تعصب حافظه در اضطراب سلامت مربوط به ظرفیت عاطفی کلمات مرتبط با سلامت است
|کد مقاله||سال انتشار||مقاله انگلیسی||ترجمه فارسی||تعداد کلمات|
|35371||2007||11 صفحه PDF||سفارش دهید||محاسبه نشده|
Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)
Journal : Journal of Psychosomatic Research, Volume 62, Issue 3, March 2007, Pages 263–274
Objectives A model based on the associative strength of object evaluations is tested to explain why those who score higher on health anxiety have a better memory for health-related words. Method Sixty participants observed health and nonhealth words. A recognition memory task followed a free recall task and finally subjects provided evaluations (emotionality, imageability, and frequency) for all the words. Hit rates for health words, d′, c, and psychological response times (PRTs) for evaluations were examined using multi-level modelling (MLM) and regression. Results Health words had a higher hit rate, which was greater for those with higher levels of health anxiety. The higher hit rate for health words is partly mediated by the extent to which health words are evaluated as emotionally unpleasant, and this was stronger for (moderated by) those with higher levels of health anxiety. Consistent with the associative strength model, those with higher levels of health anxiety demonstrated faster PRTs when making emotional evaluations of health words compared to nonhealth words, while those lower in health anxiety were slower to evaluate health words. Conclusions Emotional evaluations speed the recognition of health words for high health anxious individuals. These findings are discussed with respect to the wider literature on cognitive processes in health anxiety, automatic processing, implicit attitudes, and emotions in decision making. Gadget timed out while loading
There is evidence that health-related words are more accurately recalled and recognized than nonhealth-related words (the health-word effect) and that this effect is stronger for those who score higher on measures of health anxiety (the health anxiety moderation effect) ,  and . These findings are interpreted to be consistent with cognitive-behavioural models of health anxiety ,  and . While the health anxiety moderation effect has been replicated, less research has focused on the potential mechanisms that underpin it. Signal detection is one theoretical mechanism that has been previously explored  and this paper looks to replicate this and extend previous work on mechanisms by exploring an object-evaluation associative strength model ,  and . Object-evaluation associative strength This model suggests that the stronger the associative strength between an object and its evaluation, the greater the probability that it is activated in memory and that this may reflect automatic processing ,  and . Objects are defined, by Fazio et al.  and , in a broad sense to include a variety of categories (e.g., words); similarly, evaluations include both emotional and cognitive judgments. Furthermore, there are likely to be individual differences in associative strength that are, in part, a function of the interests and knowledge of the individual  and . The health anxious person is more likely to be interested in health-related issues and as such should have stronger associations between health-related objects and their evaluations. Psychological response time (PRT) latencies between objects and their evaluations can be used to index associative strength—shorter PRTs equate to a stronger associated strength ,  and . Psychological response times for evaluation latencies should be quicker for health words, compared to nonhealth words, for those who score higher on health anxiety. This suggests some degree of automatic processing. On the other hand, those lower in health anxiety should be slower to make evaluations of health words, as the process of making the evaluation for these people is more likely to reflect controlled and conscious processing ,  and . Exploring the idea of automaticity within the context of health anxiety allows initial links to be made to the wider literature on implicit cognitions and attitudes ,  and . For example, implicit and explicit attitudes show a double dissociation whereby implicit measures predict spontaneous behaviours and explicit attitudes more planned behaviours . While there is evidence that increased levels of health anxiety are associated with explicit beliefs and attitudes about health and illness ,  and , it is not known whether the same is true for implicit attitudes . Indeed, implicit attitudes may be a better predictor of spontaneous/automatic behaviours (e.g., anxiety reactions or checking behaviours triggered by an environmental factor) associated with health anxiety ,  and  than explicit attitudes . With this in mind, this study offers a first step