FINAL REPORT Understanding Disaster Risk Reduction and Resilience in

Urban Haiti

A Collaborative Research Study with Catholic Relief Services and the Harvard

Humanitarian Initiative

Dr. Ronak Patel, MD, MPH Kelsey Gleason, MS

2

Executive Summary ................................................................................................... 3

Introduction and Background .................................................................................... 4

Methodology .............................................................................................................. 5

Results: Community Characteristics ......................................................................... 8

Results: Main Findings ............................................................................................ 10

Citations ................................................................................................................... 16

Appendix of Tables ................................................................................................. 17

Appendix of Maps ................................................................................................... 21

Appendix of Government Opportunities on Resilience .......................................... 23

TABLE OF CONTENTS  

3

This collaborative project between CRS Haiti and the Harvard Humanitarian Initiative supports individuals in the urban Haitian communities of Solino and Tesso in becoming resilient to recurring shocks and disasters. This research supports CRS in operationalizing a true “Neighborhood Approach” for reducing urban disaster risk in Solino and Tesso by empowering communities to take a leading role in their risk mitigation measures now and in the future. The primary objective of this research was to determine if baseline social cohesion in the communities of Solino and Tesso is associated with increased community participation and/or improved resilience at the beginning and end of the study period. A secondary objective was to understand how social cohesion impacts resilience and how various household and community features impact both social cohesion and resilience. Results of this study indicate that social cohesion is a very strong predictor of community resilience. Further, there is a difference in mean social cohesion and mean resilience between the two communities, with Tesso having greater social cohesion but less resilience than Solino. Additionally, certain demographic variables, some of which vary by community, are more strongly associated with both social cohesion and resilience, indicating that demographic makeup of communities can have a large influence on community resilience and social cohesion. The findings of this research suggest that by improving social cohesion in these neighborhoods, there will be a resulting increase in community resilience as well. Programs aimed at reducing disaster risk should understand the role that social cohesion and demographics of these communities play in building resilience

EXECUTIVE SUMMARY

4

Catholic Relief Services (CRS) is an international non-governmental organization supporting relief and development work in over 99 countries around the world. CRS programs assist persons on the basis of need, regardless of creed, ethnicity or nationality. CRS works closely with local community partners in order to implement its programs effectively. For over 60 years, CRS has been meeting the basic needs of the Haitian population and supporting the government of Haiti’s development priorities. Historically, CRS focused the majority of its program implementation on food security programming in the south, however more recently; health, education, agriculture, and urban planning have emerged as priority sectors. In response to the Haiti’s Disaster Risk Reduction (DRR) initiative, particularly relevant in the Haitian context, Haiti has prioritized two specific program areas within the larger spectrum of DRR activities. The CRS program “We Protect our Life” was created in response to Program Area I: Urban Disaster Risk Reduction to support interventions that mitigate geological and hydro-meteorological hazards through the Neighborhood Approach in Port-Au-Prince. The We Protect Our Life Program empowers community members in the vulnerable and marginalized neighborhoods of Solino and Tesso to address the full range of hazards and risks they face to their well being through sustainable, community-based processes. Building upon its extensive experience in community-based disaster risk reduction (DRR) in Latin America, Africa, and Asia; CRS builds the capacities of community members to lead participatory assessments and planning and to develop, manage, and sustain DRR projects that address the range of hazards in their communities. The project supports community members in becoming resilient to recurring shocks and disasters through the use of participatory assessments, planning activities and the implementation of risk mitigation measures in the community. This project will support 71,762 individuals (11,744 families in Solino and 3,361 families in Tesso) to become resilient to recurring shocks and disasters through participatory assessment and planning activities and the implementation of risk mitigation measures in their communities. Through interventions at the household level, among community clusters or blocks, within the social institutions of schools and churches, and at the neighborhood level as a whole, CRS seeks to change the dynamic of disaster planning and response from one that is primarily external to one that is community-managed. CRS will operationalize a true “Neighborhood Approach” for reducing urban disaster risk in Solino and Tesso by empowering communities to take a leading role in their risk mitigation measures now and in the future. To facilitate the success of this program, the most at risk Solino and Tesso families (more than 2,500)1 will be targeted and benefited directly through household-level assessment and planning activities and neighborhood level risk mitigation measures, while 100% of targeted neighborhoods will benefit indirectly from the predicted community wide risk2 reduction efforts, widespread DRR mapping, planning activities and awareness-raising about DRR. CRS’s partnership with the Harvard Humanitarian Initiative (HHI) to share their data serves to investigate social cohesion and community participation in the two communities of Solino and Tesso as they relate to demographics and impact on resilience. Institutional review board approval for this 1 Data source: Family census results, CRS. October 2014. 2 CRS Haiti. Family census in Solino and Tesso. Dec. 2014.

