Healthy Eating Blogs, Helping or Hurting Women?
Study on the Effects of Blogger Expertise in Healthy Eating Blogs on the Attitudes and Behavioral Intentions of Women.
Debbie Soeters – 10389199 Master’s Thesis
Graduate School of Communication
Master’s programme Communication Science Supervisor: dr. Guda van Noort
Abstract
Online health information has multiple shortcomings, which in part stem from the authorship of this information. Specifically, medical professionals author a small amount of online health information, while non-professionals author the majority of this information. This division in authorship gives way to improper information. This is particularly worrisome, as seekers of online health information are influenced by and act upon this information. The focus here lies with diet and nutrition blogs, as these have seen considerable growth amongst health information seekers, while being under researched. The main reader- and authorship of these blogs exists of women. The aim of the current study was therefore to investigate
whether women make distinctions between professional and non-professional authors of health blogs in order to determine authors’ credibility and to adjust their attitudes and behavioral intentions accordingly. Furthermore, involvement was also taken into account, as the persuasive impact of the source may differ for women with different levels of
involvement. Tests were conducted to reveal whether the impact of source credibility as induced by source expertise on attitudes and intentions towards the health behavior was moderated by involvement. An online experiment (N = 109) revealed that source expertise enabled women to criticize bloggers’ credibility and to adapt their attitudes and behavioral intentions accordingly. Furthermore, the impact of source credibility on attitudes directly and on intention as mediated by attitude was stronger for women who were relatively low
involved with the health behavior. These results have important implications for theory, female health information seekers, professional medical bloggers, policy makers, and health practitioners.
Healthy Eating Blogs, Helping or Hurting Women?
Study on the Effects of Blogger Expertise in Healthy Eating Blogs on the Attitudes and Behavioral Intentions of Women.
Health information seekers increasingly use the Internet to find health information (Fox & Jones, 2009; Pollard, Pulker, Meng, Kerr, & Scott, 2015; Sillence, Briggs, Harris, & Fishwick, 2007). This online health information has beneficial effects for health promotion, such as enabling health consumers to make more informed health decisions (Goldsmith, 2000). However, these benefits cannot be understood without consideration of its detrimental characteristics (Berland et al., 2001; Molassiotis & Xu, 2004; Tatsioni et al., 2003).
Specifically, research revealed that shortcomings of content are typical, with part of the content being misleading, harmful, or containing poor advice (Boepple & Thompson, 2014; Carrotte, Psych, Vella, & Lim, 2015; Modave, Shokar, Peñaranda, & Nguyen, 2014;
Molassiotis & Xu, 2004; Tatsioni et al., 2003). Health information seekers can thus swiftly and easily access inaccurate information, which makes the credibility of online health information vital to providing them with high-quality information (Jung, Walsh-Childers, & Kim, 2016). This study therefore focuses on the impact of the quality and credibility of online health information.
While considering the quality and credibility, it is important to note that content only becomes harmful when people act upon, and develop opinions, thoughts, and beliefs in line with the information. Previous research revealed that health information seekers are regularly influenced by (Ostry, Young, & Hughes, 2008; Rainie & Fox, 2000; Sillence et al., 2007) and act upon health information found online (Fox & Rainie, 2002; Rosenvinge, Laugerud, & Hjortdahl as cited in Hu & Sundar, 2010; Rideout, 2001; Siow et al., 2003). Specifically, online health information impacted 70% of the treatment decisions and 28% of the decisions to visit a doctor of those influenced by online health information (Rainie & Fox, 2000).
Whereas previous studies on the impact of online health information concerned websites, or the Internet as a whole, the current study investigates the impact of blogs, which can be described as online journals in which the most recent post is displayed first (Herring et al., 2005). The focus lies with blogs as these have recently faced rapid growth (Hsu & Lin, 2008), are among the fastest expanding Internet segments (Rausch, 2006), and have become common sources of health information (Hu & Sundar, 2010). More specifically, this study investigates healthy eating blogs, as nutrition/diet is amongst the health topics with the highest online growth (Fox, 2005), and blogs have become an important medium for healthy eating information (Huovila & Saikkonen, 2016).
However, contrary to trends in blog availability and interest, research on blogs (Kaye, 2005), and specifically health and nutrition blogs has lagged behind (Lynch, 2010; Rausch, 2006). Specifically, there is a literature deficiency regarding the persuasiveness of health blogs on women and the factors underlying this effect. This literature gap and two important factors underlying the persuasion process of health blogs will hereafter be discussed.
First, previous research on food/health blogs has focused on bloggers and the content of blogs, therein neglecting the impact on blog readers (Bissonnette-Maheux et al., 2015). Discerning the impact of nutrition blogs on blog readers is important, as people are expected to become more reliant on online health information (Molassiotis & Xu, 2004), online health information seekers are likely to access blogs (Buis & Carpenter, 2009), and nutrition
misinformation may have harmful consequences (Ayoob, Duyff, & Quagliani, 2002). Determination of this impact is especially relevant for women, as they are far more inclined to engage in online health-related behaviors than men (Andreassen et al., 2007; Fallows, 2005; Fox, 2008; Rainie & Fox, 2000). Women are more likely to search for diet and
nutrition information (Pollard et al., 2015), and food bloggers’ author- and readership mainly consists of women (Lynch, 2010; Simunaniemi, Sandberg, Andersson, & Nydahl, 2011). The
general aim of the current study is therefore to understand the impact of healthy eating blogs on women.
Second, this study argues that both source and receiver characteristics play an important role in the persuasion process of health blogs. Previous research on receivers’ strategies to assess the credibility of online health information focused on source
characteristics, therein-revealing source expertise as an important and influential factor (Eastin, 2001; McCroskey & Richmond as cited in Hu & Sundar, 2010; Thon & Jucks, 2017). People differentiate between information from experts and non-experts, as has been revealed in multiple contexts (Eastin, 2001; Fogg et al., 2001; Sillence et al., 2007; Tang, 2016; Thon & Jucks, 2017), however overlooking healthy eating blogs. This while source expertise is especially important in online health information, as medical professionals author only a limited amount of this information (Eastin, 2001). This gives way to improper information (Eastin, 2001) and fraudsters (Moturu, Liu, & Johnson, 2008). The aim of this study is therefore to reveal whether women make distinctions between information from experts and non-experts in healthy eating blogs.
Third, receivers’ involvement has an important impact on the influence of source characteristics, as heuristic cues such as source characteristics are more important in low-involved processing (Petty & Cacioppo, 1986). This differential processing of source characteristics has been proven in multiple contexts (Andrews & Shimp, 1990; Reinhard & Sporer, 2010; Wilson & Sherrel, 1993). However, little is known about the interplay between involvement and source credibility as underlying mechanism of expert source persuasion in health blogs. The aim is therefore to reveal whether source expertise in healthy eating blogs enables women to judge the credibility of authors of (im)proper information, while
investigating the moderating influence of involvement. This leads to the following research question: What effect does source expertise in blogs about healthy eating have on the
attitudes and behavioral intention of women; and do these effects differ for women who are more or less involved?
Theoretical Background Source Expertise & Credibility
Authorship of online health information is not merely in the hands of sources with a professional medical background (Culver, Gerr, & Frumkin, 1997; Moturu et al., 2008), and sources without such background author the majority of messages about medical topics (Culver et al., 1997). This division in authorship can also be observed in the context of healthy eating blogs, which are authored by both bloggers with professional medical backgrounds, such as registered dietitians, and bloggers without medical backgrounds (Bissonnette-Maheux et al., 2015). Previous research revealed that the source of information is determinative of the persuasiveness of the message (Hovland, Janis, & Kelley, 1953). More specifically, due to varying sources of information, characteristics of these sources have become important to the degree of persuasiveness (Hovland et al., 1953). This effect has also been revealed in online health research, as source characteristics have been argued to
facilitate credibility assessments, therein addressing the problematic quality of online health information (Eastin, 2001; Thon & Jucks, 2017).
