Local support for conservation is associated with perceptions of good governance, social impacts, and ecological effectiveness

Local support is important for the longevity of conservation initiatives. The literature suggests that perceptions of ecological effectiveness, social impacts, and good governance will influence levels of local support for conservation. This paper examines these relationships using data from a survey of small‐scale fishermen in 11 marine protected areas from six countries in the Mediterranean Sea. The survey queried small‐scale fishermen regarding perceptions and support for conservation. We constructed composite scores for three categories of perceptions—ecological effectiveness, social impacts, and good governance—and tested the relationship with levels of support using ordinal regression models. While all three factors were positively correlated with support for conservation, perceptions of good governance and social impacts were stronger predictors of increasing support. These findings suggest that employing good governance processes and managing social impacts may be more important than ecological effectiveness for maintaining local support for conservation.

Much of the previous literature examining perceptions of conservation uses qualitative methods and is based on individual case studies. The current study aimed to build on the solid foundation and insights provided by these previous research efforts by (1) developing quantitative measures for perceptions of ecological effectiveness, social impacts, and good governance of conservation and (2) examining the relationship between these measures and levels of support for conservation. Using data collected from a survey of small-scale fishermen in 11 marine protected areas (MPAs) across six countries in the Mediterranean Sea, this paper examines how perceptions of ecological effectiveness, social impacts, and good governance are associated with small-scale fishermen's levels of support for MPAs.

Study area
The Mediterranean Sea, the largest inland sea in the world, is a highly diverse marine ecosystem and biodiversity hotspot (Coll et al., 2010). Mediterranean marine ecosystems are also among the most impacted by human development in the world (Halpern et al., 2008). This means that ecological and fisheries values are threatened by a variety of activities, including overfishing, shipping, tourism, and coastal development. There are an estimated 92,700 fishing vessels in the Mediterranean and Black Sea, employing ¼ million people, and contributing $3.09 billion in landed value (FAO, 2016). Smallscale fisheries account for 80% of the fleet, 60% of those employed, and produce 20% of landed value (FAO, 2016). As much as ∼85% of fish stocks are harvested at biologically unsustainable levels (FAO, 2016). One common management tool increasingly used across jurisdictions is MPAs. In 2016, there were a total of 1,231 MPAs covering 7.14% of the Mediterranean Sea (MedPAN, 2016). Only 76 MPAs include one or more fully protected areas, with 50% of these areas being less than 1 km 2 (Di Franco et al., 2018). As a result, only 0.04% of the area is covered by fully protected MPAs (PISCO & UNS, 2016).
We investigated small-scale fishing communities operating inside or close to 11 MPAs in six countries ( Figure 1). The MPAs were established between 1988 and 2003, ranged in size from 90 to 76,000 ha, and were between 2.7% and 100% no take (see Table 1). Key informants (e.g., MPA managers or community leaders) in each area estimated that there were between 5 and 40 small-scale fishers living in communities near or within the MPAs.

Survey design and administration
We designed and administered a quantitative survey to smallscale fishermen in communities within or near the MPAs. The  survey contained questions related to (a) basic demographic information and characteristics of individual small-scale fishermen (see Supporting Information Materials- Table S1); (b) perceptions of ecological effectiveness, social impacts, governance, and management; and (c) levels of support of smallscale fishermen for the MPA (see Supporting Information Materials- Table S2). The survey was designed by the project team, shared with project partners for feedback, pretested, finalized, and translated. Small-scale fishermen were informed about the purpose of the survey and the intended use of data, as well as how survey data from the project would be kept both anonymous and confidential, prior to being asked for verbal consent and proceeding with the survey. On average, we surveyed 69.5% (min-max = 34.2-100%) of small-scale fishermen in 11 communities (Table 1). Surveys were conducted on paper, and then returned to the project team for data entry and analysis.

Perceptions indicators and composite scores development
This paper focuses on survey questions related to small-scale fishermen's perceptions of ecological effectiveness, social impacts, and good governance as well as levels of support of small-scale fishermen for conservation (see Table 2 and Supporting Information Materials- Table S2). For each topic, we defined one or multiple items and developed indicators. Perceptions of ecological effectiveness were measured with two indicators of perceived impacts on fish abundance and habitat quality (Christie, 2005;Leleu et al., 2012). For social impacts, we developed a series of indicators that were of interest to project partners related to the categories of human well-being (Biedenweg et al., 2016;Kaplan-Hallam & Bennett 2017). In particular, we focused on perceived impacts of the MPA on individual income (economic), livelihoods (economic), food security (health), knowledge and education (social), community well-being (social), connections to nature (cultural), and fairness of distribution of impacts and benefits (economic). Finally, the survey contained indicators focused on the normative assessment of good governance (Bennett & Satterfield 2018;Borrini-Feyerabend & Hill 2015;Lockwood, 2010) including indicators related to recognition, transparency, accountability, communication, participation, consultation and consent, conflict management, trust, rule of law, and legitimacy. The indicator for recognition was constructed using four items related to the extent to which small-scale fishermen's rights, livelihoods, traditional knowledge, and culture were considered in management (Zafra-Calvo et al., 2017). Similarly, the legitimacy score was constructed using the combined means of two items related to satisfaction with decision-making processes and management actions. Support for conservation T A B L E 2 Overview of perceptions categories and associated survey items (see Supporting Information Materials- was examined with a single question regarding the level of support of small-scale fishermen for the MPA. We constructed three composite scores to represent perceptions of ecological effectiveness, social impacts, and good governance by combining items related to that topic (see Table 2). To account for different scales among the indicators, we first normalized each individual indicator to a scale of 0-2 to enable comparison. Prior to combining items, we assessed the internal coherence of the indicators in each composite score using Cronbach's alpha (>0.7 in all cases). Finally, indicators within each group were summed and then normalized on scale of 0-10 to obtain the final composite scores for ecological effectiveness, social impacts, and good governance. If any indicator was missing, no score could be calculated and that survey was omitted.

