Communicating risk without creating unintended effects.

Abstract: Health risk messages conveying susceptibility to and severity of a threat are prone to producing unintended consequences. This study examined the positive and negative effects of risk messages designed to promote sun protective behavior among Midwestern college students. The results suggest that when strong perceptions of threat are accompanied by strong efficacy perceptions, risk messages can effectively motivate behavior change. When perceived threat was high but perceived efficacy was low, college students denied the risk of skin cancer as a mere exaggeration. Ways of enhancing the effectiveness of health risk messages and of minimizing adverse reactions from college students are discussed.

**********

Motivating preventive behavior among young adults is critical in controlling the incidence of and mortality rate from skin cancer because unprotected sun exposure during young adulthood increases the risk of developing skin cancer significantly (Koh, 1991; Weinstock, Coldlitz, Willett, Stempfer, Bronstein, Mihm, & Speizer, 1989). However, many college students tend to believe that they are invincible to the consequences of risky behavior (Carmel, 1994). The risk of developing skin cancer may appear even more remote for college students in the Midwest, where exposure to the sun is limited during a number of months. However, this does not protect Midwesterners from skin cancer. The incidence of and mortality rates from skin cancer in Michigan, for example, have far surpassed the percentage increases for total cancer rates (Michigan Department of Community Health, 1996). Moreover, skin cancer was "the leading cancer among white Michigan males 25 to 44 years of age in 1994 and the third leading cancer among white females in this age group" (p. 3).

Despite these alarming facts, Midwestern youths are not engaging in appropriate precautionary behavior (Robinson, Rademaker, Sylvester, & Cook, 1997). It is easily observable on Midwestern college campuses that students often neglect regular sun protective behavior and tend to receive intermittent but intense sunburns. Perhaps in order to make up for the little sun that they have during the rest of the year, Midwestern college students tend to indulge themselves in sunlight during summer and spring break times, sometimes "frying" themselves. This pattern of behavior urgently needs to be changed because the effect of such sporadic but severe sunburns during early lifetime tends to be most deleterious (Koh, 1991; Weinstock et al., 1989).

Unfortunately, little research has been done with Midwestern college students to determine which types of interventions can effectively communicate their skin cancer risk and modify risky behavior. Communicating the seemingly unlikely risk of skin cancer to Midwestern college students can be a risky enterprise. Because the thoughts of susceptibility and severity tend to arouse fear and anxiety, health risk messages hold the potential to engender adverse reactions in which individuals defensively deny the health threat or engage in more risky behavior (i.e., "boomerang" effect) (e.g., Kleinot & Rogers, 1982; Rippetoe & Rogers, 1987; Rogers & Mewborn, 1976; Witte, 1992a). Examining such unintended and negative consequences is important, as college students are especially well known for their tendency to defy persuasion regarding behavior change.

The present study was designed to evaluate both the positive and negative effects of health risk messages and to provide practical suggestions for developing effective risk communications. This study begins by briefly reviewing its theoretical framework. Next, the methods and the results of the study are presented. Finally, implications of the results as pertaining to future skin cancer prevention efforts on college campuses are discussed.

THEORETICAL FRAMEWORK: THE EXTENDED PARALLEL PROCESS MODEL

A theoretical framework that explains and predicts the conditions of positive and negative responses to health risk messages (also known as fear appeals) is the Extended Parallel Process Model (EPPM) (Witte, 1992b). The EPPM'S theoretical postulation rests on its two key constructs: perceived threat and perceived efficacy. Perceived threat consists of perceived susceptibility--an individual's estimation of the probability of experiencing a risk--and perceived severity--an individual's estimation of the seriousness of the risk. Perceived efficacy consists of perceived self efficacy--an individual's belief in his/her ability to perform the risk preventive behavior recommended in a message--and perceived response efficacy--an individual's belief in the effectiveness of the recommendations in preventing the risk.

According to the EPPM, the success or failure of risk messages depends on the levels of perceived threat and perceived efficacy. When perceived threat is low, risk messages result in null effects. For example, if college students believe that the risk of skin cancer is remote and that unprotected sun exposure has negligible impact on them, risk messages will be ignored and will produce no effects.

The EPPM predicts that it is only when perceived threat is high that risk messages produce any effects. For risk messages advocating skin cancer prevention to have any impact, college students must believe that the consequences of unprotected sun exposure are serious and that they are likely to experience these severe consequences unless they engage in preventive behavior.

