Abstract
Introduction
Cancer survivors play a vital role in cancer control as messengers of hope and information, and advocates for prevention and screening. Understanding what makes survivor stories effective can enhance survivor-delivered programs and interventions.
Methods
By random assignment and using a cross-classified design, 200 African American women viewed videotaped stories (n = 300) from 36 African American breast cancer survivors. Analyses examined effects of story attributes (narrative quality, health message strength), participant characteristics (ways of knowing, experience with breast cancer) and identification with the survivor on women’s: (1) level of engagement in the story; (2) positive thoughts about the story; and, (3) remembering key messages about breast cancer and mammography in the story.
Results
Participant characteristics were significant predictors of all three study outcomes, accounting for 27.8, 2.6 and 22.2% of their total variance, respectively. In comparison, the variability in these outcomes that could be attributed to differences in the stories was small (0.6, 1.1 and 2%, respectively). The effects of participant characteristics on level of engagement and positive thoughts were mediated by identification with the survivor.
Conclusions
The best predictor of a woman becoming engaged in a breast cancer survivor’s story and having positive thoughts about the story was whether she liked the survivor and viewed her as similar to herself (i.e., identification).
Implications for cancer survivors
Survivor stories may be most effective when audience members identify with the survivor. Finding key characteristics that can reliably match the two will advance cancer communication science and practice.
Similar content being viewed by others
Introduction
Cancer survivors’ direct experience with the disease and demonstrated success living with cancer makes them especially attractive and credible as messengers of cancer information. Longstanding national cancer control programs like Reach to Recovery, Man to Man, and The Witness Project are built around survivors sharing their experiences to increase public awareness of cancer and use of cancer screening, and provide support and hope to other cancer patients [1–3]. In hundreds of online communities, thousands of cancer survivors share their stories everyday to inform others and build supportive virtual relationships [4]. Yet surprisingly little is known from communication science about how survivor stories are received, what factors make some stories more effective than others, and how reactions to stories differ based on audience characteristics. Answering these questions can help maximize the effects of survivor-delivered cancer control interventions.
This study sought to determine the extent to which and how different attributes of breast cancer survivor stories (six dimensions each of narrative quality and health message strength) and characteristics of audience members (ways of knowing and experience with breast cancer) affected three specific outcomes among those exposed to the stories: level of engagement in the story, positive thoughts about the story, and remembering key messages about breast cancer and mammography in the story. In addition, the study examined how a person’s level of identification with the survivor sharing her story might mediate these effects. Because of racial disparities in breast cancer mortality [5] and the particular need for involving more minority breast cancer survivors in cancer control efforts [6, 7], the survivors who shared their stories and the women who watched them were all African Americans.
This study has direct implications for cancer survivorship. For many survivors, sharing their story in order to help others is an important part of coming to terms with cancer, giving back to others, and enhancing their own quality of life. Indeed, studies have found positive psychological outcomes not just for those who receive such social support, but also those who give it [8, 9]. Although the present study evaluates effects of survivor stories in a general population sample, hearing from survivors should be equally beneficial to those recently diagnosed with cancer by modeling coping skills, helping change their appraisal of their situation, and sharing resources that are available to help them (e.g., a secondary appraisal process [10]). In the specific case of African American breast cancer survivors, relationships with other survivors—which for some may include just hearing another woman’s story [11, pp. 226–7]—have been found to motivate women to obtain follow-up care, while lack of support and information are reported as barriers to follow-up care [12].
A key assumption underlying this research is that not all stories are equally capable of stimulating desired audience reactions. Narrative communication scholars as well as those in scriptwriting and playwriting have identified attributes of well-told stories, including plot and character development, dramatic tension and suspense, emotional power and range, coherence, realism and effective use of imagery [13–15]. Generally speaking, we would expect that stories possessing these attributes of narrative quality ([11], pp. 226–7) would have greater audience impact than stories that lacked them.
Stories may also vary in the degree to which they adhere to established principles of persuasive communication and theories of behavior change. For example, stories that include explicit recommendations or repeat key points may be more persuasive than those that don’t [16, 17]. Similarly, stories that model a desired behavior and the resulting consequences or demonstrate how barriers to a behavior can be overcome should increase the motivation and ability of viewers to enact the same changes [18, 19]. We view these attributes as indicators of health message strength, and suggest their presence in survivor stories will enhance viewer learning.
Reactions to survivor stories may also vary based on differences among those watching the stories. One such difference is the extent to which a person values lived experience as a source of wisdom (i.e., ways of knowing). In an ethnological study of Black and White working class women, Lutrell [20] found that both groups distinguished between school-based knowledge and common sense, but only Black women viewed their common sense as real intelligence. In the eyes of the Black community, it was women’s ability to keep families together, resourcefulness in getting material things for their children, and general ability to survive in a White world that reflected “real intelligence.” This distinction between school-based knowledge and common sense is reported to be central among the beliefs of many poor and working class African American women [21–23]; in short, common sense and wisdom are viewed as coming from life experience outside of school and judged based on one’s ability to cope with everyday problems. Those who highly value such experience-based knowledge may be more likely to respond positively to stories from cancer survivors, precisely because a survivor has personal experience with the disease.
