Statistical Analysis of Health Hazards in Brick and Tile Industries

This report presents a statistical analysis of health hazards in the brick and tile industries, commissioned by the Health and Safety Executive (HSE). The study investigates the potential health risks faced by workers in these sectors, focusing on the hypothesis that exposure to RCS-containing dust leads to silicosis and other health issues, with risks correlating to work span. Using a sample of 65 workers, including 38 from the brick industry and 27 from the tile industry, the research addresses two key questions: whether there is a significant difference in health hazards between the two industries and if there is a correlation between health damage and service length. The report employs descriptive statistics, normality tests (Anderson Darling), two-sample t-tests, and Spearman's correlation to analyze the data, interpreting p-values, confidence intervals, and correlation coefficients. The findings reveal significant differences in health hazards between the brick and tile industries, and a moderate positive correlation between service length and cell damage, highlighting the importance of workplace safety measures.

STATISTICS
ASSIGMNMENT

Table of Contents
INTRODUCTION.........................................................................................................................................1
Population and sample size........................................................................................................................1
Research questions.....................................................................................................................................1
Hypothesis of the study..............................................................................................................................2
Descriptive statistics..................................................................................................................................2
Justifying the chosen statistical tests.........................................................................................................6
Interpreting the p-value, confidence interval and correlation coefficient..................................................8
Critically analyzing the research questions..............................................................................................10
CONCLUSION AND RECOMMENDATIONS........................................................................................11
REFERENCES............................................................................................................................................12

INTRODUCTION
The Health and Safety Executive (HSE) commissions possess interest in identifying the
health hazard of people working in different sectors in order to identify the harmful impacts of
the workplace activities. In relation to this, this assignment aims at analysing the health hazard of
the people working in the brick and tile industries. The hypothesis of this study considers that
there exist significant potential of health hazard of these workers. In addition, it considers that
the risk positively correlates with their work span. Precisely, it assumes that exposure to the
RCS-containing dust results in silicosis leading to incapacity and premature deaths. With regard
to this, it uses data collected from the workers of these industries. Further, it depends on certain
statistical tests and conducts a comparative analysis on these to industries for justifying these
assumptions.In order to analyze and examine the data set, a number of statistical tests have been
used such as descriptive statistics, correlation and others.
Population and sample size
The population consists of the people working indifferent areas of the tile and brick
industry. The study relies on a purposely selected random sampling for deciding on the sample
for the study. The sampling technique appears effective as it allows the researchers to consider
certain conditions in case of selecting the sample. In this context, the researchers ensures that the
workers have not worked in both tile and brick industry, they work in production area (as it
ensures exposure to the dust particles consisting RCS), and they are not smokers. This is useful
in identifying the real impact of the RCS-containing dust particles on the health of these workers.
Further, using random sampling ensures that the selection process is free from any bias. Based
on these criteria, the researchers used random sampling method for selecting a sample of 65
workers. Among these 65 people, 38 workers are from brick industry and 27 from the tile
industry.
Research questions
The research questions are crucial in identifying the primary cause for conducting a research.
With regard to this, this assignment sets the research questions as:
RQ: 1. Is there any significant difference in the potential health hazard of the people working in
the brick and tile industry?
1

RQ: 2 Is there any association between the potential health damage and the length of the service?
Hypothesis of the study
Hypothesis for testing research question 1
H0(Null Hypothesis):There is no significant difference between potential health hazards of the
workers employed in Bricks and tile industry.
H1 (Alternative hypothesis):There is significant difference between potential health hazards of
the workers employed in Bricks and tile industry.
Hypothesis for testing research question 2
H0 (Null Hypothesis): There is no significant relationship between length of service and health
effects?
H1 (Alternative hypothesis): There is significant relationship between length of service and health
effects? Descriptive statistics
The descriptive statistics appears useful in gathering the fundamental information on a
collected dataset. The chosen research topic requires information on the workers of the brick and
tile industry. It provides useful information about the data set and helps to summarize the whole
data set by providing the statistical results of central tendency i.e. mean, mode and median and
dispersion measurement i.e. standard deviation, variance and others.
Survey respondents
The above results provide the information on the data collected and indicates that out of
65 workers, 38 workers were employed in brick industry and 27 were engaged in tile sector.
Further, it is essential to identify the proportion of brick and tile industry workers to the selected
sample. From the sample selection it is evident that the proportion of the brick industry workers
is slightly higher to 58.46% than that of the workers in the tile industry as it is derived to
41.54%.
2

