Personality’s Prism: Personality–message Fit in Responses to Four Anti-Smoking Public Service Announcements

In the field of health communication, experts have employed a variety of strategies, one of which is the targeting strategy. This approach involves customizing health messages for specific segments of the population (Biener et al., 2004). Such personalization has been shown to enhance the effectiveness of these messages (Xu, 2017), though there are acknowledged limitations in the application of tailoring and targeting methods (Kreuter et al., 2013). Recent research has underscored the necessity of micro-level targeting in health communication, recognizing the variance in individuals’ responses to health interventions (Baumel et al., 2020).

Thus, this study examines individual differences in the interpretation and processing of health interventions, specifically in the context of smoking. We propose the use of the five-factor model (FFM or the big five personality trait model) (McCrae & John, 1992) as a tool for health communicators to categorize smokers into more distinct groups based on shared psychological and behavioral characteristics. The FFM has been widely applied in behavioral research, acknowledging that personality influences how individuals process information (McCrae, 2010). However, in health communication, there has been limited exploration of how the public’s personality traits might affect the impact of health messages (Miller et al., 2006). Previous research indicates that the FFM can be instrumental in identifying personality traits that predict addictive smoking behaviors. Despite this, there has been a lack of focus in health communication research on how the FFM can be used to anticipate responses to health-promoting messages.

Consequently, this study investigates how the five personality factors influence individuals’ perceptions of anti-smoking public service announcements (PSAs). By applying a personality-based framework to message perception, we aim to offer theoretical and practical insights into how communicators might design more psychologically tailored and effective health messages.

Literature Review

Overlooked Role of Personality in Health Communication

Research in health communication has consistently emphasized that the effectiveness of persuasive interventions depends on understanding audience diversity. A long tradition of studies on segmentation and message tailoring has demonstrated that messages are more compelling when they align with the audience’s needs, beliefs, and motivations (Dutta-Bergman, 2003; Noar et al., 2007; Rimer & Kreuter, 2006; Slater, 1996). Tailored interventions are found to enhance message involvement, perceived relevance, and behavioral intention, as individuals tend to engage more deeply with messages that address their specific health concerns (Kahlor et al., 2003; Lake et al., 2018; Lustria et al., 2016).

Despite these advances, most segmentation and tailoring approaches rely heavily on demographic or situational characteristics such as age, gender, education, or health status (Campo et al., 2012; Schmid et al., 2008). These variables help define broad audience categories but do not capture deeper psychological and dispositional differences that directly influence message interpretation and persuasion (Gaysynsky et al., 2022; Hirsh et al., 2012). Consequently, while scholars have paid considerable attention to social and contextual determinants of health communication, relatively little research has examined stable personality traits as predictors of message processing and attitudinal responses (Hakulinen et al., 2015).

This lack of attention is partly due to the practical difficulty of assessing personality in large-scale interventions and the field’s traditional focus on proximal situational factors (Kreuter & Wray, 2003; Rimer & Kreuter, 2006). Yet personality is one of the most fundamental sources of variation in how individuals attend to, evaluate, and respond to persuasive health information (Baumel et al., 2020; Halttu & Oinas-Kukkonen, 2022). Integrating personality into health communication research can therefore provide a more fine-grained understanding of audience variability, complementing demographic segmentation with psychological depth.

Considering this gap, the present study investigates how the Big Five personality traits predict individuals’ perceptions of anti-smoking PSAs. The anti-smoking context offers an especially relevant test case, as smoking is a behavior that reflects self-control, identity, and moral evaluation—domains where personality differences play a critical role in shaping message reception and persuasion.

Five-Factor Model for Personality

This study proposes that personality traits are crucial variables that can indirectly influence health behaviors. While there are various ways to conceptualize personality, the FFM provides a comprehensive framework for understanding personality (Zhang et al., 2024). The FFM includes traits of openness, conscientiousness, extraversion, agreeableness, and neuroticism, each of which can predict different behavioral patterns and decisions in similar environmental conditions (Briggs, 1992; McCrae & John, 1992). Openness to experiences reflects a tendency toward embracing new ideas, values, feelings, and behaviors. Conscientiousness indicates a propensity to adhere to social norms and rules like punctuality, reward for hard work, and sincerity. Neuroticism relates to the tendency for emotional instability and resilience in stressful situations. Extraversion describes the inclination to engage in social activities, such as conversations and events. Finally, agreeableness measures the tendency to concur with others’ opinions over personal beliefs.