A woman’s familiarity or experience with breast cancer could also affect her reactions to breast cancer survivor stories. One possible mechanism is through increased perceptions of similarity between the woman and the survivor. Perceived similarity—in attitudes, values, personality and/or background and demographics—often leads to liking (see Miller et al. [24], Chapter 3, for a review) and can affect how individuals view a behavior or attitude, enhance modeling effects and provide social validation [25]. When the qualities shared by a story source and recipient are relevant to the story topic (as would be the case with breast cancer experience when listening to a breast cancer survivor’s story), similarity can increase persuasiveness even when attractiveness of the source is controlled [26]. In addition, a woman who has friends who have had breast cancer may be more involved in the topic or view it as more personally relevant. Increased involvement and personal relevance motivate people to attend to and think about message arguments more carefully because they perceive them as having implications for their own lives [27]. Therefore, if a breast cancer survivor’s story contains compelling information and the viewer is highly motivated to think about the message, the viewer should generate more positive thoughts, which will lead to more attitude change in favor of the message arguments.
This exploratory study examined whether and how these attributes of breast cancer survivors’ stories (narrative quality and health message strength) and characteristics of women viewing the stories (ways of knowing and experience with breast cancer) affected the women becoming engaged in the stories, correctly identifying key cancer messages in the stories and generating positive thoughts about the stories.
Methods
Overview
Stories from African American women breast cancer survivors were professionally videotaped and coded for topic, narrative quality, and health message strength. A sample of 300 video clips from these stories was selected and by random assignment each of 200 African American women viewed three different clips, with each clip viewed by two different women. Cross-classified analyses examined whether and how story attributes and participant characteristics affected women’s reactions to the videos. The Saint Louis University Institutional Review Board approved the study.
Capturing survivor stories
Stories from 36 African American women who had been diagnosed with breast cancer and 13 family members of African American women who had been diagnosed with breast cancer were captured on broadcast quality video in a studio-like setting created at a centrally located community venue. Survivors and family members were identified through project partnerships with five local breast cancer support groups for African American women and received $100 each for their participation. Stories were elicited using Wengraf’s [28] biographic narrative interpretive method (BNIM). The BNIM approach includes three distinct sub-sessions within an interview: (1) the survivor or family member telling about her breast cancer experience in her own way without time limits or instruction from the interviewer; (2) the interviewer asking follow-up questions about topics introduced in sub-session 1, to elicit further stories, more detail or seek clarification; and, (3) the interviewer asking other questions of interest in the study but not introduced in sub-session 1 or 2. An African American woman on the project team (KH) conducted all but two of the interviews. Only stories from the survivors (not family members) were included in this study.
Indexing the stories
Forty-nine hours of survivor stories were recorded. To organize this information, we segmented all video into distinct “story units” (n = 1,624); the beginning and end of each story unit was determined by a consensus of project team members working in groups of three or more and adhering to definitions of narrative coherence and fidelity [29] to preserve the integrity of the stories. A story unit was defined as a continuous narrative addressing one of 10 broad topics related to breast cancer (knowledge and awareness, finding a lump, diagnosis event, coping with breast cancer, treatment issues, relationships, experiences with the health care system, follow-up care, reflection, and what I would tell others).
Coding story units
Story units were coded to identify their main breast cancer message (i.e., topic), six attributes of narrative quality, and six dimensions of health message strength. Two members of the project team independently coded each story and discrepancies were resolved by consensus. Inter-coder reliability kappas were computed for each coding variable.
Breast cancer message
Coders identified which, if any, of five key messages about breast cancer or mammography was delivered in the story: “any woman can get breast cancer” (kappa = 0.81), “getting mammograms is important” (0.88), “you can overcome obstacles to getting a mammogram” (0.56), “it’s important to discuss breast cancer with family and friends” (0.71) and “community resources are available to help with breast cancer and mammograms” (0.77).
Narrative quality
Coders examined each story unit for the presence or absence of six narrative quality indicators.
Emotion
Stories portraying strong or volatile emotions may increase an audience member’s emotional involvement in and perceived realism of a story. Based on definitions from Eagly and Chaiken [30], story units were coded for expression of feelings, moods, emotions or sympathetic nervous activity in relation to or subsequently associated with an attitude object (kappa = 0.73).
Dramatic tension
Stories that contain conflict, uncertainty or unresolved events may increase an audience member’s emotional and cognitive involvement in a story. Story units were classified as having dramatic tension if the survivor described: (1) conflict between two or more characters; (2) internal conflict within a character; (3) an event or situation with an undefined or uncertain outcome; or (4) a violation of norms for a situation or event. Inter-coder reliability for this variable was moderate (kappa = 0.52).
Character development
Audience identification and emotional and cognitive involvement with characters should be enhanced when story characters are well developed. We defined characters as individuals or groups of identifiable people that perform an affective, behavioral, or cognitive action about which the survivor talks. Coding determined whether such a person or group (besides the survivor herself) was identified in the story unit (kappa = 0.97).
