Quality of life in Lucknow: a principal component analysis.
Singh, Sanjay K.
Abstract
The main objective of this paper is to find out the key factors
that determine the quality of life in Lucknow. The study is based on a
survey of objective as well as subjective questions that are seen as
indicators of quality of life. The people of Lucknow were asked to put
their opinion on different measures of quality of life. The collected
sample of over 300 responses on a questionnaire was subjected to
principal component analysis, a statistical technique for dimensionality
reduction of the dataset. The dimensionality reduction technique
revealed seven underlying factors /components as the key determinants of
the quality of life in Lucknow. Out of these seven components, personal,
social and economic wellbeing is found to be the most important one
followed by ambient environment, transport infrastructure and police
services, housing facilities, social infrastructure, utility services,
and social environment. This shows that the state government of Uttar
Pradesh and municipal government of Lucknow needs to focus on improving
the infrastructure, both physical as well as social infrastructure, and
police services in the city to improve the quality of life of people in
Lucknow.
Keywords: Quality of life, Infrastructure, Principal Component
Analysis
JEL Classification: 131, R20, C81
1. INTRODUCTION
Quality of life is a holistic approach that not only emphasizes on
individuals' physical, psychological, and spiritual functioning but
also their connection with their environment and opportunities for
maintaining and enhancing skills (Marya et al., 2012). Theoretical
definition of quality of life can be made from many perspectives. The
definition of this complex concept, which goes beyond a single
discipline, is not uniform. The quality of life can be defined in terms
of psychology, sociology, economics, and also politics. In general,
quality of life can be defined as a product of the cooperation of
social, health, economic, and environmental conditions affecting the
development of people (Payne et al., 2005). In other words, quality of
life is the degree to which the experience of an individual's life
satisfies that individual's wants and needs, both physical and
psychological.
Measurement of quality of life is based on many different models
and approaches. Accurate, reliable, and theoretically satisfactory
measurement of quality of life which is agreed by most experts does not
exist. This is mainly due to the two relatively separated components by
which the quality of life is made of. They are objective and subjective
(Jindrova and Polackova, 2012). The objective quality of life can be
defined as the degree to which specified standards of living are met by
the objectively verifiable conditions, activities, and activity
consequences of an individual's life. On the other hand, subjective
quality of life is a person's sense of well-being, his satisfaction
or dissatisfaction with life, or his happiness or unhappiness (Kerce,
1992).
The main aim of this study is to identify the important factors
which affect the quality of life of people in Lucknow. Lucknow is the
capital city of the state of Uttar Pradesh. It is the second largest
city of northern India after New Delhi and the eleventh largest city of
India. It is also the most populated city of Uttar Pradesh with a
population of 2.82 million of which 1.47 million are male while females
constitute 1.35 million of the population. Lucknow is experiencing rapid
growth in its population; its population increased from 2.25 million in
2001 to 2.82 million in 2011. Changes in population and its structure
will have several implications one conomy, infrastructure, environment,
society, family life and consequently, quality of life of people in the
city. This study tries to identify the factors affecting the quality of
life of people in Lucknow using a survey of objective as well as
subjective questions. The collected sample of over 300 responses on a
questionnaire was subjected to Principal Component Analysis (PCA), a
statistical technique for dimensionality reduction of the dataset.
Statistical Analysis Software (SAS) is used to evaluate the survey data
by the PCA.
The study is organized into the following sections. Section 1
presents the brief introduction about the study. Data and specification
of methods are described in section 2 and 3, respectively. Section 4
presents the main results of the study. Last section concludes the
paper.
