Content-Type: text/html The Learned Helplessness Effect of Ineffective Recommendation in Threat Messages Introduction A threat message, traditionally labeled as a fear appeal message, is a "description of negative consequences that message receivers are likely to experience if they do not adopt the message recommendations" (Stiff, 1994, p. 122). Such a message is aimed at creating a perception in message receivers that they are likely to experience the described negative consequence if they do not adopt the recommended response in the message. The intuitive thinking is that the higher the magnitude of threat, the stronger the effect will be. However, empirical studies in the past few decades have produced seemingly contradictory findings. While many studies on threat messages have found a positive linear relationship between the level of threat and the attitudinal or behavioral changes (Higbee, 1969; Leventhal, 1970; Sutton, 1982; O'Keefe, 1990), others (Janis & Feshbach, 1953) found a negative relationship. Combining the positive and negative relationships, a curvilinear relationship appears. To explain these relationships, early researchers proposed the drive model (Janis, 1967) and the parallel response model (1970). Both models state that a threat message creates fear or emotional response, and that when the fear is high, people tend to be engaged in defensive maneuvers and discount the threat, minimize it or deny its existence. Thus, a high level of threat would be less effective. Recent followers have proposed an extended parallel process model (Witte, 1992), which differs little from the earlier one. Learned Helplessness The idea that the physiological response of fear has a role in people's processing of a threat message was met with criticism from its inception. Higbee (1969) and O'Keefe (1990) maintained that the physiological or emotional arousal of fear was irrelevant to attitudinal or behavioral changes, and that effects of a threat message should be explained by the cognitive appraisals people conduct after their exposure to the message. Rogers (1975, 1983) proposed the protection motivation theory which states that there are two types of cognitive appraisals people conduct when exposed to a threat message-the "threat appraisal" and the "coping appraisal" (1983, p. 168). These two appraisals are cognitive responses to the two components of a threat message-the negative consequence and a recommended response. Rogers (1983) speculated that if people are threatened but have no effective means of protecting themselves, persuasion effects are expected to be very low. This speculation has its empirical support in learned helplessness studies on both animals (Overmier & Seligman, 1967) and human beings (Hiroto, 1974; Matute, 1994). In formulating the theory of uncontrollability, Maier and Seligman (1976) pointed out that the learned helplessness effect, defined as a lack of behavioral responses of avoidance, stems from the perception of "uncontrollability of aversive events" (p. 33), and that the effect is not dependent upon the magnitude of an aversive stimulus, but upon the noncontingency between the response and the outcome. In other words, when a person or animal has learned that the outcome of eliminating the negative consequence is independent of responding, a perception of uncontrollability and, subsequently, a decrease in responses will occur. The perception of uncontrollability and response Learned Helplessness decrease are considered as two steps in the formation of learned helplessness (Maier & Seligman, 1976). The purpose of this study is to examine whether a lack of response-outcome contingency leads to perceived uncontrollability and a decrease of behavioral intention by dissecting a threat message into its two components-threat and recommendation. Specifically, this study manipulates threat and recommendation into high and low levels-high threat (serious gum diseases of gingivitis and periodontitis) versus low threat (unpleasant mouth odor), and effective recommendation (high success rate of using a fictitious brand of mouthrinse) versus ineffective recommendation (low success rate). The main and interaction effects on perceived personal efficacy and the behavioral intention will be measured. Hopefully, this study will contribute to the clarification of the seemingly contradictory relationships between a threat message and its persuasion effects, and provide theoretical guidance to communication practitioners in devising threat messages. Literature Review Negative, Positive and Curvilinear Relationships Janis and Feshbach (1953) conducted perhaps the first study on the effects of a threat message, then termed a fear appeal message, and found that it was the message with the minimal threat that produced "the greatest amount of conformity" (p. 84). The message with the strong threat was the least effective. Leventhal and Niles (1964) found Learned Helplessness the same-smokers in the high- and medium-fear conditions were less willing to stop smoking than in the low one. Janis (1967) proposed that a fear appeal message and its persuasion effects takes on a curvilinear relationship or "the form of an inverted U-shaped curve" (p. 181). He posited that a message with low threat tended to be dismissed "as inconsequential," while a message with high threat would produce "hypervigilant speculations and fantasies as well involuntary constriction of cognitive processes, resulting in marked interference with attention, comprehension, and learning" (p. 186). The optimal effect was from a message with moderate threat. However, no evidence has been found for this speculation. The majority of the studies on threat messages have