Self-disclosure
Disclosure of personal information, particularly intimate or private facts, thoughts or emotions, can build audience trust for, liking of, perceived similarity to and identification with a character. We coded as self-disclosure any information a survivor disclosed about herself, her background, life situation, feelings or challenges not related to her breast cancer experience that would help viewers understand her as a person, not just a survivor (kappa = 0.89).
Cultural appropriateness
Stories that include socio-cultural themes may be easier to understand and perceived as more pertinent by members of a given group. Group members may be more motivated to process information in stories if they perceive it as relevant to them. Drawing on our previous work exploring socio-cultural constructs and cancer in African American women [31, 32], we defined cultural appropriateness as present if the survivor expressed African American values (e.g., spirituality, kinship, resilience), beliefs (e.g., fatalism, medical mistrust), customs or traditions (e.g., prayer, not talking about cancer). Inter-coder reliability was high (kappa = 0.88).
Imagery evoking language
A survivor’s use of concrete language can aid viewers in forming mental images that may enhance memory and recall of her story. Such language may also help catch viewers’ attention. We defined imagery evoking language as that which described people, objects or events in ways that are perceptible by the senses and/or used figurative language including metaphors, similes, idioms or hyperbole (kappa = 0.66).
Health message strength
For each story unit, coders determined which, if any, of six elements of health message strength were present.
Message repetition
Coders determined whether the main breast cancer message (one of the five story topics, as coded above) was mentioned once or more than once within each story unit (kappa = 0.73).
Message centrality
The main breast cancer message was coded as “very central” if it was the primary focus or main point of a story unit, without which the story unit would have a different meaning or no meaning at all (kappa = 0.62).
Explicit recommendation
If a survivor made any recommendation explicitly stating an action to be taken to reinforce or support the main breast cancer message, that story unit was coded as including an explicit recommendation (kappa = 0.81).
Behavioral modeling
A story unit was judged to include behavioral modeling if the survivor verbalized thoughts or described actions about how she (or others) solved problems related to the main breast cancer message (kappa = 0.52).
Incentives
Coders determined whether each story unit identified personal, social or material benefits that result from performing a behavior or holding or adopting a belief that is related to the main breast cancer message (kappa = 0.66).
Overcoming obstacles
Obstacles were defined as any individual, interpersonal, organizational, environmental, social or other factor perceived as making it more difficult to act on the main breast cancer message (kappa = 0.71). When obstacles were mentioned, coders also examined whether a solution was provided for none, some or all of the obstacles.
The length (in seconds) of each story unit was also recorded.
Evaluating story effects
Of the story units that addressed one of the five key messages about breast cancer and mammography (n = 373), 300 were selected for inclusion in the study and grouped at random into 100 mutually exclusive sets of three story units each. To obtain a more precise measurement of study outcomes, two different participants viewed each set of three story units. We discovered during the study that one of the 300 story units did not in fact address one of the five key messages, so analyses include only 299 story units. Hereafter, we refer to story units as “video clips” because this is the language used with study participants.
Participants
A community sample of 200 African American women ages 40 years and older was recruited from 11 urban neighborhoods in St. Louis, MO, in which the rate of late stage breast cancer diagnosis is twice the expected rate for the state of Missouri. Neighborhood canvassing approaches were used to identify potential participants, including distribution of flyers door-to-door, individual intercepts with women at bus stops or on the street, and working with managers of apartment buildings and public housing complexes to reach their residents.
Study protocol
Eligible participants were African American women ages 40 years or older who were able to complete a self-administered questionnaire written at a fifth grade reading level and who provided informed consent. Women meeting these criteria were invited on board the Neighborhood Voice, a mobile health communication research facility that study team members drove to each recruiting site. The Neighborhood Voice was configured for this study with two separate interview areas each equipped with a 20-in. touch-screen monitor on which participants viewed the video clips and completed study questionnaires. Once on board, participants completed a brief questionnaire assessing ways of knowing and experience with breast cancer. They then viewed the three video clips one at a time, completing a thought listing procedure and answering questions about the survivor and her story after viewing each video clip.
Measures
Participant characteristics (ways of knowing, experience with breast cancer) were measured before women viewed any of the video clips. Reactions to the video clips (level of engagement in the story, positive thoughts about the story, remembering key messages about breast cancer and mammography) and the survivor (identification with survivor) were measured after women viewed each video clip. Demographics characteristics (age, education, income) were measured after women had viewed all video clips and answered all other questions.
Ways of knowing
Two items measured the extent to which participants agreed or disagreed (five-point response scale, SA-SD) that “wisdom comes from personal experience, not from learning in school” and “personal experience and common sense are more important than formal education.” These items were combined in a single index variable that reports a mean participant response.
Experience with breast cancer
Two separate yes/no items assess whether participants know anyone who has had breast cancer or have had a close friend with breast cancer. These items are combined in a single dichotomous variable (“yes” to either item/no to both items).