Next step is identifying the average values of the variables. In case of identifying the
average age of the workers and their average service length the mean calculation appears useful.
However, for identifying the average cell damage percentage, it is wise to rely on the calculation
of median. That is, it does not consider that the data follow a particular distribution. Therefore,
the mean calculation appears useful. In words of Shiet al .(2015, p.157), median appears useful
in finding a more appropriate representative value.
Descriptive Statistics: length of service (years), age (years), %
damaged cells
Interpretation: According to the results founded, on an average, every worker has
served the industry since 9 years. However, average age of the employee who took participation
in the survey has been determined to 39.52 years and average % of damaged cells in the
employee is founded to 1.541. Looking to the dispersion statistical results, standard deviation is
founded highest for the age category to 13.35 that indicates greater variability in the age of
workers surveyed. It indicates that each worker’s age varies from the mean age of 39 years.
However, on the other hand, standard deviation for the length of service period and % of
damaged cells is reported comparatively less to 7.424 and 0.841 that indicates less fluctuations &
volatility in the series.
3

70
60
50
40
30
20
10
age (years)
Boxplot of age (years)
Figure 1 Box plot of age in years
35
30
25
20
15
10
5
0
length of service (years)
Boxplot of length of service (years)
Figure 2 Box plot of length of service in years
4

4
3
2
1
0
% damaged cells
Boxplot of % damaged cells
Figure 3 Box plot of percentage of damage cells
Descriptive: Bricks & Tile industry
5

From the table, it is evident that the average age of the workers and the average service
length are higher in case of the workers of the tile industry as it has been derived to 41.70 & 9.83
years respectively. However, on the other side, in Brick sector, the values are computed to 8.77
and 37.97 years respectively slightly less than that of tile industry. However, the representative
value of average cell damage percentage is lower in case of the tile industry workers to 1.278
whilst in bricks, it is 1.728. This implies that the workers in tile sector possess lower chance of
cell damage than that of the brick industry worker.
Justifying the chosen statistical tests
The process of selecting the statistical test that would facilitate the hypothesis testing
requires identifying the distribution of the data. In this context, the normality test is the first test
to consider. The Anderson Darling normality test appears effective in commenting on the
normality test.It helps to find out that how well data follows a particular distribution. The
formula for the Anderson Darling normality test is given by:
It is selected because the Anderson-Darling test of normality is commonly used and most
reliable method for testing that whether the given data set follows a normal distribution curve or
not. However, using software eases the process of identifying whether the data follows the
normality condition. The significance level (alpha) in this case is 0.05. If the p-value is lower
than or equal to the value of alpha, the null hypothesis gets rejected. In contrary, if the p-value is
higher than the alpha value, the null hypothesis is accepted. In this context, it is essential to
identify the null and the alternative hypothesis.
H0: The data of % of damaged cells follows a normal distribution.
H1: The data of % of damaged cells do not follow a normal distribution
The findings from running software generated Anderson Darling normality test is presented in
the table below:
6
AD=−N− 2 i−1
N ( ln ( F ( Y i ) ) +ln ( 1−F ( Y N +1−i ) ) )
AD∗¿ AD ( 1 + 0 .75
N + 2. 25
N2 )

Statistical output
43210
99
95
90
80
70
60
50
40
30
20
10
5
1
Mean 1.728
StDev 0.8624
N 38
AD 0.603
P-Value 0.109
% damaged cells
Percent
Probability Plot of % damaged cells
Normal
Figure 5: Normality plot of % of damaged cells for Brick industry
7