INTRODUCTION AND BACKGROUND

5

study was processed through the Partners Human Research Committee associated with Harvard Medical School for researchers in Boston, MA USA. Primary Study Objective: The primary objective of this research is to determine if baseline social cohesion in the communities of Solino and Tesso is associated with increased community participation and/or improved resilience at the beginning and end of the study period. Secondary Objective: A secondary objective is to explore social cohesion and resilience by various household and community features. For the purposes of this research, social cohesion and resilience are defined as follows: Social cohesion: The presence of features such as social organization, trust, norms, and networks that can improve the efficiency of society by facilitating coordinated actions3.

Resilience: The capacity of individuals, families, communities, systems, and institutions to anticipate, withstand and/or judiciously engage with catastrophic events and/or experiences, actively making meaning out of adversity with the goal of maintaining ‘normal’ function without fundamental loss of identity4.

Refining the survey instrument Preliminary assessment of the communities of Solino and Tesso was achieved using existing background reports provided by CRS as well as in-depth discussions with CRS staff. This assessment then guided focus group discussions with a convenience sample of the two communities representing various stakeholders. These focus groups were then reviewed with CRS staff for proper contextualization. Participants for these discussions were at least 18 years old and were selected by community leaders. In each community, two gender mixed focus groups, and one all women focus groups were conducted, resulting in a total of 6 focus groups of 6 participants each. The key informant interview was conducted jointly with the Representative to the Mayor (female) and the First Advisor for Administration for the Mayor (male). Focus group discussion questions and key informant interview questions were authorized and asked by CRS partners, ensuring the questions were culturally appropriate, thereby ensuring high quality responses (Appendix A). CRS staff members conducted focus group discussions in Creole, and two transcribers were present in each focus group, maximizing accurate transcription. Transcripts were translated first into French, then into English by CRS staff. The key informant interview was conducted in French, followed by transcription into English. Results from focus group analyses are summarized in Appendixes B-E.

3 Adapted from Putnam (1993: 167) 4 Defined by Almedom (2008)

METHODOLOGY

6

Population Survey Local enumerators administered a systematic, randomized cross-sectional questionnaire survey in the communities of Solino and Tesso from September-October 2014. The questionnaire was designed to gather information in four main areas of interest: basic demographics; social cohesion indicators; resilience indicators; and community insecurity. The final survey comprised of 67 questions (Appendix F). Oral consent was obtained from all participants prior to administration of questionnaire (Appendix G). As there was no previously established list of households, the sampling frame for this study utilized population data. Because it can not be assumed that each individual block within its community of Tesso or Solino is homogenous between one another, this research used a conservative approach, assuming that each individual block within its community of Tesso or Solino is not homogenous between one another. Therefore, we implemented a two-stage probability proportion to size (PPS) strategy in which: (1) each block was sampled proportionate to its population size; and (2) a random sample of households within each block was selected. Further, to achieve a random and representative sample of each block, this research employed a systematic random sampling technique. This sampling technique utilized an interval approach, whereby each household had an equal probability of being selected, minimizing any potential sampling bias. Maps depicting the sampling distributions are found in the Appendix. This research was conducted using a 95% confidence level and a 2.5% margin of error. As such, following the previously described sampling strategy, a total of 1,895 participants were recruited into the study: 730 from Tesso, and 1,165 from Solino. A full description of this sampling frame can be found in Appendix H. Social Cohesion Scale To accurately measure social cohesion, this study adapted the Neighborhood Cohesion Index (NCI; Buckner, 1988). This scale was developed to capture the collective-level attribute of sense of community as it exists specifically in a neighborhood context (Buckner, 1988), making it an ideal parameter for this research. One item (“A feeling of fellowship runs deep between me and other people in this neighborhood”) was excluded because the meaning was not deemed transferrable in this Haitian culture by local staff. This social cohesion scale comprised of 17 five-point Likert-type scale items (1=strongly agree, 2=agree, 3=neither agree nor disagree, 4=disagree, 5=strongly disagree) (Table 1). Each participant’s sense of community was reported as the mean score for the 17 items.

7

Table 1. The adapted Social Cohesion scale Item Statement 1 Overall, I am attracted to living in this neighborhood. 2 I feel like I belong to this neighborhood. 3 I visit my neighbors in their homes. 4 The friendships and associations I have with other people in my neighborhood mean a lot to me. 5 If I had the opportunity I would move out of this neighborhood. 6 If the people in my neighborhood were planning something, I’d think of it as something “we” were doing,

