INTRODUCTION
Rogers' revised Protection Motivation Theory (PMT) (21) is a major health psychology theory aimed at explaining the cognitive mediation process of behavioral change in terms of threat and coping appraisal. The PMT's threat appraisal component is composed of the following: the person's estimate of the severity of the disease (perceived severity) and his or her estimate of the chance of contracting the disease (perceived vulnerability). The PMT further stipulates that the emotional state of fear arousal influences attitudes and behavior change indirectly through the appraisal of the severity of the danger. The model's coping appraisal consists of the individual's expectancy that carrying out recommendations can remove the threat (response efficacy) and belief in one's capability to execute the recommended course of action successfully (self-efficacy).
The PMT hypothesizes that the motivation to protect oneself from danger is a function of four cognitive beliefs. These are as follows: (i) the threat is severe; (ii) one is personally vulnerable to the threat; (iii) the coping response is effective in averting the threat; and (iv) one has the ability to perform the coping response. Protection motivation is the proximal determinant of protective behavior and often measured by or similar to intention (12,13). Thus, the cognitive predictors (severity, vulnerability, response efficacy, and self-efficacy) should have significant associations with intentions, which mediate their influence on behavior performance. Many studies have measured self-reported and/or observed behavior as the outcome variable of protection motivation (6,11).
Rogers' proposed full model (i.e., subtracting threat from extrinsic and intrinsic rewards and subtracting response costs from the coping appraisal) (21), however, is considered untestable (23). Indeed, no PMT study has attempted to test the full model in this way. Other aspects of the theory such as (i) the nature of the relationships between the cognitive mediators and (ii) the proposed additive principle (i.e., when combining components between the two appraisal processes, a second-order interaction effect should occur) have been considered by a number of writers to be unclear and internally inconsistent, respectively (7,13). Hence, most applications of the PMT consider only the main effects of perceptions of severity, vulnerability, response efficacy, and self-efficacy (12,13) (Fig. 1).
;)
FIGURE 1: Protection motivation theory (PMT) - conceptual model.
The PMT has been moderately successful in predicting health- and safety-related intentions and behaviors in a variety of contexts (6,11) such as smoking, alcohol consumption, and nutrition. In the two PMT meta-analyses, Floyd and colleagues (6) examined 65 studies representing more than 20 health issues (e.g., cancer prevention, AIDS prevention, adherence to medical-treatment regimes), whereas Milne and associates (11) used stricter inclusion criteria in which only empirical applications of the PMT to health-related intentions, concurrent behavior, or subsequent behavior were included. Across these studies, self-efficacy was found to be the strongest predictor of intention and behavior (6,11), and intention has been shown to be more highly correlated with behavior than any other variable from the PMT (11,12). The prediction of intention is reported to be better than behavior for current and subsequent behavior, and more variance could be explained cross-sectionally rather than longitudinally (11). Overall, the results of these studies show modest support for the threat- and coping-appraisal constructs of the model in predicting health-related intentions and behavior, with coping-appraisal components emerging as the strongest predictors (6,11,12).
The purpose of this article is to present some of our laboratory's novel hypotheses in the physical activity (PA)-PMT domain, which empirically test the following: (i) the moderating effects of age and sex and the coping appraisal constructs on the PMT's threat components; (ii) the integration of the PMT's cognitive components within a stage of change model; and (iii) an ordered, temporally sequenced model of the PMT's threat, fear, and coping components. This article also summarizes the published research in the PA domain, which has tested the PMT, which includes recent PA studies published in the past decade not included in past meta-analyses of the theory. Specifically, we briefly review the 14 published PA studies (seven of which are from our laboratory/primary author) and include samples across the general adult population, school-aged youth, cardiac patients, and adults with diabetes. Furthermore, directions for future research and a commentary on the current use of the PMT are presented.
Protection Motivation Theory Tests in the Physical Activity Domain
Nonintervention, cross-sectional tests
In a cross-sectional test, one of this article's authors, Plotnikoff, and Higginbotham (14) examined the use of the PMT with a randomly selected community sample of 800 Australian adults and found the coping appraisal constructs to have strong and significant associations with exercise intention and behavior. The threat appraisal (for heart disease) components had limited association with the exercise outcomes; fear was weakly associated with intention, whereas vulnerability was negatively associated with intentions and behavior. Vulnerability's negative association may be explained by Rogers' "boomerang effect" (21) where people who feel themselves vulnerable to a disease are more anxious about that illness and so adopt a more "defensive avoidance" style of coping, or it could be that those who are already taking precautions (e.g., exercising) are feeling less at risk from having a myocardial infarction.
