RESIDENTIAL SOLAR POWER IN UPSTATE NEW YORK:
AN ANALYSIS OF THE KEY DRIVERS OF ADOPTION
By
James C. Letzelter
A Doctoral Thesis
Presented to the Graduate Faculty of the Doctor of Law and Policy Program
at Northeastern University
In partial fulfillment of the requirements for the degree of
Doctor of Law and Policy
Under the supervision of Dr. Afi Wiggins and Dr. Chelsea Schelly
College of Professional Studies
Northeastern University
Boston, Massachusetts
August 7, 2018
2
©James C. Letzelter, 2018
3
DEDICATION
This thesis is dedicated to the lights of my life, Hannah and Emma, who have inspired me
and whom I hope to inspire…and to Kate who has always believed in me and made me believe
in myself.
4
ACKNOWLEDGEMENTS
First and foremost, I wish to thank Dr. Nancy Pawlyshyn for her tireless efforts in
assisting with every phase of producing this thesis. I also thank my advisor, Dr. Afi Wiggins, for
her expertise in research and writing and her positive guidance. I also thank my second reader,
Dr. Chelsea Schelly, whose research inspired mine and whose expertise was invaluable to my
own research. I wish to thank Dr. Tricia Gonzalez and NYSERDA for their invaluable assistance
to my data collection. Finally, I wish to thank John Antonuk, by whom I am employed, and
whom I look up to as a world class management consultant, for his generous support of this
educational endeavor.
5
ABSTRACT
This research project explores the decision factors that influenced the adoption of residential
solar power systems in upstate New York. New York has a Clean Energy Standard (CES) that
requires 50 percent of electric energy in the State be supplied by renewable resources, which
includes solar electricity, by 2030. Currently, and into the near future, New York supports the
solar industry by providing incentives in the form of subsidies and financing for solar power
systems. The program, NY-Sun, is administered by the New York State Energy Research and
Development Authority (NYSERDA). However, the subsidies are on a schedule of declining
magnitude and will be phased out by 2023. Thus, identifying the most important decision factors
may be useful in understanding potential means of promoting solar technology adoption in the
context of declining economic incentivization.
Data for this research was collected through an online survey of homeowners in upstate New
York who have installed residential solar power systems. In addition to capturing demographic
data about the respondents (e.g., gender, age, education level, income), the survey asked
participants to rate the importance of decision factors that influenced their decision to adopt a
solar power system at home. The analysis provides a demographic profile of the adopters and a
descriptive statistical analysis of the importance of the decision factors. A correlation analysis
was used to determine how the adoption factors correlate with demographic factors. Finally, the
data was analyzed to identify how factors have changed over time. The goal of this research is to
contribute new information to the field for the benefit of policymakers and solar industry actors.
Keywords: solar power, solar energy, renewable energy, renewable portfolio standard, NY-Sun
6
TABLE OF CONTENTS
Chapter 1: Introduction to the Study ....................................................................................... 12
Background ............................................................................................................................. 13
Law and Policy Review .......................................................................................................... 16
Renewable Portfolio Standards (RPS) .............................................................................. 20
Conclusions ....................................................................................................................... 21
Problem Statement .................................................................................................................. 21
Purpose of the Study ............................................................................................................... 22
Research Question and Hypothesis ......................................................................................... 22
Assumptions ............................................................................................................................ 24
Scope and Delimitations ......................................................................................................... 25
Limitations .............................................................................................................................. 25
Significance............................................................................................................................. 26
Chapter 2: Literature Review ................................................................................................... 27
Literature Search Strategy....................................................................................................... 27
Research on Renewable Portfolio Standards .................................................................... 27
Research on Solar Economics ........................................................................................... 29
Research on Customer Adoption ...................................................................................... 32
Research on Environmental Benefits of Solar Power Systems ........................................ 34
Summary and Conclusions ..................................................................................................... 35
Chapter 3: Methodology ........................................................................................................... 37
Research Design and Rationale .............................................................................................. 37
Methodology ........................................................................................................................... 37
7
Positionality Statement ........................................................................................................... 39
Data Collection ....................................................................................................................... 41
Population ......................................................................................................................... 41
Sample Size Calculation ................................................................................................... 41
Recruitment ....................................................................................................................... 42
Ethical Procedures ............................................................................................................ 42
Data Analysis .......................................................................................................................... 45
Threats to Validity .................................................................................................................. 46
Summary ................................................................................................................................. 47
Chapter 4: Results ..................................................................................................................... 48
Results of the Analysis............................................................................................................ 48
Demographic Data ............................................................................................................ 48
Descriptive Analysis: Decision Factors ............................................................................ 61
Correlations of Decision Factors to Demographics .......................................................... 73
Decision Factor Ratings Over Time ................................................................................. 76
Summary ................................................................................................................................. 78
Chapter 5: Conclusions and Recommendations ..................................................................... 83
Conclusions from the Data...................................................................................................... 83
What are the demographic parameters that describe the upstate New York residential
solar power system adopters? ........................................................................................... 84
What were the most important decision factors? .............................................................. 87
How do the importance ratings of the decision factors correlate to demographic
parameters? ....................................................................................................................... 89
8
How have the decision factors changed in importance over time? ................................... 89
Discussion ............................................................................................................................... 90
Relevance of Conclusions to the Research Questions ...................................................... 90
Relevance of Conclusions to the Literature ...................................................................... 95
Relevance of Conclusions to the Theoretical Framework ................................................ 95
Recommendations from the Study .......................................................................................... 96
Professional Use by Solar Installers ................................................................................. 96
Policy Recommendations.................................................................................................. 96
Recommendations for Further Research ........................................................................... 96
Conclusion .............................................................................................................................. 97
References .................................................................................................................................... 99
Appendix A: The Survey .......................................................................................................... 105
Appendix B: Data Cleaning ..................................................................................................... 129
Appendix C: NY-Sun Incentives ............................................................................................. 130
9
LIST OF TABLES
Table 1.1 Capacity of Solar Projects Added in New York in Megawatts (MW) .......................... 16
Table 4.1 Summary of Survey Respondents ................................................................................. 49
Table 4.2 Importance Ratings of Decision Factors and Their Ordinal Rank ............................. 62
Table 4.3 Other Decision Factors and Their Mean Importance Ratings .................................... 64
Table 4.4 Frequency of Decision Factors Selected as “Three Most Important” ........................ 65
Table 4.5 Decision Factor Importance Ratings by Race ............................................................. 66
Table 4.6 Decision Factor Importance Ratings by Education .................................................... 67
Table 4.7 Decision Factor Importance Ratings by Age ............................................................... 67
Table 4.8 Decision Factor Importance Ratings by Household Income ....................................... 68
Table 4.9 Decision Factor Importance Ratings by Home Value ................................................. 69
Table 4.10 Decision Factor Importance Ratings by Climate Change Beliefs ............................. 70
Table 4.11 Decision Factor Importance Ratings by Political Affiliation .................................... 71
Table 4.12 Decision Factor Importance Ratings by Gender ....................................................... 72
Table 4.13 Decision Factor Importance Ratings by Marital Status ............................................ 73
Table 4.14 Correlation between Decision Factors and Key Demographics ............................... 74
Table 4.15 Rule of Thumb for Interpreting the Size of a Correlation Coefficient ....................... 76
Table 4.16 Decision Factor Importance by Year of Installation ................................................. 77
Table 5.1 Summary of Demographic Findings and Conclusions ................................................ 87
Table B.1 Data Exclusions ........................................................................................................ 129
Table C.1 NY-Sun Incentives ..................................................................................................... 130
10
LIST OF FIGURES
Figure 4.1. Solar Adoption by Race. ............................................................................................ 51
Figure 4.2. Solar Adoption by Education. .................................................................................... 52
Figure 4.3. Solar Adoption by Age. ............................................................................................. 53
Figure 4.4. Solar Adoption by Household Income. ...................................................................... 54
Figure 4.5. Solar Adoption by Home Value. ................................................................................ 56
Figure 4.6. Solar Adoption by Climate Change Beliefs. .............................................................. 57
Figure 4.7. Solar Adoption by Political Affiliation. ..................................................................... 58
Figure 4.8. Solar Adoption by Gender. ........................................................................................ 60
Figure 4.9. Solar Adoption by Marital Status............................................................................... 61
Figure 4.10. Rated Importance of Decision Factors. .................................................................... 62
Figure 4.11. Importance of Peer Recommendations for Solar Over Time. .................................. 78
Figure 4.12. Importance of Solar Installer Reputation Over Time............................................... 78
Figure 4.13. Chain of Evidence. ................................................................................................... 81
11
LIST OF ACRONYMS
CES—Clean Energy Standard
GWh—Gigawatt hour
ITC—Investment Tax Credit
kWh—kilowatt-hour
LCOE—Levelized cost of electricity
NYPSC—The New York Public Service Commission
NYSERDA—The New York State Energy Research and Development Authority
PV—Photovoltaic
REV—Reforming the Energy Vision
RPS—Renewable Portfolio Standard
WACC—Weighted average cost of capital
12
Chapter 1: Introduction to the Study
“The power of the sun’s radiant energy is what makes life on earth possible. Efforts to
harness it in concentrated form and direct it to man’s ends have long been a human pursuit”
(Solar, 2018). Indeed, the photovoltaic effect, the means of harnessing the sun’s power by
transforming it into electricity, was first discovered in 1839 (Energy Education, 2018). However,
it was not until the most recent two decades that solar power systems became widely utilized for
electrical power production at both the residential and utility scale in the United States.
By providing distributed electric power produced at the site of use, residential solar
power systems partially displace the demand for electricity from utility companies. These
companies produce energy from a variety of sources, mostly through the combustion of fossil
fuels. This fossil fuel usage (especially coal) results in several kinds of harmful air emissions,
including greenhouse gases that scientists attribute to climate change (Coal and Air Pollution,
2018). Thus, solar power systems help to mitigate climate change.
At the national level, the U.S. provides financial incentive through a federal investment
tax credit (ITC) to encourage residential solar power system adoption. Individual states vary
widely in whether and how much they incentivize or subsidize installation of solar electric
systems. The state of New York has invested heavily in the promotion of solar power systems by
subsidizing solar power system costs to residential, commercial, and industrial energy users.
These incentives are spurred by aggressive targets for renewable energy production in the State,
of which solar is a component.
To add to the knowledge base on solar adoption, the goal of this quantitative research was
to examine what decision factors were most important to adopters in their decision to adopt,
13
focusing on a specific segment of the market. The research was based on a survey of upstate
New York (for practical purposes, all of New York except for New York City and Long Island)
residential solar adopters who received state-level incentives which are displayed in Appendix C.
In doing so, this research identified the demographic profiles of these constituents and the most
important factors to this group in their decisions to adopt solar electric technology at the
residential scale.
Background
The renewable energy industry in New York State has benefitted from the economic
incentives provided at the state level that subsidize the capital cost of solar power systems. This
enables solar installers to provide equipment that competes economically with traditional utility
(grid) power while providing environmental benefits. However, these economic incentives are
scheduled to decline and then disappear entirely by 2023. The decline of the incentives could
make continued adoption of solar power systems more of a challenge, thereby warranting further
investigation.
Compounding the problem of declining subsidies is that New York has a Clean Energy
Standard (CES) that requires 50 percent of its electric energy to be produced by renewable
resources, of which solar power plays a key role. “Solar energy is a key component of Governor
Cuomo’s nation-leading commitment to 50 percent renewable energy by 2030…” (NYSERDA,
2018). Through NYSERDA’s NY-Sun program, New York has a target of three gigawatts (GW)
of solar capacity by 2023 (NYSERDA, 2018). Understanding what drives customer decisions
will help maintain solar adoption rates that support the CES. This research investigates the
decision factors that influence residential solar power system adoption in upstate New York.
14
Throughout this study, the term “decision factors” is used to denote parameters that influence the
adoption of solar power systems. This study investigated the role of various decision factors in
upstate New York residential solar power system adopters’ decisions to adopt solar power for
their homes.
New York’s CES is designed to have positive environmental impacts. Solar power
systems (and other renewable energy technologies) displace electricity produced by the
combustion of fossil fuels. In doing so, many environmental benefits are produced. Most notable
of the benefits in reducing fossil fuel combustion is the reduction of air emissions including
pollutants and greenhouse gases. These emissions include sulfur dioxide (which causes acid rain
and has been a major issue for ecosystems downwind of coal-fired power plants), nitrogen
oxides (which are linked to respiratory illness and smog formation), air toxics (such as mercury,
which is deleterious to human health), greenhouse gas carbon dioxide (which scientists have
attributed to climate change), and others (Letzelter and Chupka, 1999). Since solar power
systems displace fossil fuel-fired power plant output, they reduce the harmful air emissions
associated with those power plants. As such, environmentalists have successfully pushed for
government policies that foster solar power adoption both on the federal and state levels.
The degree to which states subsidize solar power varies widely. The incentives reduce
capital costs for solar power systems, thereby making the overall cost of solar power lower and
more competitive with utility-provided electricity. In New York, the incentives are on a
declining schedule of value that decreases the subsidy as blocks of the incentive are taken by
consumers. Since NY-Sun began, the incentives have been high enough to reduce the cost of
solar power systems to a level that can make the total cost of solar power less than utility-
supplied electricity. Accordingly, solar power systems provide both economic and environmental
15
benefits, and are a desirable and commercially successful technology in New York, as shown by
the growth in solar energy in the State from 84.7 GWh in 2013 to 259.3 GWh in 2017 (NY-Sun,
2018a). Further, given the multiple benefits of solar technology, there may be multiple reasons to
choose to install a residential solar system. Those with a goal of helping the environment can
install a solar power system to reduce the fossil fuels burned by their electricity providers, while
people with an interest in saving money on the energy that is used in their home can install a
system to displace the higher-cost utility power.
Even before large solar incentives like those provided by NY-Sun, solar existed in New
York. However, it was not until the NY-Sun program launched that mass adoption in New York
took place. Once limited to the most avid and economically advantaged environmentalists, the
subsidies brought solar power systems to the masses by offering an environmentally friendly
technology that also brought cost savings to consumers. The policy benefited the recipients while
also enabling the pursuit of New York’s aggressive CES targets, as shown by the growth in solar
power between 2013 and 2017, displayed in Table 1.1.
16
Table 1.1
Capacity of Solar Projects Added in New York in Megawatts (MW)
Note. Reprinted from Data & Trends: Total Capacity, by NYSERDA, retrieved from
https://www.nyserda.ny.gov/All-Programs/Programs/NY-Sun/Data-and-Trends, Copyright 2018
NYSERDA.
Law and Policy Review
Utility regulation in New York is driven by the New York State Public Service
Commission (NYPSC) under the statutory powers granted to them. Also known as
administrative law, the primary design of utility regulation is to enforce customer protections
from a legal, natural monopoly (electric utilities). Regulated New York utilities can only charge
their customers rates that are approved by the NYPSC. For solar issues, the NYPSC works in
coordination with the New York State Energy Research and Development Authority
Year MW Added
2000 0.002
2001 0.033
2002 0.829
2003 0.987
2004 1.214
2005 1.428
2006 2.836
2007 3.863
2008 6.062
2009 12.042
2010 22.597
2011 26.798
2012 42.462
2013 74.464
2014 121.179
2015 200.018
2016 226.421
2017 222.985
2018 113.315
17
(NYSERDA) (NY-Sun, 2018). NYSERDA oversees state-level subsidies to offset the cost of
solar equipment and regulates the quantity and dollar value of solar subsidies according to a
schedule of blocks of subsidies that declines over time.
Research shows that states offering cash incentives to offset solar costs have more
success in the deployment of solar resources than states that do not (Sarzynski et al., 2012). This
suggests that as the subsidies in New York decline and, ultimately, end, it will be more
challenging to encourage adoption. Thus, it is important to understand what factors drive solar
adoption, which represents the connection between the law and policies surrounding solar power
and this research.
New York has committed to aggressive growth in the role that renewable energy plays in
the State’s overall energy resource portfolio. Through various statutory and regulatory initiatives,
New York is incentivizing the development of renewable energy projects at a rapid pace. While
there are many types of renewable energy resources, solar photovoltaic (PV) energy projects play
a key role in New York, as indicated by the launch of the NY-Sun program in 2012
(Comptroller, 2016). The adoption of solar, then, is one key factor in meeting New York’s
substantial renewable energy targets. This thesis focuses on the residential solar PV market in
upstate New York, exploring the motivators behind the choice individuals make to adopt solar
power systems.
The decision to adopt residential solar power systems in upstate New York is driven by
factors including demographics such as income and education. Economics is also a considered
factor, in the form of future expected energy bill reduction and up-front cost. Other factors
include peer influence, access to information, and environmental consciousness, discussed in the
literature review. Many of these decision factors are outside of the control of government
18
influence. However, one of the most important drivers—the cost of solar power—falls squarely
under the influence of government policy, because government agencies provide economic
incentives to defray solar costs (NY-Sun, 2018).
Indeed, New York plays a key role in the cost incurred by potential PV adopters by
offering incentives. The State has invested resources in subsidies and the administration of those
subsidies to spur adoption. This research investigates the role that various factors play in shaping
decision making in the residential market, which may prove useful to policymakers who define
subsidy levels. It is important to note that the legal aspect of solar has less to do with statutes
than regulations, as it is grounded in regulatory law governing electric utilities.
