• No results found

Public attitude towards green roofs in Watergraafsmeer

N/A
N/A
Protected

Academic year: 2021

Share "Public attitude towards green roofs in Watergraafsmeer"

Copied!
43
0
0

Bezig met laden.... (Bekijk nu de volledige tekst)

Hele tekst

(1)

1

Public attitude towards green roof

implementation

Figure 0. Green roofs (retrieved from

https://www.ecosia.org/images/?q=green%20roofing#id=5D37528097412941E61C2E927BDD3950A5 D622F4)

Lars Kagie (11621613) 28-05-2021

Bachelor Thesis

Supervisors: Jannes Willems and Rosa van Schaik Wordcount: 9929

(2)

2

Content

1. Abstract……….

3

2. Introduction………..

4

3. Theoretical Framework………..

6

3.1 Green roofs………

6

3.2 Public attitude……….

6

3.3 Implementation……….

8

3.4 Conceptual model……….

8

4. Methodology……….

9

4.1 Research design……….

9

4.1 Operationalization………

9

4.2 Data-collection……….

11

4.2 Analysis-methods………..

13

5. Results………..

16

5.1 Age and public attitude……….

16

5.2 WTP and public attitude………..

18

5.2 Environmental benefits and public attitude………..

21

5.3 Differences between neighborhoods and public attitude…

25

6. Conclusion………..

29

7. References………..

31

8. Appendix 1 (survey)...……….

34

(3)

3

Abstract

Urban climate adaptation is a term increasing in relevancy every day. Cities around the world are already experiencing the consequences of climate change and solutions need to be realized in order to mitigate impacts. One of them is green roofing, an initiative that has environmental benefits for the city, as well as for the individual. In 2020 a plan was proposed to install more green roofing in Watergraafsmeer, Amsterdam. However, due to too little public support the plan was never accomplished. This research aims to answer the question of what factors influence the public attitude towards the implementation of green roofs in Watergraafsmeer, Amsterdam. The factors researched are: age, willingness to pay, environmental benefits of green roofs, income and percentage of people that live in bought housing. To determine to what extent and what factors affect public attitude a survey was conducted across 150 residents of Watergraafsmeer.

Furthermore, data was collected from the municipality of Amsterdam and the Central Bureau of Statistics (CBS). After statistical and spatial analysis’ using the programs R and ArcGIS two

relationships were found. Background-related factors such as age and income were not found to be a significant predictor of public attitude and neither was willingness to pay. Environmental benefits such as motive for installing green roofs can however be seen as a significant predictor of public attitude, with more efficient energy use being the biggest driver. In the spatial analysis some correlation was also found between house ownership and the public attitude. Furthermore, additional research is needed to establish the link between public attitude and the implementation of green roofs.

(4)

4

Introduction

Due to climate change and rising temperatures across the globe, cities and urban regions will have to be able to manage consequences, such as: extremes and irregularities in precipitation and

temperature (urban heat island effect) and higher storm frequencies (Foster et al., 2011). As a consequence of these scenarios the term urban climate adaptation is an emerging policy domain. The combination of large populations, densely built buildings and paved surfaces makes it

increasingly apparent that cities do not represent ideal areas to be able to adapt to climate change (Derkzen et al., 2017).

Consequences of climate change can however be mitigated and adapted to, if carried out properly. One the concepts of urban climate adaptation is the implementation of green infrastructure, which Foster (2011, p. 150) explains as: “Green infrastructure approaches help to achieve sustainability and resilience goals over a range of outcomes in addition to climate adaptation”. Benefits of green in the city are often related to their ability to neutralize the impacts of extreme precipitation events or temperature rises. According to Foster (2011) benefits of green infrastructure can be divided into five categories of protection: (i) land-value, (ii) quality of life, (iii) public health, (iv) hazard mitigation and (v) regulatory compliance. One example of green infrastructure is the adaptation of buildings in the form of green roofs, a term that will stand central within this research.

As explained above, constructing green infrastructure can be done in many ways, but is often difficult to implement (Naumann et al., 2011). Naumann (2011) explains common barriers and enabling factors in regard to the implementation of green infrastructure. Barriers consist of differing priorities and points of view with stakeholders, but also economic capacity barriers such as insufficient

funding. Enabling factors consisted of embracing partnerships, acquiring sufficient knowledge and evidence and raising public support. Zhang (2021) elaborates on that by explaining that public attitude can be seen as a barrier in implementation of green infrastructure. It is therefore important to find out what drives public attitude, to be able to eliminate the barriers that Zhang (2021) and Naumann (2011) set out.

As stated above, one of the examples of green infrastructure is green roofing. As cities are getting denser, parks or urban forestry cannot be realized easily. Green roofing provides a plausible solution. Green roofs could provide a number of different urban services, e.g. (i) improve energy efficiency in buildings, (ii) reduce the urban heat island effect, (iii) retain storm water and (iv) improve biodiversity (Grunwald et al., 2017). Patchen (2006) explains that these environmental benefits could influence the public attitude via personal feedback. If people directly benefit from for instance lower energy costs, they might be more inclined to install a green roof themselves (Patchen, 2006).

In Amsterdam, the capital of the Netherlands, the green party (GroenLinks) always does relatively well (CBS, 2021). GroenLinks is a party known for its strong stance on the topic of climate change. Since, generally, many people vote for this party it is suggested that Amsterdam residents might be more concerned with the consequences of climate change than the average Dutchman. Often, the city of Amsterdam proposes green implementation plans, as it did in October 2020. They introduced a plan to implement more green roofing in Watergraafsmeer (WGM). It did however not get enough public support (Gemeente Amsterdam, 2020) and therefore was not implemented. There was no favorable public attitude towards implementation of green roofs. Because of this it is important to determine what factors might influence public attitude, leading to the following research question:

(5)

5

What factors influence the public attitude towards the implementation of green roofs in Watergraafsmeer, Amsterdam?

Because the consequences of climate change are becoming a more imminent “threat” and cities need to act accordingly, it is important to establish what factors might have an effect on the public attitude towards plausible solutions of this “threat”. According to Derkzen (2017), the population of the Netherlands has a relatively good knowledge of the consequences of climate change, but is not a front-runner when it comes to greening the society. This research might shed a light on this

paradoxical statement.

To be able to answer the above research question a quantitative research is conducted in which the influence of different factors on the public attitude is questioned via a survey and statistical/spatial analysis. Firstly, in the theoretical framework the concepts of green roofs, public attitude and implementation are briefly explained and defined. Then, the concepts are made measurable in the methodology and divided into core-indicators. The methodology will also explain how the research was conducted: what the survey entails, how the survey was spread, how participants were selected and what type of analysis will be used to answer the research questions. Thereafter the results will point out what concepts have a significant relationship with public attitude and what are the major predictors.

(6)

6

Theoretical Framework

This part of the research will explain and define certain core-concepts of the research according to a literature review. These concepts are green roofs, public attitude and implementation. Definitions will be used in operationalization of the concepts into measurable indicators.