At the same time, however, the EPPM cautions that a high level of perceived threat can create either positive or negative effects, depending on the level of perceived efficacy. On the one hand, when the levels of both perceived threat and efficacy are high, risk messages persuade individuals to adopt recommended preventive behavior. Under this condition, individuals will change their attitude, intentions, and behavior in a manner consistent with the direction of message recommendations to reduce their risk. If college students believe that the serious consequences of skin cancer are relevant to them and that they are able to perform sun-protective behavior easily, risk messages will produce the positive effects of behavior change.

On the other hand, when perceived threat is high but perceived efficacy is low, individuals exhibit adverse reactions, such as perceived manipulation and message minimization. For example, if college students believe that the effects of unprotected sun exposure are damaging (i.e., high perceived threat) but they are not sure whether they can easily perform sun-protective behavior (i.e., low perceived efficacy), they will treat the threat message as a distortion of truth and dismiss it as mere propaganda.

According to the EPPM, message minimization and perceived manipulation are reactance-type responses. Reactance refers to defiance and resentment toward a form of persuasion that appears to threaten an individual's behavioral freedom (Brehm, 1966). Further, the reactance theory predicts that when individuals perceive that their behavioral freedom is threatened, they will change their attitude to the opposite of what is advocated in the persuasion (i.e., "boomerang effect") in an attempt to reclaim the threatened freedom (Brehm, 1966).

Health risk messages urge college students to engage in preventive behavior that delimits their freedom to enjoy the sun. Wearing sunscreen, using protective hats or clothing, and staying in the shade can certainly be perceived as constraints on behavioral freedom to college students. This perception of threatened freedom can be amplified when perceived efficacy is low--when college students think that the recommended behavior is a hassle or is not effective in risk reduction. Consequently, in reaction to a message that threatens their freedom, college students may exhibit a more positive attitude toward risky behavior such as participating in outdoor activities without sun protection. On the basis of the above discussion, the following hypotheses are put forth:

H1. After reviewing a high threat message, individuals with high efficacy perceptions will exhibit more positive attitudes, intentions, and behaviors than those with low efficacy perceptions.

H2. After reviewing a high threat message, individuals with low efficacy perceptions will exhibit more message derogation, will be more likely to dismiss the message as a manipulation, and will display a more positive attitude toward risky behavior than those with high efficacy perceptions.

METHODS

OVERVIEW

This study was done as part of a larger laboratory experiment examining risk message effects on diverse populations. Although lab experiments suffer from limited ecological validity, their controlled setting offers an opportunity to evaluate the clean effects of threat and efficacy. High and low threat messages concerning the risk of skin cancer were developed (see the Skin Cancer Prevention Message section below for the details). Efficacy was intentionally not manipulated in order to avoid inducing harmful effects of risk messages, but participants' existing efficacy perceptions were measured. The experiment was conducted two weeks before the spring break of a large Midwestern university. At the immediate post-test, attitudes toward preventive and risky behavior, intentions to engage in preventive behavior, message derogation, and perceived manipulation were measured. Two weeks after the spring break at the follow-up, students reported their behavior pertaining to skin cancer prevention.

PARTICIPANTS

Undergraduate students recruited from introductory communication-related courses at the large Midwestern university were the participants of this study (N = 274). This convenience sample is justified because the purpose of this study is not to survey the Midwestern college students' skin cancer-related beliefs and behavior, but rather to identify the conditions of positive and negative effects of fear-based risk communication. The ages of the participants ranged from 18 to 37 (SD = 2.1), and their typical age was 20 (M and Mdn = 20, mode = 19). The sex and race composition of the sample resembles that of the university (indicated in parentheses). Overall, 39.4 (46) percent of the participants were male and 60.6 (53) percent of them were female. The majority of the participants were Caucasian, comprising 83.6 (84) percent. Asians, African-Americans, and Hispanics represented 6.9 (4), 6.2 (8), and 1.5 (2) percent of the participants, respectively (University data file, 2002). Four weeks later, 239 participants completed the follow-up questionnaire, producing an attrition rate of 12.7%. This attrition is attributed to students who were absent for classes at which the follow-up tests were administered.

PROCEDURE

The experiment was performed in two adjacent classrooms. Participants arrived at the experiment site in groups of varying sizes, but no more than 30 persons per group. Participants were first introduced to the purpose of the study, and then written consent was obtained. Consistent with previous health risk message studies, the experiment was presented as part of a public health campaign project--Participants were told that the messages were being developed to provide college students with facts about skin cancer and that their responses were important to further improve these messages (e.g., Rippetoe & Rogers, 1987; Witte, 1992a).

