Framing Charity Advertising: Influences ofMessage Framing, Image Valence, and
Temporal Framing on a Charitable Appeal1
Chun-Tuan Chang2
Department of Business ManagementNational Sun Yat-sen University
Kaohsiung City, Taiwan
Yu-Kang LeeDepartment of Political EconomyNational Sun Yat-sen University
Kaohsiung City, Taiwan
This study examined when and how charitable advertisements could be effective inthe context of child poverty. An experiment investigated the influences of messageframing, image valence, and temporal framing on a charitable appeal. The resultsindicate that image valence enhances framing effects on advertising effectiveness ofa charitable appeal when the image is congruent with the framed message, especiallywhen the image and the message are presented negatively. A short-term temporalframe facilitates effects of a negatively framed message with a negative pictorialpresentation. Alternatively, a long-term temporal frame increases advertising influ-ences of a positively framed message with a positive pictorial image. Relevance forinformation processing of charity advertising is discussed.jasp_555 2910..2935
In civil society, social welfare is delivered not only through governmentagencies, but also through various nonprofit organizations (NPOs). Bothtypes of organizations attempt to build a humane society and coordinatehumanitarianism activities in the interest of deprived people. In recent years,all levels of governments have faced shrinking revenues, and NPOs canno longer rely heavily on governmental funding to sustain or expand theirphilanthropic activities (Nelson, Brunel, Supphellen, & Manchanda, 2006).Charitable donations, therefore, have become vital financial sources for theongoing development of NPOs.
Individual giving is always the largest single source of donations, andmore than 60% of U.S. households give to charity (Ruotolo, 2006). This is
1The earlier findings of this research were presented at the conference of European Asso-ciation of Consumer Research and at the annual meeting of the Association of ConsumerResearch. The authors thank the conference audience and the reviewers for their valuablecomments and suggestions. The authors also appreciate the financial support of the NationalScience Council, Republic of China.
2Correspondence concerning this article should be addressed to Chun-Tuan Chang, Depart-ment of Business Management, National Sun Yat-sen University, No. 70, Lianhai Road,Gushan District, Kaohsiung City 804, Taiwan. E-mail: [email protected] or toYu-Kang Lee, Department of Political Economy, National Sun Yat-sen University, No. 70Lianhai Road, Gushan District, Kaohsiung City, 804, Taiwan. Email: [email protected]
broadly comparable with countries like the UK, Canada, and Australia(Burnett, 1993; Hibbert & Horne, 1996; Pharoah et al., 2006). Since such alarge proportion of the income for charities is provided by individual dona-tions, how to frame messages proficiently to maximize the responses amongpublic contributors is an important arena in the marketing of NPOs, espe-cially when fundraising competitions among NPOs are fierce.
Framing is a communication strategy that is used frequently by marketingcampaigners. The way information is labeled or framed may significantlyinfluence consumers’ judgments and decisions (for a review, see Levin,Schneider, & Gaeth, 1998). This phenomenon has expanded into diversifiedpromotion scenarios (e.g., health behaviors, products, public policy; Bankset al., 1995; Druckman, 2001; Ganzach & Karsahi, 1995). Nevertheless, unre-solved issues in framing research on choice behaviors still remain (e.g.,Chang, 2007a, 2007b). The present article contributes to this research streamby applying framing concepts in promoting charitable campaigns to demon-strate that message framing might not be equally persuasive in all conditions,and could be moderated by two communication format factors: imagevalence and temporal framing.
NPOs use a variety of advertising practices, some of which can be unor-thodox. An appealing ad of a charity promotion generally contains framingheuristics, such as clear and touching words, vivid images, and persuasivestatistical evidence of a public welfare issue to enhance impressions, to stimu-late sympathy, and to motivate donations. The present research tests the ideathat responses to charitable appeals could be influenced by message framingand presentations of vivid images, and determines whether statistics indifferent temporal frames would modify framing effects.
How should advertisers frame a message for promoting a charitabledonation? Should they emphasize potential gains resulting from the dona-tion or the negative consequences of not making the donation? Will theeffects of vivid pictorial information always be positive in influencingindividuals’ attitudes toward donation promotion and induce compliancewith a request? How should advertisers frame the statistics regarding thecharitable issue? Will influences of aforementioned information presenta-tions about a charitable issue cause interaction effects on advertisingeffectiveness?
Child poverty, a pressing social policy issue of the 21st century in theUnited States and elsewhere around the globe, was adopted as a charitablecontext in the present research. The United Nations Children’s Fund(UNICEF), a UN agency dedicated to improving the health and generalwelfare of children worldwide, reports that more than 30% of children indeveloping countries (about 600 million) live on less than $1 US per day.Every night, 2.7 million children go to bed hungry. Every 3.6 seconds, a
FRAMING CHARITY ADVERTISING 2911
person dies of starvation, and usually it is a child under the age of five(UNICEF Innocenti Research Centre, 2005).
Documented effects of child poverty include increased mortality andillness, impaired brain function, high rates of mental illness, and decreasedschool achievement (Lewit, Terman, & Behrman, 1997). In the United States,the child poverty rate was more than 20% in 2005 (UNICEF InnocentiResearch Centre, 2005). Swanson (2004) described child poverty as a socialtime bomb that eventually will affect taxpayers of all states in increasedcrime, welfare, and prison costs. Despite the large social costs of this crisis,remarkably little attention has been directed at this social issue in charitymarketing. Thus, how to frame the issue of child poverty effectively in an adis the focus of the current study.
Conceptual Background and Hypotheses
Message Framing and Applications in a Charitable Appeal
Framing research is grounded in the basic tenets of prospect theory(Kahneman & Tversky, 1979; Tversky & Kahneman, 1981). Framing refersto the presentation of one of two equivalent value outcomes to differentdecision makers, where one outcome is presented in positive or gain terms,and the other is presented in negative or loss terms. Labeling a glass of water“half empty” or “half full” is an illustration of message framing, as each labelpresents only one side of the information (Martin, 1995). Understandingframing effects may aid the development of creative and persuasive advertis-ing copy and layout (Arora, 2000).
Researchers endeavor to adopt framing concepts in public-matter pro-motion (e.g., Chong & Druckman, 2007; Druckman, 2001; Iyengar &Kinder, 1987; Price & Tewksbury, 1997). An issue can be framed in a delib-erately constructed message. Iyengar and Kinder indicated that people’sknowledge and justification about public affairs might be substantiallyshaped by the selection and presentation of information. Issue framingeffects occur when, in the course of describing an issue or event, a speaker’semphasis on a subset of potentially relevant considerations causes individu-als to focus on these considerations when constructing their opinions(Druckman, 2001). Various framing tactics have been extended mainly toapplications in the charity-solicitation literature, emphasizing donationrequests, such as with foot-in-the-door techniques (Schibrowsky & Peltier,1995) and framing-anchoring techniques (Smith & Berger, 1996). However,the effects of goal framing have not been explored on charitable promotionand solicitations.
