23/10/2014
Supervisor:
Dr. Ester de Waal
Master's Thesis for Master's programme Communication Science
Roeland Schoppers - 6145000
X
Creative media use for health communications:
intended and unintended persuasive effects and
the role of attitude toward the ad
2
Abstract:
This study seeks to address the question whether or not creative media use can be more effective for health communications than traditional media use. If there are possible
unintended side-effects on medium associated behavior due to using creative media. And if attitude towards the ad plays a role in facilitating the possible increased effects of creative media use. By doing an online experiment with two experimental conditions (traditional medium vs creative medium) and one control group (N = 239) we tried to find support for the effects of creative media use on health communication effects. We found no significant results between creative media use and traditional media use for recognition, attention, behavioral self-efficacy, behavioral intent, medium associated self-efficacy or medium associated intent. Results relating to the control group and the failed manipulation checks indicate a likelihood that an ineffective message was used and/or the traditional medium condition not being perceived as such. This study did find an effect of the communications on medium associated behavior (stair use) as well support for the role of attitude toward the ad in facilitating message effectiveness, being positively associated with self-efficacy and intentions. These results provide implications for possible ethical concerns regarding the use of creative media due to unintended medium associated effects as well as implications regarding the importance of considering the attitude toward the ad to facilitate effective health communications.
Introduction
The current advertising market is highly saturated, which has lead to advertising clutter, the excessive exposure to advertisements (ads) that consumers have no interest in (Rosengren, 2008). A result of this clutter is a severe decrease in advertising effects, persuasiveness and even avoidance of ads altogether (Rosengren, 2008). One strategy that has been found to overcome some of these negative effects is the use of creative media for advertising instead of traditional media (e.g., Dahlen, 2005). Traditional media are those media that are
3
creatively choosing a novel medium ,congruent to the message, to implicitly communicate a message (Dahlen, 2005; Dahlen & Edenius, 2007).
All these findings come from a product advertising context however and have not been researched in the context of health campaigns. Campaign messages are often
communicated through the same media as product ads and may therefore suffer from clutter the same way. But if creative media use can increase purchase intentions, could it also influence intentions for health behavior? Although ―there are many similarities, campaigns promoting products differ from health campaigns‖ in several ways (Flay & Burton, 1990, p. 131). Due to these fundamental differences ―the objectives and processes that are
appropriate for commercial advertising are usually inappropriate for health promotion‖ (Elliot, 1987). Nonetheless, some overlap in processeses and applicable theories does exist, such as the use of affect or celebrity endorsements (e.g., Petty, Wheeler & Tormala, 2003; Brown & Basil, 1995). Theories that also share common ground with creative media use theory (Dahlen, 2005). Creative media use has been shown to increase ad attitude (Dahlen, 2005), an aspect that has received hardly any attention in the context of health communications but according to Slater (2006) should be investigated as a potential facilitator of persuasive effects. Furthermore, results by Dahlen (2005) indicate that a creative medium choice does not only transfer meaning from the medium to the message, but also the other way around, this could open up potential ethical concerns for health campaigns using creative media. If these campaign messages lead to a negative meaning transfer to the medium, then associated behavior could be negatively impacted.
This study investigates the effects that creative media use may have on health communications. Focusing on the campaign goals of awareness and changing intentions (Rogers & Storey, 1987). In doing so it will be one of the first studies to investigate the role of ad attitude on health communications. As well as add to creative media theory by further investigating the unintentional effects that creative media use might have on medium related behavior itself.
4
RQ:
Can creative media use increase the effectiveness of health communications by increasing awareness and/or behavioral intentions?
Are there any unintended effects on the medium used for the creative medium choice and it's associated behavior?
Does attitude toward the ad (ad attitude) facilitate the potential increased effectiveness of creative media use for health campaigns?
Theoretical Framework Brief overview of creative media
Non-traditional media are those media that are not traditionally used for advertising. This can be literally anything, eggs, elevators, stairs, fire-extinguishers and so on. Creative media choice is the use of such a non-traditional medium but chosen creatively so that the desirable associations of the message are congruent with the associations evoked by the chosen medium (Dahlen, 2005). Through congruency ―the medium and the advertised brand converge and become more similar in consumers‘ minds‖ and ―by way of associative
learning, associations are [then] transferred [from the source] to the brand‖ (Dahlen, 2005, p. 90). The congruent medium associations can function as a cognitive prime, by putting the focus of our attention on these characteristics they become more salient and will play a more prominent role in our evaluation and interpretation of the message (Yi, 1990a; 1990b; 1993; Dahlen, 2005). Because we are not consciously aware of these processes taking place (Herr, 1989; Schmitt, 1994) creative media use communicates its message in an implicit way (Dahlen & Edenius, 2007), this indirect approach stimulates the audience to come to the intended meaning through their own interpretive process rather than force it onto them (Meyers-Levy and Sternthal, 1993).
Creative media use is considered a more novel, creative and unexpected form of advertising than using traditional media (Rosengren & Dahlen, 2012). It has been found to be
5
identified less as an advertising message, positively increase message attitude (Dahlen & Edeneius, 2007), elicit stronger brand associations (Dahlen, 2005), positively increase ad credibility, ad attitude (Dahlen, 2005; Dahlen & Edenius, 2007), advertising value, brand attitude (Dahlen, Granlund & Grenros, 2009; Rosengren & Dahlen, 2012; Rosengren, Modig & Dahlen, 2015), perceived ability, perceived product quality (Rosengren & Dahlen, 2012), consumer care (Rosengren, Modig & Dahlen, 2015) and purchase intent (Dahlen & Edenius, 2007; Dahlen, Granlund & Grenros, 2009; Rosengren & Dahlen, 2012; Rosengren, Modig & Dahlen, 2015).
Creative media use: avoiding clutter to increase awareness
Advertisement clutter results in reduced attention, memorability, recognition, attitude toward advertising in a medium, increased ad skepticism, mistrust and negative attitudes towards the message (e.g., Ray & Webb, 1986; MacKenzie & Lutz, 1989; Ha, 1996; Speck & Elliot, 1998; Obermiller, Spangenberg & MacLachlan, 2005; Rotfeld, 2006;Rosengren, 2008; Hammer, Riebe & Kennedy, 2009). Consumers have changed the way they deal with advertising and often try to avoid it all together (Rosengren & Dahlen, 2012). To counter these effects advertisers should find new ways to stand out and ―get consumers to approach rather than avoid ads‖ (Rosengren, Modig & Dahlen, 2015, p. 2).