Participants were also informed that they would be visited in their classrooms four weeks later for a follow-up questionnaire and that their responses to both questionnaires would remain anonymous. In order to match immediate post-test results and follow-up questionnaires without sacrificing anonymity, each of the participants was asked to provide father's, mother's, and his/her own first initials, the month he/ she was born, and the last two digits of his/her telephone number at the beginning of both immediate and follow-up questionnaires.

After the introduction, participants were randomly assigned to one of the two experimental conditions (high, low threat). Immediately after reviewing either a high or low threat message, participants completed a post-test questionnaire. Four weeks later, participants completed a follow-up questionnaire in their classrooms. After the administering of the follow-up questionnaire, participants were debriefed about the purpose and design of the study in order to clear up any possible misunderstanding caused by high vs. low threat manipulation. Also, skin cancer prevention brochures from the American Cancer Society were distributed.

SKIN CANCER PREVENTION MESSAGES

The high and low threat messages for skin cancer prevention among college students were developed with direct and indirect quotes from existing skin cancer-prevention brochures created by the American Cancer Society and the American Academy of Dermatology. Thus, both messages were based on accurate facts. To ensure that the messages were appropriate for college students, the initial versions of the messages were subjected to four-phase pilot tests in which a total of one hundred and sixty undergraduates recruited from various classes on campus participated for extra credit. Necessary revisions to the messages were made after each pilot test. In the final version, both the high and low threat messages consisted of factual messages, a fictionalized case of a skin cancer patient, and five pictures.

The high threat message underscored the skin cancer information that is significant and relevant to Midwestern college students. To highlight the relevance of sun damage to Midwesterners, the high threat message stated that one can get burned on a cloudy day and that snow reflects the suds rays. Further, the sun's damaging effect on appearance was described in order to make the consequence of risky behavior more relevant to college students (McClendon & Prentice-Dunn, 2001). The high threat message also included gruesome color photos of patients with severe cases of skin cancer. Photos of elderly individuals with skin cancer were not included in order to discourage the perception that skin cancer is relevant to only much older individuals than college students.

The low threat message did not specifically focus on college-aged persons' or Midwesterners' vulnerability to skin cancer. Instead, it merely presented the general facts about skin cancer. For example, it noted that skin cancer is often related to lifetime exposure to the sun and that it fluctuates depending upon the amount of ultraviolet radiation received in a given geographic location. It underscored the high risk among individuals in areas with year-round strong sun. No gruesome pictures of skin cancer were used in the low threat message. Instead, the low threat message included color diagrams conveying images neutral in tone, such as the structure of skin and different types of the sun's radiation.

At the end of both the high and low threat messages was the same list of recommended behavior: using sunscreen, wearing a protective hat and clothing, minimizing sun exposure at midday, and performing periodic skin self-examination. In addition, the high and low threat messages were equated for length, order, complexity, and quality, per Witte's suggestion (Witte, 1993). The threat message manipulations were validated in the main study through manipulation checks (see the results section).

MEASURES

Seven-point Likert-type scales, ranging from "strongly agree" to "strongly disagree," were used, unless otherwise noted. To generate an index for each of the constructs, items in the same construct were averaged. Cronhach's alpha was used to evaluate the reliability of the measures.

Threat. Threat (a = .89) was operationalized as susceptibility to threat and severity of threat. The Risk Diagnostic Scale (RDS) was used to measure this variable and served as a manipulation check (Witte, Cameron, McKeon, & Berkowitz, 1996). The susceptibility scale (a = .85) included three items including, "It is possible that I will develop skin cancer." The severity scale (a = .92) included three items including, "I believe that skin cancer is a severe health problem."

Fear. Fear arousal (a = .95) was measured by asking participants to rate six mood adjectives on a seven-point scale ranging from "not at all" to "very much." The six mood adjectives, "frightened," "tense," "nervous," "anxious," "uncomfortable," and "nauseous," constitute a representative verbal report measure of fear (Witte, 1992b).

Efficacy. Efficacy (a = .90), operationalized as self efficacy and response efficacy, was measured with the RDS. Self efficacy (a = .90) was assessed with fourteen items including, "Wearing sunscreen with SPF-15 or higher to prevent skin cancer when out in the sun is no problem for me" and "I can easily perform periodic skin self-examination to prevent skin cancer." Response efficacy (a = .92) was measured with seven items including, "Minimizing my exposure to the sun at mid-day is effective in preventing skin cancer."