The focus here lies with the characteristic source expertise. This is the case because source expertise has been revealed to be more impactful than other source characteristics (Busch & Wilson, 1976; Wilson & Sherrel, 1993; Woodside & Davenport, 1974), such as source trustworthiness and attractiveness (Wilson & Sherrel, 1993). Source expertise accounted for 16% of the explained variance in persuasion and had the strongest effect on attitudes and behaviors (Wilson & Sherrel, 1993). Furthermore, authorship with and without professional medical background has previously been indicated as an example of source
expertise (Hu & Sundar, 2010). The current study therefore uses previous research on source expertise to explain the impact of (non-)professional medical authorship.
Source expertise has previously been defined as either a characteristic (Braunsberger & Munch, 1998; Thon & Jucks, 2017) or a perception of the source (Hovland et al., 1953; Hu & Sundar, 2010). In the current study, source expertise is seen as a source characteristic, because it concerns bloggers whom explicitly state that they are (non-)experts. Specifically, source expertise concerns the competences and/or capability of the source to communicate the truth about a given topic (Arora & Arora, 2004; Walther, Wang, & Loh, 2004). In the context of healthy eating blogs this implies that the expert is a blogger whom is
knowledgeable about nutrition, understands the concept of healthy eating (Huovila & Saikkonen, 2016), and thus has the ability to present valid assertions concerning healthy eating. This is for example expressed via authors’ education, social ranking, knowledge, professional accomplishments (Yoon, Kim, & Kim, 1998), experience, or occupation (Arora & Arora, 2004; Thon & Jucks, 2017). This directly links to source credibility, which
concerns the perception that statements of the source are valid representations of the truth (Kelman, 1961). Thus, expert sources are perceived credible because they know the truth about the knowledge-domain in question (Kelman, 1961), possess the knowledge to back their statements (Ohanian, 1991), and are capable to disclose the truth (Arora & Arora, 2004).
The link between expertise and credibility has been attested in research. Specifically, previous non-health related research implies that people use source expertise to assess the credibility of information (Sillence et al., 2007). Therein leading to increased perceived website credibility (Fogg et al., 2001) and influencing persuasive outcomes such as the credibility of information (McCroskey & Richmond as cited in Hu & Sundar, 2010). Health communication research has similarly revealed that health information seekers viewed
(Eastin, 2001; Tang, 2016) as being more credible than non-credentialed and non-expert sources. However, Chuang (2015) and Hu and Sundar (2010) did not find an increase in source credibility for professionals as compared to laypersons. It should however be noted that these authors question their source manipulations and that the source manipulation in the study of Hu and Sundar (2010) had a high failure rate, which could explain the unexpected results. In contrast, Bissonnette-Maheux et al. (2015) called healthy eating blogs authored by qualified dieticians credible sources of information and according to Greenberg, Yaari, and Bar-Ilan (2013) expert bloggers were perceived as more credible than personal bloggers. However, the source manipulation in this study concerned both the writing style of the blog and the information about the blogger. It remains unclear whether the results can be attributed to the blogger, the content style, or a combination of these. The current study therefore aims to uncover whether women make distinctions between professional and non-professional authors solely based on information about the blogger. Thus, it is expected that expert sources are perceived as more credible than non-expert sources (see Appendix A for the conceptual model):
H1a: Expert health blog authors are perceived as more credible than non-expert health blog authors.
Source Credibility & Persuasive Outcomes
Source credibility, which is in this study seen as a perception of the source (Kelman, 1961), may in turn impact persuasive outcomes (Hovland et al., 1953). Previous research indicates that credible (expert) sources, such as physicians, are more persuasive because they are perceived to have the proper knowledge (Arora & Arora, 2004; Bednar & Levie as cited in Bansah, 2016) and objectivity to be able to communicate the truth about a certain issue (Bednar & Levie as cited in Bansah, 2016). Recipients ascribe these sources higher credibility, which leads them to be more likely to accept their messages (Arora & Arora,
2004; Hovland et al., 1953) and to have more positive attitudes, higher behavioral intentions (Arora & Arora, 2004), and higher behavioral compliance in response to messages (Crisci & Kassinove, 1973). This is the case because recipients anticipate that recommendations from well informed, knowledgeable, and honest sources can be verified via experience or would lead to rewards such as social approval and prevention of punishment (Hovland et al., 1953). The persuasiveness of source credibility has been attested in empirical research. Specifically, the influence of source credibility on attitudes will be discerned in the next paragraph
followed by a paragraph on behavioral intentions.
Credibility & attitudes. Attitudes have been defined in various ways in the literature, for example as implicit responses to communication aimed at a specific object, group, or person (Hovland et al., 1953). However, in the current study attitudes are described as relatively durable evaluations of objects, issues, and individuals (Petty, Barden, & Wheeler, 2009). This definition was adopted as it includes issues, which is vital to the attitudes in this study, which concern the issue of healthy behavior.
Previous research on the impact of source credibility on attitudes in an information technology (IT) context revealed that credibility led to more positive attitudes towards the recommended behavior (Johnston & Warkentin, 2010). Similarly, higher source credibility led to more positive attitudes in most studies included in the review on source credibility effects by Pornpitakpan (2004) and in all studies included in the meta-analysis on persuasive message cues by Stiff (1986). Besides these general and/or non-health related studies on source effects, there were also two studies that specifically focused on source credibility in nutritional contexts. These studies revealed the positive impact of source credibility on nutritional attitudes (Bansah, 2016) and attitudes towards applying proper eating habits (Arora & Arora, 2004). The designs of these studies differed from the current study in multiple ways. Specifically, participants of both genders were provided with printed
newsletters (Arora & Arora, 2004) or general nutrition messages in the context of a university course (Bansah, 2016). Moreover, their source manipulations also differed. Source credibility was conveyed via message content style (Bansah, 2016) or via a manipulation of both the author and the organization behind the author (Arora & Arora, 2004). Thus, while these studies imply that higher (versus lower) source credibility leads to more positive attitudes, it is important to test this assumption in the context and population of the current study. Specifically, this study adds to previous research by investigating blogger credibility instead of organizational or message credibility, amongst females instead of males and females, in a relatively new medium (see Appendix A).
H1b: Higher levels of source credibility induced by source expertise, lead to more positive attitudes towards the health behavior.
Credibility & behavioral intention. Besides attitudes, credibility has also been revealed to impact intentions and behavior. Highly credible sources led to behavior in line with source recommendations in most studies included in a general review on source credibility (Pornpitakpan, 2004). Next to these general results, Harris, Sillence, and Briggs (2009) suggested that source credibility positively influenced behavioral responses in the context of health communication. Source expertise, the determinant of source credibility as argued in this study, also had a positive impact on behavior (Woodside & Davenport, 1974). Whereas these studies focused directly on behavior, other studies focused on the impact of source expertise (Hong, 2006) and source credibility (Bannister, 1986; Johnston &
Warkentin, 2010) on intention, finding that higher source expertise and credibility led to increased behavioral intentions. Moreover, these results were also found in the context of healthy eating, as source credibility had a positive impact on intentions to follow healthy eating guidelines (Arora & Arora, 2004). Thus, in the current study higher credibility of the
blogger is expected to heighten behavioral intentions towards the recommended health behavior (see Appendix A):
H1c: Higher levels of source credibility induced by source expertise lead to higher behavioral intentions towards the recommended health behavior.
