Effects of a Television Drama about Environmental Exposure to Toxic Substances

Abstract

Objective.

This study assessed short-term outcomes of viewing an episode of a prime-time television drama in which a child developed cancer after environmental exposure to an illegal pesticide. The study explored the effects among viewers of feeling transported into a narrative world.

Methods.

Respondents (n=2,139) to a post-episode Internet panel survey were asked if they had seen the show and asked questions about their demographic information, their frequency of viewing the television show, the degree to which they had felt transported into a narrative world created by the drama, and their knowledge and beliefs about the health effects of environmental exposure. Conversations with key informants from federal agencies and advocacy groups were also held.

Results.

Episode viewing and narrative transportation were positively associated with knowledge of toxic exposure effects, and transported viewers reported being more likely to report an unusually high number of cancer cases to authorities. The show also appeared to have prompted a clarification of federal pesticide-testing policy.

Conclusions.

Entertainment Education is a promising strategy for disseminating key points of information about environmental health.

Entertainment education (EE) is the use of a serialized drama (e.g., a soap opera) or some other entertainment format to bring about widespread, positive change in public behavior or its prerequisites, such as knowledge and attitudes. In the United States, EE often takes the form of a health-related storyline within an episode of a regularly broadcast commercial television show.¹ Hollywood writers and producers have complete creative control over the content of the storyline, but they frequently consult with medical and public health experts to ensure the accuracy of the health information in their scripts.² There is accumulating evidence that storylines resulting from these consultations can increase health knowledge, health information-seeking behavior, and conversations about health topics, as well as intentions to perform health behaviors, among mass audiences.^3–8

BACKGROUND

Possible mechanisms of EE effects

It has long been believed that EE effects are brought about largely by observing the consequences of the behavior of a role model, a process elucidated in social cognitive theory,⁹ but other theories have begun to shed light on finer-grained causal mechanisms of EE.¹⁰ For example, Cohen¹¹ pointed to the importance of character identification in vicarious learning, and recent studies have shown that character identification can moderate EE effects.^12,13 An alternative to Cohen's explanation is that, rather than identifying with fictional characters, regular viewers form parasocial relationships with them.^14,15 Such relationships deepen the emotional impact of “interactions” with characters, and that may help to explain why stronger EE effects have been demonstrated among regular viewers of a show than among less frequent viewers.^16,17 Unsolicited viewer letters have been cited as anecdotal support for the parasocial relationship explanation.¹

Narrative transportation

The transportation-imagery model (TIM)^18,19 offers additional insight into why some televised narratives have a greater impact than others. Narrative transportation has been defined as the sense of being carried away by a narrative into a fictional world. TIM holds that a narrative increases knowledge, persuades an audience member to alter beliefs and attitudes, and prompts behavioral intentions to the degree that it transports. Unlike the mainly cognitive mechanisms believed to underlie rhetorical, argument-based persuasion, narrative persuasion is thought to engage both cognitive and emotional processes, and the engagement of head and heart is thought to be enjoyable. Enjoying a story should lead an audience member to finish it and perhaps even return to it in some way (e.g., by discussing it with others), thus creating the potential to increase the exposure dosage of any embedded health messages. In the past, scholars examining the effects of emotional arousal and involvement on health-related outcomes such as knowledge, attitudes, and behaviors tended to conclude that emotion strengthens outcomes. For example, a large body of research based on parallel- or dual-process models of attitude change^20,21 has shown that attitudes with both cognitive and emotional components are stronger predictors of intentions to take threat-reducing action.^22,23 However, narratives may have unique persuasive power. For example, they can bring about emotional arousal regardless of whether a story is framed as truth or fiction.²⁴ In addition, immersion in a story can move counterarguments from one's real life out of awareness temporarily, thus rendering the counterarguments less likely to undercut the messages the narrative conveys.²⁵

A reliable scale of transportation has been developed and used to show that written narratives are indeed more persuasive if they are transporting.¹⁸ However, a written narrative is thought to transport the reader partly by evoking vivid images; it is not clear that the TIM would apply to visual media that supply images instead of evoking them.²⁶

An opportunity to study the relationship between narrative transportation and the effects of real-world video media arose when a popular evening television drama broadcast a storyline about the health effects of environmental exposure to toxic chemicals. Experts in environmental health had consulted with writers of the show as the script was being developed. One of the major points the experts made was that it is extremely difficult and rare to link specific toxic exposures to negative health effects in individual cases. Because this notion is counterintuitive for many, an overview of the technical challenges of confirming such a linkage is provided in the next section.

