Trait anxiety, disgust sensitivity, and the hierarchic structure of fears

Journal of Anxiety Disorders 22 (2008) 1059–1074

Trait anxiety, disgust sensitivity, and the hierarchic structure of fears Scott D. McDonald a,*, Nathan S. Hartman b, Scott R. Vrana a b

a Virginia Commonwealth University, Department of Psychology, P.O. Box 842018, Richmond, VA 23284, USA John Carroll University, Boler School of Business, Office SB 215, 20700 North Park Boulevard, University Heights, OH 44118, USA

Received 20 April 2007; received in revised form 8 November 2007; accepted 15 November 2007

Abstract This paper describes an evaluation of Taylor’s (1998) hierarchic model of fears and its relationship to trait anxiety and disgust sensitivity (DS). In Study 1 (N = 420), a confirmatory factor analysis supported a hierarchic structure of fears. Next, an analysis using structural equation modeling indicated that trait anxiety is associated with claustrophobic and social fears, whereas DS is associated with all four fear subtypes examined (claustrophobic, social, blood–injection–injury and animal). However, trait anxiety and DS did not account for all variance shared by fear subtypes. The addition of a generalized ‘‘fear factor’’ accounted for significant residual shared variance between the four fear subtypes, beyond that accounted for by trait anxiety and DS. Study 2 (N = 213) generally replicated these results. Findings suggest that the hierarchic structural model of fears would benefit from inclusion of trait anxiety and DS as higher-order contributors to fearfulness. # 2007 Elsevier Ltd. All rights reserved. Keywords: Fear; Disgust; State-Trait Anxiety Inventory; Factor structure; Phobias; Psychopathology

1. Introduction There has been a growing interest in empirically derived taxonomic models of mental disorders in recent years (Brown, Chorpita, & Barlow, 1998; Hettema, Prescott, Myers, Neale, & Kendler, 2005; Kendler, Myers, Prescott, Neale, & Eaves, 2001; Kendler, Prescott, Myers, & Neale, 2003; Zinbarg & Barlow, 1996). In such models, symptoms or disorders that covary are thought to possess a common vulnerability and

* Corresponding author at: Mid-Atlantic Mental Illness Research, Education and Clinical Center (MIRECC), Durham Veterans Affairs Medical Center, 508 Fulton Street, Durham, NC 27705, USA. Tel.: +1 919 286 0411x6436; fax: +1 919 416 5912. E-mail address: [email protected] (S.D. McDonald). 0887-6185/$ – see front matter # 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.janxdis.2007.11.005

are placed within the same diagnostic class. Conversely, symptoms or disorders that are quantitatively unrelated are not placed within the same diagnostic class, and provide a demonstration of the heterogeneity of psychopathology. Furthermore, shared variance between diagnostic classes suggests higher-order classes, indicating a hierarchic structure of mental disorders. For example, major depressive disorder, dysthymic disorder, and generalized anxiety disorder have been repeatedly empirically linked, intimating a diagnostic class labeled variably as ‘‘Anxious-Misery’’ (Kendler et al., 2003; Krueger, 1999; Vollebergh et al., 2001) and ‘‘Distress Disorders’’ (Watson, 2005). Such empirically derived taxonomic models can be contrasted to the classification system in the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV-TR; American Psychiatric Association, 2000),

1060

S.D. McDonald et al. / Journal of Anxiety Disorders 22 (2008) 1059–1074

which has been criticized by several researchers for its basis in phenomenological similarity rather than empirical evidence (e.g., Brown & Chorpita, 1996; Watson, 2005). Taylor (1998) proposed a hierarchic structural model of fears based on a review of factor analyses of fear surveys and behavioral-genetic studies in the fear and anxiety literature (e.g., Kendler, Neale, Kessler, & Heath, 1992; Staley & O’Donnell, 1984). According to this model, there is a ‘‘hierarchy of causal factors’’ consisting of general and specific mechanisms that influence one’s proneness to developing fears. At the lowest level of the hierarchy are factors that are uniquely associated with specific fears (e.g., blood, snakes, and elevators). Intermediate factors are vulnerabilities associated with a set of specific fears. For example, fears of insects, snakes, and bats are thought to co-vary due to a common vulnerability to developing animal fears. Taylor identified four intermediate-level fear subtypes in his model, based on the findings of factor analytic studies of fear questionnaires: social, animal, blood–injury–illness (BII) and situational (or agoraphobic) fears (Arrindell, Pickersgill, Merckelbach, & Ardon, 1991). In addition to the four lower-order fear subtypes, Taylor’s (1998) hierarchic model of fears suggests that higher-order factors (e.g., neuroticism) represent a general proneness to developing fears. Taylor noted that several behavior genetic studies have demonstrated a two-tier hierarchy of genetic factors: specific genetic contributions to each fear or phobia type and a general genetic factor associated with all fears or phobias (e.g., Kendler et al., 1992). In addition, Taylor reviewed two factor analytic studies that supported a hierarchical structure of fears using fear surveys (Staley & O’Donnell, 1984; Zinbarg & Barlow, 1996). Subsequent evidence for a hierarchical model of fears was provided by a recent study (Cox, McWilliams, Clara, & Stein, 2003) that used exploratory and confirmatory factor analyses of 19 common fears using data from the National Comorbidity Survey (Kessler et al., 1994). Although the hierarchic model of fears is helpful in providing a descriptive framework for classification, a significant drawback is the ambiguity surrounding the meaning and identification of higher-order factors that are likely to have a bearing on fear proneness. Results of behavioral genetic studies and factor analytic studies such as Cox et al. (2003) infer that a unitary, higherorder mechanism is responsible for the shared variance between fear subtypes. However, there are indications that fear proneness is affected by an amalgamation of

several additive factors. For example, a recent factor analytic study of a fear questionnaire (Cutshall & Watson, 2004), significant and positive partial correlations were found between fear subtypes, even after controlling for neuroticism. To explain this observation, the authors contended that unmeasured factors might account for the common variance between fears that are unrelated to neuroticism. The implication of this finding is that the ‘‘general fear factor’’ identified by Cox et al. may in fact represent multiple, differentiable vulnerabilities to the development of fear. Certainly, clarifying such sources of covariance between fears would go far in developing our understanding of the structure of fears. The purpose of the current study was to test a hierarchic structural model of fears and its relationship to two likely contributors to fear proneness, trait anxiety and disgust sensitivity. Trait anxiety refers to individual differences in anxiety proneness in response to stressful situations (Spielberger, 1983). Trait anxiety, as well as its more broadly encompassing cousin neuroticism, has an established relationship with a variety of fears. For example, trait anxiety and neuroticism typically exhibit an association with agoraphobic (Kendler, Myers, & Prescott, 2002; Muris, Merckelbach, & Rassin, 2000) and social fears and phobias (Cutshaw & Watson, 2004; Shean & Lease, 1991; Stemberger, Turner, Beidel, & Calhoun, 1995). A similar relationship has been found for BII and animal fears (Olatunji, 2006; Page, 1994), although associations with these fear subtypes have generally been smaller than for social fears (Cutshaw & Watson, 2004; Kendler et al., 2002) and occasionally non-significant (Mulkens, de Jong, & Merckelbach, 1996; Thorpe & Salkovskis, 1995). The observation that disgust sensitivity is associated with fears and phobias is more recent. Disgust sensitivity (DS) is defined simply as individual differences in sensitivity to the emotion of disgust (Haidt, McCauley, & Rozin, 1994; Rozin, Fallon, & Mandell, 1984). Disgust sensitivity is conceptualized as an enduring trait, has temporal stability (Rozin, Haidt, McCauley, Dunlop, & Ashmore, 1999), and is resilient to secondary treatment effects (de Jong, Andrea, & Muris, 1997). There is also some evidence that DS is moderately heritable, in that parent and child reports of DS are moderately and positively correlated (Davey, Forster, & Mayhew, 1993; Rozin et al., 1984; but see de Jong et al., 1997). For example, in one interesting study, Davey et al. found that parental food-related DS, but not parental fear of spiders, predicted the fear of spiders in offspring.