Green roofs

Green roofs often refer to vegetative roof systems that consist of plants on top of a certain roof with two classifications: intensive and extensive (Kosareo et al., 2007). Intensive green roofs have

“intense” maintenance needs and often have a wide plant variety that usually consists of trees and shrubs that require deeper substrate layers than extensive green roofs (Getter, et al. 2006). Intensive roofs are more park-like, often require a depth up to 3 feet (Kosareo et al., 2007) and can sometimes be accessible to the public. Therefore, they are generally limited to flat roofs. In contrast, extensive green roofs require minimal maintenance. They are thin, light systems with vegetation mostly limited to plant species like herbs, grasses, mosses and succulents like sedum (Getter, 2006). Extensive green roofs can easily be built on a sloped surface and are cheaper to implement (Getter, 2006).

Cavanaugh (2008) argues that a green roof can be considered green if it is environmentally friendly. Susca (2007) however refutes this definition by arguing that if the “greenness” of a roof depends on whether or not it is environmentally friendly, a vegetation system is not the only option worthy of the definition of a green roof. To make sure the definition of a green roof does not solely depend on environmental friendliness the following will be used in this paper: “a roofing system that

is designed, constructed, maintained, rehabilitated, and demolished with an emphasis throughout its life cycle on using natural resources efficiently and preserving the global environment” (Christian &

Petrie, 1996, p. 88).

In relation to this research preservation of the environment is an important factor, because this is part of the entity that people might evaluate in favor or disfavor. Getter (2006) describes four different environmental benefits of green roofs: (i) reduced volume of stormwater runoff; within a green roof system much of the precipitation is captured in the medium or vegetation and will eventually evaporate back in to the atmosphere via transpiration. In cities this could therefore reduce the likelihood of floods. (ii) Lower temperatures in the city;- green roofs absorb sunlight because they provide shade and insulation. Green roofs can reduce indoor temperatures up to 4 oC with outside temperatures of 25-30 oC (Peck et al., 1999). (iii) Energy efficiency and conservation; because green roofs absorb sunlight and cause lower temperatures they can cool and improve energy efficiency of the building they are installed on. (iv) Increased biodiversity; extensive green roofs are inaccessible to public and can therefore harbor micro-organisms, insects and birds. Intensive green roofs can even provide habitat to bigger species if they have a substantial surface and therefore highly increase biodiversity.

Public attitude Definition

Pidgeon (2012) states that the issue of climate change, when compared to other global/societal issues, does not have the highest priority for many people. This statement indirectly indicates that measures for climate adaptation, such as green roofing, may also not have a high priority. Zhang (2021) describes that there is a certain negative public attitude amongst people to implement green roofing on their own/or other people’s building(s). It is therefore important to understand and define

(7)

7

this public attitude. Daniel Katz (1960, p. 150) was the first to define the concept of attitude: “Attitude is defined at the individual level, namely, the specific organization of feelings and beliefs according to which a given person evaluates an object or symbol positively or negatively”. Mitchell & Olson (1981) however argued that attitude is “an individual’s internal evaluation of an object” (p. 318). They state that attitude is not always a concept that originates from a person’s belief or learned values but purely an evaluation of an object. Eagly & Chaiken (1993) also explain attitude as a

“psychological tendency that is expressed by evaluating an entity with favor or disfavor” (p. 143), which now can be considered as the general agreement amongst attitude theorists. In addition, Eagly & Chaiken (2007) later also explain that there are degrees in favorability or unfavourability and people often do not solely evaluate an object with favor or disfavor.

Psychological tendencies as result of background

Psychological tendencies can originate from, for example, thoughts, feelings, intentions, behavior and background (Eagly & Chaiken, 1993). As stated above, green roofs are designed to preserve the environment, i.e. they have environmental benefits. Reinhart (2018) states that 70% of younger people (18-34) are worried about global warming, with only 56% of older people (55 and over) having concerns about global warming. Reinhart (2018) therefore argues that there could be a relation between age and people’s perception of the threat that is climate change. Therefore, age might be a partial source for the psychological tendency that develops the attitude toward green roofs. In addition to this, Bradley Jorgensen (2000) argues that income and willingness to pay (WTP) also might have an influence on the attitude towards an environmental good. It is stated that attitude toward paying is often negatively related to the price of the intervention (green roofs) if the latter is perceived to be beyond respondents’ WTP. Jorgensen (2000) also argues that there might be a positive linear relationship between income and attitude. People with lower income may consider it unfair that they must pay, seeing as they often do not have the financial means. On the other hand, higher income respondents might deem it unfair that they are targeted to pay a larger proportion of the cost of the intervention. One indicator of income is house ownership (Lind, 2009). Often, there is a causal relation between income and the type of residential housing. Wealthier people are more likely to live in bought housing, while rented housing is more common amongst residents with less money. Therefore, type of housing is a factor that is related to income and therefore could influence the public attitude.

Jorgensen also states that demographic variables (age) and environmental values could also play a role in the attitude towards the intervention, such as green roofs. Environmental values herein can be related to the environmental benefits that green roofs have. Therefore, a link between the

environmental benefits of green roofs and public attitude is proposed and considered as a factor that might influence the public attitude. Within this research, age, income, house ownership, willingness to pay and environmental benefits are defined and considered as factors that might influence the public attitude toward the intervention that is green roofing.

Attitude to public attitude

The definition of attitude by Eagly & Chaiken (1996) can be taken to a broader level, namely the public. Moffit (1992, p.18) describes the public as “a group of individuals who share a common goal or relationship to the organization”. Within this research this relationship to an organization is the fact that they all live in Watergraafsmeer. Even though a residential neighborhood is not an organization, the emphasis of this definition lies on the common relationship. The previously explained definition of attitude can be explained as public attitude if the definition includes people with a relationship to a certain “organization”, namely Watergraafsmeer. If many people evaluate a

(8)

8

certain entity in a positive or negative way it can be said there is a public attitude towards this entity (Altmann, 2008). My research reasons from Eagly & Kulesa (1997) who describe that an attitude does not exist until the entity in question has provoked an evaluative response. In this case that would mean that the public can have an attitude towards green roofing if green roofing draws out a certain evaluative response, which it does, as stated in previously mentioned research.

Implementation

Lomas (1993) described implementation as a “local process of communication in which appreciation

of the research findings is a necessary but not sufficient condition to bring about changes in decision-making that reflect the message from the research” (p. 227). Because appreciation of research

findings is not sufficient according to Durlak (2015), he defines implementation as “efforts designed

to get evidence-based programs or practices of known dimension into use via effective change strategies” (p. 395). Rather than appreciation towards research, Durlak (2015) states that there

needs to be evidence and enough research done on a certain topic for it to be implemented.

Effective change strategies originate from policies and policy change (Wlezien et al., 2007). The focus lies on different factors that might influence the public attitude, which is why implementation is integrated into the definition of public attitude, as can be seen in the conceptual model below (figure 1).