Message derogation. Message derogation was assessed by asking participants if they perceived that the message was "exaggerated," "overblown," and "overstated" (a = .86).

Perceived manipulation. To measure perceived manipulation, participants were asked if they thought that the message was "manipulative," "misleading," "distorted," and "exploitative" (a = .86).

Attitude. Attitude toward the recommended behavior was measured with a semantic-differential scale including two sets of word pairs: "undesirable--desirable" and "favorable--unfavorable" (a = .89). Attitude toward risky behavior (i.e., suntanning, sporting activities out in the sun, spending afternoons out in the sun, going to tanning parlors) was measured with two sets of word pairs: "undesirable--desirable" and "favorable--unfavorable" (a = .76).

Intentions. Intentions to engage in recommended behavior were measured with eight items including, "I intend to use sunscreen consistently when exposed to the sun" and "The next time I'm exposed to the sun, I intend to apply sunscreen with at least SPF-15 or higher to prevent skin cancer" (a = .90).

Behavior. Two weeks after the spring break at the follow-up, self-report sunscreen use behavior was measured with five items including: "Did you use sunscreen when exposed to the sun since you first participated in this study," "I have regularly used sunscreen when exposed to the sun to prevent skin cancer since I first participated in this study," and "I have applied sunscreen with at least a SPF-15 or higher when out in the sun to prevent skin cancer since I first participated in this study" (a = .92). The follow-up questionnaire also assessed the practice of skin self examination with one item ("I did skin self examination to prevent skin cancer").

Confound checks. Participants rated perceived accuracy, clarity, objectivity, quality, understandability, and amount of learning from the message (e.g., "The message was an accurate description of skin cancer").

Control variables. At the immediate post-test, extraneous variables that may have affected participants' responses to the messages in addition to threat and efficacy perceptions were measured to control for their influences. The twelve variables included: age, sex, race, skin type, prior risky behavior (i.e., frequency of outdoor leisure activities, frequency of going to a tanning parlor, frequency of sunburn resulting in peeling during the Last two years), prior preventive behavior (i.e., use of sunscreen, hat, and protective clothing; minimization of sun exposure at midday), and social influences (i.e., knowledge of other individuals who have skin cancer).

RESULTS

OVERVIEW

A 2 (threat) x 2 (efficacy) analysis of covariance (ANCOVA) method was performed to test the hypotheses. An ANCOVA method was chosen over analysis of variance in order to evaluate the clean effects of threat and efficacy while controlling for the possible influence of covariates. A total of seventeen covariates were entered in each run of ANCOVA: five significant message-confounds and twelve control variables (the results of covariates are available upon request). The unequal number of scores was adjusted for by employing a regression approach to ANCOVA, which gives equal weight to each cell mean regardless of its sample size (Tabachnick & Fidell, 1996, p. 344). A median split was done for the efficacy (4.67 on a 7-point scale).

MANIPULATION CHECKS

Threat. Participants who were exposed to the high threat message reported a greater level of susceptibility to skin cancer (M = 5.00) than those who were exposed to the low threat message (M = 3.59, t = -9.16, df= 272, p. < .001). Participants who were exposed to the high threat message perceived that skin cancer is a more severe disease (M = 5.86) than those who were exposed to the low threat message (M = 3.78, .t = -13.41, df = 272, p < .001).

Fear. Participants who were exposed to the high threat message perceived a greater level of fear (M = 4.23) than those who were exposed to the low threat message (M = 1.74, t = -17.22, df= 263, p < .001). Therefore, the results of susceptibility, severity, and fear-manipulation checks indicate that the high and low threat messages produced significantly different perceptions of threat and fear.

Confound checks. No difference was found between high and low threat-condition participants in terms of their perceived objectivity of the messages. However, high threat-condition participants indicated higher scores than low threat-condition participants in terms of perceived accuracy, clarity, quality, understandability, and amount of learning from the message (all ps < .001). To control for their possible influence on the dependent variables, the five significant confounds were controlled in the ANCOVA method.

MESSAGE DEROGATION

A significant interaction between threat and efficacy was found for message derogation, F (1,257) = 6.51, p = .01. Participants who reviewed the high threat message indicated a greater likelihood of message derogation when their perceived efficacy was low (M = 3.55) rather than high (M = 2.90).