S.D. McDonald et al. / Journal of Anxiety Disorders 22 (2008) 1059–1074

Disgust is thought to be protective, by promoting the avoidance of things that may indicate contamination and disease, such as feces or maggots (Matchett & Davey, 1991). In that sense, both the emotions of disgust and fear have functional value by motivating avoidance of potentially dangerous stimuli or situations (Woody & Teachman, 2000). Furthermore, an early study demonstrated a causal link between the experience of disgust and a subsequent development of fear (Webb & Davey, 1992). Consequently, the relationship between phobic behavior and DS has received a good deal of attention in recent years (e.g., Phillips, Senior, Fahy, & David, 1998; Woody & Teachman, 2000). For example, DS has been found to have a positive relationship to animal fears (Arrindell, Mulkens, Kok, & Vollenbroek, 1999; Klieger & Siejak, 1997; Sawchuk, Lohr, Tolin, Lee, & Kleinknecht, 2000) and BII fears (de Jong & Merckelbach, 1998; Muris, Merckelbach, Schmidt, & Tierney, 1999; Sawchuk et al., 2000). Furthermore, DS tends to be higher in individuals reporting animal phobias (Merckelbach, de Jong, Arntz, & Schouten, 1993) and BII phobias (Schienle, Scha¨fer, Walter, Stark, & Vaitl, 2005; Tolin, Lohr, Sawchuk, & Lee, 1997). Interestingly, associations have also been found between disgust sensitivity and other anxiety-relevant constructs, such as obsessive-compulsive disorder (Mancini, Gragnani, & D’Olimpio, 2001; Muris, Merckelbach, Nederkoorn, Rassin, & Horselenberg, 2000; Tsao & McKay, 2004), agoraphobia (Muris, Merckelbach, Nederkoorn, et al., 2000; Muris, Merckelbach, & Rassin, 2000), claustrophobia (Davey & Bond, 2006), height fears (Davey & Bond, 2006) and separation anxiety (Muris et al., 1999). In a review of the literature concerning disgust and fears, Woody and Teachman (2000) voiced concern that the relationship between DS and fears may be a spurious artifact due to the effects of trait anxiety on both DS and fearfulness. Certainly, several studies have reported a moderate and positive correlation between DS and trait anxiety or similar constructs such as neuroticism (Druschel & Sherman, 1999; Haidt et al., 1994, Olatunji, Sawchuk, Arrindell, & Lohr, 2005; Schienle, Scha¨fer, & Stark, 2005). Furthermore, there is evidence that the relationship between DS and fears is reduced after accounting for trait anxiety (Muris et al., 1999; cf. Mulkens et al., 1996). On the other hand, DS has been demonstrated to predict specific fears (Muris et al., 1999) and spider phobias (Olatunji, 2006) even after controlling for trait anxiety. Taken together, these findings support the contention that DS partially moderates the relationship between trait anxiety and fears (Baron & Kenny, 1986). In terms of Taylor’s

1061

(1998) hierarchic model of fears, DS as a mediator of trait anxiety would be placed on an intermediate level between first-order fear subtypes and higher-level constructs such as trait anxiety. Although one study has examined the correlations and partial correlations between DS, trait anxiety, and fears (Muris et al., 1999), no study to date has examined the relationship among these constructs in a comprehensive structural model. In the current study, we examined the relationship of DS and trait anxiety to four common fear subtypes as measured by select items from the self-report Fear Survey Schedule-III (FSS-III; Wolpe & Lang, 1964). Fifteen items from the FSS-III with one addition (‘‘long tunnels’’) were chosen to represent the four fear constructs in Taylor’s (1998) model: social fears, BII fears, animal fears, and claustrophobic fears (e.g., Arrindell et al., 2003). These 16 items were chosen to maximize factor homogeneity, based on a literature review and experience with the FSS-III in our lab. Claustrophobia is considered a component of the more broadly inclusive and relatively less-stable agoraphobia or ‘‘situational’’ factor (Arrindell et al., 1991; Taylor, 1998) and was preferred in this study in order to reduce interpretive ambiguity. The first aim of this study was to replicate prior studies (e.g., Cox et al., 2003) that have identified a hierarchic structure of fears. It was expected that a confirmatory factor analysis (CFA) of items from the FSS-III would support a hierarchic model comprised of 16 feared situations, four first-order fear factors (social, BII, animal, and claustrophobia) and one higher-order factor that is common to all fears, i.e., a ‘‘general fear factor.’’ The conceptual path diagram is illustrated in Fig. 1. The second aim was to examine whether this ‘‘general fear factor’’ can be fully accounted for by two fear-relevant traits, DS and trait anxiety. To test this possibility, two structural equation models (S.E.M.) were compared. In the first model, DS, trait anxiety, and a general fear factor were modeled as indicators of four fear subtypes. The four fear subtypes were latent variables with items from the FSS-III as indicators. DS and trait anxiety were latent variables that used parcels constructed from the Disgust Scale, Version 2 (DS2; Haidt, McCauley, & Rozin, 2002) and the Spielberger State-Trait Anxiety Inventory, Trait Form Y (STAI-T; Spielberger, 1983) as indicators, respectively. The ‘‘general fear factor’’ was a second-order factor that represented variance shared between the four fear subtypes that was not accounted for by DS or trait anxiety (Fig. 2). This model was compared to a similar model in which the paths from the general fear factor to

1062

S.D. McDonald et al. / Journal of Anxiety Disorders 22 (2008) 1059–1074

Fig. 1. Proposed hierarchic model of common fears.

the fear subtypes were set to zero (i.e., the general fear factor does not account for any variance shared between subtypes). If DS and trait anxiety fully account for the variance shared by the fear subtypes as expected, model fit will be equivalent for the two models. On the other hand, if the first model provides a significantly better fit to the data than the second model, then we can conclude that DS and trait anxiety do not account for all the variance shared by the four fear subtypes. A third aim was to test whether DS mediates the relationship between trait anxiety and fears. Following the evidence reviewed above, it was expected that DS would act as a partial mediator of the relationship between trait anxiety and the four fear subtypes. In terms of the research design, it was expected that a structural model with DS acting as a partial mediator between trait anxiety and fears would result in a superior fit in comparison to a fully mediated model and a non-mediated model (Kelloway, 1998).

2. Study 1 2.1. Methods 2.1.1. Participants Participants were 420 undergraduate students (62% female, 37% male, 1% unknown) attending a large, urban, state university in Virginia, USA and receiving class credit for participation. The majority of participants were first year students (67%; second year = 18%; third year = 11%; beyond third year = 3%; unknown < 1%). About half the students self-identified as White–Caucasian (53%), about a quarter selfidentified as Black–African-American (28%), and North Asian (7%), Hispanic (3%), Pacific Islanders (3%), ‘‘Other’’ (5%), and unidentified individuals (<1%) comprised the remaining sample. About 90% of participants were between 18 and 21 years of age with a mean age of 19.7 years.

S.D. McDonald et al. / Journal of Anxiety Disorders 22 (2008) 1059–1074

1063

Fig. 2. Structural model illustrating the relationship between trait anxiety, disgust sensitivity, a higher-order fear factor, and fears. ‘‘Trait Anxiety A’’ and ‘‘Trait Anxiety B’’ refer to parcels created from the STAI-T. ‘‘DS-A’’ and ‘‘DS-B’’ refer to the corresponding subscales of the Disgust Scale-2.

