This article was submitted to Science and Environmental Communication, a section of the journal Frontiers in Communication
This is an open-access article distributed under the terms of the
Two natural field experiments were implemented to examine the influence of framing effects on environmental behavior. The first study examined plastic bag use at a convenience store using low cost nudges: an informational message on a sign and interpersonal communication at check-out. We employed a 3 × 2 treatment design (positive message vs. negative message vs. no sign; each paired with both asking for bag vs. not asking for bag) and report the observed plastic bag use behavior. A second study was conducted using a pre-snorkel briefing with two message frames–positive, negative, and no briefing (control). Environmentally damaging snorkel behavior was anonymously observed and recorded. Both experiments show the same general result: a significant and positive difference between subjects that were exposed to an intervention compared to those in the control conditions. However, we do not find significant differences in observed environmental behavior between the negative and positive framing.
香京julia种子在线播放
Ecosystems and natural resources are humankind’s basis of life, but are generally facing pollution, degradation and overexploitation (
Large-scale changes of human behavior are crucial to diminish the underlying drivers of environmental problems and to better structure institutions (i.e., rules and norms) that provide the incentives for individuals to act in ways that do not undermine pro-environmental goals of society as a whole (
Nudges are low-cost interventions that influence decision-making without limiting freedom of choice and have been tested in the environmental realm of electricity and water saving, reduced meat consumption, recycling, and decreasing private car transportation (
The purpose of this paper is to observe real behavior changes using natural field experiments with different types of pro-environmental communication framing interventions, and to test the effectiveness of those framing intervention nudges. Due to the disparate and largely inconclusive literature on the topic in the environmental realm, this study aims to add clarity with findings from two experiments on different environmental problems 1) plastic pollution and 2) coral reef degradation on Gili Trawangan, Indonesia. The findings of these two studies contribute to the framing and environmental behavior literature and can inform design of future research, but also demonstrate practical intervention strategies for influencing human behavior in relation to local environmental management.
According to the Theory of Planned Behavior, pro-environmental behavior is thought to be influenced by one’s environmental knowledge, attitudes, social pressures, values and beliefs (
Although a multitude of studies have used interpersonal communication to examine environmental behaviors (for a review see
Nevertheless, a cross-cultural phenomenon called ‘knowledge-action’ gap occurs. Hundreds of studies have failed to explain the gap between environmental knowledge held by individuals and the resulting pro-environmental behavior they exhibit (
The framing effect, resulting from message or valence framing, is described as a cognitive bias emerging from the way information is communicated or presented (
Research has shown that communicating a descriptive norm that gives people cues about expected behavior (or inappropriate behavior) through written information can induce conformity (
For example, in a virtual reality-based intervention, it was tested whether gain or loss framing was more influential in provoking respondents to be willing to donate money and time for biodiversity conservation organizations (
Surprisingly, to the authors’ knowledge, only one pro-environmental framing experiment has been implemented as a natural field experiment using observable behavior (
The Gili islands–Gili Trawangan, Gili Meno and Gili Air–are small tropical islands off the northwest coast of Lombok, located in the strait between Lombok and Bali, Indonesia. Gili Trawangan is the largest of the three islands, at approximately six square kilometers. The population consists of around 2000 people, mainly Indonesians, but also western expatriates (
On Gili Trawangan, waste collection volumes range from 10–15 tons per day during high season and 3–5 tons per day in low season of which 42% is non-organic waste (
Regarding the marine environmental problems, coral reef degradation is one of the most severe ones. As a result, the local non-governmental organization Gili Eco Trust was established in 2002 with the aim to conserve the coral reefs. Scuba diving and snorkeling can immensely deteriorate the health of coral reefs, mainly from physically touching the reef whether accidently or purposefully (
As snorkeling and diving are the main tourist activities on Gili Trawangan, exploring management interventions that could minimize their impacts should be considered. Already in 1997,
The amount of plastic pollution ending up in the world’s oceans is increasing (
The study location was a convenience store called Coco Express, a retail chain. This was chosen because it is one of the two most common and busiest shops on Gili Trawangan. The location was at the south end of the main beachfront strip frequented heavily by tourists. The sitting area in front of the shop (bottom-right in
A 3 × 2 full factorial between-subject’s experimental design was implemented to observe differences in behavior. The first factor is message framing with three levels: positive, negative and a control with no sign. The second factor refers to whether the shop assistant asks the customer if he/she needs a plastic bag (asking) (“Do you need a plastic bag?”) or places the purchased items in a plastic bag without asking unless the customer actively refuses a plastic bag (not asking). A total of 721 observations were included for analysis (see
Sample size by treatment allocation.
