Samant
Shant Priya
Lal
Bahadur Shastri Institute of Management, India
E-mail: samantsp@gmail.com
Sushil
Kumar Dixit
Lal
Bahadur Shastri Institute of Management, India
E-mail: sushil.dixit@yahoo.com
Sajal
Kabiraj
Häme
University of Applied Sciences Ltd. (HAMK), Finland
E-mail: Sajal.kabiraj@hamk.fi
Meenu
Shant Priya
Galgotias
University, India
E-mail: meenushant@gmail.com
Ashirwad
Kumar Singh
EY
(GDS), India
E-mail: ashirwad_kumarsingh@lbsim.ac.in
Submission: 10/3/2020
Accept: 11/9/2020
ABSTRACT
This is an exploratory research highlighting the concerns and reactions of Indian working-class people towards the COVID-19. It was observed that most of the Indian working-class people were seriously concerned about the pandemic and responded well to the measures suggested by the Governments and other agencies in a big way. Most of the respondents believed the pandemic will be effectively controlled across the globe within one year. Word cloud and other data visualization techniques were used to analyze the reactions of the Indian working class towards the Central and State government’s initiatives to contain COVID-19. In the word cloud of the top 150 popular words for both central and state governments Lockdown, People and Government have taken the central stage. The word streaming analysis suggests the intense relationship among the most frequent words in the dataset. For the central government, it was social distancing and for state government, it was social distancing and relationship between central and state governments. The sentiment analysis for both central and state government was neutral, mostly. The researchers are of the view that the research will provide a deeper insight into human perception and behavior towards the measures initiated by the Central and State Governments in any similar difficult situations. Further the concerns identified may be taken into consideration by the Government while designing the policy measures and other interventions by the Government.
Keywords: Sentiment; word cloud; text mining; sentiment analysis; COVID-19
1.
INTRODUCTION
On the last day of the year 2019,
the Chinese authorities have reported to the World Health Organization (China)
about the pneumonia-like case in Wuhan city of Hubei province. Some of the
people suffering from this were reported to be the dealers and vendors at the
Huanan seafood market. On Jan. 3rd, 2020 China reported a total of
44 suspected cases (Muccari et al., 2020).
In January 2020, the outbreak was
identified to be caused by a new coronavirus. The common symptoms include
fever, cough, fatigue, shortness of breath, loss of smell, and taste. The first
death in China was reported on Jan 11, 2020. And since then, it has spread
globally. On January 30, 2020, the World Health Organization (WHO) declared the
outbreak as the global public health emergency as by then more than 9000 cases
were reported.
In February 2020, it reached to
Japan, South Korea, Egypt, France, Iran, and other countries. On March 11,
2020, the WHO declared this as pandemic (WHO,
2020), and as per the coronavirus resource center of Johns Hopkins
University, the reach of this virus spread across 188 countries as on the date
of writing this paper (CSC, 2020).
The impact of COVID 19 is
devastating as more than 4.5 million people are reported to be infected and
more than 0.3 million people have died of COVID 19. The situation in India is
not that great too as more than 90 thousand people are infected, and more than
27 hundred people have lost their lives.
As
on date, there is no medication available for COVID-19. To make COVID-19
information widely available, the Government of India has launched a dedicated
website https://www.mygov.in/covid-19, informing her citizens about its symptoms and preventive
measures. There are many myths also doing round about it, to control this and
provide accurate information regarding COVID 19, a dedicated App “Arogya Setu”
has also been launched by the Government.
Through
various addresses to the nation, the Prime Minister of India declared and
detailed about the lockdown in the country and advisory around the same were
being issued from time to time by various concerned ministries and departments.
Understanding the concerns and sentiment of the Indian people during the time
of lockdown is at the core of this work. While understanding the concerns and
sentiments of Indian working class, this work will pave ways for Government to
formulate plans and strategies to mitigate the risks associated with COVID-19
or any other similar event that may occur in future.
The Cambridge dictionary has defined sentiment
as “A thought, opinion, or idea based on a feeling about
a situation, or a way of thinking about something” (Procter, 2001). The
touch points for the research are development of word cloud, sentiment analysis
and understanding perception about changed health and sanitation behaviour in
the light of COVID-19.
2.