2912 CHANG AND LEE
Charitable donation can be promoted through positively framed mes-sages (e.g., “With your help, an unfortunate child can have an opportunityfor a bright future”) or negatively framed ones (e.g., “Without your help, anunfortunate child will remain living in the dark”) with the same goal (i.e.,donation behavior promotion). Previous studies have suggested that bothpositive and negative frames will enhance the evaluation of the issue, ascompared with information presented neutrally, but the question here iswhich type of goal framing is the more powerful enhancer (Levin et al., 1998).Consumer research studies have found negative information to be moreattention-grabbing and persuasive than positive information when peoplescrutinize and process information systematically (Banks et al., 1995;Homer & Yoon, 1992).
The phenomenon of negativity bias has been proposed as an explanationfor the greater salience and distinctiveness of loss or negative framing(Martin, 1995). Negatively framed messages are more likely to violate peo-ple’s expectations (Buda & Zhang, 2000). People are more accustomed toseeing arguments framed in positive (i.e., gain) terms, rather than negative(i.e., loss) ones. Studies have suggested that when information in a messageviolates expectations, it is subject to greater scrutiny (Levin et al., 1998).
The other possible explanation for the effectiveness of negative framing isthat it could create an interacting state. The message could arouse or emo-tionally activate consumers (Mayer, Gaschke, Braverman, & Evans, 1992).Negatively framed messages increase consumers’ need for information aboutpotential negative consequences and ways to avoid their occurrence, becausethe messages contain negative consequence information that is arousing andthat offers remedies for overcoming this arousal (Burnkrant & Sawyer, 1983;Chang, 2007a, 2007b).
Similar principles can be applied to charitable appeals. Negativelyframed messages tend to arouse viewers’ self-relevance, consciousness, andsympathy regarding the serious consequences if no action is taken. Thisincreased need for information reflects individuals’ motivation to processthe message. Many charitable appeals also exhibit the attempt to materi-alize the donor’s social responsibility by making him or her feel connected(Smith & Berger, 1996). When people face threats of undesirable socialchanges in the future, they tend to change their behaviors in an effort toavert the threat, which leads them to protect the status quo by supportingthe public policy (Miller & Krosnick, 2004; Miller, Krosnick, Lowe, & Hol-brook, 2002). Presenting outcomes that emphasize negative consequences,mistakes, or inertia may increase donors’ sense of confronting guilt andresponsibility and may cause donors to be more concerned about loss aver-sion. Donors may thus give to avoid negative beneficiary outcomes. Weproposed the following:
FRAMING CHARITY ADVERTISING 2913
Hypothesis 1. When promoting a charitable donation, messagesthat are framed negatively will be more effective than those thatare framed positively.
Moderator of Image Valence on Message Framing Effects
Although negativity bias is commonly used to explain the superioreffectiveness of negative framing (i.e., loss framing), recent meta-analyticalresearch by O’Keefe and Jensen (2007) has suggested that negative framingmay not always be persuasive. A contingent variable—image valence—isproposed to moderate message framing effects. Image display is widely usedin a charitable appeal to enhance vividness effects (Perrine & Heather, 2000;Thornton, Kirchner, & Jacobs, 1991).
Previous advertising research has illustrated the idea that “a picture isworth a thousand words” (Edell & Staelin, 1983; Goolkasian, 2000; Houston,Childers, & Heckler, 1987; Schneider et al., 2001). A negative picture isfrequently seen in emotional appeals, intending to increase altruism toward aneedy recipient, and thus facilitate charitable giving behaviors (Coke, Bat-son, & McDavis, 1978). Such appeals frequently contain a photographdepicting a person in need, presumably designed to personalize the intendedbeneficiaries, enhance compassion, and motivate the responsiveness ofpotential donors.
Thornton et al. (1991) empirically investigated photographic effects, andindicated that the use of a negative photograph was more effective in elicitinggreater contributions, compared to the non-photographic condition. In theirexperiments, a photograph of a tragic child with tears on his cheek served asan effective attention-grabber to attract people. However, some researchersspeculate that these vivid appeals with a negative picture do not necessarilyguarantee that attitudes or behaviors will be positive. Isen and Noonberg(1979), for instance, examined the influence of vivid pictures on contributionsto charity and suggested that people who see unpleasant pictures contributeless than people who do not see the same pictures. Potential donors perceivea photograph of a handicapped child as a clear demand for sympathy anddonations. Reaction against this manipulation would result in desire torestore a sense of freedom of action. The likely consequence of this reactionwould result in less sympathy and fewer donations. A possible explanation isprovided by Pratkanis and Aronson (1992) that when the vivid informationevokes a negative response, then the appeal may backfire.
Previous studies have focused mainly on the vividness effects of pictorialimages (i.e., comparison between vivid information in pictorial display andnon-vivid condition) without examining positive vivid pictorial information
2914 CHANG AND LEE
(e.g., Isen & Noonberg, 1979; Thornton et al., 1991). We now expand on thisresearch by examining image valence (i.e., presenting vivid pictorial informa-tion either positively or negatively in a framed message). Stimulus congruitybetween peripheral and central ad cues has been examined in advertisingresearch (e.g., picture–word congruency, Edell & Staelin 1983; Heckler &Childers, 1992; Houston et al., 1987; verbal–visual congruency, Schneideret al., 2001; music–message congruency, Kellaris, Cox, & Cox, 1993).
When the vivid presentation is congruent with the message content, itincreases attention (e.g., McGill & Anand, 1989). Under such circumstances,a vivid picture could create mental images that are easily retrieved andinteract with message content to facilitate information processing and reten-tion of both the images and the message arguments (Schneider et al., 2001).Thus, advertising effectiveness can be enhanced. Nevertheless, a vivid picturecan reduce the ability to process a message if the images and thoughtsbrought to mind are inconsistent. Under such circumstances, by occupyingindividuals’ working memory with information incongruent with themessage, vivid presentation may make it more difficult for individuals toprocess and remember the message arguments (Smith & Shaffer, 2000). In acharitable giving context, the use of an incongruent picture might induce abad mood, produce reactance, or be too cognitively complex, thus distractingpotential donors and thereby resulting in decreased contributions (Isen &Noonberg, 1979). Therefore, it is hypothesized that image valence mightinteract with message framing on the effectiveness of a charitable appeal. Wepropose the following:
Hypothesis 2. Congruency between a vivid picture and a framedmessage will be more effective than will incongruency between avivid picture and a framed message.
Moderator of Temporal Framing on Message Framing Effects
Consumers are sensitive to the manner in which statistics-based informa-tion is delivered (Chandran & Menon, 2004; Chang, 2006; Gourville, 2003;Wong & Kwong, 2005). Temporal framing is a common format of statisticalpresentation. Statistics can be expressed differently in terms of a temporalframe (e.g., every year, every month, every day, every minute) that objectivelyrefers to the same time period (e.g., Chandran & Menon, 2004; Gourville,2003). For example, we may frame the child poverty issue as “About 1,250children die each hour due to poverty,” or “About 11 million children dieeach year.”