2. DATA
The data was collected from 302 respondents of Lucknow city during
the period December, 2013 to February, 2014. All the respondents have
been living in the city at least for last one year. The sample data
constitute 53 per cent male and 47 per cent female. Average age of the
respondents is 34 years, with minimum age being 18 years and maximum age
being 75 years. Most of them have some level of formal education. Most
of the respondents are married; they constitute 62.5per cent of the
sample. Never married people constitute 35.8 per cent while
widowed/divorced comprise 1.7 per cent of the respondents. The sample
has people with varied employment status; 21.8 per cent of them are
employed in the public sector, 32.1 per cent in the private sector (13.6
per cent in organized private sector and 18.5 per cent in unorganized
private sector), and 12.6 per cent are self-employed. Besides employed
people, sample also constitutes people who are not in the work force
(26.5 per cent including 14.2 per cent students), who are unemployed
(5.3 per cent), and pensioners (1.7 per cent). Employment status of
female respondents is not very impressive in comparison to their male
counterparts; only 50.35 per cent females are employed whereas the
corresponding figure for males in the sample is 80.75 per cent. People
from different income groups are part of the dataset; 86 out of 302
respondents had no individual income, 72 had income up to Rs. 100,000,
111 had income between Rs. 100,000 and Rs. 500,000, and rest 33 had
income above Rs. 500,000. Most of the respondents have 4-5 members in
their household. However, 20.5 per cent respondents have less than four
members whereas 24.5 per cent have more than 5 members in their family.
Most of the respondents have been living in the city for more than 10
years. In fact, 40.7 per cent respondents have been living in the city
for more than 20 years.
A self-rated questionnaire was used to collect the data for this
study. Questionnaire was divided into two parts: first part comprised
socio-economic background of the respondents such as age, gender,
marital status, education, employment status, income, family size, etc.
whereas second part dealt with respondents' rating about quality of
life in Lucknow and the factors affecting the same. Respondents were
asked to rate their overall satisfaction and factors that influence
their satisfaction such as infrastructure and government services,
social, economic and environmental issues, and residential and personal
issues. A seven point Likert type scale with "Delighted" equal
7, "Pleased" equal 6, "Mostly Satisfied" equal 5,
"Neither Satisfied nor Dissatisfied (Neutral)" equal 4,
"Mostly Dissatisfied" equal 3, "Unhappy" equal 2,
and "Terrible" equal 1 was used to rate the satisfaction
level.
3. METHOD
To determine the key factors affecting the quality of life in
Lucknow, dimensionality reduction techniques can be used. The main aim
of the dimensionality reduction techniques is to obtain a compact and
accurate representation of the data that reduces or eliminates
statistically redundant components. Factor analysis (FA) and principal
component analysis (PCA) are the two most widely used techniques for
dimensionality reduction. These techniques are usually used when
variables are highly correlated. The factor analysis estimates factors,
which influence responses on observed variables. The factors account for
common variance in a dataset. In the FA, observed variables are a linear
combination of the underlying factors (estimated factor and a unique
factor). Squared multiplecorrelations are used as communality estimated
on the diagonals. Communality is the variance in observed variables
accounted for by a common factor. Large communality is strongly
influenced by an underlying construct. If communalities are large, close
to 1.00, the results of PCA and FA could be similar (for detail on this,
see, Suhr, 2005 and Meloun and Militky, 2006). Researchers use factor
analysis when they believe that certain latent factors exist that exert
causal influence on the observed variables.
In contrast, principal component analysis makes no assumption about
an underlying causal model. Principal component analysis is simply a
variable reduction procedure that typically results in a relatively
small number of components that account for most of the variance in a
set of observed variables. It minimizes the sum of the squared
perpendicular distances to the axis of the principal component. By
reducing a data set from a group of related variables into a smaller set
of components, the PCA achieves parsimony by explaining the maximum
amount of common variance using the smallest number of explanatory
concepts (for detail on this, see, Field, 2005).
This study uses the principal component analysis to find out the
key factors affecting the quality of life in Lucknow. Statistical
Analysis Software (SAS) is used to evaluate the survey data by PCA. In
general, principal component analysis is undertaken in cases where there
is sufficient correlation among the original variables to warrant the
factor/component representation. Also, PCA requires sample size to be
greater than 100 or at least 5 times the number of variables. It is used
for large multivariate datasets where it is often desirable to reduce
their dimensionality. The first component extracted in a principal
component analysis accounts for a maximal amount of total variance in
the observed variables. Each succeeding component will account for
progressively smaller amount of variance in the dataset and are
uncorrelated to all previous components.