LITERATURE REVIEW
Any
pandemic represents significant risk to humans because of its potential to
cause high levels of mortality and to disrupt socially and economically across
the globe. In case of outbreak of any pandemic it is essential for public
health authorities to be prepared to act. This requires careful planning on the
side of public health authorities.
An
effective plan should focus on prevention, where possible; prepare the society
to meets its health needs; respond quickly to reduce the damage; and contribute
to rapid recovery of individuals and communities (August, 2019).
Community
perception and sentiments about the associated risk and threat are essential
ingredients for an effective pandemic control plan. Most of the information
about public perception and response in case of outbreak of a pandemic is
available for SARS coronavirus and H5N1 subtype of avian influenza virus (Leung et al., 2003; Lau et al., 2003; Cava et al.,
2005).
Kulkarni et al. (2019) found that influenza
viruses have capability of causing pandemic because of their unstable nature
and a novel virus may result through the exchange of genetic material among
viruses from different animal, avian or human hosts.
Research
shows that during SARS people’s willingness to comply with risk reducing
behavior was associated with perceived urgency and seriousness of threat. In
health psychology models risk perceptions have been viewed as one of key
drivers of health behavior modifications (Brewer
et al., 2007; Brewer et al., 2004; Weinstein, 1988; Weinstein et al., 2007).
Precaution
adoption process has been found to be an orderly sequence of qualitatively
different cognitive stages (Weinstein, 1988).
Lau et al. (2003) study in Hong Kong
focused on likely compliance of public towards protective behavior in case of a
potential outbreak of H5N1. It was found that 1/3rd respondents felt
that the chances of outbreak were high, and more than half indicated that they
were worried of getting themselves of one of the family members getting
infected with the same.
De Zwart et al. (2007) compared the risk perception of Asian and European respondents towards
the spread of avian flu. The study found that the European respondents
perceived a higher risk as compared to the Asian respondents. On an average
half of the respondents indicated risk of getting infected. However, there was
wide variations in the risk perception of respondents from different countries.
Di Giuseppe et al. (2008) in knowledge and
attitude of Italians about getting infected and risk associated with the avian
flu. It was observed that 19% respondents reported higher probability of
getting infected. It was further observed that people from the lower
socio-economic strata indicated a higher probability of getting infected and
same were more likely to comply with the hygiene and safety measures suggested.
Ibuka et al. (2010) noted that the perceived
risk and precautionary behavior of people in case of spread of pandemic can be
dynamic and may change over time. It was observed that there was significant
difference between men and women about the perceived risk and precautionary
behavior. Females were found to be more concerned as compared to males.
Chew and Eysenbach (2010) did a content
analysis of 2 million Twitter posts during HINI spread and found that it was an
important source of sharing information about pandemic. Xue and Zeng (2019) studied the role of international agency like
World Health Organization (WHO) along with the initiatives taken by different
national governments in response to the threat of a global influenza pandemic
in the past.
Sentiment
analysis is also known as opinion mining is aimed at understanding people’s
sentiment towards some objects or some happenings around them. To gain an
overview of the wider public opinion around certain topics use of sentiment
analysis is extremely helpful. It has got a strong ability to extract insights
from the data. Sentiment analysis comprises a series of methods and tools aimed
at detecting and pulling out opinions from language (Indurkhya & Damerau, 2010).
The
polarity in the form of positive, neutral, or negative is found out with the
help of sentiment analysis for opinion towards something (Dave et al., 2003). Apart from sentiment
analysis, text analysis and word cloud have also been used in the study.
Analyzing texts from a data set so that machine-readable facts can be drawn
from them is the purpose of text analysis. Further, it helps create structured
data from free-text content (Ontotext, 2020).
Graphical
representation of word having greater significance and frequently appearing
into the text is a word cloud. The more common words are depicted with larger
visuals in the word cloud. It helps in understanding and identifying which word
is frequently appearing in the text is the set of responses (VisionCritical, 2020).
3.
RESEARCH METHODOLOGY
The study was conducted in urban
India during April 2020 when the county was under lockdown and had not
witnessed large scale spread of COVID-19 pandemic. An online survey form was
sent to respondents to gather their perception and behavior relating to
different dimensions of the COVID-19 pandemic. The survey questionnaire was
shared at the beginning among connected networks and the respondents were
further requested to pass on the same to people in their network. At the close,
a total of 394 valid responses were received.