2.1.2. Materials 2.1.2.1. Trait anxiety. Spielberger State-Trait Anxiety Inventory, Trait Form Y (STAI-T; Spielberger, 1983). The STAI-T is a 20-item measure indicating one’s general vulnerability to experiencing anxiety symptoms. Items are rated on how respondents generally feel on a 4-point Likert-type scale (1 = ‘‘almost never’’ to 4 = ‘‘almost always’’). Internal consistency as measured by the alpha coefficient (Cronbach, 1951) was 0.89, similar to normative samples (Spielberger, 1983). 2.1.2.2. Disgust sensitivity. Disgust Scale, Version 2 (DS2; Haidt et al., 2002). The 32-item DS2 was developed as an improvement to the Disgust Scale (DS; Haidt et al., 1994), a popular measure of disgust sensitivity. The DS2 taps several disgust-relevant domains including core disgust (e.g., food, animals, and body products), death/envelope violations, interpersonal, and sex and predicts disgust-related behaviors as well as the original DS scale (Haidt, 2004). Two 16item instruction and response sets are used with 4-point Likert-type rating scales. Scale A asks respondents to rate how much they agree with various items (e.g., If I see someone vomit, it makes me sick to my stomach.’’),

whereas Scale B asks how disgusting the respondents find various experiences (e.g., ‘‘While walking under a railroad track, you smell urine.’’). In this sample, Cronbach alpha was acceptable for the DS2 (0.87). 2.1.2.3. Fear items. Fifteen items from the 52-item Fear Survey Schedule-III (FSS-III; Wolpe & Lang, 1964) and one additional item (‘‘long tunnels’’) were used as indicators of latent variables representing fear factors. The FSS-III asks respondents to rate how much they are disturbed by a variety of fear-relevant stimuli on a 5-point scale (1 = ‘‘not at all disturbed’’ to 5 = ‘‘very much disturbed’’). 2.1.3. Procedures Participants were administered informed consent and asked to complete questionnaire packets in classrooms with approximately 10–15 other students and a research assistant. The FSS-III was counterbalanced with the DS2 and STAI-T. No order effects were found using an independent-groups t-test after Bonferroni correction for multiple comparisons. After filling out questionnaires, participants were offered a debriefing form and were thanked for their participation.

1064

S.D. McDonald et al. / Journal of Anxiety Disorders 22 (2008) 1059–1074

2.2. Results 2.2.1. The structure of fears The first aim of this study was to replicate prior studies by demonstrating a hierarchic structure of fears. A series of confirmatory factor analyses (CFA) was conducted to examine the structure of common fears using LISREL 8.54 (Jo¨reskog & So¨rbom, 2003a). Parameters were estimated using robust maximum likelihood (ML) estimation and all analyses were based on the covariance matrix. The asymptotic covariance matrix calculated with PRELIS 8.54 (Jo¨reskog & So¨rbom, 2003b) and was employed to correct for non-normality of standard errors. Model fit was assessed by several common indices: Satorra-Bentler x2 (S-Bx2) index, (Robust) Non-Normed Fit Index (NNFI; Bentler & Bonnet, 1980), (Robust) Comparative Fit Index (CFI; Bentler, 1990), and the Root Mean Square Error of Approximation (RMSEA; Browne & Cudeck, 1993). General recommendations for rejecting misspecified models were followed (Browne & Cudeck, 1993; Hu & Bentler, 1999): SRMR < 0.08, CFI > 0.95, and NNFI > 0.96. Hu and Bentler (1999) suggested that RMSEA < 0.06 be used, although Browne and Cudeck (1993) recommended RMSEA < 0.08. The Satorra-Bentler x2 difference (SBx2D) test was employed using available software (Crawford, 2007) to provide a significance test of the relative goodness of fit between nested models. The Akaike Information Criterion (AIC; Akaike, 1987) was used to compare competing non-nested models. First, a four-factor model of fears was hypothesized, each with four indicators: (1) social fears (feeling disapproved of, feeling rejected by others, being criticized, failure); BII fears (witnessing surgical operations, seeing other people injected, open wounds, human blood); animal fears (crawling insects, harmless snakes, bats, flying insects); and claustrophobic fears (being in an elevator, airplanes, enclosed places, tunnels). The four factors were allowed to co-vary with one another (intercorrelations between study variables is available from S.M.). Support was found for the hypothesized model, with analyses suggesting an adequate fit, S-Bx2 (98, N = 420) = 239.52, p < 0.01, RMSEA = 0.059, AIC = 315.52, NNFI = 0.94, CFI = 0.95. All proposed paths coefficients were significant at p < 0.05. However, modification indices suggested that a better fit would be attained with the addition of an estimated path from BII fear factor to airplanes, S-Bx2 (97, N = 420) = 232.29, p < 0.01, RMSEA = 0.058, AIC = 310.29, NNFI = 0.94, CFI = 0.95. A x2 difference test indicated that the addition of this path

significantly improved the model, S-Bx2D(1) = 5.97, p < 0.05. Following the hierarchic model of fears (Taylor, 1998), a model was tested in which a second-order factor was included (Schumacker & Lomax, 2004). In addition to the paths identified above, paths were estimated from the general fear factor to each of the four fear subtype factors. In contrast to the model tested above, fear subtype factors were not allowed to co-vary. Because a x2 difference test is not appropriate for such non-nested models, comparisons were made between the four lower-factor model and the hierarchic model using the model AIC fit index. The hierarchic model was supported and an adequate fit was found, S-Bx2 (99, N = 420) = 232.96, p < 0.01, RMSEA = 0.057, AIC = 306.96, NNFI = 0.94, CFI = 0.95, indicating a better fit than the four factor model (AIC = 306.96 vs. 310.29). All four path coefficients from the second-order factor to the four fear subtype factors were significant. Squared multiple correlations indicated that the amount of variance accounted for in each fear subtype by the higher-order factor was as follows: claustrophobia = 32%, social = 23%, BII = 32%, and animals = 63%. Standardized parameter estimates and multiple squared correlations for the hierarchical model are presented in Fig. 3. Means, standard deviations, and Pearson intercorrelations for indicators are presented in Table 1. Although existence of a second-order fear factor was demonstrated, it is less clear what it represents. In order to better identify the source of covariance between fear subtypes, two likely contributors to fearfulness were introduced into the model, trait anxiety and disgust sensitivity. 2.2.2. The relationships between disgust sensitivity, trait anxiety, and fears The second aim of this study was to examine whether DS and trait anxiety fully account for variance shared between fear subtypes, represented by the second-order ‘‘general fear factor’’ That was supported in the analysis above. First, parcels were created from items on the Disgust Scale-2 and the STAI-T to use as indicators of DS and trait anxiety, respectively (Section 2.2.2.1). Second, a confirmatory factor analysis (CFA) was conducted to establish the fit of the measurement model (Section 2.2.2.2). Finally, two structural models were compared to examine whether DS and trait anxiety fully account for variance shared by fear subtypes (Section 2.2.2.3). 2.2.2.1. Item parcels for the STAI-T and the Disgust Scale-2. Parceling has several advantages over use of single indicators (i.e., using one scale score) or the use

S.D. McDonald et al. / Journal of Anxiety Disorders 22 (2008) 1059–1074

1065

Fig. 3. Results of the hierarchic CFA. Notes. Parameter estimates are standardized. All parameter estimates are significant at p < 0.001 except for the estimate for the path from BII fears to airplanes ( p < 0.05). (Values in brackets) denote multiple correlations.

of each item in a scale as a separate indicator in structural equation modeling, in that biasing effects of measurement error are reduced and often will provide more valid estimates of relationships between latent variables (Coffman & MacCallum, 2005). For trait anxiety, items from the STAI-T were quasi-randomly distributed into two parcels of 10 items each (M = 1.44, S.D. = 0.18; M = 1.43, S.D. = 0.19, respectively). The two parcels contained five and four reversed scored items respectively, and no more than two adjoining items were allowed in each parcel. Cronbach alphas were acceptable for each parcel (0.78 and 0.81). The parcels correlated strongly with each other (r = 0.82) and with the full 20-item scale (r = 0.95 and 0.96). Each parcel had a mild positive skew which was corrected by employing a square root transformation (Tabachnick & Fidell, 2001). Modified subscales A and B of the Disgust Scale-2 were used as indicators of DS. A limitation of the DS-2 is that responses on several items may be potentiated by animal fears (e.g., ‘‘Seeing a cockroach in someone else’s house does [not] bother me.’’) and BII fears (e.g., ‘‘You see a man’s intestines exposed after an accident.’’), and as a results, correlations between the