in this direction by exploring an object-evaluation associative strength model of the health anxiety moderation effect observed for the recognition of health-related words. To test this model, participants were requested to provide evaluations of health and nonhealth words, and the PRTs to make these evaluations were recorded. Three evaluations are explored: (1) emotionality, (2) imageability, and (3) familiarity. Word emotionality There is evidence that words objectively categorized as negative are better recalled ,  and . There is also evidence that those who score higher in health anxiety are more likely to perceive the outcomes of ambiguous health scenarios as more negative  and to “catastrophize” health-related information  and . As such, it is predicted that negative emotional evaluations of health words will be higher, compared to nonhealth words, and this effect will be stronger for those who score higher on health anxiety. Word imageability Work in the area of cognitive heuristics—especially the availability heuristic—suggests that the more easily something is imagined the more likely it is to be activated in memory ,  and . The health-related words used in previous studies , ,  and  tend to be more concrete than the nonhealth words. Furthermore, it has been demonstrated that word concreteness is correlated with imageability . Therefore, ratings of the extent to which images are easily formed should be greater for health words than for nonhealth words. Word familiarity Health words are likely to be more familiar than nonhealth words given their prominence in public health campaigns as well as in the mass media  and . Also, there is some evidence that health anxious people, compared to nonhealth anxious people, seek out more information on health-related topics  and . Therefore, they are more likely to be exposed to health-related words. This increased exposure may be reflected in increased familiarity with health words and increased ease of retrieval from memory . These evaluative dimensions also have direct relevance to models of decision making that focus on emotions/feelings , , ,  and . For example, models such as ‘risk as feeling’ suggests that the extent to which a construct is imageable predicts its emotional evaluation . Furthermore, all of these models suggest that emotion, whether experienced, anticipatory, appraised, or evaluated as information, has a pivotal role to play in decision making and cognition. Given the central role ascribed to emotions, it is hypothesised that, of the three evaluative dimensions explored in this study, emotional evaluations should be the primary explanatory dimension underlying the health-word and health anxiety moderating effects. A moderated mediation model Based on the above arguments a moderated mediation hypothesis for the retrieval of health words in health anxiety is proposed (see Fig. 1). For clarity of expression, this will be described with respect to emotional evaluations (but the argument equally applies to imageability or familiarity). Full-size image (7 K) Fig. 1. Moderated-mediation model. Figure options The emotional evaluation of a word should mediate the relationship between word type (health vs. nonhealth) and hit rate. That is, subjective evaluations of emotionality should account for a significant proportion of the variance in the relationship between word type and hit rate. Furthermore, the strength of the link between the emotional evaluation and hit rate will be stronger for those with higher levels of health anxiety (i.e., moderated by health anxiety). To test whether the moderation by health anxiety is due to increased associative strength, the PRT data will be explored. The PRTs to make emotional evaluations for health words should be faster for those with higher levels of health anxiety and slower for those with lower levels of health anxiety (cf. Ref. ). Signal detection theory: recognition and recall Recognition memory research essentially compares how well individuals are able to indicate whether a presented stimulus is one that they have seen before (target) or is a new item not seen before (distracter). There are a number of indices of recognition: (1) hit rates (how many of the original targets were correctly identified), (2) sensitivity (d′)—the extent to which individual can discriminate between targets and distracters, and (3) criterion (c) or response bias (β)—the extent to which individuals, when uncertain, code stimuli as targets or distracters. These final parameters reflect not just hit rates, but also false alarms (falsely identifying a distracter as target), misses (failing to identify a target as a target), and correct rejections (correctly identifying a distracter as a distracter). A number of models indicate that the ability to recall items influences recognition memory , ,  and . These models suggest that recall ability helps in the editing of recognition memory by reducing the number of false alarms, by a process known as recall-to-reject  and . This model indicates that if distracters are similar to targets, then this similarity aids the recall of the actual targets. This improved recall of the actual target aids the rejection of the distracter. Therefore, to study recognition memory the effects of recall ability need to be controlled. As such, recall accuracy is controlled in the analyses reported in this paper. Thus, within the context of this study, levels of health anxiety are used to predict (1) hit rates, (2) d′ [which reflects the ability to discriminate whether health or nonhealth words have been seen before (targets) or are new (distracters)], and (2) c (the tendency to view all health or nonhealth words as ones they have seen before or new). Pauli and Alpers  showed that health anxious patients, compared to controls, had a less conservative criterion for pain words but do not differ on d′. That is, patients reported they had seen pain words before even if they had not. Previous memory bias studies and health anxiety A number of studies have reported the health-word effect and its moderation by health anxiety (or related constructs) for both recall ,  and  and recognition . Brown et al.  explored perceptual bias in two samples of health anxious and control subjects using degraded words (health and nonhealth) presented in four pseudo-random orders. Subjects then rated all words (nondegraded) for frequency and finally performed a free recall task. They showed a health-word effect. However, while health anxious subjects recalled more health than nonhealth words, there was no significant interaction between word type and group (heath anxious and controls). Pauli and Alpers  presented patients (health anxious and pain) and controls with a series of word sets (pain, neutral, positive and negative: in groups of three). An immediate free recall task followed each word set. After all four sets, there was a recognition memory task, and d′ and β were calculated. Finally, a delayed recall task was conducted. They showed that health anxious individuals recalled more pain words compared to those with no health anxiety and had a lower criterion for negative and pain words. Lim and Kim  had three patient groups (depression, panic, and somatoform) and a control group complete an emotional Stroop (subliminally and supraliminally), a word identification test, an awareness check (word–nonword discrimination), a distracter task (the WAIS), and, finally, an encoding and immediate free recall task. They used four groups of words: (1) physical health, (2) positive, (3) negative, and (4) category neutral (e.g., scissors). Seventy-two words (18 from each group were used in the emotional Stroop through to the awareness check) and 48 new words (12 from each group) were used in the encoding and immediate recall phase. Of relevance to this study, they showed that somatoform patients, compared to controls, recalled a higher proportion of physical health words than controls. The present study extends these theoretically by exploring the object-evaluation model and methodologically by (1) matching all words (target and distracter) for word frequency as well as word and syllable length, (2) ensuring that the presentation of words for encoding and recognition was individually randomized for each participant, (3) controlling for the effects of initial recall accuracy when predicting recognition memory performance , ,  and , and (4) extending the number of health-word categories to include cardiac, cancer, and infection words [2,p52]. Hypotheses The following hypotheses are explored. Hypothesis 1 The health-word effect. a. There will be a significantly higher hit rate for health words compared to nonhealth words. b. Health words will be evaluated as being significantly more emotionally negative, easier to imagine and having higher subjective familiarity. Hypothesis 2 Moderation by health anxiety. The hit rate for health words will be higher for those who score higher on health anxiety. Hypothesis 3 Mediated moderation effects of health anxiety. The evaluated negative emotional content (imageability or subjective familiarity) of words will mediate the relationship between word type (health vs. nonhealth) and hit rate, and this effect will be stronger for those higher on health anxiety. Hypothesis 4 Mechanisms. a. Object-evaluative associative strength—PRTs for evaluating the emotionality (imageability or familiarity) of health words should be faster for those higher in health anxiety and slower for those lower in health anxiety. b. Signal detection effect—Higher health anxiety will be related to a less stringent criterion (response bias) for health words .
