Anxiety Associated With Social Issues: the Development of a Scale to Measure an Antecedent Construct

Trina Sego, University of Texas at Austin
Patricia A. Stout, University of Texas at Austin
ABSTRACT - Emotional response to issue-related advertisements is generally studied by manipulating emotional appeals in advertisements and measuring aggregate effect on subjects. Such an approach overlooks individual response based on chronic emotional states associated with the social issues presented in the ads. In this paper, we introduce the first stage in a program of research which aims to examine how anxiety, as a chronic emotional state directed at an issue, affects how individuals respond to messages about the issue. Specifically, we develop a measure of issue-related anxiety and examine its relationship to reported behavioral intent. Results suggest that at least two dimensions of issue-related anxiety associated with AIDS and recycling are negatively related to intent to communicate about and to support those social issues.
[ to cite ]:
Trina Sego and Patricia A. Stout (1994) ,"Anxiety Associated With Social Issues: the Development of a Scale to Measure an Antecedent Construct", in NA - Advances in Consumer Research Volume 21, eds. Chris T. Allen and Deborah Roedder John, Provo, UT : Association for Consumer Research, Pages: 601-606.

Advances in Consumer Research Volume 21, 1994      Pages 601-606

ANXIETY ASSOCIATED WITH SOCIAL ISSUES: THE DEVELOPMENT OF A SCALE TO MEASURE AN ANTECEDENT CONSTRUCT

Trina Sego, University of Texas at Austin

Patricia A. Stout, University of Texas at Austin

ABSTRACT -

Emotional response to issue-related advertisements is generally studied by manipulating emotional appeals in advertisements and measuring aggregate effect on subjects. Such an approach overlooks individual response based on chronic emotional states associated with the social issues presented in the ads. In this paper, we introduce the first stage in a program of research which aims to examine how anxiety, as a chronic emotional state directed at an issue, affects how individuals respond to messages about the issue. Specifically, we develop a measure of issue-related anxiety and examine its relationship to reported behavioral intent. Results suggest that at least two dimensions of issue-related anxiety associated with AIDS and recycling are negatively related to intent to communicate about and to support those social issues.

INTRODUCTION

Educational messages sponsored by groups advocating awareness of various social causes are common. Undoubtedly, individuals sometimes respond to these messages with feelings of guilt or anxiety driven by the seriousness of the consequences associated with the issues. Often, this anxiety is further heightened by the use of "fear appeals" in advertising messages.

The "fear appeals" literature, while inconclusive, suggests that responses characterized by high degrees of anxiety hinder persuasion. The issue itself, independent of the "appeal" used in the message, may stimulate some feelings of anxiety and therefore affect response to the message.

In this exploratory study, we examine feelings associated with two social issues, AIDS and recycling. Specifically, we develop a measure of issue-related anxiety and examine its relationship to intent to communicate about and to support those social issues.

BACKGROUND

While consumer behavior researchers have shown an active interest in emotional responses to advertisements (e.g., Agres, Edell and Dubitsky 1990), most studies ignore antecedent conditions which might affect individuals' responses. Some researchers recognize that antecedent emotional states, particularly mood, can affect response to messages (Batra and Stayman 1990; Gardner 1985; Gardner and Hill 1988).

Moods are general and pervasive feeling states which are not directed toward specific objects (Gardner and Hill 1988). In this paper, we are interested in chronic feeling states which are directed toward specific objects. In particular, we are interested in anxiety because of its relevance to social issues and the degree of research attention the construct has received in psychology (see Zuckerman 1976). Because we assume that prior experience is a powerful influence on such states, the emphasis here is on measurement rather than manipulation of feelings.

Responses characterized by anxiety have long been studied within the "fear appeals" literature, where studies often examine social issues rather than products (see Rotfeld 1988; Sutton 1982). Research in this tradition generally takes a message-centered approach (focusing on messages that provoke anxiety) rather than a person-centered approach (focusing on the effects of anxiety based on individuals' experiences). Thus previous research fails to acknowledge that the individual has a history of emotional experiences related to social issues.