In PCA, most commonly used criterion to retain number of
components/factors is the eigenvalue-one criterion, also known as Kaiser
criterion (Kaiser, 1960). The eigenvalues are representations of the
variance variables share. With eigenvalue-one criterion, components with
eigenvalues greater than 1 are retained. However, there is a
considerable chance that too many components are retained (Costello and
Osborne, 2005 and Zwick and Velicer, 1986). So, it is important to
consider other criteria as well for component retention before drawing
any conclusion. With the scree test (Cattell, 1966), the eigenvalues
associated with each component is plotted. A "break" between
the components with relatively large eigenvalues and those with small
eigenvalues is found. Factor loadings above 0.40 are relevant and can be
included in the result (Hair et al., 1998). Moreover, a minimum of at
least three significant loadings are required for factor identification
(Zwick and Velice, 1986). The retained components/factors can be
interpreted on the basis of the variables that they load upon
significantly and may be named accordingly. The retained
factors/components are rotated orthogonally to make it easier to
interpret the retained components.
The rotated PCA methods rotate the PCA eigenvectors, so they point
closer to the local clusters of data points. There are several
analytical choices of rotation that were proposed in the past. One of
them is the varimax method of orthogonal rotation. The varimax rotation
criterion maximizes the sum of the variances of the squared coefficients
within each eigenvector, and the rotated axes remain orthogonal (Lin and
Altman, 2004). This study uses varimax method of orthogonal rotation for
rotated principal component analysis.
4. RESULTS AND DISCUSSION
Studies on quality of life employ both subjective as well as
objective indicators. Most of the research studies emphasize that the
quality of life is very much connected to the perceptions, feelings, and
subjective values of the persons. Satisfaction and happiness indicators
are accepted to be the most important criterion in measuring the
subjective values. The people of Lucknow were asked to put their opinion
on different measures of quality of life. These measures were based on
the following indicators - infrastructure and government services,
social and economic issues, environmental issues, residential issues,
and personal issues.
Infrastructure and government services include availability and
adequacy of power supply ([X.sub.1]), water supply ([X.sub.2]),
sanitation facilities ([X.sub.3]), transport services ([X.sub.4]),
traffic and road infrastructure facilities ([X.sub.5]), infrastructural
facilities for pedestrian and cyclist ([X.sub.6]), parking facilities
([X.sub.7]), internet services ([X.sub.8]), health services ([X.sub.9]),
quality of education services ([X.sub.10]), recreational facilities
(sports, entertainment, etc.) ([X.sub.11]), level of crowding in public
places ([X.sub.12]), and adequacy and effectiveness of police services
(for personal and property safety) ([X.sub.13]).
Social and economic issues include cooperation among neighbours
(i.e., social cohesion) ([X.sub.14]), degree of cultural integration
([X.sub.15]), sense of community (feeling that one is part of a larger
dependable and stable structure) ([X.sub.16]), availability of
employment opportunities ([X.sub.17]), and satisfaction with financial
situation of household ([X.sub.18]). Environmental issues include air
quality ([X.sub.19]), noise level ([X.sub.20]), cleanliness
([X.sub.21]), and green spaces (parks, gardens, etc.) ([X.sub.22]).
Residential issues include availability and affordability of housing
([X.sub.23]), satisfaction with overall physical condition of residence
([X.sub.24]), satisfaction with crime and fire safety in residence
([X.sub.25]), satisfaction with outdoor activity area ([X.sub.26]), and
availability of parking facility in and around residence ([X.sub.27]).
Personal issues include satisfaction with work and work environment
([X.sub.28]), relationship with friends, relatives and family members
([X.sub.29]), own health and health of family members ([X.sub.30]), food
intake ([X.sub.31]), clothing ([X.sub.32]), household items including
vehicles ([X.sub.33]), and achievement in life ([X.sub.34]).
As discussed in the previous section, we have used PC A to identify
the most important factors/components that influence the quality of life
in Lucknow. With eigenvalue-one criterion, components with eigenvalues
greater than 1 are retained. With this criterion, eight components are
retained. However, there is a chance that too many components are
retained. Using a combination of eigenvalue-one criterion, the scree
test, the proportion of variance explained criterion and the
interpretability criteria, a total of seven factors are retained. The
retained factors/components are rotated orthogonally to make it easier
to interpret the retained components. The results of the analysis of the
varimax rotated components are presented in Table 1, which succeeded in
reducing the 34 variables to 7 components. The 7 components together
explained 63 per cent of the total variance. All the seven components
have positive loadings on their respective significant variables.