Psychology and behavioral science scholars have demonstrated the relevance of personality in behavioral research (McCrae & John, 1992). The idea of the FFM is that each dimension of personality can contribute to the prediction of individuals’ decisions on an issue. For instance, a higher extroversion level can predict more social activities of individuals than individuals who have lower extroversion levels.

Health communication research has applied personality factors across various medical and health domains (Dutta-Bergman, 2003, 2005). Studies indicate that conscientiousness is predictive of favorable attitudes toward pro-health behaviors and intentions to engage in them (Bogg & Roberts, 2004), while neuroticism and extraversion have been found to have negative and positive correlations with pro-health behaviors under certain conditions (Conner & Abraham, 2001). The FFM has also been used to explore areas such as sexual health (Allen & Walter, 2018), workaholism (Andreassen et al., 2016), and mental health (Furnham & Cheng, 2019). Specifically, in the context of smoking, the FFM’s applicability in predicting smoking behaviors is well-established, with longitudinal studies showing that traits like openness to experience, neuroticism, and extraversion can be predictors of smoking behavior over a person’s lifetime (Hakulinen et al., 2015; Zvolensky et al., 2015).

However, the direct correlation between personality traits and behavioral or emotional reactions to persuasive health messages is complex (Wall et al., 2019). The effectiveness of health communication strategies is significantly influenced by the context and the nature of the message, and these influences vary based on personality traits. Noar et al. (2007) highlighted the challenges in establishing direct links between personality traits and reactions to persuasive health messages. This variability underscores the need for further research into how personality traits, particularly those outlined in the FFM, affect the reception of and reactions to health messages (Zhang et al., 2024), including anti-smoking campaigns.

Therefore, the FFM’s role in shaping responses to health communication strategies, especially in influencing the reception of anti-smoking messages, warrants further investigation. However, the role of FFM in shaping responses to health communication, especially in anti-smoking messages, is less clear, even though those studies have enlightened individual differences in persuasive message acceptance in the context of health communication. The association between individuals’ personalities and behavioral/emotional reactions to persuasive messages varies across communication situations and contexts (Zhang et al., 2024). This gap highlights the need for research exploring how each of the Big Five personality traits influence the reception of and response to health messages.

In this light, we suggest the first research question that explores how the Big Five personality traits might influence individual responses to anti-smoking campaigns. This exploration is vital for understanding how personality traits can be leveraged to enhance the effectiveness of health communication strategies.

RQ1: Do individuals’ perceptions of anti-smoking messages differ based on their five personality traits?

Perceptions of Smoking as a Factor of Reactions to Anti-Smoking Campaigns

In the specific context of anti-smoking campaigns, the influence of personality on the effectiveness of persuasive messages is likely to vary, as each message employs distinct appeals and persuasive strategies (Hirsh et al., 2012). For example, messages emphasizing anti-smoking as a social norm could be more effective for individuals with high levels of conscientiousness, as this trait is associated with a higher susceptibility to social norms (Orji et al., 2017). Similarly, it is reasonable to hypothesize that the level of neuroticism might influence the effectiveness of fear appeal strategies, which typically work by evoking negative emotions such as anxiety or distress (Ajit & Raj, 2015).

Given this potential variability, our study’s second research question investigates whether the impact of personality traits on the reception of health messages changes depending on the style of persuasion used. This question is crucial as it addresses the diverse range of persuasive strategies and aims to understand the interplay between these strategies and various personality types (Halttu & Oinas-Kukkonen, 2022). By exploring this relationship, we can gain insights into how different persuasive techniques may be more or less effective depending on individual personality characteristics.

RQ2: Is there any difference in the effect of FFM on the perception of anti-smoking messages depending on the messages’ persuasion styles?