One of this article's authors, Plotnikoff, and Higginbotham (15) also tested the cross-sectional associations of the coping and threat (i.e., heart disease complications) appraisal constructs with PA intentions and behavior among 147 cardiac patients and found that self-efficacy and fear were the only the PMT variables to emerge as significant correlates of exercise intentions. This study also examined the PMT's threat measures immediately after their myocardial infarction event and found no association with fear, vulnerability, and severity with PA intention or behavior 6-months later.
Nonintervention, longitudinal tests
The PMT in the PA domain also has been examined in seven nonintervention, prospective studies. We (18) examined the PMT's predictive ability for PA behavior related to heart disease, in a large randomly selected population sample of 1602 adults over two consecutive 6-month periods (i.e., period 1 (time 1 cognitions predicting time 2 behavior), period 2 (time 2 cognitions predicting time 3 behavior)). Self-efficacy and intentions significantly predicted subsequent PA behavior. The PMT variables and intentions explained 4% and 16% of the variance, respectively, in period 1 and 3% and 22% of the variance in PA at period 2. With PA intention, significant relationships with response efficacy, self-efficacy, and severity were observed, explaining 35% and 36% of the variance for intention at periods 1 and 2, respectively. Lippke and one of this article's authors, Plotnikoff, (9) tested and integrated the PMT and Stage Model with the above sample of 1602 adults over a 6-month period. The researchers tested whether the stages of the Transtheoretical Model (TTM) moderate the interrelation in predicting stage motivation, as well as the PMT variables' interactions in predicting stage transitions. The details of this study will be further outlined in the theory integration section of the article.
Tulloch and colleagues (26) tested the PMT in the prediction of PA intentions and behavior among 787 cardiac patients in context of secondary prevention of heart disease and found that perceived severity, response efficacy, and self-efficacy (the strongest construct) were predictors of exercise intentions and behavior explaining 23% and 20% of the variance in intention and behavior, respectively. In contrast, the PMT model was not reliable for predicting exercise behaviors at 12 months after hospitalization.
Similarly, Blanchard and associates (3) examined the PMT in explaining any significant variation in exercise intentions and behavior in 76 cardiac patients receiving a home-based cardiac rehabilitation program. Path analyses revealed that response efficacy was the main predictor of 3- and 6-month PA intentions. Self-efficacy significantly predicted 3- and 6-month exercise behavior. This study concluded that threat appraisal variables had limited motivational influence on exercise levels in home-based cardiac rehabilitation patients, whereas coping appraisal variables were useful in explaining exercise behavior in this population.
A cross-sectional and 6-month longitudinal analysis of 697 adults with type 1 diabetes and 1614 adults with type 2 diabetes was conducted by Plotnikoff et al. (17) to examine PMT in the context of diabetes management. The study revealed that self-efficacy was a stronger predictor of intention (ß = 0.64-0.68) than response efficacy (ß = 0.14-0.16) in individuals with type 1 or type 2 diabetes. Severity was significantly related to intention (ß = 0.06) in type 2 diabetes individuals only, whereas vulnerability was not significantly related to intention or PA behavior. Self-efficacy (ß = 0.20-0.28) and intention (ß = 0.12-0.30) were significantly associated with PA behavior.
In another study, we (19) separately examined the PMT constructs in predicting aerobic PA and resistance training behavior over a 3-month period in a national Canadian sample of 244 adults with type 2 diabetes. Self-efficacy and response efficacy were both significantly associated with intention (R2 = 0.43) and behavior (R2 = 0.19). In terms of resistance training, the PMT explained 56% and 20% of the variance in intention and behavior, respectively. Self-efficacy and response efficacy were both significantly associated with resistance training intention, whereas self-efficacy predicted resistance training behavior. This study was novel because research on the psychosocial predictors of this resistance training has been very limited, with no apparent published studies among adults with diabetes.
It is important to note that none of the nonintervention studies generated an R2 greater than 0.30 for explaining PA behavior, which is considered a minimum acceptable level for theory testing (2). Therefore, the integration of constructs from other theoretical models and the inclusion of moderating variables may need to be considered to facilitate the PMT in predicting a greater amount of variance for PA.