Conceptual model

The explained three concepts and its links lead to the following conceptual model:

Figure 1. Conceptual model of the theoretical framework. Certain links are defined and do not need to

be researched (dotted lines). For instance, the relation between income and WTP is already confirmed by Jorgensen (2000). Central within this research are relationships between all concepts and public attitude.

(9)

9

Methodology

Research design

As briefly stated in the introduction, the conducted research is of quantitative nature. Data is

collected, analyzed, presented and conclusions are drawn from results. Quantitative study allows for replication, direct comparison of results, large sampling and hypothesis testing (Edmonds et al., 2016). A particular group of people with common characteristics, namely the residents of Watergraafsmeer, is researched and conclusions are drawn representing them. This research is a case study of Watergraafsmeer and data collection is done via surveys. After this collection, a combination of two types of research design will be applied. First, descriptive research will give a summary of all study variables and an overview of how Watergraafsmeer is represented. It

accurately describes the population, but does not answer any why questions. To be able to do that, correlational research is applied, a type of research used to determine if there is a relationship between the dependent and independent variable. Independent variables will consist of age, willingness to pay, environmental benefits, income and house ownership, while the dependent variable will always be public attitude. Survey-answers will be made numerical and statistical analysis (in R) will give outputs for possible correlation relationships between dependent and independent variables. Because differences between neighborhoods are also researched, the last sub-question will be answered using spatial analysis (in ArcGIS Pro).

In the methodology the previously given definitions of core-concepts will be operationalized and made measurable, leading to four sub-questions that will be answered. After this the data-collection process is described and explained, after which the analysis methods are discussed on how the sub-questions will be answered.

Operationalization

It is important for this research to establish the definition of a green roof and make it measurable. In the theoretical framework the following definition was made clear: “a roofing system that is

designed, constructed, maintained, rehabilitated, and demolished with an emphasis throughout its life cycle on using natural resources efficiently and preserving the global environment” (Christian &

Petrie, 1996, p.86 ). Preservation of the global environment is the most important part of this definition and is made measurable in the operationalization table. It is important to understand that this definition can be made measurable because of probable effect of the environmental benefits on the public attitude, a possible link that was established within the theoretical framework. Large part of this definition will however not be operationalized because the emphasis of this research lies on the factors that influence public attitude and implementation, not the technical part of green roofs. The aim of this research is to determine what the public attitude towards green roofing is, creating the need for measurement of the concept of public attitude. In the theoretical framework a definition was already made clear: “a psychological tendency that is expressed by evaluating an entity with favor or disfavor” (Eagly & Chaiken, 1993, p. 143). The entity within this research would be the implementation of green roofing. Using this definition, the operationalization of the concept is straight-forward. By measuring the evaluation of favor or disfavor public attitude can be determined. However, as stated in the theoretical framework, the concepts of favor or disfavor have degrees. This will be expressed in favor, slightly favor, neutral or slightly disfavor, disfavor, (see table 1) which will be measured using surveys.

(10)

10

Possible concepts that might influence the aforementioned psychological tendency are previously defined as age, income, willingness to pay and environmental benefits of the intervention. Age is easily measurable within surveys wherein age of the respondents itself is the indicator. Within a survey-based research income often is a subject of privacy. Therefore, to determine respondents’ income the average income of their neighborhood is taken as an indicator. The central bureau of statistics (CBS) provides the average income per neighborhood in Watergraafsmeer. The municipality of Amsterdam divides Watergraafsmeer into four neighborhoods: Middenmeer, Frankendael,

Betondorp and Overamstel (see map 1). As stated before, percentage of home ownership can also be seen as an indicator of income and will be measured by asking people what type of house they live in. To determine WTP a distinction will be made between subsidized and self-pay green roofs, which can be seen as the indicators of respondents’ WTP. All aforementioned definitions, accompanying concepts and indicators can be seen set out in table 1.

Map 1. Different neighborhoods within Watergraafsmeer.

The last term defined in the theoretical framework is implementation. As suggested by the

conceptual model however, this concept is concatenated in the approach of public attitude towards implementation. The emphasis within this research lies on the social-economic factors that might influence public attitude towards implementation, which is why implementation on itself is not a term that needs to be made measurable.

As a result of the research question, theoretical framework and the operationalization of key-concepts the following sub-questions are proposed:

1. To what extent does age affect public attitude toward green roofs?

2. To what extent does willingness to pay affect public attitude towards green roofs?

3. To what extent do environmental benefits of green roofs affect public attitude towards green roofs?

(11)

11

4. To what extent do differences between the four neighborhoods of Watergraafsmeer, such as number of green roofs, income and percentage of house ownership, have an effect on the public attitude?

Table 1. Operationalization table

Data-collection

Within this part of the research it is important to mention that all four sub-questions will be

answered using the same type of data collection. Only income (CBS, 2020) and amount of green roofs in Watergraafsmeer, (Municipality of Amsterdam, 2020) are variables that were not retrieved via a survey. The following paragraphs will explain how data was collected, how participants were selected, what materials were used etc.

Participants

The aim of this research is to establish what factors explain public attitude towards green roofing and how this public attitude is formed. As stated in the definition of public attitude, if many people evaluate a certain entity in a positive or negative way it can be said that there is a public attitude towards this entity (Altmann, 2008). As a result of this, if more people evaluate a certain object, a stronger public attitude can be formed. It is therefore important to determine the evaluation of green roofs by as many people as possible. To answer sub-questions a survey was issued across Watergraafsmeer, a district with approximately 36.000 residents. To be able to represent the

population of Watergraafsmeer, a certain number of survey-respondents was set as a goal. Seeing as this research was conducted without funding or external financial means the methods to spread the survey were limited. To account for this a 10% margin of error was set in regards to the population reflection, resulting in a requirement of less respondents. Then, taking into account an average response rate of 5-15% to web surveys (Lugtig et al., 2016) and the limited survey distribution options a survey-goal of 100 was set. After distribution 150 respondents were recorded. Distribution

(12)

12

of the survey was mostly random, meaning that no distinction was made in background of the respondent. Public attitude represents the public which is why the distribution was randomized, to get a diverse group of respondents. The distribution was however not entirely random, because a partial aim of this research is to find out differences between neighborhoods within

Watergraafsmeer. Therefore, a priority was to try and assure that respondents were evenly divided across the four areas of Watergraafsmeer.

