Accepted on 29 Jul 2018
Submitted on 09 May 2017
In the Netherlands, HPV-vaccination uptake remains to be low (46%) (Van Lier et al., 2018). Hence, a need exists for evidence-based interventions aimed at increasing HPV-vaccination uptake. Several studies have provided insight in factors constituting HPV-vaccination acceptability (e.g., Alberts et al., 2017; Hofman et al., 2014; Pot et al., 2017; Van Keulen et al., 2013). Results from two Dutch studies among mothers of girls to-be-invited showed that ‘modifiable’ social-psychological determinants (e.g., perceived daughters’ susceptibility towards HPV-related risks) accounted for large proportions of variance in HPV-vaccination intention (Van Keulen et al., 2013; Pot et al., 2017). Mothers’ intention, in turn, was found to be the main and stable predictor of their daughters’ actual uptake (Pot et al., 2017). This was also found by Hofman and colleagues (2014) among parents of invited girls and is in line with social cognitive models such as the Reasoned Action Approach (Fishbein & Ajzen, 2010). Furthermore, mothers appeared most important in the process of decision-making about their daughters’ HPV-vaccination (Van Keulen et al., 2013). These insights led to the development of an innovative, interactive, web-based, tailored intervention promoting HPV-vaccination acceptability among mothers of invited girls (Pot et al., 2018).
One of the determinants to be targeted in the intervention was perceived daughters’ susceptibility towards HPV-related risks because this appeared an important determinant of mothers’ HPV-vaccination intention (Van Keulen et al., 2017; Pot et al., 2017). Besides, perceived susceptibility has been shown to be an important factor of preventive behavior according to health theory (e.g., the Health Belief Model; e.g., Schwarzer, 2001) and research (e.g., Brewer & Fazekas, 2007; Brewer et al., 2007). However, research findings are mixed about the surplus value of either using statistical or narrative risk information in affecting perceived susceptibility (for a review, see Winterbottom et al., 2008).
Therefore, this study is about an online experimental pretest of the effects of statistical versus narrative risk information on mothers’ perceptions of their daughters’ susceptibility towards HPV-related risks. Experimental pretesting enables researchers to draw reliable conclusions about the feasibility and effectiveness of developed materials in influencing predetermined objectives (Whittingham et al., 2008a; Whittingham et al., 2008b). Hence, this experimental pretest enables (a) drawing a reliable conclusion about which type of risk information is preferred when promoting mothers’ perceived daughters’ susceptibility towards HPV-related risks, and (b) making an evidence-based decision on which type of risk information to include in the intervention promoting HPV-vaccination acceptability.
Statistical risk information includes abstracts of numerical data and factual assertions (e.g., prevalence rates) about the probability of contracting a certain health-related outcome (De Wit et al., 2008). It is assumed that confrontation with the likelihood of contracting a risk, people are willing to take preventive actions. For instance, Cox and colleagues (2010) found that parents who received statistical information reported significantly stronger intentions to have their daughter vaccinated against HPV compared to parents who did not receive statistical information. However, there is also evidence for the impact of narratives. A narrative consists of a personal experience of an event, such as somebody getting a disease (Hinyard & Kreuter, 2007). Narratives are suggested to trigger the simulation heuristic (Kahnemann & Tyversky, 1982). According to this heuristic, the perceived likelihood of an event is based on the ease to picture the event mentally (Gregory, Cialdini, & Carpenter, 1982; Kahnemann & Tyversky, 1982). Narratives are suggested to increase this ease and, therefore, are likely to increase likelihood estimates (Tversky & Kahneman, 1974). De Wit and colleagues (2008) showed that among men who have sex with men, narrative risk information resulted in higher intentions to get the HBV-vaccination than information that only mentioned an increased risk of infection. Finally, Hopfer (2012) found that when college women were exposed to a hybrid message (i.e., a message containing both statistical and narrative information), their HPV-vaccination uptake was nearly double compared to when exposed to non-narrative controls (i.e., messages lacking personal decision narratives); 22% versus 12%, respectively. Nan and colleagues (2015) found that undergraduate students perceived their susceptibility towards HPV to be higher after reading a hybrid message compared to narrative-only and statistics-only messages.
Despite the lack of consensus about the surplus value of statistical versus narrative risk communication, research has pointed at conditions for effectiveness (e.g., Breakwell, 2000; Priester & Petty, 2003). Eagly and Chaiken (1993) pointed at factors moderating or mediating the effectiveness of persuasive (risk) communication. Higher comprehensibility (Eagly & Chaiken, 1993, p. 263) and credibility (Eagly & Chaiken, 1993, p. 352) of the message are suggested to improve message effects. In addition, information that is perceived as more novel, is suggested to increase the likelihood of attitude change (Wyer, 1974, p. 223). Therefore, we explore whether the effectiveness of type of risk communication is moderated by comprehensibility, credibility, and novelty of the message. In other words, we explore whether the effects of risk information depend on the information being believed as more or less comprehensible, credible, and/or novel.