found a positive linear relationship between the level of threat in a message and attitudinal or behavioral changes. Higbee (1969) noted that "most of the recent studies in the area have found a positive relationship between fear and persuasion" (p. 428). Leventhal (1970) agreed that "the overall picture presented by the data is for high fear messages to produce more persuasion than low fear messages" (p. 132). Sutton's (1982) summary of studies on effects of a threat message indicates that increases in threat are consistently associated with increases in acceptance. Boster and Mongeau's (1984) review concludes that there is no evidence of a curvilinear relationship, but a positive monotonic relationship. More recently, O'Keefe (1990) concluded that "increases in reported fear are reliably associated with increases in persuasive effectiveness" (p. 167). Empirical evidence seems in favor of this view (Quinn et al., 1992; Bennett, 1996) . Learned Helplessness Different Explanations The negative and curvilinear relationships are often explained by the drive model. In the seminal study (Janis & Feshbach, 1953), the authors explained that fear appeal messages produced "defensive avoidance arising from residual emotional tension" (p. 89). To reduce the tension, the recipients are be motivated to "ignore or to minimize the importance of the threat" (p. 90). Thus, the more severe the negative consequence is portrayed, the less effective the message is in inducing attitudinal or behavioral changes. The model predicts that, when confronted with high fear messages, people are likely to be engaged in "defensive maneuvers" or "defensive techniques," such as "avoiding the message, minimizing the severity of the threat, selectively attending the message, discounting the threat, and denying its personal relevance" (Keller & Block, 1996, p. 449). However, the drive model has little empirical support. Even in their own study (Janis & Feshbach, 1953), the authors produced no evidence of defensive avoidance. The data in their study clearly indicate that the stronger the threat, the more attention the subjects paid to the message and the less "mind-wandering" they experienced (p. 83). The messages of all three levels are equally effective in teaching the factual information. Admitting that "there was no evidence that subjects exposed to high fear messages were inattentive and failed to learn the recommendations" (p. 129), Leventhal (1970) proposed the parallel response model which states that, besides the emotional arousal leading to varied levels of effects, there is a parallel process which is cognitive. The emotional and cognitive processes are believed to be "contemporary" to each other (p. 124). The Learned Helplessness cognitive process is termed "danger control" (p. 126), while the emotional process is termed "fear control" (p. 126). He explained that "immediately after exposure to highly threatening warnings, fear will be strong and may motivate avoidance behaviors that could disrupt danger control and create resistance to persuasion" (p. 126). Therefore, the parallel response model is not much different from the drive model. Witte (1992) proposed an Extended Parallel Process Model (EPPM), differing little from the Leventhal's model. From the early period of fear appeal research, some researchers began to object to the notion of emotional arousal as the cause of attitude changes. Evidently, fear is a physiological response, not a cause. Higbee (1969) argued that "the arousal of fear" is "irrelevant" to effects of persuasion (p. 428) and that the fear measured by many researchers was not fear, but thoughts or cognitive responses to threat messages. For example, Janis and Feshbach (1953) measured fear by asking subjects how much they were worried or concerned. Clearly, it was the cognitive process, not the physiological arousal, that was examined in this study. O'Keefe (1990) argued that "the cognitive changes-not the emotional ones-might actually explain the message's effectiveness" (p. 167), and that "a purely cognitive explanation might be sufficient" (p. 168). Higbee (1969) speculated that an extremely high threat might be perceived to be unlikely to happen and it was the interaction of these two factors, se verity and probability, that might lead to "a curvilinear relationship between fear level and persuasion" (p. 441). The comprehensive theoretical guidance in fear appeal research was provided by Rogers (1975, 1983), who proposed the protection motivation theory stating that people Learned Helplessness are engaged in a number of cognitive processes when confronted with a threat message. These processes are perceptions of severity of the negative consequence, the probability, susceptibility or vulnerability, efficacy of the recommended response, and their personal capability in adopting the recommended response. The protection motivation theory has incorporated a number of important concepts from the field of psychology. The concept of perceived severity of the negative consequence corresponds to Fishbein and Ajzen's (1975) expectancy-value model which deals with beliefs of the positive attributes of a stimulus. The concept of perceived probability is based on Rotter's social learning theory that an individual's behavior is partly determined by an individual's belief of the likelihood of reinforcement (Rotter et al., 1972). The concept of perceived susceptibility is based on the concept of personal relevance or involvement (Allport, 1943; Festinger, 1957; Sherif & Hovland, 1961). The perceived response efficacy and perceived personal efficacy are the other two components in the protection motivation theory. The two components correspond to