Involvement Moderation of the Source Effects on Attitudes
The previous argumentation signifies the impact of source expertise on intentions and attitudes, as mediated by source credibility. This relationship is arguably impacted by
involvement, which is defined as the degree to which an issue is personally important (Petty & Cacioppo, 1979). Accordingly, involvement in health research has been conceptualized as the level of importance, interest, and relevance of (eating) whole grain foods (Jung et al., 2016), as involvement with personal health care (Mandl, Kohane, & Brandt, 1998), and as concern about getting cancer (Arora & Arora, 2004). The current study follows this conceptualization of issue-based involvement by investigating involvement with foods containing antioxidants.
Involvement has been argued to moderate the impact of the source on attitudes (Wilson & Sherrel, 1993). This effect can be explained by the Elaboration Likelihood Model (ELM), which provides a framework to understand the mechanisms underlying persuasive communication (Metzger, 2007) and has frequently been applied in credibility research (Arora & Arora, 2004; Eastin, Yang, & Nathanson, 2006; Homer & Kahle, 1990; Hu & Sundar, 2010). The ELM posits the existence of two distinct routes to persuasion (Petty, Cacioppo, & Goldman, 1981). Specifically, the two routes are the central route, which concerns high-involved message recipients who consciously consider arguments central to the message, and the peripheral route, in which the focus of low-involved recipients lies with peripheral cues such as source expertise (Petty et al., 1981). This peripheral processing of source characteristics has been confirmed in research (Andrews & Shimp, 1990). Moreover,
Reinhard and Sporer (2010) concluded that source expertise functioned as a peripheral cue and Jung et al. (2016) argued that this was also the case in the context of online diet/nutrition information.
Besides low-involved processing of source characteristics, the differential processing of source characteristics under low- and high-involvement should also be noted to get a more thorough understanding of source effects. Low-involved recipients tend to rely on source characteristics such as source expertise to decide whether to accept the message or not (Petty & Cacioppo, 1984; Petty et al., 1981). In contrast, high-involved recipients use source characteristics to help understand message arguments, while strong argumentation is more important to them (Petty & Cacioppo, 1984; Petty et al., 1981). This reasoning leads to the stance that source effects most likely occur when message recipients are low involved
(Wilson & Sherrel, 1993). Research supported this stance by revealing that credibility ratings (Reinhard & Sporer, 2010) and simple decisions to reject or accept the message were made based on peripheral elements such as source characteristics in case of low-involvement (Chaiken, 1980), while message based cognitions were also important to high-involved recipients (Chaiken, 1980; Reinhard & Sporer, 2010).
Involvement was also identified as a moderator in research on source effects and attitudes. Previous research revealed that source expertise (Petty et al., 1981), high source credibility (Petty & Cacioppo, 1981; Stiff, 1986), and product endorsers’ celebrity status (Petty, Cacioppo, & Schumann, 1983) led to significantly more favorable attitudes as compared to low expertise and credibility, and non-celebrity endorsers, but only for low-involved recipients. Similarly, Wilson and Sherrel (1993) reported that source characteristics had a significant influence on attitude, but only under conditions of low-involvement in 67% of the studies included in their meta-analysis. Whereas these studies reported that source characteristics only influenced attitudes under conditions of low-involvement, others argued
that source characteristics have more influence under low- than high-involvement. Specifically, Johnson and Scileppi (1969) and Stiff (1986) argued that high (versus low) source credibility led to significantly more positive attitudes under low-involvement as compared to high-involvement.
Thus, research is divided on whether source cues only (or more so) impact attitudes under low-involvement. Some studies revealed that source cues only impact attitudes under low-involvement, while others revealed that source cues more so impact attitudes under low- than high-involvement. Therefore, the current study follows the ELM, which argues that source cues impact attitudes under both high- and low-involvement (Petty et al., 2009; Petty & Cacioppo, 1986). Specifically, persuasion-relevant variables such as source expertise serve as simple cues under low-involvement, but are scrutinized as message arguments under high-involvement (Petty et al., 2009). Thus, while the persuasion process functions differently, these variables may also result in attitude change under conditions of high-involvement (Petty et al., 2009). Therefore, it is expected that source credibility, as induced by blog authors with(out) expertise, has more impact on attitudes under conditions of low-involvement as compared to high-low-involvement (see Appendix A):
H2: Higher levels of perceived credibility induced by source expertise lead to more positive attitudes towards the health behavior, but more so when people are low involved.
Involvement Moderation of the Source Effects on Behavioral Intentions
Whereas multiple studies discuss involvement as a moderator of source effects on attitudes, similar research with intention as outcome variable is rather limited. This difference might be explained by the divergent ways in which source cues are processed. As previously discussed, source cues are expected to have a stronger impact on attitudes when people are low involved and thus process the cue via the peripheral route. Attitudes formed via the peripheral route tend to be less strong (Petty et al., 2009), persistent, and resistant (Petty &
Cacioppo, 1986) than attitudes stemming from the central route and therefore less suitable to predict behavior (Krosnick & Petty as cited in Petty et al., 2009; Petty & Cacioppo, 1986). Thus, since source effects are expected to take place under low-involvement, therein leading to weak attitudes, the resulting attitudes are less likely to determine intentions and behavior. The difference between high- and low-involved people regarding the strength of the link between attitudes and intentions/behavior has been verified in research. Attitudes and issue-relevant behavior were stronger correlated for those who expected to be affected by the issue than those who did not (Sivacek & Crano, 1982). Furthermore, attitudes have been found to be better determinants of behavior for more personally relevant issues (Cialdini, Petty, & Cacioppo, 1981). Leippe and Elkin (1987) similarly found that attitudes predicted behavior, but only when the issue was relevant and participants thus processed the message via the central route. Moreover, Petty et al. (1983) found that manipulating the source’s celebrity status influenced attitudes, but not behavioral intention. The same study attested the differential impact of attitude strength on intentions. Namely, attitudes formed under high-involvement were better predictors of behavioral intentions than attitudes formed under low-involvement.
Whereas these studies concern the indirect impact of involvement on intentions via attitude, involvement arguably also moderates the direct relationship between source
credibility and intentions. This relationship has been hypothesized based on the ELM, using the same arguments as previously discussed to explain the impact of involvement on the credibility-attitude relationship (Pan & Chang, 2011; Shao, Baker, & Wagner, 2004; Yoon, Pinkleton, & Ko, 2005). The impact of involvement on the credibility-intention relationship has been verified when investigating the influence of appropriateness of bank employee clothing on purchase intention (Shao et al., 2004). However, these effects were not found for source expertise when investigating online purchase intentions of a digital camera (Pan &
Chang, 2011) or for source credibility in the context of political voting intentions (Yoon et al., 2005). As the literature is inconclusive concerning the moderating impact of involvement either directly or indirectly on the relationship between source effects and intention, the following research question is posed (See Appendix A):
RQ1: Does source credibility as induced by source expertise (a) directly or (b) indirectly via attitudes towards the health behavior lead to higher behavioral intentions for low-involved people?