Solar PV panels (key components of solar power systems) installed at the residential
scale are typically mounted on the roof of a home, and sensitive electronic control equipment is
installed inside of the home. The electricity is used in two ways. The priority is for the electricity
to be used within the home to offset electricity that would otherwise be purchased from the
electric utility company that serves the home. That is, all the electricity from the solar equipment
is utilized and is supplemented by the utility. The secondary use of the electricity occurs when
the solar power system produces more electricity than is needed within the home. In this
scenario, the excess is provided from the home to the utility for use as the utility sees fit serving
other electric demand. A net metering arrangement allows the customer to receive credit for this
electricity sold to offset the purchases it makes from the utility (What’s the Difference, 2017).
Residential solar energy competes directly with grid power, which is a generic term
describing the electrical system that traditionally provides power by electric utilities to customers
through centralized generation and transmission via electric wires. It is with this understanding
that we first encounter the term grid parity. Grid parity is the economic price at which solar
19
energy costs are on par with the cost of power from the grid (Fellows, 2017). The grid consists of
billions of dollars of electricity generation, transmission, distribution, and metering
infrastructure. Power is generated at central station power plants, powered by fossil fuels (e.g.,
coal, natural gas, and oil), water (hydroelectric dams), nuclear reactors, or large-scale renewable
resources that include wind and solar generation. The advantage of grid power is low per-unit
($/kWh) electricity costs due to economies of scale and relatively high-efficiency power plants.
Solar power system equipment costs are driven by the capital cost of the equipment,
installation cost, and financing costs, offset by government incentives. The combination of these
components is then levelized over the expected production of electricity (also in $/kWh),
resulting in levelized cost of electricity (LCOE), a standard industry term. When the LCOE is
comparable to the cost of grid power, grid parity is achieved. At that point, solar competes with
the grid, at least on the basis of economics (Branker, Pathak, and Pearce, 2011).
While economics is central to the decision to adopt solar PV equipment, many other
factors influence the prospective residential customer as shown in the literature review. One of
these drivers is the role that solar energy plays in environmental conservation. For each unit of
electricity produced by a solar installation, a unit of electricity that would have been produced
through the combustion of fossil fuels is displaced.
Solar power systems produce zero emissions, so it is preferable to power produced from
fossil fuels from an environmental standpoint. Many states, including New York, have
implemented a renewable portfolio standard (RPS) that prescribes the amount of energy to be
produced (on a percentage basis) from renewable resources, such as solar systems and wind
turbines. RPS programs outline specific renewable targets by year and are the primary drivers of
the explosive growth in renewable resources. However, since grid parity (based on utility energy
20
rates and solar power system cost) is not possible without subsidies, the RPS programs are the
drivers of the subsidies required to meet the technology adoption targets. In New York, the RPS
is known as a Clean Energy Standard (CES).
Renewable Portfolio Standards (RPS)
Some states, including NY, have a self-imposed Renewable Portfolio Standard (RPS), a
generic industry term that sets targets for percentages of a state’s electric energy that must be
produced by renewable resources (such as solar) by certain dates. New York has been subject to
an RPS since 2004 and has recently made its targets more aggressive under its Clean Energy
Standard with a goal of 50 percent by 2030 (Clean Energy Standard, 2017). For practical
purposes, New York’s Clean Energy Standard serves the role of an RPS, in that it establishes
specific renewable energy targets.
To meet renewable generation targets such as the 50 percent by 2030 mark, New York
must focus on customer adoption of renewable energy technologies. It is the RPS concept that
drives the need for solar energy resources and adoption that makes it a reality. New York’s
political environment appears favorable for the achievement of tough RPS targets. Fowler and
Breen (2013) discuss the importance of politics in the development of RPS measures in states.
RPS is a precursor and driver to solar adoption that is indirectly related to solar adoption.
Challenges in meeting RPS targets are identified and addressed by Gibson (2013), who provides
insights into what challenges and opportunities lay ahead for solar, including natural gas, energy
storage, and market mechanisms to compensate for how solar equipment burdens the electric
grid.
State RPS policies vary widely. Schelly (2014) reviews and compares RPS policies in
21
Wisconsin and Colorado by interviewing solar adopters in both states. By focusing on just New
York, this research will avoid the influence of specific state RPS requirements and will be able to
isolate the key determinants of adoption. Focus on a specific state is important since government
policy plays such an important role in adoption (Simpson & Clifton, 2015).
Conclusions
Based on the literature search, there is limited literature on the narrow niche of this
research: the drivers of residential solar adoption in upstate New York. Relevant law pertains to
renewable portfolio standards (RPS), given that solar power is a regulatory matter. However, the
law and policy review clearly shows relevant schedules and parameters for New York to meet
RPS targets. This is critical to this research, as it dictates how the decrease of financial incentives
over time points to the importance of understanding the drivers of adoption.
Problem Statement
The subsidies that enable solar power to compete with grid power are on a schedule of
reductions that could make solar a harder sell to prospective adopters and pose challenges in
meeting the CES targets. If policymakers and solar industry players understand customer
behavior, rates of solar adoption can possibly be maintained in the future.
As the literature review in Chapter Two illustrates, a substantial amount of research exists
on various factors that may influence the adoption of solar power systems. This research was
designed to investigate virtually all the reasonably expected influences on adoption (decision
factors), forcing survey respondents to rate the impact of each on their decision to adopt. The
22
decision factors selected for this study were:
● Low or no up-front cost
● Expected energy bill reduction
● Positive impact on the environment
● Leaving a positive legacy
● Recommendations for solar power from peers
● Reputation of my solar installer
● Perceived honesty of my solar sales representative
● Exposure to public information about solar
● Reduced dependency on my power company
● The timing of life events enabled my solar installation
The literature review also exposed a scarcity of solar research pertaining to New York State,
further bolstering the need for this research project.
Purpose of the Study
The purpose of this quantitative survey study was to investigate the factors that drive
adoption of residential solar power systems in upstate New York. The objective was for these
findings to be useful for New York policymakers and businesses to create policy and effectively
approach the residential solar market. The goal is continued solar power system adoption that
contributes meeting New York’s CES targets.
Research Question and Hypothesis
This research was designed to seek answers to four questions:
1. What are the demographic parameters that describe the upstate New York residential
solar power system adopters?
2. What were the most important decision factors identified by adopters as shaping their
decision to adopt?
3. How do the importance ratings of the decision factors correlate to demographic
parameters?
23
4. How have the decision factors changed in importance over time?
Mankiw (2016) notes that many parameters influence the demand for a consumer good,
but price, in particular, plays a central role. Innovation is “motivated by the prospect that the
rewards will exceed the costs” (Miller, Benjamin, and North, 2016, page 11). The net price of
solar power systems is affected by both federal and state level subsidies, thus making it a key
component of a thriving New York solar market. This net price is a key factor in the cost of
electricity and, thus, the potential to experience energy cost savings through solar power. As
such, the expected energy bill reduction is hypothesized to be an extremely important decision
factor. In fact, the research hypothesis is that the most important decision factor to upstate New
York residential solar power system adopters is an expected energy bill reduction.
Theoretical Framework
This research is guided by the theoretical framework of diffusion theory, launched first
by Ryan and Gross (1943). Diffusion is not a single, isolated theory (Rogers, 1995), but rather a
collection of theories with four primary components: the innovation-decision process theory, the
individual innovativeness theory, the rate of adoption theory, and the theory of perceived
attributes. It is the last of these (perceived attributes) that guides this research. However,
although interviews are the generally-implemented form of data collection for diffusion theory
research, this quantitative survey study uses surveys to achieve a similar result. Whereas typical
diffusion research relies on multiple disciplines within the social sciences, this research relies on
self-reported survey responses to a list of decision factors. Most importantly, though, diffusion
24
theories are used to explain adoption, and the decision factors used in this research are the
drivers of solar power system adoption.
The theory of perceived attributes is that people adopt innovations if those innovations
have the following attributes: the innovation has a relative advantage over the current product or
technology; the innovation must be compatible with existing values and practices; the innovation
cannot be too complex; the innovation must have the ability to be tested before adoption; and the
innovation must offer observable results (Rogers, 1995). For this research, the critical component
of those five attributes is the first, relative advantage. This theory is implemented within this
research by including several decision factors that define various relative advantage features of
adopting solar power systems.
Assumptions
Several fundamental assumptions were made as a foundation of the research, most
notably about the validity of survey responses. The primary assumption is that the respondents
are, in fact, among the desired target population of research. The NYSERDA contact information
used for survey distribution was carefully selected by NYSERDA staff to meet the population
sample criteria. Additionally, several survey questions were posed to validate the respondent’s
membership in the target population. Alreck and Settle (2017) describe that survey data is
subject to mistakes, errors, and oversights, but the assumption of its validity as a whole is made
for this research, as further discussed in Chapter Three. Another assumption of this quantitative
survey study is that the survey participants answer the questions truthfully. Finally, there is an
assumption that solar power system adopters are homeowners due to the financial commitment
or investment associated with installing a solar power system on a home.
25
Scope and Delimitations
The scope of this quantitative survey study is defined by the population being studied. To
set a manageable scope, the population studied was narrowly focused. The population consisted
of people who a) are residential solar power system adopters; b) are residents of upstate New
York; c) installed their solar power system between 2013 and 2017; d) received incentives from
NYSERDA; and e) had email contact information on file with NYSERDA. The survey’s scope
covers the demographics and decision factors that influenced the respondents’ decisions to install
residential solar power systems in upstate New York.
Limitations
Some sources of weakness in research are out of the researcher’s control. This research
potentially has one such notable limitation, in that while it determines correlations between
demographics and decision factors, it cannot determine causality. Citing Blalock, Creswell
(2013) points out that correlational methods do not equate to causation. This is a common
problem with quantitative analysis, and while it by no means lessens the value of the analysis,
conclusions, or recommendations, it must be noted to ensure accuracy in interpretation of results.
Other limitations of the study are due to regional and state-level differences in key solar
power system parameters. State-level solar incentives vary substantially, as do utility energy
prices. Demographics such as household income also vary regionally. All of these factors affect
the likelihood of residential solar power system adoption. This leads to the question of
generalizability of the results of this research. Due to the upstate NY-specific focus of this study,
the results may offer clues to what may be important to prospective adopters in other states or
26
regions. However, the level of information provided here is not directly a generalization for all
regions nor will it necessarily serve as a transferable understanding of what is important to solar
adopters in other geographical locations.
Significance
New York must meet its CES targets for renewable energy in an environment where
subsidies that promote the adoption of renewable energy resources are scheduled to decline and
ultimately disappear. Understanding what drives customer behavior and influences solar
adoption can be useful to maintain solar adoption rates, which is key to meeting CES targets.
This research endeavor is designed to provide that information. Also, renewable energy
resources displace fossil fuel combustion at power plants. Because of this, the research for this
thesis can have benefits to local, state, and even global interests in the form of understanding the
decision factors that motivate solar technology adoption and ultimately reduce emissions from
power plants.
Solar power systems are used throughout the U.S. and throughout the world. As such, this
research has implications for use on a global scale. However, since the regional parameters vary
widely, as mentioned previously, the actual importance of decision factors is not necessarily
applicable from place to place. However, the approach used to study the upstate New York
residential solar power system adopters may be implemented in studies elsewhere. This is
particularly important given the connection between solar power adoption and the reduction of
greenhouse gases that cause climate change. Climate change is a global issue, and solar power
systems are globally implemented. As such, this research contributes to knowledge beneficial on
a global scale.
27
Chapter 2: Literature Review
The purpose of this quantitative survey research was to investigate the factors that
influence the adoption of residential solar power systems in upstate New York. This research
relied on a foundation of knowledge gleaned in the literature review. Several of the factors that
this research investigates were researched in previous work identified in the literature review.
Literature Search Strategy
The strategy implemented for this literature research was to identify and examine existing
research on solar power adoption. Foundational literature searches were conducted between
August 2016 and November 2017 to form the basis for the research proposal. The searches were
performed primarily using Northeastern University’s library system and Google Scholar. This
ultimately led to key concepts for examination, including renewable portfolio standards, solar
economics, customer adoption, and environmental benefits of solar. This framework shaped the
research and played a key role in the development of the decision factors that were examined by
this research project.
Overall, there is a substantial body of literature on solar power, including scholarship on
policy (New York’s CES), solar economics, environmental benefits, and solar adoption. Very
little focuses specifically on New York State. This indicates that this research may add valuable
information to the field in that it focuses on residential solar power adoption in a heavily-
populated U.S. state, specifically upstate New York.
Research on Renewable Portfolio Standards
Some states, including NY, have a self-imposed Renewable Portfolio Standard (RPS), a
generic industry term for regulatory policies that set targets for percentages of a state’s electric
28
energy that must be produced by renewable resources (such as solar) by certain dates. New York
has been subject to an RPS since 2004 and has recently made more aggressive targets under its
Clean Energy Standard, to be 50 percent by 2030 (Clean Energy Standard, 2017). For practical
purposes, New York’s Clean Energy Standard serves the same purpose of an RPS in that it
requires renewable power to meet a specific target percentage of the State’s electric energy. It is
worth noting that in addition to the literature review on RPS and New York’s CES, these topics
were also the subject of the law and policy review of this quantitative survey study.
To help meet the overall renewable targets, New York has solar-specific targets through
its NY-Sun program, currently set at 3 gigawatts of solar power by 2023 (NYSERDA, 2018). It
is the RPS (or CES) concept that drives the need for solar power resources, and actual adoption
must occur to make achieving the RPS a reality. New York’s political environment appears
favorable for the achievement of tough RPS targets. Fowler and Breen (2013) address the
importance of politics in the development of RPS measures in states. RPS is a precursor and
driver to solar adoption.
Research shows that increases in renewable energy sources strain electric power systems
and that because of this, power system operators tend to create obstacles to renewable energy in
the form of increased costs imposed upon utilities that support renewable resources (Alqahtani et
al., 2016). These increased costs may ultimately be borne by solar adopters, making solar power
less economically attractive. This reinforces the need to identify the role that all factors—not just
economic ones—play in solar adoption. It is worth noting that although Alqahtani et al. (2016)
performed their research by examining North Carolina’s electric system, the same principles may
apply throughout the United States, making them applicable to New York’s system. The issue
does not deviate geographically.
29
In addition to the actual state-specific RPS targets, another factor that varies
geographically is the impact of weather and climate on solar equipment. Weather and climate
impact equipment efficiency, levelized cost, and longevity (Flowers et al., 2016), which are key
elements of solar economics. This quantitative survey study will examine factors that impact
residential solar electric technology adoption within the climate of one region only, upstate New
York.
Other challenges in meeting RPS targets are identified and addressed by Gibson (2013).
Gibson provides insights into what challenges and opportunities lay ahead for solar. One
challenge is low natural gas prices that result in low grid power prices, which makes it hard for
solar to compete. One opportunity is energy storage advancements that enable solar power to be
stored for use when needed, making it more economically advantageous.
State RPS policies vary widely (Schelly, 2014). Schelly reviewed and compared RPS
policies in Wisconsin and Colorado and interviewed solar adopters in both states. By focusing on
just New York, this research will avoid the influence of varying state RPS requirements and will
be able to isolate the key determinants of adoption within the context of one policy regime.
Focus on a specific state is relevant, since government policy is considered important to adoption
(Simpson and Clifton, 2015).
Research on Solar Economics
The levelized cost of solar power can be compared to energy purchased from the
traditional utility (Alafita and Pearce, 2014). This parameter is a key element of adoption, thus
lower levelized cost would benefit prospective adopters. A factor that can reduce levelized cost
is access to low-cost financing. Securitization (low-cost, long-term financing) is a useful tool to
decrease solar costs. Branker et al. (2011) also explore the calculation of the Levelized Cost of
30
Energy (LCOE) for solar to compare it to traditional energy costs from the power grid. The
authors show that when these two are comparable, the parity that exists makes it more likely that
there will be increased adoption of PV.
Burns and Kang (2012) examine which states do the best job of promoting solar adoption
through state-level incentives. Concepts presented by Burns and Kang (2012) are useful in
understanding the various tools that state governments can use to promote adoption, but their
study focuses only on states with Solar Renewable Energy Credits (SRECs), which New York
does not utilize. Nonetheless, their research points to the importance of state-level economic
incentives to foster solar technology adoption. Drury and Margolis (2011) stress the importance
of solar price as well, indicating “the different price thresholds for when a PV investment
becomes profitable or attractive and the different sensitivities to varying system parameters have
significant implications for policy design” (page vi).
Feldman et al. (2013) take another approach to addressing solar costs by focusing on the
non-equipment or “soft cost” components of LCOE: financing, overhead, and profit. Their study
helps to tie in the equipment costs with the soft cost to gain a total picture of all-in solar costs,
which drives what customers pay for solar power and, therefore, can influence their likelihood of
adoption. Financial incentives that drive solar costs are driven by state-level policy (Fowler and
Breen, 2014). Fowler and Breen (2014) stress the important role of policy in influencing the
adoption of renewable energy and RPS standards.