Besides, factors constituting information processing can mediate the effects of risk communication, which is imaginability, elaboration, relevance, and defensive reactions. For instance, Janssen, Van Osch, De Vries, and Lechner (2013) reported that sunbed users who were exposed to narrative information could better imagine themselves contracting skin cancer. Models of attitude change, such as the Elaboration Likelihood Model (Petty & Cacioppo, 1986), state that message relevance increases active processing of information, which enhances message effects. However, a growing body of literature also suggests that messages about health risks may evoke defensive reactions such as message derogation (i.e., denying the relevance of the message or preventing it from reaching consciousness), denial of one’s susceptibility (i.e., defensive avoidance), and/or perceived manipulation (Van’t Riet & Ruiter, 2013; Block & Wiliams, 2002; Liberman & Chaiken, 1992; Ruiter, Verplanken, de Cremer, & Kok, 2004). Defensive reactions reduce message effects because it distracts the recipient’s attention to the message (cf. Kessels, Harris, Ruiter, & Klein, 2016). Therefore, the present study also explores whether the effects of risk information are mediated by imaginability, elaboration, relevance, and defensive reactions. In other words, we explore whether the effects of risk information can be explained by how mothers processed the information (e.g., by being perceived as more imaginable). See Figure 1 for an overview of the theoretical framework.
A schematic overview of the theoretical framework of perceived susceptibility in which the direct (a), moderation (b) and mediation effects are depicted (c).
Notes: The focus of this study can be found within the bold square (direct effects; a); variables within this study: X = predictor variable; Y = outcome variable; Z = moderator; M = mediator.
To summarize, this study is an online experimental pretest of materials developed for an intervention promoting HPV-vaccination acceptability targeting mothers of invited girls. Specifically, we aim to assess the effectiveness of statistical versus narrative risk information in influencing perceived daughters’ susceptibility towards HPV and cervical cancer. To our knowledge, the effects of risk information on the latter remain to be unexplored. Another novel aspect of this study is that it explores underlying conditions and mechanisms.
Participants were recruited by postal mail via Praeventis, the Dutch vaccination register. Praeventis has the name, sex, address and birth date of all Dutch children up until 18 years old. A computer program was used to draw a random sample of mothers of girls born in 2001 from Praeventis (i.e., girls to-be-invited for the HPV-vaccination round of 2014).
Mothers were randomly assigned by the computer to one of four conditions in a 2 (statistical information: yes vs no) × 2 (narrative information: yes vs no) between-subjects factorial design.
An invitational letter was sent to selected participants’ home addresses. This included information about the study, contact details of the researchers, a link to the website, and a personal entrance code. As a reimbursement, a subscription to a magazine was raffled among eligible participants. When participants logged in to the experiment, they were first informed about the study purpose, followed by instructions for completing the experiment. After having provided informed consent, participants were assigned to one of four conditions. All participants received general information about HPV, cervical cancer, and the HPV-vaccination. Next, mothers in the experimental conditions additionally received risk information (See ‘manipulations’). After this, participants proceeded with self-report questionnaires that accounted for the dependent variables, proposed mediators and moderators, and socio-demographics. Finally, mothers were debriefed and thanked for their participation. On average, mothers needed 15 minutes to complete the experiment.
Based on the literature (e.g., Mevissen et al., 2009; De Wit et al., 2008), we expected a medium effect size on primary outcomes (η2 = .06; Cohen, 1988). In our design with 2 means per factor, this effect size value corresponds to a d-value of 0.50, and an f-value of 0.25. Power analysis showed that a minimal sample size of 32 participants per condition was needed (with power = 0.80, and two-sided = 0.05; Hintze, 2004). We took into account an attrition rate of 30% and a response rate of 5%, based on previous experiences (Van Keulen et al., 2013). Consequently, 3,680 mothers were invited.
The experimental pretest was conducted between September and October 2013.
Socio-demographics included participants’ age, gender, educational level, country of birth, and religion. Level of education accounted for mothers’ highest level of education they had completed. This was classified as low (less than secondary or vocational education), intermediate (secondary through pre-university education), or high (professional or university education) (Van Keulen et al., 2013). Mothers were asked about their religious conviction (Protestant, Roman Catholic, Muslim, Jewish, Buddhist, Hindu, other, or no religion) (Van Keulen et al., 2013). Research has shown that more Dutch Protestants refrain from vaccination compared to the other (non) religious groups (Van Keulen et al., 2013). This was confirmed in our sample, hence religion was dichotomized into ‘Protestant’ versus ‘not Protestant’. Country of birth was dichotomized into ‘Netherlands’ versus ‘other’ as in our sample only 6% participants were born in a country other than the Netherlands.
Criteria for effectiveness. The outcome criteria were mothers’ perceived daughters’ susceptibility and mothers’ HPV-vaccination intention. Mothers’ perceived daughters’ susceptibility towards HPV and towards cervical cancer was measured by 7-point scaled item, each: ‘Without the HPV-vaccination, the chances of my daughter for getting infected with HPV (getting cervical cancer) are … ‘(1 = very small; to 7 = very big) (Pot et al., 2017; Van Keulen et al., 2013). HPV-vaccination intention was measured as a composite of two items: ‘Are you planning to have your daughter vaccinated against HPV?’ and ‘What are the chances that you will get your daughter vaccinated against HPV?’ (Pot et al., 2017; Van Keulen et al., 2013). Responses were given on a 7-point scale ranging from (1) certainly not/very small/to (7) certainly/very big (Pearson’s r = .95).
Items measuring the moderating and mediating factors (except for ‘defensive avoidance, see below) referred to the specific type(s) of information the mothers had received (i.e., general and/or statistical and/or narrative risk information). Thus, mothers evaluated each specific type of information separately. For mothers in the experimental conditions, a mean score per moderating and mediating factor had to be calculated. This means that, for instance, credibility for mothers in the statistical information condition represents the mean score of perceived credibility of the general and the statistical information.