Anxiety which is directed toward specific issues, AIDS and recycling, is examined here. These issues were selected because they are timely and are likely to elicit different degrees of anxiety. Research examining responses of fear or anxiety to messages about environmental issues is scarce and warrants greater attention (Hine and Gifford 1991). Similarly, consumer researchers continue to call for additional research related to AIDS (Cooper-Martin and Stephens 1990; Hill 1989). Few studies examine anxiety as a variable in response to AIDS messages (Hill 1988; Stout 1990), and existing studies fail to account for individuals' previous experiences.

RESEARCH QUESTIONS

This study will focus on four research questions, including the following:

Question 1: What are the underlying dimensions of issue-related anxiety?

Analyses of general anxiety scales often reveal cognitive and somatic dimensions (Delmonte and Ryan 1983; Schalling, Cronholm and Asberg 1975; Schwartz, Davidson and Goleman 1978; see also Kellner 1988). The cognitive dimension typically represents worry and preoccupation while the somatic dimension represents physical symptoms such as a perspiration or nausea.

Many anxiety scales are said to have other dimensions. For example, the Symptom Rating Test is associated with four dimensions: anxiety, depression, somatic symptoms, and inadequacy (Kellner and Sheffield 1973). The Hopkins Symptom Checklist is associated with nine factor analytically validated subscales, of which the "anxiety" and "somatic" subscales are popularly used (Derogatis, Lipman and Covi 1976; Kellner and Uhlenhuth 1991). While the present study is exploratory, we expect that analysis of issue-related anxiety items may reveal cognitive and somatic factors.

Question 2: What is the relationship between issue-related anxiety and issue-related behavioral intentions?

As previously noted, the "fear appeals" literature, although inconclusive, suggests that high anxiety may discourage individuals from paying attention to a persuasive appeal, or from acting upon it (Sutton 1982). Behavioral intentions are operationalized here as intent to communicate about a social issue and intent to support a social issue. We expect that issue-related anxiety will be negatively related to these intentions.

Question 3: What is the relationship between emotional response associated with an issue and issue-related behavioral intentions?

Research on emotion in the consumer behavior literature has focused primarily on the use of emotion to create favorable brand attitudes through transfer of affect from attitudes toward the ad (Holbrook and O'Shaughnessy 1984; Holbrook and Batra 1987; Stayman and Aaker 1988), and only incidentally on behavioral intentions. Little research on emotional response and social issues exists outside of the fear appeals literature (Rotfeld 1988). Emotion typically encompasses a range of different feelings, including fear or anxiety (Izard 1977; Plutchik 1980). We expect that issue-related anxiety and emotional response are two different constructs and that emotional response will have only a weak role in predicting behavioral intentions (Stout and Leckenby 1986; 1988).

Question 4: How does the ability of issue-related anxiety to predict behavioral intentions compare to the ability of emotional response items to predict behavioral intentions?

Issue-related anxiety is a more specific construct than emotional response and as such could have more or less predictive power. Issue-related anxiety might have less predictive power because it measures only one emotion rather than several dimensions of emotional experience. However, issue-related anxiety might have more predictive power in situations where anxiety is a particularly relevant emotion. We expect that anxiety is particularly relevant to the social issues studied here, and that issue-related anxiety will be a stronger predictor of behavioral intentions than emotional response will be.

METHODOLOGY

Subjects and Procedure

A total of 103 undergraduates (44 percent male, 55 percent female) participated in the study in exchange for extra course credit. Each respondent answered all items for two issues, AIDS and recycling, resulting in a total of 206 observations for each item (issues were rotated to control for order effects). Students were considered an appropriate sample because they tend to be at moderate risk for exposure to AIDS (Edgar, Freimuth and Hammond 1988), and because they have expressed concern for the environment and will have much influence on environmental issues in their lifetimes (Buttel 1979; Dunlap, Gale and Rutherford 1973).