One way to accomplish this is by using non-traditional media (Dahlen & Edenius, 2007) such as a creative media choice. According to these authors, ad avoidance and ad skepticism comes from the identification of content as advertising (Goodstein, 1993; Friedstad & Wright, 1995; Dahlen, 2005). Through repeated exposure a person builds a frame of reference, the advertisement schema, by which they learn to identify content within a certain medium as advertising, a subconscious process (Friedstrad & Wright, 1995; Dahlen & Edenius, 2007). Because using creative media breaks away from standard advertising practices these ads do not fit within this advertisement schema and the viewer is challenged to form new
perceptions (Dahlen, 2005; Dahlen & Edenius, 2007). Not only avoiding the avoidance, reduced attention and skepticism that occurs when the schema is activated but also leading
6
to an increased interest (Goodstein, 1993; Stafford & Stafford, 2002; Dahlen, 2005). And due to a higher advertising creativity viewers are willing to pay more attention to the ad as well (Rosengren & Dahlen, 2012). Although these studies were all done within a product advertising context, health communications that use traditional media to promote their messages would suffer from the same negative effects. Because is not the ad content but the surrounding context that triggers the ad schema (Dahlen & Edenius, 2007). Which leads us to the following hypotheses:
H1: Creative media use leads to a greater (a) recognition of and (b) attention towards the ad
than traditional media use
Creative media providing a more favorable attitute toward the ad
The actual explicit message that is communicated (such as, smoking can cause lung cancer) is not the only factor that determines the effectiveness of a persuasive communication. The context of the communicated message has been shown to play a prominent role in its
effectiveness as well (Dahlen, 2005). Background music, editorial content or using celebrities to communicate the message are a few examples of these contextual cues. By influencing our perception of the communication these factors can increase or decrease the success of a particular message (e.g., Yi, 1990). These results suggests that at least to some extend the way we communicate our messages is just as important as what we want to communicate. Previous research into creative media shows that it provides the viewer with these sort of positive contextual cues while avoiding some of the negative associations of traditional media (Dahlen, 2005; Dahlen, Friberg & Nilsson, 2009; Rosengren, Dahlen & Modig, 2015) These negative cues that are associated with traditional media but not with creative media have already been discussed. The identification of traditional media as advertising along with the advertisement literacy and skepticism of the contemporary audience leads not only to avoidance but also to a greater resistance and more negative attitude towards the ad, which in turn leads to a decrease in persuasiveness (e.g.,Speck & Elliott, 1997; Obermiller,
7
Spangenberg & MacLachlan, 2005). The higher the perceived degree of persuasive intent the more resistance and skepticism will arise in audiences. Simply by working outside of this usual advertisement frame, creative media is able to avoid or lessen these negative
associations due to a lower perceived persuasive intent (Dahlen & Edenius, 2007). Since health communications often share the same space as product advertisements it is likely to assume that using traditional media would lead to the same negative effects.
By creatively choosing a non-traditional medium that is congruent with the message the audience is confronted with an unexpected form of persuasive communication. Dahlen (2005) found that due to the surprise value, creative media becomes more interesting to the audience and generates more positive evaluations. But a creative media choice is not only surprising, it is also perceived as being a more creative form of persuasive
communication than using traditional media (Rosengren & Dahlen, 2012). Creativity in advertising has been shown to increase ad attitudes (Smith, Chen & Yang, 2008) lead to greater perceptions of advertising value (Rosengren & Dahlen, 2012) and even directly increase the persuasiveness of the message (Yang & Smith, 2009) enhancing purchase intention in product advertising (Kover, Goldenberg, and James, 1995). The way that creativity may accomplish these effects has been shown to at least be partially through affective processes (Yang & Smith, 2009). The way we experience the advertisement elicits a certain affective response, this feeling is then transferred onto the target of the
advertisement (ie. the message, product, brand) (Smith, Chen and Yang, 2008). A creative advertisement might be seen as more entertaining and thus generate a positive affect, which is then transferred to the message, enhancing it‘s persuasiveness (Yang & Smith, 2009). This indicates that the way we experience an advertisement, and thus our attitude toward it, has an influence on the way we perceive the message. Creativity is one such a factor that influences our experience of an advertisement.
In product advertising the overall way we experience an advertisement (favorably or unfavorably) is often discussed as our attitude toward the ad, an extensively researched concept which has shown to be an important mediator for advertisement effectiveness
8
(MacKenzie & Lutz, 1989; Brown & Stayman, 1992). It‘s effect on advertisement
effectiveness is explained in a similar way as that of creativity, the favorable or unfavorable effects of our attitude to the ad is transferred to the message (product/brand) that it is paired with (Yang & Smith, 2009). Research on creative media use has shown that it is capable of enhancing the attitude toward the ad in a greater degree than that of traditional media use (Dahlen, 2005; Dahlen & Edenius, 2007).
Because creative media use can avoid the negative associations that have been developed for traditional media advertising, it‘s increased advertising creativity and having been shown to facilitate a more positive attitude toward the advertisement we assume that the overall experience of this form of persuasive communication will be perceived as more favorable and pleasant than with that of traditional media use. We therefore come to the following hypothesis:
H2: Creative media use leads to a more positive ad attitude than that of traditional media use
Increased behavioral intentions and self-efficacy and the role of ad attitude
As was just discussed in the previous paragraph, avoiding the negative associations of traditional media advertising, increased creativity and a more positive attitude toward the ad have all been linked to an increased message persuasiveness, enhancing attitudes and/or intention, by providing a more favorable advertisement experience. These concepts were all studied within a product advertising context however. As it stands now, the main focus of health communication research and campaign formation seems to be on constructing the most effective content for the message (Slater, 2006) through such behavioral models as the theory of planned behavior (Ajzen, 1991), integrated model of health behavior (Fishbein & Cappella, 2006), protection motivation theory (Rogers & Prentice-Dunn, 1997) or message framing (Rothman, Bartels and Salovey, 2006). These have without a doubt proven to be incredibly useful in predicting and influencing health behavior through message formation (Godin & Kok, 1996; Milne, Sheeran & Orbell, 2000).
9
For instance, the integrated model of health behavior (Fishbein & Cappella, 2006) provides insights into how health behavior can be influenced through persuasive
communications. It does this by giving an understanding of the potential determinants of behavior. In short, there are a range of background factors that may influence which determinants are most predictive of health behavior. For example, changing a specific health-related behavior may require a different approach for elder people than for teenagers. The three main determinants of intention to perform a certain behavior are attitude, norms and self-efficacy. Attitude consists out of the beliefs associated with a particular behavior and our evaluation of the outcome (favorable or unfavorable) of those behavorial beliefs, Norms consists out of the beliefs of significant others within our social network, such as parents or friends, and our motivation to comply to the norms of those significant others, Self-Efficacy consists out of the beliefs on how much control we have over performing or changing our behavior and the perceived power of those beliefs in facilitating or inhibiting the desired behavior. According to the model these three factors are most predictive of intention, while intention is most predictive of our behavior. However if there is a large intention but this does not lead to a behavioral change than there might be enviromental factors or a lack of skills and abilities impeding performing the behavior, despite the intention to do those. Which of these factors can facilitate the most change in intention and/or behavior depends on the background factors as well as the specific behavior that is targeted.