Level of engagement
After viewing each video clip, participants completed a brief scale assessing their level of engagement in the story (i.e., “transportation” into the narrative world [33]). We used nine items from Green and Brock’s [34] 11-item Transportation Scale, which had acceptable internal consistency (Cronbach’s alpha = 0.76) in an undergraduate sample. The scale measures cognitive (e.g., “I could easily picture the events”), affective (e.g., “The narrative affected me emotionally”), and imagery (e.g., “I had a vivid image of <character>”) dimensions of narrative transportation using a seven-point response scale (very much to not at all). Imagery items were not included in this study because they did not apply as well to stories presented in video form. For the same reason, “reading” was changed to “watching” where necessary in the wording of some items. Cronbach’s alpha for the nine-item version in our sample was 0.68.
Positive thoughts
Immediately after viewing each video clip, participants were asked to list on a standardized form all the thoughts, ideas or feelings they had while watching the video (i.e., thought listing [27, 35]). Upon completing their list, they were instructed to go back over their list and rate each thought as a good, bad, or neither good nor bad thing about the video (i.e., self-rated polarity). This paper analyzes the total number of positive thoughts listed, which past studies show is significantly correlated with post-message attitudes and persuasion, and much more strongly related to attitudes than simple recall of message points or arguments [35, 36].
Remembering key messages
After viewing each video clip, participants were asked in separate questions whether each of the five key messages about breast cancer and mammography was mentioned in the video. Participants responded yes, no, or not sure to each item. Responses to these items were used to create an index variable that reported the proportion of correct answers, with values ranging from 0 to 1, where 0 = no correct answers and 1 = all correct answers (i.e., participants correctly identified which messages were present and not present in a particular video clip). Thus the index variable reflects both sensitivity and specificity in recalling story content.
Identification with survivor
As specified by Cohen [37], we combined participants’ ratings of similarity to and liking of each survivor to measure level of identification. All items used the same stem (“The woman in this video clip...”), and asked about attitude/value similarity (“thinks like me”, “has values like mine”), background similarity (“comes from a background like mine”, “has an education and income like mine”), general similarity (“is like me”) and liking (“could be a friend of mine”, “would fit in my circle of friends”). All seven items were answered on a 5-point scale (strongly agree to strongly disagree), and responses were combined in a single “identification with survivor” variable.
Statistical analyses
Analyses sought to: (1) assess the effects of participant characteristics and story attributes on women’s reactions to the survivor videos; and (2) explore possible mechanisms through which such effects might occur. Cross-classified analytic models were used because participants belonged to more that one cluster rather than being strictly nested within a cluster. Figure 1 illustrates how participants (P) in the study were crossed with video clips (see panel A), not nested within video clips (as in panel B).
Cross-classified data structure in the study.
Consistent with this cross-classified data structure, general linear mixed models with random intercepts for participant and video clip effects were derived for each of the study’s three main outcomes: level of engagement, positive thoughts and remembering key messages. The components of variance for both cross-classified factors—participants and video clips—are derived using this approach and consistent parameter estimates for fixed effects are also derived simultaneously accounting for the cross-classified nature of the data. Fixed effects for participant characteristics included ways of knowing and experience with breast cancer as well as identification with the survivor. To understand how participants’ identification with the survivor may influence the impact of participant characteristics on each of the three outcomes of interest, we employed a series of sequential general linear mixed models to derive parameter estimates of regression path-type coefficients between participant characteristics (ways of knowing, experience with breast cancer) and identification with survivor, followed by models that predicted each of the three study outcomes (level of engagement, positive thoughts, remembering key messages) from participant characteristics and identification with survivor. The estimates provided on each “path” are similar to path regression coefficients. They are not standardized because they were derived using a generalized linear mixed model in order to account for multiple nesting of participants and video clips as opposed to a standard path analysis model that generally does not account for such nesting structures.
To provide a summary of the utility of focusing on participant characteristics and/or story attributes in these models, components of variance are estimated based on the unconditional null model with no additional predictors. Changes in model deviance scores were used to assess subsequent model fit with statistical significance based on comparisons with the chi-squared distribution. Pseudo R 2 values [38] were also used to summarize the contribution of participant characteristics for explaining total variation in the three study outcomes.
Findings
Participants’ characteristics
All participants (n = 200) were African American women. Forty-two percent had less than a high school education and 82% reported an annual household income of $20,000 or less. The mean age of women in the sample was 56 years. Most participants (61%) reported knowing someone who had breast cancer, and many (46%) said they’ve had a close friend with breast cancer. Sample means indicated that participants valued experience-based knowledge over that gained from formal education and schooling. Table 1 summarizes these findings.
Attributes of survivor stories
A large majority of the video clips included in the study contained attributes of narrative quality: emotions (95%), characters (84%), socio-cultural themes (83%), dramatic tension (74%) and/or self-disclosure (70%); over half (58%) included at least some imagery-evoking language. Most also included attributes of health message strength: message repetition (92%) and centrality (62%), behavioral modeling (81%) and/or incentives (67%); about half of the video clips mentioned one or more obstacles to acting on the main breast cancer message in the story, and about half of these provided some solution to the obstacle(s). On average, video clips were slightly less than 2 min long (112 s). Table 1 summarizes attributes of the survivor stories.
