Student: Tulsi Kok Student number: 10576355
Thesis supervisor: dr. J.M.F. van Oosten Submission date: 06-06-2016
Word count: 8476
frames in news-items about science on students’
trust towards these items
Abstract
Many pieces of information contain expert frames (i.e., statements that information is based on or proven by scientific research), even when the information is not (entirely) true or supported. Based on the Elaboration Likelihood Model (ELM), expert frames are thought to instigate trust due to the mental connections that they incite – especially when information is processed through the peripheral route. This study explores whether the presence of expert frames in pieces of information, specifically newspaper articles, increases students’ trust towards the information in which the frame is present. Building on ELM it is expected that level of education and level of scepticism respectively moderate any effects that expert frames might have on a students’ trust towards the framed information. The research utilizes a between-subjects experimental 2x2x2-factorial design and data from 119 Dutch students. The results demonstrate that expert frames have a positive influence on students’ trust in the framed information. Moreover, despite the fact that both education and scepticism influence trust towards news-items as well, no moderating effects were found.
Table of contents
Abstract ... 2
1. Introduction ... 4
2. Theoretical framework ... 7
2.1 Trust in science communication ... 7
2.2 Expert frames ... 8
2.3 The Elaboration Likelihood Model and trust in science ... 9
2.5 Scepticism ... 11
3. Methodology ... 12
3.1 Research design ... 13
3.2 Participants and procedure ... 14
3.3 Stimuli ... 15
3.4 Measurements ... 16
3.5 Analysis... 17
4. Results ... 18
4.1 Effects of expert frames on trust ... 18
4.2 Interaction effect of education and expert frames on trust ... 19
4.3 The effects of scepticism and expert frames on trust ... 22
5. Discussion and conclusion ... 24
6. Literature ... 29
7. Appendix ... 36
7.1 Questionnaire ... 36
1. Introduction
Science. It is what drives us forward as a society. Whether it is through breakthroughs in medicine or the research and development of all sorts of technologies – we depend on science for everyday life. Yet, for a long period of time scientific knowledge was mainly accessible to scientists (Dijkgraaf, 2011). As a result, scientists first and foremost published their work for other scientists rather than for the public at large. During this time an atmosphere prevailed in which these experts were very critical of their fellow scientists’ work. Instead of simply taking all science for granted, they would point out hiatuses and criticize methods and findings. Robbert Dijkgraaf (2011), former president of the Royal Dutch Academy of Sciences, described this period as one of “belief in the ignorance of experts”.
Nowadays, science has become accessible to virtually anyone, anywhere – partially due to the arrival of the internet (Bromme & Goldman, 2014). However, the general public usually does not come into contact with science through scientific publications or direct involvement in science, but they experience it through the media (Becker, Schnabel, & Bouman, 2000; Hijmans, Pleijter, & Wester, 2003). This is problematic for two primary reasons. Firstly, journalists are not always able and willing to paint complete and accurate pictures of scientific publications (McComas & Simone, 2003; Moynihan, 2003; Ransohoff & Ransohoff, 2001; Singer, 1990). They do not unerringly possess the time and skills necessary to accurately translate generally complex scientific publications into news articles (Wilson, 2000). Countless studies have also proven omission and sensationalism in media reporting of science – problems that both stem from conflicting agendas of science and journalism (Kleemans, van Cauwenberge, D'Haenens, & Vettehen, 2008; Wang, 2013; Zaller, 2010). Secondly, genuine expert voices are diluted by the fact that anyone can claim anything on the internet (Leurs & Kroodsma, 2014).
Often, people put their trust in information based on aspects that are unrelated to the information’s profound contents (Scheman, 2001). This phenomenon can be explained by the Elaboration Likelihood Model (ELM), which states that people can process information using two distinct routes: the central route or the peripheral route (Cacioppo & Petty, 1984). The central route can only be used to process information when people are sufficiently able and willing to process information. The profound content of information then determines their attitude towards it. When people are
unable or unwilling to process information, they use the peripheral route to form an attitude. In this case, their attitude is mainly dependant on simple cues, like the number of arguments, the length of the article, and the kind of words that are used.
A particular cue that is presumed to increase people’s trust towards information is the expert frame (Holland et al., 2011). Expert frames are claims that affirm that information is based on or proven by scientific research that can be used to actuate trust towards information – even when it is actually false, inaccurate, incomplete or unsupported. Expert frames are expected to instigate trust based on ELM. The general public predominantly has a quite low understanding of and engagement with science (Allum, Sturgis, Tabourazi, & Brunton-Smith, 2008). Therefore, people are generally presumed to process science communication using the peripheral route, which means that simple cues – like expert frames – are their key indicators of trust. These cues are assumed to be distinctly powerful, as they are effective indicators of competence (Holland et al., 2011). It is hard for lay people to assess information in fields that they know little about (Shapin, 2004). Therefore, when people determine whether or not to trust information, they tend to focus on their perceptions of characteristics indicating competence (Mascaro & Sperber, 2009; Mills & Landrum, 2012; Shafto, Eaves, Navarro, & Perfors, 2012). Consequently, people are presumed to trust expert frames, regardless of their legitimacy. Hence, supposedly, a different atmosphere prevails today – one of ‘belief in the expertise of ignorants’.
Over the past decades, quite a lot of research has focussed on the effects of science communication on the publics’ trust in science in general (Houtman, Aupers, & Achterberg, 2012; Irwin, 2015; de Keere, 2010); effective science communication (Cribb & Hartomo, 2002; Dam, Bakker, & Dijkstra, 2014); and education and trust in science (Derks, 2000; Pion & Lipsey, 1981; Ream, Lewis, Echeverria, & Page, 2014; Smetana, Wenner, Settlage, & Mccoach, 2016). None of these papers, however, focussed on trust towards specific items of (science) communication. Nor did they investigate whether expert frames influence people’s trust in information. It is important to assess whether these frames affect people’s trust towards specific news-items, as science is mostly still experienced in a mediated fashion. This paper aims to fill that gap by investigating the following research question:
RQ1: Do expert frames that are incorporated into news-items influence people’s trust towards the news-items that they are incorporated in?
Expert frames do not inevitably influence people’s trust (Entman, 1993). Inherent attributes of the receivers – among which their education and level of scepticism – are presumed to moderate the effects of expert frames on people’s trust towards the news-items containing these frames (Shapin, 2004). Scientific educations, unlike non-scientific educations, instil in people certain skills that are necessary to effectively assess news-items containing (scientific) information on a substantive level (Janssens, 2002). These skills include reflexivity, abstraction, critical thinking, and the ability to understand and value science. Consequently, it is expected that scientifically educated people, unlike non-scientifically educated people, will be able to process science communication through the central route of ELM, making them less susceptive to trivial indicators of reliability and expertise (i.e. expert frames) (Miller, 2004; Shapin, 2004). This presumption can be translated into a second research question:
RQ2: Does a scientific or non-scientific education moderate the relationship between expert frames in news-items and the trust these frames promote towards the news-items that they are incorporated in?