Beppler et al. (2017) examine the system-wide effects of solar penetration on the overall
grid. They found that residential solar technology adoption reduces adopter bills at a cost to non-
adopters, essentially a cross-subsidization. Their research is important because it stresses the
need to use this information for cost allocation to appropriately assign costs.
31
One concern not frequently mentioned by solar power researchers is the long-term
financial impact that the solar industry could have on the traditional utility business. Laws et al.
(2017) address the concept of a utility death spiral, whereby customers reduce their dependence
on their electric utility by adopting solar power. This leaves the remaining utility customers to
cover utility fixed costs in their electric rates. As the remaining customers’ bills increase, they
find it more attractive to adopt solar, and the cycle continues until the utility faces severe
financial strain. Laws et al. (2017) built a model to test how LCOE can reach parity and spur
adoption and produce this death spiral. While the authors conclude that a death spiral scenario is
highly unlikely, it is still important for both the solar and utility industries to understand their
interdependence and need for the long-term viability of both traditional utility operations and the
solar industry since solar resources can currently only provide a portion of energy needs.
Solar costs are also driven by the weighted average cost of capital (WACC) that is used
to finance equipment (Ondraczek, Komendantova, and Patt, 2015). Adoption has a lot more to do
with LCOE than actual solar radiation—and WACC is a key part of that. Work to examine the
role of WACC has been conducted at the cross-national comparative scale, but this work has
important implications for solar in New York State since New York is on the low end of annual
sunshine received and yet has high solar targets. The amount of sunshine available is not as
important as one would expect.
It is worth noting that while the proposed research and literature review focuses on state-
level policies, there is an important role played by the federal government in the form of
economic incentives (Zhai, 2013). Zhai (2013) examined the federal Investment Tax Credit
(ITC) and found that in Arizona, state and local subsidies were not sufficient to maintain
reasonable adoption rates in the absence of the ITC. The ITC allows solar customers to claim 30
32
percent of the cost of solar power systems as a credit on federal taxes.
To summarize, the literature indicates that economics is a factor in the adoption of solar
power equipment. It also indicates an important link between solar economics and government
policy. Specifically, states such as New York subsidize solar power systems and reduce the cost
of adopting a solar power system. Overall, the solar economics literature shows that levelized
cost is a key factor in solar adoption, and state and federal incentives factor into it.
Research on Customer Adoption
Although economics is a factor in solar power adoption, other research has investigated
how non-economic factors influence a prospective customer’s decision to adopt solar
technology. Understanding these factors may play a key role in understanding the viability of
solar power. For example, according to Balcombe, Rigby, and Azapagic (2014), energy
independence or not relying on the grid was a key driver for why customers opted to invest in
solar power.
Bauner and Crago (2015) found that reducing the long-term uncertainty of the economics
of energy expenditures is an important factor in solar adoption. Not only is the actual cost of
solar important, but stability in energy costs via solar installation is an important benefit to
customers as well. That is, the less variability, the better. Bollinger and Gillingham (2012)
examine the role of social interaction in the form of peer influence on the diffusion of
environmentally responsible technologies like residential solar. Although “causal peer effects are
notoriously difficult to identify” (page 3), Bollinger and Gillingham (2012) find a strong positive
correlation between peer influence and solar adoption. This finding is bolstered by Graziano and
Gillingham (2015) who investigate the diffusion of residential solar throughout Connecticut.
Their analysis shows that adoption promotes further adoption, and the diffusion patterns reflect
33
this phenomenon.
Kwan (2012) examines the environmental, social, economic, and political variables that
impact solar adoption, finding that areas with heavy insolation (the amount of solar radiation
reaching a given area) are underperforming due to environmental, social, economic, and political
factors. This bolsters the previous assertion that the amount of sunshine is not a critical driver in
solar adoption—this is very important because the other parameters can be controlled, but
sunshine cannot be.
Noll, Dawes, and Rai (2014) explored the peer effect of solar community organizations
(SCOs) on adoption. SCOs are organizations that promote solar through communication. The
study concluded that future studies should consider the role of SCOs in the adoption of solar
equipment. A study conducted by Palm (2017), which focused on solar adoption and peers,
found that direct influence of close peers is more impactful at influencing choice than passive
influence, such as seeing solar panels. The study supports the notion that peers play a very clear
and important role in influencing adoption (Palm, 2017).
Information and communication have been central themes of non-economic adoption
parameters, further supported by Rai and Beck (2017). The authors addressed the gap between
the potential of solar power adoption and reality. They attributed much of the gap to a lack of
information and misinformation. The researchers use a trivia game to disseminate information
and track the outcomes from that information in how it affects the players' views on issues that
are pertinent to solar adoption. Rai, Reeves, and Margolis (2016) found that 82 percent of
residential solar adopters also co-adopt other energy-efficient technologies. This is finding shows
that environmental concern, not just economics or peer influence, is a key driver of adopting
solar.
34
Sommerfeld et al. (2017) identified factors that motivated solar adoption in Australia by
investigating the disconnect between government policy goals and business goals in the adoption
of solar power equipment. They found that there were three types of motivators: economic (cost
of solar), social (peer influence), and environmental (reduced pollution).
Schelly (2010) correlated solar thermal adoption at the county level to socioeconomic,
environmental concern, and ecological (temperatures and solar radiation) indices and concluded
that several key correlations existed. Most notably, the socioeconomic variables household
income, home value, and education were correlated with thermal solar adoption. Schelly (2014)
explored elements that drive photovoltaic adoption, including environmental motivations,
economic considerations, and demographic characteristics. The 2014 study found that adoption
can be influenced by the timing of economic events in a homeowner’s life (e.g., an inheritance
that enables the purchase of solar equipment). The study also found that a homeowner’s concern
for the environment is by itself not typically enough to motivate solar adoption.
Based on the literature, economics appears to be an important factor influencing solar
adoption. However, several non-economic issues are also important, such as demographics (e.g.,
income, education), peer influence and communication, and environmental concern. It is a
combination of these factors that encourages solar adoption.
Research on Environmental Benefits of Solar Power Systems
The literature review so far has focused on policy, economics, and adoption. While
environmental benefits cross all three of those categories, it is worth deeper consideration due to
the importance that perceptions of environmental benefits may have on solar power system
adoption. While environmentalism is not necessarily every adopter’s primary driver of solar
adoption, there are substantial environmental benefits to be gained from solar adoption.
35
Cole et al. (2015) examine future solar costs and natural gas prices to determine the role
of solar under the draft Clean Power Plan (CPP). Their study finds that solar adoption is very
dependent on market conditions (solar and gas costs) and that, in turn, will drive environmental
benefits. This creates an important link between economics and the environment.
Kaufmann and Vaid (2016) assessed the economic viability of roof-top solar electricity
production, regarding the costs and benefits of the technology. The study focused on the impact
on the electric grid as a whole and the environmental benefits of solar. They calculated the
effective reduction in energy and greenhouse gases that would otherwise be produced by gas-
fired power plants but would instead be displaced by solar power. Their research confirmed that
the net economic impact of solar generation was positive, and that the reduction of greenhouse
gases was an additional benefit of solar.
Summary and Conclusions
There is substantial literature on solar economics, solar adoption, policy and renewable
portfolio standards (RPS), and environmental benefits on a global scale. There is a dearth of
research that focuses specifically on New York State, a state with stringent Renewable Portfolio
Standard (RPS) targets set forth by regulations and policies.
The literature shows that many factors influence solar adoption, and most research
isolates one or a few of these variables at a time. Thus, this dissertation contributes new insight
into solar technology adoption at the residential scale by investigating many known variables
simultaneously in a specific geographic region. Understanding these factors is critical to
effective policies to promote solar power. There also is a need to explore solar adoption decision
36
factors by demographics, an example of which is specifically identifying prospective customer
education levels (Varela-Margolles and Onsted, 2014).
37
Chapter 3: Methodology
The purpose of this quantitative survey study is to investigate the factors that drive
adoption of solar power systems. The objective is for these findings to be useful to state-level
policymakers and businesses to create policy and effectively approach the residential solar
market. The goal is continued solar power system adoption in a more competitive, non-
subsidized environment.
Research Design and Rationale
This survey research design was used to examine the solar adoption in upstate New York.
For homeowners in upstate New York who have adopted solar power systems, the research
investigated the factors that influenced their decisions. Descriptive statistics were analyzed to
describe participants’ decision factors. Correlation analyses were used to analyze the
relationships between decision factors and participant demographics or other indicators in the
survey; for example, this study describes the relationship between household income and the
level of importance of expected energy bill reduction (which was a decision factor). This study
did not attempt to determine causality, as only relationships between factors were examined.
Methodology
This research is quantitative, relying on numbers to identify demographic profiles,
calculate descriptive statistics, and calculate correlation coefficients. Key to a successful and
valid quantitative study is detachment of the observer (Balnaves and Caputi, 2014). That is, the
study should be performed objectively, without researcher influence, to the extent possible. In
order to gather the data for this analysis, a survey was distributed to residential solar adopters in
38
upstate New York. Objectivity of the research was kept in mind during the development of the
survey, so as not to guide or influence answers by construction of unintentionally weighted,
biased, or leading questions.
Using a survey was the logical choice for capturing the relevant data from the target
population. Balnaves and Caputi (2014) submit that the survey method is appropriate when one
cannot directly observe what is being studied and is particularly fitting for large populations such
as the one that is the subject of this research. The survey was carefully constructed to capture
demographic information and responses to how important each of ten decision factors was to the
decision to adopt, with the former being done with particular care given the sensitive nature of
demographics. Patten (2014) advises to survey about demographic information sparingly to limit
the size and perceived intrusiveness of the instrument. Alreck and Settle (2017) assert that
surveys are best used to understand or predict human behavior for academic and professional
work. They also note that, inevitably, survey data is subject to mistakes, errors, and oversights,
and that there are three key attributes of good survey questions: focus, brevity, and clarity. This
was all heeded in the development of the survey, which is attached as Appendix A.
A key element of this quantitative survey study is the calculation of statistics that
describe how important each decision factor was to the sample of upstate New Yorkers who have
adopted residential solar electric technology. This was done with a basic descriptive statistic, the
mean value of the importance ratings that ranged from one (not at all important) to five
(extremely important). Patten and Bruce (2014) note that a drawback of using the mean is that
extreme outliers skew the data. However, this is not an issue when ordinal data is used (as it was
for this survey) since the data consists of integers from one to five for an importance rating of
each decision variable.
39
Another key element of this quantitative survey study was calculating correlation
coefficients to show relationships between demographic variables and importance ratings of the
decision factors. “Correlation analysis generates a single value, the correlation coefficient, that
shows how much the two variables move together. The correlation coefficient is usually
symbolized by the letter r. It ranges from a value of zero, indicating there’s no relationship
between the variables, to a plus or minus one, indicating a perfect linear relationship” (Alreck
and Settle, 2017, page 323). Alreck and Settle (2017) describe that in a positive correlation, the
variables move in the same direction—that is, as one goes up, so does the other. In a negative
correlation, as one variable increases, the other declines.
Key to implementing an appropriate correlation analysis is choosing the correct
algorithm. While the most common correlation method is called the Pearson product-moment
correlation, it is applicable to interval data, or actual variable values. When one or more variable
is ordinal data, as is the case for this research, the Spearman rank correlation approach must be
used (Alreck and Settle, 2017). The statistical software package SPSS was used to calculate the
Spearman rank correlation to produce the correlation coefficients between the variables of this
study. Spearman’s correlation coefficient, also known as Spearman’s rho, is designated as the
Greek letter ρ.
Positionality Statement
Having worked in the energy industry for over 25 years, I have been exposed to all facets
of electric power production. This includes helping clients large and small produce electric
power for profit, testifying to help regulators keep customer costs for electricity in check,
contributing to the development of important pollution regulations by the U.S. Environmental
Protection Agency (EPA), and assisting in the development of renewable energy projects. My
40
background has been diverse, and it has shaped who I am as a consultant, a businessman, and a
member of society.
Positionality affects how researchers choose processes and interpret results (Holmes,
2014). I recognize how my experiences have shaped how I conducted my research and how this,
in turn, may have impacted how I interpreted the results. We do not live in a vacuum, so it is
important that I accepted and understood these experiences to stem any material bias that could
have resulted from them.
I chose to study the parameters that drive the adoption of residential solar power
equipment, because it touches on economics, technology, the environment, and social sciences.
In certain aspects of this undertaking, I am very much an insider. I have performed many
renewable energy consulting engagements, I am an electrical engineer, and I believe that
renewable energy should be a key component of the energy consumed in New York, the United
States, and the world. On the other hand, I am not a solar adopter. Throughout this research
endeavor, I have strived to remain cognizant of the aspects of my research that make me an
insider to maintain a balanced, scientific approach to produce valid, useful results.
It was important to maintain objectivity and eliminate bias throughout this quantitative
survey study. This included the development of the survey instrument, selection of the
population sample, data analysis, and development of conclusions. Maintaining objectivity was
an area of intense focus throughout this quantitative survey study. The choice of performing a
quantitative analysis on survey data reduced the potential for bias, since the study used numerical
data and correlations, not qualitative interpretations of open-ended questions. The approach is
pragmatic in that it is problem-centered and based on real-world practice (Creswell, 2013).
41
Data Collection
Population
The population is comprised of residents of upstate New York who adopted residential
solar technology between 2013 and 2017, who received an incentive from NYSERDA to offset
solar power system costs, and for whom NYSERDA retained email addresses. This population
includes 19,634 upstate New York residential solar power system adopters. NYSERDA
maintains a database with information pertaining to the installation, such as the system size,
electric utility, and location. Given the existence of this data, it made sense to use the same
population for this research.
Sample Size Calculation
The survey was distributed to a total of 19,634 email addresses of upstate New York
residential solar power system adopters. The response to the survey distribution was 2,093
completed surveys, a response rate of 10.7 percent. The goal was to obtain a sample that
provided a confidence level of 95 percent with a margin of error of three percent. The target
sample size is calculated by the following formula (Sample Size Calculator, 2018):
𝑆𝑆 =
𝑧2𝑝(1 − 𝑝)𝑒2
1 +𝑧2𝑝(1 − 𝑝)
𝑒2𝑁)
In this formula, N = population size (19,634), e = margin of error (0.03), z = z-score (1.96 for a
confidence interval of 95 percent), and p = population proportion. This results in a value of 1,013
for a representative sample. The sample who responded to the survey (2,093 residential solar
power system adopters) exceeded the number required for a sampling validity. About 12 percent
(251) of responses were dropped from the data for being incomplete. Records were excluded as
incomplete if they met any one of several criteria: missing any of the ratings of the ten primary
42
decision factors, declaring a solar installation year outside of the scope of this analysis, or
declaring an electric utility company not in the list of those considered by NYSERDA to be in
the upstate New York territory. The final sample size was 1,842, which is still a representative
sample of the solar adopting population at 95 percent confidence and 3 percent margin of error.
Recruitment
This study was conducted by developing a detailed online survey (Appendix A) hosted
by Qualtrics. The survey captured information about the adopters’ demographics, the physical
descriptions of the home upon which the system was installed, and, most importantly, the rating
of the importance of the factors influencing their decision to adopt. Due to concerns about
protecting the anonymity of the population, NYSERDA assisted in the distribution of the survey
rather than provide the actual email contact information. NYSERDA uploaded the survey to its
Qualtrics account and distributed the survey to its protected list of residential adopters. As such,
NYSERDA maintained ownership and access to identifiable information and did not share such
information. At the close of the survey, approximately one week after initial distribution,
NYSERDA provided the results with no identifiable information. A total of 2,093 residential
adopters completed the survey, which accounts for approximately 11 percent of the population.
This far exceeded the 1,013 participants required for a representative sample of respondents at a
95 percent confidence interval and 3 percent margin of error. NEU researchers will destroy the
de-identified data at a reasonable time after the completion of the study.
Ethical Procedures
This research relied on the input of survey participants who were contacted by
NYSERDA via email. Because human subjects were involved, ethical procedures guided the
processes by which participants were recruited and how their contact information and responses
43
were used. This was an important consideration for the research team, Northeastern University’s
Institutional Review Board (IRB), and NYSERDA management.
NYSERDA, as gatekeeper of the participant email addresses, phone numbers, and
physical addresses, was particularly concerned with protecting this information. The research
team submitted a Freedom of Information Law (FOIL) request for the solar adopter email
addresses, which was subsequently denied. However, shortly after the denial, NYSERDA gave
notice that they would reconsider providing the contact information if the Northeastern
University research team met three criteria: being a non-profit organization; using the data for
academic research; and performing research that was consistent with NYSERDA’s mission. It
was quickly determined that these three criteria were met.
Despite meeting the three criteria, NYSERDA still expressed concern over providing
solar adopters’ email addresses to an external organization. Their interest was protecting the
privacy of the solar adopters to ensure that they would not be contacted for other (especially
commercial) purposes beyond the scope of the survey to support this research. Providing the
email addresses would potentially pose that risk. An additional risk was introduced in that
Qualtrics can capture internet protocol (IP) addresses that could be used to identify individual
survey respondents from the location of their computer or mobile device. That feature was
disabled for the survey.