Intervention tests
Five studies have conducted experimental manipulations on the specific PMT variables, all by providing motivational essays. These studies related to the primary prevention of health-related issues through exercise. Courneya and Hellsten (4), Stanley and Maddux (24), and Wurtele and Maddux (30) reported the application of the PMT for PA prediction among 427, 195, and 160 university students, respectively. Fruin and colleagues (7) conducted a study with 615 adolescents, whereas Graham and colleagues (8) conducted a study with 173 teaching and school staff. Courneya and Hellsten (4) reported that perceived severity was the only variable found to have a significant effect on exercise intentions for cancer prevention. Stanley and Maddux (24) found that both self-efficacy and response efficacy were associated with subjects' PA intention, with response efficacy being the strongest construct. On the other hand, Wurtele and Maddux (30) revealed that both vulnerability and self-efficacy were associated with exercise intention and behavior. Fruin and colleagues (7) found that participants in the high self-efficacy condition presented stronger intentions to exercise, whereas those in the low response efficacy condition demonstrated more endorsement of hopelessness and fatalism than did students in the high response condition.
Graham and colleagues (8) found that persuasive message framing (presented in DVDs) was effective in manipulating participants' coping appraisal (response efficacy), which influenced their intentions to perform more exercise for colon cancer prevention, which, in turn, influenced their behavior to engage in initial exercise.
Milne et al. (11) conducted the only health education intervention based on the theory. The PMT-based health education intervention had a significant impact on intentions in a study of 248 undergraduate students but not on behavior in a 1-wk follow-up (35% vs 38% for the control group).
In sum, the findings from these 14 studies show some support for the use of the PMT's application to PA promotion (Table 1). The coping appraisal variables in predicting PA behavior is generally supported, with limited support for the theory's threat components that seem to be mainly salient for only the clinical populations with chronic diseases. This suggests threat appeals should be used judiciously (i.e., in certain clinical populations), and the promotion of the benefits and enhancing confidence of the behavior should be widely and strongly encouraged.
TABLE: Summary of research examining protection motivation theory and physical activity
Moderating effects on PMT
The PMT proposes that becoming aware of the severity of a threat that one is susceptible to will initiate protection motivation; however, the nature of the motivation will be based on coping appraisal (21). Although threat perception may contribute to precautionary motivation by provoking the consideration of outcome expectancies, perceptions of response efficacy and self-efficacy may predict intention formation and subsequent behavior change (22). However, the potentiality of the moderating effects of the PMT's coping cognitions on threat appraisal has been silent. Additional moderators to threat cognitions also may include age and sex. However, literature on potential demographic moderators of the PMT's threat cognitions also has remained limited. Our longitudinal study of a randomly selected population sample (N = 1602 adults) was designed to determine if the PMT's coping cognitions moderate threat cognitions for predicting PA intention and behavior, as well as to test if age and sex are moderators of threat cognitions for predicting PA intentions and behavior (18).
The study reported that the two coping cognitions did not emerge as significant moderators of threat cognitions in the prediction of PA intention and behavior across the study's two consecutive 6-month periods. The coping cognitions as moderators of threat explained only 1% of the variance in intention and behavior. However, it may be that threat is a precursor in formulating a high coping appraisal toward the recommended behavior, suggesting a possibility for an "ordered" PMT model (13,25) as described later.
Age and sex as moderators of threat did not provide additional variance in the models and thus were not significant moderators of threat cognitions in the prediction PA intention and behavior across the two consecutive periods. Although age and sex have been shown to significantly moderate PA in tests of other social-cognitive models, our study findings suggest that considering these two demographic factors in PMT-PA interventions may not produce favorable changes in PA behavior change. In addition to confirming the null effects of age and sex from this one study, other potential moderators (e.g., social-economic status, personality) could be examined in future research.
Theory integration
Combining and testing theoretical models may complement the explanatory ability of health behavior and guide interventions. For instance, the PMT has a clear model structure (Fig. 1), which TTM lacks; on the other hand, the TTM proposes discrete, measurable stages of behavior change, which are not included in the PMT. With the stages, there are various psychologically relevant outcomes, that is, not only behavior and motivation (as in the PMT) but more specific and qualitatively different mindsets through which individuals undergo in the process of actual behavior change. If the PMT predictions are found to be different across the TTM stages, the salient processes across the specific stages of behavior change may be identified. Weinstein et al. (28) argue that different constructs (such as the ones theorized by the PMT) could be important at different psychological stages, and therefore, stage-specific predictions exist as reported by others (16) As a result, a combination of a continuous theory and a stage model could be undertaken to test stage-specific prediction patterns.