Another inherent personality trait that is presumed to moderate the effects of expert frames in news-items on people’s trust is scepticism. Sceptical people tend to take a questioning stance towards every piece of information they encounter, rather than simply taking it for granted (Boush, Chung-Hyun, Kahle, & Batra, 1993; Pinkleton et al., 2012). It takes a certain level of engagement to be critical of information (Boush et al., 1993). Exclusively engaged people are able to process information using the central route (Cacioppo & Petty, 1984). Therefore, sceptical people are assumed to be less susceptible to expert frames than non-sceptical people. This assumption leads to the final research question:
RQ3: Does a persons’ level of scepticism moderate the relationship between expert frames in news-items and the trust these frames promote towards the news-items that they are incorporated in?
In order to answer these research questions a between-subjects experiment with two conditions was conducted. But first, a theoretical framework is given to elucidate the theories that underlie the research questions that are central to this paper.
2. Theoretical framework
In this chapter the key concepts are defined. Additionally, existing theories and various scientific findings that form the foundation of the research questions that are central to this paper are discussed.
2.1 Trust in science communication
What is trust? Previous research on trust has focused on a model of trust
based on the three types of actors that people place their trust in: themselves, other individuals, and organisations or groups of individuals that they do not know (Abdul-Rahman & Hailes, 2000; Deutsch, 1973; Kelton, Fleischmann, & Wallace, 2008). The first level of this model is ‘dispositional trust’, which focuses on an individuals’ inherent personality characteristics and experiences in making trust-related decisions (Abdul-Rahman & Hailes, 2000). The second level, ‘interpersonal trust’, describes trust relationships between actors (Deutsch, 1973; Mui, Mohtashemi & Halberstadt, 2001). This type of trust concerns a trustor (the actor that trusts) determining whether to put trust in a trustee (the actor that is trusted). The third level, ‘institutional trust’, focuses on trust towards groups of people (unknown to the trustor) and organisations (Barber, 1983; Kini & Choobineh, 1998).
Interpersonal trust can be defined as the willingness of a person to be
vulnerable to an actor based on the expectation that the actor will perform a particular task adequately, irrespective of the trustors’ ability to monitor or control that other party (Mayer, Davis, & Schoorman, 1995). The focus in this paper mainly lies on interpersonal trust, as this is the appropriate level of trust to assess a persons’ trust in news-items (Kelton et al., 2008). News-items are produced by an actor (trustee) and communicated to an individual (trustor) through the media. Therefore, when individuals trust in news-items, they actually put their trust in the actor that created
them. Whilst determining whether or not to trust an actor, people use trust antecedents that determine the perceived source credibility (Mayer et al., 1995). In doing so, they estimate certain characteristics of a trustee, deemed relevant to trustworthiness by the trustor. These typically include factors such as perceived competence, intentions, and openness.
Expectations of trust. There are two expectations of trust (Barber, 1983).
Firstly, people expect actors to be technically competent. When it comes to science communication, for instance, people place their trust in journalists to sufficiently comprehend and translate scientific publications into news-items (McChesney, 2004). Secondly, people expect actors to fulfil their obligations and responsibilities (Barber, 1983). For instance, we expect both scientists and journalists to put their moral obligations and responsibilities before their personal agendas (McChesney, 2004). Trust in epistemic matters is particularly important, as one person can never be fully informed about everything – not even in their own fields of specialization (Powys Whyte & Crease, 2010). Therefore, people inevitably and consistently have to rely on what other actors say.
2.2 Expert frames
One simple, yet powerful way to enforce people’s trust towards information provided, is by framing the information. When information is framed, it is put into a field of meaning or a certain context by selecting and emphasizing particular aspects of the information in order to drive people to subconsciously form certain mental connections (Entman, 1993). In this thesis the focus specifically lies on expert frames: scientific claims that are incorporated into pieces of communication with the purpose of conveying the fact that the information provided is based on or proven by scientific research (Holland et al., 2011). Expert frames can be incorporated into articles to persuade people of the reliability of the article’s contents (Gard & Wright, 2005; Holland et al., 2011). Expert frames are expected to have persuasive effects on people for two reasons.
Firstly, it is hard for people to assess expertise in fields that they know little about (Shapin, 2004). Hence, people tend to focus on their perceptions of characteristics indicating competence when they form an attitude towards such information (Mascaro & Sperber, 2009; Mills & Landrum, 2012; Shafto et al., 2012).
Expert frames, due to the mental connections that they incite, are effective indicators of competence (Holland et al., 2011). These frames do not communicate the expertise of the trustee, but they corroborate the trustee’s statements. Consequently, lay people are expected to take expert frames for granted.
Secondly, expertise is associated with concepts rather than words (Kendall, 2004). People, for instance, are presumed to put their trust in the concept of science in general, rather than in each individual claim of science that they encounter. When an expert frame is incorporated into a piece of information, this drives people to form mental connections that derive from the opinions and beliefs that they hold towards science in general. Despite the commonly quite low public understanding of science (PUS) and public engagement with science (PES), people generally maintain a high level of esteem for and deference to scientific research (Allum et al., 2008; Collins & Evans, 2007; Miller, 2004; Moore, 2008). Subsequently, when people read a news-item that contains an expert frame, they are assumed to rely more on the scientific claim than on the articles’ actual contents when forming trust (Kendall, 2004). Therefore, it is expected that people are more likely to put their trust in the content of an article that contains a scientific claim than in one that does not. This assumption can be explained on the basis of the Elaboration-Likelihood Model.
2.3 The Elaboration Likelihood Model and trust in science
ELM proposes that people can process information using two distinct routes: the central route or the peripheral route (Cacioppo & Petty, 1984). People’s ability (capability for critical assessment) and engagement (desire to process information) determine which of the routes they use to process information. People can only process information using the central route when they have sufficient abilities and are engaged. When people use this route to process information, they carefully and thoughtfully consider the information at hand – a process that requires relatively much time and effort. If people lack skills of engagement, they can only process information using the peripheral route. In that case, information is assessed on a more trivial basis and simple cues are key determinants of the attitudes that people form. This process is significantly less time and effort consuming.
People can use either the central route or the peripheral route to form initial trust towards news-items on science (Cacioppo & Petty, 1984). When people have a
high ability to process scientific information and they perceive the information as relevant, they are likely to use the central route. In this case, people use relevant associations and past experiences to assess the information and form an attitude with respect to news-items. However, when people lack ability or are not engaged, they tend to use the peripheral route to process news-items. When this route is used, people base their trust on simple cues in a message rather than on its substantive contents. These cues can pertain to the source credibility, the number of arguments, the length of the article, the kind of words that are used, and – more importantly – expert frames.