Method Research Design
The current study concerns an experimental design with a single between-subject factor, namely: Source expertise: (expert versus non-expert source). Participants were randomly exposed to the “about me” page of either an expert or non-expert source. Selection & Characteristics of Sample
The online experimental study was conducted from 11/05/17 to 23/05/17 amongst 144 women of 18 years or older. Participants were recruited via the personal Facebook and
LinkedIn network of the researcher. From this sample 35 women were removed because they did not currently live in the Netherlands, took less than 4 or more than 60 minutes to
complete the survey, and/or because they incorrectly answered the reading question concerning the bloggers’ daytime occupation. The answer was incorrect if for the expert it did not refer to health, nutrition, research, scientist, dietician, doctor, or hospital job (e.g. “advising & blogging”). For the non-expert it was incorrect if it did not refer to student, museum, or art/history (e.g. “health”). Answers merely stating blogger were incorrect for both conditions as this was not the bloggers’ daytime occupation. The remaining 109 participants in the analyses were on average 24 years old (SD = 5.52, MIN = 18, MAX = 55)
and the majority (84.4%) had finished higher professional education or higher. The number of participants was slightly higher than the intended sample of 100 women.
Procedure
Participants were randomized after agreeing to the informed consent and indicating their gender. Thereafter, they received either the expert or non-expert manipulation followed by a blog about antioxidants (adapted from Arora & Arora, 2004; see Appendix B) and a brief questionnaire concerning the outcome and confounding variables. The experiment ended with some questions on participants’ demographics and a message thanking them for their participation.
Pretests & Stimulus Materials
Blogger expertise was manipulated by varying the bloggers’ about me page (see Appendix C). The author’s credentials were varied in this page, because these function as expertise cues in online health information (Thon & Jucks, 2017), which blog readers may use to judge bloggers’ expertise (Buis & Carpenter, 2009). Furthermore, credentials have been used to manipulate source expertise in previous online health research (Hu & Sundar, 2010; Thon & Jucks, 2017). However, the manipulation by Hu and Sundar (2010) had a high failure rate, which was ascribed to the subtle nature of the manipulation. Therefore, the current study followed their suggestion to improve the manipulation by including a source profile. Thus, in the current study the about me pages of the bloggers differed in two ways. Specifically, the page of the expert (vs. non-expert) included medical credentials (vs. no credentials) as well as information on her medical (vs. non-health related) studies.
Two pretests were conducted to assess the differences in expertise between the expert and non-expert manipulation. In the first pretest, respondents received either the expert or the non-expert manipulation. They then indicated the extent to which they perceived the blogger to be a dietetics or nutrition expert by answering two 11-point Likert scales ranging from
strongly disagree to strongly agree (adapted from McCroskey & Richmond as cited in Hu & Sundar, 2010). The manipulation was unsuccessful. Specifically, an independent samples t-test revealed that the level of expertise in the expert condition (n = 10, M = 7.20, SD = 2.67) did not significantly differ from the non-expert condition (n = 10, M = 5.85, SD = 2.74), t(18) = -1.12, p = .279, 95% CI [-3.89, 1.19]. Therefore, the non-expert stimulus material was adapted by removing information about food-related spare-time activities and including more topics than only health in the blog description (see Appendix D). Furthermore, it was also discovered that the scale was set to 11-points instead of the intended 10-points. Therefore, the scale was adapted to 10-points for the second pretest. The second pretest, which followed the same procedure as the first pretest, revealed that the level of expertise in the expert condition (n = 9, M = 6.17, SD = 1.48) did not significantly differ from the non-expert condition (n = 6, M = 5.83, SD = 2.80), t(13) = -0.30, p = .767, 95% CI [-2.71, 2.05]. Therefore, the stimulus material for the non-expert condition was changed by making the “My mission” part of the about me page sound less professional (see Appendix E) and one additional item (e.g., concerning whether the blogger has professional nutrition/dietetics expertise) was added to the manipulation check measure (see Appendix F).
Measures
All variables in this study were observed on 10-point Likert scales ranging from strongly disagree to strongly agree unless specified elsewise (see Appendix F). To construct the scales, Principal Component Analyses and Reliability Analyses were conducted followed by computation of a mean scale if these analyses were satisfactory. Criteria for computation included: an eigenvalue equal to or higher than one, a clear bend in the scree plot after selected component(s), and all items loading unidirectional and higher than .45 on the selected factor. No items were left out to improve the scales as described below.
Attitudes towards the health behavior. Attitude towards the recommended health behavior in the blog was measured with a 5-item 10-point semantic differential scale concerning attitude towards consuming a diet high in antioxidants. The scale had endpoints such as: ‘good/bad idea’ and ‘necessary/unnecessary behavior’ (adapted from Arora & Arora, 2004). The five items together formed one factor (Eigenvalue = 4.16, EV = 83.24%) and a reliable scale (Cronbach’s α = .95, M = 7.66, SD = 1.62).
Behavioral intention towards the health behavior. Behavioral intention was measured with four 10-point Likert scales ranging from extremely unlikely to extremely likely. These included ‘How likely is it that you will act on the guidelines of the blog?’ and ‘How likely is it that you will recommend the guidelines of the blog to someone else?’ (adapted from Hu & Sundar, 2010). Sharing and recommendation intentions were included in this measure as previous research revealed that positive evaluations of online health
information led to higher recommendation intentions (Wang, Walther, Pingree, & Hawkins, 2008). The four items together formed one factor (Eigenvalue = 2.50, EV = 62.40%) and a reliable scale (Cronbach’s α = .80, M = 3.84, SD = 1.50).
Source expertise. Source expertise was measured with the three items as discussed for the pretests. The three items formed one factor (Eigenvalue = 2.92, EV = 97.21%). The source expertise scale was reliable (Cronbach’s α = .99, M = 5.21, SD = 2.99).
Source credibility. Source credibility was measured with four statements concerning the blogger (adapted from Jung et al., 2016). These included ‘The blogger is credible’ and ‘The blogger is successful’. The four items together formed one factor (Eigenvalue = 3.07, EV = 76.68%). The source credibility scale was reliable (Cronbach’s α = .90, M = 5.64, SD = 1.84).
Involvement. Involvement was measured via three statements concerning involvement with foods containing antioxidants, for example: ‘I am concerned about
including foods with antioxidants in my diet’ and ‘Information about foods with antioxidants is very relevant to me’ (adapted from Jung et al., 2016). The three items together formed one factor (Eigenvalue = 1.99, EV = 66.28%). The scale for involvement was reasonably reliable (Cronbach’s α = .73, M = 4.72, SD = 1.84).
Control variables. Eight control variables were measured (see Appendix G for statistics). Specifically, age was taken into account because it influences how Internet is used and how information retrieved from the Internet is assessed (Greenberg et al., 2013),
influences credibility of web information (Chuang, 2015), correlates with healthy eating attitudes (Sun, 2008), and use of the Internet for health purposes is more common among younger women (Hesse et al., 2005). Education was included as it impacts the use of online health information (Hesse et al., 2005; Wangberg, Andreassen, Kummervold, Wynn, & Sørensen, 2009). Individuals with higher levels of education used the Internet more frequently to find health information (Andreassen et al., 2007; Hesse et al., 2005).