Positive framing | Negative framing | Control (no poster) | Total | |
---|---|---|---|---|
Asking | Treatment1 (T1) N = 124 | Treatment2 (T2) N = 118 | Treatment3 (T3) N = 124 | N = 366 |
Not asking | Treatment4 (T4) N = 119 | Treatment5 (T5) N = 114 | Treatment6 (T6) N = 122 | N = 355 |
Total | N = 243 | N = 232 | N = 246 | N = 721 |
The treatments were displayed as an informational sign on the check-out counter of the shop (
Positively
A structured survey was created to collect personal data and supplement the interpretation of the observed behavioral intervention experiments on each subject as they exited the shop, including questions on socio-demographics, income, activities on the island, perceptions on plastic pollution on Gili Trawangan, pro-environmental behavior intentions and actions already performed on the island (see
From 2:00–8:00 pm was chosen for observations as this was the busiest time. To randomize observation times and days, each treatment was set to a period of 2 h with 5 min in between. The random assignment of treatments to observation times was done with R Studio (Package: Base) using the function sort() and order(). The target minimum sample size per treatment was 100. Interventions and observations took place on 19 days between November 14, 2018 and December 17, 2018.
Before a treatment was implemented, shop assistants were briefed about their verbal cues (asking/not asking) using Google Translate to ensure proper understanding and adherence to the treatment. The customers were observed through the glass window of the shop and the following information was recorded: gender (male or female), single item purchased, number of small items purchased (defined as items that fit in a trouser pocket/small hand bag), number of large items purchased (items that do not fit into a trouser pocket/small hand bag), and whether the customer took a plastic bag. Single items were items that are consumed right after the purchase, for example ice cream, hot beverages, beverages that were opened in the shop or right after leaving the shop. The customers that bought a single item were recorded (
After the customer’s behavior had been observed and upon exiting the store they were approached and asked whether they would have time for a three-minute survey about plastic pollution on Gili Trawangan. When a couple or a group of persons bought items together, the behavior of the one who paid was observed as he/she was in the position to make a decision about taking a plastic bag or refusing it. Consequently, only this person was asked to fill in the survey.
Of the 721 people observed, 319 were women and 402 were men. On average, 3.32 ± 1.85 items were purchased per subject, whereby the number of small purchased items is 0.43 ± 0.73 and the number of large purchased items is 2.89 ± 1.77. From the total sample, an average of 46.60% of the subjects actively refused to take a plastic bag.
The largest proportion of people who actively refused plastic bags were found under T1 (positive and asking) with 58.06% and the lowest proportion of people who refused plastic bags were exposed to T6 (no sign and not asking) with 30.03%. The percentage of actively refused plastic bags per treatment is presented in
Percentage of actively refused plastic bags per treatment including the treatment description.
Although the proportion of people refusing plastic bags was higher in the treatments with the positive framing, suggesting a trend, the two-sided proportion test revealed no significant differences between the positive and negative framing (see
Proportion test results of refused plastic bags.
Comparison | Treatments |
|
---|---|---|
…Between positive and negative framings | T1*T2 | 0.259 (0.16) |
T4*T5 | 0.457 (0.11) | |
…Between asking/not asking | T1*T4 | 0.231 (0.17) |
T2*T5 | 0.421 (0.12) | |
T3*T6 | 0.008 ** (0.36) | |
…With control for sign | T1*T3 | 0.127 (0.21) |
T2*T3 | 0.804 (0.05) | |
…With control for sign and asking/not asking | T1*T6 | <0.001 *** (0.57) |
T2*T6 | 0.003 ** (0.40) | |
T4*T6 | 0.003 ** (0.40) | |
T5*T6 | 0.044 * (0.28) |
*, **, *** indicates significance levels at the
The interaction factor (asking/not asking) only showed significant differences between treatment 3 and 6 (
The logistic stepwise regression (
Regression results for accepted plastic bags (observation data).
Time 2 (4:05–6:05 pm) | 0.145 (0.269) |
Time 3 (6:10–8:10 pm) | 0.486 (0.259) |
Treatment 2 | 0.190 (0.317) |
Treatment 3 | 0.009 (0.327) |
Treatment 4 | 0.272 (0.339) |
Treatment 5 | 0.697 * (0.352) |
Treatment 6 | 1.204 *** (0.327) |
Number of small items | 0.472 *** (0.138) |
Number of large items | 0.682 *** (0.079) |
Pseudo R-squared (Mc Fadden) | 0.164 |
No. observations | 721 |
Standard errors are reported in parentheses.