Due to their low PUS and PES, the general public is likely to process news-items about (scientific) findings using the peripheral route (Collins & Evans, 2007; Moore, 2008). Hence, they will often base their trust towards these news-items on simple cues rather than on its profound contents. The mental connections that these cues provoke instigate trust. The presence of expert frames, for instance, incites people to base their trust on mental connections that derive from the opinions and beliefs that they hold towards science in general (Kendall, 2004). People generally maintain a high level of deference to science – making expert frames highly effective tools for instigating trust in people who process information using the peripheral route (Miller, 2004). This leads to the first hypothesis.
H1: The presence of an expert frame has a positive effect on people’s trust towards a news-item containing (scientific) information.
However, frames do not perpetually influence the way people think (Entman, 1993). Nor do all people form the same mental connections when they are exposed to the same frame (Covey, & Kazak, 2014). Expert framing effects on people’s trust in items of communication therefore differ amongst individuals. Inherent attributes of the receivers are presumed to influence the effects that expert frames have on different people. Two potentially relevant attributes are a persons’ education and their level of scepticism.
2.4 Scientific education
Scientific educations, unlike non-scientific educations, teach people to become scientifically literate (Bodmer, 1985; Janssens, 2002). Scientifically literate citizens have a basic vocabulary of scientific terms and constructs as well as a general
understanding of the nature of scientific inquiry (Miller, 2004). According to ELM, people can only process information using the central route when they have sufficient abilities to do so (Cacioppo & Petty, 1984). In other words, people need to be scientifically literate in order to be able to process news-items on (scientific) findings through the central route (Kruger & Drunning, 1999; Shapin, 2004). In this study a distinction was made between university students, who are scientifically educated and therefore presumed to be scientifically literate, and all other students, who are not scientifically educated and therefore presumed to be scientifically illiterate (Bodmer, 1985; Janssens, 2002).
Scientifically illiterate people generally do not possess the abilities necessary to process news-items containing scientific information through the central route (Miller, 2004; Shapin, 2004). Regardless of their level of involvement, they can only process these news-items through the peripheral route – making them susceptible to simple cues like expert frames (Cacioppo & Petty, 1984). Scientifically educated people, on the other hand, have the ability to process news-items on scientific information through the central route – depending on their involvement (Shapin, 2004). When people process news-items through the central route, they focus on the profound contents of these items (Cacioppo & Petty, 1984). Simple cues (i.e. expert frames) have little impact on people when they use this route when forming trust towards information. Therefore, scientific education is expected to moderate the effects of expert frames on people’s trust towards an item of science communication. This assumption can be translated into the second hypothesis.
H2: People who have obtained or are currently enrolled in a scientific education are less trusting towards information containing expert frames than people who have obtained or are currently enrolled in a non-scientific education.
2.5 Scepticism
Similarly to ones level of education, scepticism is another inherent personality trait that is presumed to moderate initial trust in news-items containing expert frames. Scepticism is a constructive type of distrust that refers to a persons’ critical yet open attitude towards information (Pinkleton et al., 2012). This type of distrust manifests in
a questioning approach towards each individual piece of information that people encounter (Carr, Barnidge, Gu Lee, & Jean Tsang, 2014). In other words, sceptical people do not merely accept information because somebody claims it to be true. Rather, they will critically assess information and actively try to verify it. This verification process can happen in two ways: people can either verify information based on their previous experiences and prior knowledge or they can actively seek additional information (Pinkleton et al., 2012). Hence, a sceptical personality drives people to judge news-items on science based on their substantive contents rather than on other aspects of those news-items.
As a result, scepticism is expected to moderate the effect of expert frames on trust toward news-items for two primary reasons. Firstly, sceptical people, unlike non-sceptical people, do not put their trust in concepts over contents (Pinkleton et al., 2012). They approach each piece of science communication that they encounter in a critical manner, unchanged by any opinions that they hold towards science or the media in general. Therefore, the aforementioned mental connections that expert frames promote, are expected to be less effective when it comes to sceptical people. Secondly, in order to take a critical stance towards information provided in news-items, people need to be sufficiently engaged (Boush et al., 1993). Engaged people, provided they are able to process science related information, can process news-items on (scientific) findings trough the central route (Boush et al., 1993; Cacioppo & Petty, 1984). Simple cues like expert frames are overlooked when the central route is used to process information. Hence, it is expected that:
H3: Sceptical people are less trusting towards news-items containing expert frames than non-sceptical people.
3. Methodology
In this chapter the methodology is explained that was utilized in order to test the research questions that underlie this paper. First the sampling method and the final sample are elucidated, after which the procedure, stimuli, measurements and analyses are explained.
3.1 Research design
The only research method that can be used to examine whether a causal relationship exists between variables, is an experiment (Gravetter & Forzano, 2012). The aim of this paper was to test whether a single independent variable (expert frame) has causal effects on a single dependent variable (trust), while measuring the potentially moderating effects of two variables (education and scepticism) on this relationship. Therefore, an experiment was conducted. Each of the variables consisted of two levels: expert frame (present of absent), education (scientific or non-scientific), and scepticism (low or high scepticism). The independent variable was manipulated, whilst the moderating variables were merely measured – education and scepticism can not be manipulated. Hence the experiment consisted of two conditions, despite of the 2x2x2-factorial design.
A between-subjects experimental design was chosen. In this type of experiment each participant is subjected to only one of the treatments, which keeps testing effects like boredom and habituation to a minimum (Gravetter & Forzano, 2012). Downsides of a between-subjects design are that more respondents are needed in order to get reliable results and that it is harder to determine whether there are individual differences between respondents in both experimental conditions. These problems were accounted for by targeting a group of participants that was homogenous in age and stage of life – all participants were students at the time the study was conducted and most of them were in their twenties. In doing so, individual differences between respondents and within the experimental groups were minimized. Secondly, Qualtrics – the online survey tool that was used to conduct the experiment – automatically and randomly divided participants over the experimental group and the control group in order to keep differences between both conditions to a minimum. An online experiment was conducted. This type of experiment is relatively easy, quick, and cheap (Gravetter & Forzano, 2012). The survey was made compatible for desktops as well as mobile devices, which enabled participants to take part in the experiment virtually anywhere, making the setting of the experiment very natural. The online experiment was completely anonymous. This point was stressed to participants twice in the briefing text and the informed consent that they were shown in Qualtrics before the questions started. When surveys are anonymous this tends to
lead to less socially desirable answering, making the results more reliable (He et al., 2015).
Online experiments have some disadvantages too. Firstly, they make it hard to control external factors that influence participants while they are taking part in the experiment. Some participants may have been able to focus better on the questions and stimulus presented to them than others, due to external factors like noise, disruption, etc. Secondly the online tool and the sampling methods used, make it impossible to determine the non-response rate, as it can not be measured how many people saw the online experiment and what part of them actually participated. Thirdly, the survey was kept relatively brief in order to minimize potential problems of survey-fatigue, like sloppy responses and non-response. Additionally, reverse questions were added to the scales in order to keep people alert and as a means to determine whether people had attentively taken part in the experiment.