نتیجه گیری انگلیسی
Descriptive statistics Means, standard deviations, reliability, and intercorrelations among the Level 2 variables are presented in Table 1. All variables are reliable, and the pattern of correlations is consistent with previously reported work  and . In terms of responses to the HADS, 3% of the sample were greater than borderline cut-off for depression and 25% for anxiety. The mean for neuroticism (6.28) is slightly higher than that previously reported for men (4.95) and for women (5.90) . Table 1. Means, S.D.'s, reliabilities, and correlations for the Level 2 variables (N=60) Mean S.D. (1) (2) (3) (4) (5) HADS anxiety (1) 8.43 3.92 (.84) HADS depression (2) 3.33 2.94 .70⁎⁎⁎ (.77) SSAS (3) 30.38 5.00 .13 .17 (.65) Neuroticism (4) 6.28 2.57 .66⁎⁎⁎ .52⁎⁎⁎ .25⁎ (.69) Health Anxiety (5) 24.27 5.67 .33⁎⁎ .20 .42⁎⁎⁎ .38⁎⁎⁎ (.76) Coefficient alphas in parentheses on the diagonal (except Neuroticism which is the KR-20). ⁎ P<.05 (two-tailed). ⁎⁎ P<.01 (two-tailed). ⁎⁎⁎ P<.001 (two-tailed). Table options A series of MLMs were used to explore the basic differences between health and nonhealth words for hits, evaluations, and PRTs. Means and standard deviations and t-tests are given in Table 2. There were significant differences between health and nonhealth words for (1) hit rate, (2) imageability ratings, and (3) emotionality ratings. Participant recognition performance was better for health words, and health words were rated as being more imageable and as having more negative emotional content. Table 2. Descriptive statistics for hit rate and ratings by word type (health vs. nonhealth words) Evaluations Psychological response times Health Nonhealth t Health Nonhealth t Hits 83% (37) 71% (45) 7.9⁎ 1085.1 (945) 1150.5 (960) −1.6 ns Familiarity 3.60 (1.82) 3.62 (1.75) −0.3 ns 2516.6 (2376.6) 2638.8 (2631.0) −1.7 ns Imageability 5.05 (1.82) 4.49 (1.89) 6.3⁎ 2904.9 (2154.0) 2803.6 (2006.8) −1.4 ns Emotionality 2.75 (1.30) 3.76 (1.22) −13.6⁎ 2594.9 (1883.8) 2505.4 (1928.6) 1.06 ns ns indicates nonsignificant. Psychological response times in milliseconds (non-log-transformed data presented). Familiarity (1-7) high score=familiar; imageability (1-7) high score=easily imagined; emotionality (1-7) low score=unpleasant. ⁎ P<.001. Table options Free recall effects and intrusion errors The following MLM was used to test free-recall effects of word type and its moderation by health anxiety: View the MathML sourceCorrect recall=β0+β1(health vs. nonhealth words)+r Turn MathJax on View the MathML sourceβ1=γ10+γ11(health anxiety)+u1 Turn MathJax on The term γ10 indicates if, on average, the relationship between recall and word type (nonhealth/health) was different from zero. Whether health anxiety moderated the recall-health relationship was tested by the significance of the term γ11. In terms of immediate recall, the results showed an effect for word type, such that health words were more likely to be recalled than nonhealth words (γ10=.46, t=5.22, P<.001). This was not moderated by health anxiety (γ11=−.001, t=−0.10, P=.920). Given that there was a main effect for word type on recall, this is controlled in subsequent analyses. 2 Intrusion errors (i.e., the recall of either health or nonhealth words that were not on the original list) were observed in free recall responses. There were more health-related intrusion errors (15 across all participants) than nonhealth intrusion errors (8); however, the absolute number is small. Health-word effect: health word hits and moderation by health anxiety (H1a and H2) The following MLM was used to test hypotheses that there will be a significantly higher hit rate for health words than for nonhealth words and that the hit rate for health words will be stronger for those who score higher on health anxiety: View the MathML sourceHit=β0+β1(health vs. nonhealth words)+r Turn MathJax on View the MathML sourceβ1=γ10+γ11(health anxiety)+u1 Turn MathJax on The results are shown in Table 3. Health words were more likely to be recognized accurately (0.76). Individuals with higher health anxiety were more likely to recognize health words over nonhealth words (0.07).