The survey included questions on
demographic characteristics such as gender, age, education, income, profession,
and location. One question each was added to measure the respondent’s concern
and perception about the time that will be needed to control the spread of
COVID-19 pandemic across the globe.
Twelve questions captured the
respondent’s behaviour modifications/ initiatives to handle COVID-19
challenges. Two questions captured the respondent’s satisfaction and confidence
in the Union and State Government's measures to contain COVID-19 spread.
Further two open-ended questions captured suggestions from the respondents to
the Union and State Governments in terms of measures to be taken in short,
medium, and long term.
The research intends to understand
the perception and behavior of the Indian working class towards the COVID-19
pandemic. In the survey, 272 (69 percent) respondents were male and 122 (31
percent) were female. Since the age of working class in India, usually
comprises of people above 20, the respondents were divided into two groups:
first representing the age from 20 to 40 and the second having responses from
people of age 41 and more. 49.5 percent (195) respondents represent the first
group whereas 50.5 percent (199) were from the second group.
Qualification wise maximum
respondents (53.8 percent) have master’s degrees followed by bachelor’s degree
(35.53 percent) and the least around 1 percent were having completed their 10th
standard whereas, those completed their 10+2/diploma was 5.83 percent and
respondents having doctorate qualification represented 3.80 percent in the
sample. When it comes to income 48.22 percent of respondents were earning less
than INR 25,000 and 9.13 percent were reported in the maximum bracket earning
more than one lakh INR.
People earning between 50001-75000
and 25000 to 50000 were also represented significantly taking a pie of 17.76
percent and 16.49 percent. The geographical spread of the sample was divided as
86 percent belonging from urban places and 14 percent were from rural places.
The demographic details of the respondents are presented in Figure 1.
Figure 1: Sample Description
4.
DATA ANALYSIS AND DISCUSSION
Due to widespread publicity almost,
everyone was aware of the spread of the pandemic COVID-19. In all mediums of
mass communication, it had grabbed a large media timeshare and the mindshare of
the masses. The researcher attempted to gauge the concerns of the respondents regarding
the spread of the pandemic. It was observed that more than 80 percent of
respondents were following the information about COVID-19 because they were
seriously concerned about the consequences and spread. 17.2 percent of
respondents indicated that they were having only some concerns and so were not
very serious about the pandemic. Only 2.1 percent of respondents reported that
they were not paying much attention and so were not following information about
COVID-19.
As most of the respondents were concerned
about the spread and consequences of the pandemic, it was expected that they
will also adopt some behavioural modification to deal with the same. The table:
2 presents different voluntary behavioural measures taken by respondents to
deal with COVID-19 spread. It can be observed from the table that most
respondents adopted common safety measures suggested in the media. The adoption
of suggested measures by the respondents is very high. The measures are from
categories like avoiding access to public places and possible contact with a
COVID-19 career; higher levels of personal hygiene like washing hands and using
face masks; actively collecting more information and the inherent risks and
community protection.
Table 1: Descriptive Statistics
Behavioural
Practices New |
N |
Yes % |
No % |
Avoided crowded places such as shopping centres,
public parks, or public transportation. |
394 |
98.98 |
1.02 |
Wash your hands more often or use hand sanitizers,
and avoid direct contact with your mouth, nose, etc. |
394 |
98.98 |
1.02 |
Avoided contact with people returning from infected
or high-risk areas. |
394 |
98.22 |
1.78 |
Cover your mouth and nose with a handkerchief,
tissue or arm when coughing and sneezing. |
394 |
97.97 |
2.03 |
Avoided contact with people who have travelled
abroad in recent months. |
394 |
97.97 |
2.03 |
Avoided contact with people returning from other
cities in India |
394 |
97.72 |
2.28 |
If symptoms are suspected or confirmed, be at home
or at a designated hospital for 14 days. |
394 |
96.45 |
3.55 |
If there are suspected symptoms, you are willing to
take the initiative to go to the hospital or health station in time for the
examination. |
394 |
94.67 |
5.33 |
Bought a certain number of masks to wear while going
to public places. |
394 |
88.07 |
11.93 |
Arranged enough food and drinking water in your home
for half a month to a month. |
394 |
85.79 |
14.21 |
Discussed with a doctor or a friend about the health
topic of the COVID-19 pandemic. |
394 |
79.70 |
20.30 |
Sanitized/ventilated my locality or surroundings. |
394 |
78.43 |
21.57 |
Thus,
for individual behaviours, the modification indicated by the respondents was
around 90 percent. However, it was indicated lowest at 78.43 percent for
concerns about the neighbourhood and society by sanitizing locality or
surroundings. The same is in line with the general behaviour reported for the
Indian community which is more concerned about self rather the community.