Smokers and non-smokers, while both stakeholders in the issue of tobacco control, often exhibit divergent perceptions and reactions to anti-smoking messages. Non-smokers, though indirectly affected by smoking-related issues, are not the primary targets of anti-smoking campaigns and therefore may engage with these messages differently than smokers (Zhu et al., 2006). The effectiveness of such campaigns is thus contingent upon the audience’s smoking status and behavioral orientation toward the issue (Bauhoff et al., 2017). Prior research has also shown that differences in pre-existing attitudes and behaviors toward smoking can produce fundamentally distinct responses to health messages between these two groups (Lizama et al., 2019).

These distinctions become more critical when considered alongside individual personality traits as defined by the FFM. Personality not only influences how people form health-related attitudes but also how they process persuasive information. According to the elaboration likelihood model (ELM), smokers—whose behaviors are directly challenged by anti-smoking messages—are likely to experience higher issue involvement and thus process campaign information through a central route, carefully evaluating message arguments (Flynn et al., 2011). In contrast, non-smokers may rely on a peripheral route, attending more to superficial cues or general impressions of the campaign (Wu et al., 2024). This difference in involvement and processing depth suggests that the interaction between smoking status and personality traits may systematically shape how individuals perceive and evaluate anti-smoking messages (Ghods et al., 2024).

Building on this reasoning, examining whether the FFM differentially influences smokers and non-smokers can yield meaningful insights for health communication. Understanding these personality-based differences is essential for developing more tailored and effective anti-smoking campaigns (Halttu & Oinas-Kukkonen, 2022), ensuring that the unique motivations, perceptions, and needs of each group are adequately addressed. Such insights can ultimately advance broader public health goals by promoting more targeted and impactful communication strategies.

H1: The effects of the five-factor model on the perception of anti-smoking messages will differ between smokers and non-smokers.

Methods

Survey Data and Procedure

A pilot survey was conducted in June 2023 to review the questionnaire items prior to the main survey in South Korea, and a structured questionnaire was completed. To compare smokers and non-smokers, 1,875 adult smokers and 1,877 nonsmokers were recruited for the study through an online survey panel provided by a survey firm, Kstat, in July 2023, and the sampling error is +1.6%. Proportional quota sampling was used by gender, age, and region according to the proportion of smokers and non-smokers as presented by the 2021 Korean Community Health Survey data. The survey subjects were adult men and women aged 19 to 65. A total of 34,791 invitation emails were sent to online panels managed by the survey firm, and 3,752 completed the survey (response rate = 10.8%). Participants first answered items to assess their FFM. We then showed images of four famous anti-smoking PSAs that had aired in South Korea in the previous 20 years. We then measured participants’ perceptions of each PSA that they can recall.

Anti-Smoking Campaigns

In this survey, we used assisted recall to measure participants’ perceptions of PSA campaigns. Participants showed the pictures and taglines of television aired video-type PSA campaigns about anti-smoking. Anti-smoking advertisements in the survey were commissioned and supported by governmental organizations, such as the Korea Health Promotion Institute, an affiliated organization under the Ministry of Health and Welfare.^[1] We selected the four most significant campaigns among PSAs that were aired in the last 20 years. This selection was based on the Korea Health Promotion Institute’s internal evaluation of their anti-smoking campaigns. Each campaign reflects the changes in anti-smoking campaigns.

The first campaign emphasizes the role of expertise in quitting smoking. Actors play the smoker who wants to quit smoking, but the advertisement shows that it is difficult to quit smoking through willpower alone. The message suggests that the government can help smokers quit smoking (aired in 2022).^[2] The second PSA talks about adolescent smoking, which is very popular in South Korea. This campaign has the tagline “I hope you are #nosmoking.” The atmosphere of the advertisement is bright and cheerful (aired in 2021).^[3] The third PSA is a fear appeal message depicting the situation of purchasing cigarettes in the store (aired in 2015). Behind the smoker who is buying cigarettes, the future self of the smoker is writhing in agony due to the smoking-caused diseases.^[4] The last PSA was a celebrity endorsement (aired in 2002). In this video campaign, a male celebrity who has lung cancer testifies about the risk of smoking. He warned smokers about the damage of smoking using himself as an example.^[5]

Independent Variable

The independent and dependent variables in the study were measured with 5-point Likert-type scales.