Health effects of hazardous substances in the environment

Negative health effects of toxic environmental exposures range from short-term symptoms, such as drowsiness, headaches, rashes, and convulsions, to serious, long-term maladies, such as growth impairment, cancer, and death. The mechanisms underlying these adverse health effects are highly complex. Whether exposure to a potentially toxic environmental agent will harm health typically depends on the amount of the substance and whether it is absorbed internally; to cause harm, the agent must be able to cross the body's boundary.²⁷ The extent of health effects also depends on (a) the duration, rate, and route of exposure; (b) the health, age, family disease history, and post-exposure behaviors of the exposed individuals; and (c) the presence of and interactions with other substances also present in the environment.

Cancer is among the most widely feared of these health consequences. The World Health Organization's International Agency for Research on Cancer has identified more than 100 substances as known or probable cancer-causing agents and 300 more as suspected carcinogens.²⁸ While cancer death rates have gone down overall in recent years, rates of several cancers hypothesized to be associated with certain environmental exposures (e.g., thyroid cancer and radiation exposure) have increased.²⁹

Although these associations and trends are apparent at the population level, it is seldom possible to establish a statistically sound link between specific exposures to substances and individual cases of cancer. Among the methodological challenges in establishing defensible causal linkages are (a) the often decades-long latency period between exposure to an agent and the development of cancer, (b) the other exposures the patient may have had in the interim, and (c) the gaps in cancer-registry data needed to estimate background rates of disease occurrence.³⁰

These research challenges also pertain to confirming the existence of a suspected cancer cluster (i.e., a greater-than-expected number of cancer cases in a geographic area over a certain period of time).³¹ State and local health departments investigate more than 1,000 reports of suspected cancer clusters each year, but fewer than 15% turn out to exceed background case levels.³² When a cluster is confirmed, further epidemiologic investigation often fails to identify the underlying cause with confidence, unless the cancer is a rare type and the exposure was intense and sustained (e.g., in occupations that require use of chemicals).³³

Investigating cancer clusters is costly in terms of public health resources and credibility. Although the typical ambiguous finding presents an opportunity to educate citizens about preventable causes of cancer and effective screening methods, it frustrates community expectations.³⁴ Mass education strategies such as EE could be used to increase knowledge about the elusiveness of definitive cause-effect linkages, and such knowledge could facilitate future communication between cluster investigators and the concerned public.

The consultation

In the fall of 2006, the executive producer of the NBC television show Law and Order: Special Victims Unit (LO: SVU) learned of some pesticide-testing regulation issues.³⁵ In October, writers from the show contacted Hollywood, Health and Society (HH&S), a program of The Norman Lear Center at the University of Southern California's Annenberg School for Communication and Journalism that conducts health information outreach to the entertainment industry, to inquire about testing indoor pesticides and connections between pesticides and childhood cancers. HH&S introduced the writers to environmental health experts from the National Center for Environmental Health (NCEH) and the Agency for Toxic Substances and Disease Registry (ATSDR) at the Centers for Disease Control and Prevention and the National Cancer Institute. The experts pointed out that children exposed to hazardous chemicals are at higher risk of disease outcomes than exposed adults, but that apparent cancer clusters seldom can be distinguished statistically from background cancer rates, and that efforts to link individual cancer cases to specific environmental causes are usually unsuccessful. The director of NCEH/ATSDR also met with writers from the show in Los Angeles to provide additional information and ideas for storylines.