PERCEIVED MANIPULATION

No significant main or interaction effects were detected. However, the pattern of condition means were as expected: Participants who reviewed the high threat message indicated a greater likelihood of message derogation when their perceived efficacy was low (M = 3.09) rather than high (M = 2.69).

ATTITUDE

For attitude toward recommended behavior, a significant main effect of efficacy (F (1,258) = 9.67, p = .002) was found. Participants with high efficacy perceptions indicated a more favorable attitude toward recommended behavior (M = 4.78) than those with low efficacy perceptions (M = 4.30). For attitude toward risky behavior, no significant main or interaction effects were discovered. Participants who reviewed the high threat message showed similar attitudes toward risky behavior whether their perceived efficacy was high (M = 3.11) or low (M = 3.05).

INTENTIONS

Significant main effects of threat (F (1,255) = 12.88, p < .001) and efficacy (F (1,255) = 16.99, p < .001) were found for intentions to engage in recommended behavior. Participants who reviewed the high (M = 4.92) rather than low threat message (M = 3.46) indicated more intentions to engage in recommended behavior. Participants who reviewed the threat messages with high (M = 3.95) rather than low efficacy perceptions (M = 3.43) indicated more intentions to engage in recommended behavior. Also, the interaction between threat and efficacy almost reached significance (F (2, 255) = 3.43, p = .06). After reviewing the high threat message, participants with high (M = 4.29) rather than low efficacy perceptions (M = 3.54) indicated more intentions to engage in recommended behavior.

BEHAVIOR

Two weeks after the spring break at the follow-up, a significant main effect of threat was found across the recommended behavior of sunscreen use (F (1, 225) = 11.03, p = .001) and skin self examination (F (1,226) = 9.54, p = .002). More sunscreen use was reported by those who had reviewed the high threat message (M = 3.41) than those who had reviewed the low threat message (M = 2.66). Also, more practice of skin self examination was reported by those who had reviewed the high threat message (M = 3.01) than those who reviewed the low threat message (M = 2.27).

DISCUSSION

The positive and negative effects of health risk messages found in this study provide a guideline that college health educators can use in designing interventions for skin cancer prevention. First, the results suggest that using the threat (susceptibility, severity) component of the EPPM can be an effective way of inducing protection motivation. According to previous research, Midwestern youth tend to believe that while skin cancer can be deadly, the risk of skin cancer is irrelevant to them (Banks, Silverman, Schwartz, & Tunnessen, 1992). To address this problem of perceived remoteness of risk, the high threat message of this study emphasized the more immediate effects of unprotected sun exposure: wrinkled, sagging, dry skin. Also, the high threat message emphasized that being in the Midwest does not necessarily protect the residents from skin cancer, even in snow or on cloudy days. The results indicate that the high threat message, which highlighted Midwestern college students' susceptibility to skin cancer, was more effective in motivating sun-protective behavior than the low threat message, which contained less susceptibility-related information. Therefore, effective risk messages for college students should present significant health risk from the perspectives to which college students can relate.

Second, the results suggest that the role of efficacy is critical in maximizing the positive effects of health risk messages while minimizing the potential of the negative effects. On the one hand, college students with high threat hut low efficacy perceptions discredited the message as overblown and an exaggeration of their actual risk. On the other hand, even though the message was highly threatening, message derogation was less likely when college students' efficacy perceptions were high. Similarly, after reviewing the high threat message, college students with high efficacy perceptions indicated more intentions to engage in skin cancer-preventive behavior than those with low efficacy perceptions. Also, attitude toward skin cancer preventive behavior was the function of efficacy perceptions, as college students with high efficacy perceptions expressed a more positive attitude than those with low efficacy perceptions. Hence, enhancing efficacy perceptions is a way to maximize the positive response to health risk messages and to motivate preventive behavior.

To enhance perceived efficacy concerning skin cancer prevention, interventions should address the barriers and doubts that college students may have. For example, college students may think that it is too much of a hassle to put on sunscreen, that the feeling of sunscreen is too greasy, that long-sleeve shirts or wide-brimmed hats do not make them look fashionable, or that periodic skin self-examination is difficult. Addressing these concerns and barriers would help enhance college students' efficacy perceptions concerning skin cancer-preventive behavior.