This model and those resembling it such as the theory of planned behavior provide insights into which beliefs would prove to be the most effective target for health interventions. They tell us what to target in the message, the best content for the message, what they do not tell us though, is how to deliver this message most effectively (Slater, 2006). Creative media choice is not related to message formation but rather, message delivery. By providing a better message experience (and therefore an increased ad attitude) the persuasiveness of the message is increased, whatever that message may be. In this regard creative media use would be a complementary strategy to using health behavioral models. These models can be used to form an effective message, while creative media could, theoretically, be used to
10
enhanced the persuasiveness of this message. In a similar way as it enhances purchase intentions for product advertising by enhancing the attitude toward the ad. Within the context of health communications the attitude toward the advertisement has received hardly any attention however (Slater, 2006). Slater argues that it is something that should be
investigated in the context of health communications as it could potentially influence their effectiveness. To support his argument he points to a study by Kelly, Slater and Karan (2002) on the effects of cigarette and beer ads. This study found attitude towards the ad influenced not only brand attitude but also the overall desirability of cigarettes and beer. If our attitude toward an ad can negatively change our attitude towards health behavior then it stands to reason that the opposite could be true as well. As previously mentioned, our attitude toward the ad influences the persuasiveness of a message primarily through positive affect. By making us feel good we also feel good about the message. According to Monahan (1995), messages containing positive affect paired with a rational message (emotional benefit appeal) can enhance the persuasiveness of a campaign and create long-term change. Enhancing the compliance to the message, more easily stirring people into action. Positive affect can enhance the persuasiveness of the arguments made or make positive thoughts more salient when processing the message arguments (Petty, Gleicher and Baker, 1991). These results make it likely that creative media‘s more favorable ad attitude would also enhance the message persuasiveness of a health intervention. In the context of health communication, using the models for behavioral change, this would specifically mean that, through ad attitude, creative media use can lead to a greater attitude, norms or self-efficacy change than when the same message is communication through traditional media use. In this study, we only test one message targeting self-efficacy, which leads to the following hypothesis:
H3: Creative media use leads to a greater increase in self-efficacy regarding the behavior
11
However, as we explained, the integrated model for behavorial change stipulates that
changing one of these norms can lead to an increase in the intention to perform the behavior. So an increase in self-efficacy should then also lead to an increase in the intention to perform the behavior. But there is another reason why creative media could lead to an increase in intention. Dahlen (2005) describes how a creative media choice communicates its meaning in an implicit way. By pairing an unexpected congruent medium and the advertisement you leave the viewer to consider the associations and meaning of this combination. Through congruency between the medium and the advertisement, creative media implies, rather than explicitly states, the intended meaning. The interpretation that follows will be considered more as their own opinion rather than as one imposed upon them by the sender (Hoch, 2002). In this process there is less room for resistance and counterarguing to the message (van Gisbergen & Ketelaar, 2006). These factors can lead to a more persuasive
communication (Yi, 1990), although there is a risk that without an explicit conclusion some viewers might not come to the intended interpretation (Sawyer & Howard, 1991; Ahearne, Gruen, and Saxtion, 2000). This form of communicating should further increase the
compliance to the intention to perform the advocated behavior. Furthermore, positive affect and attitude toward the ad have been shown to have a direct effect on intention as well (Morris, Woo, Geason & Kim, 2002; Kwan & Bryan, 2010; Conner et al, 2011). We therefore hypothesize that:
H4: Creative media use will lead to a greater intention to perform the targeted behavior than
when using traditional media, partially mediated by ad attitude and self-efficacy
Unintended effects on medium related intention through association
As we already discussed, creative media use gives meaning to the message through the congruency between the medium and the message by way of associative learning (Dahlen, 2005). This transfer of meaning has thus far been discussed from one direction, the medium giving meaning to the message. The way that an elevator (uplifting) can give meaning to an
12
energy drink brand (the drink is uplifting) (Dahlen, 2005). However Dahlen, Friberg & Nilsson (2009) found that this transfer of meaning can work both ways. So that the associations of the brand/message are transferred onto the medium as well. From a product advertising perspective they interpret this finding as being interesting because the medium can serve as a brand cue, reminding consumers of the brand even when it is no longer paired with the medium. But the relevance of this finding might go further than this.
If the medium can transfer meaning to the message and thereby change our self-efficacy and intentions regarding the message then the opposite might be true as well. So if the energy drink is associated with providing us with more energy, the elevator might then be considered to also provide us with more energy. As a result people might potentially become more inclined to take the elevator over the stairs. From a health communication perspective using creative media might have to come with considerations on how it might impact the audience‘s perception of the medium that is used. If negative affect is used to dissuade a certain unhealthy behavior such as smoking then this negativity could potentially be
transferred to the medium. Using the stairs as a medium to do this, people might associate the stairs with such negative feelings or unhealthy perceptions, dissuading them to use the stairs over the escalators or elevators and potentially negatively impacting their health due to a decrease in physical activity. At the same time, a positive message might indirectly benefit such intentions and self-efficacy. If this is true then we would expect the self-efficacy and intentions toward the medium used in our study to change as well. Since we use a self-efficacy message, this might also lead to a change in self-self-efficacy toward medium related behavior. Which leads us to the following hypotheses:
H5: Creative media use increases the self-efficacy toward medium associated behavior to a greater degree than when using traditional media and this effect is partially mediated by ad attitude
13
H6: Creative media use increases the intention toward performing medium related behavior to a greater degree than when using traditional media and this effect is partially mediated by self-efficacy and ad attitude
Figure 1. Conceptual Model
Methods Design
This study used a three-group experimental design using an online survey. Group one, the traditional media group, were shown an image with the creative media version of the stimulus. Group two, the creative media group, were shown an image with the traditional media version of the stimulus. In addition to these, in contrast to previous studies on creative media choice, a control group was included. This choice was made partially due to no pre-tests, so that in the case of insignificant differences between the traditional- and creative medium group the cause could be identified as being either due to the complete absence of
14
any persuasive effect (no significant difference with the control group) or due to no increased effects of creative media over traditional media (significant difference with the control group). By still showing the control group a photograph this also allowed for the exclusion of the experimental setting as the cause for any significant effects.
Sample
A total of 239 participants were included in the final sample. They were recruited by using several different methods. First of all, the researcher requested friends, family and fellow students to fill out the survey through facebook or other digital communication channels, also asking them to share the survey with their friends. In addition to this, the researcher
requested people fill out the survey in public forums (e.g., reddit) and facebook groups. The downside of these methods is that it might not lead to an unbiased population sample. A benefit of this method was that participants had no external reason to participate other than to help, which should mean they gave high quality responses.Apart from these methods, the researcher used crowdsourcing to obtain additional respondents. These respondents were recruited using two different services: Microworker and Crowdflower. Microworker distributed the survey to an international pool of workers while Crowdflower distributed the survey to workers from native English speaking countries and the Netherlands. These respondents received $0.30 compensation for their work. Recent studies have shown that samples obtained through these services (such as MTurk, a similar service) provided equally representative and high quality samples as more conventional methods of respondent recruitment (e.g., Paolacci, Chandler & Ipeirotis, 2010; Buhrmester, Kwang & Gosling, 2011; Berinsky, Huber & Lenz, 2012).