The storyline

With 13.45 million U.S. viewers, LO: SVU was the 23rd most-watched network show in February 2007.³⁶ An episode called “Loophole” was aired February 6, 2007. It began with detectives from the New York Police Department's elite sex-crime unit receiving photographs of a little boy in his underwear. Initially assumed to be child pornography, the photographs turned out to be data from a chemical company's test of an unregistered pesticide. The test involved spraying the pesticide in an apartment building; the boy developed “cancer in his blood” as a result of repeated exposure to the hazardous substance. A corrupt federal official character had been conducting sham plant inspections, and the detectives quickly uncovered evidence that incriminated the pesticide manufacturer. However, according to the storyline, there was a loophole in federal regulations that protected the company from prosecution and damage liability. Undaunted by their lack of firm regulatory ground, the detectives bluffed their way through a confrontation with the head of the company and he agreed to pay the child's medical bills.

Health information provided

The plot dramatized the health effects of toxic chemical exposure on children, and included a heated exchange between the detectives and a female police department employee in a white lab coat. This medical authority figure argued that it is rarely possible to link a case of cancer to a particular chemical exposure with a sufficient degree of certainty, but the child's life-threatening condition was enough to convince the detectives. Several physical symptoms of toxic chemical exposure were mentioned during the show but were not depicted in action or tied integrally to the plot.

Data source

Nielsen Entertainment maintains a large (n=23,050) Internet panel of U.S. residents who expressed interest in being surveyed about entertainment-related issues when they were recruited at a public venue. To evaluate the effects of the LO: SVU storyline, Nielsen invited HH&S to add items to a wave of the survey already scheduled to be fielded for four weeks (February 21, 2007–March 26, 2007).

Hypotheses

As in previous research, we expected to find an association between knowledge and exposure to the health-related storyline. Consistent with TIM, we hypothesized that, among viewers, changes in knowledge, beliefs, and anticipated action would be associated with the degree of narrative transportation experienced.

METHODS

Participants

We invited panel households with a member who had reported sometimes watching LO: SVU and respondents to two earlier panel surveys on the topic of hazardous chemical exposure to complete an online questionnaire. The response rate was 38% (2,139/5,678). Viewing status was undetermined for 102 respondents; we excluded them from the analyses.

Procedure

We e-mailed eligible panel members invitations to “.. . help us learn more about what viewers like you think about health issues in your favorite prime-time shows.” We told them the anonymous survey would require five to eight minutes to complete and provided them a link to the survey. Panelists who did not respond initially were sent an e-mail reminder after 10 days.

Measures

Demographics.

We collected demographic information from study participants because any member of the panel household may respond to an invitation to participate. Variables included respondent age, gender, race/ethnicity (coded as white and nonwhite), highest level of education, number of children <18 years of age, and income range.

L&O: SVU viewing.

We asked participants how many episodes of LO: SVU they had seen in the past year (response options were 1, 2, 3, 4, and 5 or more). We also asked, “Did you see the LO: SVU episode about testing an unregistered pesticide in an apartment building?” If the respondent reported watching this episode, they were counted as a viewer. Participants who reported not having watched any episodes of the series in the past year were coded as non-viewers.

Symptom knowledge.

We asked participants, “How likely is it that someone who is exposed to toxic waste would develop the following problems: behavioral/learning problems, birth defects, difficulty breathing, cancer, neurological deficits, rash or skin irritation, and leukemia?” Responses were indicated by means of a 10-point, Likert-type scale (1 = not at all likely; 10 = very likely) (α=0.94).

Knowledge that children are vulnerable.

We asked participants, “On a scale from 1 to 10, how much do you agree that children are more likely than adults to develop health problems after exposure to hazardous chemicals?” (1 = strongly disagree; 10 = strongly agree).

Knowledge that a link is difficult to prove.

We asked respondents, “On a scale from 1 to 10, how much do you agree that it is difficult to prove a link between environmental toxins and cancer cases?” (1 = strongly disagree; 10 = strongly agree).

Protection belief.

We asked respondents to rate their agreement with the following statement: “Environmental policies and regulations protect me and my family from exposure to toxic substances in the environment” (1 = strongly disagree; 10 = strongly agree).

Actions participants might take.

We asked participants, “On a scale from 1 to 10, how likely would you be to do one of the following if you discovered an unusual number of cancer cases in your neighborhood: alert the media, call the health department/board, contact the companies that might be producing hazardous waste, contact the local school board, search state reports on toxic waste control, and call a law enforcement agency or official?” (1 = not at all likely; 10 = very likely) (α=0.85).

