthetically pleasing'

S.J. Thorpe, P.M. Salkovskis / Behaviour Research and Therapy 36 (1998) 877±893

891

spiders, human disease and so on, spiders were rated as generally more disgusting. However, spider phobics did not ®nd the aesthetically pleasing spiders more disgusting than the centipedes or the pictures of human disease. For the second set of pictures, both disgust and fear ratings were assessed: these measures were highly inter-correlated in all three groups. On the Stroop colour naming task, spider phobics showed evidence of colour naming interference for spider words relative to controls, but there was no sign at all of interference for disgust words. In the third study, phobics and nonphobics were compared in terms of their scores for the object they found most disgusting. In the majority of phobias this was their phobic object: comparisons between the phobic and nonphobic groups were con®ned to those phobics in whom this was the case. In this analysis, phobics did not di€er from nonphobics in the way they made ratings concerning disgust for 12 items and were signi®cantly lower in ratings of 10 items. They rated six fear statements in relation to their disgusting object signi®cantly higher than did controls and did not di€er in their responses to one item. six statements concerning ideas of avoidance were rated higher by phobics and one statement was rated the same. Taken together, these results are most consistent with the notion that being phobic of something tends to amplify feelings of disgust for things which others ®nd disgusting in any case. In the third study also it is of particular note that phobics were signi®cantly less disgusted by their phobic object (which was the thing they found most disgusting) than the nonphobics were by their most disgusting object. Results of study 1 were comparable to those found by Matchett and Davey (1991). The correlations between disgust sensitivity and fear scores in a group especially chosen as having phobias speci®cally associated with disgust were similar to Matchett and Davey's. However, this measure of disgust sensitivity may be problematic when used with spider/animal phobics because of the inclusion of grasshoppers and ¯ies in the probe questions: it is possible that only those items were associated with fear. A more recent disgust questionnaire by Haidt et al. (1994) would be more appropriate for future explorations of disgust sensitivity. It should also be noted here that data from the second study con®rmed that spider phobics, other phobics and nonphobics did not di€er signi®cantly in their disgust sensitivity scores, though spider phobics' scores were slightly lower (the lower the score, the greater the disgust sensitivity). A larger sample may have resulted in this di€erence becoming signi®cant, but there was little evidence of a trend in the present data. If disgust sensitivity is causally related to phobic responses to spiders, one might expect to ®nd an attentional bias in the processing of disgust-speci®c stimuli in phobics who show attentional bias to spider words. The Stroop test here, using disgust words, neutral control words and emotional control words found no e€ect for disgust words in the colour naming task in any of the groups, despite a spider word e€ect being found (Thorpe and Salkovskis, 1997a). With regard to the salience of spider features and the relative importance of disgust as opposed to fear-relevant features, `legs', `hairy', `creepy', `black' and `fast' were the attributes most commonly reported by spider phobics. Disgust appeared to be a constituent part of spider phobics' concept of what a spider is (`disgusting' and `ugly' were noted but were only the seventh and eighth most frequently reported adjectives on the list) but for spider phobics the most obvious and compelling characteristics spiders possess are those relevant to movement

892

S.J. Thorpe, P.M. Salkovskis / Behaviour Research and Therapy 36 (1998) 877±893

and texture rather than those characteristics more commonly associated with disgust like bodily secretions or waste. When asked during treatment what was so disgusting about spiders (Thorpe and Salkovskis, 1997b), patients' most common reply was `legs', not that the spiders were dirty or disease-ridden. Spider phobics are clearly more disgusted by spiders than are non spider phobics. However, spider phobics do not ®nd all spiders more disgusting than other repulsive and disgusting objects: spider phobics found the `revolting' spiders more disgusting than other objects but found the centipedes and human pictures at least as disgusting as the more `aesthetically pleasing' spiders. Spider phobics and other phobics were similar in their assessment of the disgustingness of centipedes, and both groups were more revolted by them than were the nonphobics. Taken together, these results suggest that the evaluation of disgust is based partly on aesthetics rather than entirely on contamination fear (although aesthetic judgement itself could be based on more primitive emotions, including, of course, contamination fear). Consistent with this, Bennett-Levy and Marteau (1984) found a high correlation between `ugliness' and `fear'. Ugliness ``incorporated elements of slimness, hairiness, colour of animal, perceived dirtiness, number of limbs and antennae projecting from the body, compactness of body and relation of eyes to the head'' (p. 41). They interpret this as support for the idea that the more discrepant an animal is from the human form, the more aversive it becomes to humans. There was a strong relationship between the reaction of disgust to a picture of a spider and the amount of fear that picture evokes. The association in the control groups may have been due to the fact that most of their ratings were series of zeros. The third study was designed to clarify the issue of whether disgust causes phobic type reactions or vice versa. The results of this study, using the disgust questionnaires, indicate that the majority of the phobics who completed the questionnaires did ®nd their phobic object disgusting. However, if disgust plays a causal role in phobias it would be predicted that `phobic disgust' (when it coincides with the persons `most disgusting object') should be more intense than disgust associated with the most disgusting object of nonphobics. The opposite was found. These results suggest that the phobic disgust response may be qualitatively di€erent from the nonphobic disgust response. The analysis of responses to the `second most disgusting object' (which was not phobic in the phobic group) showed that phobics respond in the same way to nonphobic disgusting objects as do nonphobics, apart from two statements which they rate more highly and (which are still speci®cally fear related: I am afraid; I would get help to remove it), and two which they rate lower (I think it smells unpleasant; it is biological). These ®ndings are consistent with the hypothesis that fear ampli®es general disgust tendencies: the phobic becomes more disgusted by their phobic object as a consequence of their phobia and this type of disgust is more closely allied to ideas about fear, anxiety and avoidance than those more conventionally associated with the emotion of disgust such as disease, sickness and fear of contamination. In conclusion, we suggest that when stimuli normally associated with disgust become the focus of phobic anxiety, the disgust response may be ampli®ed. The evidence is not consistent with the hypothesis that disgust plays a unique or causal role in phobic anxiety.