rather than “they” were doing 7 If I need advice about something I could go to someone in my neighborhood 8 I think I agree with most people in my neighborhood about what is important in life 9 I believe my neighbors would help in an emergency 10 I feel loyal to the people in my neighborhood 11 I borrow things and exchange favors with my neighbors 12 I would be willing to work together with others on something to improve my neighborhood 13 I plan to remain a resident of this neighborhood for a number of years 14 I like to think of myself as similar to the people who live in this neighborhood 15 I rarely have a neighbor over to my house to visit 16 I regularly stop and talk with people in my neighborhood 17 Living in this neighborhood gives me a sense of community The Likert response scale comprised the following five options: Strongly agree, Agree, Neither agree nor disagree, Disagree, Strongly disagree; Ranked 1-5 respectively. Community Resilience Scale Community resilience was measured using the Communities Advancing Resilience Toolkit (CART) Survey (Pfefferbaum, 2011). CART is a field-tested instrument, which assesses a community’s resilience and disaster management. This scale is most appropriate given the CRS project goal of building disaster resilience. This community resilience scale comprised of 18 five-point Likert-type scale items: 1=strongly agree, 2=agree, 3=neither agree nor disagree, 4=disagree, 5=strongly disagree) (Table 2). Each participant’s sense of resilience was reported as the mean score for the 18 items.

Table 2. The Communities Advancing Resilience Toolkit (CART) Survey Item Statement 1 People in my community are committed to the well-being of the community 2 People in my community have hope about the future 3 People in my community help each other 4 My community has the resources it needs to take care of community problems (resources include:

money, information, technology, tools, raw materials, and services) 5 My community has effective leaders 6 People in my community are able to get the services they need 7 People in my community know where to go to get things done 8 People in my community communicate with leaders who can help improve the community 9 People in my community are aware of community issues that they might want to address together 10 People in my community discuss issues so they can improve the community 11 People in my community work together to improve the community 12 My community looks at its successes and failures so it can learn from the past. 13 My community develops skills and finds resources to solve its problems and reach its goals 14 My community has priorities and sets goals for the future 15 My community tries to prevent disasters 16 My community actively prepares for future disasters 17 My community can provide emergency services during a disaster 18 My community has services and programs to help people after a disaster The Likert response scale comprised the following five options: Strongly agree, Agree, Neither agree nor disagree, Disagree, Strongly disagree; Ranked 1-5 respectively.

8

Data Analysis Data were analyzed using RStudio statistical software (Version 0.98.1087). Associations between individual demographics and community were examined at the p < 0.05 level of significance. Following a Principle Components Analysis, Factor Analysis was used to determine the presence of separate constructs within the NCI social cohesion Scale. Social cohesion and resilience scores were analyzed as the mean values for the items in each category of the separate scales. Two linear regression models were developed. Both models began with the same demographic covariates (questions 2-17 on the household questionnaire) and final model selections were made using stepwise selection. Model 1 examined the association between social cohesion (outcome) and the following covariates: Community (Solino/Tesso), Sex, Years in Community, Marital Status, Education Status, Residence Status, Religion, Number of Children, Household Size, and Primary Source of Income. Model 2 examined the association with resilience (outcome) and the following covariates: social cohesion score, Community (Solino/Tesso), Sex, Years in Current Household, Education Status, Residence Status, and Average Income per Month.

RESULTS: Community Characteristics

Survey Population Characteristics • 61.3% of respondents were from Solino; 38.7% were from Tesso • 64.5% of respondents were female • Average age of respondents was 38 years old • The majority (50.8%) of respondents earn between 0-5,000 Gourdes per month • 79% of respondents had either a primary or secondary education • 86% of respondents identified as either Catholic or Protestant • The majority (46.2%) of respondents were in a Domestic Partnership • The majority (48.3% of respondents primary source of income is a small business • The majority (44.7%) of respondents owned their residence, followed closely by

respondents who rent their property (43.5%) • The average number of income earners per household is less than one • The average number of years living in Solino or Tesso is 15 years • The average number of years lived in current residence is 9.4 years • The average household size in both communities is 5 people • On average, respondents reported 3.5 children per household

9

Disaggregated by community, the majority (9/13) of demographic variables considered in this analysis varied significantly between the two communities. This result, however, is not unexpected due to the high statistical power that this study had by design (95% confidence level and 2.5% margin of error). As such, we were able to detect even small differences between these two communities and the understanding of such differences should be interpreted with caution. A detailed table of these findings is presented in Table 3. An expanded version of this table is shown in Table 1 of the attached Appendix. Table 3. Demographic Variables with similar/different distributions between communities

Different Distributions

Solino Tesso

Religion Catholic 41.80% 44.20%

Protestant 44.90% 40.50%

Adventist 1.50% 3.80%

Jehovah’s Witness 0.60% 0.50%

Voodoo 2.50% 1.90%

Other 8.80% 9.00%

Average Monthly Income 0-5k HTG ($0-107 USD) 49.00% 53.70%

5-10k HTG ($107-214 USD) 30.20% 28.50%

10-15k HTG ($214-321 USD) 9.40% 13.00%

15k-20k HTG ($321-428 USD) 7.20% 2.30%

>20k HTG (>$428 USD) 4.20% 2.50%

Level of Education Primary 38.80% 42.70%

Secondary 38.10% 39.70%

University 6.70% 7.30%

None 16.40% 10.20%

Type of Residence Own 42.90% 47.40%

Rent 53.00% 28.40%

Squat 2.50% 22.40%

Other 1.50% 1.80%

Primary Source of Income

Small Business 49.30% 46.8

Permanent Employment 19.20% 16.10%

Short-Term Employment 17.50% 24.20%

Other 14.00% 12.90%

Yrs in Community (Mean) 12.3 10.3

Yrs in Current Residence (Mean) 11.2 7.6

Age (Mean) 13.9 12.3

Income Earners in HH (Mean) 0.87 0.91

Note: N = 1627.