Temporal framing can be explained with the old saying “not seeing theforest for the trees.” Across different health domains, researchers have found
FRAMING CHARITY ADVERTISING 2915
that, compared with health hazard statistics in every-year framing, every-dayframing makes risk appear more proximal and concrete and results inincreased self-risk perceptions, intentions to exercise precautionary behav-iors, concern and anxiety about the hazard, and effective risk communication(Chandran & Menon, 2004). Hence, a short-term temporal frame (e.g.,every day) through a small number size could enhance the communicationpersuasion.
However, some researchers hold opposite observations that statisticswith a large size can be more effective (Elwyn, Edwards, & Kinnersley, 1999;Tversky & Kahneman, 1974; Yamagishi, 1997). According to the concept ofbase-rate neglect, people underutilize relative information about populationstatistics (i.e., denominator of frequency information) and instead overutilizeother salient information, such as a number-size frequency (i.e., numerator;Elwyn et al., 1999; Tversky & Kahneman, 1974; Yamagishi, 1997). Yama-gishi analyzed the rankings of 11 common causes of death given by under-graduates and suggested that a risk would be judged as more serious when thedeath rates are expressed by incidence with a larger base (e.g., 1,286 out of10,000) than a smaller one (e.g., 24.14 out of 100).
The observed effects of different statistical formats violate a basic tenet ofrational decision making; that is, descriptive invariance (Tversky, Sattath, &Slovic, 1988), which argues that individuals should be indifferent to variousways of framing the same issue. Siegrist (1997) also showed that risk infor-mation via a frequency format (e.g., 600 out of 1,000 people will die) leads tomore risk-averse behavior than that via percentage (incidence rate) form(e.g., 60% of people will die). Chang (2006) further suggested that individu-als’ cognitive judgments are influenced by altering the anchoring points (i.e.,actual numbers of deaths, irrespective of actual rates), and frequency infor-mation with a larger number size may thus increase perceptions of outcomeefficacy. Based on the previous discussion, no consistent theory can be drawnto explain the main effect of temporal framing.
In the current research, we are interested in the interaction betweentemporal-framing and message-framing effects in an attempt to investigatehow statistics in different temporal frames will affect the way individualsinterpret information and influence advertising effectiveness of messageframing. Because of the norm of social responsibility, people are expected tohelp disadvantaged ones (Nelson et al., 2006). Presenting a charitable issueusing statistics in a long-term temporal frame may increase the perceivedseriousness of the social issue because of its large size. Individuals mayperceive that the issue is more threatening and may lead to potential detri-mental consequences. When the message is framed negatively (e.g., “withoutyour donation”), statistics framed in a long-term temporal frame by present-ing a large aggregate number could make the consequences severe and
2916 CHANG AND LEE
intense, enhancing the negativity bias. On the contrary, when the issue ispositively presented (e.g., “with your donation”), statistics in a short-termtemporal frame may be more effective than those in a long-term temporalframe because the small number may make the goal more attainable andfurther facilitate willingness to help.
Hypothesis 3. Temporal framing will moderate the relationshipbetween message framing and advertising effectiveness. Statis-tics in a long-term temporal frame will be more effective in anegatively framed message, but those in a short-term temporalframe will be more effective in a positively framed message.
Interrelationship Among Message Framing, Image Valence,and Temporal Framing
The last focus of the present research addresses a three-way interactionamong message framing, image valence, and temporal framing. One questionremains unanswered: Will temporal framing moderate the effects of congru-ency between a vivid picture and a framed message? The arguments aredeveloped following the logic of Hypothesis 2 that congruency betweenimage valence and message framing will be more effective than will incon-gruency. Using an appropriate temporal frame will enhance the advertisingeffectiveness of a charitable appeal when the message and image valence arecongruent. Therefore, discussion will focus on the conditions of congruentmessages.
As discussed earlier, when the donation message is negatively framed, anegative vivid picture will make the message congruent and will increaseadvertising effectiveness as a result of negativity bias. Nevertheless, recentresearch by Chang (2007b) has suggested that negative message framing maybackfire when the perceived risk is high, because people may become less ableto cope with additional negative information in a health context. Althoughstatistics in a long-term temporal frame may enhance the effects of a nega-tively framed message, those large numbers could cause boomerang effectson advertising effectiveness when the charitable issue is perceived as serious.In a charitable giving context, when the tone of a message is too negative, thealtruistic faith and perceived efficacy of donations may be reduced. Potentialdonors may feel that the problematic situation is too severe to be solvedmerely by his or her donation, resulting in the potential donor giving up onhelping. Instead, statistics in a short-term temporal frame by using a smallnumber could assist in avoiding such potential boomerang effects (caused bythe congruent negative message and picture) by providing helpful informa-
FRAMING CHARITY ADVERTISING 2917
tion that the charitable goal is attainable. Thus, it is predicted that in the caseof negative message framing with a negative picture, the advertising effectswill be increased when statistics are presented in a short-term temporalframe, as opposed to a long-term one.
There is a flip side. A positively framed message could be effective when avivid picture is also positive because the message–image congruency mayfacilitate information processing. Statistics in a short-term temporal framemay not be advantageous because it will make potential donors underesti-mate the charitable issue and be overoptimistic about the situation, whichmay decrease their intentions to help. On the other hand, statistics in along-term temporal frame may increase the perceived severity of the circum-stances through base-rate neglect. A balance between the effects of positivemessage (i.e., congruent positive message and picture) and base-rate neglectcan be maintained. Thus, the hypotheses of the three-way interaction effectare as follows:
Hypothesis 4a. Temporal framing will moderate the effects ofcongruency between image valence and message framing on theadvertising effectiveness of a charitable appeal.
Hypothesis 4b. When the message is framed negatively with anegative vivid picture, statistics in a short-term temporal framewill be more effective than will those in a long-term temporalframe.
Hypothesis 4c. When the message is framed positively with apositive vivid picture, statistics in a long-term temporal framewill be more effective than will those in a short-term temporalframe.
Method
Research Design
Moderating roles of image valence and temporal framing on messageframing effects were explored in a 2 (Message Framing: positive vs. negative)¥ 2 (Image Valence: positive picture vs. negative picture) ¥ 2 (TemporalFrame: short-term vs. long-term) factorial design. A combination ofmessages, pictures, and statistical information yielded eight versions ofadvertisements.
The experiment was a between-subjects design and was conductedthrough the Internet. An Internet survey may help to reduce social pressures
2918 CHANG AND LEE
associated with a face-to-face request that induces individuals to contribute,and may represent true responses. Potential respondents received an e-mailinvitation to participate in the study through a highlighted hyperlink toaccess the designed website instantly.
