• No results found

Effect of social comparisons on risk comprehension and behavior : the effect of social comparative information on cardiovascular risk comprehension and behavioral intentions

N/A
N/A
Protected

Academic year: 2021

Share "Effect of social comparisons on risk comprehension and behavior : the effect of social comparative information on cardiovascular risk comprehension and behavioral intentions"

Copied!
29
0
0

Bezig met laden.... (Bekijk nu de volledige tekst)

Hele tekst

(1)

The Effect of Social Comparative Information on Cardiovascular Risk Comprehension and Behavioral Intentions

Amber Elisabeth van der Meij, 10543457

Psychologie van Gezondheidsgedrag, Universiteit van Amsterdam

Supervisors: Olga Damman / Nina Bogaerts

(2)

Abstract

Research regarding comparative risk information mainly focused on the effects on screening behaviors. The aim of the present study is to investigate the effect of social comparative risk information on affective appraisals, cardiovascular risk comprehension and behavioral

intentions. Healthy participants between 45 and 69 years (N=152) were recruited via an online access panel and were randomly assigned to one of the risk comparison conditions: (1)

comparison with a person without risk factors, (2) comparison with an average person, and (3) no comparison. Participants looked at hypothetical cardiovascular risk information and while they looked at it they answered questions that assessed whether they correctly extracted the numerical information. Outcome variables were negative affect, worry, risk comprehension (including meaningful interpretations and risk perception) and behavioral intentions. Analyses of variance and mediation analyses were conducted to assess the main effects of our

manipulation on these outcome variables. Only 65% of the 152 participants correctly extracted the numbers from the risk information. Analysis of the complete sample showed that the manipulation only had an effect on people’s self-reported risk comprehension. Analysis without the people who were not able to correctly extract the numerical information revealed a significant effect on people’s self-reported risk comprehension and negative affect, worry and risk perception. Providing social comparison anchors, especially the comparison with a person without risk factors, may benefit informed decision making by improving people’s risk comprehension via worry. However, difficulties remain for people with both low numeracy and low graph literacy.

(3)

The Effect of Social Comparative Information on Cardiovascular Risk Comprehension and Behavioral Intentions

Informed decision making is becoming more important in healthcare and people more often make their own medical decisions based on information they receive (Hibbard & Peters, 2003; Kon, 2010; O’Connor et al., 2007). A lot of information regarding diseases and risk factors can be found on the internet and people often look for information about diseases (Cline & Haynes, 2001; Ybarra, & Suman, 2006). There are also various online tests to calculate one’s personal risk of getting a particular disease. In the Netherlands, an important example is the PreventieConsult risk test for calculating people’s disease risk of

cardiovascular disease, diabetes and chronic kidney damage (www.testuwrisico.nl) (Alssema et al., 2012), developed by general practitioners, occupational physicians, the Dutch Heart Foundation (Hartstichting), the Diabetes Research Foundation (Diabetes Fonds) and the Dutch Kidney Foundation (Nierstichting). The online risk test computes a personalized risk based on seven questions about age, gender, smoking behavior, BMI, waist circumference and family history of cardiovascular disease and type 2 diabetes. This risk is presented as an absolute risk number (percentage and natural frequency), accompanied by a verbal label (no elevated risk, slightly elevated risk and elevated risk) and a bar chart. Finally, a personalized advice is given; people with an elevated risk are advised to visit their GP and people with a slightly elevated risk are advised to change their lifestyle.

An important problem with this kind of information (especially the numerical

information) is that most people find it difficult to understand (Gigerenzer et al., 2007; Reyna et al., 2009). A recent study at the VUmc showed that the current numerical information from the PreventieConsult risk test means little to people, which may cause that they do not accept the information (Damman, Bogaerts, Van Dongen & Timmermans, in preparation). The study revealed that people had difficulty determining whether a risk was low or high. Some of them

(4)

were not worried by a risk of 40%, because in their opinion, it could have been worse.

Apparently they did not have a good reference point for judging the severity of their risk. The bar graph’s scale of 0-100% may have led to people considering a risk of 40% as relatively low. Additionally, the information in the risk test in its current form barely caused affective appraisals.

In practice, risk information is mainly aimed at cognitive processing of information. because it is assumed that awareness of the risk leads to healthy choices and behavior. The majority of health communications therefore focuses on evoking analytic processing (Hibbard & Peters, 2003) and creating adequate cognitive appraisals (“understanding of the facts”) in order to promote health behaviors or to help people make an informed decision (Rimer, Briss, Zeller, Chan & Woolf, 2004). However, most people not strictly follow the steps described in these models, but often construct their preferences and ideas “on the spot” while viewing information, and in doing so use both reason and emotion (Damasio, 1994; Peters, Lipkus & Diefenbach, 2006). Consequently, the way a risk is presented can have a lot of influence on people’s appraisals. In assessing risk information, people simultaneously use affective and cognitive appraisal processes (Visschers et al., 2012) and studies have indeed showed that affective processes also play an important role in risk perception (Keller, Siegrist & Gutscher, 2006; Loewenstein Weber, Hsee & Welch, 2001). This process is more generally described in dual process models of information processing, in which affective information processing is grouped under the experiential system (Slovic, Peters, Finucane & MacGregor, 2004). Since emotions also play an important role in decision making and since numerical information itself often means little to people, it may be necessary to more closely serve affective information processing by adding so-called affective cues to attain a better comprehension of the risk information among consumers. When considering affect, it is

(5)

important to make a distinction between the different types of affect and their functions as described in the literature.

