Blanchette, I. and Richards, Anne and Cross, A. (2007) Anxiety and the interpretation of ambiguous facial expressions: the influence of contextual cues. Quarterly Journal of Experimental Psychology 60 (8), pp. 1101-1115. ISSN 1747-0218.
Abstract
In 3 experiments, we investigate how anxiety influences interpretation of ambiguous facial expressions of emotion. Specifically, we examine whether anxiety modulates the effect of contextual cues on interpretation. Participants saw ambiguous facial expressions. Simultaneously, positive or negative contextual information appeared on the screen. Participants judged whether each expression was positive or negative. We examined the impact of verbal and visual contextual cues on participants’ judgements. We used 3 different anxiety induction procedures and measured levels of trait anxiety (Experiment 2). Results showed that high state anxiety resulted in greater use of contextual information in the interpretation of the facial expressions. Trait anxiety was associated with mood-congruent effects on interpretation, but not greater use of contextual information. The way we feel colours our understanding of what is happening around us. Ample empirical evidence now confirms that emotion produces systematic biases in the way people interpret ambiguous stimuli. This has been shown with anxiety in particular. An anxious person may be more likely to perceive a slight smile as a hostile sneer rather than a genuine but mild acceptance gesture. While there is research confirming the existence of such anxiety-related biases in the interpretation of ambiguous information, most of this research has focused on ambiguous stimuli presented in isolation. This is rarely the case in everyday experience. What happens if the slight smile is accompanied by a positive comment? How do anxious people process such contextual cues that could be used to resolve ambiguity? The research reported in this paper attempts to answer that question: What is the effect of anxiety on the use of contextual information in the resolution of visual ambiguity? Interpretive bias in anxiety There is much evidence that anxiety is related to interpretive biases. Anxious individuals tend to interpret ambiguous information in more threatening ways than do nonanxious individuals (Byrne & Eysenck, 1993; Calvo & Castillo, 1997; Calvo, Eysenck, & Castillo, 1997; Calvo, Eysenck, & Esteves, 1994; Eysenck, Macleod, & Mathews, 1987; Hadwin, Frost, French, & Richards, 1997; Hirsch & Mathews, 1997; MaLeod & Cohen, 1993; Mathews, Richards, & Eysenck, 1989; Richards, Reynolds, & French, 1993). Classic demonstrations of this effect include Eysenck and colleagues’ (1987) use of a homophone spelling task. Participants heard a word with two potential spellings, one threatening and one neutral (e.g., die/dye). Anxious participants tended to adopt the threatening interpretation more frequently than did nonanxious participants. Anxiety-congruent interpretation of lexical ambiguity has also been shown using word-naming tasks and lexical decision tasks (Calvo & Castillo, 1997; Calvo et al., 1994; Calvo et al., 1997; MacLeod & Cohen, 1993; Richards & French, 1992). It has also been observed with more complex verbal stimuli such as ambiguous sentences (Calvo & Castillo, 1997; Calvo et al., 1997; Calvo et al., 1994; Eysenck, Mogg, May, Richards, & Mathews, 1991) and text passages describing complex social situations (Constans, Penn, Ihen, & Hope, 1999; Hirsch & Mathews, 1997). Overall, there is abundant empirical evidence that anxious individuals are more likely to err on the side of danger alarm than are nonanxious individuals. While most of the research has examined ambiguous verbal stimuli, there is also some evidence that judgement of facial expressions may be affected by anxiety (Everhart & Harrison, 2000; Liebman & Allen, 1995; Richards, French, Young, Calder, Webb, & Fox, 2002). Ambiguous information in context The study of interpretive biases in anxiety, for both visual and verbal stimuli, has almost exclusively focused on ambiguous stimuli presented in isolation. However, in real-world situations, ambiguous information is rarely encountered out of context. For lexical ambiguity, research in text comprehension shows that context is crucial in determining what interpretation people will adopt (Sereno, 1995; Spivey-Knowlton, Trueswell, & Tanenhaus, 1995; Vu, Kellas, Metcalf, & Herman, 2000; Vu, Kellas, & Paul, 1998). Context may be particularly important in the interpretation of facial expressions, given their transient