The respondents were asked to evaluate a poster that would be used tolaunch a forthcoming charitable campaign regarding the importance ofmaking personal donations for child poverty. They were then randomlyexposed to one of eight treatment versions by viewing a poster. After viewingthe particular poster, the respondents clicked the “Continue” button andcompleted a questionnaire containing dependent measures, manipulationchecks, charity experiences, and demographics. The server detected the com-pleteness of those questions and allowed respondents to click the “Submit”button. After submission, the participants were debriefed. Special attentionwas given to ensure that no participants would seek out this fictitious NPOand the promoted donation activity depicted in the poster.
Participants
A list of potential participants was collected through the MSN member-ship directory.3 Stratified sampling of age was chosen to select participants.We sent 2,000 messages in October 2005, inviting individuals to participate inthe Web survey with a brief introduction of the research project and theauthors’ affiliations. There were 178 adults (96 males, 82 females) who par-ticipated in the experiment between October 2005 and November 2005.Response rate was 8.9%. Participants’ ages ranged from 18 to 64 years(M = 38.4 years, SD = 5.4).
Manipulations of Message Framing, Image Valence, and Temporal Framing
Prior to the experiment, a pre-test on the manipulations of the researchvariables was run on 30 undergraduates from an introductory marketingcourse who did not take part in the main experiment. The pre-test covered thefollowing sets of sentences, and students were asked to identify the tone of themessage as follows (the positive message framing condition appears first,while the negative message framing condition appears in parentheses):
If only every child was born with a silver spoon. With yourdonation, their life could become hopeful. (There are no silver
3The MSN membership directory provides an advanced member search by user name,gender, age range, and other search criteria through http://members.msn.com/
FRAMING CHARITY ADVERTISING 2919
spoons for children born into poverty. Without your donation,their life would be hopeless.)
The students reported a lower rating in identifying the tone of the nega-tively framed message (M = 1.42) than that of the positive one (M = 6.12), F(1,29) = 16.45, p < .01, on a 7-point semantic-differential scale ranging from 1(mostly negative) to 7 (mostly positive). In the main experiment, the positivemessage framing condition contained the aforementioned excerpts.
To manipulate image valence, we pre-tested six sets of photographs on thesame 30 pilot participants and assessed the images they reported to choose apositive and a negative photograph in the experiment. They were asked torate each image listed relative to the child’s well-being on a 7-point scaleranging from 1 (negative) to 7 ( positive). Two photographs were selected tobe the positive (M = 6.01) and negative (M = 1.82) images, respectively, F(1,29) = 13.98, p < .01.4 The objects in the baby’s mouth metaphorically repre-sent possible future outcomes of the baby’s life. A silver spoon in one’s mouthmeans being born into a life of opportunity and prosperity, while a cockroachsignifies poor living conditions.
Based on objective statistics from Poverty Facts and Stats (Shah, 2006),temporal framing was manipulated through child poverty’s occurrence ratein a short-term temporal frame (i.e., 1,250 children die each hour as a resultof poverty) or a long-term temporal frame (i.e., about 11 million children dieeach year as a result of poverty). The pre-test confirmed that participantsperceived the two temporal frames differently. They used a 7-point scaleranging from 1 (statistics with a small size number) to 7 (statistics with a largesize number), F(1, 29) = 10.55, p < .01 (short-term temporal frame, M = 2.39;long-term temporal frame, M = 5.44). Figures 1 and 2 present examples ofthe manipulated posters.
Post-Manipulation Measures
Thought-listing task. After exposure to the poster, participants wereasked to write their thoughts down in the form of an open-ended question.Thought-listing measures have a lower likelihood of respondents’ boredom,capture a broad range of responses, and reduce the common method cova-riation with other objective scale measures (Homer & Yoon, 1992). The taskcould help identify possible cognitive or affective responses that participantscame up with when first looking at the poster.
4The authors acknowledge the generous permission of Barnardo’s for allowing us to useimages from their award-winning 2003 and 2004 child poverty campaigns, and to modify thecontent for our research.
2920 CHANG AND LEE
Behavioral intention. Intention was assessed in terms of the likelihood ofparticipating in voluntary work, making a donation, and recommendingdonation activity to family or friends on a three-item, 7-point scale rangingfrom 1 (unlikely) to 7 (likely). The items were derived from past research(Bagozzi & Warshaw, 1990; Batra & Stephens, 1994; Maheswaran & Meyers-Levy, 1990). The three items were averaged to derive a composite rating toquantify the advertising effectiveness. Higher numbers indicate higher behav-ioral intention, resulting in higher advertising effectiveness.
Informativeness measure. Participants indicated how interesting, annoy-ing, good, informative, and appealing the poster was to them. Ratings weremade on 7-point scales (Chang, 2007a). An index was created by calculatingthe mean of the five items. Experimental versions should not differ in termsof amount and quality of information the participants perceived.
Manipulation checks. Respondents were asked to indicate how phrases inthe ad were framed on a 7-point scale ranging from 1 (negative) to 7 ( positive).
Figure 1. Positive framing: Positive vivid picture, and statistics in a long-term temporal frame.
FRAMING CHARITY ADVERTISING 2921
They also indicated how the issue of child poverty was presented in thephotograph on a 7-point scale ranging from 1 ( presenting the dark side) to 7( presenting the bright side). The midpoint was labeled neutral. Participantswere also asked to specify how the poster portrayed the statistics of the childpoverty occurrence rate on a 7-point scale ranging from 1 (frequency informa-tion with small number size) to 7 (frequency information with large number size).
Charity experiences. Respondents were asked whether they had helpedcharities during the past 12 months in the form of voluntary services ormonetary donations. A question assessing the largest monetary donationamount during the past 12 months was also included.
Demographics. Age, gender, educational level, annual income, maritalstatus, and family loading were assessed. These factors may influence thelikelihood of charitable donation (e.g., Lee & Chang, 2007). For instance,gender roles have been identified as an important determinant of reactions tocharitable appeals (Brunel & Nelson, 2000, 2003; Nelson et al., 2006).Womentend to promote social change (i.e., related to an altruistic motive) and helpothers who are less fortunate (Newman, 2000). In addition, once people
Figure 2. Negative framing: Negative vivid picture, and statistics in a short-term temporalframe.
2922 CHANG AND LEE
become parents, they tend to be more socially aware and are more likely toengage in social welfare issues (Henley Centre Research, 2003), which mightincrease the tendency of supporting a charity. Related demographics wereconsidered as potential variables that might confound the experiment’sresults.
Results
An initial analysis was conducted to determine whether any of the demo-graphic variables moderated behavioral intention. Only gender was found tobe influential in determining behavioral intention. Female participantstended to have higher behavioral intention (M = 4.66) than did theirmale counterparts (M = 3.52), F(1, 177) = 5.89, p < .05. Scale reliability waschecked, and Cronbach’s alphas of .89 and .93 were obtained for the infor-mativeness measure and behavioral intention, respectively. No statisticallysignificant difference was shown on the informativeness measure, F(1,177) = 0.51, ns, thus suggesting that the eight posters were perceived asequally informative.