First component accounts for 12.8 per cent of the total variance.
This component has high positive loadings on financial situation of
household ([X.sub.18]), work and work environment ([X.sub.28]),
relationship with friends, relatives and family members ([X.sub.29]),
own health and health of family members ([X.sub.30]), food intake
([X.sub.31]), clothing ([X.sub.32]), household items including vehicles
([X.sub.33]) and achievement in life ([X.sub.34]). These variables
describe standard of living, personal and professional relationship, and
personal achievement of individual. Thus, this component can be called
as "personal, social and economic wellbeing".
Second component explains 9.6 per cent of the total variance. This
component has high positive loadings on air quality ([X.sub.19]), noise
pollution ([X.sub.20]), and cleanliness in the city ([X.sub.21]). These
variables are related to environmental factors and thus, can be called
as "ambient environment".
Third component extracted from the analysis includes availability
and adequacy of traffic and road infrastructure facilities ([X.sub.5]),
infrastructural facilities for pedestrian and cyclist ([X.sub.6]),
parking facilities ([X.sub.7]), level of crowding in public places
([X.sub.12]), and adequacy and effectiveness of police services (for
personal and property safety) ([X.sub.13]). This component accounts for
9.6 per cent of the total variance. It mainly describes the status of
transport and police services in the city. Thus, this component can be
called as "transport infrastructure and police services".
Fourth component accounts for 8.1 per cent of the total variance.
This component includes availability and affordability of housing
([X.sub.23]), crime and fire safety in residence ([X.sub.25]), outdoor
activity area ([X.sub.26]), and parking facility in and around residence
([X.sub.27]). These variables are related to housing and facilities
available in and around houses in the city. Thus, this component can be
called as "housing facilities".
Fifth component extracted from the analysis includes availability
and quality of internet services ([X.sub.8]), adequacy of health
services ([X.sub.9]), quality of education services ([X.sub.10]), and
availability of recreational facilities (sports, entertainment, etc.)
([X.sub.11]). This component accounts for 7.8 per cent of the total
variance. It mainly describes availability and adequacy of social
infrastructure in the city. Thus, this component can be called as
"social infrastructure services".
Sixth component includes availability of power ([X.sub.1]), water
([X.sub.2]), and sanitation facilities ([X.sub.3]) in the city. These
variables are related to utility services and thus, this component can
be called as "utility services". It explains 7.6 per cent of
the total variance.
Seventh component accounts for 7.5 per cent of total variance. This
component has high positive loadings on cooperation among neighbours
(i.e., social cohesion) ([X.sub.14]), degree of cultural integration
([X.sub.15]), and sense of community (feeling that one is part of a
larger dependable and stable structure) ([X.sub.16]). These variables
describe the social relationship of the individual. Thus, this component
can be called as "social environment".
There is another component which has eigenvalue more than 1, but
this component is not presented in Table 1 because it doesn't
contain more than two variables having loadings greater than 0.40. It is
important to note that the variables which have significant loadings on
more than one component are dropped because they are not pure measures
of any one construct (for detail on scratching out any variable that
loads on more than one component, see, Rourke et al., 2005). Due to
this, availability and adequacy of transport services ([X.sub.4]),
availability of employment opportunities ([X.sub.17]), availability of
green spaces (parks, gardens, etc.) ([X.sub.22]), and overall physical
condition of residence ([X.sub.24]) are excluded from the
interpretation.