Peters, Lipkus and Diefenbach (2006) made a distinction between integral affect (“positive and negative feelings about a stimulus that are generally based on prior experiences and thoughts and are experienced while considering the stimulus”) and incidental affect (“positive and negative feelings such as mood states that are independent of a stimulus but can be misattributed to it or can influence decision processes”). According to the authors, (a combination of integral and incidental) affect has several functions in health communication. For example, affect can serve as information for evaluating risk estimates. A certain stimulus (e.g. risk information) induces certain associative feelings that people use to form a judgment or make a decision regarding the risk. Affect as information can also reflect concerns or worries that people have about getting a certain disease that they attribute to the stimulus. Generally, the more people worry about a certain disease, the higher they estimate the chance of getting that disease (Lipkus et al., 2000). According to Peters and colleagues, affect can also serve as a motivator to take action or to process (more) information. The affect that a stimulus evokes could lead to tendencies to avoid certain behaviors or to engage in them. Worry, for example, has been often associated with screening behavior and associations between affect en the amount of effort people put into making the best decision have also been found (Peters, Slovic & Gregory, 2003). Peters and colleagues suggest that this would also apply to people who experience (strong) affect with regard to a certain disease, making them more motivated to find and process information about treatment options (Peters et al, 2003). Figure 1 shows possible ways in which affect can give meaning to information or influence information seeking and other behavioral intentions.

(6)

Figure 1. Possible functions of affect. Social Comparative Information

Affective cues can be used in health communication and also in designing risk information formats. For example, information can be presented as a personal narrative (Bekker et al., 2012) or by using evaluative labels for choices (Peters et al., 2009). Such cues can give affective meaning to numbers when such numbers are not very informative to people and in that way facilitate information processing (Slovic, Peters, Finucane & MacGregor, 2005). Another affective cue that can be used is a social comparison anchor (e.g. comparing the risk to the risk of an average person; also called comparative risk information). Previous research has shown that providing people with an anchor value can influence their subsequent estimations (Jacowitz & Kahneman, 1995; Tversky & Kahneman, 1974). However, these studies have mostly focused on cognitive appraisals, while the use of social comparison anchors probably also influences affective appraisals because of the social element of the anchor. Presenting such a social comparison anchor may lead to quick associative feelings about the goodness or badness of a test result (integral affect) and whether it is worrisome or not. People have a natural tendency to evaluate their abilities (Festinger, 1954) and therefore, social comparison anchors can be seen as affective cues. Previous research shows that they do not only use objective information to evaluate their abilities or situation, but that they also make social comparisons to others in their environment (French, 2004; Klein, 1997; Klein, 2003). According to Festinger’s similarity hypothesis (1954), people prefer to compare themselves to somebody who is a lot like them; after all, such a comparison provides them

(7)

with the most relevant information. For social comparison to occur, people need to be able to identify with the comparison object. When people then make an upward comparison (i.e. with superior others), this usually induces negative affect (Schwarz, 2000), such as anger, fear, shame, worry and guilt (Watson & Tellegen, 1985). Upward social comparisons only evoke negative affect when the dimension of the comparison, in this case the result of a risk test, is relevant for the person himself (Tesser, 1988).

Until now, these different types of comparative risk information based on social comparison have not been systematically studied regarding disease risk information. Specifically the role of affect evoked by these comparisons has not yet been studied systematically. Previous research into comparative risk information in risk communication mainly focused on the effects of giving comparative information on influencing behavior and behavioral intentions, compared with absolute risk information (e.g. a risk percentage). For example, it has been investigated to what extent comparative information encourages people to engage in screening behaviors or to seek treatment (Fagerlin, Zikmund-Fisher & Ubel, 2007; Lipkus & Klein, 2006). For example, women’s intentions of seeking preventive treatment increased when people were told that their risk of getting breast cancer was above average, independent of their absolute risk numbers (Fagerlin et al., 2007). Informing people about their above average risk can therefore be used to persuade them to take preventive actions. However, from an informed decision making perspective this is not always desirable, since the goal is not per se to persuade people to take action, but rather to inform them so they can make a good decision. On the other hand, if the negative affect evoked by social

comparisons acts as a motivator to find more information, it could benefit informed decision making. In order to gain insight into this, it is important to conduct research based on the above-mentioned theories on (affective) information processing and social comparisons.

(8)

Several types of comparative risk information have been used in risk communication, including the risk of an average person and the risk of someone of the same age who does not have any risk factors (e.g. Fagerlin et al., 2007; French, Sutton, Marteau & Kinmonth, 2004). Dillard and colleagues studied women regarding breast cancer and found that, after

controlling for actual risk and absolute risk perceptions, comparative risk perceptions (higher or lower risk than the average woman) predicted worry and knowledge, but not intentions or behavior (Dillard et al., 2011). The authors therefore suggested that comparative risk

information may improve knowledge by increasing levels of worry. It also seems reasonable to provide people with their lowest possible risk (for example: their risk if they stopped smoking, see Hill et al., 2010) or to compare them to a person without risk factors in order to give them an ideal or goal to attain. These different types of comparisons have not been directly compared with each other.

The goal of the present study is to directly compare different types of comparative risk information in how they influence people’s affective risk appraisals (negative affect and worry) as well as how they influence a set of cognitive appraisals such as risk comprehension and risk perception, and behavioral intentions.