*, **, *** indicates significance levels at the
Of the 201 subjects who filled in the survey, 53.7% stated their gender as female, 44.8% as male and 1.5% as other. The age of the survey participants ranged from 17 to 71, with a median of 27. The observed behavior of survey respondents differs from the overall observed behavior: only 40.80% of the survey respondents actively refused plastic bags. Responses on statements about observation of plastic pollution on Gili Trawangan, concern about pollution, self-efficacy and contribution to it are summarized in
Survey results about plastic pollution on Gili Trawangan.
Survey statement | Agree (%) | Disagree (%) | No answer (%) |
---|---|---|---|
During my stay on Gili Trawangan I observed plastic pollution | 88.56 | 9.95 | 1.49 |
Plastic pollution is a problem on Gili Trawangan | 89.05 | 10.45 | 0.50 |
Plastic pollution on Gili Trawangan concerns me because of the environmental consequences | 94.53 | 4.47 | 1.00 |
Marine plastic pollution is a danger for both marine wildlife and humans | 98.01 | 1.49 | 1.00 |
My stay on Gili Trawangan contributes to plastic pollution | 65.67 | 33.33 | 1.00 |
I can do something against plastic pollution on Gili Trawangan | 93.53 | 4.47 | 1.00 |
Survey results suggest strong agreement that a plastic pollution problem exists and agree is negative, but also that subjects could do something about it (93%) (
Overall, all treatments were significantly more effective in discouraging the use of plastic bags compared to the control treatment (T6–no sign and not asking). Although no framing effect was detected, the treatment T1–positive and asking resulted in the largest proportion of actively refused plastic bags and the probability of people actively refusing a plastic bag is significantly higher compared to the control treatment (T6–no sign and not asking). The interaction (asking whether a plastic bag is needed) is especially important when there is no sign. However, a sign discouraging the use of plastic bags is helpful as a low-cost and low-effort approach to reduce plastic bag use, even if it serves only as a reminder to the clerk to ask each customer if they need a bag.
Despite the significant difference between treatment 5 (negative framing and not asking) and treatment 6 (no sign and not asking) resulting from the proportion test, the logistic regression indicates that both treatments cause a higher probability of people accepting a plastic bag. Consequently, when no interaction with customers is possible (i.e., through self-checkout lanes), a positively framed visual message is advisable to reduce the consumption of plastic bags.
The stepwise logistic regression results are presented in
Regression results for accepted plastic bags (survey data).
Europe and Central Asia | 0.404 (0.458) |
Latin America and Caribbean | ‒2.340* (0.949) |
Middle East and North Africa | ‒0.358 (1.413) |
North America | ‒2.174* (0.885) |
South Asia | 0.180 (1.250) |
Sub-Saharan Africa | 0.319 (2.030) |
East Asia and Pacific Islands | 0.000 |
Family income status | 0.425 (0.280) |
Accommodation cost class | 0.477* (0.231) |
Plastic pollution is a problem | ‒1.126 (0.800) |
Concern about plastic pollution | 2.420* (1.165) |
Marine plastic pollution is a danger | ‒17.520 (858.518) |
Tourists have a potential to mitigate plastic pollution | 0.615 (0.430) |
Sum of pro-environmental behavior on Gili Trawangan | ‒0.225** (0.077) |
Pseudo R-squared (Mc Fadden) | 0.239 |
No. observations | 176 |
Standard errors are reported in parentheses.
*, **, *** indicates significance levels at the
Overall, findings indicate any intervention (sign and/or asking) is more effective than none. Any sign regardless of framing, coupled with asking customers if they want a bag, can be an effective and low-cost way to reduce plastic bag use. Interestingly, no significant difference in plastic bag refusal was observed between the positive and negative framing treatments. The human interaction of asking whether a customer needs a plastic bag is particularly important when there is no sign. A sign may also serve as a reminder to the clerk to ask each customer if they need a bag.