Furthermore, source homophily, which can be described as the similarity between the source and receiver (Hu & Sundar, 2010; Wang et al., 2008), was included because it has been revealed to correlate with (Wright, 2000) and positively impact source credibility (Wang et al., 2008). Source homophily was measured with four statements concerning the blogger (adapted from McCroskey & Richmond as cited in Hu & Sundar, 2010). These included ‘The blogger is like me’ and ‘The blogger is similar to me’. Prior knowledge of the blog topic was assessed as this is linked to more systematic processing (Hu & Sundar, 2010), impacts the influence of source expertise on website credibility (Jung et al., 2016), and is a base for credibility decisions in online health communication (Thon & Jucks, 2017). Additionally, prior knowledge was measured by one self-report question. Participants had to indicate on a 10-point semantic differential scale whether they thought their knowledge about antioxidants was ‘bad/good’ (adapted from Thon & Jucks, 2017; Voedingscentrum, n.d.-a). Media
reliance was included, as it has been shown to influence perceived website expertise (Hong, 2006) and blog credibility directly (Johnson & Kaye, 2010), and behavioral intention
indirectly (Hong, 2006). Moreover, reliance has been argued to be the strongest predictor of online source credibility (Johnson & Kaye, 2010). Media experience was also included, as Internet experience is positively related to perceived Internet credibility (Flanagin & Metzger, 2000), and experience with a medium may influence the perception of and
intentions towards information from that medium (Rosenvinge et al. as cited in Hu & Sundar, 2010). Participants used the Internet most often to get information about diet and nutrition (n = 104), followed by other people (n = 68), and health professionals (n = 58; adapted from Jung et al., 2016). Those who relied on the Internet for diet/nutrition information also indicated their frequency of use on a 5-point scale ranging from never to every day (adapted from Jung et al., 2016). Most of the Internet experienced participants (61.5%) relied on the Internet for diet/nutrition information on at least a monthly basis. Lastly, dieting and BMI were measured because of their links to healthy eating. Dieting is described as the deliberate constriction of dietary intake to maintain one’s current weight (Neumark-Sztainer, Jefferey, & French, 1997) or in order to lose weight (Rideout & Barr, 2009). Participants indicated whether they were ‘not currently dieting’, ‘currently dieting to keep my weight where it is right now’, or ‘currently dieting to lose weight’ (adapted from Neumark-Sztainer et al., 1997). The answers were recoded into a yes-no format, by converting ‘not currently dieting’ into ‘not dieting’ (n = 76) and ‘dieting to lose weight’, and ‘dieting to keep my weight where it is right now’ into ‘dieting’ (n = 33) (Neumark-Sztainer et al., 1997). BMI was calculated based on participants’ self-reported weight and height. Calculations were executed following the formula [weight in kg/(2*length in m)]. Participants were then either categorized as underweight (BMI below 18.50; 51.9%), normal weight (BMI of 18.50-24.99; 46.3%), or as overweight (BMI of 25 or higher; 1.9%) (Verstuyf, Vansteenkiste, Soetens, & Soenens, 2016;
Voedingscentrum, n.d.-b). One person was considered a missing value for BMI because her weight answer was left blank.
Results Manipulation & Confound Checks
The manipulation was successful; an independent samples t-test revealed that the conditions significantly differed in the intended direction. Specifically, the level of expertise in the expert condition (n = 53, M = 7.77, SD = 1.43) was significantly higher than in the non-expert condition (n = 56, M = 2.78, SD = 1.82), t(103.46) = -15.98, p < .001, 95% CI [-5.61, -4.37].
Analyses for the control variables revealed that blog reliance was the only variable that significantly differed between the conditions t(107) = 2.15, p = .034, 95% CI [1.68, -0.07], (Mexpert = 5.02, SD = 2.10; Mnon-expert = 4.14, SD = 2.15) as revealed by an independent
samples t-test. Therefore, blog reliance was included in the analyses as a covariate. Hypotheses Testing
To investigate the impact of expertise on credibility, a one-way ANCOVA analysis with source expertise as the predictor and blog reliance as covariate revealed a significant, moderate positive effect of the covariate blog reliance, F(1, 106) = 16.73, p < .001, η2 = .14. Specifically, an increase in blog reliance led to an increase in source credibility. More importantly, the analysis revealed a significant, moderate to large effect of source expertise, F(1, 106) = 33.40, p < .001, η2 = .24. Inspection of the means revealed that expert health blog authors (M= 6.62, SD = 1.59) were perceived as more credible than non-experts (M = 4.71, SD = 1.57). Therefore, H1a was supported.
To test H1b, H1c, H2, and RQ1, moderated mediation analyses were conducted using PROCESS macro Model 8 for SPSS (Hayes, 2013). The model with 5,000 bootstrap samples had source credibility as independent variable, issue involvement as moderator, blog reliance
as covariate, attitude as mediator, and intentions as dependent variable. The analysis revealed that the model could be used to predict attitudes, F(4, 104) = 16.58, p < .001 and intention, F(5, 103) = 40.78, p < .001. Specifically, the model predicted 38.9% of the variance in attitude (R2 = .39) and 59.2% of the variance in intention (R2 = .59).
With respect to H1b, the model revealed that source credibility as induced by source expertise had a significant positive effect on attitude, b = 0.25, t(4, 104) = 3.24, p = .002, SE = 0.08, 95% CI [0.10, 0.40]. This means that higher levels of source credibility led to more positive attitudes towards the health behavior. Therefore, H1b was supported by the data. The covariate blog reliance did not have a significant impact on attitudes, b = 0.06, t(4, 104) = -0.98, p = .332, SE = 0.06, [-0.19, 0.06]. Interestingly, involvement had a significant direct impact on attitudes, b = 0.44, t(4, 104) = 6.45, p < .001, SE = 0.07, [0.30, 0.57]. This means that higher levels of involvement led to more positive attitudes towards the health behavior.
With respect to H1c, the model revealed that source credibility as induced by source expertise had a significant impact on intention, b = 0.18, t(5, 103) = 2.97, p = .004, SE = 0.06, [0.06, 0.30]. As expected higher levels of source credibility led to higher levels of behavioral intention. Thus, H1c was confirmed by the data. The covariate blog reliance also had a significant impact on intention, b = 0.17, t(5, 103) = 3.23, p = .002, SE = 0.05, [0.06, 0.27]. Interestingly, involvement had a significant direct impact on intention, b = 0.37, t(5, 103) = 6.05, p < .001, SE = 0.06, [0.25, 0.49]. This means that higher levels of involvement led to higher levels of behavioral intention.
With respect to the moderating impact of involvement on the relationship between credibility and attitudes, the analysis revealed issue involvement as a significant negative moderator of the relationship between source credibility and attitude, b = 0.06, t(4, 104) = -2.07, p = .041, SE = 0.03, [-0.12, -0.00]. This means that the impact of credibility on attitudes was higher for lower involved people as compared to higher involved people. This can also
be seen from the visual representation of these results (adapted from DeCoster & Iselin, 2015; see Appendix H), which reveals that the impact of credibility on attitudes is less strong for people who are higher involved than for women who are lower involved. Therefore, H2 was confirmed by the data.
Concerning RQ1, the model revealed that issue involvement was not a significant moderator of the relationship between source credibility and intentions b = 0.03, t(5, 103) = 1.06, p = .290, SE = 0.02, [-0.02, 0.07]. However, as previously discussed, involvement was a significant moderator of the relationship between source credibility and attitudes, and attitude was a significant mediator of the effect of source credibility on intentions, b = -0.01, SE = 0.01, [-0.03, -0.00]. The visualization of this effect (adapted from Dragt, 2017; see Appendix I) reveals that the conditional indirect effect of source credibility on intention is significant for all three levels of the moderator (e.g., high involved, averagely involved, and low involved). Furthermore, the impact of credibility on intention, as mediated by attitudes, is stronger for women who are low involved with the health behavior as compared to women who are averagely or high involved. To conclude, involvement did not moderate the
relationship between source credibility and intentions directly (RQ1a). However, the indirect effects of source credibility on intentions via attitudes were moderated by involvement (RQ1b). Specifically, the effect of source credibility on intentions was lower for higher involved women. See Appendix J for all results.