Watergraafsmeer

This research is a case study of Watergraafsmeer and aims to represent the population. In the fourth sub-question differences between neighborhoods are essential, which is why it is important to give an overview of how Watergraafsmeer is divided in terms of population, housing, area etc. As stated above, there are 36.000 residents in Watergraafsmeer: approx. 16.000 (45%) in Middenmeer, 12.000 (33%) in Frankendael, 5.000 (14%) in Overamstel and 3.000 (8%) in Betondorp (CBS, 2020). These numbers coincide with number of dwellings per neighborhood. To determine whether this an accurate representation of the population of Watergraafsmeer the percentages can be compared to the percentages of 150 respondents per neighborhoods: 65 in Middenmeer (43%), 35 in Frankendael (24%), 28 in Betondorp (19%) and 22 in Overamstel (14%). Because most percentages are close to the same number it can be stated that the population of Watergraafsmeer is well-represented. Furthermore, Watergraafsmeer has an area of approx. 11 km2, with Overamstel being the largest and Betondorp the smallest in surface-area. Noticeable herein is that Overamstel mostly consists of industrial area, which is why this is the neighborhood with the least number of residents. Materials

In order to obtain results a survey was issued amongst the residents of Watergraafsmeer. This survey was created using the survey-software Qualtrics. Within this program a survey was composed, consisting of 22 questions (see appendix 1). The first three questions were used to gather

background information on participants (neighborhood, age, house ownership). The following five questions were used to test people’s knowledge on green roofs, ranging from knowing what green roofs are made of to knowledge on types of green roof. Then, four questions (one per benefit) were asked to see if environmental benefits of green roofs could be a driver for people to install one. After this, the public attitude question was asked to find out if people thought more green roofs needed to be implemented in Watergraafsmeer. The last four questions regarded willingness to implement and pay, making distinctions between extensive and intensive green roofs (see appendix 1 for the full survey). The survey was web-based so respondents could only respond via the internet. Different questions had different answer-formats. People could for instance fill in their age via pre-made categories and answer what neighborhood of Watergraafsmeer they live in or if they live in a bought or rented house. Most questions however consisted of a 5 point multiple choice, reasoning from the 5-point Likert-scale. The Likert scale was created to ‘measure’ attitude in a scientifically accepted manner in 1932 (Joshi et al., 2015) and is now widely used as answer-models for surveys. Within this research, a 5-point Likert is taken on. This decision was made for two reasons: (i) Jungles (2013) stated in his research that knowledge of green roofs is not substantial when comparing to other climate-change mitigation-options. Because people might not have a lot of knowledge about this subject it is important to give people a limited number of options. (ii) Relating to the first reason, within surveys it is important to keep people’s attention. A 7 or even 10-point scale might be a reason for people to rush or not finish the survey. Questions were presented as statements and answer options consisted of: totally agree, agree, neutral, do not agree, totally do not agree.

(13)

13 Procedure

Data-collection was done via surveys. Distribution of said surveys was accomplished via multiple methods. As stated before, to reflect the public as effectively as possible, and get an idea of the public attitude a minimum number of respondents was set as a goal. However, the more

respondents, the better. Therefore, the surveys were distributed in a multitude of ways. Firstly, a code (provided by Qualtrics) was printed out 850 times with a piece of text, explaining where the QR-code would lead you and what the research was about. This QR-QR-code was distributed evenly across the four neighborhoods by depositing said QR-codes in mailboxes of possible respondents houses, roughly 215 per neighborhood. Secondly, a call was placed in two different Facebook-groups regarding residents of Watergraafsmeer. This resulted in a steady flow of respondents for 3 or 4 days. Thirdly, the QR-code was spread out through local supermarkets. The survey was distributed on the 9th of April 2021 and closed 2 weeks later, on the 23rd of April 2021 and resulted in a total of 150 survey respondents, which is 50 more than anticipated and enough to reflect the public attitude of Watergraafsmeer.

Analysis-methods

In contrast to the data collection, the methods for analysis will differ in regard to the proposed sub-questions. The first, second and third sub-question rely on a possible relation between dependent and independent variables, which is why a statistical analysis will be carried out using the program R. The fourth sub-question regards a spatial analysis which is why the program ArcGIS Pro will be used to answer this question.

Cronbach’s alpha

When using statistical analysis to determine relationships between variables it is important that respondents’ answers are consistent, otherwise conclusions can be invalid. Cronbach’s alpha was developed to provide a measure of consistency (Tavakol, 2011). Alpha has a value between 0 and 1. The closer to 1 the value is, the more consistent the questionnaire is and the more reliable your data. Important to note is that Cronbach’s Alpha will be calculated for survey-questions that have the same subject and will be used as a value to determine if these questions represent the same item. If Cronbach’s Alpha proves to be sufficiently high, results of all four sub-questions can be seen as valid and reliable because respondents’ answers are considered consistent. Cronbach’s alpha will be calculated for all applicable variables and is represented by α.

Simple linear regression

To quantify relationships between dependent and independent variables a simple linear regression analysis is used in sub-question one, two and three. The output of a regression model shows the change of the dependent variable (public attitude) if the independent variable changes by 1 unit and therefore explains to what extent there is a relationship between the two variables. This change of the dependent variable is expressed in the following formula (Chatterjee, et al. 2015):

y = β0 + β1X+ ε.

y refers to the dependent variable and X is the independent variable. β0 is seen as the intercept term

and β1 is regarded as the slope parameter. ε is the error that explains the inability of the data to stay

(14)

14

Subquestion 1: To what extent does age affect public attitude toward green roofs?

The first sub-question will be answered using simple linear regression-analysis methods, specifically in R. The independent variable is age. The effect of this variable on the dependent variable of public attitude will be researched. The independent variable has a certain survey-question that represents and explains this variable. The survey question and possible answers for age is: (survey questions are translated from Dutch to English)

- What is your age-category? – (<18), (18-30), (31-45), (46-60), (61-75), (75). Because of the way respondents could answer, age can be seen as an ordinal data-type wherein the different categories were given the numbers 1-7 in R. Numerical data is needed in order to run a simple regression model.

The dependent variable is public attitude. Public attitude has one survey-question that represents the variable:

- I think more green roofs should be installed in Watergraafsmeer. (Totally agree), (Agree),

(Neutral), (Do not agree), (Totally do not agree). Public attitude can also be seen as an

ordinal variable and the answer option were therefore given the value 1-5. 5 being totally agree. Because the 5-point Likert scale was used in answering options this alteration is valid. Public attitude is represented by the same question in all four sub-questions and will not be typed out for every question.

Then, one simple regression will be run to test the relationship as stated above: 1. Age vs. public attitude.

Subquestion 2: To what extent does willingness to pay affect public attitude toward green roofs?

The second sub-question will be answered using the same method as for sub-question 1, again with linear regression methods. The dependent variable of public attitude will be tested for the

independent variable of WTP, which is a variable that is represented in the survey via two questions: - I am willing to install an intensive green roof if the government pays for it. (Yes), (No),

(Neutral), (Does not apply (rental house)). VS. I am willing to install an intensive green roof if I

have to pay for it myself. (Yes), (No), (Neutral), (Does not apply (rental house)). Because of answer options, WTP can also be seen as ordinal data and the answers were given numeric values: 1-3, 3 being yes. The last option was given the value of NA because this answer cannot be taken into account when running statistical analysis. For this variable the option for intensive green roofs was taken, because intensive green roofs have a higher price and could therefore depict a better image of WTP.