398 surveys were started of which 376 were finished, unfinished survey data was removed from the data set. Exploratory data analysis lead to the removal of 37 respondents. 18 of these cases were removed due to taking an extreme amount of time to complete the survey, these fell within the 27 values SPSS identified as being extremes using a stem-and-leaf analysis. 15 minutes to 14 hours were removed, leaving 10 minutes and less. In light of
15
paying some respondents for their participation, to safe-guard reliability, eight respondents were removed because they took less than 60 seconds to complete the survey. Such times were perceived as unrealistic by the author. This put the minimum completion time at 90 seconds. Seven respondents filled out they paid attention to the material an impossible amount of time, respondents with 433 seconds and above were removed, putting the maximum value at 40 seconds. All removed values fell within the extremes identified by SPSS using a stem-and-leaf analysis. Finally, 3 participants indicated they held a very low proficiency in the language of the survey and were removed as well. This lead to a final sample size of 239 (N = 239). Participants were randomly assigned to one of the three experimental conditions: the traditional media group (n = 83), the creative media group (n = 76) and the control group (n = 80). These sizes comply with the suggested sample size per cell of ~50 by VanVoorhis & Morgan (2007). The mean age of the final sample was 31.05 (SD = 8.27). Of all participants 128 (54%) were male and 111 were female. Because of the methods employed the nationality of the respondents was diverse, with 74 (31%) being from the Netherlands, 74 (31%) from North America, 38 (16%) from the United Kingdom, 16 (7%) from Western Europe and the remaining 37 (16.%) from 19 other countries all over the world. Education wise, 38% held a bachelor degree, 24% held a vocational (MBO) degree, 24% a High School degree and 10% a masters degree. 69% of respondents filled out the survey in English, the remaining in Dutch. The mean of the perceived proficiency in the survey
language was 6.63 (SD = 0.67).
Stimulus
In this study all participants were exposed to the same basic photograph of both stairs, escalators and an elevator. The traditional media group were given this photo with the
addition of the campaign name + message placed on a billboard on the right side of the stairs (Appendix, fig. 2). The creative media group had the same campaign name + messaged placed on the stairs itself instead (Appendix, fig. 3) while the control group featured this photo without the campaign name or message (Appendix, fig. 1), to control for any influences
16
that the photograph itself might have had. The stimulus was created by the author using photoshop and a stock photograph of the stairs. The choice for this photograph was based on it‘s neutrality, showing empty stairs, a frontal photo giving no visual preference to either stairs or the escalator and providing space for a billboard next to the stairs. This material most reflects that of previous research on creative media within a product advertising context, in which a brand name + ad message were placed on a traditional and creative media (Dahlen, 2005; Rosengren & Dahlen, 2012; Rosengren, Modig & Dahlen, 2015). However, unlike these studies where a pre-test was conducted to determine the most suitable match of medium + product, this study did not test different combinations. This was due to unavailable time and resource costs that come with the creation of such material. Instead the research relied solely on the manipulation checks. Furthermore, the decision was made to place both media in a photograph of the same environment, unlike previous creative media studies which showed a photograph of only the medium itself. This was done because the simple presence of the message in the vicinity of the medium used for the creative medium choice (stairs) might have an effect on the results. For instance, research on nudging -- a small change in the environment to ‗nudge‘ people toward a specific behavior (Vallgårda, 2012) -- has shown that a poster with a health message placed near the stairs can influence whether or not participants take the stairs or not (Kerr, Eves & Carroll, 2001; Soler et al, 2010). In this way any measured increased effects of creative media use over traditional use should in fact be due to the use of creative media and not simply the close vicinity of the message and the stairs. The message that was shown on both advertisements was ―You don‘t need to climb a mountain to exercise‖. The choice for this message was based on previous research into the determinants of exercise behavior, which showed that self-efficacy was one of the most important factors to predict behavioral intentions (Fletcher & Banasik, 2001; Rhodes & Courneya, 2003). The message itself was also constructed in such a way that, in the opinion of the author, it would be likely to implicitly communicate it‘s meaning (e.g., Climbing a mountain is difficult and takes a lot of effort, stairs are much smaller and easier to climb, exercise is not that difficult). With no pre-test there was a risk
17
that the message itself would not have the intended effect.
Procedure
The data was collected through an online survey using Qualtrics and analyzed using SPSS. The survey was first tested by four people close to the author (two dutch, two english) who provided feedback on legibility and clarity of the survey content. This lead to several
changes, making questions and the context easier to read as well as more understandable. For instance, rather than measure intention and self-efficacy through questions, the decision was made to use statements. These were perceived to be less confusing by all testers as well as the author. Data collection started on Tuesday, 7th of October, and ended on the following Saturday (11th). On the starting page of the survey, participants were first briefed shortly about the purpose of the survey as well as given the choice to take the survey in either English or Dutch. The dutch intention and self-efficacy statements were taken from a dutch version of a TPB questionnaire guide (Francis et al, 2010). The other translations were done by the author himself, who is fluent in both languages. Participants were explained that the survey was part of a thesis, that it is not a test, and as such, there could be no right or wrong answers and no good or bad performance. That we were interested in their honest thoughts and opinions pertaining to the questions. Some of which would be concerning a situation that would be presented in the form of a photograph. Furthermore, they were told the survey would take approximately 5 minutes and requested to pay their full attention during this time and to not get distracted. After this introduction they were shown a form of informed consent and asked to agree to it. At this point the survey started in regards to the required data collection. Participants were told they would be presented with a photograph and asked to imagine the described situation: ―Please imagine that you just found yourself in this scenario. You are heading somewhere but you are not in a hurry to get there. Imagine how you would actually behave in a similar situation. Once you have done this you can click to continue to the questions.‖ They were given as long as they wanted to look at the image, similar to Dahlen (2005). Any mention of ‗advertisement‘ was avoided due to potential
18
influence on respondents‘ perception of the material. Instead, respondents were explained that any questions referencing ‗the material‘ refers to the material that was placed on (the billboard to the right of) the stairs. The dependent variables were asked first, starting with attention, recognition and then intentions and self-efficacy measurements. The specific order was chosen so that measures that had the most likelihood of being influenced by other measures were asked first. After these questions only participants who weren‘t in the control group and did recognize the material were asked about their ad attitude and the manipulation checks. Finally, all participants were asked the demographic and language-related control variables. At the end of the survey they were given more information on the purpose of this study. A randomly generated confirmation code was displayed with an explanation that these were for crowdworkers only. This code was then entered by the crowdworkers on the website they were recruited from to confirm their completion of the survey.
Measures
The following measurements were used in this study, listed in the order they appeared in the survey. In this study medium associated behavior was operationalized as the behavior towards taking the stairs while the actual behavior that is targeted by the health
communication message was operationalized as exercising behavior. This choice was made partially due to ethical reasons, specifically with regards to the indirect effects of the
message on the medium. If the study leads to effects on behavior these would be positive effects (taking the stairs more, rather than less). Had the main behavior been an intervention for negative behavior (such as smoking) this might have negatively impacted stair use, which could have had negative effects on respondents‘ health.