Similar Distributions

Solino Tesso

Gender Male 37.20% 32.80%

Female 62.80% 67.20%

Marital Status Married 25.20% 27.20%

Single 29.10% 24.10%

Divorced 0.80% 0.50%

Domestic Partnership 45.00% 48.20%

Num. of Children (Mean) 3.6 3.4 Household Size (Mean) 5.2 5.1

Note: N = 1627.

10

Community social cohesion On the social cohesion likert scale, whereby 1 = strongly agree through 5 = strongly disagree, mean social cohesion in Solino was 2.55, and mean social cohesion in Tesso was 2.50. Because a higher social cohesion score corresponds to lower social cohesion, Tesso seemed to have has higher social cohesion as compared to Solino. But this was not statistically significant overall. In fact, after taking into consideration demographic differences between the two communities there is a highly significant difference in mean social cohesion scores between Solino and Tesso. Analyses show that compared to Solino, Tesso actually has a statistically significantly higher social cohesion if the make-up of both communities were the same.

RESULTS: Main Findings

Finding #1: While on a whole very similar, when acknowledging demographic makeup of Solino and Tesso, there is a difference in mean social cohesion between the two communities, with Tesso having greater social cohesion than Solino if the make up of both were the same.

Operational Implication 1: Given the similar social cohesion between the two communities, when planning interventions, one community should not be considered more cohesive than another. Both communities should be treated the same with regards to overall social cohesion.

Operational Implication 2: When approaching communities, understanding their make-up may help determine their amount of social cohesion. While this study was only conducted on these two communities, this finding may be generalizable to other communities in Haiti and beyond.

Operational Implication 3: This result demonstrates that the demographic make-up of communities is an important component in overall social cohesion. As community demographics change over time, social cohesion can change as well. The demographics of communities should be monitored over time as they impact social cohesion. Large swings in demographics, for any reason, will have significant implications for overall social cohesion if operational programs are tied to cohesiveness.

11

Additionally, the following variables are important indicators of social cohesion. These results are presented in further detail in Table 3 of the attached appendix. Below they are categorized according to their strength of association with social cohesion, listed in decreasing order of strength: Very Strong • Years in Community (The greater the time in the community, the greater the social cohesion) • Education Status (The lower the education, the greater the social cohesion) • Gender (Females have better social cohesion vs. Males) Strong • Community (Tesso has more social cohesion than Solino) • Religion (Catholics have better social cohesion vs. Protestants) Moderately Strong • Marital Status (Married Residents have better social cohesion vs. Single residents) • Residence Status (Property owners have greater social cohesion compared to Renters) • Religion (Catholics have better social cohesion than Jehovah Witness and Adventist) • Number of Children (The fewer children, the better social cohesion) • Primary Source of Income (Small Business workers have better social cohesion than those with

Permanent Employment and Others)

Finding #2: Certain demographic variables are more strongly associated with social cohesion. The demographic variables listed above are significantly associated with social cohesion in Solino and Tesso. Operational Implication 1: These groups may be the best initial targets of programs aimed at engaging community members for programs improving social cohesion in the two communities. Specifically, persons with the ‘very strong’ characteristics may be the best early adopters of a program or best targets for leadership roles in operational programs reliant on social cohesion. Operational Implication 2: Similarly, special attention should be given to groups that are not significantly associated with social cohesion. Engaging these groups is critical to ensure that programs are inclusive and build social cohesion in a community rather than further marginalizing these groups.

12

Community Resilience Mean resilience in Solino was 3.07, and mean resilience in Tesso was 3.14. Because a higher score corresponds to lower resilience, this means that before adjusting for covariates5, Tesso has lower resilience than Solino. The difference in resilience score between the two communities was found to be significantly different between the two communities, meaning that before adjusting for demographic covariates, resilience differs between the two communities. Additionally, after taking mean social cohesion score and demographic differences between the two communities into consideration, there remains a highly significant difference in mean resilience scores between Solino and Tesso. Analyses show that as compared to Solino, Tesso has a statistically significant decrease in mean resilience, meaning that Tesso has less resilience than Solino.