Effects of participant characteristics and story attributes on reactions to survivor stories
For each of the three study outcomes, variation between participants accounted for a significant proportion of total variation, while variation between survivor stories did not (see Table 2). In the unconditional null model for level of engagement in the story, 62.8% of the total variation was between participants compared to 0.6% between videos. The proportions of variation accounted for were similarly imbalanced for positive thoughts about the story (56.4% between participants, 1.1% between videos) and remembering key messages about breast cancer and mammography (17.7% between participants, 2.0% between videos). These findings suggest that in this study, explaining person-to-person variability in reactions to survivor stories will contribute more to understanding story effects than would a closer examination of story attributes. Accordingly, our subsequent analyses focused on understanding how participant characteristics influenced study outcomes.
How do participant characteristics affect reactions to survivor stories?
The “path” diagrams shown in Fig. 2 illustrate results of the sequential general linear mixed models for predicting each study outcome from participant characteristics and identification with the survivor. For each path, the Figure reports a beta coefficient and (in parentheses) a p value. The “paths” from “ways of knowing” and “experience with breast cancer” to “identification with survivor” are obtained from the first of two linear mixed models; the paths from these three variables to each study outcome are obtained from the second model. The degree to which paths emanating from “ways of knowing” and “experience with breast cancer” are significant in the first model but not in the second, along with the degree to which paths from “identification with survivor” to each outcome are significant in the second model provide a gauge for the impact of “identification with survivor” as a mediating variable.
Path diagrams illustrating results of two nested general linear mixed models for predicting each study outcome from ways of knowing, experience with breast cancer and identification with the survivor. Coefficients are presented adjacent to each “path” with corresponding p values in parentheses.
Level of engagement
As shown in model A of Fig. 2, level of engagement in survivor stories is significantly predicted by identification with the survivor, even after controlling for ways of knowing and breast cancer experience—neither of which significantly predicts level of engagement on its own in this model. In other words, the impact of personal characteristics on level of engagement is mediated through identification with the survivor. Taken together, ways of knowing, experience with breast cancer and identification with the survivor explained 27.8% of the between participant variation in engagement with survivor stories, and significantly increased the amount of variability in engagement explained compared to the unconditional null model (deviance statistic: X 2(3) = 194.72, p value < 0.0001).
Positive thoughts
Positive thoughts about the survivor stories were also significantly predicted by identification with the survivor, even after controlling for ways of knowing and breast cancer experience—neither of which significantly predicted such thoughts on its own in this model (model B of Fig. 2). Taken together, ways of knowing, experience with breast cancer and identification with the survivor explained 3% of the between-person variability in positive thoughts. Overall, personal characteristics and identification with the survivor significantly improved explanation of the variability in positive thoughts over the unconditional null model (deviance statistic: X 2(3) = 238.975, p value < 0.0001).
Remembering key messages
As shown in model C of Fig. 2, experience with breast cancer was significantly and positively related to identification with the survivor, which in turn was negatively related to remembering key messages about breast cancer and mammography. At the same time, experience with breast cancer was directly and positively related to remembering key messages. This pattern of findings suggests that the direct impact of personal experience on remembering key messages is positive while the indirect effect (through identification with the survivor) is negative. The magnitude of these differences is statistically significant, however because of the scales of the identification with survivor and remembering key messages values, the net effect of personal experience on remembering key messages is still in the positive direction. Taken together, ways of knowing, experience with breast cancer and identification with the survivor explained 22.2% of the between participant variation in remembering key messages, and significantly increased the amount of variability in remembering key messages explained compared to the unconditional null model (deviance statistic: X 2(3) = 20.248, p value < 0.0002).
Discussion
Survivor stories and other forms of narrative communication are increasingly recognized as promising tools for helping achieve cancer control objectives [11, 39, 40]. Although the use of survivors as messengers of cancer information is not new, evaluations of these efforts have been relatively uncommon, and we have found no studies that sought to identify or examine specific factors that could influence the effectiveness of survivor stories. This exploratory study thus provides the first empirical findings describing the relative importance of story attributes and audience characteristics in understanding African American women’s reactions to breast cancer survivor stories, as well as identifying a pathway through which these effects occur.
Two broad findings from this study stand out. First, for all study outcomes, variation between participants accounted for a far greater proportion of total variation than did variation between the survivor stories. In other words, differences between women viewing the stories and not differences between the stories themselves best explain women’s reactions to the stories. In particular, two participant characteristics examined in the study—ways of knowing and experience with breast cancer—accounted for a good share of participant variation. While the focus of the study was on evaluating survivor stories related to breast cancer risk and screening in a sample of cancer-free women, this finding suggests that similar effects might be expected in an audience of breast cancer survivors. Previous experience with breast cancer (e.g., having a friend who was diagnosed with breast cancer) positively predicted greater impact of the stories. We would therefore expect this to be especially true for cancer survivors, who have had the ultimate experience with breast cancer. Valuing experiential ways of knowing also predicted positive reactions to the stories. If this orientation is indeed common among working class African American women [20–23], we see no reason to expect that it effects would be different among such women who have been diagnosed with breast cancer.