Five Factors

Measurement items for the five personality factors were adopted from previous research (Donnellan et al., 2006; McCrae & John, 1992). Extraversion was measured with the four items, such as "I am not talkative in meetings in general (αα = .77). Agreeableness was measured with the four items, such as “I easily feel sympathy with other people’s emotions” (αα = .72). Conscientiousness was measured with five items, such as “I am good at planning and following the plan” (αα = .70). Neuroticism was measured with a single item, “I am not emotionally stable.” Openness to experience was measured with the four items, such as “I am imaginative (αα = .76).”

Dependent Variables

Attitude toward a PSA

Individuals’ attitudes toward PSA campaigns were measured with eight items (e.g., “This message is fun”). Other items measured: how much individuals perceive the message as favorable, interesting, persuasive, effective, excellent in information and format, and whether it makes respondents feel quitting smoking (smokers)/continuing to avoid smoking (non-smokers) (αfor PSA1 α = smokers .87 / non-smokers .89, PSA2 α = .91/93, PSA3 α = .86/.89, and PSA 4 α = .85/.91).

Control Variables

In our regression models, we controlled for a set of covariates to isolate the unique contribution of personality traits on attitudes toward anti-smoking PSAs. First, individuals’ perceptions of smoking and perceived risk of smoking were included, because beliefs that smoking is beneficial or the extent of perceived risk are known to influence how smokers evaluate cessation messages (Addo et al., 2024; Dijkstra, 2009).

Second, two behavioral smoking indicators were measured with single item. Smoking frequency was assessed by asking, “On average, how many cigarettes do you smoke per day?” with four response options: (1) 10 cigarettes or fewer, (2) 11-20 cigarettes, (3) 21-30 cigarettes, (4) more than 30 cigarettes. Smoking duration was measured with single item, “How long have you been smoking?” (for former smokers: “How long did you smoke in the past?”) with six categories: (1) 5 years or less, (2) 6-10 years, (3) 11-15 years, (4) 16-20 years, (5) 21-30 years, (6) more than 30 years. These indicators capture both intensity and history of smoking behavior, which reflect dependence and involvement with the issue and are known to affect responsiveness to persuasive health communication (Addo et al., 2024).

Third, demographic factors – age, gender, and education – were included. Education was measured using a five-category item: (1) middle school or lower, (2) high school graduate or lower, (3) some college, (4) college graduate, (4) graduate school. These variables are commonly associated with differential media exposure, health literacy, and message processing in tobacco-control research (Hsieh et al., 1996; Slocum et al., 2022).

Finally, media consumption habits were controlled. TV use was measured by asking, “On average, how many hours per day do you watch TV?” with six response options ranging from (1) none to (6) 4 hours or more. Internet use was calculated as the mean of two items assessing daily Internet use on PC and mobile devices. Respondents reported their average daily usage on each device using a seven-point scale ranging from (1) none to (7) more than 5 hours. These measures account for overall media consumption patterns that may influence exposure to or engagement with anti-smoking campaigns (Slocum et al., 2022), thereby reducing potential confounding in estimating the relationship between personality and PSA attitudes

Analysis

Our research questions ask about the association between the five factors of personality and attitudes toward anti-smoking PSAs. To answer those questions, we ran four different hierarchical regressions per the data set (smokers and non-smokers). Since there were four different PSA campaigns, four multiple regressions were conducted for each sample (smokers and non-smokers). Thus, a total of eight hierarchical regression models were estimated. To ensure message-specific validity, analyses included only those participants who had previously encountered the PSA – whether on television or via digital platforms – and accurately recalled its content at the time of the survey. Consequently, the sample size (n) varied across the eight models due to this recall-based inclusion criterion. To further examine potential group differences, interaction terms between personality traits and smoking status (FFM × Smoking) were included in the pooled dataset analysis. This additional test allowed us to determine whether the effects of personality on PSA attitudes varied between smokers and non-smokers.