Regarding survey results, a large majority of respondents agreed plastic pollution is observable, a problem, and they could do something about it. Nevertheless, our findings confirm that a knowledge-action and intention-action gap exists (
The high significance of the sum of pro-environmental activities reveals an important factor. The more people already behave environmentally-friendly (i.e., refuse straws, avoid reef touching, use waste bins), the higher the likelihood is that they will perform other environmentally-friendly actions, such as refusing a plastic bag. Although the previous environmentally-friendly actions were self-reported, this finding is in line with other research revealing a positive spill-over effect from one sustainable behavior to multiple others (
Interestingly, the more money people spent on their accommodation, the more likely they were to accept a single-use plastic bag. As the family’s income status did not reveal any significant differences between observations, and was, thus, eliminated from the regression model, this factor does not correlate directly to the individual’s economic situation, but rather to the importance of luxury and or convenience while on holiday. This may also be reflected in the use of plastic bags as they are an item of convenience.
Attributing relevance to the lower likelihood of refusing bags based on the consumer’s region of origin being Latin America/Caribbean and North America compared to those from East Asia/Pacific Islands should be carefully interpreted given the uneven distribution of the number of observations per world region (both regions were only represented by a small sample size n = 11 and n = 16, respectively). Nevertheless, cultural differences in the effectiveness of differently framed messages (
Coral reefs provide important public goods and ecosystem services worldwide, but are increasingly impacted by people through climate change induced coral bleaching, pollution and overuse from fishing and tourism (
However, different framings of the briefing message, and how their effectiveness may differ, have not been previously tested. For example, if whether highlighting “what to do” (e.g., always keep 2 m distance) vs. “what to avoid doing” (e.g., do not touch marine life) leads to fewer impacts. The following experiment tests the variation in effectiveness of different pre-trip briefing framings on a random sample of snorkelers on Gili Trawangan.
The study location was the northeast beach of Gili Trawangan commonly known as “Turtle Point.” Various signs indicate the location of the area where turtles can be observed on a regular basis and snorkeling gear can be rented on multiple stands along the beach. This site was chosen because most interactions between snorkelers and turtles occur here (informal interviews with Gili Eco Trust staff and local dive businesses), it is the most frequented snorkel area around Gili Trawangan accessible from land, and, consequently, receives the most reef damage by snorkelers (
The experiment comprises two treatments and a control setting. The medium for this experiment was a printed briefing sheet informing snorkelers about either a positive (careful) or negative (damaging) snorkeling behavior (see
Briefing sheets with four key notes highlighting positive (careful) snorkeling behavior (left) and negative (damaging) snorkeling behavior (right).
Positive framing | Negative framing |
---|---|
Save the reef | Don’t harm the reef |
How to be a good guest in the underwater world | How to minimize your negative impact underwater |
You can help life underwater stay colorful and beautiful by keeping your hands to yourself and your feet off the ground as much as possible to avoid touching the reef, turtles and other marine wildlife | Your presence on the reef can have disastrous outcomes for the reef in the long-run if you touch or step on the reef, the turtles and other marine wildlife |
As a good snorkeler you are always aware of where your fins and other equipment (camera, etc.) are to avoid involuntary contacts | If you are not aware of the position of your fins and other equipment (camera, etc.) you may contact the coral reef and injure it |
Good snorkeler behavior means you keep at least 1.5 m distance to turtles at all times to ensure they remain calm in their environment and can surface to breathe. Good snorkelers enjoy watching the natural behavior of turtles and if it swims away, let it go! | Bad snorkeler behavior is when you get too close to turtles (less than 1.5 m) and they become distressed and cannot surface to breathe. Don’t chase them when they swim away! |
Take your plastic trash with you and earn extra karma by picking up more. You improve the marine animals’ chances for survival and reproduction | Plastic trash is a hazard to marine wildlife. It decreases their chances of survival and successful reproduction. Pick it up! |
Snorkelers who were preparing to get into the water were approached by one of the volunteers of the local NGO Gili Eco Trust. The volunteers introduced themselves as volunteers, briefly outlined the work of the NGO (being coral reef conservation and waste treatment) and then provided information about snorkeling behavior to reduce impacts on the reef.
