Introduction

Specific phobias are characterized by intense and persistent fear reactions towards either specific stimuli and/or situations accompanied by strong avoidant behavior [1]. According to the common DSM classification system, five different categories or subtypes of specific phobias can be differentiated: animal, natural environment, blood/injection/injury (BII), situational and other phobias [1]. Significant differences in the prevalence [2], phenomenology [3] and pathophysiology [4] of specific fears and phobia subtypes have been reported.

An influential etiological approach to understand specific phobias and anxiety disorders is derived from Bandura’s self-efficacy theory [5]. It has been proposed that the degree of perceived or subjective self-efficacy is related to successful coping with threatening stimuli and fear-related stimuli and situations [5–8]. A distinction between domain-specific self-efficacy (SSE, [9]) and general self-efficacy (GSE) has been made. SSE designates the subjective estimation of coping capabilities in situations within a specific domain, e.g. in academic or social settings [10] but also in the presence of threatening situations and stimuli [6]. In contrast, GSE is not related to a specific task or domain but rather designates a general subjective confidence in one’s ability to overcome and master difficult and stressful situations irrespective of the domain. To this end, GSE has been related to a variety of psychological constructs, including optimism, self-regulation, self-esteem, depression and anxiety [11].

Higher GSE has been linked to low trait anxiety in a non-clinical sample [12], whereas lower GSE has been related to high trait anxiety [12] and higher social anxiety symptom severity in both patients and healthy controls [13]. Moreover, reduced GSE in adolescents seems to be associated with increased anxiety [14]. In adolescents, SSE scores are correlated with corresponding phobia types (e.g. social self-efficacy with social phobia), whereas GSE is correlated with all phobias, namely social phobia, separation anxiety and school phobia [14]. Furthermore, SSE has been shown to have connections with specific phobias such as agoraphobia [15], spider phobia [16] and height phobia [17] and predicts the anxiety levels within the phobia and the degree of avoidance behavior. Finally, perceived self-efficacy seems to be an important mechanism of cognitive behavioral therapy for phobias and anxiety disorders [9]. Increases in self-efficacy predict treatment-related decreases in symptoms and increases in functioning in patients with anxiety disorders [18]. Enhancing self-efficacy during exposure therapy can improve treatment outcomes in patients with anxiety disorders [19, 20].

However, the exact relationship between GSE and/or SSE and different specific fears and phobia subtypes has not been investigated so far. This is surprising, given that the pathophysiology of some specific fears, e.g. BII-related fears and phobias, can be distinguished from other specific fears and phobia subtypes [21]. Cognitive factors related to certain specific fears seem to be important for some fear subtypes, e.g. “fear of losing control” in BII fears and phobias [21]. However, cognitive factors underlying specific fears can also differ across different fear categories [22]. The same might apply to self-efficacy. Individual’s perceived belief of being able to exercise control over the occurrence of potentially harming events contributes to fear and avoidance. Self-efficacy, thus, is an important cognitive component of specific phobia. But does the role of GSE and SSE differ depending on specific phobia subtypes? This research question has not been investigated so far.

Lifetime prevalence rates of specific fears and phobias vary from 7.69% to 19.9% [23–26], whereby women (10% - 26.5%) show higher prevalence rates relative to men (7.3% - 12.4%, [25, 26]). Furthermore, sex-specific differences seem to exist among the different specific phobia subtypes: animal, situational and natural environment phobias have the highest prevalence among women (12.1%, 3.2–5.4%, 2–7.5% respectively, [25]). However, findings regarding gender differences in prevalence rates for BII phobia are rather inconclusive [23, 24]. Specific phobia comprises distinct heterogeneous disorders which are additionally influenced by gender [27]. The complexity between gender effects and variability of specific phobia subtypes has not been sufficiently considered in etiologic research [28–31]. Interestingly, there is some evidence suggesting that GSE varies across genders, with men scoring higher in GSE than women [32]. Likewise, gender differences in domain-specific self-efficacy have been reported, e.g. academic self-efficacy (higher in male students, [33]), social self-efficacy (higher in female adolescents, [34]) and computer self-efficacy (higher in male students, [35]). The current study, therefore, examined whether gender differences in self-efficacy levels are functionally related to gender differences in specific fear subtypes.

We performed an online survey in a representative sample including questionnaires measuring symptoms of four different specific fears (animal, natural environment, situational and BII), the corresponding SSE scales and the GSE scale. We predicted that there will be a negative association between specific fears and GSE and SSE and that the association might be weaker for GSE relative to SSE. We further expected different associations between GSE and SSE depending on the specific fear category. Further exploratory analyses were performed to examine the contribution of gender effects in the above-mentioned predicted associations. In a subsample of participants, the trait anxiety questionnaire [36] was used to control for a possible influence of trait anxiety on the sense of self-efficacy and specific fears as shown earlier [14]. Finally, we predicted that all SSE scores would maintain their correlation with specific fears even when controlling for trait anxiety.