Results

Our research questions addressed how the five personality factors relate to individuals’ attitudes toward anti-smoking PSAs and whether these relationships differ across message types and audience groups. To answer these questions, we conducted a series of hierarchical regression analyses. Since the dataset contained two distinct samples—smokers and non-smokers—separate analyses were run for each group. Because participants evaluated four different PSA campaigns, four regression models were estimated per sample, resulting in a total of eight hierarchical regression models. Block 1 included the control variables (e.g., smoking perception, smoking risk, smoking frequency, smoking duration, age, gender, education, Internet use, and TV use), and Block 2 added the FFM variables to examine their incremental explanatory power.

Hypothesis Testing: Smokers

Four hierarchical regression analyses were conducted to examine how the Big Five personality traits predicted smokers’ attitudes toward four distinct anti-smoking PSAs. Block 1 included the control variables—perception of smoking, risk of smoking, smoking frequency, smoking duration, age, gender, education, Internet use, and TV use—and Block 2 added the FFM variables. Table 2 summarizes the detailed statistics for all models.

All four models were statistically significant, and the addition of the FFM variables significantly improved the explained variance across PSAs, range of ΔR² = .02–.06, all p values < .001, corresponding to small-to-medium effect sizes (f² = .03–.07). However, the significance of individual personality traits varied depending on the message framing and content of each PSA.

For PSA1 (“Smokers need help from experts”), the model was significant, F(14, 912) = 12.54, p < .001, Adj. R² = .15. Agreeableness, conscientiousness, and neuroticism were significant positive predictors (ΔR² = .05, f² = .06), suggesting that smokers high in empathy, responsibility, and emotional sensitivity were more persuaded by the expert-help appeal. For PSA2 (“Anti-smoking is trendy—targeting adolescents”), the model was significant, F(14, 977) = 6.67, p < .001, Adj. R² = .07. Only conscientiousness was a significant predictor (ΔR² = .02, f² = .03), indicating that responsible smokers were more receptive to social-norm–based appeals, whereas other traits were nonsignificant. For PSA3 (“Buying cigarettes is buying disease”), the model was significant, F(14, 700) = 8.55, p < .001, Adj. R² = .13. Conscientiousness again showed a strong positive relationship (β = .32, p < .001). This PSA produced the largest incremental effect (ΔR² = .06, f² = .07), indicating a medium-sized influence of personality on persuasion. For PSA4 (“Smoking caused my lung cancer—celebrity endorser”), the model was significant, F(14, 1078) = 16.20, p < .001, Adj. R² = .16. Both extraversion (β = .11, p =.002) and conscientiousness (β = .13, p = .002) were significant predictors (ΔR² = .03, f² = .03), suggesting that outgoing and responsible individuals responded more favorably to testimonial-based messages.

Across all models, Conscientiousness consistently predicted more favorable PSA evaluations, making it the most stable and influential personality factor. Agreeableness, extraversion, and neuroticism showed PSA-specific effects, while openness to experience did not predict any PSA attitudes. Among the control variables, risk of smoking was a robust positive predictor across all PSAs (range of β = .26–.38, all p values < .001). Perception of smoking (positive beliefs about smoking) tended to reduce PSA favorability, particularly in PSAs 1 and 4. Smoking duration and frequency were generally negative predictors, whereas TV exposure consistently showed small positive effects. Older and female smokers exhibited slightly higher favorability toward some messages, while education effects were minor.

Together, these findings indicate that personality—especially Conscientiousness—plays a meaningful role in shaping smokers’ attitudes toward anti-smoking campaigns, with the magnitude and direction of these effects varying according to the thematic framing of each PSA.

Variance inflation factors (VIFs) were examined for all predictor variables across the four hierarchical models. As shown in Table 3, all VIF values ranged between 1.0 and 2.7, well below the common cutoffs of 10.0 (Kutner et al., 2005), indicating that multicollinearity was not a concern in any of the analyses. Across models, the highest VIF values were observed for age and smoking duration (≈ 2.3–2.7), reflecting mild shared variance with other demographic variables. Personality traits showed moderate intercorrelations (VIFs ≈ 1.3–1.6), which are expected due to theoretical overlaps among the Big Five dimensions. No tolerance values fell below .40, further confirming stable estimation. Thus, all four models satisfied the assumption of independent predictors, and the reported standardized coefficients can be interpreted reliably without bias from multicollinearity.