Manipulation Checks
Manipulations were checked. As expected, participants reported a higherrating on the semantic-differential scale in identifying the positively framedmessage (M = 4.95) and stated low scores in the negatively framed one(M = 2.87), F(1, 177) = 23.02, p < .01. Respondents in the negative-photograph conditions were more likely to judge the posters as emphasizingthe dark side of child poverty (M = 3.21), whereas those in the positive-photograph conditions judged the posters as focusing on the bright side of apotential donation (M = 4.77), F(1, 177) = 18.99, p < .01.
Participants responding to the conditions of the long-term temporal frametended to perceive the statistics with a large number size (M = 5.19), whereasthose responding to the conditions of the short-term temporal frame weremore likely to judge the statistics with a small number size (M = 2.62), F(1,177) = 23.02, p < .01. The results indicate that the manipulation of temporalframing was also successful. The overall results confirm that respondentscorrectly identified the emphasis that the posters were designed to convey.
Thought-Listing Analysis
Two judges, who were blind to experimental conditions, independentlycoded the thoughts listed into either favorable (i.e., a statement expressing a
FRAMING CHARITY ADVERTISING 2923
positive reaction to the charitable donation) or unfavorable (i.e., a statementexpressing a negative reaction to the charitable donation), neutral (i.e., astatement expressing a reaction that was neither clearly positive nor clearlynegative), or unrelated (i.e., a statement not associated with reactions to thedonation) and calculated the number of thoughts.5 Interrater agreement washigh (g = .92), and disagreements were resolved through discussion.
Examples of favorable thoughts include “Helping those unfortunate chil-dren is important,” and “Children are our future, and we have to help them.”Unfavorable thoughts include “Children’s well-being should be the respon-sibility of the government, instead of us,” “There are other social problemsmore urgent to be fixed,” and “Donations don’t seem useful.” Women tendedto have more thoughts than did men after viewing the ad (women, M = 2.92;men, M = 1.95), F(1, 177) = 8.78, p < .01. Furthermore, women were inclinedto have more favorable thoughts than were men (women, M = 2.23; men,M = 1.02), F(1, 177) = 10.01, p < .01.
An ANCOVA controlling for gender yielded significant main effectsof message framing, F(1, 177) = 5.83, p < .05; and image valence, F(1,177) = 11.33, p < .01, on the number of thoughts. To be specific, the numberof thoughts individuals had after reading the negatively framed messages(M = 3.11) tended to be higher than after reading the positively framed ones(M = 2.52). Compared with a positive vivid photograph (M = 1.86), partici-pants indicated more thoughts after viewing a negative one (M = 3.44).Besides, the respondents who viewed the message–photograph congruencyposters (i.e., the positively framed message with the positive photograph orthe negatively framed message with the negative photograph; M = 3.64) iden-tified more responses than did those who viewed the incongruent message–photograph posters (M = 3.09), F(1, 177) = 5.74, p < .05. In addition, thenegative congruent version evoked more favorable thoughts than unfavor-able ones (favorable thoughts, M = 2.11; unfavorable thoughts, M = 1.52).The opposite result was found in the positive congruent condition (favorablethoughts, M = 1.11; unfavorable thoughts, M = 1.96).
Tests of Main Hypotheses
A series of ANCOVAs controlling for gender were employed to test thehypotheses. The results are presented in Table 1 and are discussed in the
5The coding scheme used here represents participants’ thought valence. Other codingschemes (e.g., positive–negative thoughts about ad execution, positive–negative thoughts aboutthe charity, self vs. others’ thoughts) were also initially considered. Unfortunately, not enoughvalid responses could be used for those analyses.
2924 CHANG AND LEE
order of the hypotheses.6 First of all, the main effect of message framing wassignificant, F(1, 177) = 7.55, p < .05. The results support Hypothesis 1, whichpredicted that negative framing (M = 5.07) would be more effective thanpositive framing (M = 4.60). No main effects of image valence or temporalframing were observed (Fs < 1). Furthermore, two 2-way interactions wereobtained. A significant two-way interaction effect between message framingand image valence was found, F(1, 177) = 5.42, p < .05. Post hoc tests withBonferroni’s adjustment were further performed.
Congruency between message framing and image valence elicited higherbehavioral intention than did incongruency between message framing andimage valence (negatively framed message/negative vivid picture, M = 5.63;positively framed message/positive vivid picture, M = 5.06; negatively framedmessage/positive vivid picture, M = 4.55; positively framed message/negativevivid picture, M = 4.23). Thus, Hypothesis 2 was confirmed. In addition, atwo-way interaction effect between message framing and temporal framingwas significant, F(1, 177) = 5.87, p < .05.
The results with post hoc tests indicate that when the message was framednegatively, the statistics presented in a long-term temporal frame (M = 5.21)were more effective than those in a short-term temporal frame (M = 4.96),
6An analysis using gender as an independent measure was also conducted. No interactionswere found between gender and researched variables.
F(1, 86) = 5.71, p < .05. The opposite findings were observed when themessage was framed positively (long-term temporal frame, M = 4.52;short-term temporal frame, M = 4.75), F(1, 90) = 5.65, p < .05. Therefore,Hypothesis 3 was supported.
As predicted, a significant three-way interaction of message framing,image valence, and temporal framing was found, F(1, 177) = 7.07, p < .05.Thus, Hypothesis 4a was supported. The results were further examined,according to differently framed messages, as depicted in Figures 3 and 4.When soliciting the donation in the negatively framed message with a nega-tive vivid photograph displayed, statistics framed in a short-term temporalframe (M = 5.81) were found to be more effective than were those in along-term temporal frame (M = 5.45), F(1, 50) = 5.34, p < .05. Conversely,when framing the donation message positively with a positive vivid photo-graph, statistics in a long-term temporal frame (M = 5.23) were more effec-tive than were those in a short-term temporal frame (M = 4.90), F(1,51) = 5.28, p < .05. The results are consistent with Hypotheses 4b and 4c.
1
2
3
4
5
6
7
Negativephoto
Positivephoto
Vivid photo
Beh
avio
ral i
nten
tion
Shorttemporalframe
Longtemporalframe
Figure 3. Interactive effect of positive message framing, image valence, and temporal framing onadvertising effectiveness of a charitable appeal. Note. Advertising effectiveness was measured onthe index of behavioral intention; higher numbers indicate higher behavioral intention, resultingin higher advertising effectiveness.
2926 CHANG AND LEE
Discussion
In the current research, framing heuristics for inducing compliance with arequest were examined as their abilities to maximize the effectiveness ofcharitable appeals. The results support the general proposal that advertisingeffectiveness depends on complicated interrelationships among messageframing, image valence, and temporal framing. Starting with goal framing,we considered negativity bias as the framing theory. Applying framing heu-ristics to charity advertising, the results show that a charitable message thatis framed negatively leads to higher advertising effectiveness than one that isframed positively. This lends qualified support to the phenomenon of nega-tivity bias (e.g., Chang, 2007a, 2007b; Martin, 1995) to provide an empiricalapplication of loss aversion in a charitable context.