Lippke and Plotnikoff (9) conducted a secondary analysis of a randomly selected population of 1602 adults to examine the interrelations of the PMT variables longitudinally to determine whether the TTM stages mediate their interplay and predictive ability of stage transitions. Mean patterns were investigated with the PMT variables for each of the five TTM stages (i.e., precontemplation, contemplation, preparation, action, maintenance). Vulnerability, severity, and response efficacy were different between precontemplation and contemplation. Self-efficacy was different between the first two and the last two stages - but not as predicted and reported in previous studies - between preparation and action. Intention was higher in contemplation than in precontemplation.
Moreover, the PMT "adequately fitted" the five stage groups. The multi-group, structural equation modeling revealed that covariances within threat appraisal and coping appraisal were invariant, and all other constrains were stage specific, that is, stage was a moderator. Intention fully mediated the relationship between severity, vulnerability, and response-efficacy and behavior. Predicting stage transitions with the PMT variables underscored the importance of self-efficacy. Only when threat appraisal and coping appraisal were high, stage movement was more likely in the preparation stage.
This study seems to be the first to investigate stage-specific interaction effects of the PMT's threat and coping appraisal and provides support for the usefulness of stage models, for the stage-specific prediction of behavior change and contradicted some of the critique of pseudo-stage assumptions (28). By integrating the two models (i.e., a stage model and a continuum model), this study demonstrated that the PMT's predictions are stage specific, which helps to explain the processes of behavior change. The TTM stages add to the PMT's view of behavior change in terms of more measurable behavior change outcomes, where not only changes in motivation and behavior are regarded as outcomes, but further cognitive changes (e.g., movements between stages and changes in stage-specific variables through preparatory acts) can serve as psychological meaningful outcomes.
An integrated PMT model may have use in further explaining PA behavior change. For example, additional constructs from social cognitive theory (e.g., social and environmental support), the theory of planned behavior (e.g., attitude), and postintentional constructs and personal/demographic characteristics found to be significant in explaining PA in other studies may provide unique concepts and enhanced explained variance in behavior. Previous PA research that has integrated additional constructs into an established theory has shown to increase the explanatory power of the behavior. For example, Rhodes and colleagues (20) successfully integrated personality, the perceived physical environment, and planning intentions into the theory of planned behavior framework to enhance the prediction of leisure-time walking.
Future research is needed in potentially broadening the use of the PMT to include additional social-cognitive variables from other theories that are deemed significant to guide the development of interventions. However, researchers need to be cautious about construct redundancy when examining the integration of theories and ground such integration on strong theoretical premises.
Ordered PMT
Compared with vulnerability and severity, the PMT's coping components of response efficacy and self-efficacy more consistently predicts PA intention. Thus, when considering the combined effect of response efficacy and self-efficacy as the coping appraisal, its strength in predicting intentions is much higher than that of the threat appraisal. Schwarzer (23) concludes that it is difficult for the threat variables to compete within the same regression equation with the coping variables. This had led some writers to suggest that threat appraisal precedes the coping appraisal (23,25). Tanner et al. (25) proposed and empirically tested several changes to the PMT model in which they term an ordered protection motivation model (OPMT) (Fig. 2). This model offers two major revisions from the existing PMT: (i) threat appraisal precedes the coping appraisal, and (ii) fear plays an essential role as a mediator between threat and coping appraisal (24). The results of an empirical study with 120 college students on sexual health behaviors provided partial support to these suggested new features that reported increases in the threat appraisal did increase fear levels (25).
FIGURE 2: Ordered protection motivation theory (PMT) conceptual model.
One of this article's authors, Plotnikoff, (13) tested the OPMT and PMT related to PA behavior in a cross-sectional questionnaire, with a representative community of 800 adults at high risk for coronary heart disease (see Plotnikoff and Higginbotham (14) for the PMT results). The models were adjusted for a host of demographic and medical characteristics. The explained variance for the coping components as dependent variables was similar between the PMT and OPMT models. However, two distinctive differences were found between the models in this study. First, the PMT produced a higher R2 for the threat variables; second, the OPMT yielded a higher R2 for fear as the dependent variable. Hence, the threat components (related to heart disease) were better explained by the PMT, whereas fear was better elucidated by the OPMT (i.e., both severity and vulnerability of heart disease were associated with fear in the regression models). The question raised by these findings highlights the theoretical issue concerning whether threat precedes fear or vice versa. A time series design would assist validation of these claims.