Analysis
Descriptive tables were used to examine demographics and characteristics of small-scale fishermen, levels of support, perceptions indicators, and composite scores. Univariate associations with level of support were assessed using chi-squared tests (categorical variables) and analysis of variance (continuous variables).
For further analysis, we limited our sample to only survey responses with complete demographics and composite scores (n = 102, ∼69% of the total questionnaires). To explore relationships between individual perceptions indicators and level of support, we recoded the indicators and support (five levels) as continuous variables and used Spearman's correlations. Last, we used ordinal regression models to predict support for the MPA as an ordinal outcome with each composite score as a predictor after controlling for the effects of other confounders. Models were adjusted for individual socioeconomic characteristics (age, education, number of years in the village, number of people in the household, relative wealth) and measures of the individuals' dependence on fishing (number of fishing gears, proportion of income from fishing, days of the week fish is eaten in the household). All analyses were conducted in SAS 9.4 and R-3.5.1 (R Core Team 2018; SAS 2018).

Survey sample
All 149 survey respondents were male small-scale fishermen (Supporting Information Materials- Table S3). Almost half (49.8%) were 50 years or older. The majority of those surveyed had completed only middle (43.2%) or elementary (29.1%) school. Households often had two (30.6%), three (18.4%), or four (33.3%) people living in them. Most smallscale fishermen came from the local village (>80%) and the mean time living in the village was 43.3 years. Among those surveyed, over half (68.6%) reported difficulty with making enough to have a good quality of life in their village (i.e., relative wealth). Specifically, 24.5% felt that "it can be challenging" and 44.1% said they are "just barely able" to make enough to live a good quality of life. They participated in an average of 2.8 different fisheries (median = 3; min-max = 1-6) and often had supplementary livelihoods (46.7%).

Level of support for marine protected areas
Overall, the small-scale fishers were relatively supportive of the MPAs (Figure 2) with 5.4% expressing strong opposition to the MPA (n = 8), 12.2% somewhat in opposition to the MPA (n = 18), 22.2% neutral toward the MPA (n = 33), 31% somewhat in support of the MPA (n = 46), and 29% voicing strong support for the MPA (n = 43). There were statistically significant differences in levels of support by MPAs (Figure 2 and Supporting Information Materials- Table S5). Levels of support was also associated with country (P < 0.001), number of people living in the household (P < 0.01), as well as for "relative wealth," "number of livelihoods," "total number of different fisheries and gears," and "number of days the week that household eats fish or seafood" (P < 0.05).

Perceptions of ecological effectiveness, social impacts, and good governance
Overall, the perceptions of small-scale fishers of individual indicators related to ecological effectiveness, social impacts, and good governance were relatively positive (Supporting Information Materials- Table S4). For ecological effectiveness, the vast majority felt that the impact was either neutral or positive for both fish abundance (neutral = 38.4%; increase = 43.8%) and habitat quality (neutral = 31.7; improve = 58.6). Most of the social impact indicatorsincome, livelihoods, food security, knowledge and education, community social well-being, and connections to naturewere skewed toward the positive. The one exception was "fairness of impacts," which was slightly skewed toward the negative. Responses to questions related to indicators for perceptions of good governance varied the most with a significant spread from negative to positive responses.