Respondents also reported lower at 79.70 percent on discussing with doctors or
a friend about the health topic relating to the COVID-19. The same also
reflects the general human tendency of discounting events and risks that seem
distanced in the future.
Table 2: perception About COVID-19 Control
Globally, how long do you think
the COVID 19 outbreak will take to be effectively controlled? |
||||
Duration |
Frequency |
percent |
Valid percent |
Cumulative percent |
2-3 Month |
111 |
28.1726 |
28.1726 |
28.1726 |
About 6 Months |
165 |
41.8782 |
41.8782 |
70.0508 |
About 12 Months |
73 |
18.5279 |
18.5279 |
88.5787 |
More than a Year |
45 |
11.4213 |
11.4213 |
100 |
Further,
the researchers explored the respondent’s perception regarding the maximum
timeframe by which the pandemic will be effectively controlled globally. The
table presents the respondents' views regarding the time it will take to
contain the COVID-19 pandemic globally. It is indicated that most respondents
with 42.3 percent think that it will take around six months to control the
pandemic around the globe. Only 10.5 percent of respondents are of the view
that it will take more than a year to control the COVID-19 outbreak globally.
Around 90 percent of the respondents believe that the pandemic will be
controlled globally in around 12 months.
Researchers also collected the
respondent’s views and suggestions regarding the consequences of the COVID-19
pandemic. To explore the same three methods were used. First, a text analysis
was conducted to identify the most frequent words in the concerns and
suggestions made by the respondents. Second, a Word Streaming was conducted to
identify not only the most frequently occurring words but also understand their
association with the other frequently discussed words. Third, a Sentiment
analysis was done to understand their sentients regarding the COVID-19
pandemic. The section below presents the same to have a deeper understanding of
the perception of the Indian working class towards COVID-19.
The dataset for the Union and State
governments were analysed separately by breaking “Dataset” into “Text Corpus”
(a collection of texts with each response being treated as one document) and
then the frequency of words were obtained by making a “Term Document Matrix”
which is a matrix where the number of rows indicates the number of distinct
words and number columns represents the number of different documents/responses
(Table” 4). Final rankings of words were obtained by calculating “tf-idf”
values for each term and each document. The resulting “Word cloud” was hence
made based upon the ranking of words-frequency for four scenarios for both the
dataset (Union Government & State Government) on “R-Programming” software
Frequently occurring 50 words, Frequently occurring 100 words, Frequently
occurring 150 words, and Frequently occurring 200 words.
Word clouds are the graphical
representation of the unstructured data in the form of text. It presents
frequently occurring words in a document or set of documents. The larger the
size of a word in a word cloud, the higher the frequency of the word in the
text under analysis. It is easy to use and inexpensive method of visualizing
the data. In the context of the COVID-19 pandemic, the word cloud may represent
people’s pain points or issues of concern. For analysis researchers used Word
Cloud with frequently occurring 150 words.
From the Word Clouds of the Union
government, it can be observed that the respondents were most frequently
discussing words like ‘lockdown’, ‘people’, ‘government’, ‘testing’,
‘distancing’, ‘medical’, ‘increase’, ‘pandemic’, etc. Thus, it may be concluded
that the respondents are concerned about things like ‘lockdown and its duration
(increase)’, ‘government action’, ‘social distancing’, ‘medical testing’,
‘medical testing of people’, social distancing by people’, etc. These issues
are of concern and being discussed by the people about the Union
government.