Recognition was measured similar to Dahlen (2009), depending on their group
respondents were asked if they recognized seeing the material that was placed on (the billboard to the right of) the stairs. They could answer with either ―yes‖ or ―no‖. ―Yes‖ was recoded as 1 and ―no‖ as 0 (M = .88, SD = .33). Those who answered ―no‖ skipped the measurement for attention, ad attitude and the manipulation checks.
19
Attention was measured using a self-reported measure of study time, respondents
were asked to write down how many seconds they estimated they paid attention to the material. Answers ranged from 1 to 44 seconds (M = 8.95, SD = 8.21).
The following measures for intention and self-efficacy were based on the
recommendations for questionnaire constructs of the theory of planned behavior by Ajzen (n.d.), the dutch statements were taken from Francis et al (2004) and exercise behavior was operationalized similar to previous studies on exercise behavior (Kwan & Bryan, 2010; De Bruijn & Rhodes, 2011). Two measures were used for medium associated self-efficacy and intention. An immediate measurement measuring self-efficacy and intent for the specific situation and a delayed measurement that measures self-efficacy and intent for similar situations the next four weeks.
Immediate medium associated (MA) intention was measured by asking respondents
how much they agreed with the statement ―In the depicted situation I intend to take the stairs instead of the elevator or escalator‖ with answers on a seven-point likert scale ranging from strongly disagree (1) to strongly agree (7) (M = 4.41, SD = 2.10).
Immediate MA self-efficacy was measured by asking respondents how much they
agreed with the statement ―If I wanted to, I feel confident that in the depicted situation I could could take the stairs instead of the elevator or escalator.‖ with answers on a seven-point likert scale ranging from strongly disagree (1) to strongly agree (7) (M = 5.83, SD = 1.55).
Delayed MA intention was measured by asking respondents how much they agreed
with the statement ―In similar situations I Intend to take the stairs instead of the elevator or escalator every time, for the next four weeks.‖ with answers on a seven-point likert scale ranging from strongly disagree (1) to strongly agree (7) (M = 4.29, SD = 2.03).
Delayed MA self-efficacy was measured by asking respondents how much they
agreed with the statement ―If I wanted to, I feel confident that in similar situations I could take the stairs instead of the elevator or escalator every time, for the next four weeks.‖ with
answers on a seven-point likert scale ranging from strongly disagree (1) to strongly agree (7) (M = 5.42, SD = 1.78).
20
Behavioral intent was measured by asking respondents how much they agreed with
the statement ―I intend to exercise for at least 20 minutes, three times per week for the next four weeks.‖ with answers on a seven-point likert scale ranging from strongly disagree (1) to strongly agree (7) (M = 4.67, SD = 2.01).
Behavioral self-efficacy was measured by asking respondents how much they agreed
with the statement ―If I wanted to, I feel confident that I could exercise for at least 20 minutes, three times per week for the next four weeks.‖ with answers on a seven-point likert scale ranging from strongly disagree (1) to strongly agree (7) (M = 5.34, SD = 1.77).
Ad attitude was measured in a similar way to previous studies on creative media, with
four semantically different items: bad/good, unfavorable/favorable (Dahlen, 2005;
Rosengren, Modig & Dahlen, 2015), negative/positive (Rosengren, Modig & Dahlen, 2015) unpleasant/pleasant (Dahlen, 2005) on a seven-point scale (α = .92). These items were averaged out to create a new scale (M = 5.57, SD = 1.10).
Three manipulation checks were included to verify whether or not the experimental stimuli had successfully been perceived as traditional medium vis a vis creative medium advertising. Similar to Dahlen (2009) congruency between the message and the medium was measured by two items on a seven-point semantically different scale: ―do not match/match‖, ―do not fit well/fit well‖. These two items significantly correlated r = .91, p < .01 and were averaged out to create a new scale (M = 5.79, SD = 1.09). Ad creativity was measured similar to Rosengren & Dahlen (2012), participants were asked ―To what extent do you think that the material you just saw is creative?‖ with answers on a seven-point likert scale ranging from (1) not at all creative to (7) very creative (M = 5.57, SD = 1.10). Finally, similar to
Dahlen & Edenius (2007) we measured ad perception to see if respondents categorized the material as advertising or as not. However in this study we used this measurement as a manipulation check. As it was theorized that, unlike traditional media, creative media would fall outside of an individual‘s advertising schema. Respondents answered the question ―How do you perceive the material you just saw? I perceive it as:‖ on a semantic differential where 1 = advertising, and 7= information (M = 5.01, SD = 1.70 ).
21
Results
Preliminary analyses on the demographics using a MANOVA showed that there were no significant differences in these variables between the groups. This should mean that these variables should not lead to significantly different effects across the conditions. Subsequently none were included as covariates.
Manipulation check
Three manipulation checks were included to see if the experimental stimuli had the desired effect in creating a traditional media- and creative media condition. A multivariate analysis of variance (MANOVA) was carried out to see if these were successful. The experimental condition was selected as independent variable (fixed factor) and ad creativity, ad perception and congruency were selected as dependent variables. We found no significant effects of medium type for either ad creativity F (1, 138) = 1.02, p = .31, η2 =.007 , ad perception F (1, 138) = 0.41, p = .52, η2 =.003 or congruency F (1, 138) = 0.08, p = .78, η2 =.001. The
assumption of equal variances in the population was not violated. These results suggest that the experimental stimuli did not create the desired effects and the traditional medium
stimulus was perceived to be no different than the creative medium stimulus. Surprisingly though, as can be seen in table 1. it seems not the case that the creative medium condition failed, but that the traditional medium stimulus was perceived more positive than expected. Considering 4 as neutral point, both conditions were on average considered to be moderately creative, to be slightly more informational than advertising and perhaps most surprisingly in both conditions the medium and the message were considered to be fairly congruent. Compared to similar research these evaluations are not low for the creative condition but high for the traditional condition, in one study (Dahlen, 2009) the traditional medium was on average perceived fairly neutral (M = 4.20) and in an earlier study (Dahlen, 2005) it was perceived even more poorly (M = 2.43). To confirm that there was no difference in effect between the two media types additional analysis were carried out to test the hypotheses.
22 Table 1 – Medium type effects on manipulation checks.
Mean Standard Deviation N Ad creativity: - Traditional media 5.23 1.39 74 - Creative media 5.47 1.42 66 Ad perception: hhhh hhhhhhhh - Traditional media 5.09 1.75 74 - Creative media 4.91 1.64 66 Congruency: - Traditional media 5.82 1.01 74 - Creative media 5.77 1.18 66 Hypotheses testing
Two separate MANOVAs were carried out with the experimental groups as independent variable (fixed factor), the first one included all dependent variables, this excluded the control group from the analyses. The second MANOVA included all dependent variables except for recognition, attention and ad attitude, which lead to the inclusion of the control group in these analyses. The first MANOVA reported an insignificant effect of medium type on any of the dependent variables F(8, 131) = 1.43, p = .19, Wilk‘s Λ = 0.92, partial η2 = .08. The second
MANOVA showed an insignificant effect across the conditions on most dependent variables F(12, 462) = 1.59, p = .09, Wilk‘s Λ = 0.92, partial η2 = .04 but the tests of between-subject effects showed a very weak significant effect of medium type on both immediate F(2, 236) = 5.99, p = .003, partial η2
=.05 and delayed medium associated intentions F(2, 236) = 3.90, p = .022, partial η2
=.03.
