SCHOOL OF MATHEMATICAL SCIENCES, USM PENANG

MSG469: REGRESSION ANALYSIS 19/20

ASSIGNMENT 4

Q1A. PASSENGAR CAR MILEAGE

Similar to the development in social media, we are living in a golden age of automotive
technology. Cars are developed to be lighter, more powerful, and perhaps most importantly
more fuel-efficient. Not only accidently become the topics of conversation, studies have also
shown that fuel economy is the most important factor in the decision process of vehicle
purchases. However, even the most fuel-efficient cars on the market can be held back by a
number of different factors. Most drivers know regular maintenance will help keep their
vehicle’s efficiency numbers high, but there are many more surprising things that can have a
significant effect on the fuel economy of your car. These factors are the weather (hot or
cold), car shape and styling, driving style/pattern, in-car gadgets and hidden weights.
This assignment looks into how variables: top speed, TOPS (miles per hour), engine
horsepower, HPWR and vehicle weight, WGHT (in 100 pounds) affects mileage of the
vehicle, MPG (average miles per gallon).

a) Produce scatterplot between the response variable and each of the explanatory variables.
Discuss on the linearity relationship the response and the explanatory variables.
b) Estimate a simple linear regression model relating the impact of TOPS on MPG and the
impact of HPWR on MPG. Examine the appropriateness of linearity assumption on the
two relationships.
c) Investigate the appropriate linearized model relating MPG with TOPS and MPG with
HPWR. Interpret on the estimated coefficients measuring the impact of top speed and
horsepower on the vehicle mileage. Also, discuss on the validity of the error
assumptions.
d) Investigate the appropriate linearized multiple regression model. Provide interpretation
on the estimated coefficients.
e) Discuss on the fulfillment of homoscedasticity assumption for the estimated multiple
regression model above. Suggest the appropriate analysis that should be conducted to
correct any issue.
f) Briefly suggest and discuss a few ways to increase efficiency of fuel usage. Your
discussion should consider few types of vehicle such as cars, buses and lorries.

1
Q1B. WAGE
While salary (gaji) is more popularly known, especially in Malaysia, certain type jobs
requiring physical skills or lower education level are typically paid a wage (upah). Someone
receiving a salary is usually in a managements or professional positions, requiring
higher education level. The essential difference between a salary and wages is that a
salaried person is paid a fixed amount per pay period and a wage earner is paid by the
hour. Typically, wage is paid in weekly basis, for the work completed in the previous
week. Someone who is paid a salary is paid a fixed amount in each pay period, in most
cases it is paid monthly, with the total of these fixed payments over a full year summing
to the amount of the salary. As such, in western countries, salary is defined as the
amount received per year. Another difference is that wage earner may receive overtime
pay while salary earner often receive bonus for the extra work completed.

This assignment investigates how education level and three profiling status: non-white,
female and married affect the amount of hourly wage received.
a) Produce a scatterplot between the response variable, WAGE and the quantitative
explanatory variable of education level, EDUC. Briefly discuss the variation of wage
received across the different level of education.
b) Estimate a multiple linear regression model relating the impact of EDUC, NON-WHITE,
FEMALE AND MARRIED on WAGE. Interpret the estimated coefficients as well as
the quality of the fitted model.
c) Conduct residual analysis to examine for the homoscedasticity assumption for the errors.
The analysis should include graphical plot as well as statistical tests. Note: for White’s
test, use the version that utilizes the fitted values.
d) Conduct analysis to obtain an estimate for the weights to be use in FGLS approach.
Consequently, estimate a multiple regression model by using the FGLS approach.
Compare your results with those obtained in parts (b) and (c) above.
e) Repeat the analysis in parts (b) and (c) above by using log(WAGE) as the explanatory
variable. Provide interpretation of the estimated coefficients from this log-lin model.
Discuss and compare your findings, in particular discuss the impact of conducting log
transformation on the explanatory variable.

