Matching the Message 1
Running head: Effect of Framing on Pro-Environmental Behavior
Matching the Message to the Audience:
The Effect of Framing on Pro-Environmental Behavior
Andria M. Cimino
Columbia University
Abstract
The gap between American pro-environmental attitudes and actual pro-environmental behavior (PEB) poses a challenge to organizations seeking to encourage Americans to embrace a more conservation-minded lifestyle. Theories abound on the possible causes of this gap. As part of a 2 x 2 between-groups study, 85 Columbia University students participated in two PEB surveys, one framed to appeal to those with self-enhancing (SE) values and the other framed to appeal to those with self-transcending (ST) values. Although the results fell short of significance, unexpected main effects of gender were found. The findings suggest that conservation organizations should make an effort to match their messages to target audiences to realize their goal of increasing PEB in America.
Matching the Message to the Audience:
The Effect of Framing on Pro-Environmental Behavior
Based on response to nationwide public opinion surveys, it would seem that most Americans are concerned about the environment (Kluger, 2006; PollingReport.com, 2006), even ranking protection of the environment as more important than economic growth (PollingReport.com, 2006). Nearly three out of four (74%) of Americans surveyed for a Harris poll conducted in August 2005 agreed with the statement, “Protecting the environment is so important that requirements and standards cannot be too high, and continuing environmental improvements must be made regardless of cost” (PollingReport.com, 2006). Yet there is a large “gap” between this environmental concern and actual pro-environmental behavior (PEB) (Kollmuss & Agyeman, 2002), especially as measured by American energy consumption—with only 5% of the world’s population, the U.S. consumes more than a quarter of the world’s oil, coal, and natural gas resources, contributing more than a quarter of the world’s total carbon dioxide emissions (Scherr, 2005). That China and India seek to emulate America’s unsustainable lifestyle only adds to the urgency of conservation organizations’ efforts to encourage the American public to replace their environmentally damaging behaviors with PEBs.
Social and environmental psychologists have explored numerous theories to explain the gap between America’s attitude and behavior toward the environment and to offer ideas for how to bridge it. Studies have examined the connection between values and PEB, including the ability to predict future PEB based on values (Schultz & Zelezny, 2003); the negative fallout of the environmental movement’s “doomsday” approach to communicating with the public, coupled with its emphasis on sacrifice (Kaplan, 2002; Sale, 1993); and the importance of message framing to effectively promote an environmental social norm (Cialdini, 2003).
Values, according to Schwartz (1992), are important life goals, standards that serve as guiding principles in a person’s life (Schultz & Zelezny, 2003). In his Model of Human Values, Schwartz (1994b) classified 56 values according to two core dimensions: self-transcendence versus self-enhancement and openness to change versus conservatism. He defined self-transcendence in terms of life goals that “transcend the individual and instead promote the interests of other persons and the natural world.” Examples include being broad-minded, helpful, honest, forgiving, and loyal. This is in contrast to self-enhancement goals such as social power, authority, wealth, success, ambition, and influence, which he defined as promoting one’s own interests “regardless of others’ interests.” The 56-item measure is now known as the Schwartz Human Values Measure and is considered by many to be universal and applicable across cultures, providing “a broad theoretical framework for understanding human values at the level of the individual” (Schultz & Zelezny, 2003).
In his data from 86 independent samples obtained from 38 countries, with approximately 44,000 participants, Schwartz found that the U.S. ranked high on self-endorsing values, particularly those concerned with mastery and competition (i.e., successful, capable, independent, choosing one’s own goals). On values of self-transcendence, which includes core items such as “protecting the environment” and “unity with nature,” the US ranked relatively low (Schwartz, 1992).
In their survey of numerous studies that have used the Schwartz Human Values Measure to examine environmentalism in terms of how it correlates with self-transcendent/self-enhancing values, Schultz and Zelezny (2003) discovered that there seems to be a strong, positive correlation between self-transcendent (ST) values and environmental concern. They also found a negative correlation between such concern and those who have more self-enhancement (SE) values. The same relationships were found in studies examining correlations between these value sets and PEB (as opposed to environmental values or attitudes) (Schultz & Zelezny, 2003).
Schultz and Zelezny then explored two questions: why do people care about the environment and why do SE values correlate negatively with environmental behavior? To answer the first question, they conducted a qualitative assessment of the types of concerns people have about the environment and asked them why they care. They coded these responses and found three different values-based attitude groups: egoistic concerns focused on the self and self-oriented goals (e.g., health, quality of life, prosperity, convenience); social-altruistic concerns focused on other people (e.g., children, family, community, humanity); and biospheric concerns focused on the well-being of living things (e.g., plants, animals, trees). So, for example, if the environmental topic was “water pollution,” the reasons for caring that were cited included “I don’t want to drink dirty water” (egoistic), “I don’t want my children to drink polluted water” (social-altruistic), or “Every living creature on the planet needs clean water to survive” (biospheric). The research also revealed, not surprisingly, positive correlations between ST values and biospheric concerns and between SE values and egoistic concerns. The correlation between PEB and SE values was zero or negative (Schultz 2001, 2002).