H1 stated that a creative medium choice would lead to a greater (a) recognition of and (b) attention toward the ad than a traditional medium choice. Recognition was higher for the traditional medium (M = .89 , SD = .31) than the creative medium (M = .87 , SD = .34 ) while attention was higher for the creative medium (M = 9.05 , SD = 9.01) than the traditional medium (M = 8.89 , SD = 7.50). For both these effects the differences between the two
23
media choices were found to be non-significant. Both the traditional medium choice as well as the creative medium choice lead to equal recognition F(1, 157) = 0.20, p = .66 and attention F(1, 138) = 0.02, p = .90. H1 is therefore unsupported.
H2 stated that a creative media choice (M = 5.47, SD = 1.16) would lead to a more positive ad attitude than a traditional medium choice (M = 5.66, SD = 1.05). The differences in the effects on ad attitude between the two media choices were found to be non-significant however F(1, 138) = 1.07, p = .30 and the hypothesis is therefore unsupported.
H3 stated that creative media use (M = 5.43, SD = 1.72) would lead to a greater increase in self-efficacy for the behavior (exercise) than when using traditional media (M = 5.39, SD = 1.70) and this effect would be partially mediated by the ad attitude. However the differences in self-efficacy between both the media choices and the control group (M = 5.21, SD = 1.89) were found to be non-significant F(2, 236) = 0.34, p = .71. Indicating that the
experimental stimuli had no effect on exercise self-efficacy at all. Since H2 found no
significantly different effect on ad attitude between the media choices and the main effect on self-efficacy was non-significant the mediation effect of ad attitude could not be tested. H3 is therefore unsupported.
H4 stated that creative media use (M = 4.72, SD = 3.04) would lead to a greater intention to perform the targeted behavior (exercise) than when using traditional media (M = 4.73, SD = 1.96), partially mediated by ad attitude and self-efficacy. The differences in behavioral intent between both medium choices and the control group (M = 4.55, SD = 2.07) were found to be non-significant F(2, 236) = 0.20, p = .82. It would seem the experimental stimuli had no effect on exercise intentions at all. With no main effect on intention and H2 and H3 showing no significantly different effect on ad attitude or self-efficacy between the media choices these meditation effects could not be tested. H4 is therefore unsupported.
H5 stated that through association with the message, creative media use increases the self-efficacy toward medium associated behavior to a greater degree than when using traditional media and that this effect would be partially mediated by ad attitude. There were no significantly different effects between the creative medium choice (M = 5.99, SD = 1.27),
24
the traditional medium choice (M = 5.73, SD = 1.48) or the control group (M = 5.78, SD = 1.83) for immediate MA self-efficacy (stair use) F(2, 236) = 0.60, p = .55. The control group
(M = 5.35, SD = 1.99), traditional medium choice (M = 5.53, SD = 1.65) and the creative medium choice (M = 5.38, SD = 1.68) did not significantly differ for the delayed MA self-efficacy F(2, 236) = 0.24, p = .79 either. This seems to indicate that the experimental conditions failed to have an effect on MA self-efficacy entirely. Because the main effect is non-significant and H2 showed no significantly different effect on ad attitude between the media choices the meditation effect of ad attitude could not be tested. H5 is therefore unsupported.
Finally, H6 stated that creative media use increases the MA intention (stair use) to a greater degree than when using traditional media and that this effect would be partially mediated by self-efficacy and ad attitude. For the immediate MA intention participants in the traditional medium choice reported the highest intention to take the stairs (M = 4.76, SD = 1.97), followed by the creative medium choice (M = 4.72, SD = 2.00) while those in the control group scored the lowest (M = 3.76, SD = 2.22). For the delayed MA intention participants in the creative medium choice reported the highest intention to take the stairs every time for the next four weeks (M = 4.62, SD = 1.87), followed by the traditional medium choice (M = 4.48, SD = 1.93) while those in the control group scored the lowest (M = 3.79, SD = 2.21). The experimental conditions had a very weak significant effect on both
immediate F(2, 236) = 5.99, p = .003, η2 =.05 and delayed F(2, 236) = 3.90, p = .022, η2 =.03 medium associated intentions. To compare the differences across the three conditions for these two significant effects a bonferroni post-hoc analyses was performed. For the immediate MA intention the only significant difference found was that between the control group and the traditional medium choice (Mdifference = -1.00, p = .007, d = .45) and the creative media choice (Mdifference = -.96 , p = .012, d = 0.48). No significant difference between the traditional medium choice and creative media choice were detected. This indicates that compared to the control group both the traditional medium choice and the creative medium choice had a moderate effect on immediate intention to take the stairs. But the effects did not
25
differ between the two media choices. For the delayed MA intention the only significant difference found was that between the control group and the creative medium choice (Mdifference = -.83, p = .031, d = .41). No significant difference between the control group and
the traditional medium choice or the traditional medium choice and the creative media choice were detected. This indicates that compared to the control group the creative medium choice had a moderate effect on intention to take the stairs every time for the next four weeks. However, the difference between the control group and the traditional medium choice did approach significance (Mdifference = -.69, p = .085). Considering the means of these two groups are closest it is more likely that the traditional medium choice and the creative medium choice did not differ than that the control group and traditional medium choice did not differ. We can therefore not find support that the creative medium choice leads to a greater MA intention than the traditional medium choice. Since self-efficacy and ad attitude were not found to be significant the mediation effect could not be tested. H6 can therefore not be supported.
Additional analyses
Although it was not hypothesized, we further investigated the effects that ad attitude and self-efficacy may have on behavioral and immediate MA intentions, unrelated to the effects of creative media use. This choice was made because one of our research questions asked whether or not ad attitude could have a relevant effect on the effectiveness of health communications. This question was posed within the context of creative media use but it is equally relevant for health communications in general (Slater, 2006). In the discussion of the theory we predicted that ad attitude would have an effect on both self-efficacy as well as intention while self-efficacy itself would also have an effect on intention. To test the direct effect as well as the indirect effect through self-efficacy of ad attitude on behavioral intention and the two measures of MA intentions, three separate conditional process analyses using Model 4 of Hayes‘s (2013) PROCESS macro were conducted (for a conceptual model see Appendix, fig. 4). The regression model with behavioral intention as dependent variable,
26
immediate MA self-efficacy as partial mediator and ad attitude as independent variable is significant, F(2, 137) = 17.21, p < .001. The regression model can therefore be used to predict the intention to take the stairs. The strength of the prediction is moderate: 20 per cent of the variation in intention to take the stairs can be predicted by ad attitude and self-efficacy (R2 =.20). Ad attitude, b* = 0.62, t = 4.37, p<.001, 95% CI [0.34, 0.90] has a direct significant, strong association with immediate intention to take the stairs while controlling for the indirect effect through immediate MA self-efficacy. Immediate MA self-efficacy, b* = 0.32, t = 2.56, p < .01, 95% CI [0.73, 0.57] has a significant moderate association with immediate intention to take the stairs. Ad attitude, b* = 0.32, t = 3.47, p<.001, 95% CI [0.14, 0.51] also had a significant moderate effect on immediate MA self-efficacy.