Conclusion & Discussion
The aim of this study was to investigate whether source expertise influences female health blog readers and if these effects differ for women who are more or less involved. Analyses were conducted to reveal whether the impact of source credibility as induced by source expertise on attitudes and intentions is stronger for women who are lower involved with the health behavior.
The results of an online experimental study in which women viewed the about me page of either an expert or a non-expert blogger, followed by a blog about foods with antioxidants, led to three conclusions. First, the results revealed that source expertise as disclosed via the blogger’s professional (non-)medical background enabled women to discern credible from less credible authors. Second, higher levels of blogger credibility in turn led to higher behavioral intentions and more positive attitudes towards the health behavior as disclosed by the blogger. These results seem to indicate that bloggers’ self-disclosed
professional background enables women to make distinctions between more and less credible authors, as well as adapt their attitudes and behavioral intentions accordingly.
Third, from the analyses it appeared that the recipient’s involvement with the health behavior moderated the direct relationship between credibility and attitudes and the indirect
relationship between credibility and intention via attitudes. The results indicate that higher blogger credibility leads to more positive attitudes and higher behavioral intentions, but more so for lower involved women. However, it should be noted that involvement did not
moderate the source effects on intention directly, as some researchers argued (Pan & Chang, 2011; Shao et al., 2004; Yoon et al., 2005). This result may be explained from the theory used to argue for the influence of involvement on the credibility-intention relationship. Specifically, the influence of involvement on intentions was based on the ELM reasoning concerning attitude (Petty et al., 1981). Thus, it may be the case that this reasoning is not applicable to intentions. To summarize, the results imply that health blogs by unprofessional sources should not necessarily be harmful to women’s attitudes and behavioral intentions, as source expertise enables them to differentiate between credible and less credible authors, and to adjust their attitudes and behavioral intentions accordingly.
The results also reveal two unexpected, but interesting, findings. First, the extent to which women rely on blogs for health significantly impacts credibility and intentions.
Specifically, reliance on blogs for health information leads to higher credibility and intentions. The higher credibility amongst women who rely on blogs may stem from their higher familiarity with the form and objectives of blogs (Johnson, Kaye, Bichard, & Wong, 2007). This higher credibility may in turn lead to higher behavioral intentions (Hong, 2006).
The second finding concerns the direct influence of involvement on attitudes and intention. Specifically, involvement caused the largest increase in both variables. These results may be explained via the link between involvement and prior knowledge. According to the ELM, simple cues such as source credibility affect persuasion in case of low prior knowledge, while high prior knowledge enables people to reinforce or counter message arguments (Petty & Cacioppo, 1986). Jung et al. (2016) provided evidence for this reasoning, finding that participants with low (versus high) prior knowledge were more likely to rely on the expertise of the source. Thus, in the current study it may have been the case that higher involved women had higher prior knowledge about antioxidants, which may have enabled them to judge the quality of the information in the blog regardless of bloggers’ expertise. Limitations & Future Research
There are four limitations of this study, which call for future research. First, the results of this study are based on a convenience sample. This type of sampling does not adhere to the precise standards of random probability sampling. Therefore, it may be the case that the results are representative of a specific subsample instead of the entire population (Babbie as cited in Johnson & Kaye, 2010). The results can therefore not be generalized to the entire population of female health blog readers (Johnson & Kaye, 2010). Future research should aim to replicate the results with a different sample.
Second, this study investigated a situation in which participants were explicitly asked to pay close attention to information about the blogger. However, it is not clear whether blog readers pay attention to information about the blogger in real-life. The substantial amount of
incorrect answers to the occupation question in this study implies that certainly not all
women pay attention to bloggers’ expertise. Similarly, Eysenbach and Köhler (2002) reported that none of the participants in their study looked at the “about us” page of the source.
Furthermore, Buis and Carpenter (2009) established that bloggers credentials are not always clearly stated, which further complicates the differences between this study and realistic settings. Future research should therefore investigate the impact of source expertise in a more naturalistic setting to be able to further generalize the results.
Third, this study focused on a controversial blog topic for three reasons. Specifically, to highlight bloggers’ expertise (Jung et al., 2016), make for an even distribution of
preexisting opinions (Hovland & Weiss, 1951), and to spur participants to attempt to verify the information and question the credibility (Hu & Sundar, 2010). However, it may be the case that the effects of the current study differ for less controversial topics. In this line of reasoning, absence of source expert effects has been attributed to participants’ familiarity with the study topic (Jung et al., 2016). Future research should therefore further explore source expert effects for blog topics with different levels of controversy.
Lastly, the source factors in this study explained a fair amount of variation in attitudes and intentions. However, future research should investigate more factors, as source effects have been revealed to have a small impact on persuasive outcomes (Wilson & Sherrel, 1993). Eastin (2001) similarly argued that although the source is often used to make credibility judgments, other content features could also affect perceptions and should therefore also be investigated. Message accuracy would be an example of such content feature. Specifically, the current study provided participants with a blog containing accurate information. Effects may however differ for blogs containing inaccurate information, as previous research revealed interactions between message accuracy and source expertise (Jung et al., 2016), as well as between message accuracy, source characteristics, and involvement (Reinhard &
Sporer, 2010). Future research should thus aim to investigate message accuracy as well as other features that possibly underlay the persuasive influence of health blogs.
Implications
Besides these limitations, the findings of this study have important practical and theoretical implications. The current study extends the theory on source persuasion by including healthy eating blogs, therein also furthering the limited research on blogs (Kaye, 2005) and health, and nutrition blogs (Lynch, 2010; Rausch, 2006). Specifically, this study was the first to underline that source credibility as induced by expertise impacts intentions and attitudes stemming from healthy eating blogs. The results reveal that providing women with information on bloggers’ expertise enables them to judge bloggers’ credibility and to develop their attitudes and intentions towards the health behavior accordingly. Thus, source credibility is an important factor that helps explain the impact of source expertise on
persuasive outcomes. It is therefore important to include source credibility in the theoretical underpinnings of persuasion in health blogs. Furthermore, this study was the second to successfully manipulate blogger expertise in a health blog. The results therefore offer two contributions to the theory on source manipulations in health blogs. Firstly, the findings reveal that subtle source manipulations are less effective in conveying expertise than more undisguised manipulations. Secondly, the findings indicate that providing the message recipient with a source profile is an effective way to convey source expertise, as was previously suggested by Hu and Sundar (2010).
These results also have important implications for female health information seekers and policy makers. Specifically, previous studies revealed that the quality of online health information is questionable (Boepple & Thompson, 2014; Carrotte et al., 2015; Modave et al., 2014) and that health information seekers are influenced by the content nonetheless (Ostry et al., 2008; Sillence et al., 2007). Thus, the ability to judge the quality of online health
information is important as an inability to do so increases the risk of believing and
implementing health misinformation (Yi, Yoon, Davis, & Lee, 2013). The findings of this study imply that in order to enable quality judgments, it is important to ensure that bloggers provide source expertise information, as well as to make blog readers aware of this
information. For policy makers it is important to translate these findings into public policy regarding health blogs. Furthermore, the majority (86%), of health information seekers has previously expressed concerns about getting health information from unreliable sources (Rainie & Fox, 2000). The results of this study reassure female health information seekers of their ability to judge the quality of information based on bloggers’ expertise.