To assuage the concerns of misuse of the email addresses and capturing IP addresses, the
research team proposed an alternative solution to NYSERDA. The alternative was to have
NYSERDA send the email introducing the study and soliciting participation. In this manner, no
third party would have the email addresses. The other component of the alternative approach was
that NYSERDA would host the survey on its own Qualtrics account. This ensured that the survey
44
responses would be controlled by NYSERDA and delivered to the researchers without any
identifiable information such as email or IP addresses. This alternative plan was accepted and
implemented by NYSERDA.
In alignment with IRB requirements, the de-identified data will be stored in a password-
protected Microsoft Excel file for a period of no more than five years. The data will be stored on
a secure computer with firewall protection. After this period, the data will be permanently
deleted.
Prospective participants were emailed a link to the survey with a request for their
participation in the research study. Those interested in participating clicked a link that opened a
browser for either PC-based or smart phone-based use and began with an embedded informed
consent form. The survey was designed to be completed in under five minutes to minimize the
inconvenience of any participant.
The survey consisted of 31 questions, most of which were multiple choice, with a few
spaces for open-text explanations. The survey was piloted before release by sending it to NEU
DLP students and other peers. The goal of the pilot was to test the distribution email, the link, the
performance of the survey platform, the readability of the survey items, and the collection of the
results into a database. The pilot indicated that all phases were satisfactory, and the survey was
launched by NYSERDA to their contact list. At the close of the survey, NYSERDA exported the
responses from Qualtrics to an Excel worksheet, de-identified the data, and provided the data to
the researcher for analysis.
45
Data Analysis
Once the results were provided by NYSERDA, the analysis phase began. The data were
tabular in nature and were managed exclusively in Microsoft Excel in a password-protected
workbook file on a password-protected laptop computer. Excel was used for the secure storage
of the raw data, the statistical analyses, and the creation of graphics.
Part of data analysis is cleaning the data to ensure that only valid data are analyzed. Data
were cleaned to eliminate survey responses that appeared to be generated from respondents who
are not members of the target sample and incomplete data. Typically, either type can be deleted,
as long as doing so does not affect the statistical significance of the sample (Osborne, 2013).
Deleting erroneous data did not impact the sampling validity of the data, given the large sample
collected for the study. Details of the data cleaning process are provided in Appendix B.
This quantitative analysis was performed on survey data in four phases. In the first phase,
descriptive statistics show the demographic makeup of the respondents, all of whom are adopters
of residential solar power systems in upstate New York. In the second phase, descriptive
statistics were calculated on the ratings of the importance of the decision factors that influenced
participants to adopt solar power systems. In the third phase, correlations were calculated to
identify key relationships between demographic characteristics and ratings of the decision
factors. Finally, in the fourth phase, the importance ratings were grouped by the year of solar
system installation.
The first phase of the analysis was to conduct descriptive statistics to describe key
elements of the sample and survey responses. Each of the ten decision factors that were rated in
importance to the solar power system adoption were put in ordinal format of integers from 1 (not
important at all) to 5 (extremely important). The minimum, maximum, mean, and median values
46
were calculated for each for the entire sample. For this analysis, the key determinant of
importance was the mean score. The mean is appropriate for use with “equal interval data” such
as the ordinal data used for the importance ratings of decision factors (Patten and Bruce, 2014).
Based on the mean scores, it was immediately apparent that certain variables were of far greater
importance to the decision to adopt than others.
The next step was to calculate the same descriptive statistics groups within the sample.
These were created by the demographic questions in the survey; for example, age, race, and level
of education. This enabled a view of how people in various demographic groups appraised the
importance of the decision factors. Finally, statistics for each individual year of the five years of
the study were calculated, for use in a later stage of reporting. The hypothesis of this research
was that the most important decision factor would be expected energy bill reduction. The
analysis was designed to use these data to test the hypothesis.
Threats to Validity
Ensuring validity in this quantitative survey study required obtaining and maintaining
valid data from survey respondents. Key to this was that all survey respondents were certified by
NYSERDA as solar adopters. This limits some of the concern about validity of the data. Of
course, threats to validity can come from bias in the data as well. One concern was that the
responses would not show a diverse geographical dispersion across all upstate New York, but a
review of the data showed that all counties that comprise upstate New York were represented. Of
the 62 counties in New York State, 55 of them are considered part of NYSERDA’s definition of
upstate New York, and each was represented in the sample.
47
Creswell (2013) highlights internal and external threats to validity that should be
considered in a research endeavor. Of those threats, those associated with history (an internal
threat) are relevant to this quantitative survey research. History-related threats are related to the
elapsing of time from when the solar installation occurred and the time of the survey. The survey
was designed to capture the upstate New York solar power system adopters’ view on solar power
at the time of their installation. Since up to 5 years may have elapsed since completion of the
installation (installations were between 2013 and 2017), a survey participant’s actual experience
with solar since the installation may influence his recollection of his view of solar at the time of
installation. This was dealt with at the time of the survey, as it was designed with specific
language to highlight the temporal anchor of aiming to receive information regarding
respondent’s thoughts at the time of installation.
Summary
The primary purpose of this quantitative survey study was to investigate the decision
factors that drive adoption of residential solar power systems. The objective was for the findings
to be useful to state-level policymakers and businesses to create policy and effectively approach
the residential solar market for continued solar power system adoption in a more competitive,
non-subsidized environment. Through a tailored survey design and implementation of
established analytical techniques, the research project met those objectives.
NYSERDA played an important role in recruiting solar adopters to participate in this
research. Their participation enabled the collection of a large sample of valuable data from solar
adopters while maintaining the anonymity of the survey participants and ensuring that their
identifiable information was not disclosed to any outside party.
48
Chapter 4: Results
The primary question for this survey research was “What are the key factors that
influence residential solar power system adoption in upstate New York?” Through descriptive
and correlational analysis of survey responses, this research considered the factors that
influenced homeowners’ decisions to adopt solar power systems. The research hypothesis was
that the most important decision factor, regardless of participants demographics, is expected
energy bill reduction.
Results of the Analysis
The analysis was organized into four components, each designed to examine a specific
research question. The components were Demographic Data, Descriptive Analysis: Decision
Factors, Correlations of Decision Factors to Demographics, and Decision Factors Ratings Over
Time.
Demographic Data
The survey captured demographics of the respondents, including race, education level,
age, and household income. The demographics of the sample are presented in Table 4.1. To
summarize the key demographics, 94 percent of the sample is White, most of the sample is over
50 years of age and hold at least a Bachelor’s degree and believe that climate change is
happening.
49
Table 4.1
Summary of Survey Respondents
Demographic Parameter
Race
American Indian or Alaska Native
Asian
Black or African American
Native Hawaiian or Other Pacific Islander
Other
White
Mixed
Education Level
No high school diploma or GED
High school diploma or GED
Associates Degree
Bachelor’s Degree
Master’s Degree
Doctoral Degree
Age
18-29
30-39
40-49
50-59
60-69
70 or older
Household Income
$0 to $50,000
$50,000 to $100,000
$100,000 to $200,000
$200,000 to $300,000
$300,000 to $400,000
$400,000 to $500,000
Above $500,000
Home Value
Under $100,000
$100,000 to $199,999
$200,000 to $299,999
$300,000 to $399,999
$400,000 to $499,999
$500,000 to $599,999
$600,000 to $699,999
Over $700,000
Climate Change Beliefs
%
0.1%
1.5%
0.9%
0.1%
2.1%
94.2%
1.1%
0.3%
10.3%
12.6%
30.5%
31.8%
14.6%
2.3%
12.4%
19.4%
28.0%
29.3%
8.5%
7.6%
35.7%
43.6%
8.6%
2.1%
0.8%
1.6%
4.3%
29.6%
29.9%
17.7%
7.7%
4.5%
2.5%
3.7%
50
It's not real
It's real but is not caused by humans
It's real and is caused by humans
Political Affiliation
Democrat
Republican
Other
Multiple
Gender
Female
Male
Other
Marital Status
Single
Married
3.8%
11.8%
84.4%
48.7%
21.9%
25.2%
4.2%
26.8%
73.0%
0.3%
16.7%
83.3%
Solar adoption and race. The first demographic category analyzed was race of the
respondent. The survey provided respondents with the ability to select between primary race
categories but also allowed for selecting not to answer and for selecting multiple categories to
reflect mixed race. It is worth noting that 1,771 of the 1,842 survey responses (96 percent)
identified their race. Figure 4.1 uses 1,771 as the basis for its racial ratios. That is, the percentage
makeup by race is of those who provided race data, not all 1,842 responses.
The data shows that most solar adopters in upstate New York identify as White,
indicating an overwhelming majority of White people in the survey response. It is worth noting
that the mixed/multiple race category was comprised almost completely of respondents who
identify with White as one of their multiple racial categories.
51
Figure 4.1. Solar Adoption by Race.
The high ratio of White respondents reflects the general population of upstate New York
that is largely White (82 percent in 2016) according to 2016 U.S. Census Bureau data (US
Census Bureau, 2018). Thus, it stands to reason that solar adoption in the region would have a
large White representation. The 94 percent of the sample who are White show that Whites make
up a larger portion of the solar adopters than they do of the general population. The opposite is
true of other races in the sample. For example, Census Bureau data for upstate New York
indicates that Asians comprise 3.2 percent of the general population but only 1.5 percent of the
population who adopted solar. Likewise, African Americans comprise 8.8 percent of the general
population in upstate New York, but just 0.9 percent of the population of who adopted solar.
Solar adoption and educational attainment. Educational Attainment was another key
demographic component. The survey provided respondents with the ability to select between
educational attainment, but also allowed for selecting not to answer. 99 percent of respondents
identified their education level. Figure 4.2 uses 1,822 its education ratios. That is, the percentage
makeup by education is of those who provided education data, not all 1,842 responses.
52
The data shows that solar adopters in upstate New York are generally college educated,
with 77 percent holding a Bachelor’s degree or higher. About fifteen percent (slightly more than
one out of seven) of the respondents hold a Doctorate degree.
Figure 4.2. Solar Adoption by Education.
Clearly, the results show the high ratio of college to non-college-educated respondents.
This is high from a relative standpoint when compared to the general population of upstate New
York. Approximately 32 percent of the general population of upstate New York (of those 25
years of age or older) hold a Bachelor’s degree or greater (US Census Bureau, 2018). The survey
respondents have well more than twice that rate (77 percent). Also, in the U.S., approximately 2
percent of the population has achieved a Doctoral level degree (US Census Bureau, 2018). With
a rate of nearly 15 percent, upstate New York’s solar adopters are more than seven times as
likely to hold a Doctorate than the general population.
Age and Solar Adoption. Age was another key demographic component of the survey.
The survey provided respondents with the ability to select between age ranges, but also allowed
for selecting not to answer. 99 percent of the sample identified their age range. Figure 4.3 uses
53
1,826 as the basis for its age range ratios. That is, the percentage makeup is of those who
provided age data, not all 1,842 responses.
The data show that most of the solar adopters in upstate New York are between the ages
of 30 and 69 years, with almost two-thirds being ages 50 to 69. This data is displayed graphically
in Figure 4.3, which helps to highlight the substantial makeup of older respondents among solar
adopters. The data shows that the lowest and highest age ranges add very little to the makeup of
this sample of solar adopters. In fact, 89 percent of the solar adopters fall into the age range of 30
to 69. Below that range accounts for just 2 percent, and above that range accounts for just 9
percent.
Figure 4.3. Solar Adoption by Age.
Clearly, the results show the high ratio of respondents in the middle age range categories.
This is high from a relative standpoint when compared to the general population. For the general
population of upstate New York, the makeup of the population between 30 and 69 years of age is
approximately 61 percent (US Census Bureau, 2018). However, since this group accounts for 89
54
percent of the sample, it shows a much higher proportion of this age group represented in the
sample as compared to the general population.
Household Income and Solar Adoption. Income was another demographic component
of the survey. The survey provided respondents with the ability to select between income ranges
but also allowed for selecting not to answer. 90 percent of respondents (1,655 of the 1,842)
identified their income range. Figure 4.4 uses 1,655 as the basis for its age range ratios. That is,
the percentage makeup by household income is of those who provided household income data,
not all 1,842 responses.
The data shows that solar adoption in upstate New York is largely comprised of
households with annual income between $50,000 and $200,000, totaling 79 percent of the solar
adopting population. Only 4.5 percent of respondents had household incomes more than
$300,000. The data is displayed graphically in Figure 4.4, which helps to highlight the
substantial makeup of respondents with incomes between $50,0000 and $200,000. The data
shows that the lowest and highest income ranges add very little to the makeup of the sample.
Figure 4.4. Solar Adoption by Household Income.
55
Clearly, the results show the high ratio of respondents in the $50,000 to $200,000 income
range categories, and shows that the lowest income bracket, $0 to $50,000 accounts for just 7.6
percent. The median household income of the general population of upstate New York was
$59,918 in 2016 (US Census Bureau, 2018). This indicates that the sample is comprised of
higher earners, on average, than the general population.
Home Value and Solar Adoption. Home value was another demographic component of
the survey. The survey provided respondents with the ability to select between income ranges but
also allowed for selecting not to answer. Ninety-seven percent (1,783 of the 1,842) of
respondents indicated their home value range. Figure 4.5 uses 1,783 as the basis for its age range
ratios. That is, the percentage makeup by home value is of those who provided home value data,
not all 1,842 responses.
The data shows that solar adoption in upstate New York is largely comprised of homes
with values between $100,000 and $300,000 which account for approximately 60 percent of
installations. The data is displayed graphically in Figure 4.5, which helps to highlight the
substantial makeup of respondents with home values in the lower value ranges. The data shows
that the lowest and highest ranges add very little to the makeup of the solar adopter population.
56
Figure 4.5. Solar Adoption by Home Value.
Clearly, the results show the high ratio of respondents in the $100,000 to $299,999 home
value range categories.
Climate Change Beliefs and Solar Adoption. Respondents’ beliefs about climate
change was another demographic component of the survey. The survey provided respondents
with the ability to select from three options: that climate change is not real, that climate change is
real but is not caused by humans, and that climate change is real and is caused by humans.
Respondents were also allowed to choose not to answer. Ninety-eight percent of respondents
(1,797 out of 1,842) answered this question. Figure 4.6 uses 1,797 as the basis for its age range
ratios. That is, the percentage makeup by climate change belief is of those who provided answers
to this question, not all 1,842 responses.
The data show that solar adoption in upstate New York is largely comprised of people
who believe climate change is caused by humans (84.4 percent). Another 11.8 percent believe
that climate change is real but that humans do not cause it. Combined, this shows that an
overwhelming majority (96.2 percent) of solar adopters in upstate New York believe that climate
57
change is real. The data is displayed graphically in Figure 4.6, which helps to highlight the
substantial makeup of respondents who believe that climate change is happening and, to a
slightly lesser degree, is caused by humans.
Figure 4.6. Solar Adoption by Climate Change Beliefs.
While the survey for this research shows 96.2 percent belief in climate change, it is
important to put this into proper context. To that end, a recent survey by Yale University shows
that 77 percent of adult New York State residents believe that global warming is happening
(Yale Climate, 2018). Two things should be noted about this. First, the Yale survey specifically
refers to global warming, just one component of climate change. Second, the Yale survey does
not distinguish between upstate New York and the rest of New York, but rather considers the
State in its entirety. Despite these differences, when comparing 96.2 percent of upstate New
York solar adopters to 77 percent New York State residents, an inference may be drawn that
solar adopters are more likely to believe in climate change than other New York State residents.
Political Affiliation and Solar Adoption. Political Affiliation was another demographic
component analyzed. The survey provided respondents with the ability to select between
58
Democrat, Republican, other, and prefer not to answer. Eighty-five percent (1,568 of the 1,842)
of respondents indicated a political affiliation. Figure 4.7 uses 1,568 as the basis for its political
affiliation ratios. That is, the percentage makeup by political affiliation is of those who provided
political affiliation data, not all 1,842 responses.
The data shows that nearly half (48.7 percent) of the responses identified as Democrats as
compared to approximately one in five (21.9 percent) who identify as Republican and one in four
(25.2 percent) who declare an “other” political affiliation. This data is displayed graphically in
Figure 4.7, which helps to highlight the substantial makeup of respondents who are Democrats.
Figure 4.7. Solar Adoption by Political Affiliation.
Based on the survey, two times more respondents identify as Democrats than as
Republicans. This is important when compared to the political affiliation of the general
population of upstate New York. New York’s voter registration data was analyzed to determine
the political party affiliation for the active voter registrations for the 55 counties that comprise
the upstate New York solar market. The registration data shows that 50 percent of the upstate
voters were Democrats, nearly identical to the percentage of Democrats who adopted solar power
59
systems. However, the registration data showed that 40 percent of the upstate voters were
Republicans, nearly twice the percentage of solar power system adopters who identified as
Republicans (Board of Elections, 2018). Per capita, this sample includes a lower percentage of
Republicans who have installed solar power systems. The group who identified their political
affiliation as other than purely Democrat or Republican was also surprisingly large. The other
group made up just eleven percent of the voter registrations (Board of Elections, 2018), but over
25 percent of the adopters.
Gender and Solar Adoption. Gender was another demographic component analyzed.
The survey provided respondents with the ability to select between male, female, other, or prefer
not to answer. Ninety-nine percent (1,822 of the 1,842) of respondents indicated their gender.