Research Questions

The main research question of this study was: What is the effect of giving comparative risk information on people’s information processing when using the information in a disease risk test? The main question was divided into the following sub questions: Does providing people with an upward social comparison anchor in disease risk information lead to more motivation to search for information (1), does providing people with an upward social comparison anchor in disease risk information lead to more intention to engage in screening behaviors (2) and/or does providing people with an upward social comparison anchor in

(9)

disease risk information improve risk comprehension (3)? If so, does negative affect (in this case: a form of integral affect) or more specifically worry also play a role?

Methods

Design

This study had an experimental between-subjects design. Participants saw fictitious information about their cardiovascular disease risk, and were randomly assigned to one of the following three different comparison conditions; (1) comparison with a person of the same age without risk factors, (2) comparison with an average person of the same age and (3) a no comparison control condition. This study was part of a broader project about different

intuitive/affective cues to design disease risk information. In that project, an additional fourth condition with a comparison with someone who was 10 years older with the same risk was studied, as well as a set of different manipulations using analogies. These manipulations were no object of the study reported here. Participants first filled out the questionnaire for the present study and subsequently the questionnaire regarding analogies.

Participants

Participants from the target population of cardiovascular health checks were recruited via a Dutch online access research panel (FlyCatcher Internet Research, 20,000 panel

members in total, ISO 20252 and ISO 26362 certified). We approached people between 45 and 69 years old, because of the typical higher cardiovascular disease risk in older people and the corresponding Dutch guideline to invite people for cardio metabolic health checks from about 40 to 45 years of age. People were excluded if they reported having cardiovascular disease, diabetes or chronic kidney failure. The survey was sent out to 580 panel members, and the final sample consisted of 152 participants (see Figure 2 for the selection process). The sample was stratified by gender, age and education level.

(10)

Figure 2. Selection process of the participants.

Procedure

Participants received an e-mail from Flycatcher inviting them to participate in the study. All participants were presented with the same result of a risk test for cardiovascular diseases and asked to imagine that this test result was their own test result. Participants had to answer questions that assessed whether they correctly extracted the numbers from the

information while looking at the information. Subsequently, they were asked questions about the test result. Afterwards they were thanked for their cooperation and were again explicitly told that the presented cardiovascular risk was not their own risk and that they could calculate their own risk at www.testuwrisico.nl.

Materials

The experimental materials provided fictitious information about people’s

cardiovascular risk. The risk information was presented in a bar graph (see Figure 3) with either two bars for the two experimental conditions or one bar for the control condition. Participants had to answer the following questions that assessed whether they correctly extracted the numbers from the information: “Based on the graph pictured above, what is:

(11)

your risk of developing cardiovascular disease? _%” and for both experimental conditions the same question was asked again for “the risk of a person your age who does not have any risk factors? _%” and “the risk of an average person your age? _%”. The questionnaire consisted of multiple choice questions. Except for the negative affect items, the items had a 7 point Likert scale where only the two most extreme options had verbal labels.

Figure 3. Example of a comparison graph. Main Dependent Variables

Negative affect. In previous research negative affect was mostly assessed with one or

two items about worry, but we also wanted to measure negative affect more in general, because upward social comparisons have been associated with different kinds of negative affect (Schwarz, 2000; Watson & Tellegen, 1985). Therefore, we used the PANAS scale (Watson, 1995) to measure negative affect (Cronbach’s alpha of .938 for the negative scale items and .919 for the all the items).

Worry. A separate question was added to assess worry as a more specific aspect of

negative affective risk appraisals mostly used in risk communication and perception research. Participants had to indicate whether they thought the test result was worrisome on a 7 point Likert scale.

(12)

Behavioral intentions. For behavioral intentions 2 separate items were formulated,

one regarding lifestyle change and one regarding visiting the GP. Participants had to indicate how likely it was that they would change their lifestyle or visit their GP based on the test result, on a 7 point Likert scale.

Information seeking. Information seeking was assessed by asking participants

whether they would search for information regarding cardiovascular disease, information on cholesterol and blood pressure values, information on their risk at cardiovascular disease and information about possibilities to lower their risk. An average score was calculated from the answers to these questions on 7 point Likert scales (Cronbach’s alpha .981).

Risk comprehension. Risk comprehension was assessed by asking participants to

indicate on a 7 point Likert scale whether they had a good picture of their risk, whether they had the idea that something is going on and whether the result told them nothing (with the last item being reverse-coded).

Other Dependent Variables

Risk perception. Risk perception was assessed by asking participants to indicate their

chance and likelihood of getting CVD and the severity of the disease. Background Variables

Numeracy. For measuring numeracy we used a single item version of the Berlin

Numeracy Test (Cokely et. al, in press). Participants were allowed to skip this question if they did not want to answer it.

Graph literacy. Graph literacy was measured by using three questions regarding bar

graphs from a graph literacy scale (adapted from Galesic and Garcia-Retamero, 2011). Participants were also allowed to skip these questions if they did not want to answer them. Participants were allowed to skip these questions, because previous experience of the VUmc

(13)

research group revealed that participants generally find these types of questions very bothersome.

Statistical analyses

We conducted a MANOVA to assess the main effects of the manipulation on the outcome variables. Mediation analyses were conducted to determine whether negative affect or worry was a mediator for the effect of the manipulation on the outcome measures. We conducted all of the analyses twice; once for the entire sample and once for a selected sub sample, based on whether participants were able to correctly extract the numerical

information from the risk graph.