Then, two simple linear regressions will be run:

1. Willingness to implement if government pays vs. public attitude 2. Willingness to implement if self-pay vs. public attitude

The differences between outputs of the regression analyses will determine to what extent willingness to pay has an effect on the public attitude.

(15)

15

Subquestion 3: What environmental benefit of green roofs influences public attitude the most?

This second sub-question will also be answered using a regression analysis in R. Independent variables will describe the environmental benefits and one dependent variable will describe public attitude. The four survey-questions that explain the environmental benefits of green roofs are:

- Improved biodiversity would be a motive for me to install a green roof. (Totally agree),

(Agree), (Neutral), (Do not agree), (Totally do not agree)

- Lower temperatures in the city would be a motive for me to install a green roof. (Totally

agree), (Agree), (Neutral), (Do not agree), (Totally do not agree)

- Lower flood risks (green roofs retain water) would be a motive for me to install a green roof.

(Totally agree), (Agree), (Neutral), (Do not agree), (Totally do not agree)

- More efficient energy-use would be a motive for me to install a green roof. (Totally agree),

(Agree), (Neutral), (Do not agree), (Totally do not agree)

All four statements and answer options follow the 5-point Likert-scale. These variables are ordinal and therefore can be given the numeric value of 1-5 within the program R. 1 being totally do not agree and 5 being totally agree.

First, all independent variables will be combined into one variable to establish if environmental benefits as a whole have a significant relationship with public attitude and if it can be seen as a predictor of the dependent variable. Then, a simple linear regression will be run:

1. Environmental benefits vs. public attitude

Then, to determine what motive is the biggest driver of public attitude, four simple linear regressions will be run:

1. Biodiversity as motive vs. public attitude

2. Lower temperatures as motive vs. public attitude 3. Lower flood-risks as motive vs. public attitude

4. More efficient energy-use as motive vs. ublic attitude

Subquestion 4: To what extent do differences between the four neighborhoods of Watergraafsmeer, such as number of green roofs, income and percentage of house ownership, have an effect on the public attitude?

In contrast to the first two sub-questions this one will be answered using geographical information science (GIS), namely the program ArcGIS Pro. A program that is used for the visualization of data and allows you to convert raw data into actionable information from which you can create maps for instance. Answers from survey-questions and other data will be compared between the four neighborhoods of Watergraafsmeer, resulting in three visual maps that show differences, via spatial analysis. The survey-questions and consequently control-variables that result in these maps are: (questions are again translated from English)

- What type of house do you live in? (Bought), (Rented), (Rather not say). - Average income per neighborhood (retrieved from CBS, 2019)

- Number of green roofs per neighborhood (retrieved from the municipality of Amsterdam, 2020)

These three variables will be visualized per neighborhood, together with the public attitude, using ArcGIS Pro whereafter they will be combined in three different maps to determine if there is correlation between the variables. To be able to make public attitude a visible variable the mean of

(16)

16

the answers will be used. Per neighborhood there will be a percentage that lives in rented housing and a percentage that lives in bought housing. Average income is expressed in euros per year and the number of green roofs can be visualized via point-data. Steps taken can be seen in the flowchart presented in appendix 2.

Results

Within this part of the research results are presented and conclusions will be drawn per sub-question. In the first sub-question age will be tested against public attitude (via a simple linear regression) to determine whether there is a significant relationship between the two variables. Secondly, the variable of WTP will be tested with public attitude, to establish a possible relationship. Then, to see if environmental benefits of green roofs are a predictor of public attitude a simple linear regression will be run, after which every environmental benefit individually will be tested against public attitude, to determine the biggest driver. Lastly, a spatial analysis (via visualization in map-form) will determine whether any correlation can be found between the number of green roofs, income and percentage of house ownership and the dependent variable of public attitude (mean). All variables in the spatial analysis are displayed per neighborhood. After calculating different simple linear regressions and producing three maps the main research question is answered.

Age and Public attitude

To determine to what extent age has an effect on the public attitude of green roofing one simple linear regression is run (as stated in the methodology). In order to understand and interpret regression analysis’ better it is important to first give an overview/summary of used data. Age

Age was divided in 6 classes: (1) <18, (2) 18-30, (3) 31-45, (4) 46-60, (5) 61-75, (6) >75. Figure 2 shows a bar plot of the age variable:

(17)

17

The bar plot shows that most respondents are middle-aged. Age classes of under 18 and over 75 are not represented well with only two respondents per class. To determine the average age a summary is shown in the table below: (the variable was made numeric for the summary; 1 corresponds to <18 and 6 to >75)

Minimum 1st quarter Median Mean 3rd quarter Maximum Cronbach’s alpha

1.0 3.0 4.0 3.764 4.0 6.0 NA

Table 2.

The most important aspect of this table is the mean: 3.764, roughly corresponding to an average age of 43. According to the CBS (2020) the average age in WGM is approximately 38. Therefore, the average age of survey-respondents corresponds well with the actual average age in WGM and is represented by the questionnaire. Cronbach’s alpha is not calculated for the variable of age, since this is a value that needs to be calculated for a set of items and cannot be done for one variable. Public attitude

Public attitude is expressed in the statement: I think more roofs should be installed in

Watergraafsmeer. Answer options ranging from totally disagree (1) to totally agree (5). Figure 8 shows how this public attitude is divided across the respondents:

Figure 3.

In the bar plot it can be seen that most people have a positive public attitude towards the implementation of more green roofs in Watergraafsmeer. To determine the mean a summary is shown in the table below:

Minimum 1st quarter Median Mean 3rd quarter Maximum Cronbach’s alpha

1.0 4.0 5.0 4.271 5.0 5.0 NA

(18)

18

The table shows that the average public attitude is very high, namely almost 4.3 on a 5-point scale, resulting in a very positive public attitude in WGM.

Regression analysis age vs. public attitude

To determine the relationship between age and public attitude and to see if age significantly predicts public attitude a simple linear regression is run in R:

Public attitude Predictors Estimate p Age -0.09 0.258 Observations 150 R2 0.009 Table 4.

The first value to look at when interpreting a simple linear regression is the significance value. This level is the probability of rejecting the null hypothesis when it is true (Labovitz, 1968). The null hypothesis within this test is the proposition that there is no relationship between age and public attitude. To be able to reject a null hypothesis the significance level has to be below 0.05, a value that was set by Labovitz in 1968 and still the widely accepted significance level. The program of R does not give a value for significance, but a p-value, a measure of probability that an observed difference could have occurred by random change. If this p-value is below the significance level of 0.05, the null hypothesis can be rejected. However, as can be seen in table 4, the p-value is 0.258. The null hypothesis can therefore not be rejected and it cannot be said that a definite relationship between age and public attitude towards the implementation of green roofs exists.

In the theoretical framework Reinhart (2018) stated that younger people tend to be more concerned regarding consequences of climate change than older people. This can however not be confirmed within this research due to a too high p-value. A reason for this could be the skewedness of the age-variable as can be seen in figure 5. Even though the average age roughly corresponds with the actual average age in Watergraafsmeer, the survey is skewed towards the older generation, possibly causing the high p-value. Reinhart (2018) defines young people as 18-35, but if these people are not represented well in the survey it is difficult to draw conclusions. To confirm or deny the link further research is needed.