Other person-level characteristics not examined in this study may still prove to be important predictors of women’s reactions to survivor stories. For example, some have proposed that people with limited numeracy skills, lower health literacy, lower self-efficacy for understanding information, and those who mistrust medical authorities might all be especially responsive to information presented through stories [11]. To the extent that these characteristics are more common among socio-economically disadvantaged and medically underserved populations, it’s possible that survivor stories might hold special promise for helping eliminate cancer disparities. Practically speaking, identifying characteristics of individuals who respond more favorably to survivor stories could help those planning cancer control efforts to better match communication strategies with specific audiences.
The second notable finding is that effects of these participant characteristics on level of engagement in the story and positive thoughts about the story was mediated by the extent to which participants identified with the breast cancer survivor sharing her story. Simply put, the best predictor of whether women became engaged in a survivor’s story and had positive thoughts about the story was whether they viewed her as similar to themselves and liked her. This finding is consistent with results from a qualitative evaluation of the Witness Project, a community-based cancer education and screening program in which African American breast and cervical cancer survivors share faith-based cancer stories in church settings. In the study, Bailey and colleagues (2000) found that the survivors were trusted and deemed truthful largely because participants viewed them as having similar cultural values [41]. The finding has implications for survivors who might be considering becoming involved in breast cancer education, support or advocacy efforts, but who have reservations about publicly sharing their stories. In short, it suggests that audiences are not critiquing such survivors on the basis of their storytelling ability, but rather on the basis of how similar they seem to be. A reasonable interpretation of this might be, “be yourself, and chances are some people will identify with you.”
The important role of identification with the survivor is also consistent with theoretical explanations of how stories influence an audience. In contrast to didactic and expository forms of cancer communication in which people are presented with reasons, facts and scientific evidence designed to convince them to think or act in a certain way, stories work by engaging an audience and modeling behaviors and their consequences. According to social cognitive theory, effects of modeling are enhanced when the model (the breast cancer survivor, in this case) is viewed as similar or is admired [42]. This has been demonstrated in experimental studies of breast cancer screening education among African American women [43], and suggested in evaluation findings from community-based breast cancer control efforts. For example, the strongest finding in an evaluation of Reach to Recovery was that volunteers (who are breast cancer survivors) were seen as role models [3]; similarly, the use of peer educators—many of whom were breast cancer survivors – to reach women in need was reported to be one of the most successful aspects of Avon-funded breast cancer programs in community-based organizations ([44], p. 276). Although the current study did not examine behavioral responses to survivor stories (e.g., getting a mammogram), social cognitive theory suggests that all else being equal, participants who identified strongly with a survivor would be more likely to emulate her actions.
The mediating effect of identification also has practical implications for cancer communication efforts involving survivors as messengers of health information. The fundamental challenge may be pairing audiences with survivors they are most likely to view as similar to themselves. If reliable indicators of similarity and liking can be identified, such matches could be engineered using approaches such as computer tailored health communication [45]. Alternatively, audiences might be offered a choice of stories from multiple cancer survivors to increase the likelihood that they find at least one with whom they identify. Web- and other technology-based cancer information resources seem especially well suited to provide such options.
The finding that identification with the survivor negatively affected remembering key messages about breast cancer and mammography was unexpected. We speculate that personal experience with cancer (which directly and positively affected remembering) may help women identify key messages in a survivor’s story, perhaps because personal experience increases involvement in a topic, which motivates people to think carefully about message content [27]. As with several other study findings, this suggests that an audience of survivors—who have direct experience with breast cancer—might also be more likely to remember information from such stories. In contrast, identification with a survivor may make it more difficult to identify messages not included in a story, perhaps because a survivor who was well liked and seen as similar was given the benefit of the doubt (e.g., “she must have talked about this topic”).
It was also unexpected that differences between survivor stories would account for such a small percentage of the total variability in each of the three outcomes. Thus, including attributes of the survivor stories in explanatory models would contribute little to understanding women’s reactions to the videos. We see two possible explanations for this finding. First, because the video clips were short in length, neither scripted nor edited, and showed talking heads only, it is possible that story elements like plot and character development—while measurable—were not so relevant to viewers. In other words, short-form survivor videos that show only the survivor on screen likely accentuate reactions to the survivor, perhaps at the expense of story attributes. Second, the stories included in the study may not have been sufficiently variable to detect differences between them (see Table 1), and our cross-classified design may have benefited from having more than two participants view each story.
Findings from this exploratory study indicate that a woman’s personal experience with breast cancer, experiential ways of knowing, and level of identification with a storyteller are important determinants of her level of engagement in breast cancer survivor stories, positive thoughts about the stories, and remembering key messages in the story. One limitation of the study is that while these outcomes can be important precursors to attitude and behavior change, they are just that, precursors. To help fill this gap, we are now comparing the effects of survivor stories (longer compilations of selected video clips from the study reported here) to videos presenting identical information in a non-narrative way on use of mammography in an intervention trial among a community sample of African American women. This ongoing study will also build upon the current findings by exploring effects of additional participant characteristics (numeracy, medical mistrust, self-efficacy for mammography) as well as exploring how these characteristics might predict differences in reactions, attitudes and behavior to narrative vs. non-narrative communication.