3.2 Participants and procedure
Two sampling methods were used in order to reach the participants: a convenience sample and a snowball sample. First, potential participants – acquaintances of the researcher – were approached via email and social media. They were asked to take part in the online experiment and to share it within their social circles. Both sampling methods were primarily used for convenience purposes, as this study had to be conducted in a short period of time with limited means and funding. Since potential moderating effects were inspected, a minimum of 50 participants was required per condition in order to get reliable results (Gravetter & Forzano, 2012). Thus, the aim was to have a minimum of 100 students participate in the experiment.
All participants were shown an informed consent form in Qualtrics that they had to agree to before they could take part in the experiment, which informed them on their rights to decline or cease taking part in the survey at any given time without having to explain themselves, their full anonymity, and the fact that taking part in the survey would in no way harm them or put them in danger. Since giving participants prior insight into the true purpose of the experiment would have impaired the reliability of their responses, participants were told that the purpose of the experiment was to test what people remembered from newspaper articles. After the experiment ended, the debriefing explained the true purpose of the experiment.
163 people took part in the experiment. Participants who did not finish the experiment were removed from the dataset (N = 36). A manipulation check was included in the survey to test whether participants read, understood, and paid attention to the manipulation. This was a closed-ended question with three answering options about the topic of the stimuli. Of the remaining 127 respondents, 119 respondents answered it correctly (93.7%). The other 8 participants were removed from the dataset and were not taken into account during any of the analyses. The final sample consisted of 119 people, aged 18 to 28 (M = 21.96, SD = 2.30), 79 of which were women (66.3%). An overview of the final sample can be found in table 1. The number of women relative to men in this sample differed significantly from the variation in gender amongst Dutch students, t (118) = -1142.05, p < 0.001, d = 0.78 (Wobma, 2011). Participants in the experimental condition and control condition did not differ significantly in gender, t (116) = -0.07, p = 0.949, age, t (117) = 1.45, p = 0.151, and level of education t (117) = 1.20, p = 0.233.
Men Women Total
Mbo 11 14 25
Hbo 14 25 39
Wo 15 40 55
Total 40 79 119
Table 1 - education level and gender of the final sample
3.3 Stimuli
The stimuli were forged newspaper articles made to look like they were screenshots from the digital edition of the NRC Handelsblad – a Dutch quality newspaper (Bakker & Scholten, 2011). All the information in the stimuli was fictional – the articles explained the influence of genes on people’s preferences when it comes to appearances. This topic and these findings were chosen based on the confirmation bias theory, which states that people tend to believe things that they already think (Howgego, 2015). Therefore, if the findings in the stimuli had been to obvious, people would not need proof (i.e., expert frames) to believe them and take them for granted rather easily. In the stimulus for the experimental condition an expert frame was
incorporated (i.e., a claim that the discussed findings were based on scientific research by researchers from Utrecht University). In the stimulus for the control condition no such frame was present. The stimuli were otherwise identical. Due to the sampling methods it was expected that most scientifically educated participants would be enrolled at the University of Amsterdam. Therefore, a Dutch University of a similar ranking, was chosen for the expert frame stimulus. The stimuli were kept relatively brief in order to retain participants’ focus and motivation to finish the experiment. Snippets of the stimuli are listed in the appendix.
3.4 Measurements
The dependent variable trust in the stimuli was measured using five indicators of trust that were previously proven to make a reliable scale: whether it is fair, unbiased, tells the whole story, is accurate, and can be trusted (Meyer, 1988; Trumbo & McComas, 2003). This scale measures trust towards a specific item rather than trust in general. The five indicators formed separate questions in the survey, each of which was to be answered on a 7-point scale ranging from 1 (completely disagree) to 7 (completely agree). Sample items included “the information in the article is biased” (reverse coded; M = 4.34, SD = 1.42), “the information in the article is accurate” (M = 4.01, SD = 1.31), and “the article can be trusted” (M = 4.19, SD = 1.51). To test whether these items made a valid and reliable scale a factor analysis with varimax rotation was carried out. The items loaded on one component, with an explained variance of 53.5%. The reliability of the items was determined using Cronbach’s Alpha, which appeared to be acceptably reliable α = 0.77. The reliability of the scale could have been improved to α = 0.78 if the question ‘the information in the article is biased’ had not been included. It was chosen not to, however, as the increase in the reliability of the scale was relatively small, but would have led to a significant loss in information – the scale would still have been acceptably reliable. Participant responses to the five items were averaged in an overall measure of trust towards the newspaper article (M = 4.04, SD = 1.01). The composed scale of trust appeared to follow a normal distribution.
The first moderating variable, education level, was measured by a closed-ended question asking participants what their highest level of education was (1 = no education, 2 = primary school, 3 = high-school, 4 = mbo, 5 = hbo, and 6 = wo). The
responses to this variable were recoded into scientific education (N = 55, 46.2%) and non-scientific education (N = 64, 53.8%), as the focus in this paper lies on that distinction.
The second moderating variable, scepticism, was measured using a 7-point scale (1 = hardly sceptical, 7 = very sceptical). Seven scale questions were formulated based on a scale that was created to measure scepticism towards advertisements (Obermiller & Spangenberg, 1998), and included questions on whether newspapers tell the truth, inform us, see informing us as their main goal, are reliable, give a complete picture, make us feel like we are well informed, and provide us with valuable information. Some of these scale questions were “we can assume that most newspaper articles tell the truth” (M = 4.15, SD = 1.68) and “newspapers are a trustworthy source of information on science and research” (M = 4.08, SD = 1.55). To test whether these items made a valid and reliable scale, a factor analysis with varimax rotation was carried out. The items loaded on one component, with an explained variance of 51.1%. Cronbach’s Alpha was α = 0.83 – hence, the scale was reliable. A mean scale was computed (M = 3.50, SD = 1.02). The composed scale of scepticism seems to be normally distributed. Scepticism was recoded into two categories using median split (Md = 3,57). All responses lower than the median were labelled ‘low scepticism’ (N = 58, 48.7%) and all responses higher than the median were labelled ‘high scepticism’ (N = 61, 51.3%).
3.5 Analysis
All three hypotheses were tested using several two-way analyses of variance. Hypothesis one was tested by analysing the main effects of the independent variable expert frame on the dependent variable trust towards the item of science communication. Interaction effects show the effects that a moderating variable has on the main effect between the independent variable and the dependent variable. Therefore, hypotheses two and three were tested by inspecting the interaction effects of education and scepticism as the moderating variables in two respective two-way analyses of variance that both had expert frame as the independent variable and trust as the dependent variable.