Following this interaction, and allowing for time for the subjects to adjust their gear and become comfortable in the water, the underwater observer watched the snorkeler(s) for 10 min. Gender, use of snorkel, fins, camera and other equipment as well as the number of people surrounding the individual snorkeler (>2 m distance to subject) during the 10-minute observation were noted. The position in the water was recorded every 60 s as the average position over the 10-minute period can act as a proxy for the snorkelers’ level of experience. The snorkelers’ proximity to the reef was also assessed every 60 s given the obvious correlation between reef proximity and contact. Several different kinds of impacts on the reef and marine life were observed, including type of contact (kick, touch, stand, etc.), type of coral (branching, foliaceous, massive, soft, table) or sponge, body part/equipment (foot, hand, leg, fin, camera), type of impact (sediment suspension, single abrasion, multiple abrasion, breakage), voluntary vs. involuntary impact (voluntary included standing on the reef and grabbing the reef as opposed to accidental brushing against the reef and accidental fin kicks), coming closer than 1.5 m distance to a turtle, touching, obstructing and feeding (see
During data analysis, the number of impacts was included, as well as, a calculated impact score which was created to account for the different severity of impacts on marine wildlife. These were calculated by the type of impact with the coral reef (sediment suspension = 1, single abrasion = 2, multiple abrasion = 3, breakage = 4), the type of impact with a turtle (chasing, distance closer than 1.5 m, obstructing = 2, touching/feeding = 4) and the intention of the subject (impact was voluntary = 1, impact was involuntary = 0). For subjects with multiple impacts, the scores of each impact were summed up. With the ten observations of the subject’s proximity to the reef, a median was calculated. A position index was created to use as a proxy for one’s level of experience and comfortability snorkeling. For this, the observation every 60 s was assigned to a number (standing on the reef = 1, standing on sand = 2, swimming vertically = 3, floating on the water surface = 4, swimming horizontally = 5, diving down = 6) and the ten observations were averaged.
In total, 184 snorkelers were observed. As one of the observed snorkelers noticed that he was observed, this subject was removed from analysis, leaving 183 observations for data analysis. Overall, 144 impacts on marine wildlife by 73 snorkelers were noted, of which 28 impacts were involuntary. Impacts included standing and walking on the reef, touching the reef, breaking/kicking corals, suspension of sand close to corals, coming closer than 1.5 m to sea turtles, obstructing, chasing, and touching sea turtles. Although the Gili Eco Trust staff had reported feeding of turtles by snorkelers, this was not observed during the sampling days. The mean number of impacts per snorkeler is highest with 1.34 ± 1.76 per 10-minute observation period in the control setting. In the positive (0.39 ± 0.78) as well as in the negative framing (0.62 ± 1.03), the mean number of impacts was lower. The mean position index is 4.17 ± 0.50. The number of impacts is not normally distributed, and the variances are not homogeneous (all:
Concerning the weighted impact score, similar results were found. The mean impact score is highest in the control (3.69 ± 5.47), and lower in the positive (0.97 ± 2.07) and the negative treatment (1.42 ± 2.44). Likewise, the impact scores are not normally distributed, and the variances are not homogeneous (all:
The calculated impact score per treatment including the
Regarding the Zero-inflated Poisson (ZIP) regression with the number of impacts and the Zero-inflated negative-binomial (ZINB) regression with the impact score, both indicate that the treatments, as well as a higher position index, are significantly negatively related with the number of impacts and the impact score (see
Regression results for number of impacts (ZIP) and impact score (ZINB).
Count model with number of impacts | Count model with impact score | |
---|---|---|
Negative treatment | ‒0.544 * (0.219) | ‒0.527** (0.188) |
Positive treatment | ‒0.798 ** (0.251) | ‒0.720*** (0.207) |
Position index | ‒0.927 *** (0.146) | ‒0.783*** (0.129) |
Sex: male | 0.374 * (0.182) | – |
Logistic model with number of impacts | Logistic model with impact score | |
Median proximity to reef | 1.953 *** (0.461) | 1.718*** (0.332) |
Standard errors are reported in parentheses.
*, **, *** indicates significance levels at the
As the regression results reveal, neither the use of a camera, a life vest, fins or the tide and current had significant effects on the number of impacts or the impact score, as these variables were eliminated from the model due to insignificance.
The impact of snorkel related damages on the coral reef at “Turtle Point” was already reported in 1997 (
A gender effect was revealed by the ZIP of the number of impacts. Men were more likely to have a higher number of impacts compared to women. This effect was also detected by a study that had exposed snorkelers to a pre-trip media-based intervention in Puerto Rico (
The position index can be used as a proxy for the subject’s snorkel experience. Individuals that stand on the reef or sand for long time periods during their snorkel trip may feel uncomfortable swimming with snorkel and mask which indicates that they might be beginners. This highly correlates with the number of impacts and the impact score. The lower a subject’s position index, the higher the number of impacts on the marine environment.