DS2 and fear scales may be artificially inflated. Therefore, for the current study, a 21-item modified DS2 was constructed that excluded items mentioning animals or mutilation-related stimuli (items 2, 4, 5, 6, 8, 10, 12, 13, 14, 15, and 16 from Scale A; items 2, 4, 5, 6, 8, 10, 12, 13, 14 and 15 from Scale B). The correlation between the original DS2 and modified DS2 was high (r = 0.96) and Cronbach alpha for modified DS2 was acceptable (0.83). Furthermore, when mutilationrelated items were removed from the DS2, the correlation between the DS2 and FSS BII fear ratings were mildly diminished (from 0.40 to 0.33), whereas the correlation were virtually unchanged between the DS2 and FSS animal fear (from 0.47 to 0.46), FSS claustrophobia (unchanged at 0.28), and social fear ratings (from 0.21 to 0.20). A similar finding was observed when animal items were removed from the DS2 (results available from S.M.). Therefore, removal of items from the DS2 that overlap with FSS items appears to reduce the inflation of correlations with disgust-relevant fear subtypes while preserving correlations with non-disgust-relevant fear subtypes and the scale’s internal consistency. Scale A (M = 2.33, S.D. = 0.55) and Scale B (M = 2.40, S.D. = 0.58) of

1066

1. Elevators (item 35) 2. Enclosed places (item 40) 3. Airplanes (Item 42) 4. Long tunnels 5. Failure (item 9) 6. Being criticized (item 33) 7. Feeling rejected by others (item 41) 8. Feeling disapproved of (item 44) 9. Open wounds (item 1) 10. Seeing other people injected (item 22) 11. Witnessing surgical operations (item 36) 12. Human blood (item 38) 13. Bats (Item 16) 14. Flying insects (item 21) 15. Crawling insects (item 29) 16. Harmless snakes (item 45) *

p < 0.05.

**

p < 0.01.

Mean (S.D.)

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

1.45 1.95 1.87 1.79 3.39 2.62 2.91 2.96 2.26 1.68 2.49 1.86 2.04 2.02 2.86 2.27

– 0.44** 0.33** 0.33** 0.04 0.10* 0.15** 0.11* 0.11* 0.21** 0.08 0.13** 0.12* 0.15** 0.20** 0.13**

– – 0.24** 0.49** 0.09 0.16** 0.15** 0.13** 0.12* 0.24** 0.08 0.17** 0.23** 0.18** 0.26** 0.18**

– – – 0.37** 0.10* 0.16** 0.18** 0.17** 0.16** 0.06 0.22** 0.26** 0.18** 0.17** 0.21** 0.18**

– – – – 0.016 0.23** 0.18** 0.21** 0.08 0.15** 0.10* 0.20** 0.25** 0.19** 0.22** 0.17**

– – – – – 0.37** 0.31** 0.47** 0.12* 0.11* 0.11* 0.15** 0.13** 0.12* 0.14** 0.11*

– – – – – – 0.59** 0.57** 0.17** 0.18** 0.25** 0.20** 0.17** 0.17** 0.24** 0.16**

– – – – – – – 0.69** 0.18** 0.12* 0.21** 0.13** 0.17** 0.21** 0.20** 0.12*

– – – – – – – – 0.16** 0.15** 0.20** 0.13** 0.14** 0.21** 0.26** 0.13**

– – – – – – – – – 0.35** 0.49** 0.46** 0.20** 0.26** 0.27** 0.15**

– – – – – – – – – – 0.45** 0.54** 0.11* 0.25** 0.23** 0.09

– – – – – – – – – – – 0.64** 0.17** 0.16** 0.29** 0.22**

– – – – – – – – – – – – 0.25** 0.19** 0.27** 0.29**

– – – – – – – – – – – – – 0.34** 0.39** 0.52**

– – – – – – – – – – – – – – 0.57** 0.23**

– – – – – – – – – – – – – – – 0.44**

(0.82) (1.10) (1.27) (1.05) (1.17) (1.06) (1.11) (1.09) (0.96) (1.00) (1.34) (1.08) (1.11) (1.00) (1.35) (1.42)

S.D. McDonald et al. / Journal of Anxiety Disorders 22 (2008) 1059–1074

Table 1 Means, standard deviations, and intercorrelations between fear indicators

S.D. McDonald et al. / Journal of Anxiety Disorders 22 (2008) 1059–1074

the 21-item DS2 were used as indicators in analyses. The scales correlated moderately with each other (r = 0.61) and strongly with the full 21-item scale (r = 0.90 and 0.89, respectively). 2.2.2.2. The measurement model for disgust sensitivity, trait anxiety, and fears. The measurement model specifies how indicators are related to latent variables, whereas the structural model specifies how latent variables are related. It is useful to test the measurement model prior to fitting the complete model to examine your hypotheses concerning relationships between indicators and latent variables (Kelloway, 1998). The measurement model can then be refined, if needed, before testing the structural model. For the current study, the measurement model was specified by allowing the paths between the latent variables representing DS (two indicators), trait anxiety (two indicators), and the fear subtypes (four indicators each) to correlate freely. Scale of measurement was specified by fixing one of the paths from an indicator to equal 1.0 for each latent variable. The measurement model demonstrated an acceptable fit to the data S-Bx2 (155, N = 420) = 327.23, p < 0.01, RMSEA = 0.051, AIC = 437.23, NNFI = 0.95, CFI = 0.96, and all paths between indicators and their associated latent variables were significant. In addition, as was the case for the second-order CFA of the fear items, modification indices suggested fear of airplanes is an indicator of both claustrophobic and BII fears. Correlations were positive and significant between the four fear subtypes and disgust sensitivity (Table 2). Trait anxiety was also positively and significantly correlated with claustrophobic and social fears, but unexpectedly, was not significantly correlated with animal or BII fears. Another unanticipated result was that the correlation between DS and trait anxiety was nonsignificant (r = 0.11, ns). 2.2.2.3. The structural model for disgust sensitivity, trait anxiety, and fears. The structural model was modified as suggested by the measurement model. That is, the paths between DS and trait anxiety, the path

1067

between trait anxiety and BII, and the path between trait anxiety and animal fears were not estimated. In addition, fear of airplanes was modeled as an indicator of both claustrophobic and BII fears. Support was found for the hypothesized model, with analyses suggesting an adequate fit, S-Bx2 (159, N = 420) = 324.48, p < 0.01, RMSEA = 0.050, AIC = 426.48, NNFI = 0.95, CFI = 0.96. Standardized parameter estimates and multiple squared correlations for the structural model are illustrated in Fig. 4. This model proved superior to a competing model in which paths from the general ‘‘fear factor’’ to the fear subtypes were constrained to zero, S-Bx2 (163, N = 420) = 401.12, p < 0.01, RMSEA = 0.059, AIC = 495.12, NNFI = 0.93, CFI = 0.94, S-Bx2D (4) = 74.36, p < 0.0001. In other words, the addition of a second-order ‘‘fear factor’’ improved the model’s ability to accurately represent the observed relationships between common fears, DS, and trait anxiety. 2.2.3. Disgust sensitivity as mediator of the relationship between trait anxiety and fears The third aim of this study was to test whether DS is a mediator of the relationship between trait anxiety and fears. Results of the measurement model demonstrated that DS and trait anxiety were not correlated. A significant correlation between an independent variable and its mediator is an essential condition for identification of a mediated relationship (Baron & Kenny, 1986). Subsequently, it became unnecessary to compare mediated, unmediated, and partially mediated models because DS was clearly not a mediator of trait anxiety and fears. 2.3. Summary of findings for Study 1 Findings from Study 1 can be summarized according to the three aims of this project. First, results of Study 1 replicated prior studies (e.g., Cox et al., 2003) that have indicated a hierarchic structure of fears. Confirmatory factor analysis supported a model consisting of one second-order or ‘‘general fear

Table 2 Correlations between disgust sensitivity, trait anxiety, and fears in the measurement model in Study 1 Social fears Claustrophobia Social fears BII fears Animal fears Disgust sensitivity **

p < 0.01.