Contrary to hypothesis 2, the high threat message did not create a boomerang effect on college students' attitudes toward risky behavior. One possible reason for this finding is that taking precautionary measures before going out in the sun may be perceived as less of a threat to freedom than other preventive behavior such as practicing safe sex or reducing alcohol use, both of which require more efforts and social skills. As a result, college students with low efficacy perceptions complained that the messages on skin cancer risk were exaggerated and distorted, but they did not react further to form a more positive attitude toward risky behavior such as tanning. These results may suggest that the harmful effects of risk messages can be effectively reduced when behavioral recommendations are presented as less of a threat to freedom. Freedom-threatening messages gained more compliance when behavioral choices were provided than not (Heilman & Garner, 1975). Therefore, a way to control the boomerang effect may be to present behavioral recommendations in the context of choice, so that college students can perceive that they have freedom to choose among different options by which to engage in risk prevention. For example, instead of stating, "To reduce the risk of skin cancer, you should minimize your exposure to the sun at midday," risk messages may state, "Here are some of the ways that you can choose to reduce the risk of skin cancer. First, you may consider minimizing your exposure to the sun at midday."

Overall, this study provides useful guidelines for designing skin cancer-prevention messages for college students. The results suggest that fear-amusing health risk messages, if used with caution, can effectively motivate college students' sun-protective behavior. To be effective, messages should first present the significant consequences of risky behavior that are relevant to college students. Second, in order to enhance efficacy perceptions and minimize the potential of unintended effects, the significant and relevant threat should be accompanied by messages that specifically address the barriers to college students' skin cancer preventive behavior.

REFERENCES

Banks, B. A, Silverman, R. A., Schwartz, R. H., Tunnenssen, W. W. (1992). Attitudes of teenagers toward sun exposure and sunscreen use. Pediatrics, 89, 40-42.

Brehm, J.W. (1966). A theory of psychological reactance. New York: Academic Press.

Carmel, S. (1994). The role of age and an expanded Health Belief Model on predicting skin cancer protective behavior. Health Education Research, 9, 433-447.

Heilman, M.E., & Gardner, K.A. (1975). Counteracting the boomerang: The effects of choice on compliance to threats and promises. Journal of Personality and Social Psychology 31, 911-917.

Kleinot, M. C., & Rogers, R. W. (1982). Identifying effective components of alcohol misuse prevention programs. Journal of Studies on Alcohol, 43, 802-811.

Koh, H. K. (1991). Cutaneous melanoma. New England Journal of Medicine, 325, 171-182.

McClendon, B. T., Prentice-Dunn, S. (2001). Reducing skin cancer risk: An intervention based on protection motivation theory. Journal of Health Psychology 6, 321-328.

Michigan Department of Community Health (1996). Cancer incidence and mortality, Michigan 1994. Lansing, MI: Author.

Rippetoe, P.A., & Rogers, R. W. (1987). Effects of components of protection-motivation theory on adaptive and maladapdve coping with a health threat. Journal of Personality and Social Psychology, 52, 596-604.

Robinson, J.K., Rademaker, A. W., Sylvester, J.A., & Cook, B. (1997). Summer sun exposure: Knowledge, attitudes, and behaviors of midwest adolescents. Preventive Medicine, 26, 364-372.

Rogers, R. W., & Mewborn, C. R. (1976). Fear appeals and attitude change: Effects of a threat's noxiousness, probability of occurrence, and the efficacy of coping responses. Journal of Personality and Social Psychology 34, 54-61.

Tabachnick, G.B., & Fidell, L.S. (1996). Using multivariate statistics. New York: Harper Collins.

University data file (2002). Head counts by racial/ethnic Federal category, gender and student category.

Witte, K. (1992a). The role of threat and efficacy in MDS prevention. International Quarterly of Community Health Education, 12, 225-248.

Witte, K. (1992b). Putting the fear back into fear appeals: The extended parallel process model. Communication Monographs, 59, 330-349.

Witte, K. (1993). Message and conceptual confounds in fear appeals: The role of threat, fear, and efficacy. Southern Communication Journal, 58,147-155.

HEALTH EDUCATION RESPONSIBILITY AND COMPETENCY ADDRESSED

Responsibility IV--Evaluating Effectiveness of Health Education Programs

Competency D--Infer implications from findings for future program planning.

Sub-competency 2--Recommend strategies for implementing results of evaluation.

Hyunyi Cho, Ph.D. is an Assistant Professor in the Department of Marketing Communication at Emerson College. Address all correspondence to Hyunyi Cho, Ph.D. 120 Boylston Street, Boston, MA 02116, PHONE: [phone] FAX: [phone], E-MAIL: [e-mail]