One of this article's authors, Plotnikoff, (13) also tested the OPMT against the PMT with 147 cardiac patients during hospitalization and 6 months after their cardiac event to determine the model's strength in determining exercise intentions and behaviors. The threat components (fear, severity, and vulnerability) were measured at the time of hospitalization (baseline). All the PMT's constructs and behavior were then assessed 6 months later (see Plotnikoff and Higginbotham (15)). The OPMT was not fully supported in this study. Baseline (time 1) fear had no impact on either of the coping efficacies at 6 months. Furthermore, baseline threat had no mediating effect on baseline fear. The explained variances for the intention and behavior equations were virtually the same for both the PMT and OPMT models indicating that the PMT's time two baseline threat (vulnerability and severity) measures had no additional effect on the PMT intention and behavior outcomes at 6 months. Furthermore, the R2 for baseline fear and the cognitive components were all identical for the remainder of each model's equations. These results fail to confirm Tanner's modification to the PMT (25). As discussed above in the previous study, this again highlights the theoretical question of whether threat precedes fear or vice versa. Interestingly, however, the emotional construct of fear of further heart problems (measured at baseline) was salient for predicting exercise intention at 6 months. Moreover, this observation was not apparent for low-fat dietary behavior at 6 months in the same sample, which could suggest that exercise may be a stronger motivator than dietary behavior after a myocardial infarction.
Other conclusions from this study reveal that fear's correlations with both severity and vulnerability were greater than the associations between severity and vulnerability. Hence, fear should be considered an equal component with severity and vulnerability within the threat appraisal. As previously mentioned, the threat measures had a strong effect on fear and intention. In contrast, fear's hypothesized influence on the coping appraisal was generally not supported. Future studies may want to further investigate potential interrelationships between the threat and coping variables in temporally sequenced, longitudinal designs to test moderation effects.
FUTURE RESEARCH
One intriguing example of further research would be the addition of bidirectionality dimension to the PMT model. Bidirectionality could be another possible explanation for some of the negative threat effects on the PMT outcomes (14). This suggests that changes in behavior lead to changes in threat perceptions. It can be assumed, according to Weinstein and Nicolich (27), that when people adopt precautions, they also perceive their threat to be lowered. However, the PMT and other major models (e.g., health belief model) assume that perceptions of high personal risk increase the likelihood of precaution adoption (threat's causal role). Hence, considering Weinstein and Nicolich's (27) assertion, models such as the PMT could be bidirectional in nature.
CONCLUSION
The PMT across all health behavior contexts has generally received strong support from correlational and longitudinal studies that have used the PMT as a social cognition model and, to a lesser extent, from experimental studies that have manipulated specific PMT variables (12). However, from a population health perspective, the use of PMT for PA promotion does not hold very well for the threat appraisal components; the coping appraisal variables (consistent with social cognitive theory (1)) have been consistently stronger predictors of PA intention and behavior.
However, there is some evidence that threat perceptions do hold some value in adopting health behaviors in general. The PMT proposes that becoming aware of the severity of a threat that one is susceptible to will initiate protection motivation; however, the nature of the motivation will be based on coping appraisal (21). Whereas threat perception may contribute to precautionary motivation by provoking the consideration of outcome expectancies, perceptions of response efficacy, response costs, and self-efficacy may predict intention formation and subsequent behavior change (22). Health promotion campaigns often combine messages highlighting threats to well-being with recommendations to adopt a protective behavior. This strategy is based on implicit assumptions about the impact of fear arousal and perceived threat on decision making and action regulation. Research efforts targeting these processes reveal that threat perceptions are shown to have weaker relationships than coping perceptions in terms of intention and behavior (6,11). This implies that it is not the ability to arouse fear that is likely to have influence on behavior, especially because fear may hinder the establishment of precautionary motivation through the initiation of fear control processes. Rather, it is the precautionary information or reassurance included in the message that will have the greatest impact on behavior (27,29).