Respondents were told they were participating in a research project asking their feelings, opinions and behavior in general and related to certain social issues. Respondents completed the questionnaire at their own pace, which took approximately 10-15 minutes.

Independent Measures

Current state of anxiety, issue-related anxiety, and issue-related emotional responses were measured as independent measures. Current state of anxiety was assessed using the 20-item state anxiety form from Spielberger's (1983) State-Trait Anxiety Inventory (STAI). Items were scored using a four-point scale (1=not at all, 4=very much so). The STAI is widely used and considered reliable (Levitt 1980).

Issue-related anxiety was measured using fifteen Likert items (1=strongly agree, 5=strongly disagree). Some items were developed by adapting items from existing scales, including Taylor (1953), Cattell and Scheier (1963), Kellner and Sheffield (1973), and Spielberger (1983), which are used to assess general and clinical anxiety. Such scales are widely available; however, items had to be adapted to measure issue-related anxiety (for reviews of anxiety scales, see Kellner and Uhlenhuth 1991; Thompson 1989).

Many items from existing scales were not applicable because they contained clinical terms, they were not adaptable to Likert scaling, their wording seemed cliched or dated (e.g., "I am often afraid that I am going to blush" Taylor 1953), they were too general ("feeling inferior to other people" Kellner and Sheffield 1973), or they lacked face validity ("I am inclined to let my actions get swayed by jealousy" Cattell and Scheier 1963). Items dealing with cognition, negative affect, arousal, physiological symptoms, and general statements about anxiety were included in the questionnaire.

The respondent's emotions about the issue were measured using eighteen semantic-differential items developed by Russell, Ward, and Pratt (1981). The scale, which is designed to measure three dimensions of emotion (i.e., pleasure, arousal and dominance), was developed using multidimensional scaling techniques, and has been applied in studies on environmental psychology (Russell 1989; Russell, Ward, and Pratt 1981). Sample items include word pairs like "happy-unhappy" (to measure the pleasure dimension), "stimulated-relaxed" (to measure arousal), and "controlling-controlled" (to measure dominance).

Data on knowledge about each issue, sexual behavior, recycling behavior, the degree to which respondents consider the issue a personal threat as well as a threat to society, and demographics were also collected. An open-ended question asked respondents to recall incidents related to the social issue.

Dependent Measures

Intentions to communicate about and to support the issues were measured using Likert scales (5=strongly agree, 1=strongly disagree). Intent to communicate was measured by two items: "I try to avoid conversation about AIDS [recycling]" and "I always pay attention to reports about AIDS [recycling] in the newspaper and on television." The former item was reverse coded. Behavioral intention was measured by one item: "I intend to take actions to combat AIDS [support recycling]."

RESULTS

A state anxiety score was calculated for each individual based on responses to Spielberger's State Anxiety Inventory using the procedures outlined in Spielberger (1983). This score provides a baseline measure of respondents' current levels of general anxiety. The means (and standard deviations) for state anxiety scores found for females and males are 43.53 (11.20) and 42.57 (14.28) respectively.

Spielberger (1983) reports scores (and standard deviations) for female and male college students of 38.76 (11.95) and 36.47 (10.02) under normal circumstances. He reports that mean scores for female and male college students under exam conditions (when students are anxious about exams) of 60.51 and 54.99. State anxiety scores found here appear to be slightly high for normal conditions, but not as high as scores expected for exam conditions.

To examine the underlying dimensions of issue-related anxiety, the fifteen issue-related anxiety items were analyzed using principal components analysis, with varimax rotation, retaining factors with eigenvalues greater than 1.00. The analysis resulted in four factors, labeled "uncomfortable," "tense," "rational," and "decisive" (see Table 1).