4. Results
Several analyses of variance were performed in order to find correlations and test the hypotheses that underlie this paper. In this chapter the research findings are discussed.
4.1 Effects of expert frames on trust
In order to test the first two hypotheses a two-way analysis of variance with the composed scale of trust as the dependent variable and expert frame and education as independent variables was conducted. To test the first hypothesis – that expert frames have a positive effect on people’s trust in news items on scientific information – exclusively the main effect of expert frames on trust was analysed. The two-way analysis of variance with trust as the dependent variable and expert frame and education as independent variables showed that the presence of an expert frame in a newspaper article had a significant, small to medium effect on participants’ trust in the stimuli, F (1, 115) = 3.62, p < 0.01, η2 = 0.03. Participants who read the newspaper article with an expert frame were significantly more trusting of the article (M = 4.21, SD = 0.13) than participants who read the newspaper article without the expert frame (M = 3.86, SD = 0.13). Figure 2 shows the difference in trust between the experimental condition and the control condition. These findings supported the hypothesis that the presence of an expert frame has a positive effect on people’s trust towards a news item on (scientific) findings. Levene’s test of equality of variances indicated that the size of all testing groups featured in this analysis did not significantly differ, Levene’s F (3, 115) = 2.05, p = 0.111. Therefore, the two-way analysis of variance could be conducted reliably.
Figure 2 - the effect of expert frames on trust
4.2 Interaction effect of education and expert frames on trust
In the second hypothesis it was postulated that people with a non-scientific education would be more trusting of information containing an expert frame than people with a scientific education. To test this hypothesis, the interaction effect of education and expert frames on trust was examined. The two-way analysis of variance with trust as the dependent variable and expert frame and education as independent variables showed that expert frame and participants’ education did not have a significant interaction effect on trust towards specific items of information, F (1, 115) = 0.48, p = 0.828. Participants with a non-scientific education who were shown the newspaper article with an expert frame were slightly more trusting (M = 4.26, SD = 0.19) of the contents of the article than participants with a scientific education (M = 4.16, SD = 0.18). However, this difference was not significant. Similarly, participants with a non-scientific education who were shown the newspaper article without an expert frame were slightly – but insignificantly – more trusting (M = 3.95, SD = 0.17) of the contents of the article than participants with a scientific education (M = 3.77, SD = 0.17). Therefore, the hypothesis that a (scientific) education moderates the effect that an expert frame has on a participants’ trust in a news item on (scientific) findings was not supported. Figure 3 shows a graphic representation of these results.
4.21 3.86 1 2 3 4 5 6 7
Expert frame No expert frame
T rust in the sti mul us
The same two-way analysis of variance showed that a scientific education did not have any significant main effects on participants’ trust in news-items, regardless of the presence or absence of expert frame in those articles, F (1, 115) = 0,58, p = 0,449. Participants with a scientific education were slightly but insignificantly less trusting of the stimuli (M = 3,97, SD = 0,14) than participants with a non-scientific education (M = 4,11, SD = 0,13).
Figure 3 - the effect of scientific or non-scientific education on the influence of expert frames on trust
No significant main or interaction effects of education were found on participants’ trust towards newspaper articles when a distinction was made between scientific or non-scientific education. A subsequent two-way analysis of variance was conducted in order to determine whether education and expert frames had a significant interaction effect on participants trust when level of education (mbo/hbo/wo) was used as an independent variable instead of scientific or non-scientific education. A two-way analysis of variance with level of education and expert frames as the independent variables and trust as the dependent variable, showed that a participants’ level of education and expert frames did not have a significant interaction effect on participants’ trust towards the stimuli, F (2, 113) = 0,55, p = 0,580. Participants with an mbo education who saw the news-item with the expert frame (M = 4,77, SD = 0,27) were slightly more trusting of its contents than those who saw the news-item without the expert frame (M = 4,23, SD = 0,28).
3.95 3.77 4.26 4.16 1 2 3 4 5 6 7
Scientific education Non-scientific education
T rust in the sti mul us Expert frame No expert frame
Similarly, participants with an hbo education who saw the manipulation with the expert frame (M = 3,85, SD = 0,24) were somewhat more trusting than those who saw the one without the expert frame (M = 3,80, SD = 0,20). Participants with a wo education who were shown the manipulation with an expert frame (M = 4,16, SD = 0,18) were also slightly more trusting of the contents of the article than those who were shown the manipulation without an expert frame (M = 3,77, SD = 0,20). None of these differences, however, were significant. This analysis could be reliably conducted, as the variances in this analysis did not significantly differ from each other, Levene’s F (5, 123) = 2,44, p = 0,038. An overview of these findings is shown in figure 4.
When the main effects of the same two-way analysis of variance with level of education and expert frames as the independent variables and trust as the dependent variable were inspected, they displayed that education did have significant, small to medium effects on participants’ trust towards news-items, F (2, 113) = 3,83, p < 0,05, η2
= 0,06. Participants whose highest level of education is mbo are the most trusting towards the manipulation (M = 4,50, SD = 0,20). Participants whose highest level of education is hbo (M = 3,83, SD = 0,16) are slightly less trusting than participants whose highest level of education is wo (M = 3,97, SD = 0,13). A post-hoc Bonferroni test showed that there is only a significant difference in trust between participants with an mbo education and participants with either an hbo (MDIFFERENCE = 0,69, p <
0,05) or a wo education (MDIFFERENCE = 0,52, p < 0,1). There is no significant
difference in trust between participants whose highest level of education is hbo or wo (MDIFFERENCE = 0,17, p = 1,0). A factor that may have impacted the reliability of these
results, however, is the fact that the variances in this analysis were not equal, Levene’s F (5, 113) = 2,23, p < 0,1. Figure 5 shows the difference in trust between mbo, hbo and wo students.
Figure 4 - the effect of education level on the influence of expert frames on trust
Figure 5 - the effect of education level on trust
4.3 The effects of scepticism and expert frames on trust
To test the hypothesis that people with high scepticism are less trusting towards information containing expert frames than people with low scepticism a two-way analysis of variance with scepticism and expert frame as the independent variables and trust as the dependent variable was conducted. The variances in this analysis did not significantly differ from each other, Levene’s F (3, 115) = 1.16, p =
4.77 3.85 4.16 4.23 3.80 3.77 0 1 2 3 4 5 6 7 mbo hbo wo T rust in the sti mul us expert frame no expert frame 4.50 3.83 3.97 1 2 3 4 5 6 7 mbo hbo wo T rust in the sti mul us
0.329. The analysis could therefore be conducted reliably. The analysis showed that there was no interaction effect between scepticism and expert frames on participants’ trust, F (1, 115) = 2.51, p = 0.116. Participants who scored high on the scale for scepticism and were shown the manipulation with the expert frame (M = 3.83, SD = 0.15) were slightly more trusting of the manipulation than those who were shown the manipulation without the expert frame (M = 3.57, SD = 0.19). Likewise, participants with low scepticism that were shown the news-item with the expert frame (M = 4.87, SD = 0.20) were somewhat more trusting of the manipulation than those who were shown the manipulation without the expert frame (M = 4.07, SD = 0.15). Neither one of these interaction effects, however, proved to be significant. The results of this analysis are shown in figure 6. Hence, the hypothesis that people with high scepticism were less trusting towards information containing expert frames than people with low scepticism was not supported by these findings.