Table 3: Depiction of 2-D "Term Document Matrix" (values
indicate the number of times that term was present in document)
TERMS
à |
DOCUMENTSà |
|||||
|
Response-1 |
Response- 2 |
Response- 3 |
Response- 4 |
…… |
|
“awesome” |
0 |
2 |
1 |
0 |
…… |
|
“business” |
0 |
0 |
0 |
1 |
|
|
“yes” |
1 |
0 |
2 |
0 |
…… |
|
…….. |
……. |
……. |
……. |
…… |
…… |
Figure 2: Word Cloud from Central Government
Dataset (Top 150 Popular words)
From the Word Clouds of the State
governments, it can be observed that the respondents were most frequently
discussing words like ‘lockdown’, ‘government’, ‘people’, ‘distancing’,
‘testing’, ‘medical’, ‘state’, ‘social, ‘provide’, ‘health’, ‘measures’,
‘proper’, etc. Thus, it may be concluded that the respondents are concerned
about things like ‘lockdown’, ‘measures by the state government’, ‘social
distancing’, ‘medical testing’, ‘provide food’, social distancing by people’,
etc. These issues are of concern and being discussed by the people about the
state governments. In comparison with the issues relating to the Union
government, it can be concluded that the respondents do not differ much on
issues relating to Union and State governments.
Figure 3: Word Cloud of State
Government Dataset (Top 150 Popular words)
For linguistic and grammatical
reasons people use different forms of the word to convey ideas. Additionally,
there are families of derivationally related words with similar meanings. In
many situations, it seems as if it would be useful for a search for one of
these words to return documents that contain another word in the set. The goal
of stemming is to reduce inflectional forms and sometimes derivationally
related forms of a word to a common base form to make it easy to comprehend the
underlying idea.
For the present research, “Porters
Word Stemming” was carried out on the “R-Programming” platform using the
“snowballC” package. The result for both the dataset: Central Government and
State Government is a pictorial display of most frequently occurring words that
are also associated in some way with other words. A thicker line indicates a
more intense relation among the most frequently occurring words in the dataset.
Figure 4: Central Government Dataset
with associated words and connections
Word streaming diagram for the Union
government indicates that the people are highly concerned about ‘social
distancing’ and ‘social distancing in public places’. The same can be inferred
from the thickest lines between these words. Second, an important concern of
people is that of an ‘increasing number of testing’ having a line less thick
than the earlier one. People are also talking about ‘other countries’ while
discussing COVID-19 issues.
This means they are comparing the
Indian situation with the situation in other countries. Further people are
connecting ‘government lockdown with strict action’. This may be interpreted as
the people are concerned about the strictness of the lockdown measures by the
government. Another important connection is between ‘short’, ‘long’, and
‘government’ and ‘lockdown’. This simply means that the people are concerned
about the long-term and short-term lockdown by the government. Which may be
referred to as the people’s concern about the duration of the
government-imposed lockdown.
Figure 5: State Government Dataset
with associations and connections
Word streaming diagram for the State
governments indicates that the people are highly concerned about ‘social
distancing’. The same can be inferred from the thickest lines between these
words. Maybe the people are talking about the social distancing measures
adopted and enforced by the State governments. Second, the thickest line
connects words Central, State, and Governments. This simply means that people
are seeing State and Central governments as a collective whole when it comes to
their COVID-19concers.
People are also talking about
‘health care facilities’ while discussing COVID-19 issues. This means they are
concerned about the state of the health care facilities in their state. Further
people are connecting ‘measures to provide proper food and health care
facilities to people’. This may be interpreted as the people are concerned
about the government measures to provide proper food and health care facilities
to people in the state concerned.
5.
SENTIMENT ANALYSIS
Sentiment analysis or opinion mining
is the interpretation and classification of emotions (positive, negative, and
neutral) within text data using text analysis techniques. Sentiment analysis
allows researchers to identify the respondent’s sentiment toward the object
under study. Sentiment analysis is the area which deals with judgments,
responses as well as feelings, which is generated from texts because sentiments
are the most essential characteristics to judge the human behaviour (Chakraborty et al., 2019). To understand the
respondent’s sentiments towards COVID-19 a sentiment analysis was conducted.
Sentiment analysis was done for the
394 responses received from the respondents. The respondents have voiced their
opinion on what the central government and the state government should do to
tackle COVID 19. “R-Programming” platform was used for carrying out the
sentiment analysis. There is a recognized library in “R-Programming” named
“tidytext” which is extensively used for text mining and sentiment analysis.
The library (tidytext) consists of three inbuilt lexicons (lexicons are a
dictionary of words) named “afinn”, “nrc” and “bing” which are used to define
the sentiment of a word.
“afinn”
lexicon classifies the sentiment of a word by scoring an integral number
between -5 to +5 with -5 representing extreme negative sentiment and +5
representing extreme positive sentiment. At the time of writing this research,
a total of 2477 words were rated with a sentiment score in the “afinn” lexicon.