Moderating factors. Comprehensibility was assessed by a composite of three items accounting for the extent to which mothers understood the information, and whether they considered the text to be clear and readable (Mevissen et al., 2009). Responses were given on a seven-point scale ranging from (1) not at all understandable/very unclear/very readable to (7) fully understandable/very clear/very readable (Cronbach’s alpha = .87). Credibility was measured with the single-item: ‘To what extent do you believe the content of the text is credible?’ (1 = very incredible; to 7 = very credible) (Mevissen et al., 2009). Novelty was measured with a single-item: ‘To what extent does the text contain new information for you?’ (1 = much new information; to 7 = much known information) (Mevissen et al., 2009).
Mediating factors. Imaginability was measured with one item for HPV and one item for cervical cancer: ‘Because of the text, I could imagine my daughter getting HPV/cervical cancer’ (1 = totally disagree; to 7 = totally agree) (Broemer, 2004; Janssen et al., 2013; Mevissen et al., 2010). Elaboration was assessed by a single item ‘How well did you read the information?’ (1 = not carefully at all; to 7 = very carefully) (Chaiken, 1980). Personal relevance was measured with the item ‘I considered the text to be personally relevant’ (1 = totally disagree; to 7 = totally agree) (Mevissen et al., 2009). Defensive avoidance was accounted for by two 7-point scaled items (one referring to HPV and the other to for cervical cancer): ‘I’d rather not think about my daughter getting infected with HPV/developing cervical cancer’ (1 = totally disagree; to 7 = totally agree) (Van’t Riet & Ruiter, 2013). Perceived manipulation of the information was measured by the item ‘I considered the text to manipulate my beliefs’ (1 = totally disagree; to 7 = totally agree) (Mevissen et al., 2009; Van’t Riet et al., 2010).
The text messages were pretested among a sample of mothers (N = 10) to assure comprehensibility, and, for the narrative message, to assure that mothers were able to imagine the situation.
General information (see Appendix A). All participants received general information about HPV, cervical cancer, and the HPV-vaccination, to ensure their background knowledge was similar.
Statistical risk information (see Appendix B). The statistical message was based on data from the Dutch National Institute for Public Health and Environment and data released by the Dutch Health Council (Gezondheidsraad, 2008). It presented factual epidemiological information about the prevalence of HPV and cervical cancer among women and information about the reduction in cervical cancer cases to be expected when all girls would obtain the HPV-vaccination. The latter was based on a study by Cox and colleagues (2010): “when no girl gets the HPV-vaccination, 600 women per year will get cervical cancer. However, when all girls get the HPV-vaccination, 300 women per year (half of 600) will be saved from getting cervical cancer”.
Narrative risk information (see Appendix C). The narrative message was based on a personal story derived from an Internet health forum of a woman who had found out she had cervical cancer, caused years ago by an HPV-infection. It was reported in the first person, because research has shown this to be more effective in influencing perceived susceptibility than the third person (Nan et al., 2015). We changed the name of the woman into ‘Anne’, because it is relatively common in The Netherlands, and not specific to individuals with certain backgrounds. She was 30 years old, as it typically takes 10–20 years for an HPV-infection to develop into cancer (Weinberg, 1999). It was emphasized that she got infected with HPV, despite having had an overall safe sex life (Manhart & Koutsky, 2002; Moscicki, 2005). To ensure mothers that the message concerns their daughters’ HPV-vaccination, Anne ended her story by noting: “With the HPV-vaccination that is available now, I sometimes can’t help but wonder: ‘what if I had gotten myself vaccinated against HPV 18 years ago…’”.
The statistical and narrative information were approximately of the same length (i.e., ±300 words).
A two-way analysis of variance (ANOVA) was conducted to examine the impact of the different types of risk information on perceived susceptibility and intention. Mediation and moderation analyses were examined using SPSS PROCESS macro (2.12.1 release), a widely used regression-based approach (Hayes, 2013).
For moderation analysis, the independent variable and the proposed moderator were entered in the model in the first step, followed by the interaction between the independent variable and the moderator in Step 2. The interaction (i.e., moderation) was assumed significant when p < .05. For mediation analyses, the PROCESS macro first conducted the following hierarchical regressions: the mediators (M) were regressed onto the independent variable (X) and the dependent variable (Y) was regressed onto both the independent variable and the mediator. We assumed mediation when the following conditions were met: (1) X has an effect on M, (2) M has an effect on Y when controlling for X, and (3) the effect of X on Y becomes significantly smaller or non-significant when controlling for M. Then, in order to test the indirect effect for significance, the bootstrapping method was utilized. The indirect effect represents the influence of the mediating variable on the original relation between X and Y. A value of 0 indicates no impact. A bias-corrected 95% bootstrap model with confidence intervals resampled 5,000 times for each analysis was used to judge significance (Shrout & Bolger, 2002); if the bootstrap confidence interval did not include zero, mediation was supported. This method was chosen as it is considered to maximize power and is robust against non-normality (Hayes, 2013).
The power calculation was conducted with PASS software (Hintze, 2004). The online experiment was developed using TailorBuilder © computer software. Main, moderation, and mediation analyses were conducted using SPSS version 22 (IBM Corp., 2013).
This experiment was part of a larger study, for which ethical approval was obtained from the ethical committee of the VU Medical Center in Amsterdam.
Of the invited individuals (n = 3,680), 467 initiated the study (13%). Inclusion criteria were that participants had to (1) provide informed consent online (excluded: n = 1), (2) fully complete the questionnaire (excluded: n = 72), and (3) be female (excluded: n = 19). This resulted in a final sample of 375 (11%) mothers.
The mean age of mothers was 43 years (SD = 4.05). Almost all mothers were born in the Netherlands (94%) and almost half of the mothers were high in education (42%). A relatively small percentage was Protestant (16%).