5. CONCLUSION
The main aim of this paper is to identify the important factors
which affect the quality of life of people in Lucknow. We used principal
component analysis for the same. The principal component analysis
reduced the 34 variables in 7 principal components. The 7 components
together explained 63 per cent of the total variance in quality of life
in Lucknow. All the seven components have positive loadings on their
respective significant variables. Out of these seven components,
"personal, social and economic wellbeing" is the most
important one; it explains the highest variance (12.8 per cent). Other
important components are "ambient environment" (9.6 per cent)
followed by "transport infrastructure and police services"
(9.6 per cent), "housing facilities" (8.1 per cent),
"social infrastructure services" (7.8 per cent), "utility
services" (7.6 per cent), and "social environment" (7.5
per cent). Although, in general, issues related to individuals are
beyond the control of the government, government can improve the quality
of life of people by improving the transport infrastructure, social
infrastructure, utility services, ambient environment, housing
facilities, and police services. Therefore, state government of Uttar
Pradesh and municipal government of Lucknow needs to focus on improving
the infrastructure, both physical as well as social infrastructure, and
police services in the city to improve the quality of life of people in
Lucknow.
Acknowledgement
This paper is part of a seed money project sponsored by the Indian
Institute of Management, Lucknow, India. I am thankful to the Director
and Dean (Academic Affairs) of the institute for providing us with an
initiation grant for the study.
References
Cattell, R. B. (1966), "The scree test for the number of
factors", Multivariate Behavioral Research, 1(2), pp. 245-276.
Costello, A.B. and Osborne, J.W. (2005), "Best practices in
exploratory factor analysis: four recommendations for getting the most
from your analysis", Practical Assessment, Research and Evaluation,
10(7), pp. 1-9.
Field, A. (2005), "Discovering statistics using SPSS",
(2nd ed.), London: SAGE Publication.
Hair et al. (1998), "Multivariate data analysis with
readings", (5th ed.). Englewood Cliffs, NJ: Prentice-Hall.
Jindrova, A. and Polackova, J. (2012), "Dimensionality
reduction of quality of life indicators", ACTA Univesitatis
Agriculturae Et Silviculture Mendeleanae Brunensis, 60(7), pp. 147-154.
Kaiser, H. F. (1960), "The application of electronic computers
to factor analysis", Educational and Psychological Measurement,
20(1), pp. 141-151.
Kerce, E. W. (1992), "Quality of life: Meaning, measurement,
and models" (NPRDC- TN-92-15). San Diego: Navy Personnel Research
and Development Center.
Lin, Z. and Altman, R. B. (2004), "Finding haplotype tagging
SNPs by use of principal components analysis", American Journal of
Human Genetics, 75(5), pp. 850-861.
Marya et al. (2012), "Oral health related quality of life and
its impact on elderly people", Indian Journal of Dental Sciences,
4(4), pp. 127-132.
Meloun, M. and Militky, J. (2006), "Compendium of statistical
data processing", Prague: Academia.
O'Rourke, N., Hatcher, L. and Stepanski, E. J. (2005), "A
step by step approach to using SAS[R] for univariate and multivariate
statistics", (2nd ed.), Cary, NC: SAS Institute Inc.
Payne et al., (2005), "Quality of life and health",
Prague: Triton.
Suhr, D. (2005), "Principal component analysis vs. exploratory
factor analysis", Proceedings of the Thirtieth Annual SAS[R] Users
Group International Conference, Cary, NC: SAS Institute Inc.
Zwick, W. R., and Velicer, W. F. (1986), "Comparison of five
rules for determining the number of components to retain",
Psychological bulletin, 99(3), pp. 432-442.
SANJAY K. SINGH, Professor, Indian Institute of Management Lucknow,
Prabandh Nagar, Off Sitapur Road, Lucknow 226013, India, E-mail:
sanjay@iiml.ac.in
Table 1
Total Variance Explained by Different Components
Component Initial Eigenvalues
Total % of Cumulative
Variance %
1 9.685 28.5 28.5
2 3.453 10.2 38.7
3 2.306 6.8 45.5
4 1.972 5.8 51.3
5 1.514 4.5 55.8
6 1.383 4.1 59.9
7 1.068 3.1 63.0
Component Rotation Sums of Squared
Loadings
Total % of Cumulative
Variance %
1 4.366 12.8 12.8
2 3.255 9.6 22.4
3 3.248 9.6 32.0
4 2.738 8.1 40.1
5 2.643 7.8 47.9
6 2.590 7.6 55.5
7 2.539 7.5 63.0