Hypothesis Testing: Non-Smokers

We conducted four hierarchical regression analyses predicting attitudes toward four types of anti-smoking PSAs among nonsmokers. Block 1 included control variables (perception of smoking, risk of smoking, age, gender, education, Internet use, and TV use), and Block 2 added the Big Five personality factors. Detailed statistics are presented in Table 2.

The overall models were statistically significant for all four PSAs. For the first PSA (smokers need help from experts), the full model explained 12% of the variance in PSA attitude, F(12, 849) = 9.70, p < .001, ΔR² = .04, f² = .05. Among the control variables, perception of smoking (b = .07, p < .01), Internet use (b = .06, p < .001), and TV use (b = .04, p < .05) significantly predicted positive attitudes. Within the personality block, agreeableness (b = .10, p < .05) and conscientiousness (b = .11, p < .01) emerged as significant predictors, indicating that nonsmokers higher in agreeableness or conscientiousness tend to hold more favorable attitudes toward this expert-oriented PSA.

For the second PSA (anti-smoking as a social trend for adolescents), the model accounted for 9% of the variance: F(12, 977) = 6.67, p < .001, ΔR² = .02, f² = .02). Among the Big Five, only agreeableness (b = .21, p < .001) was a strong and consistent predictor, implying that individuals who are more cooperative and sympathetic perceived this socially driven PSA more favorably.

For the third PSA (buying cigarettes is buying disease), the model explained 9% of the variance: F(12, 1129) = 9.31, p < .001, ΔR² = .05, f² = .06). From the personality traits, agreeableness (b = .21, p < .001) again showed a strong positive association, indicating that empathic and prosocial individuals responded more favorably to the moral message framing of this PSA.

For the fourth PSA (celebrity endorsement: “smoking caused my lung cancer”), the full model explained 8% of the variance: F(12, 1092) = 8.41, p < .001, ΔR² = .03, f² = .03). Within the Big Five, agreeableness (b = .18, p < .001), conscientiousness (b = .08, p < .05), and neuroticism (b = .06, p < .01) were significant predictors, suggesting that emotionally and socially attuned individuals may be more persuaded by a testimonial-style PSA featuring a credible celebrity.

In sum, agreeableness consistently emerged as the strongest personality predictor across all four PSAs, while conscientiousness and neuroticism showed conditional effects depending on the PSA framing.

In four hierarchical regressions, VIFs were examined for all predictors in both blocks. Across the four PSA models, all VIF values ranged from 1.03 to 1.51, well below the conventional threshold of 5, indicating no problematic multicollinearity. Specifically, in Block 1, VIFs for control variables ranged between 1.05 and 1.23, while in Block 2 (including personality traits), the highest observed VIF was 1.51 for agreeableness in PSA4. These results confirm that the predictor variables were sufficiently independent and that collinearity did not distort the regression estimates.

Interaction Between Smoking Behavior and FFM

For each PSA, we estimated three hierarchical regression models using the full sample (smokers and nonsmokers combined), as the effects of smoking status were tested through Smoking × Personality interaction terms (see Table 4). Here, Block 1 included control variables (age, gender, education, Internet use, TV use, perception of smoking, and perceived risk of smoking). Block 2 added the Big Five personality traits. Block 3 included the interaction terms between smoking status and each personality trait (smoking coded as 1 = smoker, 0 = nonsmoker).

Hierarchical regression analyses indicated that adding Block 3 significantly improved model fit across all four PSAs. For PSA1, the inclusion of personality traits produced a significant increase in explained variance, ΔF(6, 1812) = 4.51, p < .001. A similar pattern emerged for PSA2, ΔF(6, 2158) = 11.22, p < .001, and PSA3, ΔF(6, 1514) = 3.87, p < .001. The model for PSA4 also showed a significant improvement, ΔF(6, 2214) = 4.29, p < .001. These results demonstrate that personality traits contributed meaningful explanatory power beyond smoking perceptions, risk beliefs, and demographic factors. The following section summarizes the key statistical results for each PSA type.