0.31

***

p < 0.001. yp < 0.10.

***

BII fears ***

0.32 0.29***

Animal fears ***

0.48 0.36*** 0.46***

Disgust sensitivity ***

0.31 0.23*** 0.39*** 0.52***

Trait anxiety 0.17** 0.37*** 0.10y 0.04 0.11y

1068

S.D. McDonald et al. / Journal of Anxiety Disorders 22 (2008) 1059–1074

Fig. 4. Results of the structural model in Study 1. Notes. Parameter estimates are standardized. All parameter estimates are significant at p < 0.001 except where noted by a superscript. *p < 0.05. **p < 0.01. y = ns. nt Parameter estimates constrained to ‘‘1’’ for model identification purposes and therefore were not tested. (Values in brackets) denote multiple correlations.

factor,’’ four first-order fear factors (claustrophobic, social, BII, and animal), and a lower tier of specific common fears. Secondly, after accounting for variance associated with trait anxiety and DS, significant and positive correlations were still observed between the four fear subtypes. This study extends findings of Cutshall and Watson (2004), by establishing that fears continue to share variance beyond what is accounted for by two fear-relevant constructs, trait anxiety and DS. These findings suggest that trait anxiety and DS are two, but not the only, higher-level factors in Taylor’s (1998) hierarchic model of fears. Thirdly, results indicated that trait anxiety and DS are not correlated. Thus, it can be concluded that DS does not act as a mediator for the relationship between trait anxiety and fears. 3. Study 2 It is generally desirable to replicate a theoretical model, particularly when post hoc modifications have been made (Kelloway, 1998). Two archival data sets that included comparable measures were combined to replicate the model developed in Study 1 (see Fig. 4). We predicted that Study 2 would replicate

the structural model for disgust sensitivity, trait anxiety, and fears that was supported in Study 1. 3.1. Participants Two archival data sets (N = 96, N = 117) collected in two consecutive semesters were combined to create this sample of undergraduate students (N = 213; 77% female, 22% male, 1% unknown). The students were from the same university and curriculum as those participating in Study 1 and received class credit for participation. About half the students self-identified as White–Caucasian (48%), about a quarter self-identified as Black–African-American (25%), and North Asian (12%), Hispanic (4%), Pacific Islanders (2%), ‘‘Other’’ (6%), and unidentified individuals (3%) comprised the remaining sample. Data was not collected on year in school or age, but generally such research participants are 18–22-years-old and first or second year students. 3.2. Materials The State-Trait Anxiety Inventory, Trait Form Y (STAI-T; Spielberger, 1983) and 52-item Fear Survey Schedule-III (FSS-III; Wolpe & Lang, 1964) as

S.D. McDonald et al. / Journal of Anxiety Disorders 22 (2008) 1059–1074

1069

described above were administered. An 8-item version of the Disgust Scale, Version 2 (DS2-S; Haidt et al., 2002) was employed to represent DS. The DS2-S was developed by Haidt et al. as a brief measure of disgust sensitivity, and is comprised of two representative items from each of the four domains that are used in the full DS-2 scale. The authors report that the DS2-S correlates at 0.90 with the 32-item version. In our sample, Cronbach alpha for the DS2-S was an acceptable 0.70. The eight items were distributed into two parcels to use as indicators of disgust sensitivity.

model, the path from trait anxiety to claustrophobia was not significant (0.16, p < 0.10). Modification indices did not recommend additional paths between latent variables or between latent variables and indicators. Standardized parameter estimates and multiple squared correlations for the structural model are presented in Fig. 5 (means, standard deviations, correlations between study variables, and path coefficients in the measurement model for Study 2 are available by contacting S.M.).

3.3. Procedure

The structural model developed in Study 1 was generally supported by replication in Study 2. However, whereas the path between trait anxiety and claustrophobia was significant in Study 1 (0.14, p < 0.05), it was not in Study 2 (0.16, p < 0.10). Considering that their standardized parameter estimates were very similar, this difference between studies is likely due to the fact that the estimates are small and were both near the threshold for significance at p < 0.05.

The procedure was the same as used in Study 1, with the exception that the presentation of measures was not counterbalanced in Study 2. 3.4. Results The measurement model demonstrated good fit to the data, S-Bx2 (155, N = 213) = 241.04, p < 0.01, RMSEA = 0.051, AIC = 351.04, NNFI = 0.95, CFI = 0.96, and all estimated paths from latent variables to indicators were significant. Consistent with Study 1, BII fears, animal fears and disgust sensitivity were not correlated with trait anxiety (Table 3). However, modification indices suggested the estimation of paths from claustrophobia to ‘‘human blood’’ and from animals to ‘‘witnessing surgical operations.’’ Results of the test of the measurement model informed the design of the structural model. The structural model for Study 2 tested the model supported in Study 1 with the following exceptions: the paths from trait anxiety to both BII and animal fears were not estimated, the path from claustrophobia to ‘‘human blood’’ was estimated, and the path from animals to ‘‘witnessing surgical operations’’ was estimated. Results indicated good fit, S-Bx2 (158, N = 213) = 229.10, p < 0.01, RMSEA = 0.046, AIC = 333.10, NNFI = 0.96, CFI = 0.97. All estimated paths from latent variables to indicators were significant. However, in contrast to results from the measurement

3.5. Summary of findings for Study 2

4. General discussion The current study had three aims: (1) replicate the hierarchic structure of fears found in previous studies using confirmatory factor analysis (CFA); (2) test whether DS and trait anxiety would fully account for shared variance between fear subtypes; and (3) examine whether disgust sensitivity (DS) mediates the relationship between trait anxiety and fears. Concerning the first aim, findings replicate prior studies (e.g., Cox et al., 2003) and support a hierarchic structural model of fears. A CFA indicated that 16 common fears loaded on four identifiable fear subtypes (i.e., social, BII, animal, and claustrophobia) that, in turn, loaded on a second-order, general ‘‘fear factor.’’ At first blush, these results may appear to indicate a single, perhaps genetic or characterological vulnerability to fears. However, the nature of this higher-order factor has not been thoroughly identified in prior studies. At best, research has indicated that covariance among fears

Table 3 Correlations between disgust sensitivity, trait anxiety, and fears in the measurement model in Study 2 Social fears Claustrophobia Social fears BII fears Animal fears Disgust sensitivity *

p < 0.05.

0.25

*

p < 0.001. yp < 0.10.

***

BII fears ***

0.40 0.14y

Animal fears ***

0.47 0.35*** 0.43***

Disgust sensitivity ***

0.41 0.19* 0.42*** 0.44***

Trait anxiety 0.20* 0.40*** 0.02 0.15y 0.01

1070

S.D. McDonald et al. / Journal of Anxiety Disorders 22 (2008) 1059–1074

Fig. 5. Results of the structural model in Study 2. Notes. Parameter estimates are standardized. All parameter estimates are significant at p < .001 except where noted by a superscript. *p < 0.05. **p < 0.01. y = ns. nt Parameter estimates constrained to ‘‘1’’ for model identification purposes and therefore were not tested. (Values in brackets) denote multiple correlations.