In addition to message framing, image valence with vivid picture presen-tation was found to be influential in charity advertising. Although previous
1
2
3
4
5
6
7
Negativephoto
Positivephoto
Vivid photo
Beh
avio
ral i
nten
tion
Shorttemporalframe
Longtemporalframe
Figure 4. Interactive effect of negative message framing, image valence, and temporal framingon advertising effectiveness of a charitable appeal. Note. Advertising effectiveness was measuredon the index of behavioral intention; higher numbers indicate higher behavioral intention,resulting in higher advertising effectiveness.
FRAMING CHARITY ADVERTISING 2927
findings have indicated that vivid pictures can enhance communication effec-tiveness (Edell & Staelin, 1983; Houston et al., 1987), we went one stepfurther to investigate the influences of image valence (i.e., positive vs. nega-tive). A pictorial image corresponding to the message was shown to increasethe impact of message framing, especially when both were presented nega-tively. The thought-listing analysis provided evidence of information process-ing that the number of thoughts and the thought valence of participantsdepend on whether the message and the image are congruent. Compared withmessage–image incongruency, message–image congruency enabled partici-pants to generate more thoughts. Participants who viewed the negative con-gruent messages were more likely to have favorable thoughts, and those whoread the positive congruent message were inclined to have unfavorablethoughts. The negative image seemed to boost the persuasion power ofnegative message in the experiment, which is consistent with previous find-ings that negative images increase persuasive appeals in promoting charitableappeals (e.g., animal shelters, Pratkanis & Aronson, 1992; world hunger,Thornton et al., 1991). Our results also confirm the findings of Smith andShaffer (2000) that vividness may undermine message effectiveness. Incon-gruency between image valence and message framing actually reduced theadvertising effects.
Interestingly, it appears that some NPOs have caught onto the idea ofusing vividness techniques to promote their causes in practice. For instance,Barnardo’s (one of the most prominent NPOs eliminating child poverty inthe UK and around the globe) has utilized hard-hitting images, which werearguably controversial, to draw public attention. Those ads can be verypowerful tools to highlight the damage child poverty can cause for children,which encourages policymakers to search for solutions (Morgan, 2000). Anegatively framed message, especially with a matched image attached, couldmore concretely illustrate miserable outcomes when help is not given. Undersuch circumstances, potential donors may feel sympathy and a responsibilityto help.
Temporal framing was found to moderate the message framing effects.Previous studies of base-rate neglect focused on areas including public healthcommunication (Elwyn et al., 1999; Yamagishi, 1997) and consumer productevaluation (Chang, 2006; Wong & Kwong, 2005). This study is one of thefirst to explore the impact of statistical framing in the field of charity adver-tising. Statistics in a long-term temporal frame (with a large number size)could enhance the influences of negativity bias, and thus facilitate the effectsof negative framing. However, we found that the advantageous effects oflarge number size may not always occur. A three-way interaction amongtemporal framing, message framing, and vivid pictorial information wasfound.
2928 CHANG AND LEE
A noteworthy finding of our experimental results is that the effects ofbase-rate neglect may become limited when the framed message and the vividphotograph are both negative. Statistics in a long-term temporal frame couldhave exaggerating effects on the severity of the charitable issue and couldbackfire on potential donors’ behavioral intentions when people may alreadysee the issue pessimistically. Using statistics in a short-term temporal framecan avoid such boomerang effects and maintain a balance between perceivedseverity and intention to help.
This article makes noteworthy theoretical and managerial contributionsto charity advertising, information processing, decision making, and prac-tice. It integrates the message framing, negativity bias, vividness effects,message congruency, and base-rate neglect literature to examine the effects offraming heuristics on individuals’ responses to persuasive ads promotingcharitable donations. In addition to going beyond simple demonstrations ofmessage effectiveness, the article attempts to clarify when framing effects arelikely to be observed, reversed, or eliminated with considerations of factorsfrom pictorial display and statistical presentation. Understanding the impor-tance of advertising framing is extended by providing evidence that differentinformation formats can affect advertising effectiveness in different ways.
The limitations inherent in this study present opportunities for futureresearch. First, research should examine whether the nature of charitablecontexts would moderate framing effects. Generalizing our results acrosscharity/nonprofit areas, such as underprivileged groups of patients withserious diseases, environmental protection, and science and art promotionwould be desirable. The nature of charitable issues might determine whethernegative pictorial display can strengthen or weaken communication effective-ness, depending on how people perceive whether the situation can beimproved or remedied. For instance, differences between our findings andthose reported by Isen and Noonberg (1979) may reflect the specific nature ofcharitable appeals and the particular photographs employed. With theMarch of Dimes’ poster child in Isen and Noonberg’s study, people may havereacted negatively to a negative photograph because they perceived that littleor nothing could be done to help the person depicted when requests are madefor the purpose of prevention of birth defects, rather than treating them. Aperceived incongruency between the depicted person and the stated purposeof a solicitation may contribute to the photograph being counterproductive.
The nature of child poverty may be different from that of being borndisabled. The issue of child poverty can be remediable and treatable. Nega-tive vivid photographs might stimulate more sympathy and compassiontoward the figure in the ad (i.e., underprivileged little baby in our case) andprovoke approving attitudes toward the promoted charitable appeals. Suchdifferentiation may be validated in future research.
FRAMING CHARITY ADVERTISING 2929
Second, different manipulations of vivid information and statistical pre-sentation may be considered in future research. Case stories are recognizedas a popular alternative to create vividness effects (e.g., Rook, 1986, 1987).Vivid events that are easier to remember are perceived as being more frequentor more probable (Tversky & Kahneman, 1974). Taylor and Thompson(1982) indicated that the audience tends to be more interested in and influ-enced by case stories with anecdotal evidence than abstract information withstatistical evidence. Episode narratives “outpull” statistics (Smith & Berger,1996).
Concrete personal stories could improve comprehension and impression.Small and Loewenstein (2003) found that people contributed more to acharity when the contributions were framed to benefit a family that hadalready been selected from a list than when told that a family would beselected from the same list. Examination with case stories will assistresearchers in assessing whether the interaction effect between framing effectsand vivid presentation can be replicated. The other common way to framestatistics is to present frequency information with different sizes of numera-tors (e.g., Chang, 2006; Wong & Kwong, 2005). For example, “640 millionout of 1.9 billion children from the developing world live without adequateshelter” is statistically equivalent to “1 in 3 children from the developingworld live without adequate shelter.” Frequency with a large-size numeratormay cause the illusion of base-rate neglect. Different manipulations willfacilitate the generalizability of our findings in understanding how charitableappeals work.