Results

Table 1 describes the characteristics of the 152 participants. Participants in the final sample were men and women between the ages of 45 and 69 (M = 56.330, SD = 7.099). Differences can be seen between the two sub samples when the sample is divided based on whether participants were able to correctly extract numerical information from the test result graph. As can be seen, participants who incorrectly extracted the numerical information were more often low educated than participants who correctly extracted the information

(respectively 55% and 27%). From the first group only 23% answered the numeracy question correctly, while 43% of the second group did so. The first group also scored lower on graph literacy; 17% answered all questions correctly compared to 58% of the second group.

(14)

Table 1

Description of the Study Population

Variable

Complete sample (N=152)

Correctly extracted numbers from risk information (N=99)

Incorrectly extracted numbers from risk information (N=53) No (%) Age 45-49 37 (24) 24 (24) 13 (25) 50-54 29 (19) 23 (23) 6 (11) 55-59 30 (20) 20 (20) 10 (19) 60-64 30 (20) 18 (18) 12 (23) 64-69 26 (17) 14 (14) 12 (23) Gender Male 80 (53) 54 (55) 26 (49) Female 72 (47) 45 (45) 27 (51) Education Low 56 (37) 27 (27) 29 (55) Medium 56 (37) 44 (44) 12 (23) High 40 (26) 28 (28) 12 (23)

Berlin Numeracy Test (1 item)

Correct 55 (36) 43 (43) 12 (23)

Incorrect 77 (51) 43 (43) 34 (64)

Not answered 20 (13) 13 (13) 7 (13)

Graph literacy (number of questions answered correctly)

0 19 (13) 9 (9) 10 (19) 1 10 (7) 4 (4) 6 (11) 2 27 (18) 19 (19) 8 (15) 3 67 (44) 58 (58) 9 (17) Not answered 29 (19) 9 (9) 20 (38) Analyses

Overall, negative affect scores were low for all the participants (M = 1.786, SD = .787, on a scale of 5) while the worry scores were medium (M = 4.546, SD = 1.380, on a scale of 7) as seen in Table 2.

First, we employed he analyses for the complete sample. The MANOVA only revealed a significant effect of the manipulation on the picture participants had of their risk (F(2, 149) = 6.366, p = .002). Participants who had been presented with the risk of a person without risk factors reported having a better picture of their risk compared to the participants in the control condition (see Table 2). No other effects of the manipulation were found.

(15)

Table 2

Results of the complete sample

Without risk factors Average person No comparison

Variable M SD M SD M SD

Behavioral intentions

Intention to visit GP 5.359 2.001 5.375 1.931 5.373 1.847 Intention to change lifestyle 5.283 1.364 5.188 1.734 5.314 1.530 Information seeking (average) 4.788 1.762 5.151 1.850 5.020 1.557 More meaningful interpretation

Having a good picture of one’s risk 4.830ᵃ 1.297 4.208 1.487 3.922ᵃ 1.197 Having the idea that something is

wrong

4.811 1.520 4.646 1.839 4.745 1.440 The test result tells one nothing 2.981 1.647 2.875 1.872 3.059 1.630 Affect

Negative affect (average) 1.900 0.841 1.850 0.829 1.606 0.787

Worry 4.491 1.502 4.688 1.518 4.471 1.102

Risk perception

Risk estimation 3.849 1.392 4.083 1.471 3.863 1.096

Chance of getting CVD 3.660 1.372 3.854 1.255 3.608 1.218

Severity 6.019 1.168 6.042 0.967 6.098 0.922

Note. Means with the same superscript within rows differ significantly at the .05 level (with Bonferroni correction).

Second, we performed the same analyses only in the subgroup of participants who were able to correctly extract the numerical information from the risk information graph (see Table 3). Once again the MANOVA revealed a significant effect of the manipulation on the picture the participants had of their risk (F(2, 96) = 3.349, p = .039) with the participants who saw the comparison with a person without risk factors scoring higher than the participants in the control condition. Apart from this effect, we demonstrated a number of other effects of our manipulation. Participants who had seen the comparison with a person without risk factors or with an average person both scored higher on the negative affect (F(2, 96) = 5.811, p = .004) and worry (F(2, 96) = 6.085, p = .003) scales than the participants in the control condition. Participants in the two experimental conditions with comparative risk information also estimated their risk at developing cardiovascular disease higher than the participants in the control condition (F(2, 96) = 4.696, p = .011) and participants who saw the comparison with a person without risk factors thought their chance at developing cardiovascular disease was higher than the participants in the control condition (F(2, 96) = 4.190, p = .018). Mediation

(16)

analysis revealed that worry mediated the relationship between providing comparative risk information and the picture participants had of their risk, while negative affect did not (see Figure 4).

Table 3

Results of people who correctly extracted numerical information from the graph

Without risk factors Average person No comparison

Variable M SD M SD M SD

Behavioral intentions

Intention to visit GP 5.867 1.833 5.867 1.676 5.385 1.680 Intention to change lifestyle 5.600 1.276 5.533 1.432 5.103 1.553 Information seeking (average) 5.408 1.647 5.475 1.678 4.878 1.589 More meaningful interpretation

Having a good picture of one’s risk 4.700ᵃ 1.264 4.367 1.426 3.897ᵃ 1.210 Having the idea that something is

wrong

5.467 .973 5.267 1.660 4.718 1.538 The test result tells one nothing 2.233 1.382 2.500 1.737 3.000 1.638 Affect

Negative affect (average) 2.087ᵃ .897 2.000ᵇ .905 1.510ᵃᵇ .488

Worry 5.067ᵃ 1.112 5.300ᵇ 1.236 4.385ᵃᵇ 1.091

Risk perception

Risk estimation 4.633ᵃ .964 4.567ᵇ 1.331 3.897ᵃᵇ 1.046 Chance of getting CVD 4.300ᵃ 1.055 4.133 1.074 3.600ᵃ 1.094

Severity 6.200 .805 6.167 .913 6.077 .929

Note. Means with the same superscript within rows differ significantly at the .05 level (with Bonferroni correction).