We also recognize the great potential of such stories as a resource for African American women recently diagnosed with breast cancer. Of the 1,624 story units coded, the large majority are not about breast cancer risk and mammography as featured in this study, but rather post-diagnosis issues such as treatment, decision-making, coping, personal relationships, interactions with the health care system, and finding meaning in one’s cancer experience. Building on the findings from this study, we have developed a prototype interactive computer system on a tablet PC that allows women to browse a library of survivor stories by either picking a survivor (i.e., identification with survivor) or selecting a topic. The system is in the third phase of usability testing among African American breast cancer survivors, and will be ready for a randomized trial in early 2008.
Finally, although the unique nature of the participant sample (very low income African American women from a handful of neighborhoods), survivor sample (African American women), and story materials (unedited, unrehearsed personal narratives delivered via video) limits generalizability of the findings, within the context of these conditions there are clear implications for cancer communication research and practice, and particular insight for those seeking to address breast cancer disparities in African American women.
References
Coreil J, Behal R. Man to man prostate cancer support groups. Cancer Pract. 1999;7(3):122–9.
Erwin D, Spatz T, Stotts R, Hollenberg J, Deloney L. Increasing mammography and breast self-examination in African American women using the witness project model. J Cancer Educ. 1996;11(4):210–5.
Rogers TF, Bauman LJ, Metzger L. An assessment of the Reach to Recovery program. CA: A Cancer Journal for Clinicians. 1985;35(2):116–24.
Eysenbach G. The impact of the Internet on cancer outcomes. CA: A Cancer Journal for Clinicians. 2003;53:356–71.
Ghafoor A, Jemal A, Ward E, Cokkinides V, Smith R, Thun M. Trends in breast cancer by race and ethnicity. CA: A Cancer Journal for Clinicians. 2003;53:342–55.
Kaur JS. The potential impact of cancer survivors on Native American cancer prevention and treatment. Cancer. 1996;78(7):1578–81.
Lythcott N, Green BL, Brown ZK. The perspective of African American breast cancer survivor advocates. Cancer. 2002;97(S1):324–8.
Schwartz C, Meisenhelder JB, Ma Y, Reed G. Altruistic social interest behaviors are associated with better mental health. Psychosom Med. 2003;65:778–85.
Thoits PA, Hewlett LN. Volunteer work and well-being. J Health Soc Behav. 2001;42:115–41.
Lazarus RS, Folkman S. Stress, appraisal, and coping. New York: Springer; 1984.
Kreuter MW, Green MC, Cappella JN, Slater MD, Wise ME, Storey D, et al. Narrative communication in cancer prevention and control: a framework to guide research and application. Annals Behav Med. 2007;33(3):221–35.
Thompson HS, Littles M, Jacob S, Coker C. Posttreatment breast cancer surveillance and follow-up care experiences of breast cancer survivors of African descent: an exploratory qualitative study. Cancer Nurs. 2006;29(6):478–87.
Bruner J. Actual minds, possible words. Massachusetts: Harvard University Press; 1986.
Hatcher J. The art and craft of playwriting. Cincinnati, OH: Storey; 1996.
Hicks ND. Screenwriting 101: The essence and craft of feature film writing. New York, NY: McNaughton and Guner; 1999.
Cacioppo JT, Petty RE. Effects of message repetition and position on cognitive responses, recall and persuasion. J Pers Soc Psychol. 1979;37:97–109.
O’Keefe DJ. Standpoint explicitness and persuasive effect: a meta-analytic review of the effects of varying conclusion articulation in persuasive messages. Argum Advocacy. 1997;34:1–12.
Bandura A. Social cognitive theory in cultural context. J Appl Psychol, An International Review. 2002;51:269–90.
Janz NK, Champion VL, Strecher VJ. The health belief model. In: Glanz K, Lewis F, Rimer B, editors. Health behavior and health education. 3rd ed. San Francisco: Jossey-Bass; 2002. pp. 45–66.
Lutrell W. Working-class women’s ways of knowing: effects of gender, race, and class. Sociol Educ. 1989;62:33–6.
Abrums M. “Jesus will fix it after awhile”: meanings and health. Soc Sci Med. 2000;50:89–105.
Banks-Wallace J. Womanist ways of knowing: theoretical considerations for research with African American women. Adv Nurs Sci. 2000;22(3):33–5.
Collins P. (Ed.) (2000). Black feminist thought. 2nd ed. New York: Routledge.
Miller RS, Perlman D, Brehm SS. Intimate relationships. Boston: McGraw-Hill; 2007.
Cialdini RB, Sagarin BJ. Principles of interpersonal influence. In: Brock TC, Green MC, editors. Persuasion: psychological insights and perspectives. 2nd ed. Newbury Park, CA: Sage; 2005. pp. 143–69.