The first factor, "uncomfortable," explained nearly 23 percent of the variance. Items about discomfort and irritation loaded positively on this factor.

The second factor, "tense," explained 22 percent of the variance. Items such as those describing felt anxiety and physiological symptoms loaded positively on this factor.

The third factor, "rational," explained nearly 10 percent of the variance. Items such as those describing rationality and energy when dealing with the issue loaded positively on this factor.

The fourth factor, "decisive," explained nearly 10 percent of the variance, and was the most difficult to interpret. An item about difficulty in thinking and deciding about the issue loaded negatively on the factor (-.705), while an item about becoming excited about the issue loaded positively on the factor (.638).

TABLE 1

FACTOR LOADINGS OF ISSUE-RELATED ANXIETY ITEMS

Mean factor scores and standard deviations for each issue are presented in Table 2. According to the results of t-tests comparing these scores, the mean scores for the two issues differed significantly on only two factors: "uncomfortable" and "rational." This suggests that respondents associated AIDS with significantly less "uncomfortable" feelings and significantly less "rational" feelings than they associate with recycling.

A series of regression analyses used as predictors the state anxiety scores and four factor scores from the issue-related anxiety items (see Table 3). First, regression analyses were performed separately for each issue, which generated some differential results (available from the first author upon request). Then, the analyses were performed with items pooled across issues in order to increase generalizability of the findings.

TABLE 2

T-TESTS ASSESSING DIFFERENCES IN MEAN FACTOR SCORES FOR AIDS AND RECYCLING

The regression equation predicting reported likelihood of conversing about the issue was significant (R2=.190, p=.001). Three independent variables, "tense" (p=.009), "rational" (p=.001) and "decisive" (p=.001), were significant. This suggests that individuals who reported being less "tense," more "rational," and less "decisive" about a social issue were more likely to report a high likelihood of conversing about the issue.

The regression analysis predicting reported likelihood of attending to media reports about the issue was also significant (R2=.214, p=.001). All five independent variables, state anxiety (p=.008), "uncomfortable" (p=.006), "tense" (p=.014), "rational" (p=.020) and "decisive" (p=.001), were highly significant. This suggests that individuals who reported being in a more anxious state and who reported being less "uncomfortable," less "tense," less "rational," and less "decisive" about a social issue were more likely to report a high likelihood of attending the media reports about the issue.

The regression analysis predicting reported behavioral intent to take action to support the issue was significant (R2=.155, p=.001). Two independent variables, "tense" (p=.001) and "decisive" (p=.004), were highly significant. This suggests that subjects who reported being less "tense" and less "decisive" about a social issue were more likely to report strong intent of taking action to support recycling.

The emotional response scale was factor-analyzed (principal components analysis, varimax rotation, minimum eigenvalue=1.00), resulting in four factors. As expected, the first three factors represented pleasure (variance explained=27.7 percent), arousal (variance explained=14.3 percent) and dominance (variance explained=14.1 percent) (Russell, Ward and Pratt 1981). Only two items loaded significantly on the fourth factor (variance explained=7.3 percent). The item with the highest loading on the fourth factor, "autonomous-guided," may have been confusing to undergraduate respondents. (Complete factor loadings are available from the first author upon request.)

The factor scores for four emotional response factors were then used in a series of regression analyses to predict behavioral intent (see Table 3). The dimensions of the emotion scale proved significant predictors of reported likelihood of conversing about the issue (R2=.049, p=.039), reported likelihood of attending to media reports about the issue (R2=.051, p=.032) and reported intent to support the issue (R2=.047, p=.047). However, the strength of the relationship between the emotion dimensions and the dependent variables is weak as evidenced by the small R2. The predictive power of issue-related anxiety appears to exceed that of emotional response.