Figure 4. Mediation analyses for worry and negative affect. Standardized coefficients are presented, * = significant at the .05 level.

Discussion

The main goal of the current study was to determine the effect of giving social comparative risk information in the risk communication in an online cardiovascular risk test.

(17)

We looked at the effect of providing consumers with such information on negative affect, worry, risk comprehension (including meaningful interpretations and risk perception) and behavioral intentions. For the complete sample in this study, we only found an effect of the manipulation on the self-reported quality of the picture participants had of their risk (in other words: their meaningful interpretation). When the participants who were not able to correctly extract the numerical information from the risk test result were excluded from the analyses, we also demonstrated effects of the manipulation on negative affect, worry and risk

perception.

An unexpected and important finding of this study was that approximately one third of the participants was not able to correctly extract the numerical information from the test result graph. We cannot know for sure whether these people did not understand the information or whether they simply did not put much effort in this task. Since we gave participants their hypothetical 10 year risk at developing cardiovascular disease, it is possible that they thought they had to calculate their current risk based on the given 10 year risk. However, since the people who were not able to correctly extract the numerical information overall tended to have a lower level of education, lower numeracy scores and lower graph literacy scores, we have good reason to believe that they simply did not adequately understand the presented risk information in the graph. The fact that less numerate people have difficulty using numerical information has been shown in several previous studies (e.g. Garcia-Retamero & Galesic, 2009). In particular, Peters and colleagues found that less numerate individuals benefited less from numeric information (Peters, Hart, Tusler & Fraenkel, 2014). Notably, we only found the expected effects of our manipulation on negative affect, worry and risk perception in the subgroup of participants who were able to correctly extract the numerical information. From this we conclude that it seems that a certain level of understanding of the risk information is necessary for comparative risk information to induce higher levels of negative affect, worry

(18)

and to heighten risk perception. This is unfortunate because individuals who have trouble understanding risk information are the ones who need more help to make a good and informed decision. The question remains what types of formats would help these people in making informed decisions. Using visual aids could benefit people with low levels of numeracy but relatively high levels of graph literacy, as suggested by Garcia-Retamero and Galesic (2010). For people with both low numeracy and low graph literacy analogies could be helpful, as suggested by Galesic and colleagues (Galesic, Barton, Wegwarth, Gaissmaier, & Gigerenzer, 2008).

An interesting finding of this study was that when we look at the effect of the

manipulation of the comparative risk information on the picture participants had of their risk, only the scores in the condition with the comparison with a person without risk factors

significantly deviated from the control condition. Participants who saw the comparison with a person without risk factors reported having the best picture of their risk at cardiovascular disease. A comparison with a person without risk factors may serve as an ideal or baseline risk and could therefore be considered the most informative.

Another important finding was that the manipulation of the comparative risk

information had no effect on behavioral intentions regarding GP visits, lifestyle change and looking for more information about cardiovascular disease and its risk factors. Although these findings differ from findings in some other published studies (Lipkus & Klein, 2006;

Schmiege, Klein & Bryan, 2009), they are broadly consistent with those of Dillard et al. (2011). Lipkus and Klein found that giving people social comparative information regarding colorectal cancer (i.e. that they had more than the average number of risk factors) led to both higher screening intentions and higher screening rates (2006). This is in line with the results of the study from Schmiege, Klein and Bryan (2009), who found that providing people with inflated estimates of peer flossing behavior led to increased behavioral intentions and

(19)

increased flossing behavior (2009). Dillard and colleagues studied women at an increased risk of breast cancer and found no associations between comparative risk perceptions and

treatment- related intentions or behaviors, but they did find an association between

comparative risk perception and feelings of anxiety and knowledge about treatment (2011). The authors therefore suggested that influencing comparative risk perceptions could increase informed decision making. There are several differences between the previous studies en the present study. First of all, most of the previous studies have investigated the effect of giving another type of numerical information than chance information, such as the amount of risk factors instead of the risk itself (Lipkus & Klein, 2006) or the behavior of others (Schmiege et al., 2009) and this might have different effects on people than providing them with

comparative information regarding their risk itself. For example, people may downplay the contribution of those separate risk factors to their risk. Furthermore, the discrepancies with previous research may also be partly explained by the fact that most previous research studied different types of cancer, instead of CVD. Cancer may be different from cardiovascular disease in the sense that people may think most causes of cancer are uncontrollable and/or caused more often by environmental or genetic factors than by lifestyle factors (Maskarinec, Gotay, Tatsumura, Shumay & Kakai, 2001; Willcox, Stewart, Sitas, 2011; Wold, Byers, Crane, Ahnen, 2005), while lifestyle-related factors like overweight, physical inactivity, smoking, high cholesterol and chronic stress are considered the most important causes of cardiovascular disease (French, Senior, Weinman & Marteau, 2001; Mosca, Jones, King, Ouyang, Redberg, & Hill, 2000). Furthermore, cancer is often considered a more life threating and more scary disease than cardiovascular disease (Mosca, Jones, King, Ouyang, Redberg, & Hill, 2000). In conclusion, the effects found in studies regarding cancer risk may not

(20)

scared of developing cardiovascular disease and may therefore be less concerned when they have a higher risk than another person.