Third, what other boundary conditions may come into play? Framinginfluences on mood that might act as a moderator deserve further attention.Researchers have suggested that positively framed appeals in charitablesolicitations influence more favorably than do those framed negatively sincegift giving is generally associated with positive affect (Smith & Berger, 1996).Smith and Berger’s findings were explained based on the thesis of Arrow(1974), which argued that donors give altruistically to benefit or enhancethe satisfaction of others. This also fits the notion of warm-glow giving(Andreoni, 1990). The positive affect literature (Isen, 1993) suggests thatsituations associated with positive feelings give rise to a general state ofpositive affect that motivates people to maintain the positive feeling state.
A recent study by Chang (2007b) further suggests that framing effects areinfluential on people in a positive mood, but are attenuated for those in anegative mood in the context of healthcare advertising. Thus, one possibleproposition could be that a positive mood might moderate framing effects oncharitable advertising effectiveness and increase the effects of positiveframing. In addition, during the past few years, charity fundraising acrosscountries has developed from a predominantly nonprofessional, voluntary-
2930 CHANG AND LEE
based exercise to a highly competitive, professionally managed marketingoperation (Perrine & Heather, 2000). Further research could consider theeffectiveness of various media channels (e.g., broadcasts on radio, television,or Internet) for charitable appeals, which will have direct applications in theexecutions of format and the layout of charity advertising used to grab publicattention.
In summary, the present study has provided new insights into presenta-tion formats while promoting charitable donation messages to the public, inaddition to suggestions of specific attributes associated with informationformats commonly used in charitable communication. How information andstatistics are expressed makes little factual and mathematical difference, butit does create a psychological significance. The findings suggest that appro-priately framed messages should be incorporated into charity advertising topromote public endorsement, leading to increased donations for nonprofitorganizations.
References
Andreoni, J. (1990). Impure altruism and donations to public goods: Atheory of warm-glowing giving. Economic Journal, 100, 464–477.
Arora, R. (2000). Message framing and credibility: Application in dentalservices. Health Marketing Quarterly, 18, 29–44.
Arrow, K. J. (1974). Gifts and exchanges. Philosophy and Public Affairs, 1,343–362.
Bagozzi, R. P., & Warshaw P. R. (1990). Trying to consume. Journal ofConsumer Research, 17, 127–140.
Banks, S. M., Salovey, P., Greener, S., Rothman, A. J., Moyer, A., Beauvais,J., et al. (1995). The effects of message framing on mammography utili-zation. Health Psychology, 14, 178–184.
Batra, R., & Stephens, D. (1994). Attitudinal effects of ad-evoked moods andemotions: The moderating role of motivation. Psychology and Marketing,11, 199–215.
Brunel, F. F., & Nelson, M. R. (2000). Explaining gendered responses to“help-self” and “help-others” charity ad appeals: The mediating role ofworld-views. Journal of Advertising, 29, 15–28.
Brunel, F. F., & Nelson, M. R. (2003). Message order effects and genderdifferences in advertising persuasion. Journal of Advertising Research, 43,330–341.
Buda, R., & Zhang, Y. (2000). Consumer product evaluation: The interactiveeffect of message framing, presentation order, and source credibility.Journal of Product and Brand Management, 9, 229–242.
FRAMING CHARITY ADVERTISING 2931
Burnett, K. (1993). Relationship fundraising. London: White Lion Press.Burnkrant, R. E., & Sawyer, A. G. (1983). Effects of involvement and
message content on information processing intensity. In R. J. Harris(Ed.), Information processing research in advertising (pp. 43–64). Hillsdale,NJ: Lawrence Erlbaum.
Chandran, S., & Menon, G. (2004). When a day means more than a year:Effects of temporal framing on judgments of health risk. Journal ofConsumer Research, 31, 375–389.
Chang, C. (2007a). Health-care product advertising: The influences ofmessage framing and perceived product characteristics. Psychology andMarketing, 24, 143–169.
Chang, C. (2007b). Interactive effects of message framing, product perceivedrisk, and mood: The case of travel healthcare product advertising. Journalof Advertising Research, 47, 51–65.
Chang, C. (2006). Is a picture worth a thousand words? Influences of graphi-cal illustration on framed advertisements. In C. Pechmann & L. L. Price(Eds.), Advances in consumer research (Vol. 33, pp. 104–112). Duluth,MN: Association for Consumer Research.
Chong, D., & Druckman, J. N. (2007). Framing theory. Annual Review ofPolitical Science, 10, 103–126.
Coke, J. S., Batson, D., & McDavis, K. (1978). Empathic mediation ofhelping: A two-stage model. Journal of Personality and Social Psychology,36, 752–766.
Druckman, J. N. (2001). On the limits of framing effects: Who can frame?Journal of Politics, 63, 1041–1066.
Edell, J. A., & Staelin, R. (1983). The information processing of pictures inprint advertisements. Journal of Consumer Research, 10, 45–61.
Elwyn, G. J., Edwards, A. G. K., & Kinnersley, P. (1999). Shared decisionmaking in primary care: The neglected second half of the consultation.British Journal of General Practice, 49, 477–482.
Ganzach, Y., & Karsahi, N. (1995). Message framing and buying behaviors:A field experiment. Journal of Business Research, 32, 11–17.
Goolkasian, P. (2000). Pictures, words, and sound: From which format arewe best able to reason? Journal of General Psychology, 127, 439–459.
Gourville, J. T. (2003). The effects of monetary magnitude and level ofaggregation on the temporal framing of price. Marketing Letters, 14,125–135.
Heckler, S. E., & Childers, T. L. (1992). The role of expectancy and relevancyin memory for verbal and visual information: What is incongruency?Journal of Consumer Research, 18, 475–492.
Henley Centre Research. (2003). The 2005 New Life Campaign. RetrievedSeptember 14, 2009, from www.barnardos.org.uk/resources
2932 CHANG AND LEE
Hibbert, S., & Horne, S (1996). Giving to charity: Questioning the donordecision process. Journal of Consumer Marketing, 13, 4–13.
Homer, P. M., & Yoon, S. (1992). Message framing and the interrelation-ships among ad-based feelings, affect, and cognition. Journal of Advertis-ing, 21, 19–33.
Houston, M., Childers, T. L., & Heckler, S. E. (1987). Picture–word con-sistency and the elaborative processing of advertisements. Journal ofMarketing Research, 24, 359–369.
Isen, A. M. (1993). Positive affect and decision making. In M. Lewis & J.Haviland (Eds.), Handbook of emotions (pp. 413–435), New York:Guilford.
Isen, A. M., & Noonberg, A. (1979). The effect of photographs of thehandicapped on donation to charity: When a thousand words may be toomuch. Journal of Applied Social Psychology, 9, 426–431.
Iyengar, S., & Kinder, D. R. (1987). News that matters. Chicago: Universityof Chicago Press.