This study also sheds light on the influence of source expertise on source credibility, attitudes, and intentions. These findings have both theoretical and practical implications. Specifically, the findings provide theoretical insight into the impact of online features on health-related perceptions, attitudes, and behaviors by revealing the influence of source expertise on health-related persuasive outcomes. This contribution is important because of peoples’ increased interest in online health information and the growing concerns about this content (Rains, 2007). Furthermore, these findings have practical implications for health practitioners. Specifically, according to Rains (2007), practitioners should develop programs to train health information seekers in evaluating online health information and Silberg, Lundberg, and Musacchio (1997) argue that authorship is one of the content factors to do so. Furthering these arguments, the results of the current study imply that authorship is indeed an important evaluative factor. However, these results in combination with previous research imply that instead of training women to judge authorship, it may be more helpful to make source expert information salient and confirmable (Thon & Jucks, 2017) and to train women to access authorship information. These findings are also important for professional medical bloggers. Specifically, in order to be seen as credible, they should adhere to professional
standards such as disclosing their professional medical credentials (Helm & Jones, 2006) and make this information clearly visible.
Lastly, this study contributes to the involvement literature, as it was the first to formally hypothesize that involvement and source credibility form the underlying mechanism of expert source persuasion in health blogs. It was also the first to comparatively investigate the direct and indirect involvement moderation of source effects on intentions. The results reveal that credibility and involvement both underlay expert source persuasion and that involvement moderates the impact of source effects on intentions indirectly via attitudes. Specifically, credibility strengthens expert source persuasion, whereas involvement dials the effect down.
To conclude, online health misinformation may have harmful consequences. The results of this study however indicate that source expertise enables women to criticize source credibility and to adjust their attitudes and behavioral intentions accordingly. Therein
maximizing help from healthy eating blogs authored by professional sources and minimizing harm provoked by unprofessional healthy eating bloggers.
Reference List
Andreassen, H. K., Bujnowska-Fedak, M. M., Chronaki, C. E., Dumitru, R. C., Pudule, I., Santana, S., ... Wynn, R. (2007). European citizens' use of E-health services: Q study of seven countries. BMC Public Health, 7(1), 1-7. doi:10.1186/1471-2458-7-53 Andrews, J. C., & Shimp, T. A. (1990). Effects of involvement, argument strength, and
source characteristics on central and peripheral processing of advertising. Psychology & Marketing, 7, 195-214. doi:10.1002/mar.4220070305
Arora, R., & Arora, A. (2004). The impact of message framing and credibility: Findings for nutritional guidelines. Services Marketing Quarterly, 26(1), 35-53.
Ayoob, K. T., Duyff, R. L., & Quagliani, D. (2002). Position of the American Dietetic Association: Food and nutrition misinformation. Journal of the American Dietetic Association, 102, 260-266. doi:10.1016/s0002-8223(02)90062-3
Bannister, B. D. (1986). Performance outcome feedback and attributional feedback: Interactive effects on recipient responses. Journal of Applied Psychology, 71, 203-210. doi:10.1037//0021-9010.71.2.203
Bansah, A. K. (2016). The effect of message credibility on attitude change as measured by argumentation style: In an introductory nutrition class (Doctoral dissertation). Retrieved from https://etd.ohiolink.edu/
Berland, G. K., Elliott, M. N., Morales, L. S., Algazy, J. I., Kravitz, R. L., Broder, M. S., … McGlynn, E. A. (2001). Health information on the Internet. JAMA: Journal of the American Medical Association, 285, 2612-2621. doi:10.1001/jama.285.20.2612 Bissonnette-Maheux, V., Provencher, V., Lapointe, A., Dugrenier, M., Dumas, A. A., Pluye,
P., ... Desroches, S. (2015). Exploring women’s beliefs and perceptions about healthy eating blogs: A qualitative study. Journal of Medical Internet Research, 17, e87. doi:10.2196/jmir.3504
Boepple, L., & Thompson, J. K. (2014). A content analysis of healthy living blogs: Evidence of content thematically consistent with dysfunctional eating attitudes and behaviors. International Journal of Eating Disorders, 47, 362-367. doi:10.1002/eat.22244 Braunsberger, K., & Munch, J. M. (1998). Source expertise versus experience effects in
hospital advertising. Journal of Services Marketing, 12, 23-38. doi:10.1108/08876049810202348
Buis, L. R., & Carpenter, S. (2009). Health and medical blog content and its relationships with blogger credentials and blog host. Health Communication, 24, 703–710. doi:10.1080/10410230903264014
Busch, P., & Wilson, D. T. (1976). An experimental analysis of a salesman's expert and referent bases of social power in the buyer-seller dyad. Journal of Marketing
Research, 13, 3-11. doi:10.2307/3150896
Carrotte, E. R., Psych, B., Vella, A. M., & Lim, M. S. (2015). Predictors of “liking” three types of health and fitness-related content on social media: A cross-sectional study. Journal of Medical Internet Research, 17, e205. doi:10.2196/jmir.4803
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:10.1037//0022-3514.39.5.752
Chuang, C. P. (2015). Effects of audience characteristics and sources of information on perceived credibility of web information. 2015 IEEE/WIC/ACM International
Conference on Web Intelligence and Intelligent Agent Technology (WI-IAT), 3,
110-113. doi:10.1109/wi-iat.2015.261
Cialdini, R. B., Petty, R. E., & Cacioppo, J. T. (1981). Attitude and attitude change. Annual
Review of Psychology, 32, 357-404. doi:10.1146/annurev.ps.32.020181.002041 Crisci, R., & Kassinove, H. (1973). Effect of perceived expertise, strength of advice, and
environmental setting on parental compliance. The Journal of Social Psychology, 89, 245-250. doi:10.1080/00224545.1973.9922597
Culver, J. D., Gerr, F., & Frumkin, H. (1997). Medical information on the Internet. Journal of
General Internal Medicine, 12, 466-470. doi:10.1046/j.1525-1497.1997.00084.x DeCoster, J., & Iselin, A.-M. (2015). Simple slopes plot for a 2-way interaction [Excel
spreadsheet]. Retrieved from http://www.stat-help.com/spreadsheets.html
Dragt, M. (2017). Graphing conditional indirect effects v1-3 [Excel spreadsheet]. Retrieved from http://www.md2c.nl/graphing-conditional-indirect-effects/
source expertise and knowledge of content. Journal of Computer-Mediated Communication, 6(4). doi:10.1111/j.1083-6101.2001.tb00126.x
Eastin, M. S., Yang, M. S., & Nathanson, A. I. (2006). Children of the net: An empirical exploration into the evaluation of Internet content. Journal of Broadcasting & Electronic Media, 50, 211-230. doi:10.1207/s15506878jobem5002_3
Eysenbach, G., & Köhler, C. (2002). How do consumers search for and appraise health information on the World Wide Web? Qualitative study using focus groups, usability tests, and in-depth interviews. BMJ, 324, 573-577. doi:10.1136/bmj.324.7337.573 Fallows, D. (2005). How women and men use the Internet (Research report). Retrieved from
Pew Research Center: Internet & Technology website:
http://www.pewinternet.org/2005/12/28/how-women-and-men-use-the-internet/ Flanagin, A. J., & Metzger, M. J. (2000). Perceptions of Internet information
credibility. Journalism & Mass Communication Quarterly, 77, 515-540. doi:10.1177/107769900007700304
Fogg, B. J., Swani, P., Treinen, M., Marshall, J., Laraki, O., Osipovich, A., … Shon, J.