The findings from the PA-PMT overview of studies presented in this article have theoretical, methodological, and practical implications for providing direction for theory building and intervention development. Theoretically, the comparison of the PMT with the OMPT highlights the OPMT's superiority in explaining fear as the dependent variable. Because of its demonstrated effect on other PMT variables, fear should be re-evaluated as more than a mediator of severity to confirm its place in the PMT. Furthermore, combining two different types of health behavior change models demonstrated the strength of differentiating stage-specific mechanisms in a continuum model. Integrating the PMT with tenants from other social cognitive models (e.g., social cognitive theory, theory of planned behavior) including implementation intentions also may be incorporated in a promising theory that needs improvement.
Methodologically, moderation of the demographic and coping appraisal variables was shown to not exist when these analyses were included in testing the PMT. However, other moderating variables (e.g., ethnicity and income) may contribute to PA behavior that warrants further investigation. In terms of measurement issues, not all of the PMT variables (i.e., response costs, rewards of maladaptive response) were tested in the reviewed studies. The inclusion of these variables along with the employment of congruent and better validated measures of the core PMT across many of the studies may have yielded stronger results. Future research also may consider using objective measures for examining PA (i.e., pedometry, accelerometry) as the literature in the PMT-PA domain has focused on self-report measures.
From a practical standpoint, Fishbein and colleagues (5) advocate identifying salient beliefs from the target population, developing persuasive messages around the beliefs, and then developing appropriate material based on the elicited beliefs. Based on our study results, these interventions would need to apply strategies for increasing the salience of self-efficacy and response costs for PA intention and behavior. For self-efficacy, interventions should focus on addressing how to overcome barriers to PA, for example. In terms of response efficacy, interventions should focus on highlighting the effectiveness of regular PA in reducing their risk for developing health complications (e.g., cardiovascular disease).
The implications of all of these study findings will help guide needed research toward the operationalization and testing of the PMT's coping appraisal variables in PA interventions. The findings suggest that the PMT in its current form may not be an overly useful theory for predicting and promoting PA behavior, given that the theory's threat appraisal variables have limited contribution in explaining and promoting PA in the general population. However, the PMT's threat appraisal may have relatively more (albeit limited) salience for PA strategies in clinical populations, but such threat and fear arousal strategies should be used only in conjunction with coping appraisal approaches.
Acknowledgments
R.C. Plotnikoff is supported by a Salary Award from the Canadian Institutes of Health Research (Applied Public Health Chair Program). L. Trinh is supported by Full-Time Health Research Studentships from the Alberta Heritage Foundation for Medical Research.
References
1. Bandura A.
Social Foundations of Thought and Action: A Social Cognitive Theory. Englewood Cliffs (NJ): Prentice Hall; 1986.
2. Baranowski T, Anderson C, Carmack C. Mediating variable framework in physical activity interventions: How are we doing? How might we do better?
Am. J. Prev. Med. 1998; 15(4):266-97.
3. Blanchard CM, Reid RD, Morrin LI, et al. Does protection motivation theory explain exercise intentions and behavior during home-based cardiac rehabilitation?
J. Cardiopulm. Rehabil. Prev. 2009; 29:188-92.
4. Courneya KS, Hellsten LAM. Cancer prevention as a source of exercise motivation: an experimental test using protection motivation theory.
Psychol. Health Med. 2001; 6:59-64.
5. Fishbein M, Von Haeften I, Appleyard J. The role of theory in developing effective interventions: implications from project SAFER.
Psychol. Health Med. 2001; 6(2):223-38.
6. Floyd DL, Prentice-Dunn S, Rogers RW. A meta-analysis of research on protection motivation theory.
J. Appl. Soc. Psychol. 2000; 30(2):407-29.
7. Fruin DJ, Pratt C, Owen N. Protection motivation theory and adolescents' perceptions of exercise.
J. Appl. Soc. Psychol. 1992; 22(1):55-69.
8. Graham SP, Prapavessis H, Cameron LD. Colon cancer information as a source of exercise motivation.
Psychol. Health 2006; 21(6):739-55.
9. Lippke S, Plotnikoff RC. The protection motivation theory within the stages of the transtheoretical model-stage-specific interplay of variables and prediction of exercise stage transitions.
Br. J. Health Psychol. 2009; 14:211-29.
10. Milne S, Orbell S, Sheeran P. Combining motivational and volitional interventions to promote exercise participation: protection motivation theory and implementation intentions.
Br. J. Health Psychol. 2002; 7:163-84.