“bing” lexicon classifies the
sentiment of a word by classifying it either as “positive” or “negative”.
Similarly, “nrc” lexicon decides the sentiment of a word by assigning it with
words like: “trust”, “fear”, “negative”, “sadness”, “anger”, “fear”,
“surprise”, “positive” etc.
In our
research, we have used “afinn” lexicon for carrying out sentiment analysis.
However, before carrying the sentiment analysis, proper data cleaning was done
by removing punctuation marks, white spaces, some stopwords (like: ‘in’, ‘of’,
‘the’ etc.) and some other words would not make any sense otherwise.
Any word which was present in the
dataset and for which there was no sentiment score in “afinn” lexicon was
scored a ‘0’. Also, in this research, a single response was broken into
individual words and the sentiment score for each of the words was added to
obtain a final sentiment score of the whole response. This has been illustrated
below:
a) Suppose, RESPONSE_1: “Increase the
lockdown”.
Overall “RESPONSE_1’s Sentiment
Score” = Sentiment score of “increase” + Sentiment score of “lockdown” (the
word ‘the’ was removed in data cleaning and hence it’s value was not computed).
Table 4: Sample Words with Sentiment Values in “afinn” Lexicon
Word |
Value |
Word |
Value |
Abductions |
-2 |
Admire |
3 |
Ability |
2 |
Affected |
-1 |
Abuse |
-3 |
Afraid |
-2 |
Accept |
1 |
Amazing |
4 |
Acrimonious |
-3 |
Anger |
-3 |
Audacious |
3 |
Awesome |
4 |
Award |
3 |
Bastard |
-5 |
Bitch |
-5 |
Breathtaking |
5 |
Similarly,
an overall sentiment score was obtained for each of the 394 responses and the
results have been summarized for the Central and State governments both.
Figure 6: Sentiments for Central
Government
Table 5: Frequency Distribution Sentiments for Central Government
Sentiment Scores of “Responses” |
Frequency (Total = 394) |
-4 |
2 |
-3 |
2 |
-2 |
12 |
-1 |
55 |
0 |
227 |
1 |
63 |
2 |
19 |
3 |
7 |
4 |
2 |
5 |
0 |
6 |
1 |
7 |
0 |
Figure 7: Positive, Negative, and
Neutral Sentiments for Central Government
From the figures and Table, it can
be concluded that most of the responses scored an overall sentiment score of
‘0’ meaning they hold neutral sentiments (sentiment score = 0). But the count
of positive sentiments (sentiment score > 0) reported at 24 percent exceeded
over negative sentiments (sentiment score < 0) reported at 18 percent by a
thin margin of 6 percent. Hence, it can be concluded that overall people have a
positive reaction towards the measures and initiatives taken by the Central
Government.
So, the measure initiated may be
continued. But the margins between the positive and negative sentiments are not
very large. This demands a need to analyse the issues which are causing a
negative reaction towards the measures and initiatives. In the light of
findings of this research, the measures and initiatives taken by the Central
Government need to be modified to make them more appropriate for people’s
requirements.
Figure 8: Sentiments for the State Governments
Table
6: Frequency Distribution Sentiments for the State Governments
Sentiment Scores of “Responses” |
Frequency (Total = 394) |
-4 |
0 |
-3 |
2 |
-2 |
6 |
-1 |
50 |
0 |
235 |
1 |
71 |
2 |
22 |
3 |
4 |
4 |
2 |
5 |
0 |
6 |
1 |
7 |
0 |
8 |
1 |
Figure 9: Positive, Negative, and
Neutral Sentiments for the State Governments
From the figures and Table relating
to sentiment analysis for the state governments, it can be concluded that most
of the responses scored an overall sentiment score of ‘0’ meaning they hold
neutral sentiments (sentiment score = 0). But the count of positive sentiments
(sentiment score > 0) reported at 26 percent exceeded over negative
sentiments (sentiment score < 0) reported at 14 percent by a comfortable
margin of 12 percent.
Hence, it can be concluded that
overall people have a positive reaction towards the measures and initiatives
taken by the State Government. So, the measure initiated may be continued and
augmented further. But there is still a need to analyse the issues which are
causing a negative reaction towards the measures and initiatives. In the light
of findings of this research, the measures and initiatives taken by the State
Governments need to be modified to make them more appropriate for people’s
requirements.