Willingness to pay and public attitude

The second relationship researched is between the independent variable of willingness to pay and the dependent variable of public attitude. A simple linear regression will be the end product and show to what extents a relationship exists. First, Cronbach’s alpha was calculated for the two statements of WTP together to determine if the two questions actually represent WTP and if questions were interpreted accordingly:

Chronbach’s alpha (WTP) 0.75

Table 5.

The value for Chronbach’s alpha is 0.75 and it can therefore be stated that the variables representing WTP are considered to be answered consistently and therefore reliable.

(19)

19

Then, a bar plot displays how willingness to pay if subsidized is divided across respondents:

Figure 4.

This bar plot shows the spread of answers to the question if people are willing to implement a green roof if the government should pay for it. It can be stated that people on average are inclined to implement a green roof if subsidized by the government. The following table gives a further overview of how the variable is spread:

Minimum 1st quarter Median Mean 3rd quarter Maximum

1.0 2.0 3.0 2.465 3.0 3.0

Table 6.

Table 3 shows (as expected) that the mean of answers is 2.465, confirming the statement that generally, people in WGM are prepared to implement a green roof if subsidized.

In order to determine whether WTP is a predictor of public attitude a differentiation will be made between willingness to implement if subsidized and willingness to implement if self-paid. The figure below shows the spread of the latter:

(20)

20

Figure 5.

When comparing the two bar plots of WTP it is visible that there is a difference in WTP. If people have to pay for the green roofs themselves they are far less willing to implement. This can also be seen in the summary:

Minimum 1st quarter Median Mean 3rd quarter Maximum

1.0 1.0 1.0 1.604 2.0 3.0

Table 7.

On a scale of 3 the mean now is 1.604 against the 2.465 of the first summary. It can be stated that the overall willingness to implement is significant but if people have to pay for it themselves willingness declines rapidly.

Regression analysis willingness to pay vs. public attitude

To determine if WTP can predict public attitude and if there is a relationship between the two variables two simple linear regression analyses are run. Firstly, a regression is run regarding the first question and public attitude (see table 8).

Public attitude

Predictors Estimate p

Willingness to implement if subsidized 0.43 0.00019

Observations 150

R2 0.094

Table 8.

The first thing to notice is a p-value of 0.00019, a value that is significantly lower than the proposed significance level of 0.05. The null hypothesis can therefore be rejected and it can be said that there is a (strong) relationship between the two used variables. However, the value for R2 is considered very low, resulting in low explanatory power of the model. This has no immediate consequences for possible conclusions, but is important to keep in mind when looking at results. The second (and most

(21)

21

important) parameter to look at is the estimate value of 0.43, a value that represents the β1

coefficient. Therefore, if the answer to the statement that people are willing to implement a green roof if the government pays for it goes from (for instance) neutral (2) to yes (3) the answer to the statement that people want more green roofs in Watergraafsmeer generally goes up by 0.43. It can therefore be said that there is a relevant relationship between the independent and dependent variable and that willingness to implement if subsidized is a predictor for public attitude.

To determine a possible relationship between WTP and public attitude a second regression model is run:

Public attitude

Predictors Estimate p

Willingness to implement self-payed 0.18 0.057

Observations 150

R2 0.025

Table 9.

As with age vs. public attitude, this regression model has a p-value that is higher than the proposed significance level of 0.05. The null hypothesis that willingness to implement if self-pay has a

relationship with public attitude can therefore not be rejected with certainty.

Overall, WTP cannot be seen as predictor of public attitude, because no distinction can be made between willingness to implement if subsidized and self-pay. Both regression models also have very low R2-values resulting in low explanatory power. Important to notice nevertheless is the difference that can be seen in the previously shown histograms: 80 people are willing to implement if the government pays for it, to only 30 if they have to pay themselves. There is definitely a difference in willingness to pay among respondents. Relating back to the literature, Jorgensen (2000) explained that attitude towards an environmental intervention tends to become more negative when this intervention is beyond a respondent’s WTP. Even though WTP is definitely an existing factor when it comes to green roofs in Watergraafsmeer, the proposed relationship of Jorgensen (2000) cannot be confirmed within this research. A reason for this could be that the survey-questions that represented WTP regarded implementation of intensive green roofs. Intensive green roofs are bigger in size and require more maintenance, something that might have played a role in how people chose to answer the survey-questions, and therefore might have had an influence on the results. Furthermore, both regression-models have very low R2-values, resulting in low explanatory power of the models and a possible cause for significance levels.

Environmental benefits and Public attitude

The third question aims to answer what environmental benefit of green roofs influences public attitude the most. These environmental benefits are explained as improved biodiversity, reduced temperatures, reduced flood-risk and higher energy-efficiency. Respondents were asked to what extent either of these benefits would be a motive for installing a green roof. All answers reason from a 5-point Likert scale and are altered to display numeric answers: from totally disagree (1) to totally agree (5). First, all four variables concerning environmental benefits as motive are concatenated into one variable to learn to what extent environmental benefits are related to public attitude. When adjoining multiple variables, the mean is taken from all four variables and taken on by the one variable representing environmental benefits. Therefore no bar plot could be made but only a summary of the variable as descriptive statistic:

(22)

22

Minimum 1st quarter Median Mean 3rd quarter Maximum Alpha

1.50 4.00 4.25 4.26 5.00 5.00 0.89

Table 10.

Table 10 shows that environmental benefits of green roofs are considered major motives for potential implementation. Most people agree or totally agree. Also, a value for Cronbach’s alpha of 0.89 was found, meaning that respondents to the survey answered the questions regarding

environmental benefits consistently and that the questions represent the item accurately.

Now a regression is run between the variable representing environmental benefits as motive and the dependent variable of public attitude to establish to what extent there is a relation between these two variables: Public attitude Predictors Estimate p Environmental benefits 0.96 <0.0001 Observations 150 R2 0.6 Table 11.

Due to the low p-value the null-hypothesis can be rejected and it can be stated that there is a relationship between environmental benefits as motive to implement and the public attitude. The estimate value is 0.96 which indicates that this relationship is also very strong. Reinhart (2018) and Jorgensen (2000) argued that there might be a link between environmental values and the attitude towards the intervention, a connection that can be strongly confirmed by these results. The strength of this relationship can be attributed to the general concern for the environment of people in Watergraafsmeer. A statement also asked in the survey was: “I am concerned with the consequences

of climate change”. Because answers to this statement might partially explain the strength of the

relationship, a summary of answers to this question is given in the table below:

Minimum 1st quarter Median Mean 3rd quarter Maximum

1.00 4.00 5.00 4.48 5.0 5.0

Table 12.

The answers to this statement were also following the 5-point Likert scale, meaning that the average resident of WGM is very concerned with the consequences of climate change (mean of 4.48), a possible reason for the previously found relationship.