is only partially explained by measured higher-order factors (i.e., neuroticism), suggesting that multiple underlying mechanisms may be present (Cutshall & Watson, 2004). Thus, the second aim of this study was to examine the relative contributions of trait anxiety and DS to fearfulness. Results of the current study support prior reports that trait anxiety and DS are both positively correlated with a variety of fears. Trait anxiety had a significant (and strong) positive relationship with social fears and a modest relationship with claustrophobic fears, whereas trait anxiety was not related to animal and BII fears. These findings are consistent with prior studies that have found a strong and positive relationship between trait anxiety (and neuroticism) and social fears, compared with a small or insignificant relationship with animal and BII fears (e.g., Thorpe & Salkovskis, 1998). The small correlations between trait anxiety and claustrophobia were unexpected, considering trait anxiety is generally found to be correlated with agoraphobic fears (Muris, Merckelbach, Nederkoorn, et al., 2000; Muris, Merckelbach, & Rassin, 2000). For the current project, claustrophobia was used as a fear subtype, a gambit intended to maximize construct validity. It is possible that trait anxiety is more strongly

related to other agoraphobic or situational fear stimuli, such as open spaces or heights. Disgust sensitivity, on the other hand, demonstrated a significant and positive correlation with each of the four fear subtypes, consistent with prior studies that have demonstrated the broad effects of DS on a variety of fears. Perhaps not surprisingly, DS was most strongly correlated with animal and BII fears, which are commonly found to be associated with DS in the fear and phobia literature (Muris et al., 1999). One possible criticism of studies that used the original Disgust Scale (Haidt et al., 1994) is that of the 32 items, nine mention animals and four mention injuries, potentially inflating the correlations between DS and fears of these stimuli. To account for this possibility in Study 1, animal and injury-related items were removed from the revised Disgust Scale-2 (Haidt et al., 2002). Results indicated a strong relationship between the modified Disgust Scale2 and the original Disgust Scale-2 (r = 0.96), and a preserved albeit mildly diminished positive correlation to animal and BII fears. These results support the association between DS and fears beyond that accounted for by the overlap in scale item content. Several studies have found positive correlations between DS and trait anxiety (e.g., Olatunji et al.,

S.D. McDonald et al. / Journal of Anxiety Disorders 22 (2008) 1059–1074

2005). However, in the current study, correlations between DS and trait anxiety were positive but not significant in both the primary and replication samples. Consequently, results did not support the contention that DS mediates the relationship between trait anxiety and fear, as has been suggested (Davey & Bond, 2006; Woody & Teachman, 2000). It is unclear why a significant relationship between DS and trait anxiety was not replicated in the current study, although two possibilities stem from methodological variations. First, it is possible that the parceling procedure to identify the DS and trait anxiety constructs may have resulted in different correlations than have been found with singlescale indicators. To examine this possibility, Study 1 scale scores were computed for the 21-item Disgust Scale-2 and trait anxiety, and their correlations were examined. Again, the correlation was small and positive, but not significant (r = 0.09, p = 0.062). This supports the contention that the parceling of scales did not substantially alter the correlations between DS and trait anxiety. Another possible explanation for the small and nonsignificant correlation between DS and trait anxiety is that removal of the eleven animal and mutilation-related items from the Disgust Scale-2 reduced the correlation between DS and trait anxiety in Study 1. The correlation between the full 32-item Disgust Scale-2 and trait anxiety was small, positive, and significant (r = 0.10, p = 0.048). Nonetheless, the difference between correlations for trait anxiety and the two Disgust Scale-2 scales was not significant (z = 0.09, ns), suggesting that removal of these items had little substantive effect on the correlation between DS and trait anxiety. More importantly, a correlation of approximately 0.10 is still substantially smaller than those reported in published studies using various measures of DS and trait anxiety or neuroticism (0.21–0.45 for the studies noted in the introduction). Moreover, results of Study 2 also demonstrated that trait anxiety is not related to DS using a related but abridged measure of disgust sensitivity (the 8-item Disgust Scale-2). It may be the case that whereas both the Disgust Scale-2 and the original Disgust Scale are good indicators of disgustrelevant avoidance (Haidt et al., 2002), the Disgust Scale-2 has the advantage of being less sensitive to a general propensity to experience anxiety. Certainly, future research on the validity of the Disgust Scale-2 scale is necessary to garner support for this contention. That DS demonstrated significant correlations with each of the four fear subtypes supports prior findings, but does little to resolve the question of why DS would be related to fear stimuli that would not be expected to

1071

elicit a disgust reaction (i.e., social and claustrophobic fears). One possibility is that individual differences in both DS and fear surveys represent aspects of ‘‘harm avoidance’’ (Cloninger, 1987). This would be consistent with Haidt et al. (1994) contention that ‘‘highly disgust sensitive people appear to be guarding themselves from external threats’’ (p. 711) and Cook’s (1999) observation that fearfulness is strongly associated with human startle valence modification, commonly regarded to be ‘‘a protective or defensive response to a stimulus’’ (p. 198). Future studies are needed to clarify whether the relationship between DS and a broad range of fears is explained by harm avoidance or other proposed mitigating factors, such as self-disgust or shame (Davey & Bond, 2006). Finally, the contention that DS and trait anxiety would fully account for covariance between fear subtypes was not supported. Although DS accounted for significant covariance between all fear subtypes and trait anxiety accounted for significant covariance between social fears and claustrophobic fears, a general ‘‘fear factor’’ was identified that accounted for additional variance shared by the four fear subtypes. What is particularly striking is that the amount of variance this construct accounts for is not trivial, with total effects accounting for nearly as much if not more variance in the four fear subtypes (Study 1: 16–42%; Study 2: 8–52%) as does trait anxiety and DS. This finding appears to provide evidence for a ‘‘unique component’’ of fears that may differentiate fears from other psychiatric disorders (Mineka, Watson, & Clark, 1998; Watson, 1999). However, it is cautioned that this ‘‘fear factor’’ may, in part, also represents method variance (Campbell & Fiske, 1959), as well as individual differences in negative affectivity (Watson & Clark, 1984), a construct that is more generally related to a broad range of psychopathology. Future studies will benefit from inclusion of measures representing other disorders (e.g., Penn State Worry Questionnaire for Generalized Anxiety Disorder) and using multi-method designs to control for error variance. The primary implication of this study concerns the suitability of Taylor’s (1998) structural model of fears. In Taylor’s hierarchic model, common fears, at the lowest level, share variance associated with several intermediate-level fear subtypes (e.g., social, animal, BII, claustrophobic), which in turn share common variance attributable to one or more higher-order factors that contribute to fear proneness. In the current study, a confirmatory factor analysis of a three-level model of fears supported a hierarchic model of fears. Further-

1072

S.D. McDonald et al. / Journal of Anxiety Disorders 22 (2008) 1059–1074

more, results indicate that trait anxiety and DS are each unique and identifiable higher-order factors that influence fearfulness. The critical inference of these findings regards the interpretation of single, higherorder fear factors observed in behavioral genetic (Kendler et al., 1992, 2001) and factor analytic studies (Staley & O’Donnell, 1984). Results of our first analysis supported a general, higher-order fear factor. However, integrating DS and trait anxiety into the model revealed that this apparently unitary higher-order factor was, in fact, a blend of multiple, higher-order contributors to fear proneness. Future research concerned with the taxonomy of fears and phobias are expected to benefit from inclusion of trait anxiety and DS into their studies. The current study has several limitations that should be noted. First, the samples were comprised of undergraduate students in the United States of America. Although the structure of fears has been found to be stable across cultures, age, nationality, and clinical status (Arrindell et al., 2003), the proposed model should be replicated in other samples. Secondly, only a selection of common fears and fear subtypes were used for this study. It is possible that including additional common fears would increase the number of fear subtypes and identify lower-order fear subtypes (e.g., ‘‘claustrophobia’’ could have two further subtypes, ‘‘fear of suffocation’’ and ‘‘fear of physical restriction’’). Third, recent work (van Overveld, de Jong, Peters, Cavanagh, & Davey, 2006) suggests that differentiating ‘‘disgust propensity’’ from ‘‘disgust sensitivity’’ may provide additional information concerning the relationship between disgust and phobias. Future research may be strengthened by examining both of these disgust-relevant traits. And finally, the structural equation modeling approach utilized in this study provided evidence of correlational relationships but not causality. However, as reviewed in the introduction, there is substantive evidence that DS and trait anxiety are relatively stable traits that provide a vulnerability to the development of fears and phobias. In summary, a hierarchic structural model of fears was supported in this study consisting of three levels: lower-order common fears, four intermediate fear subtypes, and a general ‘‘fear factor.’’ Trait anxiety was correlated with claustrophobic and social fears, whereas disgust sensitivity was correlated with each of the four fear subtypes. The idea that DS mediates the relationship between trait anxiety and fears was not supported, as correlations between DS and trait anxiety were not significant. Furthermore, a general fear factor unrelated to DS and trait anxiety was observed. Together, findings support the hierarchic structural