the concepts of "outcome expectation" and "efficacy expectation" in Bandura's social learning theory (1977, p. 79). Rogers (1983) believed that these perceptions constitute the motivation to protect oneself from danger. The motivation is a positive linear function of these perceptions. He emphasized that "attitude change is not mediated by or a result of an emotional state of fear, but rather is a function of the amount of protective motivation aroused by the cognitive appraisal processes" (Rogers, 1983, p. 158). Learned Helplessness Research Question and Hypotheses: In his protection motivation theory, Rogers (1983) combined five perceptions into two types of appraisals. The "threat appraisal" consists of perceptions of severity, probability and susceptibility, and the "coping appraisal" consists of perceived response efficacy and perceived personal efficacy (p. 168). He predicted a significant interaction between the threat appraisal and the coping appraisal. High threat combined with low efficacy would result in a boomerang effect: If people are threatened by a severe negative consequence but see no effective recommendation, their behavioral intention will drop. Beck and Frankel (1981) also speculated that the negative relationship between threat and persuasion in Janis and Feshbach's study (1953) might be "the result of a lack of perceived threat control" (p. 211). These speculations may find their support in one learning theory-the theory of uncontrollability or the learned helplessness hypothesis. The learned helplessness effect is defined as a lack of avoidance or escape behavior "resulting from the uncontrollability of aversive events" (Maier & Seligman, 1976, p. 33). Early experiments (Overmier & Seligman, 1967; Seligman & Maier, 1967) found that dogs that had been exposed to inescapable electric shocks during treatment were unable to initiate behavioral responses to escapable shocks. Seligman and Maier (1967) concluded that the dogs "learned as a consequence of inescapable shock that its responding was independent of shock termination, and therefore the probability of response initiation during shock decreased" (p. 8). The same effect has been observed in human beings who are exposed to inescapable noise or unsovable problems (Hiroto, 1974; Hiroto & Seligman, 1975; Witkowski, 1997). Maier and Seligman (1976) Learned Helplessness explained that "like the dog, cat, rat, and fish, when a human being is faced with noxious events that it cannot control, its motivation to respond seems to be reduced" (p. 9). The lack of behavioral responses is attributed to the perceived "uncontrollability" or a lack of "perception of control" (Maier & Seligman, 1976, p. 13). Maier and Seligman (1976) explained that "men and animals actively learn that responses and outcomes are independent of each other" (p. 14). When human beings or animals have learned that the outcome is independent of responding, they stop responding. Briefly, "a belief in uncontrollability undermines the incentive to initiate responses" (Maier & Seligman, 1976, p. 18). The concept of perceived controllability in learned helplessness hypothesis is renamed perceived self-efficacy in Bandura's social learning theory (Bandura, 1977, 1990). It is widely believed that perceived self-efficacy, or perceived personal capability in controlling the environment, is "the most powerful influence on both the initiation of a behavior and persistence in the face of frustration or failure" (Maddux & Stanley, p. 250). On the other hand, people "see little point to even trying if they believe they cannot exercise control over their own behavior" (Bandura, 1990, p. 11). It is also an essential part of the theory of uncontrollability that learned helplessness effect is not dependent upon the magnitude of an aversive stimulus, but upon the nature of inescapability. Experiments showed that the learned helplessness effect is "essentially unaffected by the use of an increased level of shock" (Overmier & Seligman, 1967, p. 31). In the words of Maier and Seligman (1976), "learning that shock is uncontrollable, and not shock per se, causes helplessness" in dogs (p. 6). Experiments on Learned Helplessness human beings (Matute, 1994) provided the same conclusion that it is the "perception of uncontrollability" that determines the learned helplessness effect, not the negative consequences (p. 229). Furthermore, the learned helplessness effect is divided into two stages-The first stage is the formation of the perception that "outcomes are uncontrollable" and the second stage is the decrease in behavioral responses (Maier & Seligman, 1976, p. 18). In mass communication, a threat message has two components: negative consequence and recommendation. The recommendation is the suggested response to eliminate the negative consequence. If the recommendation is ineffective, the response is independent of the outcome of eliminating the negative consequence. Thus, the perception of uncontrollability is likely to occur and subsequently a reduction in the behavioral intention. Therefore, this study hypothesizes: Hypothesis One: The ineffective recommendation, rather than the low threat, will reduce perceived personal efficacy. Hypothesis Two: The ineffective recommendation, rather than the low threat, will reduce the behavioral intention. As the level of threat is not relevant to inducing perceived personal efficacy or the behavioral intention, this study also hypothesizes: Hypothesis Three: The high threat followed by the ineffective recommendation will be less effective in inducing perceived personal efficacy and the behavioral intention than the low threat followed by the effective recommendation. Learned Helplessness