Environmental benefits separately

Originally, the environmental benefits as motive were four separate variables. These are used to find out what benefit is the largest predictor for the public attitude. Therefore, descriptive statistics are displayed to give an overview of how these individual variables are spread across the respondents:

(23)

23 Figure 6.

These bar plots confirm the notion that environmental benefits can be seen as a significant motive for people to implement a green roof. Additionally, the mean, median and other factors were calculated for the four motives:

Minimum 1st quarter Median Mean 3rd quarter Maximum

Biodiversity 1.0 4.0 5.0 4.299 5.0 5.0

Temperature 1.0 4.0 4.0 4.153 5.0 5.0

Flood-risk 2.0 4.0 4.0 4.253 5.0 5.0

Energy-efficiency 1.0 4.0 4.5 4.347 5.0 5.0

Table 13.

The mean is the highest for the motive regarding better energy efficiency, meaning that that could be the biggest environmental driver in the implementation of green roofs. To find out relations between environmental benefits and the public attitude four regression analyses are run: one for every environmental benefit.

(24)

24

Higher biodiversity and Public attitude

Public attitude

Predictors Estimate p

Higher biodiversity as motive 0.76 <0.0001

Observations 150

R2 0.540

Table 14.

Due to the very low p-value it can be stated that there is a strong relationship between higher biodiversity as motive and public attitude. This relationship is expressed in the estimated value of 0.76. This means that if answers to the statement that higher biodiversity is a motive for installing a green roof goes up by on the 1-5 scale the unit of public attitude rises with 0.76 on a 1-5 scale, indicating the fact that higher biodiversity as a motive can be seen as a predictor for public attitude.

Lower temperatures vs Public attitude

Public attitude

Predictors Estimate p

Lower temperatures as motive 0.63 <0.0001

Observations 150

R2 0.432

Table 15.

As can be seen in table 15, lower temperatures as a motive also has a strong relationship with the public attitude, resulting from the low p-value and the relatively high estimate-value. Lower temperatures as a motive for installing can therefore also be seen as a predictor for public attitude. However, the relationship is weaker than in the first regression-model regarding environmental benefits.

Lower flood risks vs Public attitude

Public attitude

Predictors Estimate p

Lower flood-risks as motive 0.75 <0.0001

Observations 150

R2 0.420

Table 16.

With a p-value of below 0.0001 it can be said that there is a strong relationship between lower flood risks as a motive for installing green roofs and public attitude. The relationship is again expressed in the estimated value: 0.75 . Lower flood risks as a motive for installing can therefore be seen as a stronger predictor for public attitude than lower temperatures in the city but slightly weaker than a higher biodiversity.

Better energy efficiency vs Public attitude

Public attitude

Predictors Estimate p

Better energy efficiency as motive 0.78 <0.0001

Observations 150

R2 0.437

(25)

25

With a p-value of below 0.0001 it can be said that there is a strong relationship between better energy efficiency as a motive for installing green roofs and public attitude. The relationship is again expressed in the Estimate-value: 0.78 . Better energy efficiency as a motive for installing can therefore be seen as the largest predictor for public attitude.

All four regression models have a very low p-value, meaning that environmental benefits overall have a very strong relationship with public attitude. Estimate values of the regression models lie between 0.6 and 0.8. Because of the fact that in all regarded statements the 5-point Likert scale was used these values can be considered as relatively high, resulting in strong relationships. The highest estimated value was that of the variable regarding a better energy-efficiency, namely 0.78, resulting in the largest predictor of public attitude. The outcomes of the regression models confirm the previously explained theory of Patchen (2006) regarding personal feedback. The largest predictor of public attitude is better energy-efficiency, a concept that includes immediate consequences for the person installing a green roof, namely lower costs. The weakest relationship is between lower temperatures in the city and public attitude, which also can be attributed to the same theory. Even though reduced temperatures are considered the most important benefit of green roofing, a 1 0C (exemplar number) temperature-drop is not noticeable for the individual. It can therefore be stated that personal feedback that is received from installing a green roof is a very important factor in change in public attitude towards implementation.

Differences between neighborhoods and public attitude

In order to determine how differences between neighborhoods affect public attitude multiple spatial analyses are used, containing the following variables: amount of green roofs in Watergraafsmeer (Gemeente Amsterdam, 2020) , average income (CBS, 2019), percentage house ownership (survey) and the mean public attitude (survey). To perform the spatial analyses the first three variables are displayed in maps along with the mean of the public attitude. First, a bar plot is presented to show how the respondents are spread across Watergraafsmeer’s neighborhoods:

(26)

26 Figure 7.

Figure 7 shows that most respondents originate from Middenmeer and the least from Overamstel. This correlates with the number of dwellings per neighborhoods so this can be seen as an accurate representation of residents per neighborhood.

Then, to see if there is a correlation between the number of green roofs and the public attitude both variables are presented in map 1:

Map 2.

According to the municipality of Amsterdam there are 39 green roofs in Watergraafsmeer; figure 15 shows how they are spread across different neighborhoods. Betondorp has no green roofs,

Overamstel has 4, Frankendael 17 and Middenmeer has the most with 19 green roofs installed. These numbers can be attributed to the number of dwellings within every neighborhood: Betondorp has 1960, Overamstel has 3530, Frankendael has 6116 and Middenmeer has the most with 8000 (CBS, 2020). It is logical that the neighborhood with the highest number of houses also has the highest number of green roofs. The expectation is that the neighborhood with the highest number of green roofs would also have the highest average public attitude; this is however not the case.

Therefore it can be stated that the number of green roofs in a neighborhood is not positively correlated with the public attitude. This could be due to a multitude of reasons. Firstly, the number of green roofs could be attributed to the number of dwellings per neighborhood, therefore exist by chance and not have any relation with public attitude. Other external factors could also play a role, such as average income or percentage of house ownership per neighborhood, which will be discussed in the following two maps.

(27)

27 Map 3.

The average annual income per neighborhood differs highly, as can be seen in figure 18. From 23.000 in Betondorp to 33.900 in Middenmeer. There is however hardly any correlation between income and public attitude. The neighborhood with the lowest income is 2nd in public attitude, while the neighborhood with the highest income is lowest in public attitude. It can therefore be stated that within this research average income is not a factor that influences public attitude. Jorgensen (2000) stated in his research that survey-respondents with a lower income may consider it unfair that they have to pay for the intervention. On the other hand, higher income respondents might deem it unfair and feel that they might be targeted to pay a larger proportion of the cost. The main reason for the fact that no relationship was found and therefore Jorgensen’s theory could not be confirmed is that the income-variable is an average, taken from the Central Statistical Bureau (CBS), resulting in possible faulty income values that might not correlate with the income of the actual respondents. The last variable that is discussed, and might, partially, explain the difference in number of green roofs and public attitude is percentage of house ownership within the different neighborhoods (map 4).