State anxiety was not included in the regression equation with emotional response items since the antecedent state of anxiety is not theoretically applicable to the emotional response model. The analogous antecedent state of mood was not measured in this study. Two additional regression analyses were performed using as predictors issue-related anxiety without state anxiety, and the dimensions of emotional response with state anxiety, to allow direct comparison of the two affective measures with and without the antecedent state anxiety. State anxiety did not greatly contribute to the predictive power of the emotional response items. In predicting intent to support the issue, the dimensions of emotional response without state anxiety performed better than the same equation including state anxiety (complete results are available from the first author).

In addition, items measuring the perceived threat to self, perceived threat to society, age, sex, marital status, and ethnicity were added to the issue-related anxiety regression analysis. While these items increased the R2 (by an average .053 for R2 and .029 for adjusted R2), none of these additional variables were significant predictors of all three dependent variables. The added variables also had little effect on the significance or directionality of the other independent variables (complete results are available from the first author).

DISCUSSION AND CONCLUSIONS

In this exploratory study, we report the first stage in a program of research which aims to examine how anxiety, as a chronic affective state directed at an issue, relates to how individuals respond to messages about the issue. Of the factors which emerged from the factor analysis of issue-related anxiety items, all four were significant predictors of likelihood to attend to media reports about social issues, while three were significant predictors of likelihood to converse about social issues. Two factors were significant predictors of intent to support social issues.

TABLE 3

RESULTS OF REGRESSION ANALYSES USING STATE ANXIETY AND ISSUE-RELATED ANXIETY TO PREDICT INTENT TO CONVERSE, TO ATTEND TO MEDIA REPORTS, AND TO SUPPORT ISSUE

The factor labeled "tense" was a significant predictor of all three dependent variables. Similar to somatic dimensions found in other studies of anxiety, this factor represents symptoms such as "butterflies in my stomach" and loss of appetite associated with anxiety (Delmonte and Ryan 1983; Schalling, Cronholm and Asberg 1975; Schwartz, Davidson and Goleman 1978). The results suggest that an absence of such symptoms is related to a higher reported likelihood of conversing about an issue, reported likelihood of attending media reports about an issue, and reported intent to support an issue.

The factor labeled "rational" resembles a cognitive dimension that has emerged in other anxiety studies (Schalling, Cronholm and Asberg 1975; Schwartz, Davidson and Goleman 1978). This factor, which represents the lack of ability to think rationally and steadily relative to an issue, was a significant predictor of two of the three dependent variables. These results suggest that individuals who report an ability to think rationally about an issue are less likely to attend to media reports about that issue, but more likely to converse about the issue.

The factor labeled "decisive" explained little variance in the item responses, but was a significant predictor of all three dependent variables. Decisiveness and arousal associated with an issue appears to discourage individuals from conversing about an issue, attending to media reports about an issue, and intending to support an issue. Although results for this factor were not as expected, they are consistent with traditional consumer behavior theory which would suggest that decisiveness, in many situations, would not lead to information search (Solomon 1992). The ambiguity of the factor analysis results and the general measure of intent to support the issue may also have influenced the overall results.

The factor labeled "uncomfortable" accounted for a substantial amount of the variance of responses to issue-related anxiety items, but was a significant predictor of only one dependent measure. The authors suspected that the "uncomfortable" factor might be associated with a general societal concern (rather than perceived threat to self) that is not strongly related to action. However, the addition of items about perceived social threat and perceived personal threat to the regression equations did not affect the direction or significance of the other variables, including "uncomfortable."

Relative to emotional response, issue-related anxiety appears to be a stronger predictor of intent to communicate about and to support social issues. In research situations where a specific feeling (e.g., anxiety) is expected to be relevant, use of a measure that better captures that particular feeling may be a more fruitful approach than use of an emotional response scale (such as those that tap the pleasure-arousal-dominance dimensions).

As with all research studies, the methodology applied here suffers from some limitations. For example, the three dependent variables were each measured using a single item, and intention measures were not time-specific. Future studies will correct these limitations as well as employ larger samples, use more diverse social issues, and examine test-retest reliabilities. In addition, the results presented here offer only correlational evidence; no directionality can, in fact, be determined.