and dynamic nature. To fully understand the influence of anxiety on interpretation in more realistic settings, we need to examine whether anxious individuals differ in the way they use contextual cues in the interpretation of ambiguous facial expressions. Based on previous theoretical and empirical work on anxiety, conflicting predictions could be made concerning the effect of anxiety on the use of contextual information. According to Easterbrook's influential cue-utilization theory (1959), increased emotional arousal such as that felt in anxiety will lead to a restriction in attentional and processing capacity. Anxious participants would be expected to have a narrowed range of attention and process a more limited set of information. Thus, when ambiguous information is presented in context, we would expect anxious participants to be less influenced by contextual information than would nonanxious participants. The cue-utilization theory is consistent with some findings on the effect of negative emotion on memory, for instance the weapon-focus effect (Christianson, 1992). In contrast, other research suggests that anxiety may be associated with a broadening of attention and of the range of information processed. For instance, Koegh and French (1999) found that participants induced to feel anxious were more likely to process peripheral stimuli than were nonanxious participants. While this experiment examined basic perceptual and attentional processes, conceptually related effects have been found with higher level cognitive processes. Studying how people reason about political issues, Marcus and colleagues (MacKuen, Marcus, Neuman, Keele, & Wolak, 2001; Marcus, Sullivan, Theiss-Morse, & Wood, 1995) found that when participants were anxious, they tended to seek out more information, and a wider range of information, before reaching a decision. Thus, very different tasks have provided evidence suggesting that anxiety may lead to a broadening of attention and information processing. This would lead to the prediction that anxious participants should use contextual information to a greater extent than would nonanxious participants. Our own recent research shows that anxiety has a strong effect on the use of contextual cues in the resolution of verbal ambiguity (Blanchette & Richards, 2003). In a series of experiments, participants heard ambiguous homophones (e.g., die/dye) while simultaneously reading associated contextual cues. These contextual cues could be related to the emotional (e.g., death) or neutral (e.g., hair) interpretation of the homophone. We measured the extent to which participants adopted the emotional or neutral interpretation. Results consistently showed that participants who had been made to feel anxious were more likely to be influenced by contextual cues than were control participants. That is, if the context was emotional, anxious participants were more likely to adopt the emotional interpretation than were control participants. However, if the context was neutral, they were more likely to adopt the neutral interpretation. This was true not only for threat/neutral homophones, but also for positive/neutral and neutral/neutral homophones. Overall, anxiety led to increased sensitivity to contextual information in the interpretation of ambiguous information. In a subsequent study (Richards, Blanchette, & Munjiza, in press), we found the same contextual effects in a group of patients about to undergo dental treatment. The effect therefore seems to generalize to more externally valid moods. In the experiments reported here, we examine whether the effect of anxiety on the use of context in interpretation extends to visual stimuli, using ambiguous facial expressions of emotion. We asked participants to judge images created by blending different facial expressions (morphs). Participants viewed these morphs simultaneously with contextual words (either positive or negative). We examined whether context had an impact on participants’ interpretations and, crucially, whether this impact was the same for participants who had been induced to feel anxious and participants in a control group.
Metadata
| Item Type: | Article |
|---|---|
| School: | Birkbeck Faculties and Schools > Faculty of Science > School of Psychological Sciences |
| Depositing User: | Sarah Hall |
| Date Deposited: | 21 Jan 2020 15:05 |
| Last Modified: | 02 Aug 2023 17:57 |
| URI: | https://eprints.bbk.ac.uk/id/eprint/30648 |
Statistics
Additional statistics are available via IRStats2.
Archive Staff Only (login required)