(28)

28 Map 4.

Map 3 shows that percentage of house ownership differs highly between neighborhoods. Of the respondents in Overamstel more than 70% of the people live in bought housing versus only a little more than half in Betondorp. There is a weak correlation between the two variables. Overamstel has the highest public attitude of all neighborhoods and the highest percentage of house ownership. Frankendael has the 3rd highest public attitude and is 2nd in bought housing and Middenmeer has the lowest public attitude and the 3rd highest percentage of bought housing. Because of this it can be stated that there is, to some extent, a positive correlation between public attitude per neighborhood in Watergraafsmeer and percentage of house ownership. Public attitude towards green roofing might therefore be more positive if people live in a bought home, rather than a rented home. Because no definitive link can be established, and house ownership is a variable that was part of the survey, for this particular spatial analysis a statistical analysis is carried out as well, to be able to fully confirm or deny the link. This spatial analysis will be in the form of a simple linear regression

between the variables of house ownership and public attitude and results in the following table:

Public attitude Predictors Estimate p House ownership -0.29 0.063 Observations 150 R2 0.02 Table 18.

The p-value of the regression model is 0.063 and the null hypothesis cannot be rejected, meaning that a relationship between house ownership and public attitude cannot be confirmed via a statistical analysis. The spatial analysis shows some correlation, but further research is needed to fully confirm or deny the link.

(29)

29

Conclusion

The consequences of climate change are becoming a more relevant topic for cities around the globe, and measures need to be taken in order to mitigate these. One of these measures is green

infrastructure, namely green roofs. A plan for green roofs was not implemented in the neighborhood of Watergraafsmeer, Amsterdam, due to too little public support. Therefore, the main goal of this research was to establish what factors (and to what extent) influence the implementation of green roofs in Watergraafsmeer, Amsterdam. To answer this a survey was conducted across

Watergraafsmeer, totaling 150 respondents. In the theoretical framework multiple possible

relationships between concepts and public attitude were proposed, such as: age, willingness to pay, environmental benefits of green roofs and factors that relate to differences between neighborhoods, such as income, number of green roofs and percentage of house ownership.

Firstly, when examining an overview of the data it can be stated that the overall public attitude towards the implementation of more green roofs in Watergraafsmeer is extremely positive,

corresponding to an average public attitude of 4.271, on a scale of 1-5. The overall consensus seems to be that people think more green roofs should be installed. Then, the question of what factors influence the public attitude was explored.

A simple regression was run using age as independent and public attitude as dependent variable. Due to a high p-value the null hypothesis could not be rejected. Even though research suggests that younger people are generally more concerned with the environment, this link could not be established within this particular research. Reasons for this could be the skewedness of the age variable or distribution type of the survey.

Then, the variable willingness to pay was tested to find out a possible relationship in regard to public attitude. Two regression models were run: (i) willingness to implement if subsidized vs public attitude and (ii) willingness to implement if self-pay vs public attitude. A relationship was found in the first model, but not in the second model (high p-value). Because the aim was to be able to differentiate between these two models, no definite conclusion can be attached to this proposed relationship and earlier research could not be confirmed.

The second sub-question aims to answer if environmental benefits of green roofs can be seen as a predictor for public attitude and what benefit of green roofs is the largest. Five simple regressions were run to determine this. The first thing to notice is that the environmental benefits of green roofs overall are a driver for the public attitude, with a low p-value and an estimate value of 0.96. The separate regression models showed estimate values between 0.6 and 0.8 and very low p-values, generally considered as signs of a significant relationship. The lowest estimate value was that of lower temperatures in the city as a motive, and the highest calculated estimate value was that of better energy efficiency as a motive. A difference that can be explained via the immediate feedback people receive from these benefits. Lower temperatures in the city are a phenomenon that people maybe view as “not a big deal”, something they ultimately do not tie to the dangerous consequences of climate change. Better energy efficiency for your own house is something that directly benefits respondents in the form of lower bills for instance.

Lastly, three spatial analyses were conducted to determine if differences amongst neighborhoods affect the public attitude in different ways. Firstly, the spread of already installed green roofs was mapped versus the public attitude and no immediate correlation was found. This could be due to the amount of people that responded to the survey or the number of houses/apartments there are in a certain neighborhood. The second spatial analysis was conducted between income and public

(30)

30

attitude, where the same type of findings was established. Mainly because the area with the lowest public attitude mean has the highest income, and the area with the 2nd highest public attitude has the lowest income. These results might be related to the fact that the average income was taken from CBS and not directly surveyed. The last spatial analysis however shows some correlation. The percentage of bought housing per neighborhood was mapped out together with the public attitude. The area with the highest public attitude, Overamstel, also has the highest percentage of bought housing, resulting in the conclusion that people with bought housing often have a more positive public attitude towards the implementation of green roofing than people in rented housing. A relation that could not be confirmed further by an additional statistical analysis.

Concluding, factors that influence the public attitude are mainly the environmental benefits of green roofs and whether or not people live in bought houses, to some extent. The overall public attitude is very high, with a mean of 4.271 in Watergraafsmeer, while the highest public attitude is present in the neighborhood of Overamstel (almost 4.5) and the lowest in Middenmeer (4.1). While literature suggests that factors relating to background influence public attitude, this does not seem to be the case. Parameters such as age or income do not influence the public attitude towards the

implementation of green roofs in Watergraafsmeer.

Finally, some further research is proposed. In the introduction and methodology, a conceptual model was given, consisting of links and factors that might have an influence on public attitude. Read research suggest however that public attitude might have an effect on the implementation of green roofs, resulting in a feedback-loop. Because this research was survey and GIS-based there was no room for a documentation-analysis that confirms or denies this link. If that link exists the conceptual model would be as the one in figure 20:

Referenties

GERELATEERDE DOCUMENTEN

De voordelen van EMIGMA zijn een centrale database met gegevens (trenddata, maar ook, referentiewaarden, streefwaarden, ondergrenzen, bronvermelding en rapporten), snelle

Furthermore, there has been no comprehensive study in this region to analyze the influence of green roof design parameters such as LAI, plant height, plant albedo, soil

Respondent Maar goed, dat is voor heel veel mensen -daar is subsidie een hele mooie manier voor om mensen toch over de streep te trekken denk ik. Subsidie Interviewer

In addition, the number of macrophages positive for TRAP activity was higher in lung tissue from patients with fatal asthma as compared to control subjects (Fig.  3c ).. The number

The refractive index change induced by the electronic Kerr effect will result in a larger resonance frequency shift in comparison to the cavity linewidth with high-Q cavities..

Volgens Troskie, in ’n onderhoud met Cassie Carstens in Vir die Musiekleier (Carstens 2011:45), het hy tydens streekkursusse wat deur SAKOV aangebied is besef dat daar

The effect of independent variables Fear, Attitude and Subjective norm on the variable, intention to seek information was measured using a hierarchical linear regression..

A threefold methodology was used that consisted of (1) an extensive review of scientific literature complemented by (2) a thorough examination of policy documents