2
Q2. GARBAGE INCINERATOR & HOUSE PRICE
In most villages, people continue to bakar sampah. The burn pile is one of the simplest and
earliest forms of waste disposal, essentially consisting of a mound of combustible materials
piled on the open ground and set on fire.
Garbage from homes, schools and businesses around the globe amounted to 2.2 billion tons
in 2016, with majority of the amount discarded by people in North America, Europe and
Central Asia. Projections show that the number will reach 3.7 billion tons in 2050.
In urban and city areas, incinerator is built to take care of garbage and other types of waste.
An incinerator is a furnace for burning waste. Incinerators reduce the solid mass of the
original waste by 80%–85% and the volume by 95%–96%, depending on composition and
degree of recovery of materials such as metals from the ash for recycling. This means that
while incineration does not completely replace landfilling, it significantly reduces the
necessary volume for disposal.
Incineration is a waste treatment process that involves the combustion of organic substances
contained in waste materials. Incineration of waste materials converts the waste into ash, flue
gas and heat. The ash is mostly formed by the inorganic constituents of the waste and may
take the form of solid lumps or particulates carried by the flue gas. Without adequate flue gas
cleaning, these outputs may add a significant amount of pollution to the environment.
Dioxins and furans produced from incineration are considered by many to be serious health
hazards. These days, modern incinerators include equipment to mitigate pollution and the
heat generated by incineration can be used to generate electric power.
For your information, there are only five (small) incinerators in Malaysia, the largest being in
Pulau Langkawi with the others located at Labuan (second largest), Cameron Highland,
Pulau Pangkor and Pulau Tioman (smallest). The first waste-to-energy incinerator in
Malaysia was built in Port Dickson, and have started operation in June 2019.
Despite modern technology, higher safety level and environmental laws there are still
concerns from experts and local communities about the environmental effect of incinerators.
This assignment looks into the impact of nearby incinerator to house price, in addition to
other standard variables. The response variable is price of the house in dollars, PRICE while
the explanatory variable of main interest is the distance of house from the incinerator
measured in feet, DIST. Other variables are INTST that is the distance from house to
interstate (highway connecting states in the US), AREA is square footage of the house,
LAND is the lot size in square feet, ROOMS is total number of rooms in the house and
BATHS is the number of bathrooms.
a) Produce scatter plot between each of the explanatory variables and the response variable.
Discuss on the observation, particularly on the possible issues of heteroscedasticity and
quadratic relationships.
Also discuss on the benefit/advantage of using response variable in the log form,
log(PRICE) to the regression analysis.
b) Estimate an OLS regression model, relating PRICE and BATHS and interpret the results.
1) Conduct residual analysis as well as heteroscedasticity tests to examine the
homoscedasticity assumption for the error term.
2) Estimate a WLS regression model using weights obtain from the hard way, that is
using each number of bath room as a group/range. Compare and discuss similarities
and differences between the estimated models using the OLS and WLS approaches.
c) Estimate an OLS regression model, relating PRICE and ROOMS and interpret the
results.

3
 1) Estimate a quadratic OLS regression model to investigate any diminishing effect of
ROOMS on PRICE. What is the optimal number of rooms in the house with respect
to price.
2) Estimate a quadratic log-level regression model relating log(PRICE) and ROOMS.
Compare and discuss similarities and differences of the results with those in part (1).
Comparison should include discussion on the validity of error assumption. Relate
your findings with the discussion given at the end of part (a) above.
d) Estimate a log-log regression model relating log(PRICE) and log(DIST) and interpret the
results. Conduct residual analysis as well as heteroscedasticity tests to examine the
homoscedasticity assumption for the error term.
e) Expand your model by estimating multiple log-(log+level) regression model consisting
of all the six explanatory variables. Use ROOMS and BATHS in its original form.
Provide interpretation of the estimated coefficients. In addition, discuss the significance
and quality of the estimated model as compared to previous models estimated in parts (c)
and (d) above.
f) Based on your observations in part (a) above, expand further your model by
incorporating quadratic terms of the appropriate variables. Discuss the diminishing effect
of the incorporated variables, emphasizing on the impact of garbage incinerator on house
price in the nearby area.
Note: quadratic term of log variable should be included as  log x  term rather than
2

log  x 2  term.

g) After fighting Barisan Nasional relentlessly over their decision to build the harmful
incinerators in 2011, the Pakatan Harapan has made a U-turn and seems to love the idea
so much and supporting the built of Waste-to-Energy incinerator near Port Dickson.
Explain your opinion on the issue of incineration by highlighting the advantage and
disadvantage of using garbage-and-waste incinerators in Malaysia for waste
management.

ALL THE BEST AND GOOD LUCK