As expected, the median proximity to the coral reef correlates with the probability of impact. This is rather unsurprizing given that the closer the proximity, the higher the probability of impact. Subsequently, the designation of snorkel areas that are easily accessible without close contact to the reef, thus, sandy entry areas and deeper reefs, would make direct damaging impacts of snorkelers less likely. For this site, clear instructions on where and how to enter and exit the water without contacting the reef would likely decrease the probability of damaging impacts. The effect of the personal interaction would have to be considered as influential, as other studies have shown observed behavior change differences between mediums where the same message was conveyed. For example, having a real person convey the message has been shown to be more effective than just a sign. Nonetheless, a sign whether positive or negative, can very likely be an effective low-cost management tool.
Overall, the findings of these two experiments are difficult to directly compare. However, they provide important and congruent insights on the effectiveness of informational messaging as an effective way to nudge pro-environmental behavior. They also both show that there are only small differences between positively and negatively framed informational messages. Both messages are effective as interventions, with both experiments indicating a trend toward a positive framing being slightly more effective but not statistically different. Future research could examine similar interventions, to confirm their effectiveness in other contexts. From a policy perspective, the most important conclusions from this study would be to have any nudging intervention rather than none, regardless of the negative or positive framing. However, our understanding of this may differ across contexts with further studies in the future.
It is clear from this study that informational messaging strategies can be implemented as effective management tools to reduce harmful environmental behavior at the individual level. A key reflection is that the interventions tested here are very low cost and easy-to-implement. These results have broader practical applications for encouraging pro-environmental behavior, particularly in situations where there is already a human interaction (i.e., entrance to a park, check-out counter, transportation, rental equipment, etc.). Regarding the situation in a grocery store, asking whether a customer needs a bag is effective in reducing the number of plastic bags used and, consequently, this should be implemented to decrease the consumption of single-use plastic bags where necessary.
Concerning snorkel tourism, we show that pre-trip briefings can be very effective. We suggest implementing a briefing upon rental of snorkel gear, or before entering the water during regular SCUBA dive briefings. Due to the finding that impacts are highest from beginners, an introductory lesson would be helpful to improve snorkeling skills of novices. In the case of Gili Trawangan, information sheets could be offered to rental stall owners and require mandatory reading of the briefing sheet before renting the equipment. Likewise, snorkel and SCUBA boat tour operators could present a briefing sheet before snorkelers enter the ocean.
Reflecting broadly, understanding how and why human behavior changes, or not, is essential for effective environmental conservation and management (
From a research perspective, and specifically in the environmental context, more experiments with real observed behavior are needed to test the effectiveness of the many proposed interventions, mediums for implementation and framings that have been suggested (
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.
The studies involving human participants were reviewed and approved by Ethics Review Board of Leibniz Centre for Tropical Marine Research: Marion Glaser and Agostino Merico. Written informed consent for participation was not required for this study in accordance with the national legislation and the institutional requirements.
Conceptualization, KN; methodology, KN; validation, MB; formal analysis, MB; investigation, MB and KN; resources, KN and SP; data curation, MB; writing—original draft preparation, MB and KN; writing—review and editing, KN, MB, and SP; visualization, MB; supervision, KN and SP; funding acquisition, KN, SP, and MB. All authors have read and agreed to the published version of the manuscript.
This study was funded by a Rapid Ocean Conservation (ROC) grant from the Waitt Foundation. Open access fees are paid by the Leibniz Centre for Tropical Marine Research (ZMT).
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
The authors would like to thank Delphine Robbe, Sian Williams and interns at the Gili EcoTrust for assisting with this research. We send extended thanks to local residents, staff and owners of the many businesses on the Gili Islands for their generosity, hospitality and willingness to participate in this study. We also thank Achim Schlüter for discussions and comments on earlier drafts of the manuscript. The development of this paper was additionally supported by the European Cooperation in Science and Technology (COST) action network on Ocean Governance, and the Land-Sea interactions working group in particular.
Date | Sex (f/m) | Camera (y/n) |
---|---|---|
Snorkel (y/n) | Fins (y/n) | Crowd |
Tide (l/m/h) | Current (l/m/h) |
Minutely observation | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 |
---|---|---|---|---|---|---|---|---|---|---|
Position in water* | ||||||||||
Proximity to reef (in m) |
* standing on the reef = 1, standing on sand = 2, swimming vertically = 3, floating on the water surface = 4, swimming horizontally = 5, diving down = 6.