Figure 4.8 uses 1,822 as the basis for its gender ratios. That is, the percentage makeup by gender
is of those who provided gender data, not all 1,842 responses.
The data shows that just more than one in four (26.8 percent) of the responses identified
as female as compared to approximately three out of four (73.0 percent) who identify as male,
and three responses (0.2 percent) who declare an “other” gender. This data is displayed
graphically in Figure 4.8, which helps to highlight the majority of respondents who are male.
60
Figure 4.8. Solar Adoption by Gender.
Based on the survey, respondents identify as male at a rate of almost three times that of
female. Clearly, the results show the high ratio of respondents who are male. This is high from a
relative standpoint when compared to the general population, since females actually outnumber
males. For the general population of upstate New York, the makeup of the population is 49
percent male and 51 percent female (US Census Bureau, 2018).
Marital Status and Solar Adoption. Marital status was the final demographic
component analyzed. The survey provided respondents with the ability to select between
married, single, or prefer not to answer. Ninety-seven percent (1,795 of the 1,842) of respondents
indicated their marital status. Figure 4.9 uses 1,795 as the basis for its marital status ratios. That
is, the percentage makeup by marital status is of those who provided marital status data, not all
1,842 responses.
The data shows that approximately 83 percent of the responses identify as married as
compared to approximately 17 percent who identify as single. This data is displayed graphically
in Figure 4.9, which helps to highlight the majority of respondents who are married.
61
Figure 4.9. Solar Adoption by Marital Status.
Based on the survey, respondents identify as married at a rate of almost five times that of
single. Clearly, the results show the high ratio of respondents who are married. This is high from
a relative standpoint when compared to the general population, since in upstate New York, just
49 percent of the population is married (US Census Bureau, 2018).
Descriptive Analysis: Decision Factors
Participants were asked to rank from 1 to 5 the importance of 10 decision factors that
represent elements that influenced their decision to acquire a solar power system for their home.
Additionally, respondents were provided the option to name another decision factor of their own
creation and rate it as well. Table 4.2 describes the ranking scale.
62
Table 4.2
Importance Ratings of Decision Factors and Their Ordinal Rank
With the ordinal value assigned to the importance of each factor for each survey
response, the mean value of each decision factor was calculated. These are displayed graphically
in Figure 4.10.
Figure 4.10. Rated Importance of Decision Factors.
The calculation of means shows that the most important overall factor in the decision to
adopt solar power systems was that solar power has a positive impact on the environment
Ordinal
Importance Rating Rank
Not at all important 1
Slightly important 2
Moderately important 3
Very important 4
Extremely important 5
63
(μ=4.25). This was followed closely by the expectation of a reduction in energy bills (μ=4.18).
This strikes a nearly fair balance between environmental and economic considerations, with the
environment ranked as slightly more important. These two factors are considered the highest tier
factors in importance.
The factors in the next tier of importance were perceived honesty of the solar sales
representative (μ=3.90) and reduced dependency on the power company (μ=3.89). The next tier
was comprised of leaving a positive legacy (μ=3.59) and reputation of the solar installer
(μ=3.58). The factors fall off in importance after that, with recommendation for solar power from
peers being the least important factor (μ=2.30).
In addition to the 10 specific decision factors for which the survey collected importance
ratings, respondents were given the opportunity to rank another non-listed decision factor of their
own choosing. Respondents who chose to do this wrote in a description of the other decision
factor and provided an importance rating. Approximately 17 percent of respondents (307 of
1,842) indicated another factor. For those who did this, the other decision factor descriptions
varied widely, and many of those descriptions were exact matches to or closely related to the ten
decision factors listed in the survey. The 307 other responses had a high mean importance score
of 4.08. This leads to the inference that the category was used by some respondents to emphasize
specific decision factors. The other factors were grouped into three categories: solar economics,
environmental impact, and undefined. These were coded based on the text provided by
respondents indicating certain themes, which were generally either cost (economic) related or
environmental benefits related. Examples of the economics category include “good investment“
and “tax breaks.” Examples of the environmental impact category include “global warming” and
“environmental.” The unspecified factors were diverse and generally unfit for analysis.
64
Examples include “I signed the loan agreement the day Donald Trump was elected” and
“retired”. Table 4.3 shows the number of other factors in each of these categories and the mean
importance rating of each.
Table 4.3
Other Decision Factors and Their Mean Importance Ratings
In a follow-up question, survey participants were asked to select the top three most
important decision factors. Participant responses were summed for each decision factor. The top
two most common selections were also the two decision factors with the highest mean
importance ratings. The first and second decision factors are reversed when compared to the
mean of the importance ratings. These are shown in Table 4.4.
Other Factor Count Mean
Solar Economics 150 4.06
Environmental Impact 24 4.04
Unspecified 133 4.12
Total 307 4.08
65
Table 4.4
Frequency of Decision Factors Selected as “Three Most Important”
Expected energy bill reduction and positive impact on the environment are far more
important than the other factors based on this second approach. Most of the factors had similar
importance between both approaches to measuring importance, with one key exception. In the
overall ratings shown in 4.8, low or no up-front cost ranked eighth. However, in the top three
approach, it ranked as the third most selected decision factor. This is discussed in the discussion
section at the end of Chapter 5.
The bar chart discussed earlier in Figure 4.10 displays the mean importance ratings for
the entire pool of 1,842 responses. The next step was to calculate the mean importance ratings
for the factors, grouped by demographic factors. In this manner, any difference in the importance
of decision factors between demographic groups can be identified. Table 4.5 displays the
importance factors by race. Since the racial makeup of the pool of solar adopters is largely White
(94.2 percent), the other race groups do not have material impact on the overall importance
rankings. That is, the average rankings for the White race group drive the average ratings for the
Rank Decision Factor Count
1 Expected energy bill reduction 1,452
2 Positive impact on the environment 1,298
3 Low or no up-front cost 757
4 Reduced dependency on my power company 667
5 Leaving a positive legacy 369
6 The timing of life invents enabled my solar installation 272
7 Reputation of my solar installer 158
7 Perceived honesty of my solar sales representative 158
9 Other (if designated above) 148
10 Exposure to public information about solar 113
11 Recommendation for solar power from peers 71
66
entire pool. Indeed, the ranking of the importance for the White adopters is in the exact order of
the overall respondent rankings.
Table 4.5
Decision Factor Importance Ratings by Race
Table 4.6 shows importance ratings by education. The results of this show that a
reduction of energy bills is the most important for respondents with an Associate’s degree or
less. For those with Bachelor’s degrees, it is a virtual tie between energy bill reduction and
positive impact on the environment. However, for those with a Master’s degree or higher, the
environment is the most important factor.
Decision Factor Am
eric
an I
ndia
n o
r
Ala
ska
Nat
ive
Asi
an
Bla
ck o
r A
fric
an
Am
eric
an
Nat
ive
Haw
aiia
n o
r
Oth
er P
acif
ic I
slan
der
Oth
er
Whit
e
Mix
ed
n (sample size) 2 27 16 1 37 1,669 19
Low or no up-front cost 2.50 3.52 4.31 5.00 3.46 3.10 2.58
Expected energy bill reduction 4.50 4.07 4.38 4.00 4.08 4.18 4.16
Positive impact on the environment 3.50 4.15 4.31 1.00 4.16 4.27 4.42
Leaving a positive legacy 3.50 3.81 3.94 1.00 3.73 3.60 3.53
Recommendations for solar power from peers 2.00 2.33 2.44 1.00 2.27 2.30 2.74
Reputation of my solar installer 3.50 3.37 3.50 4.00 3.22 3.60 3.74
Perceived honesty of my solar sales rep 4.00 3.63 3.63 5.00 3.62 3.91 4.32
Exposure to public information about solar 3.00 3.30 4.00 2.00 3.32 3.32 3.58
Reduced dependency on my power company 5.00 3.37 4.19 5.00 3.81 3.89 4.05
The timing of life events enabled my solar installation 2.50 2.44 3.13 5.00 2.95 3.07 3.32
Race
67
Table 4.6
Decision Factor Importance Ratings by Education
Table 4.7 shows importance ratings by age of the solar adopter. The results of this show
that for older adopters, the environment is more important than energy bill reduction.
Table 4.7
Decision Factor Importance Ratings by Age
Decision Factor No
hig
h s
cho
ol
dip
lom
a o
r G
ED
Hig
h s
cho
ol
dip
lom
a o
r G
ED
Ass
oci
ates
Deg
ree
Bac
hel
ors
Deg
ree
Mas
ters
Deg
ree
Do
cto
ral
Deg
ree
n (sample size) 6 187 229 555 579 266
Low or no up-front cost 4.67 3.52 3.42 3.14 3.03 2.67
Expected energy bill reduction 4.83 4.35 4.43 4.22 4.14 3.82
Positive impact on the environment 3.83 3.87 3.94 4.21 4.39 4.57
Leaving a positive legacy 3.83 3.26 3.29 3.56 3.72 3.86
Recommendations for solar power from peers 3.67 2.13 2.38 2.23 2.39 2.21
Reputation of my solar installer 4.50 3.66 3.76 3.56 3.55 3.41
Perceived honesty of my solar sales rep 4.67 4.11 4.08 3.88 3.81 3.79
Exposure to public information about solar 3.83 3.20 3.35 3.35 3.35 3.27
Reduced dependency on my power company 4.83 4.12 4.10 3.86 3.89 3.55
The timing of life events enabled my solar installation 3.67 3.18 3.14 3.00 3.15 2.85
Education Level
Decision Factor 18-29 30-39 40-49 50-59 60-69 70+
n (sample size) 42 227 355 511 535 156
Low or no up-front cost 3.64 3.36 3.27 3.09 2.96 2.93
Expected energy bill reduction 4.29 4.20 4.20 4.20 4.17 4.04
Positive impact on the environment 3.76 4.08 4.19 4.24 4.35 4.44
Leaving a positive legacy 3.19 3.41 3.49 3.58 3.70 3.79
Recommendations for solar power from peers 2.55 2.23 2.36 2.25 2.32 2.24
Reputation of my solar installer 3.43 3.52 3.51 3.53 3.69 3.67
Perceived honesty of my solar sales rep 3.95 3.84 3.78 3.89 3.99 4.02
Exposure to public information about solar 2.88 3.15 3.20 3.40 3.41 3.44
Reduced dependency on my power company 3.86 3.78 3.81 3.97 3.90 3.89
The timing of life events enabled my solar installation 2.79 3.04 2.88 3.09 3.20 3.05
Age
68
Table 4.8 shows importance ratings by household income. The results show that adopters
in the lowest and highest income ranges rated the importance of positive impact on the
environment lower than other groups did.
Table 4.8
Decision Factor Importance Ratings by Household Income
Table 4.9 shows importance ratings by home value. The results show a close mean rating
for bill reduction and the environment across the ranges of home value.
Decision Factor $0
to
$5
0,0
00
$5
0,0
00
to
$1
00
,00
0
$1
00
,00
0 t
o
$2
00
,00
0
$2
00
,00
0 t
o
$3
00
,00
0
$3
00
,00
0 t
o
$4
00
,00
0
$4
00
,00
0 t
o
$5
00
,00
0
Ab
ov
e
$5
00
,00
0
n (sample size) 125 591 722 142 34 14 27
Low or no up-front cost 3.51 3.23 3.10 2.77 2.21 2.57 2.37
Expected energy bill reduction 4.24 4.17 4.22 3.96 3.91 3.79 4.00
Positive impact on the environment 4.22 4.26 4.23 4.31 4.59 4.14 4.15
Leaving a positive legacy 3.49 3.66 3.57 3.50 3.79 3.64 3.52
Recommendations for solar power from peers 2.39 2.32 2.33 2.11 2.32 1.79 1.93
Reputation of my solar installer 3.46 3.59 3.58 3.49 3.53 3.14 3.33
Perceived honesty of my solar sales rep 3.87 3.96 3.85 3.74 4.09 3.79 3.56
Exposure to public information about solar 3.42 3.38 3.33 3.18 3.35 3.21 2.89
Reduced dependency on my power company 4.11 3.95 3.85 3.65 3.59 3.21 3.56
The timing of life events enabled my solar installation 3.22 3.27 3.01 2.68 2.79 2.79 2.59
Annual Household Income
69
Table 4.9
Decision Factor Importance Ratings by Home Value
Table 4.10 shows importance ratings by climate change beliefs. The results show that
those who believe that climate change is real and caused by humans rank the importance of the
environment higher than the other factors. All other respondents rank energy bill reductions as
the most important.
Decision Factor Under
$10
0,0
00
$10
0,0
00
to
$19
9,9
99
$20
0,0
00
to
$29
9,9
99
$30
0,0
00
to
$39
9,9
99
$40
0,0
00
to
$49
9,9
99
$50
0,0
00
to
$59
9,9
99
$60
0,0
00
to
$69
9,9
99
Over
$70
0,0
00
n (sample size) 77 528 534 315 138 81 44 66
Low or no up-front cost 3.49 3.24 3.12 2.98 3.07 2.95 3.09 2.79
Expected energy bill reduction 4.14 4.20 4.22 4.07 4.26 4.01 4.36 4.18
Positive impact on the environment 4.34 4.24 4.19 4.31 4.28 4.26 4.34 4.33
Leaving a positive legacy 3.77 3.56 3.51 3.70 3.54 3.70 3.70 3.59
Recommendations for solar power from peers 2.47 2.38 2.26 2.39 2.23 2.14 2.11 2.09
Reputation of my solar installer 3.42 3.56 3.63 3.59 3.64 3.52 3.70 3.30
Perceived honesty of my solar sales rep 3.97 3.91 3.89 3.90 4.01 3.88 4.05 3.48
Exposure to public information about solar 3.35 3.41 3.34 3.31 3.17 3.15 3.45 3.15
Reduced dependency on my power company 4.09 3.94 3.90 3.85 3.92 3.68 3.86 3.65
The timing of life events enabled my solar installation 3.62 3.24 3.03 2.92 2.87 2.63 2.98 2.73
Home Value
70
Table 4.10
Decision Factor Importance Ratings by Climate Change Beliefs
Table 4.11 shows importance ratings by political affiliation. The data indicates three key
things: Democrats rate the importance of the environment higher than the other factors,
Republicans rate the importance of bill reductions higher than the other factors, and those in the
other category rank these two factors equally, in a tie for the most important decision factor.
Decision Factor It's
no
t re
al
It's
rea
l b
ut
is n
ot
cau
sed
by h
um
ans
It's
rea
l an
d i
s
cau
sed
by h
um
ans
n (sample size) 68 212 1,517
Low or no up-front cost 3.54 3.53 3.05
Expected energy bill reduction 4.56 4.40 4.13
Positive impact on the environment 3.19 3.26 4.44
Leaving a positive legacy 2.51 2.59 3.78
Recommendations for solar power from peers 2.16 2.13 2.33
Reputation of my solar installer 3.53 3.62 3.58
Perceived honesty of my solar sales rep 3.91 3.85 3.91
Exposure to public information about solar 3.07 3.04 3.38
Reduced dependency on my power company 3.99 3.86 3.88
The timing of life events enabled my solar installation 2.90 3.08 3.07
Climate Change Beliefs
71
Table 4.11
Decision Factor Importance Ratings by Political Affiliation
Table 4.12 shows importance ratings by gender. The data indicates a fundamental
difference between the male and female respondents. Females in the sample rate positive impact
on the environment as the most important decision factor. Men in the sample rate expected
energy bill reduction as the most important decision factor.
Decision Factor Dem
ocr
at
Rep
ub
lica
n
Oth
er
Mult
iple
n (sample size) 763 344 395 66
Low or no up-front cost 2.98 3.39 3.04 3.15
Expected energy bill reduction 3.99 4.47 4.17 4.23
Positive impact on the environment 4.63 3.63 4.17 4.09
Leaving a positive legacy 3.98 2.96 3.48 3.42
Recommendations for solar power from peers 2.36 2.21 2.21 2.18
Reputation of my solar installer 3.59 3.62 3.50 3.48
Perceived honesty of my solar sales rep 3.90 3.97 3.81 3.94
Exposure to public information about solar 3.42 3.17 3.27 3.12
Reduced dependency on my power company 3.79 4.05 3.90 3.80
The timing of life events enabled my solar installation 3.03 2.99 3.14 3.21
Political Affiliation
72
Table 4.12
Decision Factor Importance Ratings by Gender
Table 4.13 shows importance ratings by marital status. The data indicates a difference
between married and single respondents. Married members of the sample indicate that expected
energy bill reduction is the most important decision factor. Single respondents find positive
impact on the environment to be the most important.
Decision Factor Fem
ale
Mal
e
Oth
er
n (sample size) 489 1,330 3
Low or no up-front cost 3.22 3.08 3.67
Expected energy bill reduction 4.16 4.19 4.67
Positive impact on the environment 4.56 4.14 3.33
Leaving a positive legacy 3.90 3.47 2.00
Recommendations for solar power from peers 2.55 2.20 2.00
Reputation of my solar installer 3.72 3.54 2.33
Perceived honesty of my solar sales rep 4.01 3.86 2.67
Exposure to public information about solar 3.55 3.24 2.00
Reduced dependency on my power company 4.06 3.83 3.33
The timing of life events enabled my solar installation 3.22 3.01 2.33
Gender
73
Table 4.13
Decision Factor Importance Ratings by Marital Status
Correlations of Decision Factors to Demographics
In addition to identifying the most important decision factors to solar adopters, this
analysis was designed to calculate correlations between demographic parameters (independent
variables) and the importance of decision factors (dependent variables). Table 4.14 shows the
correlation coefficient ρ-values between four demographic factors (age, education, home value,
and income) and the importance assigned to the decision factors. The table also shows the
number of samples used in each calculation (N) and the statistical significance calculated with
the two-tailed t-test all produced by SPSS.