Interestingly, we demonstrated that the relationship between the type of comparative risk information and the self-reported quality of their picture participants had was mediated by worry. In this case worry may have served as information to give more meaning to the risk information, as suggested by Peters and colleagues (Peters et al., 2006). This finding

corroborates the ideas of Dillard and colleagues (2011), who suggested that comparative risk perceptions may increase knowledge via anxiety. In general, therefore, it seems that social comparison anchors can benefit informed decision making by improving the picture people have of their risk via increased levels of worry. However, we used a self-report measure for the picture people had of their risk and therefore we do not know whether it was actually improved. If people’s more objective risk comprehension is not actually improved,

heightening levels of worry may be an unnecessary negative consequence of providing people with social comparative risk information.

Unfortunately, we were not able to demonstrate that the relationship between the type of comparative risk information and the better picture participants reported was mediated by negative affect in general. It might be that the way we measured negative affect was too broad for risk communication purposes, since it also includes items like hostile, irritable and guilty, which may not be applicable. It would be interesting to look at the different items separately to see whether specific items such as afraid, nervous, upset and distressed do in fact mediate the relationship between comparative risk information and self-reported risk comprehension. Limitations

The generalizability of our results is subject to certain limitations. First, the risk we presented to the participants was hypothetical and possibly very different from their actual

(21)

risk. Therefore, some participants may have had more difficulty imagining the presented risk was theirs and would have reacted differently to their own risk.

Second, to recruit participants we used a Dutch online access research panel. This type of online panels are increasingly used for research (Evans & Mathur, 2005), but the question remains whether their use is always valid. The use of such panels may limit the

generalizability of the results. While the sample was stratified by gender, age and education level based on the population information from the Netherlands, the sample might not accurately represent the entire population since people need to have internet access and have to sign up for the panel themselves. This is a general problem with online access that was also pointed out in previous studies (e.g. Baker et al., 2010; Loosveldt & Sonck, 2008).

Furthermore, we do not know how committed panel members are to fill out these

questionnaires to their best knowledge or whether they rush through these questionnaires to earn their points. This may not have been the case in this study since Flycatcher checked the quality of the data and removed data of very poor quality.

Conclusions

The results from this study indicate that giving comparative risk information can contribute to cardiovascular risk comprehension and heightened risk perceptions. Importantly, the effect on self-reported risk comprehension occurred through influencing people’s levels of worry. Since comparative risk information, especially the comparison with a person without risk factors, is able to heighten levels of worry and risk perception and can improve risk comprehension, it could be used in risk communication to improve informed decision making for consumers. However, people with low numeracy and low graph literacy still have trouble understanding the (numerical) risk information in general, while they are the people who need more help to increase their risk comprehension. Future research is therefore needed to

(22)

different levels of numeracy and graph literacy and different needs regarding risk information, there may not be one ideal format to present risk information. The solution may lie in tailoring formats for groups of individuals with different levels of numeracy and graph literacy or by combining different formats.

(23)

References

Alssema, M., Newson, R. S., Bakker, S. J., Stehouwer, C. D., Heymans, M. W., Nijpels, G., ... & Dekker, J. M. (2012). One risk assessment tool for cardiovascular disease, type 2 diabetes, and chronic kidney disease. Diabetes Care, 35(4), 741-748.

doi:10.2337/dc11-1417

Baker, R., Blumberg, S. J., Brick, J. M., Couper, M. P., Courtright, M., Dennis, J. M., ... & Zahs, D. (2010). Research Synthesis: AAPOR Report on Online Panels. Public Opinion Quarterly, 74(4), 711–781. doi:10.1093/poq/nfq048

Bekker, H. L., Winterbottom, A. E., Butow, P., Dillard, A. J., Feldman-Stewart, D., Fowler, F. J., ... & Volk, R. J. (2013). Do personal stories make patient decision aids more effective? A critical review of theory and evidence. BMC Medical Informatics and Decision Making, 13(Suppl 2), S9. doi:10.1186/1472-6947-13-s2-s9

Cline, R. J., & Haynes, K. M. (2001). Consumer health information seeking on the Internet: the state of the art. Health Education Research, 16(6), 671-692.

doi:10.1093/her/16.6.671

Damasio, A. (1994). Descartes' Error: Emotion, Reason and the Human Brain. New York: Putnam.

Damman, O. C., Bogaerts, N. M. M., Van Dongen, D., & Timmermans, D. R. M. (in preparation). Barriers in using cardio-metabolic risk information among consumers with low health literacy.

Dillard, A. J., Ubel, P. A., Smith, D. M., Zikmund-Fisher, B. J., Nair, V., Derry, H. A., ... & Fagerlin, A. (2011). The distinct role of comparative risk perceptions in a breast cancer prevention program. Annals of Behavioral Medicine, 42(2), 262-268. doi:10.1007/s12160-011-9287-8

(24)

Evans, J. R., & Mathur, A. (2005). The value of online surveys. Internet Research, 15(2), 195-219. doi:10.1108/10662240510590360

Fagerlin, A., Zikmund-Fisher, B. J., & Ubel, P. A. (2007). “If I’m better than average, then I’m ok?”: Comparative information influences beliefs about risk and benefits. Patient Education and Counseling, 69(1), 140-144. doi:10.1016/j.pec.2007.08.008

Festinger, L. (1954). A theory of social comparison processes. Human relations, 7(2), 117-140.