Methods

Results

Descriptive statistics

Demographic information and test results of the sample are shown in Table 1. A comparison of the specific fear scores (animal: M = 17.04, SD = 5.67; natural environment: M = 15.32, SD = 5.17; situational: M = 12.32, SD = 4.22; BII: M = 23.06, SD = 8.60) suggests that fear was more prevalent in situations related to animals and natural environment, whereas it was the lowest in situation-related fear. For SSE, the results show that animal-related SSE had the lowest scores while situation-related SSE had the highest scores.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Descriptive statistics.

https://doi.org/10.1371/journal.pone.0283660.t001

Relationships between GSE, SSE and specific fear

Spearman correlation analyses were used for all demographical data and questionnaire scores, namely gender, age, GSE, SSE and specific fear (see Table 2). The correlations between the different fear categories revealed that fear related to natural environment and situations correlated highly with each other (r(715) = .64, p < .001) and BII-related fear seemed to be more distinct as it only had moderate correlations with the other fear categories (see Table 2). Furthermore, GSE scores correlated negatively with all fear categories, among which situation-related fear correlated the highest (r(715) = -.35, p < .001) and BII-related SSE the lowest (r(715) = -.23, p < .001). This difference was confirmed through a significant result of a Fisher’s z-Test (z = 2.48, p = .007). This outcome implies that higher GSE scores are related to less fearfulness, although these relations seem to be weak. The analysis of data related to avoidance of phobic situations followed the same pattern. In particular, avoidance was highly correlated with specific fear (animal fear: r(715) = .81, p < .001; situational fear: r(715) = .77, p < .001; natural environment fear: r(715) = .81, p < .001; BII fear: r(715) = .80, p < .001) and showed the same pattern with respect to the association with GSE (animal avoidance: r(715) = -.23, p < .001, z = -.60, p = .27; situational avoidance: r(715) = -.30, p < .001, z = -1.06, p = .15; natural environment avoidance: r(715) = -.29, p < .001, z = -.21, p = .42; BII avoidance: r(715) = -.24, p < .001, z = -.20, p = .42) and SSE (animal avoidance: r(715) = -.53, p < .001, z = -.26, p = .40; situational avoidance: r(715) = -.39, p < .001, z = -.68, p = .25; natural environment avoidance: r(715) = -.47, p < .001, z = 0, p = .5; BII avoidance: r(715) = -.52, p < .001, z = .79, p = .21). The correlations between the different specific fears and SSE were higher than the correlations with GSE and actually followed an opposite pattern. Specifically, BII-related fear had the highest negative correlation with SSE (r(715) = -.55, p < .001), whereas situation-related fear showed the lowest correlation (r(715) = -.42, p < 0.001). This difference was supported by a significant result of a Fisher’s z-Test (z = 3.23, p = .001).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Correlations table.

https://doi.org/10.1371/journal.pone.0283660.t002

Regression models for all specific fear categories

To approach the question of which factors have the highest predictive power for the specific fears, multiple linear regressions with the subsample were conducted (see Table 3). In the multiple linear regression model for animal-related fear, gender (β = -.05, p = .30), age (β = -.02, p = .73) and GSE (β = -.03, p = .67) were not significant predictors. The overall regression model was statistically significant (R² = .23, F(5, 345) = 20.43, p < .001) and accounted for 23% of the variation in animal-related fear. SSE (β = -.39, p < .001) and trait anxiety (β = .18, p = .004) were significant predictors with SSE predicting the largest amount. Next, a regression analysis was performed to evaluate the prediction of natural environment-related fear. The analysis yielded three insignificant predictors, which were GSE (β = -.05, p = .43), age (β = -.01, p = .84) and gender (β = -.01, p = .85). Trait anxiety (β = .21, p = .001) and SSE (β = -.37, p < .001) significantly predicted environment-related fear and yielded a significant model, R² = .24, F(5, 345) = 21.79, p < .001, that can explain 24% of the variance in natural environment -related fear with SSE explaining the largest amount. For situation-related fear, the analysis showed that all variables except gender (β = .03, p = .49) and age (β = .02, p = .74) significantly predicted situation-related fear and together explained 24% of the variance, R² = .24, F(5, 345) = 22.11, p < .001. Among the predictors, SSE turned out to be the most influential (β = -.29, p < .001) and trait anxiety the least (β = .15, p = .01). Lastly, a linear regression model for BII-related fear was conducted which showed that gender (β = -.02, p = .75), age (β = -.02, p = .63), GSE (β = -.06, p = .34) and trait anxiety (β = .08, p = .18) did not predict BII-related fear. In total, the model explained 30% of the variance of BII-related fear (R² = .30, F(5, 345) = 29.17, p < .001) with SSE being the only significant predictor (β = -.51, p < .001).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Regression models for all specific fears.