PSA1 (Smokers Need Help from Experts)

The Non-Smoking × FFM interaction reached conventional significance. The Neuroticism × Smoking interaction showed a marginal trend, B = .06, SE = .03, p = .088. Consistent with expectations, smokers evaluated the PSA less favorably overall, as indicated by a negative main effect of smoking, B = −.19, SE = .04, p < .001.

PSA2 (Anti-Smoking as a Social Trend for Adolescents)

Three personality interactions were significant. Agreeableness × Smoking was negative, B = −.16, SE = .06, p = .013, indicating that agreeableness was a stronger positive predictor among non-smokers than smokers. Conscientiousness × Smoking was positive and robust, B = .21, SE = .06, p < .001, suggesting that conscientiousness more strongly predicted favorable attitudes among smokers. Openness × Smoking was negative, B = −.12, SE = .06, p = .034, reflecting an attenuated openness effect among smokers. The main effect of smoking was again negative, B = −.29, SE = .04, p < .001.

PSA3 (Buying Cigarettes Is Buying Disease)

Conscientiousness × Smoking was significant, B = .24, SE = .07, p < .001, showing that conscientious individuals who smoke responded particularly favorably to this threat-framed message. Openness × Smoking showed a marginal negative pattern, B = −.11, SE = .06, p = .085. Smokers again expressed less favorable attitudes overall, B = −.12, SE = .05, p = .011.

PSA4 (Celebrity Endorsement: “Smoking Caused My Lung Cancer”)

Extraversion × Smoking was significant, B = .10, SE = .05, p = .027, indicating that extraversion more strongly predicted positive attitudes among smokers. Agreeableness × Smoking showed a marginal negative trend, B = −.10, SE = .06, p = .081. As with other messages, smoking predicted lower attitudes overall, B = −.15, SE = .04, p < .001.

Across PSAs, agreeableness—and in some cases openness—more strongly predicted positive evaluations among non-smokers, whereas conscientiousness (PSA2 and PSA3) and extraversion (PSA4) were more predictive among smokers. Variance-inflation diagnostics indicated no multicollinearity concerns in Block 3 models (all VIFs ≤ 3.28).

Discussion and Implications

This study examined how the FFM of personality traits predicts individuals’ attitudes toward anti-smoking PSA campaigns. By comparing smokers and non-smokers across four message types—expert-centered guidance (PSA1), adolescent social-trend framing (PSA2), smoking as self-harm/disease (PSA3), and celebrity-endorsed narrative appeals (PSA4)—we identified how personality-based differences shape responses to persuasive health communication. The hierarchical regression analyses confirmed that personality traits contribute explanatory value beyond conventional predictors such as smoking perception, perceived risk, and demographics.

Across the full models, conscientiousness emerged as the most consistent predictor of favorable evaluations, particularly among smokers. Individuals high in conscientiousness—organized, self-disciplined, and future-oriented—responded more positively to anti-smoking messages, suggesting that appeals emphasizing control, planning, or long-term health consequences may resonate strongly with this group. Agreeableness, in contrast, was the strongest and most stable predictor among non-smokers, indicating that empathetic and prosocial individuals are especially receptive to messages framed around social norms, collective well-being, or moral concern.

Extraversion showed selective effects, enhancing positive responses to socially framed or celebrity-endorsed messages, whereas neuroticism had a modest but meaningful influence on reactions to expert-centered campaigns that provide emotional reassurance. Notably, interaction analyses demonstrated that the strength and direction of these personality effects varied by smoking status, underscoring that personality operates as a context-dependent determinant of PSA persuasion that interacts with audience involvement and message framing.

Theoretical Implications

This study advances the theoretical understanding of persuasive health communication by empirically demonstrating that personality traits—particularly conscientiousness, agreeableness, and extraversion—serve as stable dispositional factors that meaningfully shape audience responses to anti-smoking messages. Even after controlling for conventional predictors such as perceived risk, smoking perception, and demographic variables, personality explained additional variance in message evaluation. This indicates that personality exerts an independent and systematic influence on persuasive processing, broadening the theoretical focus of health communication beyond situational or cognitive determinants.