French, D. P., Senior, V., Weinman, J., & Marteau, T. M. (2001). Causal attributions for heart disease: a systematic review. Psychology and Health,16(1), 77-98.

doi:10.1080/08870440108405491

French, D. P., Sutton, S. R., Marteau, T. M., & Kinmonth, A. L. (2004). The impact of personal and social comparison information about health risk. British Journal of Health Psychology, 9(2), 187-200. doi:10.1348/135910704773891041

Galesic, M., Barton, A., Wegwarth, O., Gaissmaier, W., & Gigerenzer, G. (2008). Do analogies help elderly people understand medical information? International Journal of Psychology, 43, 32.

Galesic, M., & Garcia-Retamero, R. (2011). Graph Literacy A Cross-Cultural Comparison. Medical Decision Making, 31(3), 444-457.

doi:10.1177/0272989x10373805

Garcia-Retamero, R., & Galesic, M. (2009). Communicating treatment risk reduction to people with low numeracy skills: A cross-cultural comparison.American Journal of Public Health, 99(12), 2196-2202. doi:10.2105/ajph.2009.160234

Garcia-Retamero, R., & Galesic, M. (2010). Who proficts from visual aids: Overcoming challenges in people's understanding of risks. Social science & medicine, 70(7), 1019-1025. doi:10.1016/j.socscimed.2009.11.031

(25)

Gigerenzer G., Gaissmaier W., Kurz-Milcke E., Schwartz L. M. , Woloshin S. (2007). Helping doctors and patients to make sense of health statistics. Psychological Science in the Public Interest , 8, 53-96. doi:10.1111/j.1539-6053.2008.00033.x

Hill, S., Spink, J., Cadilhac, D., Edwards, A., Kaufman, C., Rogers, S., ... & Tonkin, A. (2010). Absolute risk representation in cardiovascular disease prevention:

comprehension and preferences of health care consumers and general practitioners involved in a focus group study. BMC Public Health, 10(1), 108. doi:10.1186/1471-2458-10-108

Jacowitz, K. E., & Kahneman, D. (1995). Measures of anchoring in estimation tasks.

Personality and Social Psychology Bulletin, 21, 1161-1166. doi:10.1037/e722982011-058

Keller, C., Siegrist, M., & Gutscher, H. (2006). The role of the affect and availability heuristics in risk communication. Risk Analysis, 26(3), 631-639. doi:10.1111/j.1539-6924.2006.00773.x

Klein, W. M. (1997). Objective standards are not enough: affective, self-evaluative, and behavioral responses to social comparison information. Journal of Personality and Social Psychology, 72(4), 763. doi:10.1037/0022-3514.72.4.763

Klein, W. M. (2003). Self-prescriptive, perceived, and actual attention to comparative risk information. Psychology and Health, 18(5), 625-643.

doi:10.1080/0887044031000118465

Kon, A. A. (2010). The shared decision-making continuum. JAMA, 304(8), 903-904. doi:10.1001/jama.2010.1208

Kunda, Z. (1987). Motivated inference: Self-serving generation and evaluation of causal theories. Journal of Personality and Social Psychology, 53(4), 636-647.

(26)

Lipkus, I. M., & Klein, W. M. (2006). Effects of communicating social comparison information on risk perceptions for colorectal cancer. Journal of Health Communication, 11(4), 391-407. doi:10.1080/10810730600671870

Lipkus, I. M., Kuchibhatla, M., McBride, C. M., Bosworth, H. B., Pollak, K. I., Siegler, I. C., & Rimer, B. K. (2000). Relationships among breast cancer perceived absolute risk, comparative risk, and worries. Cancer Epidemiology Biomarkers & Prevention, 9(9), 973-975.

Loewenstein, G. F., Weber, E. U., Hsee, C. K., & Welch, N. (2001). Risk as

feelings. Psychological Bulletin, 127(2), 267-286. doi:10.1037/0033-2909.127.2.267 Loosveldt, G., & Sonck, N. (2008). An evaluation of the weighting procedures for an online access panel survey. In Survey Research Methods (Vol. 2, No. 2, pp. 93-105). Wiley. Maskarinec, G., Gotay, C. C., Tatsumura, Y., Shumay, D. M., & Kakai, H. (2001). Perceived

Cancer Causes. Use of Complementary and Alternative Therapy. Cancer Practice, 9(4), 183–190. doi:10.1046/j.1523-5394.2001.94006.x

Mosca, L., Jones, W. K., King, K. B., Ouyang, P., Redberg, R. F., & Hill, M. N. (2000). Awareness, perception, and knowledge of heart disease risk and prevention among women in the United States. Archives of Family Medicine, 9(6), 506-515.

doi:10.1001/archfami.9.6.506

O’Connor, A. M., Wennberg, J. E., Legare, F., Llewellyn-Thomas, H. A., Moulton, B. W., Sepucha, K. R., ... & King, J. S. (2007). Toward the ‘tipping point’: decision aids and informed patient choice. Health Affairs, 26(3), 716-725. doi: 10.1377/hlthaff.26.3.716 Peters, E., Dieckmann, N. F., Västfjäll, D., Mertz, C. K., Slovic, P., & Hibbard, J. H. (2009).