3.02.52.01.51.00.50.0-0.5
99
95
90
80
70
60
50
40
30
20
10
5
1
Mean 1.278
StDev 0.7475
N 27
AD 0.701
P-Value 0.059
% damaged cells
Percent
Probability Plot of % damaged cells
Normal
Figure 4 Normality plot of % of damaged cells for Brick industry
As per the output derived, in both the industries, Bricks and Tiles, P-value is above the
alpha value of 0.05. As in Brick, P-value is founded to 0.109>0.05 that accept ull hypothesus and
rejects alternative hypothesis. Thus, it reflects that the given data set regarding % of damaged
cells follows a normal distribution. Similarly, in Tile sector, AD test’s P value is derived to
0.059>0.05 that supports null hypothesis and clearly represents that % of damaged cells of the
workers in such industry is normally distributed (Gelmanet al. 2014, p.268).
Interpreting the p-value, confidence interval and correlation coefficient
As the data follows a normal distribution, it is effective to use two sample t-test to
examine that whether there is a significant difference or not between potential health hazards of
the employees in tile and bricks industry.
8

In accordance with the summary statistics, the mean % of damaged cells was higher in
bricks industry to that of employees engaged in tile sector. However, as per the results of two-
sample t-test, the p-value to 0.029 is lower than the alpha value of 0.05 indicating rejection of
null hypothesis and accept alternative hypothesis. It implies there is significant statistical
difference between potential health hazards of the workers in bricks and tile sector. The
confidence interval was taken 95% and derived to 0.048 to 0.851. The purpose of setting a
confidence interval is to ensure that there exists probability that the confidence interval will hold
true parameter value. The justification of selecting the 95% confidence interval is it allows the
researchers to ensure that the interval cover 95% area of the curve. At the other extreme, the 95%
confidence interval is commonly chosen for its capability to produce justifies conclusion
(Siegmund, 2013, p.39).
In this context, the researchers ensures that the workers have not worked in both tile and
brick industry, they work in production area (as it ensures exposure to the dust particles
consisting RCS), and they are not smokers. This is useful in identifying the real impact of the
RCS-containing dust particles on the health of these workers. The hypothesis of this study
considers that there exist significant potential of health hazard of these workers. In addition, it
considers that the risk positively correlates with their work span. Precisely, it assumes that
exposure to the RCS-containing dust results in silicosis leading to incapacity and premature
deaths. With regard to this, it uses data collected from the workers of these industries.
9

In order to identify the interrelationship between the service length and the health hazard;
it is effective to run a correlation test on these to variables. This would allow the researchers to
identify the degree of association between the two variables. In order to test the correlation,
Spearman’s correlation has been used. The purpose of this test was to identify if the service span
of the workers has any relation with the potential health hazards. A low value indicates that the
service span of the workers do not have any significant association on the health hazards to the
workers. In this context, it is essential to identify the way of interpreting the correlation
coefficient for commenting on the degree of association between two variables. The value of
correlation coefficient varies from -1 to +1. If the correlation coefficient is negative it indicates
negative association between variables. Further, if the value is close to zero, it indicates a poor
association whereas; a value closer to 1 or -1 indicates strong correlation among variables
(Shevlyakovand Smirnov, 2016, p.149). If the correlation coefficient is 0, it indicates that there is
no correlation among the variables. On the other extreme if the correlation value is positive; it
indicates positive relation between the variables.
From the data collected, it is evident that the correlation coefficient is determined to
0.593 indicates moderate level of association between both the series as it lies under the limit of
0.25 to 0.75. In this particular case, the correlation coefficient is derived positive to 0.593 that
represents that with the increasing period of service length of the employee in the sector, the %
of damaged cells goes high. The P-value is 0.00>0.05 that fully supports alternative hypothesis
and presents that there is correlation exists between service length duration and % of damaged
cells.
Critically analyzing the research questions
The first research question appears useful in identifying the differences in the potential
risks of the workers working in different sectors; namely the brick and the tile industry.
Similarly, the second question ensures an overall analysis of the data collected. This allows the
researchers to analyse the interrelation between the service length and the potential health
hazard. However, the third question requires a critical analysis. The procedure of collecting data
10