Building on dual-process frameworks such as the ELM (Petty & Cacioppo, 1986) and message-tailoring theory (Rimer & Kreuter, 2006), our results position personality as a dispositional antecedent of the message-processing route. Individuals high in conscientiousness or agreeableness appear more likely to engage in deliberate, argument-based elaboration, whereas extraverted or neurotic individuals may rely on affective and social cues. This reframes personality as a pre-cognitive filter that determines how audiences attend to and interpret persuasive appeals, enriching the conceptualization of involvement and motivation to process in existing persuasion models.

Importantly, the inclusion of interaction analyses revealed that the role of personality is context dependent. The predictive patterns of personality differed between smokers and non-smokers—conscientiousness emerged as the most consistent predictor among smokers, while agreeableness dominated among non-smokers. These results imply that personality interacts with audience involvement and message relevance in shaping attitudinal responses, providing empirical support for trait-by-context perspectives in persuasion theory. Future theoretical frameworks should therefore model these interactive dynamics rather than treating individual differences as fixed or uniform predictors.

Finally, by identifying personality as a key antecedent of message interpretation, this study contributes to the emerging body of research connecting psychographic segmentation and algorithmic personalization in health communication. As AI-assisted message design becomes increasingly prevalent, integrating personality constructs into predictive models of persuasion offers a more psychologically grounded, human-centered approach to tailoring health messages—one that recognizes the interplay between enduring traits and contextual engagement.

Practical Implications

Practically, our findings offer actionable insights for designing personality-sensitive anti-smoking campaigns. First, message framing should reflect dominant personality tendencies. Conscientious individuals—who value order, responsibility, and long-term control—may respond better to efficacy- and goal-oriented appeals that emphasize discipline and future benefits (e.g., “Stay committed—your future self will thank you”). In contrast, those higher in neuroticism may engage more with emotionally charged or fear-based narratives that address immediate health threats and anxiety reduction.

Second, message tone and appeals can be aligned with social orientation. Campaigns highlighting empathy, cooperation, and collective well-being may resonate more strongly with agreeable audiences (particularly non-smokers), while extraverted individuals may prefer socially interactive or testimonial-based content such as influencer- or celebrity-driven PSAs. Openness-oriented audiences, though less predictive in this dataset, may still respond favorably to creative, autonomy-focused messaging that frames quitting as self-exploration.

Third, personality can guide channel and format selection. Conscientious or introverted audiences may prefer reflective, information-dense formats, such as personalized quit-plan apps or educational videos, whereas extraverted and socially oriented individuals may engage more through participatory social media challenges or peer-involvement campaigns.

Crucially, these strategies should differentiate smokers and non-smokers. Smokers—characterized by higher personal involvement and behavioral relevance—benefit from detailed, efficacy-based, and self-regulatory messaging, whereas non-smokers respond better to concise, socially framed messages reinforcing anti-smoking norms and collective responsibility. Together, these personality- and status-based insights can inform multi-channel campaign design that maximizes both cognitive engagement and emotional resonance.

Limitations and Future Directions

This study offers novel insights but also several limitations. First, attitudes toward anti-smoking campaigns were measured through recall-based responses, capturing long-term impressions rather than immediate reactions. Because the sample was not longitudinal, we could not compare original exposure effects with current evaluations. Future work should employ real-time or repeated-exposure designs to examine attitude stability over time.

Second, although the hierarchical and interaction analyses strengthen the explanatory power of our model, causality cannot be inferred. Experimental manipulations varying both message type and personality salience would help clarify the mechanisms through which personality traits influence persuasion and whether these effects generalize to other health domains (e.g., diet, vaccination, mental health).

Lastly, extending this trait-by-context analytical framework to cross-cultural or AI-mediated health communication settings could reveal how cultural norms or algorithmic personalization further moderate these relationships, advancing a more integrated and adaptive understanding of personality in persuasive health messaging.