The same analysis of variance with scepticism and expert frame as the independent variables and trust as the dependent variable showed that scepticism had a significant, medium to large main effect on trust, F (1, 115) = 19.39, p < 0.001, η2 = 0.14. There are significant differences in participants’ trust towards the manipulation between participants with high scepticism (M = 3,70, SD = 0,12) and those with low scepticism (M = 4,47, SD = 0,13).
Figure 6 - the effect scepticism on the influence of expert frames on trust 4.87 3.83 4.07 3.57 0 1 2 3 4 5 6 7
Low scepticism High scepticism
T rust in the sti mul us expert frame no expert frame
5. Discussion and conclusion
This study explored whether the presence of expert frames in pieces of information, specifically newspaper articles, increased students’ trust towards that information. Furthermore, the influence of scientific education and scepticism on this relationship was investigated. In line with expectations, it was found that expert frames influence students’ trust in newspaper articles that these frames are integrated in. This effect can be explained on the basis of ELM. The general public commonly has a low understanding of and engagement with science – causing them to process information on (scientific) findings through the peripheral route. Simple cues are key determinants of trust formation when this route is used to process information. Expert frames are potentially powerful cues, due to the mental connections that they incite. They drive people to put their trust in the concept of science, which people generally hold in a high regard, rather than in an individual piece of science communication. Hence, expert frames increase students’ trust in news-items on (scientific) findings.
According to ELM it depends on people’s engagement and abilities whether they process information through the central route or the peripheral route. It was expected that scientifically educated people were scientifically literate, making them sufficiently able to process scientific news-items through the central route. Additionally, sceptical people critically assess information based on internal or external verification before they form an attitude towards it. A critical assessment of such requires engagement. Therefore, it was expected that sceptical people would be sufficiently engaged to process news-items on science through the central route. Depending on the route that people use to process information, distinctive information characteristics are prevalent in conveying credibility. Simple cues, like expert frames, become less important when the central route is used. Accordingly, it was expected that a scientific education and a high level of scepticism would moderate the influence that expert frames have on students’ trust. Contrary to what was expected, no moderating effects were found for education or scepticism. Both intrinsic personality characteristics, however, have a direct effect on trust.
It could be the case that no moderating effect of education was found due to the fact that scientifically literate students were not engaged enough to process the information in the news-items through the central route. The fact that students are, theoretically speaking, able to process science communication through the central
route, does not mean that they do. They might still process pieces of science communication through the peripheral route, making them as susceptible to the influences of expert frames as scientifically illiterate students. Another possible explanation is that, although technically speaking exclusively university students are enrolled in a scientific education, students of other education levels are also scientifically literate to a certain extent. The findings suggest that there is hardly any difference in trust towards items of science communication between hbo and wo students, whereas there is a rather large difference in trust between hbo and wo students relative to mbo students. Therefore a part of the non-scientifically educated students might have been able to process the news-item through the central route.
Similarly, perhaps no moderating effect of scepticism on the influence of expert frames on trust was found due to students’ lack in scientific literacy. When sceptical people are unable to process information through the central route, they are as vulnerable to expert frames as all other students who process information through the peripheral route. Another possible explanation lies in the students’ verification process. When sceptical people determine whether or not to trust news-items, they verify the information in these items internally or externally. The internal verification process happens on the basis of existing knowledge and past experiences. Expert frames, though they are considered to be simple cues, might give legitimacy to statements in news-items. They transform an unsupported statement into a statement that was proven and validated by scientific research. Expert frames might therefore actually pose as a means of validation for sceptical people. Hence, expert frames may have notable effects on students who process information in news-items through the central route – even though they were deemed irrelevant in the case that students process information through this route.
Both education and scepticism have direct effects on students’ trust towards information. Additionally, while the moderating effect of scepticism on the influence that expert frames have on students’ trust was not significant, non-sceptical people scored almost a point higher on a seven-point scale of trust towards information due to the presence of an expert frame. It seems possible that perhaps level of education and scepticism do moderate the relationship between expert frames and trust, yet this experiment was unable to prove such effects due to the inability to control for the route that people used to process the news-item. Therefore, future research is recommended to determine whether moderating effects of education as well as
scepticism on the influence that expert frames have on trust towards information exist. Since it cannot be excluded that non-scientifically educated students are scientifically illiterate perse, future research should use scales that specifically measure scientific literacy. Furthermore, a distinction should be made between levels of education (i.e., mbo, hbo, wo) or between higher education (i.e., hbo and wo students) and lower education (i.e., mbo students), rather than between scientific and non-scientific education. Additionally, in order to account for the effects that expert frames might have when sceptical people process information through the central route, statements should be incorporated in the stimulus for the control condition that claim that the findings are based on some form of non-scientific research, like a research that was conducted by the newspaper itself. This way both stimuli contain similar cues, one of which is an expert frame and one of which is not.
The main limitation in this study is the problems with measuring which route people used to process the information presented to them in the stimuli. The fact that students are, theoretically speaking, able to process science communication through the central route due to their education or scepticism, does not mean that they do. Future research should account for this problem in two ways. Firstly, since level of education may not have been an accurate indicator of scientific literacy, more reliable results might be found if variables are included that specifically determine one’s scientific literacy. Secondly, variables that measure people’s level of engagement with the information should also be included in the experiment. Future research can, for instance, take into account the time that participants spend reading the stimuli – engaged people will likely spend more time reading them (Greene, 2009). Additionally, if the stimulus is a snippet from the digital edition of a newspaper, a clickable link can be added to the stimulus that leads to additional information or to the academic paper that the newspaper article was based on. Engaged people are more likely to seek external verification through additional information, which makes them more likely to click the link. Consequently, it should be possible to determine whether people are engaged and capable enough to process information in stimuli through the central route or the peripheral route. This way, if people are able but unwilling or unable but willing to process information through the central route it will be clear that they did not.
A second limitation concerns the generalizability of these findings to other sources of information, contexts, and to a general population. The problem with
expert frames and ungrounded scientific claims is much bigger than just news-items. Anyone can write anything on the internet and people can integrate expert frames in any type or form of information to convey trust. Based on this paper it cannot be determined whether expert frames have similar effects on people’s trust in those situations. Nor can we be certain that a general public responds in a similar way that students do. Additionally, people might usually assess science communication differently. Participants knew they were being studied when they read the news-item. Therefore, along the lines of the problem of demand characteristics, people may have been more critical than they would normally have been because they knew they were being studied.