6.
SUMMARY AND CONCLUSION
A pandemic like COVID-19 is
unprecedented in recent human memory. It presented not only a threat to social
and economic functioning but also human lives. So, it required ordinary people
to modify their behaviour. Governments across the nations initiated unique
measures like social-distancing, lock-down controlling different social and
economic activities, planning for social and economic activities, increasing
emphasis on public health initiatives, etc. Some governments failed to meet
public expectations, but some were able to live up to their expectations.
The present research, in the first
part, explored the concerns of the Indian working class toward the COVID-19
pandemic and associated responses in the form of behaviour modifications. It
was observed that most of the Indian working-class people were seriously
concerned about the pandemic and responded to the measures suggested by the
Governments and other agencies in a big way. However, the response observed for
the individual behaviour modifications were high as compared to group behaviour
modifications. Most of the respondents believed the pandemic will be
effectively controlled across the globe within one year.
To understand the perception and sentiments text analytical methods like word-cloud, word-streaming, and sentiment analysis were used. In the word cloud of the top 150 popular words for both central and state governments lockdown, people and government have taken the central stage. The word streaming analysis suggests the intense relationship among the most frequent words in the dataset.
For
the central government, it was social distancing and for state government, it
was social distancing and relationship between central and state government. It was observed that the majority
was having a ‘neutral’ sentiment towards the initiatives and measures taken by
the Central and State Governments, both. However, the respondents who indicated
a positive sentiment were marginally higher as compared to those who indicated
a negative sentiment for the Central and State Governments out of those who
expressed their sentiments.
7.
IMPLICATION OF THE RESEARCH
It is evident for the findings that
the sentiment of Indian working class is neutral in both cases of Central and
State Governments. The word cloud suggests that the most important concern of
people at the time of COVID-19 is lockdown and social distancing. The
behavioural practices as suggested by government and other agencies in eight
cases were reported to be followed in high numbers (more than 90% cases), so
reported in table 1.
The compliance to the measures
suggested by the WHO and Government of India was much higher as compared to
earlier works in Hong Kong and Italy done at the time of spread of H5N1
pandemic (Lau et al., 2003, Di Giuseppe et al.,
2008, Fielding et al., 2005). Probably the reason for the same may be
widespread availability of information about the seriousness of the COVID-19
resulting in loss of human lives.
A large extent (88.58%) of the
respondents were sure that it will take around a year’s time to control this
pandemic. The findings of the research can be used Government to create and
adopt various measures to not only come out with an effective plan to control
COVID-19 and associated risks but also help her to control pandemic like this
or some new and different health emergency in future.
REFERENCES
Australian (2019). Australian Health Management Plan for Pandemic Influenza. Online
ISBN: 978-1-74186-151-8. Available in:
https://www1.health.gov.au/internet/main/publishing.nsf/Content/519F9392797E2DDCCA257D47001B9948/$File/w-AHMPPI-2019.PDF.
Access in: 17/11/2020
Brewer, N. T., Chapman, G. B.,
Gibbons, F. X., Gerrard, M., Mccaul, K. D., & Weinstein, N. D. (2007).
Meta-analysis of the relationship between risk perception and health behavior:
the example of vaccination. Health
Psychol, 26, 136-45.
Brewer, N. T., Weinstein, N. D.,
Cuite, C. L., & Herrington, J. E. (2004). Risk perceptions and their
relation to risk behavior. Ann Behav Med, 27,
125-30.
Cava, M. A., Fay, K. E., Beanlands,
H. J., Mccay, E. A., & Wignall, R. (2005). Risk perception and compliance
with quarantine during the SARS outbreak. J
Nurs Scholarsh, 37, 343-7.
Chakraborty, K., Bhattacharyya, S.,
Bag, R., & Hassanien, A. A. (2019). 7 - Sentiment Analysis on a Set of
Movie Reviews Using Deep Learning Techniques. In: Dey, N., Borah, S., Babo, R., & Ashour, A. S. (eds.) Social Network Analytics. Academic Press.
Chew, C., & Eysenbach, G.
(2010). Pandemics in the Age of Twitter: Content Analysis of Tweets during the
2009 H1N1 Outbreak. PLOS ONE, 5,
e14118.