Kahneman, D., & Tversky, A. (1979). Prospect theory: An analysis of deci-sion under risk. Econometrica, 47, 263–291.
Kellaris, J. J., Cox, A. D., & Cox, D. (1993). The effects of background musicon ad processing: A contingency explanation. Journal of Marketing, 57,114–125.
Lee, Y. K., & Chang, C. T. (2007). Who gives what to charity? Characteris-tics affecting donation behavior. Social Behavior and Personality, 35,1173–1180.
Levin, I. P., Schneider, S. L., & Gaeth, G. J. (1998). All frames are notcreated equal: A typology and critical analysis of framing effects. Orga-nizational Behavior and Health Decision Processes, 76, 149–188.
Lewit, E. M., Terman, D. L., & Behrman, R. E. (1997). Children andpoverty. Future of Children, 7, 4–24.
Maheswaran, D., & Meyers-Levy, J. (1990). The influence of message framingand issue involvement. Journal of Marketing Research, 27, 361–367.
Martin, B. A. S. (1995). Is the glass half-empty or half-full? How the framingof advertising messages affects consumer responses. New Zealand Journalof Business, 17, 95–105.
Mayer, J. D., Gaschke, Y. N., Braverman, D. L., & Evans, T. W. (1992).Mood-congruent judgment is a general effect. Journal of Personality andSocial Psychology, 63, 119–132.
McGill, A. L., & Anand, P. (1989). The effects of vividness attributes on theevaluation of alternatives: The role of differential attention and cognitiveelaboration. Journal of Consumer Research, 16, 188–196.
Miller, J. M., & Krosnick, J. A. (2004). Threat as a motivator of politicalactivism: A field experiment. Political Psychology, 25, 507–523.
FRAMING CHARITY ADVERTISING 2933
Miller, J. M., Krosnick, J. A., Lowe, L., & Holbrook, A. (2002). The impactof policy change threat on financial contributions to interest groups. Unpub-lished manuscript, University of Minnesota, Minneapolis–St. Paul, MN.
Morgan, J. (2000). Comments on Barnardo’s advertisements of 2000 advertis-ing campaigns. Retrieved January 1, 2006, from http://barnardos.org
Nelson, M. R., Brunel, F. F., Supphellen, M., & Manchanda, R. V. (2006).Effects of culture, gender, and moral obligations on responses to charityadvertising across masculine and feminine cultures. Journal of ConsumerPsychology, 16, 45–56.
Newman, R. (2000). Gender differences in philanthropy. FundraisingManagement, 31, 28–29.
O’Keefe, D. J., & Jensen, J. D. (2007). The relative persuasiveness of gain-framed and loss-framed messages for encouraging disease preventionbehaviors: A meta-analytic review. Journal of Health Communication, 12,623–644.
Perrine, R. M., & Heather, S. (2000). Effects of picture and even-a-penny-will-help appeals on anonymous donations to charity. PsychologicalReports, 86, 551–559.
Pharoah, C., Walker, C., Goodey, L., Wilding, K., Wainwright, S., & Piper,G. (2006). UK giving 2005/06: Results of the 2005/06 survey of individualcharitable giving in the UK. Kent, UK: Charities Aid Foundation.
Pratkanis, A., & Aronson, E. (1992). Age of propaganda: The everyday useand abuse of persuasion. New York: Freeman.
Price, V., & Tewksbury, D. (1997). News value and public opinion: A theo-retical account of media priming and framing. In G. Barnett & F. J.Boster (Eds.), Progress in the communication sciences (pp. 173–212).Greenwich, CT: Ablex.
Rook, K. S. (1986). Encouraging prevention behavior for distant and proxi-mal health threats: Effects of vivid versus abstract information. Journal ofGerontology, 41, 526–534.
Rook, K. S. (1987). Effects of case history versus abstract information onhealth attitudes and behaviors. Journal of Applied Psychology, 17, 533–553.
Ruotolo, G. C. (2006). July 2006: Special edition detailing giving USA 2006findings. Retrieved February 18, 2007, from http://www.ruotoloassoc.com/newsletter/NetlinksJuly06/NETLINKS_July2006specialfnl.htm
Schibrowsky, J. A., & Peltier, J. W. (1995). Decision frames and directmarketing offers: A field study in a fundraising context. Journal of DirectMarketing, 9, 8–16.
Schneider, T. R., Salovey, P., Pallonen, U., Mundorf, N., Smith, N. F., &Steward, W. (2001). Visual and auditory message framing effects ontobacco smoking. Journal of Applied Social Psychology, 31, 667–682.
2934 CHANG AND LEE
Shah, A. (2006). Poverty facts and stats-global issues. Retrieved September14, 2009, from www.globalissues.org/TradeRelated/Facts.asp
Siegrist, M. (1997). Communicating low risk magnitudes: Incidence ratesexpressed as frequency versus rates expressed as probability. Risk Analy-sis, 17, 507–510.
Small, D. A., & Loewenstein, G. (2003). Helping a victim or helping thevictim: Altruism and identifiability. Journal of Risk and Uncertainty, 26,5–16.
Smith, G. E., & Berger, P. D. (1996). The impact of direct marketing appealson charitable marketing effectiveness. Journal of the Academy of Market-ing Science, 24, 219–231.
Smith, S. M., & Shaffer, D. R. (2000). Vividness can undermine or enhancemessage processing: The moderating role of vividness congruency.Personality and Social Psychology Bulletin, 26, 769–779.
Swanson, A. (2004). Analysis: Child poverty a life sentence? The WashingtonTimes. Retrieved September 14, 2009, from http://washingtontimes.com/upi-breaking/20041020-035218-9606r.htm
Taylor, S. E., & Thompson, S. C. (1982). Stalking the elusive “vividness”effect. Psychological Review, 89, 155–181.
Thornton, B., Kirchner, G., & Jacobs, J. (1991). Influence of a photographon a charitable appeal: A picture may be worth a thousand words when ithas to speak for itself. Journal of Applied Social Psychology, 21, 433–445.
Tversky, A., & Kahneman, D. (1974). Judgment under uncertainty: Heuris-tics and biases. Science, 185, 1124–1131.
Tversky, A., & Kahneman, D. (1981). The framing of decisions and thepsychology of choice. Science, 211, 453–458.
Tversky, A., Sattath, S., & Slovic, P. (1988). Contingent weighting in judg-ment and choice. Psychological Review, 95, 371–384.
UNICEF Innocenti Research Centre. (2005). Child poverty in rich countries,2005: Innocenti Report Card (No. 6). Florence, Italy: Author.
Wong, E. K. F., & Kwong, J. Y. Y. (2005). Comparing two tiny giants or twohuge dwarfs? Preference reversals owing to number size framing. Organi-zational Behavior and Human Decision Processes, 98, 54–65.
Yamagishi, K. (1997). When a 12.86% mortality is more dangerous than24.14%: Implications for risk communication. Applied Cognitive Psychol-ogy, 11, 495–506.