Despite its limitations, these results are a promising first step in a program of research which aims to examine how issue-related anxiety affects individual response to messages about social issues. The results suggest that antecedent affective states may have a powerful effect on response to messages. Future research should continue to address the role of antecedent affective states.

The findings reported here further suggest that campaign planners should take great care when designing messages about social issues such as AIDS and environmentalism. Antecedent feelings associated with the issue, in addition to those provoked by the message appeal, may lead to negative response. The results reported here offer a possible explanation for the contradictory findings evident in the "fear appeals" literature.

REFERENCES

Agres, S. J., J. A. Edell and T. M. Dubitsky (Eds.) (1990), Emotion in Advertising: Theoretical and Practical Explorations. Westport CT: Quorum Books.

Batra, Rajeev and Douglas M. Stayman (1990), "The Role of Mood in Advertising Effectiveness," Journal of Consumer Research, 17, 203-214.

Buttel, F. H. (1979), "Age and Environmental Concern: A Multivariate Analysis," Youth and Society, 10, 237-256.

Cattell, R. B. and I. H. Scheier (1963). Handbook for the IPAT Anxiety Scale. Champaign, IL: Institute for Personality and Ability Testing.

Cooper-Martin, Elizabeth and Debra L. Stephens (1990), "AIDS Prevention through Consumer Communication: Ideas from Past and Current Research," In Advances in Consumer Research, Vol. 17, Eds. Marvin E. Goldberg, Gerald Gorn and Richard W. Pollay, Chicago: Association for Consumer Research, 288-293.

Delmonte, M. M. and G. M. Ryan (1983), "The Cognitive-Somatic Anxiety Questionnaire," British Journal of Clinical Psychology, 22, 209-212.

Derogatis, L. R., R. S. Lipman and L. Covi (1976), "An Outpatient Psychiatric Rating Scale," In ECDEU Assessment Manual for Psychopharmacology, Rockville, MD: U.S. Department of Health, Education, and Welfare, 320.

Dunlap, R. P., R. P. Gale, and B. M. Rutherford (1973), "Concern for Environmental Rights among College Students," American Journal of Economics and Sociology, 32, 45-60.

Edgar, Timothy, Vicki S. Freimuth and Sharon L. Hammond (1988), "Communicating the AIDS Risk to College Students: The Problem of Motivating Change," Health Education Research, 3, 59-65.

Gardner, Meryl P. (1985), "Mood States and Consumer Research: A Critical Review," Journal of Consumer Research, 12, 281-300.

Gardner, Meryl P. and Ronald P. Hill (1988), "Consumers' Mood States: Antecedents and Consequences of Experiential versus Informational Strategies for Brand Choice," Psychology and Marketing, 5, 169-182.

Hill, Ronald P. (1988), "An Exploration of the Relationship between AIDS-related Anxiety and the Evaluation of Condom Advertisements," Journal of Advertising, 17, 35-42.

Hill, Ronald P. (1989), "The Growing Threat of AIDS: How Marketers Must Respond," Journal of Health Care Marketing, 9, 5-12.

Hine, Donald W. and Robert Gifford (1991), "Fear Appeals, Individual Differences, and Environmental Concern," Journal of Environmental Education, 23, 36-41.

Holbrook, Morris B. and Rajeev Batra (1987), "Assessing the Role of Emotions as Mediators of Consumer Responses to Advertising," Journal of Consumer Research, 14, 404-420.

Holbrook, Morris B. and John O'Shaughnessy (1984), "The Role of Emotion in Advertising, Psychology and Marketing, 1, 45-64.

Holbrook, Morris B. and Richard A. Westwood (1989), "The Role of Emotion in Advertising Revisited," In Cognitive and Affective Responses to Advertising, Eds. Patricia Cafferata and Alice M. Tybout, Lexington, Mass.: Lexington, 353-372.