https://doi.org/10.1371/journal.pone.0283660.t003

Discussion

Our results are in line with the general proposition of the self-efficacy concept of anxiety that the perceived ability to cope with threatening stimuli and situations contributes to specific fears and phobias [6]. Although GSE was related to specific fears, as previously reported for anxiety and negative affect [11], a relatively stronger association between SSE and specific fear and avoidance was found. GSE seems to be related to a broad range of personality factors (such as optimism or self-esteem, [11]), but all in a moderate range. Likewise, we could show that GSE is related to all different fear categories to a small significant extent, whereas SSE had moderate to high relations to its corresponding fear. This is in accordance with previous findings by Muris [14], who showed that SSE is most strongly related to specific types of anxiety disorders in adolescents. Our findings support the conclusion that this pattern also exists in adults and may be applied to specific fears. To further examine the relationships between SSE and specific fears, we have demonstrated that this relationship stays robust even when controlling for trait anxiety and gender. Interestingly, GSE lost its significance or was the smallest predictor for specific fear, while SSE was the strongest predictor for all specific fears. These findings support the assumption that SSE is more linked to specific fears than trait anxiety. Our results indicate that trait anxiety and self-efficacy may display unique connections to specific fears and phobias.

BII-related fear demonstrated the highest association with SSE compared to all other specific fears. At the same time, BII-related fear was not related to GSE after controlling for SSE. In contrast, natural environment-related fear showed the lowest overall connection with self-efficacy. These findings demonstrate that certain irrational cognitions and beliefs are more tied to certain phobic subtypes [22]. Thorpe & Salkovskis [22] for example showed that spider phobics exhibit higher levels of belief in the amount of harm than non-spider phobics and non-phobic individuals. In contrast to other phobia subtypes, BII-related phobia is associated with strong diphasic physiological reactions [21] which may result in fainting. Fear to faint in public situations is central to BII phobia [42] and thus differs from beliefs and concerns which are more characteristic of other phobia subtypes [22]. It is possible that BII-related fears, i.e. beliefs to have a decreased control over bodily reactions and a belief to be unable to exert control over strong physiological reactions (fainting, [21]) are linked to lower SSE. While this conclusion is speculative, our results nevertheless support the conclusion that the cognitive determinants (and thereby different associations with self-efficacy) related to different fears might vary.

Women are more affected by specific phobias than men (e.g. [43]). The most common forms of phobias in women are those related to animals and heights [44]. In line with this, we found that animal- and situation-related fear are higher in women relative to men. This gender effect, however, was not found for other specific fears. In accordance, gender effects were detected in animal-related SSE, whereas no gender differences were found for other specific fears. Gender effects in situation-related fear were not evident in situation-related SSE, probably because a combination of multiple factors predicts situation-related fear. However, the current findings suggest that gender effects in animal-related fears might be explained by gender differences in animal-related SSE. Animal phobias often have their main focus on feelings of disgust [3] and are generally associated with higher disgust sensitivity [45]. Women show higher levels of disgust sensitivity [46] and therefore might be more disgusted by animals such as spiders. High disgust reactions during animal encounters might cause a sense of inability to cope with these situations. Consequently, this low perceived self-efficacy might lead to avoidance of animals and higher levels of fear towards them [16].

The exact mechanisms by which cognitive determinants such as SSE contribute to specific fears and phobias remain elusive. Lower self-efficacy can have an effect on the individual’s capacity to learn from potentially threatening situations. We have shown that lower perceived self-efficacy is related to impaired discriminative fear learning [47], i.e. the ability to distinguish between safety and danger cues. Similarly, higher perceived self-efficacy can promote fear extinction learning [48], i.e. the ability to learn that a stimulus is no longer associated with potential harm. Fear learning and extinction constitute important mechanisms in the development of anxiety disorders [49]. Psychological treatments based on extinction learning and the promotion of self-efficacy are very effective in reducing fear and avoidance related to specific phobias (e.g. [19, 20]). Our findings support the conclusion that interventions aimed to alleviate fear and avoidance related to specific stimuli and situations should target self-efficacy as a central component.

The current study examined a non-clinical sample, which gives a good understanding of the relationship between specific fears and self-efficacy. However, a comparable study in a clinical sample with phobic participants is warranted. Furthermore, our results do not allow any causal explanations of the relationship between self-efficacy and specific fears and phobias. Decreased self-efficacy might potentiate fear and vice versa. Here, the use of experimental paradigms to increase or decrease perceived self-efficacy (see [19, 47, 48]) and examine the effect of self-efficacy interventions on fear and avoidance related to specific fears might be interesting.

In conclusion, our results indicate that SSE strongly contributes to fear and avoidance related to different fear categories. In support of the self-efficacy concept of anxiety, we showed that perceived inefficacy to cope with threatening situations might represent an etiological factor which contributes to the development and maintenance of specific fears and phobias. SSE seems to be especially relevant for certain types of fears, e.g. BII-related fears which are characterized by a decreased control over bodily reactions and/or belief of being unable to exert control over strong physiological reactions (fainting, [21]). A gender-dependent influence on animal-related fears and corresponding self-efficacy was also found. Our findings can help to understand and enhance psychological treatments based on learning and the promotion of self-efficacy in the context of phobias and anxiety disorders.

Supporting information