model of fears (Taylor, 1998), and highlight the importance of considering multiple, higher-order contributors to fear proneness. And finally, results of this study provide additional evidence that DS is a general diathesis for specific (and social) phobias. As DS is not intuitively associated with non-disgust-related stimuli (i.e., claustrophobic and social fears), it may be the case that DS is acting as a proxy for the trait of harm avoidance (Cloninger, 1987). Certainly, further research is necessary to lend support to this contention. Acknowledgement The authors would like to thank Lisa Jobe for her assistance with data collection and management. References Akaike, H. (1987). Factor analysis and AIC. Psychometrika, 52, 317– 332. American Psychiatric Association. (2000). Diagnostic and statistical manual of mental disorders (4th ed. text rev.). Washington, DC: Author. Arrindell, W. A. (1980). Dimensional structure and psychopathology correlates of the Fear Survey Schedule (FSS-III) in a phobic population: a factorial definition of agoraphobia. Behaviour Research and Therapy, 18(4), 229–242. Arrindell, W. A., Eisemann, M., Richter, J., Oei, T. P. S., Caballo, V. E., van der Ende, J., et al. (2003). Phobic anxiety in 11 nations part I: dimensional constancy of the five-factor model. Behaviour Research and Therapy, 41, 461–479. Arrindell, W. A., Mulkens, S., Kok, J., & Vollenbroek, J. (1999). Disgust sensitivity and the sex difference in fears to common indigenous animals. Behaviour Research and Therapy, 37, 273– 280. Arrindell, W. A., Pickersgill, M. J., Merckelbach, H., & Ardon, A. l. M. (1991). Phobic dimensions: III. Factor analytic approaches to the study of common phobic fears: an updated review of findings obtained with adult subjects. Advances in Behaviour Research & Therapy, 13(2), 73–130. Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable distinction in social psychological research: conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51(6), 1173–1182. Bentler, P. (1990). Comparative fit indices in structural equation models. Psychological Bulletin, 107, 238–246. Bentler, P. M., & Bonnet, D. G. (1980). Significance tests and goodness-of-fit in the analysis of covariance structure. Psychological Bulletin, 88(3), 588–606. Brown, T. A., Chorpita, B. F., & Barlow, D. H. (1998). Structural relationships among dimensions of the DSM-IV anxiety and mood disorders and dimensions of negative affect, positive affect, and autonomic arousal. Journal of Abnormal Psychology, 107(2), 179– 192. Brown, T. A., & Chorpita, B. F. (1996). On the validity and comorbidity of the DSM-III-R and DSM-IV anxiety disorders. In: R. M. Rapee (Ed.), Current controversies in the anxiety disorders (pp. 48–52). New York: Guilford Press.

S.D. McDonald et al. / Journal of Anxiety Disorders 22 (2008) 1059–1074 Browne, M. W., & Cudeck, R. (1993). Alternative ways of assessing model fit. In: K. A. Bollen & G. S. Long (Eds.), Testing structural equation models (pp. 136–162). Newbury Park, CA: Sage. Campbell, D. T., & Fiske, D. W. (1959). Convergent and discriminant validation by the multitrait-multimethod matrix. Psychological Bulletin, 56, 81–105. Chambless, D. L. (1985). The relationship of severity of agoraphobia to associated psychopathology. Behaviour Research and Therapy, 23(3), 305–310. Cloninger, R. C. (1987). A systematic method for clinical description and classification of personality variants: a proposal. Archives of General Psychiatry, 44, 573–588. Coffman, D. L., & MacCallum, R. C. (2005). Using parcels to convert path analysis models into latent variable models. Multivariate Behavioral Research, 40(2), 235–259. Cook, E. W. (1999). Affective individual differences, psychopathology, and startle reflex modification. In: M. E. Dawson, A. M. Schell, & A. H. Bo¨hmelt (Eds.), Startle modification: implications for neuroscience, cognitive science, and clinical science (pp. 187– 208). Cambridge, UK: Cambridge University Press. Cox, B. J., McWilliams, L. A., Clara, I. P., & Stein, M. B. (2003). The structure of feared situations in a nationally representative sample. Journal of Anxiety Disorders, 17(1), 89–101. Crawford, J. R. (2007). sbdiff.exe [Computer software], Retrieved June 15, 2007 from http://www.abdn.ac.uk/psy086/dept/psychom.htm. Cronbach, L. J. (1951). Coefficient alpha and the internal structure of tests. Psychometrika, 16, 297–333. Cutshall, C., & Watson, D. (2004). The phobic stimuli response scales: a new self-report measure of fear. Behaviour Research and Therapy, 10, 1193–1201. Davey, G. C. L. (1994). Self-reported fears to common indigenous animals in an adult UK population: the role of disgust sensitivity. British Journal of Psychology, 85, 541–554. Davey, G. C. L., & Bond, N. (2006). Using controlled comparisons in disgust psychopathology research: the case of disgust, hypochondriasis and health anxiety. Journal of Behavior Therapy and Experimental Psychiatry, 37, 4–15. Davey, G. C., Forster, L., & Mayhew, G. (1993). Familial resemblances in disgust sensitivity and animal phobias. Behaviour Research and Therapy, 31(1), 41–50. de Jong, P. J., Andrea, H., & Muris, P. (1997). Spider phobia in children: disgust and fear before and after treatment. Behaviour Research and Therapy, 35(6), 559–562. de Jong, P. J., & Merckelbach, H. (1998). Blood–injection–injury phobia and fear of spiders: domain specific individual differences in disgust sensitivity. Personality and Individual Differences, 24(2), 153–158. De Jongh, A., Bongaarts, G., Vermeule, I., Visser, K., De Vos, P., & Makkes, P. (1998). Blood–injury–injection phobia and dental phobia. Behaviour Research and Therapy, 36(10), 971–982. Druschel, B. A., & Sherman, M. F. (1999). Disgust sensitivity as a function of the Big Five and gender. Personality and Individual Differences, 26(4), 739–748. Haidt, J. (2004). The Disgust Scale Home Page. Retrieved January 20, 2004, from http://wsrv.clas.virginia.edu/jdh6n/disgustscale.html. Haidt, J., McCauley, C., & Rozin, P. (1994). Individual differences in sensitivity to disgust: a scale sampling seven domains of disgust elicitors. Personality and Individual Differences, 16(5), 701–713. Haidt, J., McCauley, C., & Rozin, P. (2002). The Disgust Scale, Version 2. Retrieved January 20, 2004, from http://www.people.virginia.edu/jdh6n/disgustscale.html.