The regression model with behavioral intention as dependent variable, behavioral self-efficacy as partial mediator and ad attitude as independent variable is significant, F(2, 137) = 32.25, p < .001. The regression model can therefore be used to predict the behavorial intention (exercise for 20 minutes a day, 3 times a week for the next four weeks). The
strength of the prediction is moderate: 32 per cent of the variation in behavioral intention can be predicted by ad attitude and self-efficacy (R2 =.32). Ad attitude, b* = 0.67, t = 5.34,
p<.001, 95% CI [0.42, 0.91] has a direct significant, strong association with behavioral
intention while controlling for the indirect effect through behavioral self-efficacy. Behavioral self-efficacy, b* = 0.37, t = 4.35, p < .001, 95% CI [0.20, 0.53] has a significant moderate association with behavioral intention. Ad attitude, b* = 0.40, t = 3.27, p<.001, 95% CI [0.16, 0.64] also had a significant moderate effect on behavioral self-efficacy. These results support the notion that ad attitude has both a direct effect on intentions as well as an indirect effect through self-efficacy.
Discussion and conclusion
The main purpose of this study was to investigate the effects creative media use could have on the persuasiveness of health communications, while also considering the possibility of unintended effects towards the medium and the role of attitude towards the ad in facilitating
27
these effects. Creative media use has been shown to result in more persuasive effects than traditional media use in a product advertising context (e.g, Dahlen, Friberg & Nilsson, 2009), but these effects had not yet been tested within the context of health-related behavior. We theorized that health communications using traditional media would suffer from the same negative effects of advertising clutter as product advertising. By avoiding these effects we hypothesized that creative media use would lead to a greater ad recognition and attention (h1). Furthermore the more positive ad evaluations through avoiding clutter, congruency and increased creativity was expected to lead to a more positive attitude toward the ad (h2). This, combined with other aspects of creative media such as it‘s implicit form of communicating were expected to lead to a greater message persuasiveness and increase self-efficacy (h3) and intentions (h4). We found no support for these hypothesizes in this study however. Based on the non-significant results between the two medium conditions and the control group and the non-significant results of all three manipulation checks it is a good possibility that there were two causes for these results. First of all, it seems likely that the message ―You don‘t have to climb a mountain to exercise‖ failed to have a persuasive effect on self-efficacy and intention. As we discussed in the theory, the effects from creative media use come from enhancing the message that is used. With an ineffective message, there would be no increased effect possible. In hindsight the author acknowledges that with no pre-test a message that had already been proven effective in a previous study would have been a better choice. The second cause could explain why there was no difference between the traditional or creative medium even if the message was persuasive to some degree. The manipulation checks were not significantly different and as such both the traditional medium choice as the creative medium choice seem to have been perceived in a similar way.
Surprisingly, the mean values of both groups were relatively positive for all three
manipulation checks (refer back to table 1). It seems illogical that a billboard and an exercise message would be perceived to match/fit well. It might be, then, that by placing the traditional medium choice within the environment that also features the medium for the creative medium choice (in this case the stairs) respondents linked the two together. Thereby causing similar
28
effects for the traditional medium choice as for the creative media choice. Perhaps partially made possible because the message was not extremely implicit and because the
experimental setting would likely have caused a greater cognitive processing of the material than in a real-life scenario. Causing respondents to ponder the meaning of the traditional medium choice more thoroughly than they might have when actually faced with such a scenario. In addition, the negative effects of advertising clutter are partially due to ad avoidance (Rosengren, 2008), but the experimental setting would likely have caused respondents to pay more attention to the billboard than they would normally. Because of these reasons, although the hypotheses found no support, it is not conclusive that creative media use has no increased beneficial effects over traditional media use for health
communications. We make recommendations for future studies at the end of this section. The secondary research aim was to investigate the possibility for unintended effects on the medium itself due to the same associative transfer of meaning that makes the creative media use give meaning to the message. As Dahlen (2005) explained, this transfer may work both ways. These unintended effects could cause unwanted effects on behavior associated with the medium. For instance a message with negative associations could potentially lead to a negative attitude toward the medium and it's associated behavior. In our study we
investigated this associative effect for a positive association, to not risk any adverse changes that might negatively impact our respondents. The message targeted exercise and used the stairs, so using the stairs was considered the medium associated behavior. We hypothesized that creative media would lead to a greater effect on medium associated self-efficacy (h5) and intentions (h6). There was no significant effect found for self-efficacy but we found a significant effect on medium associated intentions. This effect was only significant between the control group and the two media choice groups however. As such there is no support found for the notion that a creative media choice would have a greater unintended effect on the medium associated intentions than a traditional medium choice. But as was stated with the earlier hypotheses, the manipulation checks seem to indicate that the traditional medium choice was perceived similar to a creative medium choice, most likely due to the
29
experimental setting. As such, were the traditional medium choice perceived as expected, and as in previous studies, a significantly different effect from the creative medium choice would not be improbable. We therefore recommend further research into these unintended effects is needed, for both product advertising as well as health communications.
The third research aim of this study was to investigate the role of attitude toward the ad in enhancing message persuasiveness for health communications in the context of creative medium use. Specifically, it was hypothesized that a creative media choice would lead to a greater attitude toward the ad than a traditional media choice and that this would lead to a greater self-efficacy and intention. But as has already been discussed these findings turned out insignificant. However, based on recommendations by Slater (2006) to study the potential effects of ad attitude on health communications effectiveness we decided to perform additional analyses without consideration for the medium type. Even so, these findings might still have possible implications for creative media theory. Specifically, we tested the direct effect of ad attitude and the indirect effect through behavioral self-efficacy on behavioral intention (exercise) as well as the direct effect of ad attitude and the indirect effect through immediate medium associated self-efficacy on immediate medium associated intention (stair use). In both cases ad attitude had a moderate to strong direct and indirect association through self-efficacy with intention. These findings support the notion that ad attitude can enhance the effectiveness of messages in health communications (Slater, 2006). Health communication ads that are perceived more favorably could lead to more favorable messages evaluations as well. Had the experimental manipulation been successful in creating a difference between the traditional and creative medium choice, there could have potentially been a difference in ad attitude between these groups as well.
Managerial implications
Besides the academic implications of our findings there are also some potential implications for health and advertising practitioners. Considering we have found some indications that health-related messages may have unintended effects on health-related behavior associated
30
with the used medium, health campaign practitioners might want to take into consideration the possible effects their intervention may have on behavior associated to the medium. If future studies confirm these effects, policymakers responsible for advertising guidelines might want to consider commissioning research into these effects if future studies support these notions to ensure no negative effects on health-related behavior are unintentionally taking place due to creative media use.