(2001). What makes Web sites credible? In J. Jacko & A. Sears (Eds.), Proceedings of the SIGCHI Conference on Human Factors in Computing Systems (pp. 61-68). New York, NY: ACM. doi:10.1145/365024.365037
Fox, S. (2005). Health information online (Research Report). Retrieved from Pew Research Center: Internet & Technology website: http://www.pewinternet.org/2005/05/17/ health-information-online/
Fox, S. (2008, August 7). Three-quarters of Internet users seek health info [Web log message]. Retrieved from http://thehealthcareblog.com/blog/2008/08/07/three-quarters-of-internet-users-seek-health-info/
Retrieved from Pew Research Center: Internet & Technology website: http://www.pewinternet.org/ files/old-media/Files/Reports/2009/PIP_ Health _2009.pdf
Fox, S., & Rainie, L. (2002). Vital decisions: A Pew Internet health report (Research Report). Retrieved from Pew Research Center: Internet & Technology website:
http://www.pewinternet.org/2002/05/22/vital-decisions-a-pew-internet-health-report/ Goldsmith, J. (2000). How will the Internet change our health system? Health Affairs, 19(1),
148–156. doi:10.1377/hlthaff.19.1.148
Greenberg, S., Yaari, E., & Bar-Ilan, J. (2013). Perceived credibility of blogs on the Internet – the influence of age on the extent of criticism. Aslib Proceedings, 65, 4-18.
doi:10.1108/00012531311297159
Harris, P., Sillence, E., & Briggs, P. (2009). The effect of credibility-related design cues on responses to a web-based message about the breast cancer risks from alcohol: Randomized controlled trial. Journal of Medical Internet Research, 11, e37. doi:10.2196/jmir.1097
Hayes, A. F. (2013). Introduction to mediation, moderation, and conditional process analysis: A regression-based approach [eBook]. Retrieved from
https://www.ebscohost.com
Herring, S. C., Kouper, I., Paolillo, J. C., Scheidt, L. A., Tyworth, M., Welsch, P., ... Yu, N. (2005). Conversations in the blogosphere: An analysis "From the bottom up".
Proceedings of the 38th Annual Hawaii International Conference on System Sciences, 9, 107b. doi:10.1109/HICSS.2005.167
Hesse, B. W., Nelson, D. E., Kreps, G. L., Croyle, R. T., Arona, N. K., Rimer, B. K., & Viswanath, K. (2005). Trust and sources of health information. The Impact of the Internet and its implications for health care providers: Findings from the first Health
Information National Trends Survey. Archives of Internal Medicine, 165, 2618-2624. doi:10.1001/archinte.165.22.2618
Homer, P. M., & Kahle, L. R. (1990). Source expertise, time of source identification, and involvement in persuasion: An elaborative processing perspective. Journal of Advertising, 19(1), 30-39. doi:10.1080/00913367.1990.10673178
Hong, T. (2006). Contributing factors to the use of health-related websites. Journal of Health Communication, 11, 149-165. doi:10.1080/10810730500526679
Hovland, C. I., Janis, I. L., & Kelley, H. H. (1953). Communication and persuasion: Psychological studies of opinion change. New Haven, CT: Yale University Press. Hovland, C. I., & Weiss, W. (1951). The influence of source credibility on communication
effectiveness. Public Opinion Quarterly, 15, 635-650. doi:10.1086/266350 Hsu, C. L., & Lin, J. C. C. (2008). Acceptance of blog usage: The roles of technology
acceptance, social influence and knowledge sharing motivation. Information & Management, 45, 65-74. doi:10.1016/j.im.2007.11.001
Hu, Y., & Sundar, S. S. (2010). Effects of online health sources on credibility and behavioral intentions. Communication Research, 37, 105-132. doi:10.1177/0093650209351512 Huovila, J., & Saikkonen, S. (2016). Establishing credibility, constructing understanding: The
epistemic struggle over healthy eating in the Finnish dietetic blogosphere. Health: An
Interdisciplinary Journal for the Social Study of Health, Illness, and Medicine, 20,
383-400. doi:10.1177/1363459315595849
I’m a Foodie. (n.d.). Marijke Berkenpas [Web log message]. Retrieved from https://www.iamafoodie.nl/author/marijke-berkenpas/
Johnson, H. H., & Scileppi, J. A. (1969). Effects of ego-involvement conditions on attitude change to high and low credibility communicators. Journal of Personality and Social
Johnson, T. J., & Kaye, B. K. (2010). Choosing is believing? How Web gratifications and reliance affect Internet credibility among politically interested users. Atlantic Journal
of Communication, 18, 1-21. doi:10.1080/15456870903340431
Johnson, T. J., Kaye, B. K., Bichard, S. L., & Wong, W. J. (2007). Every blog has its day: Politically‐ interested Internet users’ perceptions of blog credibility. Journal of
Computer‐ Mediated Communication, 13, 100-122.
doi:10.1111/j.1083-6101.2007.00388.x
Johnston, A., & Warkentin, M. (2010). The influence of perceived source credibility on end user attitudes and intentions to comply with recommended IT actions. Journal of
Organizational and End User Computing, 22(3), 1-21.
doi:10.4018/joeuc.2010070101
Jung, E. H., Walsh-Childers, K., & Kim, H. S. (2016). Factors influencing the perceived credibility of diet-nutrition information web sites. Computers in Human Behavior, 58, 37-47. doi:10.1016/j.chb.2015.11.044
Kaye, B. K. (2005). It’s a blog, blog, blog world: Users and uses of weblogs. Atlantic Journal of Communication, 13, 73-95. doi:10.1207/s15456889ajc1302_2
Kelman, H. C. (1961). Processes of opinion change. Public Opinion Quarterly, 25, 57-78. doi:10.1086/266996
Leippe, M. R., & Elkin, R. A. (1987). When motives clash: Issue involvement and response involvement as determinants of persuasion. Journal of Personality and Social
Psychology, 52, 269-278. doi:10.1037//0022-3514.52.2.269
Lynch, M. (2010). Healthy habits or damaging diets: An exploratory study of a food blogging community. Ecology of Food and Nutrition, 49, 316-335.
Mandl, K. D., Kohane, I. S., & Brandt, A. M. (1998). Electronic patient-physician
communication: Problems and promise. Annals of Internal Medicine, 129, 495-500. doi:10.7326/0003-4819-129-6-199809150-00012
Metzger, M. J. (2007). Making sense of credibility on the Web: Models for evaluating online information and recommendations for future research. Journal of the American
Society for Information Science and Technology, 58, 2078-2091.
doi:10.1002/asi.20672
Modave, F., Shokar, N. K., Peñaranda, E., & Nguyen, N. (2014). Analysis of the accuracy of weight loss information search engine results on the Internet. American Journal of Public Health, 104, 1971-1978. doi:10.2105/ajph.2014.302070
Molassiotis, A., & Xu, M. (2004). Quality and safety issues of web-based information about herbal medicines in the treatment of cancer. Complementary Therapies in Medicine,
12, 217-227. doi:10.1016/j.ctim.2004.09.005
Moturu, S. T., Liu, H., & Johnson, W. G. (2008). Trust evaluation in health information on the World Wide Web. Proceedings of the 30th Annual International Conference of
the IEEE Engineering in Medicine and Biology Society, 1525-1528.
doi:10.1109/IEMBS.2008.4649459
Neumark‐ Sztainer, D., Jeffery, R. W., & French, S. A. (1997). Self‐ reported dieting: How should we ask? What does it mean? Associations between dieting and reported energy intake. International Journal of Eating Disorders, 22, 437-449.
doi:10.1002/(sici)1098-108x(199712)22:4<437::aid-eat9>3.3.co;2-x
Ohanian, R. (1991). The impact of celebrity spokepersons’ perceived image on consumers’ intention to purchase. Journal of Advertising Research, 31(1), 46-54.