ANOVA revealed a significant main effect of statistical risk information on mothers’ perceived daughters’ susceptibility towards HPV (F(1, 371) = 7.56, p < .01). Mothers who received statistical information perceived their daughters’ susceptibility towards HPV to be higher (M = 4.11, SD = .10) than mothers who did not receive statistical information (M = 3.73, SD = .09). There was no main effect of narrative risk information on perceived daughters’ susceptibility towards HPV (F(1, 371) = .00, p = .97), nor was there a significant interaction-effect between statistical and narrative information on perceived daughters’ susceptibility towards HPV (F(1, 371) = 1.54, p = .22). No significant main or interaction effects were found on mothers’ perception of their daughters’ susceptibility towards cervical cancer (F’s < 2.10; p’s > .15) or on HPV-vaccination intention (F’s < .83; p’s > 36). See Tables 1, 2, 3 for the mean scores on the outcomes per condition.
Table 1
Mean scores (SD) on susceptibility towards HPV (1–7; N = 375).
| Statistical risk information | No statistical risk information | |
|---|---|---|
| Narrative information | 4.02 (1.28)a,b | 3.82 (1.23)a,b |
| No narrative information | 4.19 (1.43)a | 3.65 (1.26)b |
Notes: A higher score represents a higher perceived daughters’ susceptibility towards HPV. Values with different superscripts differ significantly (p < .05).
Table 2
Mean scores (SD) on susceptibility towards cervical cancer (1–7; N = 375).
| Statistical risk information | No statistical risk information | |
|---|---|---|
| Narrative information | 3.65 (1.18) | 3.81 (1.17) |
| No narrative information | 3.79 (1.31) | 3.60 (1.08) |
Notes: A higher score represents a higher perceived daughters’ susceptibility towards cervical cancer. Differences between the groups were nonsignificant.
Table 3
Mean scores (SD) on HPV-vaccination intention (1–7; N = 375).
| Statistical risk information | No statistical risk information | |
|---|---|---|
| Narrative information | 6.01 (1.42) | 5.92 (1.42) |
| No narrative information | 5.76 (1.62) | 5.88 (1.56) |
Notes: A higher score represents a higher HPV-vaccination intention. Differences between the groups were nonsignificant.
Moderation and mediation analyses were performed using statistical risk information (yes versus no) as the independent variable and mothers’ perception of their daughters’ susceptibility towards HPV as the dependent variable. No significant interaction effect was found with any of the proposed moderators (comprehensibility, credibility and novelty; p’s > .05). Also, we didn’t find any support for an indirect impact of statistical risk information on mothers’ perception of their daughters’ susceptibility towards HPV via the proposed mediators (i.e., imaginability, elaboration, perceived personal relevance, defensive avoidance and perceived manipulation). Overall, mean scores on the moderators and mediators were high (except for defensive avoidance and perceived manipulation). Differences between the conditions were small and nonsignificant (See Table 4).
Table 4
Mean scores (SD) on moderators and mediators of those who did versus those who did not receive statistical risk information (N = 375).
| No Statistical Risk Information (n = 191) | Statistical Risk Information (n = 184) | |
|---|---|---|
| Proposed moderators* | ||
| Comprehensibility (1–7) | 5.83 (.91) | 5.80 (.86) |
| Credibility (1–7) | 5.60 (1.11) | 5.58 (.99) |
| Novelty (1–7) | 4.70 (1.40) | 4.92 (1.20) |
| Proposed mediators* | ||
| Imaginability HPV (1–7) | 5.76 (.91) | 5.74 (.81) |
| Imaginability cervical cancer (1–7) | 5.59 (1.04) | 5.63 (.90) |
| Elaboration (1–7) | 5.93 (.97) | 5.86 (.84) |
| Perceived personal relevance (1–7) | 5.24 (1.10) | 5.42 (1.04) |
| Defensive avoidance HPV (1–7) | 3.65 (2.02) | 3.42 (2.00) |
| Defensive avoidance cervical cancer (1–7) | 3.68 (2.06) | 3.64 (2.10) |
| Perceived manipulation (1–7) | 3.70 (1.64) | 4.09 (1.50) |
Notes: *A higher score represents a higher score on the moderator/mediator. Differences between the groups were nonsignificant.
This study was an experimental pretest of materials targeting mothers of invited girls developed for a web-based, tailored intervention promoting HPV-vaccination acceptability. We tested the effectiveness of statistical versus narrative risk information on affecting how mothers’ perceived their daughters’ susceptibility towards HPV and cervical cancer. Results showed that statistical information increased mothers’ perceptions of their daughters’ susceptibility towards HPV. Such a positive effect was also found by Mevissen and colleagues (2009), within the context of Chlamydia: participants felt more susceptible to a Chlamydia infection after they read statistical risk information compared to those exposed to narrative or no risk information. These results made us decide to include statistical information as the mode of communication about HPV-related risks in the web-based intervention.
We did not find statistical risk information to affect mothers’ perceptions of their daughters’ susceptibility towards getting cervical cancer. This might be related to the actual difference in the likelihood of contracting HPV (high probability), relative to contracting cervical cancer (low probability). In the present study, we found that mothers who received no risk information highly overestimated the probability of their daughter contracting cervical cancer when looking at the actual incidence in the Netherlands (Integraal Kankercentrum Nederland, 2013). It seems unlikely that exposure to statistical information –presenting the actual low probability rate– would increase mothers’ perceptions of the daughters’ susceptibility for contracting cervical cancer.