To explore the second question, Schultz and Zelezny examined the way in which environmentalism has been marketed in the U.S. as “a backlash against the mainstream American lifestyle (materialism, pursuit of personal wealth, self-interest).” They found that traditional environmental messages were framed in such a way as to appeal only to those with ST values because they stressed self-sacrifice in order to protect the environment. Indeed, even research on environmental behavior followed suit, looking to models of altruism to explain why it even existed, tacit acknowledgement of the inherent cost involved in PEBs, at least as proposed by traditional environmental messages (Schultz & Zelezny, 2003).
The problem with the environmental movement’s altruistic approach to encouraging PEB is that “it attempts to put aside the issue of gain, of self interest, in human behavior” (Kaplan, 2002). Emphasizing actions that carry a cost to self in terms of convenience or loss of comfort or hard-won security, all for the benefit of a vague goal or nebulous others, is a tough sell to a person with mostly SE values. Altruistic approaches are also hampered by the large scale of most environmental problems and by the fact that most Americans have yet to feel their effect. Theoretically speaking, this makes it easy for those with SE values to shrug off messages framed in biospheric (i.e., ST) terms.
To justify its requests for sacrifice, the environmental movement inadvertently set up another negative feedback loop, painting such dire pictures of the state of the world that it created a sense of helplessness (as well as guilt, fear, shame, and resentment) in the American public. This is made clear in nationwide public opinion surveys, in which a large majority of Americans say they believe that problems like global warming can only be solved by government or industry, not by individuals (Gallup, 2006; Dunlap & Saad, 2001).
Recent research by Cialdini (APA Online, 2006) supports the importance of framing in messages aimed at shaping PEB. Working with two graduate students, Cialdini helped a hotel create an effective program to encourage lodgers to use their towels more than once. The researchers randomly assigned cards with one of five different messages to 260 guest rooms. The message that described a social norm, “Join your fellow citizens in helping to save the environment,” proved the most successful, with 41% of the guests who received this message opting to re-use their towels. The next best messages stressed environmental protection and the benefit for future generations (31%). The least successful was the rather vague message “Help the hotel save energy.” The results were in keeping with the social psychology theory that, when in a new situation, people take their cue from what seems to be other people’s normal behavior. It also underscored that it is important for normative messages to emphasize that the PEB is normal behavior and beware of inadvertently making it seem like everyone is engaging in the unsustainable behavior (Cialdini, 2003).
Although they pursue different theories, most researchers agree on one point: that any attempt to move the American public in a greener direction must take into account the underlying values of the American culture. This is not the same as advocating that Americans change their values. Rather, it is an acknowledgement that, in trying to encourage the development of more PEBs, conservation organizations must work with people’s core values as they are now. In terms of global warming, there simply isn’t time to create fundamental change at the values level. The hope and intent is that when an individual performs a new PEB it will result in a values change and become permanent a la Bem’s (1972) self-perception theory, in which people observe their behavior and then infer internal reasons for it (Cialdini, 2000).
Taken together, Cialdini’s research on message framing, within the context of Schultz and Zelezny’s work on SE-ST values, suggest that framing PEB in terms of SE values should elicit a greater response than if those same PEBs were couched in traditional environmental language. For the current research, this led to the following hypotheses: (1) Environmental messages will elicit a more positive response overall if framed to appeal to those with SE rather than ST values because (2) although SEs and STs would respond differently to different environmental messages, showing a preference for those that reflect their value set, all but the most self-sacrificing STs would also respond positively to messages framed in SE values, resulting in a higher overall response rate to SE messages. And the inverse would also be true, (3) overall SE response to environmental messages should be lower than overall ST response. In other words, because of their predisposition toward the environment, ST supporters of any PEB would always outnumber SEs, assuming equal representation of both types in a population.
Method
To explore the effect of framing on pro-environmental behavior (PEB), two surveys were developed. Although the surveys inquired about the same 14 PEBs, the ST version ascribed traditional biospheric motivations for them, whereas the SE version ascribed egoistic motivations. The surveys were randomly distributed to Columbia University students who agreed to participate in a “consumer survey for a research project conducted in conjunction with EarthDay 2006” over a two-day period in April 2006.