Decision Factor Mar
ried
Sin
gle
n (sample size) 1,495 300
Low or no up-front cost 3.11 3.14
Expected energy bill reduction 4.17 4.20
Positive impact on the environment 4.22 4.37
Leaving a positive legacy 3.57 3.66
Recommendations for solar power from peers 2.27 2.40
Reputation of my solar installer 3.59 3.55
Perceived honesty of my solar sales rep 3.89 3.97
Exposure to public information about solar 3.31 3.39
Reduced dependency on my power company 3.85 4.04
The timing of life events enabled my solar installation 3.04 3.15
Marital Status
74
Table 4.14
Correlation between Decision Factors and Key Demographics
Note. A single asterisk (*) indicates statistical significance at the 0.05 level. A double asterisk
(**) indicates statistical significance at the 0.01 level.
Interpreting this table relies on an understanding of both correlation and significance.
“Virtually all of the measures of relationships between variables are designed not only to show
the degree of association between the items, but also to report the statistical significance of the
Variable Parameter Education Age Income Home Value
Correlation Coefficient -.177**
-.119**
-.143**
-.087**
Sig. (2-tailed) 0.000 0.000 0.000 0.000
N 1822 1826 1655 1783
Correlation Coefficient -.178** -0.044 -.049
* -0.017
Sig. (2-tailed) 0.000 0.062 0.046 0.469
N 1822 1826 1655 1783
Correlation Coefficient .210**
.115** -0.001 0.004
Sig. (2-tailed) 0.000 0.000 0.978 0.856
N 1822 1826 1655 1783
Correlation Coefficient .150**
.099** -0.023 0.010
Sig. (2-tailed) 0.000 0.000 0.346 0.673
N 1822 1826 1655 1783
Correlation Coefficient 0.011 -0.007 -0.048 -.055*
Sig. (2-tailed) 0.647 0.770 0.052 0.019
N 1822 1826 1655 1783
Correlation Coefficient -.089**
.073** -0.016 0.006
Sig. (2-tailed) 0.000 0.002 0.523 0.788
N 1822 1826 1655 1783
Correlation Coefficient -.116**
.074**
-.058* -0.021
Sig. (2-tailed) 0.000 0.001 0.017 0.384
N 1822 1826 1655 1783
Correlation Coefficient 0.007 .096**
-.054*
-.054*
Sig. (2-tailed) 0.772 0.000 0.027 0.024
N 1822 1826 1655 1783
Correlation Coefficient -.124** 0.025 -.100
**-.050
*
Sig. (2-tailed) 0.000 0.284 0.000 0.033
N 1822 1826 1655 1783
Correlation Coefficient -0.038 .065**
-.127**
-.135**
Sig. (2-tailed) 0.106 0.005 0.000 0.000
N 1822 1826 1655 1783
Reputation of my
solar installer
Perceived honesty
of my solar sales
representative
Exposure to public
information about
solar
Reduced
dependency on my
power company
The timing of life
events enabled my
solar installation
Independent Variable
Low or no up-front
cost
Expected energy bill
reduction
Positive impact on
the environment
Leaving a positive
legacy
Recommendations
for solar power
from peers
75
relationship. The word significance has a very special and consistent meaning when used in this
context. It does not mean important! If a relationship between two variables is statistically
significant, this simply means it signifies or signals there’s a good chance the two items are
actually related to one another in the population, just as they are in the sample” (Alreck and
Settle, 2017, page 304). Significance levels below 0.05 indicate that there is less than a five
percent chance that the relationship between two variables is caused by random chance.
Significance levels below 0.01 indicate that there is less than a one percent chance that the
relationship between two variables is caused by random chance.
There were 40 variables tested for correlation. Alreck and Settle (2017) describe that a
positive correlation coefficient indicates that the variables move in the same direction—that is, as
one goes up, so does the other. In a negative correlation, as one variable increases, the other
declines. Of the 40 variables tested, 25 were shown to be statistically significant. However, the
correlation coefficients were low, ranging from -0.178 to 0.210. While there is no universal
criteria that assigns a label of weak or strong to a correlation factor value, the closer to zero that
the coefficient is, the weaker the correlation. The closer to +1 that the coefficient is, the stronger
the positive correlation is, and the closer to -1 that the coefficient is, the stronger the negative
correlation is. Hinkle, et al. (2002) provide guidelines for interpreting the strength of correlation
according to correlation coefficient values in Table 4.15.
76
Table 4.15
Rule of Thumb for Interpreting the Size of a Correlation Coefficient
Note. From Hinkle et al. (2002).
Based on the guidelines in Table 4.15, each correlation coefficient tested in Table 4.14 lies in the
category of little, if any, correlation. This indicates that none of the correlations are strong
enough to indicate a meaningful relationship between the demographic parameters and the
importance ratings of the decision factors.
Decision Factor Ratings Over Time
The wealth of data from the large number of survey responses enabled this analysis to
consider how importance of decision factors has evolved over time. The study horizon
considered survey adopters for the five-year period from 2013 through 2017. Table 4.16 shows
the average importance rating of each decision factor for each year of installation represented by
the sample.
Size of Correlation Interpretation Size of Correlation Interpretation
.90 to 1.00 (-.90 to –1.00) Very high positive (negative) correlation
.70 to .90 (-.70 to -.90) High positive (negative) correlation
.50 to .70 (-.50 to -.70) Moderate positive (negative) correlation
.30 to .50 (-.30 to -.50) Low positive (negative) correlation
.00 to .30 (.00 to -.30) Little if any correlation
77
Table 4.16
Decision Factor Importance by Year of Installation
There are three key takeaways from this data. First, positive impact on the environment
and expected energy bill reduction maintained their respective number one and number two
rankings throughout the study period. Second, the rankings for most decision factors remained
relatively flat over the study horizon. Third, two of the less important decision factors over the
entire period showed the largest change between 2013 and 2017. Recommendations for solar
power from peers and reputation of my solar installer were the fastest growing decision factors
in terms of importance as shown in the shaded area of the Change column of Table 4.16. These
are also displayed graphically in Figures 4.11 and 4.12, respectively.
Decision Factor 2013 2014 2015 2016 2017 Change
Low or No Up-Front Cost 3.02 3.11 3.08 3.17 3.14 4.0%
Expected Energy Bill Reduction 4.15 4.20 4.21 4.20 4.13 -0.5%
Positive Impact on the Environment 4.19 4.23 4.22 4.28 4.26 1.8%
Leaving a Positive Legacy 3.47 3.56 3.54 3.65 3.62 4.2%
Recommendations for Solar Power from Peers 2.11 2.22 2.30 2.34 2.36 11.7%
Reputation of My Solar Installer 3.34 3.56 3.63 3.60 3.62 8.4%
Perceived Honesty of My Solar Sales Representative 3.83 3.88 3.91 3.91 3.90 2.0%
Exposure to Public Information About Solar 3.33 3.38 3.37 3.32 3.24 -2.7%
Reduced Dependency on My Power Company 3.77 3.78 3.89 3.97 3.89 3.2%
The Timing of Life Events Enabled My Solar Installation 2.97 3.07 3.04 3.07 3.11 4.9%
Other 3.07 2.78 3.06 2.60 2.89 -5.9%
78
Figure 4.11. Importance of Peer Recommendations for Solar Over Time.
Figure 4.12. Importance of Solar Installer Reputation Over Time.
Summary
Based on the analysis described earlier in this chapter, several key findings were
established. These are broken down into the type of analysis performed: demographic
79
assessment, descriptive statistics on importance of decision factors, correlation analysis, and time
series analysis.
In the demographic assessment, race stands out. While the population of upstate New
York is 82 percent White, 94 percent of the solar adopters surveyed were White. Another key
element is the role of education, which showed that this sample of solar adopters in upstate New
York is far more educated than the general population of the region. Whereas 33 percent of the
general population of upstate New York has a Bachelor’s degree or higher, 77 percent of the
solar-adopting population in upstate New York holds a Bachelor’s degree or higher. Moreover,
14 percent of upstate New York solar adopters surveyed held a Doctorate level degree, compared
to approximately two percent of the general population of upstate New York holding a
Doctorate.
Another component of demographics was beliefs related to climate change. The data
clearly showed that the overwhelming majority of solar adopters (96 percent) believe that
climate change is real. This compares to 77 percent of the general population with the same
belief. Finally, the political affiliation of adopters in the sample indicates an underrepresentation
of Republicans, when compared to demographics of the state, and overrepresentation of people
who indicate other as their political affiliation. The makeup of Democrats is approximately 50
percent in both the solar-adopting and general populations.
Descriptive statistics were used to assess the importance ratings of decision factors. The
hypothesis was that expected energy bill reduction would be the most important. However, this
hypothesis was not supported. Overall, the positive impact that solar has on the environment was
consistently more important than the economic factors. The expectation of lower energy costs
was the second most important factor over all, but the up-front cost factor came in a distant
80
eighth. Positive impact on the environment and expected energy bill reduction were consistently
the top two in importance across demographic groups.
Correlations were calculated and proved disappointing. The linkage between the
demographic data and importance ratings of decision factors was weak. Finally, the importance
ratings of decision factors were grouped by year to identify trends or changes. Throughout the
five years of the period studied for solar adoption (2013-17), the top two decision factors
remained unchanged.
These conclusions are the culmination of the chain of evidence as displayed graphically
in Figure 4.13.
81
Figure 4.13. Chain of Evidence.
This analysis applied quantitative analysis techniques to data collected in a survey of
solar adopters in upstate New York. Several key findings were developed based on the
82
demographic analysis, descriptive statistics, correlation analysis, and time series analysis. These
findings are the basis for the discussion and recommendations made in Chapter 5.
83
Chapter 5: Conclusions and Recommendations
“…more energy from the sun falls on the earth in one hour than is used by everyone in
the world in one year” (NREL, 2018).
A tiny portion of the sun’s energy is captured by solar power systems that turn it into
electric energy. New York has aggressive targets for renewable energy as a percentage of total
electric energy production, per the CES. The CES has a stated goal of 50 percent of electrical
energy within the state to be produced by renewable resources by 2030, in which solar energy is
expected to play a key role. To that end, the NY-Sun program offers financial incentives and
informational support to New York residents and business owners to promote solar adoption.
This quantitative survey study investigated one specific sector of the renewable energy
industry: residential solar power system adoption (solar equipment installed for private residents
on their homes, and not including any commercial or industrial solar applications). The study
was designed to determine the importance of the factors that influence the homeowners to power
their homes with solar equipment and to determine how those factors differed by race, age,
educational level, and income. For upstate New York homeowners, this study shows that positive
impact on the environment is the most important decision factor when it comes to adopting
residential solar power systems, with expected energy bill reduction as a secondary factor. This
research also showed that the typical adopter is a highly educated White person who believes
that climate change is real.
Conclusions from the Data
The research questions that guided this investigation were:
1. What are the demographic indicators that describe the upstate New York residential
solar power system adopters?
84
2. What were the most important decision factors in solar power adoption?
3. How do the importance ratings of the decision factors correlate to demographic
indicators?
4. How have the decision factors changed in importance over time?
This analysis has answered those questions with clear—yet in some cases surprising—
findings that lead to conclusions about the upstate New York residential solar adopter population
that are backed by data.
Of the four research questions, the most important (and the primary reason for this
quantitative survey study) was what the most important decision factors are. This question served
as the basis for the hypothesis of the study, which is that the most important decision factor was
the economic factor expected energy bill reduction, which defines how important saving money
from solar energy was to the survey respondents. Thus, the hypothesis was that this decision
factor would be rated as the most important factor by respondents, as defined as having the
highest mean score of importance ratings. Of course, this hypothesis was not meant to suggest
that all respondents would rank this decision factor highest, but rather that when the mean rating
for the decision factor was calculated (Figure 4.10), this decision factor would have the top
rankings. The findings and conclusions related to this hypothesis will be discussed under the
second research question on the importance of decision factors.
What are the demographic parameters that describe the upstate New York residential
solar power system adopters?
The first research question (demographics) required determination of the demographic
makeup of the survey sample. The demographic parameters gathered by the survey were
compared to U.S. Census Bureau data for upstate New York, allowing a comparison of solar
85
adopters to the general public. The demographic findings covered race, age, education level
attained, household income, size of the home, beliefs about climate change, and political
affiliation.
Over 94 percent of the respondents were White, while the general population is 82
percent White. Based on these findings, the first conclusion is that not only is upstate New York
largely comprised of White people, but White people comprise an even larger share of this
sample of solar adopters.
Another key demographic finding is related to educational attainment. While 32 percent
of the general population of upstate New York holds a Bachelor’s degree or higher, 77 percent of
the sample have this level of education; the sample is more educated than the general population.
Further to the point that educational achievement is a characteristic of solar adopters, nearly 15
percent of the survey respondents held a Doctoral level degree, as compared to approximately
two percent for the general population of upstate New York.
Another key demographic finding is related to age. The survey respondents were asked to
provide their age range at the time of installation. More than half (57 percent) of the respondents
fell into the 50-59 and 60-69-year-old age ranges. This compares to just 34 percent of the general
population belonging to those age groups. Another age-related finding is that the youngest and
oldest age groups (18-29 and over 70) added little to the solar-adopting ranks, with a combined
10.8 percent of adopters in those categories. Census data show that these two groups make up
more than a third (35 percent) of the general population.
Another key demographic finding is related to household income. The annual income
range most represented in this sample of solar adopters was the $100,000-$200,000 category,
accounting for approximately 44 percent of the sample. However, the mean income for the
86
general population of upstate New York is approximately $60,000. These income-related
findings are similar to the findings associated with home value in that the highest and lowest
ranges are not well represented in the sample. The two survey ranges of home value that together
comprise a range of $100,000-$300,000 account for approximately 60 percent of the
installations.
The extent to which upstate New Yorkers believe that climate change is real also factors
into the solar adopter decision. While 77 percent of New Yorkers believe that climate change is
real, 96 percent of upstate New York solar adopters do. The logical conclusion is that believers
in climate change make up a much higher percentage of the sample than they do of the general
population. This is consistent with the fact that solar power systems play a part in reducing
greenhouse gas emissions, the assumed cause of global warming.
Another demographic finding is related to political affiliation, with a much higher
percentage of solar adopters being Democrats than Republicans by more than two to one. The
percentage of Democrats in the sample is almost the same as the percentage of Democrats in the
general population (approximately 50 percent). However, the percentage of solar adopters who
identify as Republicans (22 percent) is substantially less than Republicans in the general
population (40 percent). The demographic findings and conclusions are summarized in Table
5.1.
87
Table 5.1
Summary of Demographic Findings and Conclusions
What were the most important decision factors?
This quantitative analysis used survey data associated with how important the various
decision factors were to solar adopters. The participants were asked to rate how important each
of the following decision factors was to them, on a 1-5 scale from not important at all (1) to
extremely important (5):
▪ Low or no up-front cost
▪ Expected energy bill reduction
▪ Positive impact on the environment
Demographic Parameter Findings Conclusions
Race96 percent of the sample is White while 82
percent of the general population is White.
The general population is predominantly White;
the sample has an even higher ratio of whites
than the general population
Education
77 percent of the sample have a bachelors
degree, compared to 32 percent for the general
population. 15 percent have doctorates, as
compared to 2 percent of the general population.
The sample is far more educated than the
general population.
Age
57 percent of the respondents are between 50
and 69 years old. 34 percent of the general
population are in that range.
The sample has more people in the 50-69 age
range than the general population.
Household Income
Adopters with income between $100,000 and
$200,000 account for 44 percent of the sample,
while the mean income of the general population
is $60,000.
The population has a higher level of household
income than the general population.
Climate Change Beliefs
96 percent of upstate New York solar adopters
believe that climate change is real. 77 percent of
New Yorkers agree.
The sample has more believers in climate
change than the general population.
Politcal Affiliation
49 percent of the sample are democrats, about
the same as the general population. About 22
percent of the sample is Republican, compared
to 40 percent of the general population.
Sample members who are Democrats outweigh
Republicans by more than two to one due to low
representation of Republicans.
Marital Status83 percent of the sample is married and 49
percent of the general population is married.
The sample is comprised of far more married
people per capita than the general population.
88
▪ Leaving a positive legacy
▪ Recommendations for solar power from peers
▪ Reputation of my solar installer
▪ Perceived honesty of my solar sales representative
▪ Exposure to public information about solar
▪ Reduced dependency on my power company
▪ The timing of life events enabled my solar installation
The first two decision factors of the survey comprise economic factors, designed to
capture how important money-related factors were to the decision-making process. The other
eight factors are non-economic. The literature review suggested that the most important factors
in the decision to adopt solar power systems would be expected energy bill reduction—that
money played the most important role in the decision. For that reason, the hypothesis of this
research was that the most important decision factor to upstate New York solar adopters was
expected energy bill reduction. This hypothesis, however, was not supported based on the
calculation of mean importance rating of the ten decision factors.