Methods Subjects 154 undergraduate students from a mid-western university participated in the experiment for extra points toward their grades. The subjects were recruited from an introductory class of mass communication. Most of them are freshmen and have not participated in an experiment before. Experimental Design This study used a two-factor design. The first factor, threat in a message, was manipulated into two levels: high and low. The second factor, the recommendation (recommended response to avoid the threat), was manipulated into two levels: effective and ineffective. The manipulation produced a total of four conditions: high-threat-effective-recommendation, high-threat-ineffective-recommendation, low-threat-effective-recommendation, and low-threat-ineffective-recommendation. Stimulus Materials Four short articles were devised for the four experimental groups. All articles have two elements: a description of the negative consequences of poor dental hygiene and the introduction of a fictitious antiseptic mouthrinse called Super Fresh. The high-threat-effective-recommendation article has 220 words on one page. It contains a description of the severe negative consequences of poor dental hygiene-gum diseases, gingivitis (the buildup of plaque, calculus or tartar, toxins released from bacteria between teeth, and unpleasant odor) and periodontitis (destroyed bones and ligaments supporting the teeth, and the detachment of teeth from the gums). The Learned Helplessness description is followed by the introduction of a new antiseptic mouthrinse, Super Fresh, which is clinically proven by the American Dental Association to be effective in killing germs between teeth-preventing/reducing plaque buildup by 98 percent, tartar by 95 percent, and unpleasant odor (bad breath) by almost 100 percent. The high-threat-and-ineffective-recommendation article is of the same length and contains the same description of the severe negative consequences, gingivitis and periodontitis, followed by the introduction of the same product, a new antiseptic mouthrinse, Super Fresh. But the article says that the product is proven by ADA to be ineffective in killing germs between teeth-preventing/reducing plaque buildup by less than 15 percent, tartar by less than 10 percent, and unpleasant odor (bad breath) by less than 20 percent. The low-threat-effective-recommendation article has 92 words on one page. It briefly describes that bad breath (unpleasant odor) occurs as a result of the accumulation of food debris between teeth. This is followed by the introduction of a new antiseptic mouthrinse, Super Fresh, which is clinically proven by ADA to be effective in killing germs between teeth-reducing unpleasant odor by almost 100 percent. The low-threat-ineffective-recommendation article is of the same length and contains the same description of bad breath, followed by the introduction of the same product, which is clinically proven by ADA to be ineffective-reducing unpleasant odor by less than 20 percent. Learned Helplessness Dependent Variables and Measurement As the learned helplessness hypothesis or theory of uncontrollability predicts, the immediate result of being exposed to an inescapable threat is the lack of perceived control Therefore, the first dependent variable is personal efficacy, which was defined by Bandura as an individual's belief in his capability of adopting the recommendation in avoiding the negative consequence. According to the learned helpless hypothesis and Bandura, the lack of personal efficacy leads to reduction in the intention to respond. Therefore, behavioral intention is the next dependent variable. As manipulation checks, perceived severity of the threat and perceived response efficacy are also measured. For each of the dependent variables, one question is asked, followed by three 1-to-7 semantic differential scales. The questions are: For severity, how serious do you think gingivitis (bad breath) is? (not at all serious/very serious, not at all critical/very critical, not at all grave/very grave); for response efficacy, how effective do you think Super Fresh will be in preventing gingivitis (bad breath)? (not at all effective/very effective, not at all capable/very capable, not at all successful/very successful); for personal efficacy, how capable do you think you yourself will be in using Super Fresh to prevent gingivitis (bad breath)? (not at all capable/very capable, not at all effective/very effective, not at all successful/very successful), and for the behavioral intention, how likely will you use Super Fresh to prevent gingivitis (bad breath)? (not at all likely/very likely, not at all interested/very interested, not at all willing/very willing). To avoid the potential resp onse habit of selecting one extreme on the continuum, the anchoring semantic terms, "very" and "not at all", are reversed for half of the scales. Learned Helplessness Procedure The experiment was conducted at the end of a class session. The subjects were instructed that they were to evaluate a piece of writing. The articles corresponding to the number of subjects participating in the experiment were randomly distributed among the subjects. Each subject received one article, read it for three to five minutes and filled out a questionnaire. The whole experiment took about 15 minutes. At the end of the experiment, the subjects were thanked for their participation. Results Main Effects This study has a 2 x 2 factorial design, resulting in four experimental groups: high-threat-effective-recommendation (N = 41), high-threat-ineffective-recommendation (N = 38), low-threat-effective-recommendation (N = 37), and low-threat-ineffective-recommendation (N = 38). The