S.J. Thorpe, P.M. Salkovskis / Behaviour Research and Therapy 36 (1998) 877±893

893

References Beck, A. T., Ward, C. H., Mendelsohn, M., Mock, J., & Erdbaugh, J. (1961). An inventory for measuring depression. Archives of General Psychiatry, 4, 561±571. Bennett-Levy, J., & Marteau, T. (1984). Fear of animals: what is prepared?. British Journal of Psychology, 75, 37±42. Davey, G. C. L. (1992a). An expectancy model of laboratory preparedness e€ects. Journal of Experimental Psychology, General, 121, 24±40. Davey, G. C. L. (1992b). Classical conditioning and the acquisition of human fears and phobias: a review and synthesis of the literature. Advances in Behaviour Research & Therapy, 14, 29±66. Davey, G. C. L. (1992c). Characteristics of individuals with fear of spiders. Anxiety Research, 4, 299±314. Davey, G. C. L. (1994). The `disgusting' spider: The role of disease and illness in the perpetuation of fear of spiders. Society & Animals, 2, 17±24. Davey, G. C., Forster, L., & Mayhew, G. (1993). Familial resemblances in disgust sensitivity and animal phobias. Behaviour Research & Therapy, 31, 41±50. Garcia, J., & Koelling, R. (1966). Relation of cue to consequence in avoidance learning. Psychonomic Science, 4, 123±124. Garcia, J., Rusiniak, K.W. & Brett, L.P. (1977) Conditioning food-illness aversions in wild animals. In Davis, H. & Hurwitz, H.M.B. (Eds.) Operant pavlovian interactions. New York: Whiley. Haidt, J., McCauley, C. R., & Rozin, P. (1994). A scale to measure disgust sensitivity. Personality & Individual Di€erences, 16, 701±713. Matchett, G., & Davey, G. C. L. (1991). A test of a disease-avoidance model of animal phobias. Behaviour Research & Therapy, 29, 91± 94. Mulkens, S. A. N., De Jong, P., & Merckelbach, H. (1996). Disgust and spider phobia. Journal of Abnormal Psychology, 105, 464±468. Preston-Mafham, R. (1991) Spiders: An Illustrated Guide. Cassell. Rozin, P., & Fallon, A. E. (1987). A perspective on disgust. Psychological Review, 94, 23±41. Rozin, P., Fallon, A. E., & Mandell, R. (1984). Family resemblances in attitudes to food. Developmental Psychology, 20, 309±314. Rozin, P., Millman, L., & Nemero€, C. (1986). Operation of the laws of sympathetic magic in disgust and other domains. Journal of Personality & Social Psychology, 50, 703±712. Seligman, M. E. P. (1971). Phobias and preparedness. Behaviour Therapy, 2, 307±321. Spielberger, C.D. (1983) State trait anxiety inventory. Palo Alto, CA: Consulting Psychologists Press. Thorpe, S.J. (1990) Unpublished pilot study. Thorpe, S. J., & Salkovskis, P. M. (1995). Phobic beliefs: do cognitive factors play a role in speci®c phobias?. Behaviour Research & Therapy, 33, 805±816. Thorpe, S. J., & Salkovskis, P. M. (1997a). Information processing in spider phobics: the Stroop colour-naming task may indicate strategic but not automatic attentional bias. Behaviour Research & Therapy, 35, 131±144. Thorpe, S. J., & Salkovskis, P. M. (1997b). The e€ect of one-session treatment for spider phobia on attentional bias and beliefs. British Journal of Clinical Psychology, 36, 225±245. Watts, F. N. (1986). Cognitive processing in phobias. Behavioural Psychotherapy, 14, 295±301. Watts, F. N., & Sharrock, R. (1984). Questionnaire dimensions of spider phobia. Behaviour Research & Therapy, 22, 575±580. Watts, F. N., Trezise, L., & Sharrock, R. (1986). Processing of phobic stimuli. British Journal of Clinical Psychology, 25, 253±259. Wolpe, J., & Lang, P. J. (1964). A fear survey schedule for use in behaviour therapy. Behaviour Research & Therapy, 2, 27.