Mothers who received no risk information perceived the chance of their daughter contracting HPV or cervical cancer to be almost similar, while in fact, the chance of contracting HPV is much higher (Integraal Kankercentrum Nederland, 2013). It seemed that mothers underestimated the risk for HPV, and overestimated the risk for cervical cancer. The tendency to overestimate small frequencies and underestimate larger ones is well known (Fischhoff, Bostrom & Quadrell, 1993; Lichtenstein et al., 1978). Correcting mothers’ misperceptions about the risks of their daughter contracting HPV and cervical cancer, would provide a more realistic and bases for initiating processes of informed decision-making.
In this study, narrative risk information had no effect on how mothers perceived their daughters’ susceptibility towards HPV or cervical cancer. We hypothesized narratives to have an effect by increasing the ease of imagination (Kahnemann & Tyversky, 1982). However, secondary analysis indicated that mothers could imagine their daughter contracting both HPV and cervical cancer less easily after reading the narrative compared to the statistics. The absence of an effect of narrative information may also have been associated with difficulty for mothers to identify their adolescent daughter with the 30-year-old woman getting cervical cancer. For narrative risk information to be effective, it is important that one can identify him- or herself with the narrative (Cohen, 2001; de Graaf et al., 2012). Difficulties with mothers’ identification of their daughter may also have been associated with the relatively young age at which the character had sex for the first time (i.e., 16 years old). In 2017, 28% of Dutch girls aged 15–17 has had sex for the first time, compared to 67% of girls aged 18–20 (De Graaf et al., 2017). Finally, mothers may not have been ‘transported’ into the narrative enough. Transportation into narrative worlds, or immersion into a story, is a primary mechanism of narrative persuasion (Green & Brock, 2000). Unfortunately, we did not assess the degree of mothers’ identification of the daughter with the story’s character nor did we measure the extent of transportation into the narrative.
Although mothers who received statistical information felt their daughters were more susceptible towards HPV, this did not result in a higher HPV-vaccination intention. An explanation comes from social cognitive models of health behavior (e.g., the Health Belief Model and the Health Action Process Approach; Schwarzer, 2001, 2008) and earlier research (Pot et al., 2017; Van Keulen et al. 2013), which indicate that perceived daughters’ susceptibility is not the sole determinant of mothers’ intention to have their daughter vaccinated against HPV. Thus, to influence intention, communication should as well target other relevant determinants of HPV-vaccination uptake (e.g., attitudinal beliefs) (Pot et al., 2017; Van Keulen et al., 2013). In the web-based intervention, other relevant determinants of HPV-vaccination uptake were also targeted.
The effect of statistical risk information on mothers’ perceptions of their daughters’ susceptibility towards HPV was not moderated by message characteristics (i.e., comprehensibility, credibility and novelty), nor was it mediated by factors constituting information processing (i.e., imaginability, elaboration, perceived personal relevance, and defensive reactions). This indicates that, in this study, statistical risk information increased mothers’ perceptions of their daughters’ susceptibility towards HPV irrespective of the information being believed as more or less comprehensible, credible or novel. Moreover, the increase in mothers’ perceptions of their daughters’ susceptibility towards HPV could not be explained by how mothers processed the statistical information (e.g., by being perceived as more imaginable). This absence of effects may be related to the fact that we were forced to compute a mean for moderating and mediating factors when mothers received more than one type of information. Thus, for the combination condition, the evaluations of three pieces of risk information (i.e., the general, statistical and narrative information) were averaged. Moderation and mediation was tested between the subgroups ‘statistical risk information’ (i.e., statistical and combination condition) and ‘no statistical risk information’ (i.e., no risk information and narrative risk information condition). It could well be that actual differences between the statistical versus narrative information were diluted because of this averaging over the separate pieces of risk information. An indication for this comes from secondary analyses, in which we only compared the statistical (n = 102) and narrative condition (n = 89) (i.e., the general and combination condition were excluded) on the moderator and mediator evaluations of the experimental information only (i.e., the evaluations of the general information was excluded). This way, the scores for moderating and mediating factors did not have to be averaged. Mean scores on moderating and mediating factors differed significantly between the two conditions for the moderator comprehensibility (i.e., mothers understood the narrative better) and for the mediators imaginability (i.e., mothers could imagine their daughter contracting both HPV and cervical cancer better after reading statistics) and personal relevance (i.e., mothers judged the statistics as more relevant). In addition, we found personal relevance to mediate the effect of statistical information on mothers’ perceived daughters’ susceptibility towards HPV: statistical information was judged more personal relevant, which increased mothers’ perceptions of their daughters’ susceptibility towards HPV. This is in line with the Elaboration Likelihood Model (Petty & Cacioppo, 1986), which suggests that message relevance increases active processing of messages, which in turn enhances message effectiveness if persuasive arguments are perceived as strong. Moreover, this implicates that statistical risk information may be particularly effective for persons who perceive the information as being personally relevant. Hence, statistical messages should be designed in a way that is personally relevant to the message recipient. This can be done, for instance, by tailoring the message to the recipient, as tailored messages have shown to improve personal relevance (Kreuter & Wray, 2003). Finally, this raises the question whether the effects of statistical information would be even stronger in influencing one’s own perceived susceptibility instead of someone else’s susceptibility. Future studies could include dyads (mother/father-daughter) to examine this.