Composite metrics for ecological effectiveness, social impacts, and good governance
The composite scores (n = 102 respondents) for each of the three categories were as follows (scale 0-10): (a) the perceptions of ecological effectiveness composite had a mean score of 7.6 (median = 8.0), (b) the perceptions of social impacts composite had a mean score of 6.0 (median = 5.9), and (c) the good governance composite had a mean score of 5.8 (median = 5.6) ( Table 3). The composite scores for ecological effectiveness, social impacts, and good governance differed significantly by MPAs (Supporting Information Materials- Figure S1).
The distribution of scores for all three composites aligned with an increase in levels of support for the MPA (Figure 4). Using ordinal regression (Table 4), there is a statistically significant (P < 0.001) effect of increasing support for the MPA 3.9 (4.0) 5.9 (6.1) 5.1 (4.9) 6.4 (6.6) 7.1 (7.0) 6.0 (5.9) Min-Max 2.0-5.8 2.2-8.3 1.3-8.3 2.9-9.3 4.6-9.3 1.3-9.3 Good governance combined score (0-10) based on increasing ecological effectiveness score (OR = 1. 26), social impacts score (OR = 1.75), or governance score (OR = 1.80) (ORs reported for a 1 unit change in score) (all P < 0.001) in unadjusted models. These represent roughly a 25% increased probability of fishers' increasing support for the MPA with each unit increase in the ecological impacts score, a 75% increased probability for the social impacts score, and a 80% increased probability for the governance

DISCUSSION
This research provides a multi-sited study of how people's perceptions can be monitored using a quantitative survey to understand the social impacts of conservation (Jones et al., 2017;Kaplan-Hallam & Bennett 2017;McNeill, Clifton, & Harvey, 2018;Svensson, Rodwell, & Attrill, 2010), ecological effectiveness (Leleu et al., 2012;Webb et al., 2004;Yasué et al., 2010), and conservation governance (Bennett & Satterfield 2018;Borrini-Feyerabend & Hill 2015;Lock-wood, 2010). This study extends previous research through employing quantitative methods and analysis to examine how perceptions are associated with support for conservation. Overall, results show that small-scale fishermen are mostly supportive of MPAs in the northern Mediterranean and held generally positive perceptions of the ecological effectiveness and social impacts of MPAs but more varied perceptions of governance. We found agreement with the general hypothesis that positive perceptions are associated with increased support for MPAs (Bennett, 2016). While all three factors were positively correlated with levels of support for conservation, perceptions of good governance and social impacts were stronger predictors of increasing support. These findings suggest that employing good governance processes and managing social impacts may be more important than ecological effectiveness for maintaining local support for conservation.
The results presented here have a number of important implications for conservation policy makers and practitioners. First, these results confirm the worth of understanding people's perceptions (Bennett, 2016). Monitoring people's perceptions can help identify management actions-for example, relationship building, conducting outreach activities, communicating science, and increasing transparencythat will improve perceptions and increase support. Second, the analysis of individual indicators against levels of support points to specific factors related to conservation that may be more important determinants of support-these include transparency, conflict management, recognition, trust, and knowledge and education. Third, these results suggest that conservation practitioners and managers need to be attentive to the quality of governance (Bennett & Satterfield, 2018;Borrini-Feyerabend & Hill 2015;Lockwood, 2010) and the social impacts of conservation (Jones et al., 2017;Kaplan-Hallam & Bennett 2017;McNeill et al., 2018;Svensson et al., 2010)-just as they need to attend to the ecological effectiveness of conservation. Fostering support through greater attention to good governance and social impacts may also have knock-on benefits through increasing compliance, decreasing enforcement costs, and improving ecological outcomes (Bergseth, Gurney, Barnes, Arias, & Cinner, 2018;Rohe et al., 2017).
We recognize some limitations of this work. First, we caution here that this study may have limited generalizability to other settings as it focuses on one context, stakeholder group, and timeframe. Second, the overall sample was relatively small (n = 149) and was further restricted for regression models. Third, items in the composite scores were summed but not weighted.
Thus, we encourage improving upon the approach taken here in future research. In particular, we recommend refining the indicators and composite scores, weighting items within composite scores, sampling a broader group of stakeholders  Survey data from 102 small-scale fishermen with complete responses in the survey. 1 OR = odds ratio, odds of increasing support from one category to the next, for 1 unit increase in associated score. 2 All models adjusted for respondents' age (categorical), education level (categorical), household size (continuous), number of years living in the village (continuous), number of days/week household eats fish (continuous), relative wealth (categorical), number of fishing gears owned (continuous), and the proportion of income from fishing (categorical). There was no evidence of lack of model fit using both Hosmer-Lemeshow and Lipsitz tests for the adjusted models. Pseudo-R 2 (McFadden's) were 0.13 (ecological), 0.18 (social), and 0.22 (governance).
to enable comparative analysis, and engaging with both subjective and objective measures. Finally, similar approaches to hypothesis testing would be beneficial for research on other topics, such as satisfaction, social license, attitudes, commitment, and compliance.

CONCLUSION
The support of local people is important for the longevity and success of conservation. This study demonstrates that small-scale fishermen's perceptions of ecological effectiveness, social impacts, and good governance were all positively correlated with levels of support of small-scale fishermen for MPAs. However, perceptions of good governance and social impacts were stronger predictors of increasing support. These results suggest that conservation practitioners need to be attentive to all three factors-ecological effectiveness, social impacts, and good governance-during the implementation and ongoing management of conservation initiatives. Greater attention to the means of conservation could increase the likelihood that conservation initiatives will be successful in the end.