Further, the following variables are important indicators of resilience. These results are presented in further detail in Table 4 of the attached appendix. Below they are categorized according to their strength of association with resilience, listed in decreasing order of strength: Very Strong • Social cohesion score (An increase in mean social cohesion is associated with an increase in mean

resilience) • Community (Solino has greater resilience vs. Tesso) • Gender (Females have better resilience vs. Males) • Number of Income Earners (More income earners in the household, better resilience)

Strong • Years in House (Residents who have lived in their community for 0-3 years have better resilience

vs. 3-5 years) • Education Status (Respondents who have a University Education have better Resilience vs. those

with No education) • Residence Status (Squatters have better resilience vs. Owners) Moderately Strong • Years in House (Residents who have lived in their homes 0-3 years have better resilience than

those who have lived in their homes for 21+ years) • Education Status (Those with No Education have better resilience compared to those with a

Primary Education) • Average Income per Month (Residents who earn 5-10k per month have more resilience vs. 0-5k

Gourdes)

5 A covariate is a secondary variable that can affect the relationship between the dependent variable (resilience score) and other independent variables (community) of primary interest.

Finding #3: Tesso has less resilience as compared to Solino. This conclusion remains true before and after adjusting for demographic differences between the two communities and despite similar overall social cohesion.

Operational Implication: Programs should aim at customizing activities towards improving resilience or even disaster risk reduction activities in Solino and Tesso.

13

A univariate analysis of the association between social cohesion and resilience demonstrates that social cohesion in both Solino and Tesso Solino is associated with resilience. The Multivariate analysis demonstrates that the magnitude of this association is stronger in Tesso, but is significant in both communities. From the above analyses the following two conclusions can be drawn:

Finding #4: Certain demographic variables are more strongly associated with resilience. The above listed demographic variables are significantly associated with resilience in Solino and Tesso. Operational Implication 1: These groups may be the best initial targets of programs aimed at engaging community members for programs improving resilience or reducing disaster risk in the two communities. Specifically, persons with the ‘very strong’ characteristics may be the best early adopters of a program or best targets for leadership roles in operational programs intended to reduce risk and improve community resilience Operational Implication 2: Similarly, special attention should be given to groups that are not significant associated with resilience. Engaging these groups is critical to ensure that programs are inclusive and build overall community reliance rather than leaving them at increased risk during disasters.

Finding #5: Social cohesion is a very strong predictor of community resilience. The degree of social cohesion in Solino and Tesso is strongly tied to the degree of resilience in these communities. As such, increasing social cohesion can increase community resilience. Despite this, Tesso has similar social cohesion but worse resilience as compared to Solino. It is important to note that this is simply a baseline measure relative to Solino. Therefore, the overall finding that improving social cohesion will result in improved resilience is still valid. Operational Implication: This finding is extremely important operationally as it highlights that an increase in social cohesion will result in an increase in resilience. Therefore, programs aimed at improving resilience in these two communities should also consider social cohesion as a pathway towards greater resilience. Program activities should always consider the impact they have on social cohesion. By improving social cohesion in these neighborhoods, there will be a resulting increase in community resilience as well. Programs aimed at reducing disaster risk should understand the role that social cohesion and demographics of these communities play in building resilience.

14

While providing valuable quantitative insight into these two communities, this study could not account for their different histories or their relationship with government. These are very different communities in these two regards and it was impossible to measure these differences or incorporate them into such a quantitative analysis. Nonetheless, these factors play a very important role in both social cohesion and resilience. We suspect that had the role of government been something that could be included it would also be a variable strongly associated with resilience. Government has an important role to play in helping communities achieve resilience. Policies that enable communities and empower them to achieve resilience make them a positive force in the equation. Conversely, a poorly functioning government and over-reliance on it for otherwise community level activities can hinder the development of community resilience. Aside from the demographic make-up of these two communities, we suspect that the role of government is an important influence on their resilience. Further qualitative study through case studies or outcome mapping exercises may help explore and characterize this relationship. Community narratives and histories also play a role in developing or hindering social cohesion and, as this study proves, this cohesion impacts resilience. Again, a non-quantifiable factor in this study, community histories could explain some of the variance between the communities that these data show. Although the initial focus groups were of limited value, future focus groups may help elucidate how histories and narratives impact these two communities and their cohesion and resilience. Perception of and intrinsic quality of community leaders and community based organizations (CBO) can have either a positive or adverse impact on community social cohesion and resilience. Additionally, community leaders and CBOs interact with government and other outside agencies such as aid organizations. How these play out can override the intrinsic social cohesion and resilience measured in these communities and mitigate or amplify the impact that government, identified above, can have on them. Unfortunately, we could not use this quantitative study to evaluate these interactions among community leaders and CBOs with outside entities and their own communities. As noted in this report, targeting marginalized groups is important. Those demographics that associate less with social cohesion and resilience may be an issue of perspective but they are likely the same groups that are not as well integrated into the community or feel served by it. Ensuring that future programs ensure that they are not excluded either by design or by accident is very important. If programs and policies fail to engage these groups, specific effort should be made for outreach and ensuring participation. Having these groups interact with groups that associate with higher social cohesion and resilience may also be another strategy to explore. This study also demonstrates some interesting findings that can be explored further with qualitative study. Despite having more of the demographics that are associated with lower social cohesion, Tesso still has very similar social cohesion to Solino overall. There may be other characteristics about Tesso and it’s recent history that explain why it has more almost equal social cohesion. Additionally, Solino demonstrated better resilience overall than Tesso even though it had similar social cohesion. This shows that social cohesion is only part of the equation in building resilience. Further study in Solino may help reveal what other factors help improve it’s overall resilience.