unequal sizes of the groups were due to unusable or unreturned questionnaires. ANOVA results (see Table 3) indicate that there are significant differences among the four groups on all dependent variables: perceived severity [F (3, 150) = 5.88, p = .001], perceived response efficacy [F (3, 150) = 40.06, p = .000], perceived personal efficacy [F (3, 150) = 11.14, p = .000], and behavioral intention [F (3, 150) = 18.61, p = .000]. Significant differences are also found between the marginal means of high and low threats, and between the marginal means of effective and ineffective recommendations (see details in Tests of Hypotheses below). However, no significant Learned Helplessness interactions are found between threat and recommendation on any of the dependent variables. Manipulation Checks Two factors are manipulated in this study. The factor of threat is manipulated into high (N = 79) and low (N = 75), while the factor of recommendation is manipulated into effective (N = 78) and ineffective (N = 76). The dependent variables of perceived severity of the threat and perceived response efficacy are used as manipulation checks for threat and recommendation respectively. ANOVA results show that the mean of perceived severity for the low threat subjects (M = 3.65, N = 75) is 3.65 is lower than the mean for high threat (M = 4.50, N = 79) is 4.50. The difference is significant [F (1, 152) = 16.86, p = .000] (see Table 2 for means and Table 3 for significance tests). The mean of perceived response efficacy for the ineffective recommendation subjects (M = 2.86, N = 76) is lower than the mean for effective recommendation (M = 5.02, N = 78). The difference is significant [F (1, 152) = 117.74, p = .000] (see Table 1 for means and Table 3 for significance tests). Therefore, both manipulations are successful. Reliability Checks Since this study used multi-item measurements for the dependent variables, a reliability check was run and the reliability coefficients are: perceived severity (Alpha = .8238), perceived response efficacy (Alpha = .9658), perceived personal efficacy (Alpha = .9285), and behavioral intention (Alpha = .9197). Therefore, the measurements have acceptable reliability. Learned Helplessness Tests of Hypotheses Hypothesis One states that the ineffective recommendation, rather the low threat, will reduce perceived personal efficacy. This is supported by data of both main effects and specific contrasts. The main effect of recommendation on perceived personal efficacy (see Table 1 for means and Table 3 for significance tests; Figure 1) show that the mean for the subjects receiving the effective recommendation (N = 76) is 4.84, while the mean for the subjects receiving the ineffective recommendation (N = 78) is lower at 3.36. The difference is significant [F (1, 152) = 33.29, p = .000]. On the contrary, data of the main effect of threat indicate that there is no significant difference between the low threat (M = 4.16, N = 75) and the high threat (M = 4.06, N = 79) on perceived personal efficacy [F (1, 152) = .272, p = .602] (see Table 2 for means and Table 3 for significance tests). Figure 1. Main Effect of Recommendation Figure 2. Main Effect of Recommendation Dependent Var': Personal Efficacy Dependent Var': Behavioral Intention 6 6 5 5 4 4 3 3 Ineffective R' Effective R' Ineffective R' Effective R' Data of specific contrasts (see Table 4 for means of specific groups and Table 5 for significance tests) show that the ineffective recommendation reduces perceived Learned Helplessness personal efficacy in both the high threat situation (3.31 < 4.76, p = .000) and the low threat situation (3.41 < 4.93, p = .000). On the contrary, no significant difference exists between the low threat and the high threat in either the ineffective recommendation situation or the effective recommendation situation. Hypothesis Two states that the ineffective recommendation, rather than the low threat, will reduce the behavioral intention. This is supported by data of both main effects and specific contrasts. The main effect of recommendation (see Table 1 for means and Table 3 for significance tests; Figure 2) indicates that the mean of the behavioral intention for the subjects receiving the effective recommendation (N = 76) is 4.04, while the mean for the subjects receiving the ineffective recommendation (N = 78) is much lower at 2.32. The difference is significant [F (1, 152) = 55.32, p = .000]. On the contrary, the main effect of threat indicates that the difference between the low threat (M = 3.17, N = 75) and the high threat (M = 3.25, N = 79) on the behavioral intention is not significant [F (1, 152) = .034, p = .854] (see Table 2 for means and Table 3 for significance tests). Data of the specific contrasts also support the hypothesis (see Table 4 for means of specific groups and Table 5 for significance tests). In both the high threat and low threat situations, the ineffective recommendation reduces the behavioral intention. In the high threat situation, the mean for the ineffective recommendation (M = 2.29, N = 38) is significantly lower (p = .000) than that of the effective recommendation (M = 4.13, N = 41). In the low threat situation, the mean for the ineffective recommendation (M = 2.34, N = 38) is also significantly lower (p = .000) than that of the effective recommendation Learned Helplessness (M = 3.99, N = 37). On the contrary, no significant difference exists between the low threat and the high threat in either the ineffective recommendation situation or the effective recommendation situation. Data of correlation among variables also support the hypothesis. The bivariate correlation between threat and the behavioral intention (r = .077) is