Narrative transportation.

We created a measure of narrative transportation by modifying an existing scale developed for written narratives by Green and Brock.¹⁹ Respondents indicated their degree of agreement with seven statements describing their viewing experience: found my mind wandering while watching; show affected me emotionally; wanted to find out how the episode ended; when the show ended, I found it easy to put out of my mind; events in the show are relevant to my everyday life; could picture myself in the scene of the events in the show; and events in the show seemed realistic (1 = not at all; 7 = very much) (α=0.72).

Analysis plan

We computed descriptive statistics and bivariate comparisons of viewers and non-viewers of the LO: SVU episode “Loophole.” The mean response of the items each individual completed was used for multi-item scales. We used multiple regression to estimate (a) effects of exposure to LO: SVU “Loophole” for the sample as a whole and (b) effects of narrative transportation among members of the viewer subsample. We adjusted all models for age, gender, education, race/ethnicity, income, and number of children aged <18 years. Analyses were performed using SAS® version 9.1.³⁷

RESULTS

Demographics of viewers and non-viewers

With the exception that non-viewers had more children aged <18 years, viewers and non-viewers of LO: SVU “Loophole” were similar demographically (Table 1). The typical respondent was a white female, >40 years of age, with an annual family income of ≥$70,000.

Table 1.

Frequencies of and differences between demographic and outcome variables for viewers vs. non-viewers of the “Loophole” episode of Law and Order: Special Victims Unit

^aSignificance of F tests adjusted for unequal numbers

^bp<0.10

^cp<0.01

^dScore based on a scale of 1 to 10 where 1 = strongly disagree and 10 = strongly agree

^ep<0.001

^fScore based on a scale of 1 to 10 where 1 = not at all likely and 10 = very likely

^gScore based on mean total of respondents' degree of agreement (where 1 = not at all and 7 = very much) with seven statements describing their viewing experience: found my mind wandering while watching; show affected me emotionally; wanted to find out how the episode ended; when the show ended I found it easy to put out of my mind; events in the show are relevant to my everyday life; could picture myself in the scene of the events in the show; and events in the show seemed realistic

SD = standard deviation

F = F statistic

df = degree of freedom

NA = not applicable

Bivariate contrasts between viewers and non-viewers

Viewers had higher scores for all three knowledge variables than non-viewers, although the difference on the scale of symptom knowledge was significant only for a one-sided test. Viewers and non-viewers did not differ on either the protection belief item or the action likelihood scale.

Regression analyses for all respondents

Being a regular viewer of LO: SVU and having seen the “Loophole” episode were independently and positively associated with knowing that it is hard to prove a link between environmental toxins and cancer cases (Table 2). Higher levels of education, being nonwhite, having more children aged <18 years, and having seen “Loophole” were positively associated with knowing that children are more vulnerable to toxic exposure. Older age and being female were positively associated with symptom knowledge. There were no significant predictors of the protection belief. Age (positive), education (negative), and number of children aged <18 years (positive) were all associated with the action likelihood scale.

Table 2.

Multiple regression of demographics, regular Law and Order: Special Victims Unit viewership, and viewing of the “Loophole” episode on knowledge, attitude toward protective policy, and estimated likelihood of taking action, for the full study sample (n=2,139)

^ap<0.01

^bp<0.05

SEB = standard error of b

L&O: SVU = Law and Order: Special Victims Unit

Regression analyses for “Loophole” viewers

Among viewers, no variables tested were significantly associated with endorsement of the statement that proving a link between environmental toxins and cancer cases is difficult or the protection belief. Being nonwhite, having more children aged <18 years, and reporting higher levels of narrative transportation were positively associated with agreeing that children are more vulnerable to toxic exposure (Table 3). Narrative transportation was positively associated with scores on both the symptom knowledge and anticipated action scales.

Table 3.