DISCUSSION AND LIMITATIONS

15

In sum, this quantitative analysis reveals much about the social cohesion and resilience in these two communities to guide operational efforts. Further qualitative study can help reveal a more nuanced understanding of role that non-quantifiable factors play.

16

Almedom AM, Tumwine JK: resilience to disasters: a paradigm shift from vulnerability to strength. African Health Sciences 2008, 8(S):1-4. Buckner J. The Development Of An Instrument To Measure Neighborhood Cohesion. Am J Commun Psychol 1988, 16(6): 771-791.

Pfefferbaum RL, Pfefferbaum, B, Van Horn FL. Communities Advancing resilience Toolkit (CART). The CART Integrated System. 2011.

Putnam, Robert with Robert Leonardi and Raffaella Y. Nanetti. 1993. Making Democracy Work: Civic Traditions in Modern Italy. Princeton: Princeton University Press.

CITATIONS

17

Table 1. Demographic Variables: Summary Statistics n (%) n (%) Solino Tesso p Sex

Male 671 (35.5) 431 (37.2) 240 (32.8) 0.060 Female 1219 (64.5) 728 (62.8) 491 (67.2)

Religion 0.017* Catholic 812 (42.7) 487 (41.8) 325 (44.2)

Protestant 821 (43.2) 523 (44.9) 298 (40.5) Adventist 45 (2.4) 17 (1.5) 28 (3.8)

Jehovah’s Witness 11 (0.6) 7 (0.6) 4 (0.5) Voodoo 43 (2.3) 29 (2.5) 14 (1.9)

Other 168 (8.8) 102 (8.8) 66 (9.0) Marital Status 0.105

Married 492 (26.0) 292 (25.2) 200 (27.2) Single 514 (27.1) 337 (29.1) 117 (24.1)

Divorced 13 (0.7) 9 (0.8) 4 (0.5) Domestic Partnership 876 (46.2) 552 (45.0) 354 (48.2)

Average Monthly Income <<0.001*** 0-5k HTG ($0-107 USD) 959 (50.8) 567 (49.0) 392 (53.7)

5-10k HTG ($107-214 USD) 557 (29.5) 349 (30.2) 208 (28.5) 10-15k HTG ($214-321 USD) 204 (10.8) 109 (9.4) 95 (13.0)

15k-20k HTG ($321-428 USD) 100 (5.3) 83 (7.2) 17 (2.3) >20k HTG (>$428 USD) 66 (3.5) 48 (4.2) 18 (2.5)

Level of Education 0.002** Primary 763 (40.3) 449 (38.8) 314 (42.7)

Secondary 732 (38.7) 440 (38.1) 292 (39.7) University 131 (6.9) 77 (6.7) 54 (7.3)

None 265 (14.0) 190 (16.4) 75 (10.2) Type of Residence << 0.001***

Own 849 (44.7) 500 (42.9) 349 (47.4) Rent 827 (43.5) 618 (53.0) 209 (28.4)

Squat 194 (10.2) 29 (2.5) 165 (22.4) Other 31 (1.6) 18 (1.5) 13 (1.8)

Primary Source of Income 0.004** Small Business 912 (48.3) 570 (49.3) 342 (46.8)

Permanent Employment 340 (18.0) 222 (19.2) 118 (16.1) Short-Term Employment 380 (20.1) 203 (17.5) 177 (24.2)

Other 256 (13.6) 162 (14.0) 94 (12.9) Mean (sd)

Mean (sd) Solino Tesso p Age 38.26 (13.45) 39.16 (13.9) 36.62 (12.3) <0.001*** Income Earners in Household 0.75 (0.80) 0.77 (0.87) 0.67 (0.91) <0.001*** Number of Children 3.54 (1.60) 3.6 (1.7) 3.4 (1.4) 0.052 Years in Community 15.15 (12.17) 17.8 (12.3) 10.8 (10.3) <<0.001*** Years in Current Residence 9.37 (10.28) 10.8 (11.2) 6.7 (7.63) <<0.001*** Household Size 5.14 (2.27) 5.2 (2.3) 5.1 (2.2) 0.391 Note: N = 1627. Significance codes: *0.05, **0.01, ***0.001. Statistical tests: Wilcoxon Rank-Sum test was used for continuous outcomes; Chi-Square test was used for dichotomous outcomes.