Bringing meaning to numbers: the impact of evaluative categories on decisions. Journal of Experimental Psychology: Applied, 15(3), 213-227. doi:10.1037/a0016978

(27)

Peters, E., Hart, P. S., Tusler, M., & Fraenkel, L. (2014). Numbers Matter to Informed Patient Choices A Randomized Design across Age and Numeracy Levels. Medical Decision Making, 34(4), 430-442. doi:10.1177/0272989x13511705

Peters, E., Lipkus, I., & Diefenbach, M. A. (2006). The functions of affect in health communications and in the construction of health preferences. Journal of Communication, 56(s1), S140-S162. doi:10.1111/j.1460-2466.2006.00287.x Peters, E., Slovic, P., & Gregory, R. (2003). The role of affect in the WTA/WTP disparity.

Journal of Behavioral Decision Making, 16(4), 309-330. doi:10.1002/bdm.448 Reyna V. F., Nelson W. L., Han P. K., Dieckmann N. F. (2009). How numeracy influences

risk comprehension and medical decision making. Psychological Bulletin, 135, 943-973. doi:10.1037/a0017327

Rimer, B. K., Briss, P. A., Zeller, P. K., Chan, E. C., & Woolf, S. H. (2004). Informed decision making: What is its role in cancer screening? Cancer, 101(S5), 1214-1228. doi:10.1002/cncr.20512

Schmiege, S. J., Klein, W. M., & Bryan, A. D. (2010). The effect of peer comparison information in the context of expert recommendations on risk perceptions and subsequent behavior. European Journal of Social Psychology, 40(5), 746-759. doi:10.1002/ejsp.645

Schwarz, N. (2000). Emotion, cognition, and decision making. Cognition & Emotion, 14(4), 433-440. doi:10.1080/026999300402745

Sherman, D. A., Nelson, L. D., & Steele, C. M. (2000). Do messages about health risks threaten the self? Increasing the acceptance of threatening health messages via self-affirmation. Personality and Social Psychology Bulletin, 26(9), 1046-1058.

(28)

Slovic, P., Finucane, M. L., Peters, E., & MacGregor, D. G. (2004). Risk as analysis and risk as feelings: Some thoughts about affect, reason, risk, and rationality. Risk Analysis, 24(2), 311-322. doi:10.1111/j.0272-4332.2004.00433.x

Slovic, P., Peters, E., Finucane, M. L., & MacGregor, D. G. (2005). Affect, risk, and decision making. Health psychology, 24(4S), S35. doi:10.1037/0278-6133.24.4.s35

Tesser, A., Millar, M., & Moore, J. (1988). Some affective consequences of social

comparison and reflection processes: the pain and pleasure of being close. Journal of Personality and Social Psychology, 54(1), 49. doi:10.1037/0022-3514.54.1.49 Tversky, A., & Kahneman, D. (1974). Judgment under uncertainty: Heuristics and biases.

Science, 185(4157), 1124-1131. doi:10.1126/science.185.4157.1124

Visschers, V. H. M., Wiedemann, P. M., Gutscher, H., Kurzenhäuser, S., Seidl, R., Jardine, C. G., & Timmermans, D. R. M. (2012). Affect-inducing risk communication: current knowledge and future directions. Journal of Risk Research, 15(3), 257-271.

doi:10.1080/13669877.2011.634521

Watson, D., & Clark, L. A. (1999). The PANAS-X: Manual for the positive and negative affect schedule-expanded form.

Watson, D., & Tellegen, A. (1985). Toward a consensual structure of mood. Psychological Bulletin, 98(2), 219. doi:10.1037/0033-2909.98.2.219

Willcox, S. J., Stewart, B. W., & Sitas, F. (2011). What factors do cancer patients believe contribute to the development of their cancer? (New South Wales, Australia). Cancer Causes & Control, 22(11), 1503-1511. doi:10.1007/s10552-011-9824-6

Witte, K., & Allen, M. (2000). A meta-analysis of fear appeals: Implications for effective public health campaigns. Health Education & Behavior, 27(5), 591-615.

(29)

Wold, K. S., Byers, T., Crane, L. A., & Ahnen, D. (2005). What do cancer survivors believe causes cancer? (United States). Cancer Causes & Control,16(2), 115-123.

doi:10.1007/s10552-004-2414-0

Ybarra, M. L., & Suman, M. (2006). Help seeking behavior and the Internet: A national survey. International Journal of Medical Informatics, 75(1), 29-41.

Referenties

GERELATEERDE DOCUMENTEN

We sum- marize the findings from a study of myocardial infarction in young women that suggests that the joint effects of a prothrombotic risk factor, factor V Leiden, and other

In het kort geeft dit boek als antwoord op deze dreiging: ga voor ‘real-time’ evaluatie; verleg uw aandacht van ex-post evaluatie ten behoeve van verantwoording achteraf, naar

Gekeken zal worden naar in hoeverre deze specifieke criteria anders worden toegepast dan de algemene eisen van proportionaliteit en subsidiariteit en tot wat voor soort praktijken

In this section I will be presenting prior research and formulating hypotheses in association with the effects of ERM implementation with regard to CEO

On 23 February 2009 the European Commission adopted the 'EU Strategy for Supporting Disaster Risk Reduction in Developing Countries’, together with the related ‘Community Approach on

To provide more insight in the relationship between social capital of a country and risk-taking behaviour in this thesis I will use two measurements (The Legatum Institute

A concern with regression 2 is that banks may have changed their credit derivative activities in response to the crisis. The crisis interaction term in regression 2 relates to

Contradictory, current study did find a significant effect for peer popularity and self-esteem on selection when comparing a high and low self-esteem group, which suggests that