Furthermore, health campaign practitioners might want to take into consideration not only the message that is used but also the context of the way this message is
communication. Considering the possible effects of attitude toward the ad on message persuasiveness. An effective message could potentially become ineffective or less effective due to a negative attitude toward the ad. While a positive attitude toward the ad may make the message more effective.
Limitations & recommendations for future studies
As has been previously mentioned, due to a lack of pre-testing the effectiveness of the used message could not be guaranteed. This potentially lead to insignificant results on some of the findings. Future studies are recommended to pre-test the messages or alternatively to use messages from previous studies that were proven to be effective. Furthermore there are some ecological validity considerations to take into account. By administering an online survey it was not possible to measure actual behavior. The experimental setting may have lead to unnatural behavior in the respondents that do not represent actual behavior in such a situation, affecting the obtained results. This could potentially be why the traditional medium was perceived unexpectedly positive in regards to creativity and congruency with the
message. Referred to as the intention-behavior gap, there is evidence that intentions
regarding health-related behavior do not always predict the actual behavior very well (Fennis, Adriaanse, Stroebe & Pol, 2011). Similarly, all variables were measured using self-report scales, which may not always accurately measure the intended construct due to response biases (Podsakoff, MacKenzie, Lee & Podsakoff, 2003).We therefore recommend future
31
studies to use a field-experiment as research design to measure actual behavior. Or when using an online survey to use a video of a person walking through the environment, rather than a static photo, to create a more ecologically valid setting.
References
Ahearne, M., Gruen, T., & Saxton, M. K. (2000). When the product is complex, does the advertisement's conclusion matter?. Journal of Business Research,48(1), 55-62 Ajzen, I. (1991). The theory of planned behavior. Organizational Behavior and Human
Decision Processes, 50(2), 179-211.
Ajzen, I. (n.d.). Constructing a theory of planned behaviour questionnaire, from
http://people.umass.edu/aizen/pdf/tpb.measurement.pdf.
Berinsky, A. J., Huber, G. A., & Lenz, G. S. (2012). Evaluating online labor markets for experimental research: Amazon. com's Mechanical Turk. Political Analysis, 20(3), 351-368.
Brown, S. P., & Stayman, D. M. (1992). Antecedents and consequences of attitude toward the ad: A meta-analysis. Journal of consumer research, 34-51.
Brown, W. J., & Basil, M. D. (1995). Media celebrities and public health: Responses to'Magic'Johnson's HIV disclosure and its impact on AIDS risk and high-risk behaviors. Health Communication, 7(4), 345-370.
Buhrmester, M., Kwang, T., & Gosling, S. D. (2011). Amazon's Mechanical Turk a new source of inexpensive, yet high-quality, data?. Perspectives on Psychological
Science, 6(1), 3-5.
De Bruijn, G. J., & Rhodes, R. E. (2011). Exploring exercise behavior, intention and habit strength relationships. Scandinavian journal of medicine & science in sports, 21(3), 482-491.
Conner, M., Rhodes, R. E., Morris, B., McEachan, R., & Lawton, R. (2011). Changing exercise through targeting affective or cognitive attitudes.Psychology and Health,
32 26(2), 133-149.
Dahlén, M. (2005). The medium as a contextual cue: Effects of creative media choice.
Journal of Advertising, 34(3), 89-98.
Dahlén, M., & Edenius, M. (2007). When is advertising advertising? Comparing responses to non-traditional and traditional advertising media. Journal of Current Issues &
Research in Advertising, 29(1), 33-42.
Dahlén, M., Rosengren, S., & Törn, F. (2008). Advertising creativity matters.Journal of
Advertising Research, 48(3), 392-403.
Dahlén, M., Friberg, L., & Nilsson, E. (2009). Long live creative media choice.Journal of
Advertising, 38(2), 121-129.
Dahlén, M., Granlund, A., & Grenros, M. (2009). The consumer-perceived value of non-traditional media: effects of brand reputation, appropriateness and expense. Journal
of Consumer Marketing, 26(3), 155-163.
Elliott, B. (1987). A re-look at why its so hard to sell brotherhood like soap.Australian Journal
of Communication, 11, 20-39.
Fishbein, M., & Cappella, J. N. (2006). The role of theory in developing effective health communications. Journal of Communication, 56(s1), S1-S17.
Flay, B. R., & Burton, D. (1990). Effective mass communication strategies for health campaigns.
Fletcher, J. S., & Banasik, J. L. (2001). Exercise self-efficacy. Clinical excellence for nurse
practitioners: the international journal of NPACE, 5(3), 134-143.
Francis, J. J., Eccles, M. P., Johnston, M., Walker, A., Grimshaw, J., Foy, R., Kaner, E. F. S., Smith, L., & Bonetti, D. (2004). Constructing questionnaires based on the theory of planned behaviour. A manual for health services researchers, 2010, 2-12.
Friestad, M., & Wright, P. (1995). Persuasion knowledge: Lay people's and researchers' beliefs about the psychology of advertising. Journal of Consumer Research, 62-74. Godin, G., & Kok, G. (1996). The theory of planned behavior: a review of its applications to
33
Hammer, P., Riebe, E., & Kennedy, R. (2009). How clutter affects advertising effectiveness.
Journal of Advertising Research, 49(2), 159-163.
Herr, P. M. (1989). Priming price: Prior knowledge and context effects. Journal of consumer
research, 67-75.
Hoch, S. J. (2002). Product experience is seductive. Journal of Consumer Research, 29(3), 448-454.
Kerr, J., Eves, F. F., & Carroll, D. (2001). Getting more people on the stairs: the impact of a new message format. Journal of health psychology, 6(5), 495-500.
Kelly, K. J., Slater, M. D., & Karan, D. (2002). Image advertisements‘ influence on
adolescents‘ perceptions of the desirability of beer and cigarettes. Journal of Public
Policy & Marketing, 21(2), 295-304.
Kover, A. J., Goldberg, S. M., & James, W. L. (1995). Creativity vs. effectiveness? An integrating classification for advertising. Journal of Advertising Research, 35, 29-40. Kwan, B. M., & Bryan, A. D. (2010). Affective response to exercise as a component of
exercise motivation: Attitudes, norms, self-efficacy, and temporal stability of intentions. Psychology of Sport and Exercise, 11(1), 71-79.
MacKenzie, S. B., & Lutz, R. J. (1989). An empirical examination of the structural antecedents of attitude toward the ad in an advertising pretesting context. The
Journal of Marketing, 48-65.
Meyers-Levy, J., & Sternthal, B. (1993). A two-factor explanation of assimilation and contrast effects. Journal of marketing research, 359-368.
Monahan, J. L. (1995). Thinking positively: Using positive affect when designing health messages.
Morris, J. D., Woo, C., Geason, J. A., & Kim, J. (2002). The power of affect: Predicting intention. Journal of Advertising Research, 42(3), 7-18.
Milne, S., Sheeran, P., & Orbell, S. (2000). Prediction and Intervention in Health‐Related Behavior: A Meta‐analytic Review of Protection Motivation Theory. Journal of