This study had several limitations worth mentioning. First, we did not include identification nor transportation as mediators, which could have helped in disentangling why the narrative used in this study, was not effective since these factors have been shown to be important effect conditions for narratives (Cohen et al., 2001; de Graaf et al., 2017). Second, we cannot exclude that the general information affected mothers’ perceived daughters’ susceptibility. However, general information is an essential element in communication about the HPV-vaccination, as sufficient knowledge is a prerequisite for an informed decision (Marteau, Dormandy & Michie 2001). Third, the use of single-item measures represents a limitation. These are suggested to be less stable, reliable, and precise compared to multi-item measures (Bowling, 2005). The questionnaire used for this experiment included additional questions about design features of the developed intervention. In order to reduce the time burden for participants, single-items were used. Finally, response rate was low (11%). Although it was higher than found in a previous study which recruited mothers through Praeventis (Van Keulen et al., 2013), the low response rate could be explained by the use of a written invitation letter. Research has shown written invitations to result in lower response rates compared to e-mail invitations in studies using online questionnaires (Manfreda et al., 2008). Due to the low response rate, caution is needed in generalizing the study results.
We recommend future research to investigate how to best present statistical risk information to maximize its effectiveness. For instance, Cox and colleagues (2010) found that parents who viewed a graphic presentation of statistical information had a higher HPV-vaccination intention compared to parents who viewed a nongraphic presentation. Furthermore, we recommend future research to investigate how to make narratives about HPV-related risks more effective, as we still believe narratives to be a promising method in communicating about HPV-related risks (e.g., De Wit et al., 2008; Janssen et al., 2013; Mevissen et al., 2012). To improve effectiveness, future research could use strategies such as multiple narrative messages (Mevissen et al., 2010), self-construed narratives (Mevissen et al., 2012), and tailoring the narrative to the recipients characteristics (Kreuter et al., 2007).
Statistical risk information was most effective in increasing mothers’ perceived daughters’ susceptibility towards HPV and was therefore implemented in the web-based, tailored intervention promoting HPV-vaccination acceptability. Future research is needed on (a) how to best present statistical risk information and (b) how to make narratives about HPV-related risks more effective.
The additional files for this article can be found as follows:
Appendix AGeneral information. DOI: https://doi.org/10.5334/hpb.7.s1
Appendix BStatistical risk information. DOI: https://doi.org/10.5334/hpb.7.s1
Appendix CNarrative risk information. DOI: https://doi.org/10.5334/hpb.7.s1
We would like to thank Petra Oomen (RIVM) for her help with data collection.
The authors have no competing interests to declare.
Alberts, C., Schim van der Loeff, M. F., Hazeveld, Y., de Melker, H. E., van der Wal, M. F., Nielen, A., El Fakiri, F., Prins, M., & Paulussen, T. H. G. W. M. (2017). A longitudinal study on determinants of HPV vaccination uptake in parents/guardians from different ethnic backgrounds in Amsterdam, the Netherlands. BMC Public Health, 17, 220. DOI: https://doi.org/10.1186/s12889-017-4091-4
Block, L. G., & Williams, P. (2002). Undoing the effects of seizing and freezing: decreasing defensive processing of personally relevant messages. Journal of Applied Social Psychology, 32(4), 803–833. DOI: https://doi.org/10.1111/j.1559-1816.2002.tb00243.x
Bowling, A. (2005). Just one question: if one question works, why ask several? Journal of Epidemiology and Community Health, 59(5), 342–345. DOI: https://doi.org/10.1136/jech.2004.021204
Breakwell, G. M. (2000). Risk communication: Factors affecting impact. British Medical Bulletin, 56(1), 110–120. DOI: https://doi.org/10.1258/0007142001902824
Brewer, N. T., Chapman, G. B., Gibbons, F. X., Gerrard, M., & McCaul, K. D. (2007). Meta-analysis of the relationship between risk perception and health behavior: The example of vaccination. Health Psychology, 26, 136–145. DOI: https://doi.org/10.1037/0278-6133.26.2.136
Brewer, N. T., & Fazekas, K. I. (2007). Predictors of HPV vaccine acceptability: a theory-informed, systematic review. Preventive Medicine, 45(2–3), 107–114. DOI: https://doi.org/10.1016/j.ypmed.2007.05.013
Broemer, P. (2004). Ease of imagination moderates reactions to differently framed health messages. European Journal of Social Psychology, 34, 103–119. DOI: https://doi.org/10.1002/ejsp.185
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, 752–766. DOI: https://doi.org/10.1037/0022-3514.39.5.752
Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ: Lawrence Earlbaum Associates.
Cohen, J. (2001). Defining identification: A theoretical look at the identification of audiences with media characters. Mass Communication and Society, 4, 245–264. DOI: https://doi.org/10.1207/S15327825MCS0403_01
Cox, D. S., Cox, A. D., Sturm, L., & Zimet, G. (2010). Behavioral interventions to increase HPV vaccination acceptability among mothers of young girls. Health Psychology, 29(1), 29–39. DOI: https://doi.org/10.1037/a0016942
De Graaf, A., Hoeken, H., Sanders, J., & Beentjes, J. W. J. (2012). Identification as a mechanism of narrative persuasion. Communication Research, 39, 802–823. DOI: https://doi.org/10.1177/0093650211408594
De Graaf, H., Van der Borne, M., Nikkelen, S., Twisk, D., & Meijer, S. (2017). Seks onder je 25e. Seksuele gezondheid van jongeren in Nederland anno 2017/Sex under the age of 25. A population study of Dutch adolescent’s sexual health. Delft: Eburon.
De Wit, J. B. F., Das, E., & Vet, R. (2008). What works best: Objective statistics or a personal testimonial? An assessment of the persuasive effects of different types of message evidence on risk perception. Health Psychology, 27, 110–115. DOI: https://doi.org/10.1037/0278-6133.27.1.110
Eagly, A. H., & Chaiken, S. (1993). The psychology of attitudes. Fort Worth, Texas: Harcourt Brace Jovanovich.