APPENDIX - Tables

18

Table 2. Demographic Variables: Total and Community Mean social cohesion (SC) and resilience scores Mean SC Score Mean resilience Score Total Solino Tesso Total Solino Tesso Sex

Male 2.382 2.427 2.301 3.115 3.067 3.202 Female 2.608 2.614 2.599 3.088 3.071 3.113

Religion Catholic 2.459 2.439 2.409 3.089 3.042 3.158

Protestant 2.572 2.576 2.564 3.095 3.089 3.106 Adventist 2.804 3.143 2.599 3.115 3.108 3.119

Jahovas Witness 3 2.815 3.324 3.157 2.960 3.500 Vaudou 2.46 2.477 2.427 3.120 3.075 3.213

Other 2.559 2.55 2.573 3.140 3.103 3.198 Marital Status

Married 2.51 2.556 2.444 3.126 3.094 3.172 Single 2.589 2.561 2.642 3.061 3.022 3.135

Divorced 2.593 2.536 2.721 3.145 3.148 3.139 Domestic Partnership 2.504 2.536 2.457 3.104 3.088 3.128

Average Monthly Income 0-5k HTG ($0-107 USD) 2.477 2.483 2.47 3.099 3.054 3.166

5-10k HTG ($107-214 USD) 2.541 2.547 2.532 3.055 2.996 3.153 10-15k HTG ($214-321 USD) 2.652 2.727 2.566 3.119 3.095 3.146

15k-20k HTG ($321-428 USD) 2.629 2.653 2.513 3.149 3.194 2.931 >20k HTG (>$428 USD) 2.616 2.702 2.386 3.278 3.477 2.747

Level of Education Primary 2.518 2.552 2.469 3.141 3.137 3.147

Secondary 2.581 2.623 2.519 3.114 3.091 3.150 University 2.639 2.585 2.715 2.938 2.826 3.098

None 2.345 2.326 2.394 2.988 2.929 3.136 Type of Residence

Own 2.423 2.409 2.443 3.078 3.015 3.168 Rent 2.659 2.657 2.666 3.128 3.104 3.196

Squat 2.408 2.542 2.385 3.042 3.161 3.021 Other 2.644 2.531 2.801 3.204 3.238 3.158

Primary Source of Income Small Business 2.495 2.501 2.486 3.074 3.006 3.186

Permanent Employment 2.626 2.656 2.569 3.085 3.068 3.116 Short-Term Employment 2.432 2.456 2.405 3.116 3.106 3.127

Other 2.64 2.652 2.618 3.181 3.247 3.065 Social cohesion (SC) score is available only for dichotomous outcomes, whereby 1 = strongly agree through 5 = strongly disagree.

19

Table 3. Multivariate Associations Between Demographic Variables And social cohesion scores

β p Community

Solino Reference Tesso -0.088 0.009**

Sex Male Reference

Female -0.215 <<0.001*** Years in Community

0-3 Reference 3-5 -0.175 0.002**

6-10 -0.262 <<0.001*** 11-20 -0.395 <<0.001***

21+ -0.515 <<0.001*** Marital Status

Married Reference Single 0.080 0.046*

Divorced 0.0833 0.611 Domestic Partnership -0.012 0.742

Education Status None Reference

Primary 0.186 <0.001*** Secondary 0.185 <0.001*** University 0.308 <0.001***

Residence Status Own Reference Rent 0.077 0.027*

Squat -0.070 0.171 Other 0.221 0.046*

Religion Catholic Reference

Protestant 0.087 0.007** Adventist 0.205 0.034*

Jehovah Witness 0.460 0.027* Vaudou 0.079 0.410

Other 0.056 0.289 Number of Children 0.0217 0.034* Household Size -0.010 0.142 Primary Source of Income

Small Business Reference Permanent Employment 0.086 0.035* Short-Term Employment -0.030 0.444

Other 0.113 0.013* Note: Linear regression models. N for each model = 1627. Outcomes: social cohesion (SC), and resilience, whereby 1 = strongly agree through 5 = strongly disagree. Significance codes: *0.05, **0.01, ***0.001.  