CSC. (2020). COVID-19 Dashboard by the Center for Systems Science and Engineering
(CSSE) at Johns Hopkins University [Online].
John Hopkins University. Available: https://coronavirus.jhu.edu/map.html [Accessed May 18 2020].
Dave, K., Lawrence, S., &
Pennock, D. M. (2003). Mining the peanut gallery: opinion extraction and
semantic classification of product reviews. Proceedings of the 12th international conference on World Wide Web. Budapest, Hungary: Association for
Computing Machinery.
De Zwart, O., Veldhuijzen, I. K.,
Elam, G., Aro, A. R., Abraham, T., Bishop, G. D., Richardus, J. H., & Brug,
J. (2007). Avian influenza risk perception, Europe and Asia. Emerg Infect Dis, 13, 290-3.
Di Giuseppe, G., Abbate, R.,
Albano, L., Marinelli, P., & Angelillo, I. F. (2008). A survey of
knowledge, attitudes and practices towards avian influenza in an adult
population of Italy. BMC Infect Dis, 8,
36.
Fielding, R., Lam, W. W., Ho, E.
Y., Lam, T. H., Hedley, A. J., & Leung, G. M. (2005). Avian influenza risk
perception, Hong Kong. Emerg Infect Dis, 11,
677-82.
Ibuka, Y., Chapman, G. B., Meyers,
L. A., Li, M., & Galvani, A. P. (2010). The dynamics of risk perceptions
and precautionary behavior in response to 2009 (H1N1) pandemic influenza. BMC Infectious Diseases, 10,
296.
Indurkhya, N. & Damerau, F. J.
(2010). Handbook of natural language
processing, Boca Raton, FL,
Chapman & Hall/CRC.
Kulkarni, S., Narain, J., Gupta,
S., Dhariwal, A., Singh, S. & Macintyre, C. (2019). Influenza A (H1N1) in
India: Changing epidemiology and its
implications. 32, 107-108.
Lau, J. T., Yang, X., Tsui, H.,
& Kim, J. H. (2003). Monitoring community responses to the SARS epidemic in
Hong Kong: from day 10 to day 62. J
Epidemiol Community Health, 57, 864-70.
Leung, G. M., Lam, T. H., Ho, L.
M., Ho, S. Y., Chan, B. H., Wong, I. O., & Hedley, A. J. (2003). The impact
of community psychological responses on outbreak control for severe acute
respiratory syndrome in Hong Kong. J
Epidemiol Community Health, 57, 857-63.
Muccari, R., Chow, D. A., &
Murphy, J. (2020). Coronavirus timeline:
Tracking the critical moments of COVID-19 [Online]. NBC News. Available: https://www.nbcnews.com/health/health-news/coronavirus-timeline-tracking-critical-moments-covid-19-n1154341 [Accessed May 18 2020].
Ontotext. (2020). What is Text Analysis? [Online]. Ontotext. Available: https://www.ontotext.com/knowledgehub/fundamentals/text-analysis/ [Accessed May 18 2020].
Procter, P. (2001). Cambridge international dictionary of
English. Cambridge (UK): Cambridge University Press.
Visioncritical. (2020). The Pros and Cons of Word Clouds as
Visualizations [Online]. Vision
Critical. Available: https://www.visioncritical.com/blog/pros-and-cons-word-clouds-visualizations [Accessed May 18 2020].
Weinstein, N. D. (1988). The
precaution adoption process. Health
Psychol, 7, 355-86.
Weinstein, N. D., Kwitel, A.,
Mccaul, K. D., Magnan, R. E., Gerrard, M., & Gibbons, F. X. (2007). Risk
perceptions: assessment and relationship to influenza vaccination. Health Psychol, 26, 146-51.
WHO. (2020). WHO Director-General's opening remarks at the media briefing on
COVID-19 - 11 March 2020 [Online].
World Health Organization. Available: https://www.who.int/dg/speeches/detail/who-director-general-s-opening-remarks-at-the-media-briefing-on-covid-19---11-march-2020 [Accessed May 18 2020].
Xue, L., & Zeng, G. (2019).
Global Strategies and Response Measures to the Influenza A (H1N1) Pandemic. A Comprehensive Evaluation on Emergency
Response in China: The Case of Pandemic Influenza (H1N1) 2009. Singapore: Springer Singapore.