Fischhoff, B., Bostrom, A., & Quadrell, M. J. (1993). Risk perception and communication. Annual Review of Public Health, 4, 183–203. DOI: https://doi.org/10.1146/annurev.pu.14.050193.001151
Gezondheidsraad. (2008). Vaccinatie tegen baarmoederhalskanker, 154. Den Haag: Gezondheidsraad.
Green, M. C., & Brock, T. C. (2000). The role of transportation in the persuasiveness of public narratives. Journal of Personality and Social Psychology, 79, 701–721. DOI: https://doi.org/10.1037/0022-3514.79.5.701
Gregory, W. L., Cialdini, R. B., & Carpenter, K. M. (1982). Self-relevant scenarios as mediators of likelihood estimates and compliance: Does imagining make it so? Journal of Personality and Social Psychology, 43, 89–99. DOI: https://doi.org/10.1037/0022-3514.43.1.89
Hayes, A. (2013). Introduction to mediation, moderation, and conditional process analysis: A regression-based approach. New York, NY: The Guilford Press.
Hintze, J. 2004. NCSS and PASS. Number cruncher statistical systems. Kaysville, Utah. Available at: www.ncss.com.
Hinyard, L., & Kreuter, M. W. (2007). Using narrative communication as a tool for health behavior change: A conceptual, theoretical, and empirical overview. Health Education & Behavior, 34, 777–792. DOI: https://doi.org/10.1177/1090198106291963
Hofman, R., van Empelen, P., Richardus, J. H., de Kok, I. M. C. M., de Koning, H. J., van Ballegooijen, M., et al. (2014). Predictors of HPV vaccination uptake: a longitudinal study among parents. Health Education Research, 29(1), 83–96. DOI: https://doi.org/10.1093/her/cyt092
Hopfer, S. (2012). Effects of a narrative HPV vaccination intervention aimed at reaching college women: A randomized controlled trial. Prevention Science, 13, 173–182. DOI: https://doi.org/10.1007/s11121-011-0254-1
IBM Corp. Released 2013. IBM SPSS Statistics for Windows, Version 22.0. Armonk, NY: IBM Corp.
Integraal Kankercentrum Nederland. (2013). Cijfers over kanker. Retrieved from: http://www.cijfersoverkanker.nl/.
Janssen, E., van Osch, L., de Vries, H., & Lechner, L. (2013). The influence of narrative risk communication on feelings of cancer risk. British Journal of Health Psychology, 18, 407–419. DOI: https://doi.org/10.1111/j.2044-8287.2012.02098.x
Kahnemann, D., & Tversky, A. (1982). The simulation heuristic. In: Kahnemann, D., Slovic, P., & Tversky, A. (eds.), Judgment under uncertainty: Heuristics and biases, 201–208. Cambridge, MA: Cambridge University Press. DOI: https://doi.org/10.1017/CBO9780511809477.015
Kessels, L. T. E., Harris, P. R., Ruiter, R. A. C., & Klein, W. M. P. (2016). Attentional effects of self-affirmation in response to graphic antismoking images. Health Psychology, 35(8), 891–897. DOI: https://doi.org/10.1037/hea0000366
Kreuter, M. W., Green, M. C., Cappella, J. N., Slater, M. D., Wise, M. E., Storey, D., et al. (2007). Narrative communication in cancer prevention and control: a framework to guide research and application. Annals of Behavioral Medicine, 33(3), 221–235. DOI: https://doi.org/10.1007/BF02879904
Kreuter, M. W., & Wray, R. J. (2003). Tailored and targeted health communication: Strategies for enhancing information relevance. American Journal of Health Behavior, 27, S227–S232. DOI: https://doi.org/10.5993/AJHB.27.1.s3.6
Liberman, A., & Chaiken, S. (1992). Defensive processing of personally relevant health messages. Personality and Social Psychology Bulletin, 18(6), 669–679. DOI: https://doi.org/10.1177/0146167292186002
Lichtenstein, S., Slovic, P., Fischhoff, B., Layman, M., & Combs, B. (1978). Judged frequency of lethal events. Journal of Experimental Psychology: Human Learning and Memory, 4(6), 551–578. DOI: https://doi.org/10.1037/0278-7393.4.6.551
Lier, E. A., Geraedts, J. L. E., Oomen, P. J., Giesbers, H., van Vliet, J. A., Drijfhout, I. H., de Melker, H. E., et al. (2018). Vaccinatiegraad en jaarverslag Rijksvaccinatieprogramma Nederland 2017. Retrieved from: https://www.rivm.nl/dsresource?objectid=30c7c6ab-197d-44a0-a901-f9719f916bf9&type=pdf&disposition=inline.