The most important decision factor was positive impact on the environment, with a mean
importance score of 4.25. This was followed by expectation of lower energy bills, one of the two
economic factors. However, the other economic decision factor, low or no up-front cost, was the
eighth most important decision factor. The conclusion of decision factors is that the most
important solar decision factor was positive impact on the environment, followed by expected
energy bill reduction. Most importantly, this quantitative survey study concludes that the
hypothesis that the most important decision factor was expected energy bill reduction failed. It is
worth noting that when asked to identify the top three most important decision factors, expected
energy bill reduction was the most frequently selected, followed closely by positive impact on
the environment. Based on looking at the importance in two different ways, these top two
decision factors were far more important than the others.
89
Another important finding is the importance of perceived honesty of my solar sales
representative, which ranked third in importance when respondents were forced to select their
top three reasons for adopting. This decision factor has nothing to do with the environment,
economics, or solar itself, other than offering prospective adopters comfort that what the sales
representative tells them about those things is true. This decision factor had a mean importance
rating of 3.90.
How do the importance ratings of the decision factors correlate to demographic
parameters?
At the onset of this quantitative analysis, there were expectations that the rated
importance of decision factors would show strong correlations to the demographic data. For
example, one might expect a strong negative correlation between income and the importance of
expected energy bill reduction, because saving money would be more important in lower income
households. Four of the demographic parameters were tested for correlation to the decision
factors: age, education, household income, and home value (other demographic variables such as
race could not be properly assigned numerical, ordinal values with which to perform
correlations). The analysis produced correlation coefficients for the relationship between these
four demographic independent variables and the rated importance of the ten decision factors,
which are displayed in Table 4.14 of the previous chapter, resulting in 40 correlation coefficients
(ρ-values).
How have the decision factors changed in importance over time?
The survey data analyzed for this research reflected solar power system installations over
the five-year period of 2013-2017. As previously concluded, the two most important decision
factors over the entire period for all demographics were positive impact on the environment and
90
expected energy bill reduction, respectively. This research also sought to identify how the
importance factors have changed over time, since each survey respondent identified the year in
which their solar power system was installed.
The mean importance score for each decision factor was calculated for each year, and
these mean values were ranked from highest to lowest (one to ten). Over the five-year period,
these rankings remained virtually unchanged. Only two of the lower-ranked decision factors
(recommendations for solar power from peers and reputation of my solar installer) increased
materially in importance over time. The two most important decision factors maintained their
importance. In short, the key decision factors did not change substantially over time. This serves
to bolster the conclusion that positive impact on the environment and expected energy bill
reduction were clearly the most important decision factors for residential solar power system
adoption for the study period.
Discussion
This research resulted in conclusions about the demographic makeup of the population of
solar adopters, the importance of each decision factor, correlations between demographics and
decision factors, and decision factor trends. Each provides insights on the upstate New York
residential solar power system market.
Relevance of Conclusions to the Research Questions
This section elaborates on how the conclusions are relevant to the research questions that guided
the analysis.
Demographics. For demographics, one of the most striking features of the population of
solar power adopters was race. Specifically, this sample of solar adopters are overwhelmingly
91
White, representing over 94 percent. There are several possible contributors to that. First, as
noted previously, upstate New York has a large White population of 82 percent. However, other
factors must comprise the difference in the makeup of Whites in the general population and the
solar-adopting population. There are several potential reasons for this that this research does not
cover, which may be an opportunity for future research to help explain the gap. For one, it might
be shown that Whites have a higher percentage of home ownership, a very important factor in
solar adoption. It may also show that Whites have a higher level of educational attainment than
other races, a demographic parameter that is consistent with solar adoption.
Another key finding was related to household income, which showed that neither low
income nor high-income households are likely to adopt solar power systems. One potential
explanation for the lack of lower income adopters is that prospective adopters need either up-
front capital to pay for a solar power system or good credit to finance or lease a system. It can be
theorized that lower-income households are less likely to have either of those available to foster
solar power system installation. On the other hand, higher income households are also
underrepresented in this sample of solar adopters. A potential explanation for this is that high-
income individuals do not need the energy cost savings associated with solar power systems. The
expectation of lower energy bills decision factor was shown to be the second most important of
all factors overall, but it stands to reason that high income households would feel that saving
money on utility bills is less important than lower income families. In other words, high income
families do not need solar energy savings, and low income families are precluded from adopting
solar due to their financial standing.
Education played another key role in solar adoption, with a conclusion that solar adopters
are far more educated than the general population. One theory to explain this is the potential
92
(and likely) correlations between education and income and education and good credit.
Households with educated individuals, then, would be less likely to be part of the lower income
groups that can be disqualified as prospective adopters for income or credit reasons.
Another key demographic showed that solar adopters overwhelmingly (96 percent)
believe that climate change is real. This conclusion is logical and expected, since climate change
has been linked to greenhouse gases produced by fossil fuel combustion, which is reduced when
solar power systems displace traditional utility power service. That is, solar power systems are
generally accepted as contributing to the reduction of greenhouse gases and, therefore, climate
change.
Decision Factors. The rating of the importance of decision factors proved that overall,
and for each year of the analysis, positive impact on the environment was the most important to
solar adopters. This was followed by the expectation of lower energy bills. These are key
findings of this research due to their potential implications for policymaking and marketing by
solar installers.
The analysis suggests that economic factors were not the most influential consideration,
since the economic decision factors were not the most highly rated for importance. It is worth
noting that this research does not consider their true importance to the solar decision, because it
excluded non-adopters. Additional research might show that non-adopters did not adopt solar
due to unfavorable economics, which would make solar economics an extremely important
factor for non-adopters. The solar adopters in this research responded that positive impact on the
environment was the most important consideration to them.
Another important finding was the importance of perceived honesty of my solar sales
representative, which ranked third, closely trailed by reduced dependency on my power
93
company. The former is most intriguing, because it is the one that can be most directly controlled
by solar power companies. While policymakers likely cannot use this information, solar
installers likely can. Solar companies can focus their efforts on training sales representatives to
project an image of honesty to prospective solar adopters.
The environmental decision factor was not only the highest rated by the mean score for
the entire survey but was the highest in each year of installation over the analysis period. This
bolsters the conclusion that the environmental decision factor was clearly the most important. It
is worth noting that only two of the decision factor importance ratings were substantially
different over time (recommendations for solar power from peers and reputation of my solar
installer). Both of these two factors grew in importance between 2013 and 2017. There are a few
possible explanations.
First, earlier adopters (e.g., 2013) experienced higher NY-Sun incentives than the later
adopters (e.g., 2017), which may have led to better economics of solar power systems for the
earlier adopters (Appendix C). Based on this premise, later adopters would be those less
influenced by economics and more influenced by other factors, including recommendations for
solar power from peers and reputation of my solar installer. At the same time, as the number of
solar installations grew over time, there were more peers from whom to get recommendations
and more installations that shaped the reputations of solar installers. This would explain why
these two decision factors grew in importance over time during the analysis period. It is worth
noting that even though these two factors grew in importance over time, they were both not very
important to adopters. Their mean rankings even at their highest in 2017 were tenth and fifth out
of ten, respectively. It would be wise for policymakers and marketers to monitor how decision
factors change as incentives change and the industry grows.
94
It is worth noting the unexpected results of the importance of low or no up-front cost.
When asked to rate the importance of each decision factor, it was the eighth most important.
However, it was the third most frequently selected decision factor when survey participants were
asked to identify the top three reasons. While such a result is unexpected, it does not alter the
conclusion that the top two decision factors (positive impact on the environment and expected
energy bill reduction) are much more important than the others. The unexpected result for this
decision factor is potentially caused by a flaw in the survey that could lead to misleading results
about the decision factor. The issue is that not all survey participants have access to acquiring a
solar power system at zero or minimal up-front cost. In fact, there is no survey question to
capture whether such an option was available to the adopter at the time of their installation. All
other decision factors apply to all participants and can be truly rated for importance. The up-front
cost decision factor only applies to certain adopters. It may be useful for those doing future
research to acknowledge this and first inquire whether low or no up-front cost was applicable or
available before asking if it was important.
Correlations. The correlation coefficient calculations produced ρ-values that were low,
which indicated that there was not a tight relationship between the demographic factors
(independent variables) and decision factors (dependent variables). There are several potential
reasons for this.
First, the survey sample consisted solely of solar adopters, not a mix of adopters and non-
adopters. This was an essential part of the research because the key element was understanding
which of the decision factors were most important to adopters, and such a question would not
apply to a non-adopter. As such, the homogeneous pool of solar adopters possibly led to low
correlation coefficients. They share many demographic characteristics: the majority are White,
95
educated, believe in climate change, and fall into a relatively high income range. Therefore, they
may feel similarly about the importance of solar power decision factors with only slightly
different ratings based on the independent variables. For example, the ρ-value for the
relationship between household income and the importance of environmental impact was -0.001,
indicating practically no correlation. Given that environmental impact was the most important
decision factor overall, the logical explanation is that households of all income levels find
environmental impact to be very important to the solar adoption decision.
Relevance of Conclusions to the Literature
The primary conclusion of this quantitative survey study is that the most important
decision factor to upstate New York residential solar power system adopters is positive impact
on the environment followed by expectation of lower energy costs as a close second. The
literature suggested that both of these decision factors were important to adopters’ decisions, and
that was shown in this quantitative survey study.
Relevance of Conclusions to the Theoretical Framework
The key aspect of the theoretical framework of this quantitative survey study was the role
that relative advantage played in the upstate New York residential solar power system adopters’
importance ratings. The most important decision factors proved to be positive impact on the
environment, followed by expectation of lower energy cost. When viewing solar power systems
as an innovation relative to traditional grid power, relative advantage of this innovation drove
customer choice. For adopters, solar power systems offered the advantage of expected energy
bill reductions, environmental benefits, or both.
96
Recommendations from the Study
Certain conclusions of this quantitative survey study may be useful for both
policymaking purposes and for professional use by solar installers to assist in effective marketing
to prospective solar customers. Additionally, this research also may serve as the basis for
additional research. Recommendations for all of these are as follows:
Professional Use by Solar Installers
1. Since the most important decision factor for the sample was positive impact on the
environment, followed by expectation of lower energy costs, develop marketing
materials that focus on the environment as the primary focus and cost savings as a
secondary focus.
2. Focus on underserved races. The study indicated that non-White races are under-
represented in the sample. Focus marketing efforts to reach specific groups that are
currently underserved.
3. Consider targeting Democrats as prospective customers since they have been shown
to make up a portion of the sample consistent with the general population, whereas
Republicans do not.
4. Train sales representatives to project honesty to prospective solar adopters, given the
importance of honesty as a decision factor.
Policy Recommendations
1. Consider bolstering public information about solar energy incentives, environmental
benefits, and cost savings associated with solar. Public information was shown to be a
relatively unimportant solar decision factor, but this may be due to the lack of such
public information being provided, not its ineffectiveness if provided.
2. Focus on delivering public information to underserved populations such as non-
Whites.
3. Consider providing solar economic incentives or financing for low income families
who may not have credit to qualify for solar lease plans.
Recommendations for Further Research
1. Perform a similar study based on a survey of other geographic areas.
97
2. Perform a research project focusing on why non-solar adopters have not adopted
solar.
Conclusion
This research project studied the decision factors that influenced the adoption of
residential solar power systems in upstate New York. New York’s Clean Energy Standard (CES)
requires 50 percent of the State’s energy to be supplied by renewable resources by 2030, in an
effort to curb greenhouse gases and other deleterious emissions. The State fosters solar power
system adoption through incentives in the form of subsidies and financing. These subsidies,
however, are scheduled to be phased out by 2023. Thus, identifying the most important decision
factors is important.
Through an online survey of upstate New York residential solar power system adopters,
this research captured demographic data about the respondents and asked them to rate the
importance of the decision factors that influenced their decision to adopt a solar power system.
The analysis produced a demographic profile of the adopters and provided a descriptive
statistical analysis of the importance of the decision factors. Correlation coefficients were
calculated to determine the relationship between decision factors and demographic data. Finally,
the data was analyzed to identify how factors have changed over time. The goal of this research
was to contribute new information to the field for the benefit of policymakers and industry
players.
The results showed that the upstate New York residential solar power system adopters
were mostly White and college-educated. The analysis showed that the most important decision
factors were expected energy bill reduction and positive impact on the environment. The analysis
also showed that there were only minimal correlations between demographics and the
98
importance of the decision factors. Finally, the analysis showed that over the course of the five-
year analysis horizon, the importance ratings stayed practically the same. These insights may be
useful to policymakers and industry players to further adoption of residential solar power
systems in upstate New York.
99
References
Alafita, T., & Pearce, J. M. (2014). Securitization of residential solar photovoltaic assets: Costs,
risks and uncertainty. Energy Policy, 67, 488-498.
Alqahtani, B. J., Holt, K. M., Patiño-Echeverri, D., & Pratson, L. (2016). Residential Solar PV
Systems in the Carolinas: Opportunities and Outcomes. Environmental Science &
Technology, 50(4), 2082-2091.
Alreck, P. & Settle, R. (2017). The survey research handbook. Boston: McGraw-Hill/Irwin.
Balcombe, P., Rigby, D., & Azapagic, A. (2014). Investigating the importance of motivations
and barriers related to microgeneration uptake in the UK. Applied Energy, 130, 403-418.
Balnaves, M. & Caputi, P. (2001). Introduction to quantitative research methods : an
investigative approach. London Thousand Oaks, Calif: SAGE.
Bauner, C., & Crago, C. L. (2015). Adoption of residential solar power under uncertainty:
Implications for renewable energy incentives. Energy Policy, 86, 27-35.
Board of Elections. (n.d.). Retrieved May 23, 2018, from
http://www.elections.ny.gov/enrollmentcounty.html
Bollinger, B., & Gillingham, K. (2012). Peer Effects in the Diffusion of Solar Photovoltaic
Panels. Marketing Science, 31(6), 900-912.
Branker, K., Pathak, M. J. M., & Pearce, J. M. (2011). A review of solar photovoltaic levelized
cost of electricity. Renewable and Sustainable Energy Reviews, 15(9), 4470-4482.
Burns, J. E., & Kang, J.-S. (2012). Comparative economic analysis of supporting policies for
residential solar PV in the United States: Solar Renewable Energy Credit (SREC)
potential. Energy Policy, 44, 217-225.
100
Clean Energy Standard. (n.d.). Retrieved June 1, 2017, from https://www.nyserda.ny.gov/All-
Programs/Programs/Clean-Energy-Standard/Renewable-Portfolio-Standard
Coal and Air Pollution. (n.d.). Retrieved May 23, 2018, from https://www.ucsusa.org/clean-
energy/coal-and-other-fossil-fuels/coal-air-pollution#.WxrWBopKhPY
Cole, W., Mai, T., Eurek, K., Steinberg, D. C., & Margolis, R. (2015). Considering the Role of
Solar Generation under Rate-Based Targets in the EPA's Proposed Clean Power Plan.
The Electricity Journal, 28(8), 20-28.
Comptroller, N. Y. S. O. o. t. S. (2016). NY-Sun Incentive Program, Report 2015-S-91.
Creswell, J. W. (2013). Research Design. Qualitative, Quantitative, and Mixed Methods
Approaches. California: Sage Publications.
EnergySage. (2018, April 27). 2018 Cost of Solar Panels in New York State | EnergySage.
Retrieved June 30, 2018, from https://news.energysage.com/are-solar-panels-worth-the-
investment-in-new-york/
Feldman, D., Friedman, B., & Margolis, R. (2013). Financing, Overhead, and Profit: An In-
Depth Discussion of Costs Associated with Third-Party Financing of Residential and
Commercial Photovoltaic Systems - NREL/TP-6A20-60401: National Renewable Energy
Laboratory (NREL), Golden, CO.
Fellows, U. O. (2017, April 04). Despite Claims Of 'Grid Parity,' Wind And Solar Are Still More
Expensive Than Fossil Fuels. Retrieved May 23, 2018, from
https://www.forbes.com/sites/uhenergy/2017/04/04/despite-claims-of-grid-parity-wind-
and-solar-are-still-more-expensive-than-fossil-fuels/#65f447ae40e4
Flowers, M. E., Smith, M. K., Parsekian, A. W., Boyuk, D. S., McGrath, J. K., & Yates, L.
(2016). Climate impacts on the cost of solar energy. Energy Policy, 94, 264-273.
101
Fowler, L., & Breen, J. (2013). The Impact of Political Factors on States’ Adoption of
Renewable Portfolio Standards. The Electricity Journal, 26(2), 79-94.
Fowler, L., & Breen, J. (2014). Political Influences and Financial Incentives for Renewable
Energy. The Electricity Journal, 27(1), 74-84.
Gibson, B. (2013). What's Next for Solar. The Electricity Journal, 26(2), 51-57.
Graziano, M., & Gillingham, K. (2015). Spatial patterns of solar photovoltaic system adoption:
The influence of neighbors and the built environment ‡. Journal of Economic Geography,
15(4), 815-839.
Heeringa, S., West, B. & Berglund, P. (2017). Applied survey data analysis. Boca Raton, FL:
CRC Press, Taylor & Francis Group.