Izard, Carroll E. (1977), Human Emotions, New York, NY: Plenum.

Kellner, Robert (1988), "Anxiety, Somatic Sensations and Bodily Complaints," In Handbook of Anxiety, Vol. 2: Classification, Etiological Factors and Associated Disturbances, Ed. R. Noyes, M. Roth and G. D. Burrows, New York: Elsevier Sciences, 213-237.

Kellner, Robert and Brian F. Sheffield (1973), "A Self-Rating Scale of Distress," Psychological Medicine, 3, 88-100.

Kellner, Robert and Eberhard H. Uhlenhuth (1991), "The Rating and Self-rating of Anxiety," British Journal of Psychiatry, 159, 15-22.

Levitt, E. E. (1980), The Psychology of Anxiety, Hillsdale, NJ: Erlbaum.

Plutchik, Robert (1980), Emotion: A Psychoevolutionary Synthesis, New York: Harper and Row.

Rotfeld, H. J. (1988), "Fear Appeals and Persuasion: Assumptions and Errors in Advertising Research," Current Issues and Research in Advertising, 11, 21-40.

Russell, J. A. (1989). Measures of Emotion. In Emotion: Theory, Research, and Experience, Eds. R. Plutchik and H. Kellerman, San Diego: Academic, 83-111.

Russell, J.A., L. M. Ward, and G. Pratt (1981), "The Affective Quality Attributed to Environments: A Factor Analytic Study," Environment and Behavior, 13, 259-288.

Schalling, D., B. Cronholm and M. Asberg (1975), "Components of State and Trait Anxiety Related to Personality and Arousal," In Emotions: Their Parameters and Measurement, Ed. L. Levi, New York: Raven, 603-617.

Schwartz, G., R. Davidson and D. Goleman (1978), "Patterning of Cognitive and Somatic Processes in the Self-regulation of Anxiety," Psychosomatic Medicine, 40, 321-328.

Solomon, Michael R. (1992), Consumer Behavior: Buying, Having, and Being, Boston: Allyn and Bacon.

Spielberger, Charles D. (1983), Manual for the State-Trait Anxiety Inventory (Form Y), Palo Alto, CA: Consulting Psychologists Press, Inc.

Stayman, Douglas M. and David Aaker (1988), "Are All the Effects of Ad-Induced Feelings Mediated by Aad?" Journal of Consumer Research, 15, 368-373.

Stout, Patricia (1990), "Positive and Negative Emotional Responses to Threat Appeals in AIDS Messages," Paper presented at the conference of the International Communication Association, Dublin, Ireland.

Stout, Patricia A. and John D. Leckenby (1986), "Measuring Emotional Response to Advertising," Journal of Advertising, 15, 35-42.

Stout, Patricia A. and John D. Leckenby (1988), "The Nature of Emotional Response of Advertising: A Further Examination," Journal of Advertising, 17, 53-57.

Sutton, S. B. (1982), "Fear-arousing Communications: A Critical Examination of Theory and Research," In Social Psychology and Behavioral Medicine, Chichester, England: John Wiley & Sons, 303-337.

Taylor, J. A. (1953), "A Personality Scale of Manifest Anxiety," Journal of Abnormal and Social Psychology, 48, 285-290.

Thompson, Chris (1989), "Anxiety," In The Instruments of Psychiatric Research, Ed. Chris Thompson, Chichester, England: John Wiley & Sons, 127-155.

Zuckerman, Marvin (1976), "General and Situation-specific Traits and States: New Approaches to Assessment of Anxiety and Other Constructs," In Emotions and Anxiety: New Concepts, Methods, and Applications, Eds. M. Zuckerman and C. D. Spielberger, Hillsdale, NJ: Lawrence Erlbaum, 133-174.

----------------------------------------