Hass RG. Effects of source characteristics on cognitive responses and persuasion. In: Petty RE, Ostrom TM, Brock TC, editors. Cognitive responses in persuasion. Hillsdale, NJ: Lawrence Erlbaum Associates; 1981. pp. 141–72.
Petty RE, Cacioppo JT, Strathman AJ, Priester JR. To think or not to think: exploring two routes to persuasion. In: Brock TC, Green MC, editors. Persuasion: psychological insights and perspectives. 2nd ed. Needham Heights, MA: Allyn & Bacon; 2005. pp. 81–116.
Wengraf T. Qualitative research interviewing: biographic narrative and semi-structured methods. London: Sage; 2001.
Gubrium J, Holstein J. Narrative practice and the coherence of personal stories. Sociol Q. 1998;39(1):136–87.
Eagly AH, Chaiken S. Attitude structure and function. In: Gilbert D, Fiske S, Lindzey G, editors. The handbook of social psychology. 4th ed. New York: McGraw-Hill; 1998. pp. 269–322.
Deshpande AD, Sanders-Thompson VL, Vaughn K, Kreuter MW (2007) Use of socio-cultural constructs in cancer research among African Americans: A review. Cancer Control (in press).
Kreuter MW, Lukwago SN, Bucholtz D, Clark EM, Sanders-Thompson V. Achieving cultural appropriateness in health promotion programs: targeted and tailored approaches. Health Educ Behav. 2003;30(2):133–46.
Green MC. Transportation into narrative worlds: the role of prior knowledge and perceived realism. Discourse Process. 2004;38(2):247–66.
Green MC, Brock T. The role of transportation in the persuasiveness of public narratives. J Pers Soc Psychol. 2000;79(5):701–21.
Cacioppo JT, Harkins SG. The nature of attitudes and cognitive responses and their relationship to behavior. In: Petty RE, Ostrom TM, Brock TC, editors. Cognitive responses in persuasion. Hillsdale, NJ: Lawrence Erlbaum Associates; 1981. pp. 31–4.
Petty RE, Wegner DT. Attitude change: Multiple roles for persuasion variables. In: Gilbert D, Fiske S, Lindzey G, editors. The handbook of social psychology. 4th ed. New York: McGraw-Hill; 1998. pp. 323–89.
Cohen J. Defining identification: a theoretical look at the identification of audiences with media characters. Mass Commun Soc. 2001;4(3):245–64.
Singer JD, Willett JB. Applied longitudinal data analysis. Oxford: Oxford University Press; 2003.
Green MC. Narratives and cancer communication. J Commun. 2006;56:163S–83S.
Hinyard L, Kreuter MW. Using narrative communication as a tool for health behavior change: a conceptual, theoretical, and empirical overview. Health Educ Behav. 2007;34(5):777–92.
Bailey E, Erwin D, Belin P. Using cultural beliefs and patterns to improve mammography utilization among African-American women: The Witness Project. J Natl Med Assoc. 2000;92(3):136–42.
Bandura A. Self-efficacy: toward a unifying theory of behavior change. Psychol Rev. 1977;84(2):191–215.
Anderson RB, McMillion PY. Effects of similar and diversified modeling on African American women’s efficacy expectations and intentions to perform breast self-examination. Health Commun. 1995;7:327–43.
Antman K, Abraido-Lanza AF, Blum D, Brownfield E, Cicatelli B, Debor MD, et al. Reducing disparities in breast cancer survival: a Columbia University and Avon Breast Cancer Research and Care Network symposium. Breast Cancer Res Treat. 2002;75:269–80.
Kreuter MW, Strecher VJ, Glassman B. One size does not fit all: the case for tailoring print materials. Annals Behav Med. 1999;21(4):276–83.
Acknowledgement
This study was supported by a grant from the National Cancer Institute’s Centers of Excellence in Cancer Communication Research program (CA-P50-95815). The authors thank JoAnne Banks Wallace and Shanti Parikh for assistance with formative research, Jean Freeman, Hollie Milam and Kim Vaughn for assistance capturing survivor stories, Dana Weston, Maria Hamtil-Hockenhull, Delores Dotson, Keri Jupka, Catina Scott, Kim Kennedy and Leslie Hinyard for indexing and coding survivor stories, Christine Dao and Kia Davis for administering the research protocol, and Sharon Homan, Qiang Fu, Melanie Green and Mike Slater for helping with research design and analytic methods. We are especially grateful to the 49 breast cancer survivors and family members who shared their stories, without whom this research would not be possible. The Murchison Tabernacle Cancer Support Group, Sistah Connection, The Breakfast Club, Inc., The Witness Project and Mustard Seed were key partners in identifying breast cancer survivors.
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Kreuter, M.W., Buskirk, T.D., Holmes, K. et al. What makes cancer survivor stories work? An empirical study among African American women. J Cancer Surviv 2, 33–44 (2008). https://doi.org/10.1007/s11764-007-0041-y
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11764-007-0041-y