11. Milne S, Sheeran P, Orbell S. Prediction and intervention in health-related behavior: a meta-analytic review of protection motivation theory.
J. Appl. Soc. Psychol. 2000; 30(1):106-43.
12. Norman P, Boer H, Seydel E. Protection motivation theory. In: Conner M, Norman P, editors.
Predicting Health Behavior. 2nd ed. London: Open University Press; 2005, p. 81-126.
13. Plotnikoff R. An Application of the Protection Motivation Theory to Coronary Heart Disease Risk Factor Behavior in Three Australian Samples: Community Adults, Cardiac Patients and School Children. Published Technical Report; The University of Newcastle; 1994.
14. Plotnikoff RC, Higginbotham N. Protection motivation theory and exercise behavior change for the prevention of heart disease in a high-risk, Australian representative community sample of adults.
Psychol. Health Med. 2002; 7(1):87-98.
15. Plotnikoff RC, Higginbotham N. Protection motivation theory and the prediction of exercise and low-fat diet behaviors among Australian cardiac patients.
Psychol. Health 1998; 13:411-29.
16. Plotnikoff R, Hotz S, Birkett N, Courneya K. Exercise and the transtheoretical model: a longitudinal test of a population sample.
Prev. Med. 2001; 33:441-52.
17. Plotnikoff R, Lippke S, Trinh L, Courneya K, Birkett N, Sigal R. Protection motivation theory and the prediction of physical activity among adults with type 1 or type 2 diabetes in a large population sample.
Br. J. Health Psychol. DOI: 10.1348/135910709X478826. Accessed November 13, 2009.
18. Plotnikoff RC, Rhodes RE, Trinh L. Protection motivation theory: a longitudinal test among a representative population sample of Canadian adults.
J. Health Psychol. 2009; 14:1119-34.
19. Plotnikoff RC, Trinh L, Courneya KS, Karunamuni N, Sigal RJ. Predictors of aerobic physical activity and resistance training among Canadian adults with type 2 diabetes: an application of the protection motivation theory.
Psychol. Sport Exerc. 2009; 10(3):320-28.
20. Rhodes RE, Courneya KS, Blanchard CM, Plotnikoff RC. Prediction of leisure-time walking: An integration of social cognitive, perceived environmental, and personality factors.
Int. J. Behav. Nutr. Phys. Act. 2007; 4:51.
21. Rogers R. Cognitive and psychological processes in fear appeals and attitude change: A revised theory of protection motivation. In: Cacioppo J, Petty R, editors.
Social Psychology: A Sourcebook. New York (NY): Guilford Press; 1983, p. 153-76.
22. Ruiter RAC, Abraham C, Kok G. Scary warnings and rational precautions: a review of the psychology of fear appeals.
Health Psychol. 2001; 16(6):613-30.
23. Schwarzer R. Self-efficacy in the adoption and maintenance of health behaviors: theoretical approaches and a new model. In: Bennett P, Weinman J, Spurgeon P, editors.
Current Developments in Health Psychology. Switzerland: Harwood Academic Publishers; 1992. p. 219-43.
24. Stanley MA, Maddux JE. Cognitive processes in health enhancement: investigation of a combined protection motivation and self-efficacy model.
Basic Appl. Soc. Psych. 1986; 7(2):101-13.
25. Tanner JF, Hunt JB, Eppright DR. The protection motivation model: a normative model of fear appeals.
J. Mark. 1991; 55:36-45.
26. Tulloch H, Reid R, Slovenic D'Angeloa M, et al. Predicting short and long-term exercise intentions and behavior in patients with coronary artery disease: a test of protection motivation theory.
Psychol. Health 2009; 24(3):255-69.
27. Weinstein ND, Nicolich M. Correct and incorrect interpretations of correlations between risk perceptions and risk behaviors.
Health Psychol. 1993; 12:324-33.
28. Weinstein ND, Rothman AJ, Sutton SR. Stage theories of health behavior: conceptual and methodological issues.
Health Psychol. 1998; 17:290-99.
29. Witte K, Allen M. A meta-analysis of fear appeals: implications for effective public health campaigns.
Health Educ. Behav. 2000; 27(5):591-615.
30. Wurtele SK, Maddux JE. Relative contributions of protection motivation theory components in predicting exercise intentions and behavior.
Health Psychol. 1987; 6(5):453-66.