This thesis contributes to the existing knowledge on science communication, expert framing, and trust in science. Unlike previous papers, this thesis focused on trust in science on a micro level rather than trust in science in general – it zooms in on trust in specific pieces of information containing scientific claims. However, future research is needed in order to determine whether expert frames have similar effects on trust in different contexts on a general population. Future research in this field is advised to investigate the effects of expert frames on trust in different sources of information, like blogs, webshops, social media posts, etcetera, on a broader population. Additionally, the stimuli that were used in this paper were quite brief and the expert frames were subtle. Future research can investigate whether the same effects are found in more elaborate newspaper articles and when the expert frames are more explicitly present in news-items.
Over the past decades, science has become accessible to the general public. Most people still experience science through the media. The media, however, is an institution that pursues different goals than science. Therefore, pieces of science communication are not always immaculate translations of scientific papers. Yet, because people generally hold science in a high regard, the presence of expert frames in news-items drives people to trust them. The influence of science communication on people’s trust in science in general has been researched numerous times, even in conjunction with education. This paper expanded on this research by showing that expert frames in pieces of (science) communication affect students’ trust towards specific pieces of information. The findings also suggest that, while no interaction effects were found, both education and scepticism directly influence students’ trust towards information. So even though it is a great development that science is
becoming more accessible to the general public, it is important for people to stay critical of information that contains expert claims. After all, no one wants to believe in the expertise of ignorants.
6. Literature
Abdul-Rahman, A., & Hailes, S. (2000). Supporting trust in virtual communities. IEEE Conference Publications, 32, 52-61.
Allum, N., Sturgis, P., Tabourazi, D., & Brunton-Smith, I. (2008). Science knowledge and attitudes across cultures: A meta-analysis. Public Understanding of Science, 17, 35-54.
Bakker, P., & Scholten, O. (2011). Communicatiekaart van Nederland. Amsterdam: Kluwer.
Barber, B. (1983). The logic and limits of trust. New Brunswick: Rutgers University Press.
Becker, J., Schnabel, P., & Bouman, M. (2000). Het beeld van de wetenschap. Den Haag: NWO.
Bodmer, W. (1985). The Public Understanding of Science. London: Royal Society.
Boush, D., Chung-Hyun, K., Kahle, L.R., & Batra, R. (1993). Cynicism and conformity as correlates of trust in product information sources. UNSW Journal of Current Issues & Research in Advertising, 15(2), 71-79.
Bromme, R., & Goldman, S.R. (2014). The public's bounded understanding of
science. Educational Psychologist, 49(2), 59-69. doi:10.1080/00461520.2014.921572.
Cacioppo, J.T., & Petty, R.E. (1984). The elaboration likelihood model of persuasion. Advances in Consumer Research, 11(1), 673-675.
Carr, D.J., Barnidge, M., Gu Lee, B., & Jean Tsang, S. (2014). Cynics and skeptics: Evaluating the credibility of mainstream and citizen journalism. Journalism & Mass Communication Quarterly, 91(3), 452-470. doi: 10.1177/1077699014538828.
Collins, H.M., & Evans, R. (2007). Rethinking expertise. Chicago, IL: University of Chicago Press.
Covey, J., & Kazak, A.E. (2014). The Role of Dispositional Factors in Moderating Message Framing Effects. Health Psychology, 33(1), 52-65.
Cribb, J., & Hartomo, T.S. (2002). A Guide to Effective Science Communication. Melbourne: CSIRO Publishing.
Dam, F. van, Bakker, L. de, & Dijkstra, A. M. (2014). Wetenschapscommunicatie, een kennisbasis. Den Haag: Boom Lemma.
Derks, A. (2000). Individualisme zonder verhaal. Een onderzoek naar de verspreiding en de betekenis van individualistische vertogen in Vlaanderen. Brussel: VUB-Press.
Deutsch, M. (1973). The resolution of conflict. New Haven and London: Yale University Press.
Dijkgraaf, R. (2011). Machiavellilezing. Wetenschap, Publiek, Politiek – Over feiten
en meningen. Retrieved on February 15th 2016 from http://www.robbertdijkgraaf.com
Entman, R.M. (1993). Framing: Toward clarification of a fractured paradigm. Journal of Communication, 43(4), 51-58.
Gard, M., & Wright, J. (2005). The obesity epidemic: Science, morality and ideology. Abingdon: Routledge.
Gravetter, F. J., & Forzano, L. B. (2012). Research methods for the behavioural sciences. Wadsworth: Cengage Learning.
Greene, J. (2009). Aligning video with content to increase engagement: by embedding videos next to related articles, publishers are able to increase page views and time spent for both mediums. Audience Development, 24(5), 14-22.
He, J., Van De Vijver, F.J.R., Dominguez Espinosa, A., Abubakar, A., Dimitrova, R., Adams, B.G., Aydinli, A., Atitsogbe, K., Alonso-Arbiol, I., Bobowik, M., Fischer, R., Jordanov, V., Mastrotheodoros, S., Neto, F., Ponizovsky, Y.J., Reb, J., Sim, S., Sovet, L., Stefenel, D., Suryani, A.O., Tair, E., & Villieux, A. (2015). Socially desirable responding. Cross-Cultural Research, 49(3), 227-249.
Hijmans, E., Pleijter, A., & Wester, F. (2003). Science Communication, 25(2), 153-176. doi: 10.1177/1075547003259559.
Holland, K.E., Warwick Blood, R., Thomas, S.I., Lewis, S., Komesaroff, P.A., & Castle, D.J. (2011). Our girth is plain to see: An analysis of newspaper coverage of Australia's future ‘fat bomb’. Health, Risk & Society, 13(1), 31-46. doi:10.1080/13698575.2010.540648.
Houtman, D., Aupers, S., & Achterberg, P. (2012). Alledaags postmodernisme: De gezagscrisis van de hedendaagse wetenschap. Beleidswetenschap, 33(4), 143-156.
Howgego, J. (2015). Confirmation bias: We only believe what we already think. New Scientist, 227(3034), 29-31.
Irwin, A. (2015). Science, public engagement with. International Encyclopedia of the Social & Behavioral Sciences, 255-260.
Janssens, J. (2002). Hoe moet een wetenschappelijke opleiding binnen de bachelor-master structuur worden ingericht? Pedagogiek, 22(2), 105-108.
Keere, K., de. (2010). Wantrouwen in wetenschap: een kwestie van reflexiviteit of maatschappelijk onbehagen? Sociologie, 6,(1), 45-62.
Kelton, K., Fleischmann, K.R., & Wallace, W.A. (2008). Trust in digital information. Journal of the American Society of Information Science and Technology, 59(3), 363-374.