Hinkle, Wiersma, & Jurs. (2002). Applied Statistics for the Behavioral Sciences (5th ed.).
Houghton Mifflin.
Holmes, Andrew GD. (2014, March). Researcher Positionality - a consideration of its influence
and place in research. Retrieved from Research Gate.
https://www.researchgate.net/publication/260421552_Researcher_positionality_a_consid
eration_of_its
Kaufmann, R. K., & Vaid, D. (2016). Lower electricity prices and greenhouse gas emissions due
to rooftop solar: empirical results for Massachusetts. Energy Policy, 93, 345-352.
Kwan, C. L. (2012). Influence of local environmental, social, economic and political variables on
the spatial distribution of residential solar PV arrays across the United States. Energy
Policy, 47, 332-344.
Laws, N. D., Epps, B. P., Peterson, S. O., Laser, M. S., & Wanjiru, G. K. (2017). On the utility
death spiral and the impact of utility rate structures on the adoption of residential solar
102
photovoltaics and energy storage. Applied Energy, 185, Part 1, 627-641.
Letzelter, J. & Chupka, M. (1999). Surviving the SIP Call: Fossil Plant Economics Under NOx
Control. Public Utilities Fortnightly, 137(9), 44-51.
Mankiw, N. (2016). Principles of microeconomics. Australia: Cengage Learning.
Miller, R., Benjamin, D. & North, D. (2016). The economics of public issues. Boston: Pearson.
Noll, D., Dawes, C., & Rai, V. (2014). Solar Community Organizations and active peer effects in
the adoption of residential PV. Energy Policy, 67, 330-343.
NYSERDA. (n.d.). Retrieved July 08, 2018, from
https://www.nyserda.ny.gov/About/Newsroom/2018-Announcements/2018-06-18-
NYSERDA-Announces-Redesign-of-NY-Suns-Megawatt-Block-Program
NY-Sun (2018a). (n.d.). Retrieved May 23, 2018, from https://www.nyserda.ny.gov/All-
Programs/Programs/NY-Sun/Data-and-Trends
NY-Sun (2018b). (n.d.). Retrieved July 08, 2018, from https://www.nyserda.ny.gov/All-
Programs/Programs/NY-Sun/Contractors/Upstate-Dashboard
Ondraczek, J., Komendantova, N., & Patt, A. (2015). WACC the dog: The effect of financing
costs on the levelized cost of solar PV power. Renewable Energy, 75, 888-898.
Osborne, J. W. (2013). Best practices in data cleaning: A complete guide to everything you need
to do before and after collecting your data Thousand Oaks, CA: SAGE Publications Ltd
doi: 10.4135/9781452269948
Palm, A. (2017). Peer effects in residential solar photovoltaics adoption—A mixed methods
study of Swedish users. Energy Research & Social Science, 26, 1-10.
Patten, M. (2014). Questionnaire research : a practical guide. Glendale, CA: Pyrczak
Publishing.
103
Patten, M. & Bruce, R. (2014). Understanding research methods : an overview of the
essentials. Glendale, Calif: Pyrczak Publishing.
Rai, V., & Beck, A. L. (2017). Play and learn: Serious games in breaking informational barriers
in residential solar energy adoption in the United States. Energy Research & Social
Science, 27, 70-77.
Rai, V., Reeves, D. C., & Margolis, R. (2016). Overcoming barriers and uncertainties in the
adoption of residential solar PV. Renewable Energy, 89, 498-505.
Sample Size Calculator: Understanding Sample Sizes. (n.d.). Retrieved June 6, 2018, from
https://www.surveymonkey.com/mp/sample-size-calculator/
Sarzynski, A., et al. (2012). "The impact of state financial incentives on market deployment of
solar technology." Energy Policy 46: 550-557.
Schelly, C. (2010). Testing Residential Solar Thermal Adoption. Environment and Behavior,
42(2), 151-170.
Schelly, C. (2014a). Residential solar electricity adoption: What motivates, and what matters? A
case study of early adopters. Energy Research & Social Science, 2, 183-191.
Schelly, C. (2014b). Implementing renewable energy portfolio standards: The good, the bad, and
the ugly in a two state comparison. Energy Policy, 67, 543-551.
Simpson, G., & Clifton, J. (2015). The emperor and the cowboys: The role of government policy
and industry in the adoption of domestic solar microgeneration systems. Energy Policy,
81, 141-151.
Solar. (n.d.). Retrieved May 05, 2018, from
http://instituteforenergyresearch.org/topics/encyclopedia/solar/
Sommerfeld, J., Buys, L., & Vine, D. (2017). Residential consumers’ experiences in the adoption
104
and use of solar PV. Energy Policy, 105, 10-16.
US Census Bureau. (2018). Data. Retrieved April 22, 2018, from
https://www.census.gov/data.html
Varela-Margolles, A., & Onsted, J. (2014). Do Incentives Work?: An Analysis of Residential
Solar Energy Adoption in Miami-Dade County, Florida. Southeastern Geographer,
54(1), 18-35.
What’s the Difference Between Net Metering and Feed-In Tariffs? (n.d.). Retrieved June 1,
2017, from http://energyinformative.org/net-metering-feed-in-tariffs-difference
Yale Climate Opinion Maps - U.S. 2016. (n.d.). Retrieved May 23, 2018, from
http://climatecommunication.yale.edu/visualizations-data/ycom-us-
2016/?est=happening&type=value&geo=state&id=36
Zhai, P. (2013). Analyzing solar energy policies using a three-tier model: A case study of
photovoltaics adoption in Arizona, United States. Renewable Energy, 57, 317-322.
105
Appendix A: The Survey
Survey_of_Upstate_New_York_Residential_Solar_Customers
Survey Flow
Standard: Consent Form (1 Question)
Block: Section 1: Solar Home Information (15 Questions)
Standard: Section 2: Why Did You Install A Solar Power System? (4 Questions)
Standard: Section 3: Homeowner Information (11 Questions)
Page
Break
106
Start of Block: Consent Form
Q1
SURVEY OF NEW YORK STATE SOLAR CUSTOMERS
UNSIGNED CONSENT DOCUMENT FOR WEB-BASED ONLINE SURVEYS
Request to Participate in Research
We would like to invite you to participate in a web-based online survey. The survey is part of a
research study whose purpose is to investigate the factors that influence people to install solar
power systems on their homes. This survey should take about five minutes to complete.
We are asking you to participate in this study because you are a New York resident who installed
a solar power system on your home. You must be at least 18 years old to take this survey.
The decision to participate in this research project is voluntary. You do not have to
participate and you can refuse to answer any question. Even if you begin the web-based online
survey, you can stop at any time.
107
There are no foreseeable risks or discomforts to you for taking part in this study.
There are no direct benefits to you from participating in this study. However, your responses
may help us learn more about the factors that are important to people in deciding to install a solar
power system.
You will not be paid for your participation in this study.
Your participation in this study is anonymous to the researcher(s). However, because of the
nature of web based surveys, it is possible that respondents could be identified by the IP
address or other electronic record associated with the response. Neither the researcher nor
anyone involved with this survey will be capturing those data. Any reports or publications
based on this research will use only group data and will not identify you or any individual
as being affiliated with this project. The information you provide will be kept confidential
to the extent permitted by law.
If you have any questions regarding electronic privacy, please feel free to contact Mark
Nardone, NU’s Director of Information Security via phone at 617-373-7901, or via email at
If you have any questions about this study, please feel free to contact James Letzelter at 518-
727-0144 or by email at [email protected], the person mainly responsible for the
108
research. You can also contact Dr. Afi Wiggins at 334-309-6060 or by email at
[email protected], the Principal Investigator.
If you have any questions regarding your rights as a research participant, please contact
Nan C. Regina, Director, Human Subject Research Protection, 960 Renaissance Park,
Northeastern University, Boston, MA 02115. Tel: 617.373.4588, Email: [email protected].
You may call anonymously if you wish.
By clicking on the “I accept” button below, you are indicating that you consent to
participate in this study. Please print out a copy of this consent form for your records.
Thank you for your time.
Dr. Afi Wiggins
James Letzelter
o I accept (start survey)
o I do not accept (exit survey)
Skip To: End of Survey If SURVEY OF NEW YORK STATE SOLAR CUSTOMERS UNSIGNED CONSENT DOCUMENT FOR WEB-BASED ONLINE SURVEYS... = I do not accept (exit survey)
End of Block: Consent Form
Start of Block: Section 1: Solar Home Information
109
Q2 SECTION 1 of 3: HOME INFORMATION
The following questions are about the residence at which you installed a solar power system.
Please respond with answers valid for when the solar power system was installed.
Q3 What is the zip code of the residence?
________________________________________________________________
110
Q4 What is the county of the residence?
o Albany
o Allegany
o Bronx
o Broome
o Cattaraugus
o Cayuga
o Chautaua
o Chemung
o Chenango
o Clinton
o Columbia
o Cortland
o Delaware
o Dutchess
o Erie
o Essex
o Franklin
o Fulton
o Genesee
o Greene
111
o Hamilton
o Herkimer
o Jefferson
o Kings
o Lewis
o Livingston
o Madison
o Monroe
o Montgomery
o Nassau
o New York
o Niagara
o Oneida
o Onondoga
o Ontario
o Orange
o Orleans
o Oswego
o Otsego
o Putnam
o Queens
112
o Rensselaer
o Richmond
o Rockland
o Saratoga
o Schenectady
o Schoharie
o Schuyler
o Seneca
o St. Lawrence
o Steuben
o Suffolk
o Sullivan
o Tioga
o Tompkins
o Ulster
o Warren
o Washington
o Wayne
o Westchester
o Wyoming
o Yates
113
Q5 What is your electric utility company?
o Central Hudson Gas & Electric
o Consolidated Edison
o National Grid
o New York State Electric & Gas
o Orange & Rockland Utilities
o PSEG Long Island
o Rochester Gas & Electric
o Unknown/Other
Q6 In what year was the solar power system installed?
________________________________________________________________
Q7 What is the approximate above-ground square footage of the home?
________________________________________________________________
114
Q8 What is the approximate market value of the home ($)?
o Under $100,000
o $100,000 to $199,999
o $200,000 to $299,999
o $300,000 to $399,999
o $400,000 to $499,999
o $500,000 to $599,999
o $600,000 to $699,999
o Over $700,000
o Prefer not to answer
Q9 How many adults (18 year of age or older) lived in the home at the time of the solar
power system installation?
________________________________________________________________
115
Q10 How many children (under the age of 18) lived in the home at the time of the solar
power system installation?
________________________________________________________________
Q11 What energy source is used for your major equipment/appliances?
Electric Gas/Oil/Other
Not
Applicable
Heating/Furnace
o o o Hot Water
Heater o o o
Stove/Oven
o o o Clothes Dryer
o o o Pool Heater
o o o Other (describe
below if electric) o o o
116
Display This Question:
If What energy source is used for your major equipment/appliances? = Other (describe below if electric) [ Electric ]
Q12 Please describe the "other" major electric equipment in the previous question.
________________________________________________________________
________________________________________________________________
________________________________________________________________
________________________________________________________________
________________________________________________________________
Q13 Do you have air conditioning?
o Yes, central air conditioning
o Yes, window units
o Yes, both central and window units
o No
117
Q14 Was your initial contact with a solar company initiated by you or by a solar
company representative?
o By me
o By a solar company representative
o Other
Display This Question:
If Was your initial contact with a solar company initiated by you or by a solar company representative? = Other
Q15 If your answer to the previous question was "Other", please describe.
________________________________________________________________
Q16 How many solar companies did you get quotes from before deciding on a company?
________________________________________________________________
End of Block: Section 1: Solar Home Information
Start of Block: Section 2: Why Did You Install A Solar Power System?
Q17 SECTION 2 of 3: WHY DID YOU INSTALL A SOLAR POWER SYSTEM?
118
We would like to find out which factors were most important in your decision to install a solar
power system in your home.
119
Q18 Please rate the following factors in terms of their importance to your decision to
install a solar power system at your home.
120
N
ot at all
importan
t
Slig
htly
important
Modera
tely important
V
ery
important
Extre
mely
important
Low or no
up-front cost o o o o o
Expected
energy bill reduction o o o o o
Positive
impact on the
environment o o o o o
Leaving a
positive legacy o o o o o
Recommend
ations for solar
power from peers o o o o o
Reputation of
my solar installer o o o o o
Perceived
honesty of my solar
sales representative o o o o o
121
Exposure to
public information
about solar o o o o o
Reduced
dependency on my
power company o o o o o
The timing
of life events
enabled my solar
installation
o o o o o
Other
(describe below) o o o o o
Display This Question:
If Please rate the following factors in terms of their importance to your decision to install a sola... [ Other (describe below) ] (Recode) Is Not Empty
Q19 If you rated the importance of an "other" reason to install a solar power system
above, please describe the reason.
________________________________________________________________
122
Q20 Of the factors rated on importance above, which were the three most important in
your decision to install a solar power system?
▢ Low or no up-front cost
▢ Expected energy bill reduction
▢ Positive impact on the environment
▢ Leaving a positive legacy
▢ Recommendation for solar power from peers
▢ Reputation of my solar installer
▢ Perceived honesty of my solar sales representative
▢ Exposure to public information about solar
▢ Reduced dependency on my power company
▢ The timing of life invents enabled my solar installation
▢ Other (if designated above)
End of Block: Section 2: Why Did You Install A Solar Power System?
Start of Block: Section 3: Homeowner Information
123
Q21 SECTION 3 of 3: HOMEOWNER INFORMATION
The following questions are designed to gather demographic information about the primary
decision-maker responsible for installing a solar power system. Please respond to the questions
with answers that were valid at the time of the solar power system installation.
Q22 What was your age at the time of installing a solar power system?
o 18-29
o 30-39
o 40-49
o 50-59
o 60-69
o 70 or older
o Prefer not to answer
124
Q23 What was your marital status at the time of installing a solar power system?
o Single
o Married
o Prefer not to answer
Q24 What was your education level at the time of installing a solar power system?
o No high school diploma or GED
o High school diploma or GED
o Associates Degree
o Bachelors Degree
o Masters Degree
o Doctoral Degree
o Prefer not to answer
125
Q25 At the time of the solar power equipment installation, what was your total annual
household income range?
o $0 to $50,000
o $50,000 to $100,000
o $100,000 to $200,000
o $200,000 to $300,000
o $300,000 to $400,000
o $400,000 to $500,000
o Above $500,000
o Prefer not to answer
Q26 What is your gender?
o Male
o Female
o Other
o Prefer not to answer
126
Q27 What best describes your race (choose all that apply)?
▢ White
▢ Black or African American
▢ Asian
▢ American Indian or Alaska Native
▢ Native Hawaiian or Other Pacific Islander
▢ Other
▢ Prefer not to answer
Q28 Are you Hispanic or Latino?
o No
o Yes
o Prefer not to answer
127
Q29 What is your political affiliation (check all that apply)?
▢ Democrat
▢ Republican
▢ Other
▢ Prefer not to answer
Q30 Please select the statement that most closely aligns with your thoughts on climate
change:
o It's real and is caused by humans
o It's real but is not caused by humans
o It's not real
Q31 Thank you for your participation in this survey. If there is any other information you
would like to share regarding your solar power system, please submit it here:
________________________________________________________________
________________________________________________________________
________________________________________________________________
________________________________________________________________
128
________________________________________________________________
End of Block: Section 3: Homeowner Information
129
Appendix B: Data Cleaning
The survey resulted in 2,093 responses. Some of the responses were incomplete or
indicated that the respondent’s solar power system installation did not meet the sample criteria,
resulting in exclusion of 251 responses. There were four general reasons for exclusion: not
accepting the terms of the survey; installation year not between 2013 and 2017; respondent was
outside of the upstate New York region; and respondent did not complete all decision factor
importance ratings. Eliminating the 251 responses resulted in 1,842 responses used for analysis.
The eliminations are displayed in Table B.1.
Table B.1
Data Exclusions
Reason for Exclusion Count
Did not accept the terms of the survey 17
Installation year unknown or was outside of study scope 107
Utility company was outside of the Upstate New York scope 30
Incomplete decision factor ratings 97
Total Exclusions 251
130
Appendix C: NY-Sun Incentives
The NY-Sun program has provided incentives to offset the cost of residential solar power
systems since 2014 according to Table C.1 (NY-Sun, 2018b). These incentives compare to an
average 2018 residential solar power system cost of $3.41 per watt in New York (EnergySage,
2018). These are New York incentives above the federal incentives that are common to all states.
Table C.1
NY-Sun Incentives
Block From To Size (kW) Incentive ($/W) Status
1 1/1/2014 9/23/2014 40,000 1.00 Closed
2 9/23/2014 11/12/2014 15,000 0.90 Closed
3 11/12/2014 2/18/2015 19,000 0.80 Closed
4 2/18/2015 6/12/2015 22,000 0.70 Closed
5 6/12/2015 9/24/2015 24,000 0.60 Closed
6 9/24/2015 1/28/2016 35,000 0.50 Closed
7 1/28/2016 9/21/2017 85,000 0.40 Closed
8 9/21/2017 TBD 75,000 0.35 41% complete
9 TBD TBD 148,000 0.20 Future