Participants
Approximately 120 students were recruited, with 91 consenting to participate. Of these, 6 were dropped because of incomplete or incorrectly filled out surveys, so the final number of participants was 85. More than 55% of the participants were post-baccalaureate or graduate students, 22% were juniors or seniors, and 20% were sophomores or freshmen. Ages ranged from 18 to 43, with a mean of 24.5. The largest group of participants was pursuing a Masters of Business (20%) degree, followed by those in pre-med (7%), psychology (6%), and architecture (6%) programs. Gender split as follows: 48 women and 37 males. More than 50% reported their political affiliation as Democrat, followed by Independent (12.5%), Republican (10%), undecided (7%), Green (3.5%), and apolitical (3%). Other political affiliations included Libertarian and parties from other countries such as the U.K. and India. Religious affiliation was also surveyed, with the most frequent being agnostic (23%), followed by Catholic (22%), Jewish (12.5%), Protestant (11.5%), atheist (8%), and Hindu (3.5%).
Students were recruited in mostly outdoor settings, but also in the business school library and cafeteria. The experimenter greeted prospective participants, explained she was conducting a survey as per above, and asked if the participants had “5 to 7 minutes to answer 15 questions.” She also mentioned that participation would be compensated “either in the form of chocolate or a $1 bill.” Upon agreement, the experimenter provided a survey from the pre-mixed stack in her arms and also offered the participant a pen. Stating that she would “circle back around in 5 minutes or so,” the experimenter would try to find additional participants while keeping an eye on those filling in surveys in case they had questions and to ensure she was available to collect the surveys as soon as they were completed.
Design and Materials
A 2 (SE values vs ST values) x 2 (SE survey vs ST survey) between subjects design was used. Each survey included 14 questions regarding the same key PEBs such as buying organic food, taking public transportation, eating a vegetarian diet, voting for higher emissions standards, etc. Participants indicated their agreement or disagreement with the statement using a 5-point scale that ranged from strongly disagree (1 point) to strongly agree (5 points). The two sets of questions are included in Table 1. Surveys were scored by averaging each participant’s ratings across all 14 PEBs and then coded with either a 1 to indicate they took the SE survey or a 2 to indicate they took the ST survey.
To determine whether participants had SE or ST values, the survey included a modified version of the Schwartz Human Values Model. Instead of using all 56 values, the survey included only those 12 pertaining to the SE-ST continuum. Six values corresponded to SE values, including wealth, authority, influence, and choosing one’s own goals, and six corresponded to ST values, including unity with nature, honesty, and protecting the environment. Participants were asked to rank the 12 values in terms of “how important they are as guiding principles in your life,” with 1 as least important and 12 as most important. To detect a framing effect for SEs, it first had to be established which participants were truly SEs. Of the 85, 42 scored as SEs and 43 scored as STs, with 45 (23 SEs and 22 STs) taking the SE survey and 40 (19 SEs and 21 STs) taking the ST version. The mean ranking was found for each participant by subtracting their SE rankings from their ST rankings, resulting in either a positive (ST) or negative (SE) score. These scores were then recoded into new values, either 0 for SEs or 1 for STs.
To take into account the possibility of a gap between pro-environmental attitudes and behavior, as documented by Kollmuss and Agyeman (2002), both surveys were written in a conditional, future-oriented tense, i.e., “I would be more likely to do x if y.” The y statement contained the ST/SE variable, with the x holding the PEB in question constant between surveys. This tense was also chosen as a result of a pilot survey, which offended some subjects with its absolute statements (“I consider myself an environmentalist”). Unaware of the framing hypothesis, some pilot subjects considered the ST framing of the first set of pilot questions as indicative of experimenter bias and reported strong negative feelings such as guilt and anger. Other pilot subjects expressed concern with the terminology; for instance, the term organic was considered offensive or vague by some. Given the hypothesis for this study, it was not appropriate to remove or overly explain such “loaded” terminology. Instead, the experimenter added a 15th question to gauge whether taking either survey caused a negative emotional response in the participant. The question, “Did this survey make you feel anything unpleasant?” was followed by a list of possible negative emotions including anger, sadness, guilt, shame, fear, anxiety, helplessness, boredom, loneliness, confusion, and resentment. Participants could check off as many as applied. This question was identical on both surveys.
After the participants completed their surveys, they were thanked for their time and offered a small chocolate bar or $1 as compensation for participating.
Results
It will be recalled that two main effects of framing were predicted. The first looked at how people scored on the value inventory, collapsed across survey type. It was predicted that the mean ST response to all questions would be higher than the mean SE response. The second prediction concerned the combined response to each survey, collapsed across people’s score on the inventory. It was predicted that the mean general response to the SE survey would be higher than the mean general response to the ST survey. In addition, an interaction between value score and survey type was predicted as well, resulting in the mean SE response to the SE survey being significantly higher than mean SE response to the ST survey. And conversely, it was predicted that the mean ST response to the ST survey would be significantly higher than the mean ST response to the SE survey. None of the predictions were supported by the data; the analyses did not provide strong enough evidence to support the framing hypothesis.