negligible, but perceived response efficacy is highly correlated with the behavioral intention (r = .632, p < .001). Hypothesis Three states that the high threat followed by the ineffective recommendation will be less effective in inducing perceived personal efficacy and the behavioral intention than the low threat followed by the effective recommendation. This is supported. Figure 3. Specific Contrasts of Four Groups Effective Recommendation: ___________ Ineffective Recommendation: - - - - - - - - - - Dep' Var': Perceived Personal Efficacy Dep" Var': Behavioral Intention 6 6 5 5 4 4 3 3 High Threat Low Threat High Threat Low Threat Data of the specific contrasts among the four experimental groups (see Table 4 for means and Table 5 for significance tests; Figure 3) reveal that the mean of perceived Learned Helplessness personal efficacy for the high-threat-ineffective-recommendation subjects (M = 3.31, N = 38) is significantly lower (p = .000) than the mean for the low-threat-effective-recommendation subjects (M = 4.93, N = 37). When the high threat is combined with the ineffective recommendation, it also produces less behavioral intention than the low threat combined with the effective recommendation, creating a negative relationship between a threat and the behavioral intention. The mean of behavioral intention for the high-threat-ineffective-recommendation subjects (M = 2.29, N = 38) is significantly lower (p = .000) than the mean for the low-threat-effective recommendation subjects (M = 3.99, N = 37). Conclusion and Discussion This study examines the roles of the two components in a threat message: threat and recommendation. Applying the theory of uncontrollability (Maier & Seligman, 1976), this study has found that it is the recommendation component, rather than the threat component, that determines the effects of a threat message. Specifically, by manipulating the threat into high and low levels, and the recommendation into effective and ineffective types, this study has found that it is the ineffective recommendation, rather the low threat, that reduces perceived personal efficacy and the behavioral intention. This study has also found that the high threat followed by the ineffective recommendation is less effective than the low threat followed by the effective recommendation, producing a negative relationship between a threat message and its effects. Learned Helplessness This study has provided empirical evidence for the speculation that the negative relationship is a result of "a lack of perceived threat control" (Beck & Frankel, 1981, p. 211), and for the notion of "boomerang effect" (Rogers, 1983, p. 170). This study indicates that the negative relationship is attributed to a cognitive process in which people learn and establish that there is a lack of contingency between the recommended response and the outcome of avoiding the negative consequence. People's knowledge of the response-outcome independence leads to their perception of uncontrollability (Maier & Seligman, 1976, p. 33), or reduced perceived self-efficacy (Bandura, 1977), and subsequently, a reduction in the behavioral intention. By establishing the causal relationship between ineffective recommendation and the learned helplessness effect-a reduction in perceived personal efficacy and the behavioral intention, this study indicates that the recommendation component in a threat message plays a crucial role, while the threat component is inconsequential. The results of this study also imply that fear, a physiological response caused by the threat component, has little impact on people's behavioral intention. This is contrary to the belief that the fear response has a significant role in people's processing of threat messages (Witte, 1992; LaTour & Rotfeld, 1997). The implication for communication practitioners is significant. A message merely containing a threat component is unlikely to cause behavioral changes. An effective recommendation must be available to induce the perception of controllability. Furthermore, raising the level of threat in the absence of an effective recommendation Learned Helplessness will only weaken the response-outcome contingency and diminish the behavioral intention. Learned helplessness is a complex phenomenon. It may be "general," "specific," "chronic" or "acute" (Abramson et al., 1978). According to the revised theory of learned helplessness, human beings, when exposed to inescapable aversive stimuli, make attributions, either internal or external. These attributions may moderate the learned helplessness effect. Future communication studies may attempt to find its applicability in the mass communication setting. References Abramson, L., Seligman, M., & Teasdale, J. (1978). Learned helplessness in humans: Critique and reformulation. Journal of Abnormal Psychology, 87, 49-74. Bandura, A. (1977). Self-efficacy: Toward a unifying theory of behavioral change. Psychological Review, 84, 191-215. Bandura, A. (1977). Social learning theory. Englewood Cliffs, NJ: Prentice-Hall. Bandura, A. (1982). Self-efficacy mechanism in human agency. American Psychologist, 37, 122-147. Bandura, A. (1990). Perceived self-efficacy in the exercise of control over AIDS infection. Evaluation and Program Planning, 13, 9-17. Beck, K., & Frankel, A. (1981). A conceptualization of threat communications and protective health behavior. Social Psychology Quarterly, 44, 204-217. Bennett, R. (1996). Effects of horrific fear appeals on public attitudes towards AIDS. International Journal of Advertising, 15, 183-202. Boster, F., & Mongeau, P. (1984). Fear-arousing persuasive messages. In R. Bostrom (Ed.), Communication yearbook 8 (pp. 330-375). Beverly Hills, CA: Sage. Chaiken, S. (1980). Heuristic versus systematic information processing and the use of source versus message cues in persuasion. Journal of Personality and Social Psychology, 39 (5), 752-766. Dabbs, J., & Leventhal, H. (1966). Effects of varying the recommendations in a fear-arousing communications. Journal of Personality and Social Psychology, 4, 525-531. Fishbein, M., & Ajzen, I. (1975). Beliefs, attitude, intention, and behavior: An introduction to theory and research. Reading, MA: Addison-Wesley. Higbee, K. (1969). Fifteen years of fear arousal: Research on threat appeals: 1953-1968. Psychological Bulletin, 72, 426-444. Hiroto, D. (1974). Locus of control and learned helplessness. Journal of Experimental Psychology, 102, 187-193. Hiroto, D., & Seligman, M. (1975). Generality of learned helplessness in man. Journal of Personality and Social Psychology, 31, 311-327. Insko, C., Arkoff, A., & Insko, V. (1965). Effects of high low fear arousing communication upon opinions toward smoking. Journal of Experimental and Social Psychology, 1, 256-266. Janis, I. (1967). Effects of fear arousal on attitude change: Recent developments in theory and research. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 3, pp. 166-224). New York: Academic Press. Janis, I., & Feshbach, S. (1953). Effects of fear-arousing communications. Journal of Abnormal and Social Psychology, 48, 78-92. Keller, P., & Block, L. (1996). Increasing the persuasiveness of fear appeals: The effect of arousal and elaboration. Journal of Consumer Research, 22, 448-459. LaTour, M., & Rotfeld, H. (1997). There are threats and (maybe) fear-caused arousal: Theory and confusion of appeals to fear and fear arousal itself. Journal of Advertising, 26, 45-59. Leventhal, H. (1970). Findings and theory in the study of fear communications. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 5, pp. 119-186). New York: Academic Press. Leventhal, H., & Niles, P. (1964). A field experiment on fear-arousal with data on the validity of questionnaire measures. Journal of Personality, 32, 459-479. Leventhal, H., & Watts, J. (1966). Sources of resistance to fear-arousing communications on smoking and lung cancer. Journal of Personality, 34, 155-175. Maddux, J., & Stanley, M. (1986). Self-efficacy theory in contemporary psychology: An overview. Journal of Social and Clinical Psychology, 4, 249-255. Maier, S., & Seligman, M. (1976). Learned helplessness: Theory and evidence. Journal of Experimental Psychology: General, 105, 3-46. Matute, H. (1994). Learned helplessness and superstitious behavior as opposite effects of uncontrollable reinforcement in humans. Learning and Motivation, 25, 216-232. Mikulincer, M. (1994). Human learned helplessness: A coping perspective. New York: Plenum Press. Miller, G. (1963). Studies in the use of fear appeals: A summary and analysis. Central States Speech Journal, 14, 117-125. Miller, G. (1980). On being persuaded: Some basic definitions. In M. Roloff & G. Miller (Eds.), Persuasion: New directions in theory and research (pp. 12-15). Beverly Hills, CA: Sage. O'Keefe, D. (1990). Persuasion: Theory and research. Newbury Park, CA: Sage. Overmier, J., & Seligman, M. (1967). Effects of inescapable shock upon subsequent escape and avoidance responding. Journal of Comparative and Physiological Psychology, 63, 28-33. Petty, R., & Cacioppo, J. (1986). Communication and persuasion: Central and peripheral routes to attitude change. New York: Springer-Verlag. Quinn, V., Meenaghan, T., & Brannick, T. (1992). Fear appeals: Segmentation is the way to go. International Journal of Advertising, 11, 355-366. Rogers, R. (1975). A protection motivation theory of fear appeals and attitude change. Journal of Psychology, 91, 93-114. Rogers, R. (1983). Cognitive and physiological processes in fear appeals and attitude change: A revised theory of protection motivation. In J. Cacioppo & R. Petty (Eds.), Social psychophysiology (pp. 153-176). New York: Guilford Press. Rogers, R. (1985). Attitude change and information integration in fear appeals. Psychological Reports, 56, 179-182. Rogers, R., & Deckner, C. (1975). Effects of fear appeals and physiological arousal upon emotion, attitudes, and cigarette smoking. Journal of Personality and Social Psychology, 32, 222-230. Rogers, R., & Mewborn, C. (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. Rotter, J., Chance, J., & Phares, E. (1972). Applications of a social learning theory of personality. New York: Holt, Rinehart and Winston. Seligman, M., & Maier, S. (1967). Failure to escape traumatic shock. Journal of Experimental Psychology, 74, 1-9. Stiff, J. (1994). Persuasive communication. New York: The Guilford Press. Sutton, S. (1982). Fear-arousing communications: A critical examination of theory and research. In J. Eiser (Ed.), Social psychology and behavioral medicine (pp. 303-337). London: Wiley. Sutton, S., & Eiser, J. (1984). The effect of fear-arousing communications on cigarette smoking: An expectancy-value approach. Journal of Behavioral Medicine, 7, 13-33. Sutton, S., & Hallett, R. (1988). Understanding the effects of fear-arousing communications: The role of cognitive factors and amount of fear aroused. Journal of Behavioral Medicine, 11, 353-360. Tanner, J., Hunt, J., & Eppright, D. (1991). The protection motivation model: A normative model of fear appeals. Journal of Marketing, 55, 36-45. Witkowski, T. (1997). Performance level in situations of helplessness threat and group affiliation: Egotistic mechanisms in helplessness deficits. Journal of Social Psychology, 137, 229-234. Witte, K. (1992). Putting the fear back into fear appeals: The extended parallel process model. Communication Monographs, 59, 329-349. Table 1 Means of Ineffective (N = 76) and Effective (N = 78) Recommendations Dependent Variable Condition Mean perceived response efficacy ineffective recom 2.85967 effective recom 5.01709 perceived severity ineffective recom 4.17104 effective recom 4.00000 perceived personal efficacy ineffective recom 3.35965 effective recom 4.84188 behavioral intention ineffective recom 2.31579 effective recom 4.06410 Table 2 Means of Low (N = 75) and High (N = 79) Threats Dependent Variable Condition Mean perceived severity low threat 3.64888 high threat 4.49789 perceived response efficacy low threat 4.08000 high threat 3.83123 perceived personal efficacy low threat 4.15999 high threat 4.06330 behavioral intention low threat 3.15556 high threat 3.24473 Table 3 Tests of Main Effects and Interaction 1. Dependent Variable: perceived severity Source Type III Sum of Squares df Mean Square F Sig. FOUR GROUPS 29.281 3 9.760 5.878 .001 THREAT 27.992 1 27.992 16.858 .000 RECOMMEN 1.452 1 1.452 .874 .351 THREAT * RECOMMEN .117 1 .117 .070 .791 Error 249.066 150 1.660 Total 2847.442 154 2. Dependent Variable: perceived response efficacy Source Type III Sum of Squares df Mean Square F Sig. FOUR GROUPS 183.404 3 61.135 40.060 .000 THREAT 3.597 1 3.597 2.357 .127 RECOMMEN 179.671 1 179.671 117.735 .000 THREAT * RECOMMEN .677 1 .677 .444 .506 Error 228.910 150 1.526 Total 2818.002 154 3. Dependent Variable: perceived personal efficacy Source Type III Sum of Squares df Mean Square F Sig. FOUR GROUPS 85.302 3 28.434 11.144 .000 THREAT .695 1 .695 .272 .602 RECOMMEN 84.940 1 84.940 33.291 .000 THREAT * RECOMMEN 3.282E-02 1 3.282E-02 .013 .910 Error 382.710 150 2.551 Total 3069.889 154 4. Dependent Variable: behavioral intention Source Type III Sum of Squares df Mean Square F Sig. FOUR GROUPS 118.088 3 39.363 18.606 .000 THREAT 7.184E-02 1 7.184E-02 .034 .854 RECOMMEN 117.027 1 117.027 55.317 .000 THREAT * RECOMMEN .353 1 .353 .167 .683 Error 317.337 150 2.116 Total 2013.666 154 Table 4 Mean of Specific Groups Dependent Variable Condition N Mean perceived severity high threat effective recom 41 4.43091 high threat ineffective recom 38 4.57016 low threat effective recom 37 3.52251 low threat ineffective recom 38 3.77193 perceived response efficacy high threat effective recom 41 4.93494 high threat ineffective recom 38 2.64038 low threat effective recom 37 5.10811 low threat ineffective recom 38 3.07896 perceived personal efficacy high threat effective recom 41 4.76424 high threat ineffective recom 38 3.30703 low threat effective recom 37 4.92792 low threat ineffective recom 38 3.41227 behavioral intention high threat effective recom 41 4.13008 high threat ineffective recom 38 2.28947 low threat effective recom 37 3.99099 low threat ineffective recom 38 2.34211 Table 5 Multiple Comparisons (Bonferroni t-test) of Means of Specific Groups Dependent Variable (I) groups (J) groups Mean Difference (I-J) Sig. perceived severity high threat effective recom high threat ineffective recom -.13925 1.000 low threat effective recom .90840 .013 low threat ineffective recom .65898 .147 high threat ineffective recom High threat effective recom .13925 1.000 low threat effective recom 1.04764 .003 low threat ineffective recom .79823 .046 low threat effective recom High threat effective recom -.90840 .013 high threat ineffective recom -1.04764 .003 low threat ineffective recom -.24942 1.000 low threat ineffective recom High threat effective recom -.65898 .147 high threat ineffective recom -.79823 .046 low threat effective recom .24942 1.000 perceived response efficacy high threat effective recom high threat ineffective recom 2.29457 .000 low threat effective recom -.17317 1.000 low threat ineffective recom 1.85599 .000 high threat ineffective recom high threat effective recom -2.29457 .000 low threat effective recom -2.46773 .000 low threat ineffective recom -.43858 .743 low threat effective recom high threat effective recom .17317 1.000 high threat ineffective recom 2.46773 .000 low threat ineffective recom 2.02915 .000 low threat ineffective recom high threat effective recom -1.85599 .000 high threat ineffective recom .43858 .743 low threat effective recom -2.02915 .000 perceived personal efficacy high threat effective recom high threat ineffective recom 1.45721 .000 low threat effective recom -.16368 1.000 low threat ineffective recom 1.35196 .001 high threat ineffective recom high threat effective recom -1.45721 .000 low threat effective recom -1.62089 .000 low threat ineffective recom -.10525 1.000 low threat effective recom high threat effective recom .16368 1.000 high threat ineffective recom 1.62089 .000 low threat ineffective recom 1.51565 .000 low threat ineffective recom high threat effective recom -1.35196 .001 high threat ineffective recom .10525 1.000 low threat effective recom -1.51565 .000 behavioral intention high threat effective recom high threat ineffective recom 1.84061 .000 low threat effective recom .13909 1.000 low threat ineffective recom 1.78798 .000 high threat ineffective recom high threat effective recom -1.84061 .000 low threat effective recom -1.70152 .000 low threat ineffective recom -5.26316E-02 1.000 low threat effective recom high threat effective recom -.13909 1.000 high threat ineffective recom 1.70152 .000 low threat ineffective recom 1.64889 .000 low threat ineffective recom high threat effective recom -1.78798 .000 high threat ineffective recom 5.2632E-02 1.000 low threat effective recom -1.64889 .000 Table 6 Bivariate Pearson Correlation (N = 154) BI Severity Res' Efficacy Severity .077 Response Efficacy .632** -.074 Personal Efficacy .530** .003 .534** **. Correlation is significant at the 0.01 level (2-tailed). *. Correlation is significant at the 0.05 level (2-tailed).