20

Table 4. Multivariate Associations Between Demographic Variables And resilience scores

β p Social cohesion score 0.338 <<0.001*** Community

Solino Reference Tesso 0.127 <0.001***

Sex Male Reference

Female -0.123 <<0.001*** Years in House

0-3 Reference 3-5 0.103 0.004**

6-10 -0.016 0.699 11-20 0.028 0.521

21+ 0.117 0.011* Education Status

None Reference Primary 0.095 0.014*

Secondary 0.052 0.199 University -0.170 0.005**

Residence Status Own Reference Rent 0.024 0.448

Squat -0.125 0.009** Other 0.070 0.471

Average Monthly Income 0-5k HTG ($0-107 USD) Reference

5-10k HTG ($107-214 USD) -0.059 0.044* 10-15k HTG ($214-321 USD) -0.035 0.416

15k-20k HTG ($321-428 USD) 0.044 0.454 >20k HTG (>$428 USD) 0.124 0.080

Number of Income Earners -0.042 0.005** Note: Linear regression models. N for each model = 1627. Outcomes: Social cohesion (SC), and resilience, whereby 1 = strongly agree through 5 = strongly disagree. Significance codes: *0.05, **0.01, ***0.001. Table 5. Logistic Regression Analysis of social cohesion and resilience, disaggregated by Community β p Social cohesion: Solino 0.348 <<0.001

Social cohesion: Tesso 0.261 0.028*

21

SOLINO SAMPLING DISTRIBUTION:

APPENDIX - MAPS

22

TESSO SAMPLING DISTRIBUTION:

23

The following is a list of essential components in making cities resilient. This list was obtained from The United Nations Office for Disaster Risk Reduction (UNISDR) Resilience Handbook, which can be found at the following link: http://www.unisdr.org/files/26462_handbookfinalonlineversion.pdf. • Essential 1: Put in place organization and coordination to understand and reduce disaster risk,

based on participation of citizen groups and civil society. Build local alliances. Ensure that all departments understand their role to disaster risk reduction and preparedness.

• Essential 2: Assign a budget for disaster risk reduction and provide incentives for homeowners, low-income families, communities, businesses and public sector to invest in reducing the risks they face.

• Essential 3: Maintain up-to-date data on hazards and vulnerabilities, prepare risk assessments and use these as the basis for urban development plans and decisions. Ensure that this information and the plans for your city's resilience are readily available to the public and fully discussed with them.

• Essential 4: Invest in and maintain critical infrastructure that reduces risk, such as flood drainage, adjusted where needed to cope with climate change.

• Essential 5: Assess the safety of all schools and health facilities and upgrade these as necessary. • Essential 6: Apply and enforce realistic, risk compliant building regulations and land use

planning principles. Identify safe land for low-income citizens and develop upgrading of informal settlements, wherever feasible.

• Essential 7: Ensure education programs and training on disaster risk reduction are in place in schools and local communities.

• Essential 8: Protect ecosystems and natural buffers to mitigate floods, storm surges and other hazards to which your city may be vulnerable. Adapt to climate change by building on good risk reduction practices.

• Essential 9: Install early warning systems and emergency management capacities in your city and hold regular public preparedness drills.

• Essential 10: After any disaster, ensure that the needs of the survivors are placed at the center of reconstruction with support for them and their community organizations to design and help implement responses, including rebuilding homes and livelihoods.

Future Resources Future follow-up on this study may be achieved using various resources and tools. One such recommended tool is The Local Government Self-Assessment Tool designed by UNISDR. This self-assessment provides key questions and measurements guided by the aforementioned essential components to making cities resilient. This tool will help cities and local governments determine baseline resilience measurements, and can be used in the future to measure progress from baseline. This tool can be found at the following link: http://www.unisdr.org/campaign/resilientcities/toolkit/howto.

APPENDIX – Government Opportunities on Resilience

24

Additionally, government officials should devise a community action framework aimed at improving resilience. One such framework presented below is adapted from the RAND Corporation Community Resilience Action List. The following are suggested components of a community action framework:

1. Focus on Assets in your community. Consider the following assets available in your community: Competencies, money, infrastructure, equipment, services, relationships, data, and other assets that your community has that have not been identified. Conducting an assessment focused on community-level assets can identify assets.

2. Improve your community’s plan for long-term recovery. Include the following items in your plan: Recovery activities that extend beyond the acute phase of disaster, recovery activities that extend until normalcy resumes. Additionally, develop a community-wide long-term recovery plan that includes data for tracking progress, long-term behavioral health needs, and adaptations for your community. Pilot test this recovery plan through a planned exercise.

3. Help community members plan for the social and economic impacts of disaster. Consider including the following items in your framework: Information on rapidly providing economic supports to disaster-affected individuals, ways to estimate long-term economic effects and strategies for accessing relevant funding, support for those who may be considered socially isolated (via distance, vulnerabilities, etc.). It is also critical to develop a plan for disseminating information about economic preparedness and recovery, and map and plan those at greatest “social risk”.

4. Develop data systems to track preparedness and recovery planning. These systems should include: Sources of data on community assets and vulnerabilities, plans for accessing organization and community-level information on how resources are distributed after a disaster, information on the health and well being of individuals before and after a disaster. Use assets and vulnerabilities data to inform community planning and decision-making and synchronize data systems that will provide information on community recovery.

5. Plan for sustainable development after disaster. Include in your plan: Discussions on how your community will rebuild service lines and more-resilient infrastructure, partnerships that support redevelopment of relevant services, infrastructure, etc., and development of a community rebuilding team.