Multiple regression of demographics, regular viewership, and narrative transportation on knowledge, attitude toward protective policy, and estimated likelihood of taking action for viewers of the “Loophole” episode of Law and Order: Special Victims Unit

^ap<0.05

^bp<0.001

SEB = standard error of b

L&O: SVU = Law and Order: Special Victims Unit

DISCUSSION

This study may be the first to have demonstrated an association between viewing a fictional televised narrative about exposure to hazardous chemicals and knowledge about the effects of toxic chemical exposure. One thing that viewers of the episode learned—that identifying cancer clusters and linking them to specific toxic agents is inherently difficult—was a central learning objective for the public health experts who served as consultants to writers of the drama. Perhaps because of deeper engagement with the characters and storylines fostered by following them over time, regular viewers of the show learned the most; this is a common finding in EE.

Of three measures of knowledge, scores on the symptom knowledge scale had the weakest association with viewing the episode. One explanation for the weakness of this finding is that symptom information was provided but not actually dramatized in the storyline. For example, there was no pregnant woman in the show, but participants were asked about the connection between birth defects and toxic chemicals in the environment. The integration of health facts into the action of a fictional narrative is an EE mechanism that deserves further study.

Among viewers of the show, those who felt transported by the narrative were more likely to have retained the undramatized symptom information, as well as to learn that children are more vulnerable to toxic exposures and to indicate that they would take action if they suspected a cluster of cases. These observations were consistent with the TIM. Although narrative transportation was not associated with knowing that causal linkages are elusive, there may have been a ceiling effect on this measure among episode viewers. Overall, our findings support extending the TIM from its original focus on written narratives, which require a reader to generate mental images, to visual media, which supply images for a viewer to perceive and process.

Given the strain on cluster investigation resources and the high proportion of inconclusive investigation findings, reinforcing the tendency to take action if one suspects a cancer cluster might be considered an unintended consequence of the EE stimulus. Offsetting this effect was the finding that, although the storyline portrayed a corrupt government official and cast a federal pesticide-testing policy in a negative light, there were no exposure or transportation-related drops in the belief that “environmental policies and regulations protect me and my family from exposure to toxic substances in the environment.” Viewers may have assumed that, like the corrupt official, the policy was a fictional device intended to add drama to the plot.

It has been charged, however, that a pesticide-testing loophole really did exist.³⁸ After the show aired, the U.S. Environmental Protection Agency posted a clarification of its policy on a Web page entitled “Response to Television Show Depicting Illegal Pesticide Human Study.”³⁹ We regard this clear articulation of the limits of allowable human pesticide testing as a positive, policy-level outcome of EE outreach.

Another higher-order effect may also be attributable to the broadcast. The storyline spurred environmental advocacy organizations to circulate an e-mail promoting the show and to organize viewing groups comprising concerned individuals and families of children who had been exposed to pesticides. The families called the national prime-time broadcast of the storyline empowering (Personal communication, Martha Arguello, MD, director, Physicians for Social Responsibility, May 2008).

Limitations

Survey respondents were not asked if they had accessed an informational Web page (www.cdc.gov/nceh/clusters/faq.htm) recommended by the public health consultants and posted on the show's Web page shortly before the broadcast, so the opportunity to study dosage and channel effects was missed. This study used a modified version of an existing scale developed for written narratives by Green and Brock¹⁸ that has demonstrated reliability and validity. The modified version was not subjected to pretesting prior to being used in this study; fortunately, the Chronbach's alpha reliability of the modified scale proved to be acceptable. Other study limitations included use of single-item measures and a sample of volunteers from a Web panel that was skewed toward well-educated, affluent individuals. However, the method was internally valid in that the demographic profiles of exposed and unexposed respondents were not markedly different, and major demographic variables were held constant statistically.

Conclusions

From a practical intervention perspective, these findings suggest that health outreach to Hollywood can facilitate the effective delivery of technical and even counterintuitive information about an important environmental health topic. Despite the fact that public health officials did not have creative control over the content of the network broadcast, the EE offering addressed some of the aims of environmental health advocates without lowering confidence in (and potentially compromising the effectiveness of) agencies that promulgate and enforce environmental regulations. From a theoretical perspective, the most interesting finding was that those who felt more transported by a video narrative were likely to have higher levels of knowledge. This may mean that they learned more from the show and were more primed for action. Future research should explore narrative factors that make all kinds of narratives more transporting.