Kendall, S. (2004). Framing authority: Gender, face, and mitigation at a radio network. Discourse & Society, 15(1), 55-79. doi:10.1177/0957926504038946.
Kini, A., & Choobineh, J. (1998). Trust in electronic commerce: definition and theoretical considerations. IEEE Conference Publications, 4, 51-61.
Kleemans, M., van Cauwenberge, A., D'Haenens, L., & Vettehen, P. (2008). Explaining sensationalism in television news. A comparison of Dutch, Flemish, Walloon, and French news stories. Tijdschrift Voor Communicatiewetenschap, 36(4), 301-319.
Kruger, J., & Dunning, D. (1999). Unskilled and unaware of it: how difficulties in recognizing one's own incompetence lead to inflated self-assessments. Journal of personality and social psychology, 77(6), 1121-1132.
Leurs, A., & Kroodsma, D. (2014). Science communication in the post-expert digital age. Eos, 95(24), 203-204.
Mascaro, O., & Sperber, D. (2009). The moral, epistemic, and mind reading
components of children’s vigilance towards deception. Cognition, 112(3), 367-380. doi:10.1016/j.cognition.2009.05.012.
Mayer, R.C., Davis, J.H., & Schoorman, F.D. (1995). An integrative model of organizational trust. Academy of Management Review, 20(3), 709–734.
McChesney, R. (2004). The problem of the media: US communication politics in the 21st century. New York: Monthly Review Press.
McComas, K.A, & Simone, L.M. (2003). Media coverage of conflicts of interest in science. Science Communication, 24, 395-419.
Meyer, P. (1988). Defining and measuring credibility of newspapers: Developing an index. Journalism Quarterly, 65, 567-574.
Miller, J.D. (2004). Public understanding of, and attitudes toward, scientific research: What we know and what we need to know. Public Understanding of Science, 13, 273-94.
Mills, C.M., & Landrum, A.R. (2012). Judging judges: how do children weigh the importance of capability and objectivity for being a good decision maker? British Journal of Developmental Psychology, 30(3), 393-414.
doi:10.1111/j.2044835X.2011.02047.
Moore, K. (2008). Disrupting science: Social movements, american scientists, and the politics of the military. Princeton, NJ: Princeton University Press.
Moynihan, R. (2003). Making medical journalism healthier. Lancet, 361, 2097-2098.
Mui, L., Mohtashemi, M., & Halberstadt, A. (2002). A computational model of trust and reputation. IEEE Conference Publications, 35, 62-69.
Obermiller, C., & Spangenberg, E.R. (1998). Development of a scale to measure consumer scepticism toward advertising. Journal of Consumer Psychology, 7(2), 159-186.
Pinkleton, B.E., Weintraub Austin, E., Zhou, Y., Fitts Willoughby, J., & Reiser, M. (2012). Perceptions of news media, external efficacy, and public affairs aparthy in political decision making and disaffection. Journalism & Mass Communication Quarterly, 89(1), 23-39.
Pion, G., & Lipsey, M. (1981). Public attitude towards science and technology. What have the survey told us? Public Opinion Quarterly, 45(1), 303-316.
Powys Whyte, K., & Crease, R.P. (2010). Trust, expertise, and the philosophy of science. Synthese, 177, 411-425. doi: 10.1007/s11229-010-9786-3.
Ransohoff, D.F., & Ransohoff, R.M. (2001). Sensationalism in the media: When scientists and science writers may be complicit collaborators. Effective Clinical Practice, 4, 185-188.
Ream, R.K., Lewis, J.L., Echeverria, B., & Page, R.N. (2014). Trust matters: Distinction and diversity in undergraduate science education. Teachers College Record, 116(5). 128-142.
Scheman, N. (2001). Epistemology resuscitated: Objectivity as trustworthiness. Albany, NY: SUNY Press.
Shafto, P., Eaves, B., Navarro, D.J., & Perfors, A. (2012). Epistemic trust: modelling children’s reasoning about others’ knowledge and intent. Developmental Science, 15(3), 436-447. doi:10.1111/j.1467-7687.2012.01135x.
Shapin, S. (2004). Trust me, I'm a scientist. London: The British Council.
Singer, E. (1990). A question of accuracy: How science writers and scientists report research on hazards. Journal of Communication, 1990(40), 102-116.
Smetana, L.K., Wenner, J., Settlage, J., & Mccoach, D.B. (2016). Clarifying and capturing “trust” in relation to science education: dimensions of trustworthiness within schools and associations with equitable student achievement. Science Education, 100(1), 78-95.
Trumbo, C.W., & McComas, K.A. (2003). The function of credibility in information processing for risk perception. Risk Analysis, 23(2), 243-253.
Wang, T. (2013). Presentation and impact of market-driven journalism on
sensationalism in global TV news. International Communication Gazette, 75(1), 138-138.
Wilson, K.M. (2000). Drought, debate, and uncertainty: Measuring reporters’ knowledge and ignorance about climate change. Public Understanding of Science, 9(1), 1-13.
Wobma, E. (2011). Mannen en vrouwen in Nederland. Retrieved on April 25th 2016 from www.cbs.nl.
Zaller, J. (2010). A new standard of news quality: Burglar alarms for the monitorial citizen. Political Communication, 20(2), 109-130. doi: 10.1080/10584600390211136.
7. Appendix 7.1 Questionnaire
Q1 Welkom!
Ik wil je graag uitnodigen om deel te nemen aan een onderzoek dat wordt uitgevoerd onder verantwoordelijkheid van de Graduate School of Communication, onderdeel van de Universiteit van Amsterdam. Het doel van dit onderzoek is om te achterhalen welke aspecten van krantenartikelen door mensen worden onthouden. Het invullen van de vragenlijst duurt ongeveer 5 minuten.
Omdat dit onderzoek wordt uitgevoerd onder de verantwoordelijkheid van ASCoR, Universiteit van Amsterdam, heb je de garantie dat:
- Je anonimiteit is gewaarborgd en dat je antwoorden of gegevens onder geen enkele voorwaarde aan derden worden verstrekt, tenzij je hiervoor van te voren uitdrukkelijke toestemming hebt verleend.
- Je zonder opgaaf van redenen kunt weigeren mee te doen aan het onderzoek of je deelname voortijdig kunt afbreken. Ook kun je achteraf (binnen 24 uur na deelname) je toestemming intrekken voor het gebruik van je antwoorden of gegevens voor het onderzoek.
- Deelname aan het onderzoek geen noemenswaardige risico’s of ongemakken met zich meebrengt, geen moedwillige misleiding plaatsvindt, en je niet met expliciet aanstootgevend materiaal zult worden geconfronteerd.
- Je uiterlijk vijf maanden na afloop van het onderzoek de beschikking kunt krijgen over een onderzoeksrapportage waarin de algemene resultaten van het onderzoek worden toegelicht.