1073

Hettema, J. M., Prescott, C. A., Myers, J. M., Neale, M. C., & Kendler, K. S. (2005). The structure of genetic and environmental risk factors for anxiety disorders in men and women. Archives of General Psychiatry, 62, 182–189. Hu, L., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling, 6(1), 1–55. Jo¨reskog, K. G., & So¨rbom, D. (2003a). LISREL (Version 8.54). Chicago: Scientific Software International. Jo¨reskog, K. G., & So¨rbom, D. (2003b). PRELIS (Version 8.54). Chicago: Scientific Software International. Kelloway, E. K. (1998). Using LISREL for structural equation modeling: a researcher’s guide. Sage Publications Inc. Kendler, K. S., Myers, J., & Prescott, C. A. (2002). The etiology of phobias: an evaluation of the stress-diathesis model. Archives of General Psychiatry, 59(3), 242–248. Kendler, K. S., Myers, J., Prescott, C. A., Neale, M. C., & Eaves, L. J. (2001). The genetic epidemiology of irrational fears and phobias in men. Archives of General Psychiatry, 58(3), 257–265. Kendler, K. S., Neale, M. C., Kessler, R. C., & Heath, A. C. (1992). The genetic epidemiology of phobias in women: the interrelationship of agoraphobia, social phobia, situational phobia, and simple phobia. Archives of General Psychiatry, 49(4), 273–281. Kendler, K. S., Prescott, C. A., Myers, J., & Neale, M. C. (2003). The structure of genetic and environmental risk factors for common psychiatric and substance use disorders in men and women. Archives of General Psychiatry, 60, 929–937. Kessler, R. C., McGonagle, K. A., Zhao, S., Nelson, C. B., Hughes, M., Eshleman, S., et al. (1994). Lifetime and 12-month prevalence of DSM-III-R psychiatric disorders in the United States: results from the National Comorbidity Survey. Archives of General Psychiatry, 51, 8–19. Klieger, D. M., & Siejak, K. K. (1997). Disgust as the source of false positive effects in the measurement of ophidiophobia. Journal of Psychology, 131(4), 371–382. Krueger, R. F. (1999). The structure of common mental disorders. Archives of General Psychiatry, 56, 921–926. Mancini, F., Gragnani, A., & D’Olimpio, F. (2001). The connection between disgust and obsessions and compulsions in a non-clinical sample. Personality & Individual Differences, 31, 1173–1180. Matchett, G., & Davey, G. C. L. (1991). A test of a disease-avoidance model of animal phobias. Behaviour Research & Therapy, 29, 91– 94. Merckelbach, H., de Jong, P. J., Arntz, A., & Schouten, E. (1993). The role of evaluative learning and disgust sensitivity in the etiology and treatment of spider phobia. Advances in Behaviour Research and Therapy, 15, 243–256. Mineka, S., Watson, D., & Clark, L. A. (1998). Comorbidity of anxiety and unipolar mood disorders. Annual Review of Psychology, 49, 377–412. Mulkens, S. A. N., de Jong, P. J., & Merckelbach, H. (1996). Disgust and spider phobia. Journal of Abnormal Psychology, 105(3), 464– 468. Muris, P., Merckelbach, H., Nederkoorn, S., Rassin, I. C., & Horselenberg, R. (2000a). Disgust and psychopathological symptoms in a nonclinical sample. Personality & Individual Differences, 29, 1163–1167. Muris, P., Merckelbach, H., & Rassin, E. (2000b). Monitoring, trait anxiety, and panic disorder symptomatology in normal subjects. Journal of Behavior Therapy and Experimental Psychiatry, 31(1), 21–28.

1074

S.D. McDonald et al. / Journal of Anxiety Disorders 22 (2008) 1059–1074

Muris, P., Merckelbach, H., Schmidt, H., & Tierney, S. (1999). Disgust sensitivity, trait anxiety and anxiety disorders symptoms in normal children. Behaviour Research and Therapy, 37(10), 953–961. Olatunji, B. O. (2006). Evaluative learning and emotional responding to fearful and disgusting stimuli in spider phobia. Journal of Anxiety Disorders, 20(7), 858–876. Olatunji, B. O., Sawchuk, C. N., Arrindell, W. A., & Lohr, J. M. (2005). Disgust sensitivity as a mediator of the sex differences in contamination fears. Personality and Individual Differences, 38(3), 713–722. Page, A. C. (1994). Blood-injury phobia. Clinical Psychology Review, 14(5), 443–461. Phillips, M. L., Senior, C., Fahy, T., & David, A. S. (1998). Disgust: the forgotten emotion of psychiatry. British Journal of Psychiatry, 172, 373–375. Rozin, P., Fallon, A., & Mandell, R. (1984). Family resemblance in attitudes to foods. Developmental Psychology, 20(2), 309–314. Rozin, P., Haidt, J., McCauley, C., Dunlop, L., & Ashmore, M. (1999). Individual differences in disgust sensitivity: comparisons and evaluations of paper-and-pencil versus behavioral measures. Journal of Research in Personality, 33(3), 330–351. Sawchuk, C. N., Lohr, J. M., Tolin, D. F., Lee, T. C., & Kleinknecht, R. A. (2000). Disgust sensitivity and contamination fears in spider and blood–injection–injury phobias. Behaviour Research and Therapy, 38(8), 753–762. Schienle, A., Scha¨fer, A., & Stark, R. (2005a). Relationship between disgust sensitivity, trait anxiety and brain activity during disgust induction. Neuropsychobiology, 51, 86–92. Schienle, A., Scha¨fer, A., Walter, B., Stark, R., & Vaitl, D. (2005b). Elevated disgust sensitivity in blood phobia. Cognition & Emotion, 19(8), 1229–1241. Schumacker, R. E., & Lomax, R. G. (2004). A beginner’s guide to structural equation modeling. Mahwah, NJ: Lawrence Erlbaum. Shean, G., & Lease, C. (1991). The relationship between interaction patterns and agoraphobic fears among college students. Journal of Psychology: Interdisciplinary and Applied, 125(3), 271–278. Spielberger, C. D. (1983). Manual for the State-Trait Anxiety Inventory (STAI). Palo Alto, CA: Consulting Psychologists Press. Staley, A. A., & O’Donnell, J. P. (1984). A developmental analysis of mothers’ reports of normal children’s fears. Journal of Genetic Psychology, 144, 165–178. Stemberger, R. T., Turner, S. M., Beidel, D. C., & Calhoun, K. S. (1995). Social phobia: an analysis of possible developmental factors. Journal of Abnormal Psychology, 104(3), 526–531.

Tabachnick, B. G., & Fidell, L. S. (2001). Using multivariate statistics (4th ed.). Boston: Allyn and Bacon. Taylor, S. (1998). The hierarchic structure of fears. Behaviour Research and Therapy, 36(2), 205–214. Thorpe, S. J., & Salkovskis, P. M. (1995). Phobia beliefs: do cognitive factors play a role in specific phobias? Behaviour Research and Therapy, 33(7), 805–816. Thorpe, S. J., & Salkovskis, P. M. (1998). Studies on the role of disgust in the acquisition and maintenance of specific phobias. Behaviour Research and Therapy, 36, 877–893. Tolin, D. F., Lohr, J. M., Sawchuk, C. N., & Lee, T. C. (1997). Disgust and disgust sensitivity in blood-injection-injury and spider phobia. Behaviour Research and Therapy, 35(10), 949–953. Tsao, S. D., & McKay, D. (2004). Behavioral avoidance tests and disgust in contamination fears distinctions from trait anxiety. Behaviour Research and Therapy, 42, 207–216. van Overveld, W. J. M., de Jong, P. J., Peters, M. L., Cavanagh, K., & Davey, G. C. L. (2006). Disgust propensity and disgust sensitivity: separate constructs that are differentially related to specific fears. Personality and Individual Differences, 41(7), 1241–1252. Vollebergh, W. A. M., Iedema, J., Bijl, R. V., de Graff, R., Smit, F., & Ormel, J. (2001). The structure and stability of common mental disorders. Archives of General Psychiatry, 58, 597–603. Watson, D. (1999). Dimensions underlying the anxiety disorders: a hierarchic perspective. Current Opinions in Psychiatry, 12(2), 181–186. Watson, D. (2005). Rethinking the mood and anxiety disorders: a quantitative hierarchic model for DSM-V. Journal of Abnormal Psychology, 114(4), 522–536. Watson, D., & Clark, L. A. (1984). Negative affectivity: the disposition to experience aversive emotional states. Psychological Bulletin, 96(3), 465–490. Webb, K., & Davey, G. C. L. (1992). Disgust sensitivity and fear of animals: effects of exposure to violent or revulsive material. Anxiety, Stress, and Coping, 5, 329–335. Wolpe, J., & Lang, P. J. (1964). A Fear Survey Schedule for use in behaviour therapy. Behaviour Research and Therapy, 2, 27–30. Woody, S. R., & Teachman, B. A. (2000). Intersection of disgust and fear: normative and pathological views. Clinical Psychology: Science and Practice, 7, 291–311. Zinbarg, R. E., & Barlow, D. H. (1996). Structure of anxiety and the anxiety disorders: a hierarchic model. Journal of Abnormal Psychology, 105(2), 181–193.