Manfreda, K. L., Bosnjak, M., Berzelak, J., Haas, I., & Vehovar, V. (2008). Web surveys versus other survey modes: A meta-analysis comparing response rates. International Journal of Market Research, 50(1), 79–104. DOI: https://doi.org/10.1177/147078530805000107
Manhart, L. E., & Koutsky, L. A. (2002). Do condoms prevent genital HPV infection, external genital warts, or cervical neoplasia? A meta-analysis. Sexually Transmitted Diseases, 29(11), 725–35. DOI: https://doi.org/10.1097/00007435-200211000-00018
Marteau, T. M., Dormandy, E., & Michie, S. (2001). A measure of informed choice. Health Expectations, 4, 99–108. DOI: https://doi.org/10.1046/j.1369-6513.2001.00140.x
Mevissen, F. E. F., Meertens, R. M., & Ruiter, R. A. C. (2012). Bedtime stories: The effects of self-constructed risk scenarios on imaginability and perceived susceptibility to sexually transmitted infections. Psychology & Health, 27, 1036–1047. DOI: https://doi.org/10.1080/08870446.2011.648935
Mevissen, F. E. F., Meertens, R. M., Ruiter, R. A. C., Feenstra, H., & Schaalma, H. (2009). HIV/STI Risk communication: The effects of scenario-based risk information and frequency-based risk information on perceived susceptibility to Chlamydia and HIV. Journal of Health Psychology, 14, 78–87. DOI: https://doi.org/10.1177/1359105308097948
Mevissen, F. E. F., Ruiter, R. A. C., Meertens, R. M., & Schaalma, H. P. (2010). The Effects of Scenario-Based Risk Information on Perceptions of Susceptibility to Chlamydia and HIV. Psychology & Health, 25, 1161–1174. DOI: https://doi.org/10.1080/08870440903055901
Moscicki, A. B. (2005). Impact of HPV infection in adolescent populations. Journal of Adolescent Health, 37, S3–9. DOI: https://doi.org/10.1016/j.jadohealth.2005.09.011
Nan, X., Dahlstrom, M. F., Richards, A., & Rangarajan, S. (2015). Influence of evidence type and narrative type on HPV risk perception and intention to obtain the HPV vaccine. Health Communication, 30, 301–308. DOI: https://doi.org/10.1080/10410236.2014.888629
Petty, R. E., & Cacioppo, J. T. (1986). The elaboration likelihood model of persuasion. In: Berkowitz, L. (ed.), Advances in experimental social psychology, 19, 123–205. Orlando, Florida: Academic Press.
Pot, M., Ruiter, R. A. C., Paulussen, T. W. G. M., Heuvelink, A., de Melker, H. E., van Vliet, H. J. A., & van Keulen, H. M. (2018). Systematically Developing a Web-Based Tailored Intervention Promoting HPV-Vaccination Acceptability Among Mothers of Invited Girls Using Intervention Mapping. Front Public Health, 6, 226. DOI: https://doi.org/10.3389/fpubh.2018.00226
Pot, M., van Keulen, H. M., Ruiter, R. A. C., Eekhout, I., Mollema, L., & Paulussen, T. W. G. M. (2017). Motivational and contextual determinants of HPV-vaccination uptake: A longitudinal study among mothers of girls invited for the HPV-vaccination. Preventive Medicine, 100, 41–49. DOI: https://doi.org/10.1016/j.ypmed.2017.04.005
Priester, J. R., & Petty, R. E. (2003). The influence of spokesperson trustworthiness on message elaboration, attitude strength, and advertising effectiveness. Journal of Consumer Psychology, 13, 408–421. DOI: https://doi.org/10.1207/S15327663JCP1304_08
Ruiter, R. A. C., Verplanken, B., de Cremer, D., & Kok, G. (2004). Danger control and fear control in response to fear appeals: the role of need for cognition. Basic and Applied Social Psychology, 26, 13–24. DOI: https://doi.org/10.1207/s15324834basp2601_2
Schwarzer, R. (2001). Social-cognitive factors in changing health-related behavior. Current Directions in Psychological Science, 10, 47–51. DOI: https://doi.org/10.1111/1467-8721.00112
Schwarzer, R. (2008). Modeling health behavior change: How to predict and modify the adoption and maintenance of health behaviors. Applied Psychology, 57, 1–29. DOI: https://doi.org/10.1111/j.1464-0597.2007.00325.x
Shrout, P. E., & Bolger, N. (2002). Mediation in experimental and nonexperimental studies: new procedures and recommendations. Psychological Methods, 7, 422–445. DOI: https://doi.org/10.1037//1082-989X.7.4.422
Tversky, A., & Kahneman, D. (1974). Judgment under uncertainty: Heuristics and biases. Science, 185, 1124–1131. DOI: https://doi.org/10.1126/science.185.4157.1124
van Keulen, H. M., Otten, W., Ruiter, R. A. C., Fekkes, M., van Steenbergen, J., Dusseldorp, E., & Paulussen, T. W. G. M. (2013). Determinants of HPV vaccination intentions among Dutch girls and their mothers: A cross-sectional study. BMC Public Health, 13, 111. DOI: https://doi.org/10.1186/1471-2458-13-111
Van’t Riet, J., & Ruiter, R. A. C. (2013). Defensive reactions to health-promoting information: An overview and implications for future research. Health Psychology Review, 7, S104–S136. DOI: https://doi.org/10.1080/17437199.2011.606782
Weinberg, R. A. (1999). One renegade cell: how cancer begins (1st ed.). New York: Basic Books.
Whittingham, J., Ruiter, R. A. C., Zimbile, F., & Kok, G. (2008a). Experimental pretesting of public health campaigns: A case study. Journal of Health Communication, 13, 216–229. DOI: https://doi.org/10.1080/10810730701854045
Whittingham, J. R., Ruiter, R. A. C., Castermans, D., Huiberts, A., & Kok, G. (2008b). Designing effective health education materials: experimental pre-testing of a theory-based brochure to increase knowledge. Health Education Research, 23, 414–426. DOI: https://doi.org/10.1093/her/cym018
Winterbottom, A., Bekker, H., Conner, M., & Mooney, A. (2008). Does narrative information bias individual’s decision making? A systematic review. Social Science & Medicine, 67, 2079–88. DOI: https://doi.org/10.1016/j.socscimed.2008.09.037
Wyer, R. S. (1974). Cognitive organization and change: An information-processing approach, 223. Hillsdale, NJ: Lawrence Erlbaum Associates.
Health Psychology Bulletin has blind peer review, which is unblinded upon article acceptance. The editorial history of this article can be downloaded here:
PR 1Editorial History. DOI: https://doi.org/10.5334/hpb.7.pr1