See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/6379317

Relation of waste generation and composition to socio-economic factors: A

case study

Article in Environmental Monitoring and Assessment · November 2007

DOI: 10.1007/s10661-007-9705-3 · Source: PubMed

CITATIONS READS

295 10,394

4 authors, including:

Nilanthi Bandara Joseph P A Hettiaratchi

University of Sri Jayewardenepura The University of Calgary
37 PUBLICATIONS 646 CITATIONS 140 PUBLICATIONS 2,593 CITATIONS

SEE PROFILE SEE PROFILE

S.C. Wirasinghe
The University of Calgary
242 PUBLICATIONS 3,670 CITATIONS

SEE PROFILE

All content following this page was uploaded by S.C. Wirasinghe on 21 May 2014.

The user has requested enhancement of the downloaded file.

Environ Monit Assess (2007) 135:31–39
DOI 10.1007/s10661-007-9705-3

Relation of waste generation and composition

to socio-economic factors: a case study
Nilanthi J. G. J. Bandara &
J. Patrick A. Hettiaratchi & S. C. Wirasinghe &
Sumith Pilapiiya

Received: 28 June 2006 / Accepted: 15 January 2007 / Published online: 21 April 2007
# Springer Science + Business Media B.V. 2007

Abstract To develop an effective waste management data to various socio-economic factors. Over 400
strategy for a given region, it is important to know the sample households were selected for the study using
amount of waste generated and the composition of the a stratified random sampling methodology based on
waste stream. Past research has shown that the amount municipal wards and property values. A technique that
of waste generated is proportional to the population considers both the number of households in a particular
and the average mean living standards or the average income group (property value range) and the standard
income of the people. In addition, other factors may deviation of property values within a given income
affect the amount and composition of waste. These are group was used to determine the appropriate sample
climate, living habits, level of education, religious and size for each municipal ward. Through category and
cultural beliefs, and social and public attitudes. This regression analyses, the quantities of waste and waste
paper presents the findings of a study carried out in a composition were related to several socio-economic
suburban municipal area in Sri Lanka to determine the factors. The paper describes the basis for the sample
solid waste generation rate and waste composition selection, the methodology adopted for data col-
based on field surveys and to determine the related lection, the socio-economic parameters used for the
socio-economic factors. A database was developed that analysis, and the relationships developed from the
included information on the quantity and composition analysis.
of waste generated in a sample of households in the
study area over a time period. The collected data was Keywords Solid waste management .
analysed to relate waste generation and composition Waste generation . Socio-economic factors

Introduction

N. J. G. J. Bandara (*)
To plan a municipal solid waste (MSW) management
University of Sri Jayewardenepura, Nugegoda, Sri Lanka
e-mail: nbandara@sltnet.lk strategy for a given region, it is essential to know the
quantity of waste generated and its composition.
J. P. A. Hettiaratchi : S. C. Wirasinghe Various authors have shown that the amount of waste
University of Calgary, Calgary, Canada
generated by a country is proportional to its popula-
S. Pilapiiya tion and the mean living standards of the people
The World Bank Office, Colombo, Sri Lanka (Wertz 1976; Grossmann et al. 1974). Medina (1997)
32 Environ Monit Assess (2007) 135:31–39

related waste generation rates to income levels of was determined by weighing randomly selected trucks
people. However, it has been shown that these are with and without waste.
not the only governing factors. Amongst other socio- King and Murphy (1996) describe the application
economic factors that have been said to influence of the survey sampling theory to the estimation of the
MSW generation are persons per dwelling, cultural annual amount of solid waste generated by an average
patterns, education, and personal attitudes (Al- residential unit in a municipality. They conducted a
Momani 1994; Grossmann et al. 1974). In recognition field survey covering a 12 month period to account
of the importance of a reliable tool to predict the for seasonal variations and incorporated all of the
MSW characteristics, various researchers have districts in the target area. Their approach included
attempted to construct models to predict these the following; selection of a probability sample of
parameters. They found that relationships obtained truckloads, measure the amount of waste collected
between various parameters vary by country. This has and the number of residences covered by each
been attributed to variations in consumer behaviour truckload in the sample, divide the total waste by
and lifestyles. the total residences in the sample to determine a ratio
Although models are available to predict waste gen- of weight per residence. This ratio was used to obtain
eration patterns in developed countries (Daskapoulous an annual total weight per residence. Taking into
et al. 1998), very little research has been done so far to account the effect of socio-economic characteristics,
develop models applicable in developing countries. time of the week and seasonal factors on the amount
Models are available to predict the gross waste of waste left by the curb side by a residence, truck
generation capacity of countries. But, these are not loads were stratified by district, position in the week
adequate for developing integrated waste manage- (first or last collection) and quarter of the year. During
ment plans for municipalities or regions, since waste the week, two samples, at the beginning and at the
generation patterns are unique to regions. There is a end of the week, were taken considering the increase
necessity for waste generation prediction models for in waste loads after weekends. The total number of
suburban municipalities in developing countries, such strata they sampled was 64. The collections per year
as Sri Lanka. Such municipalities are fast growing at each residential unit were 104.
and lack basic infrastructure for waste management. Gay et al. (1993) proposed an alternative to the
To develop an integrated waste management plan, traditional waste characterization studies. Their meth-
a variety of waste specific data is required. To od of estimating waste composition and generation is
determine the resources required for waste manage- based on converting economic sales data for a region
ment and for sizing of waste management facilities, into estimates of solid waste generation. The meth-
an accurate estimate of current and future total waste odology, termed Economic Input/Output Analysis
generation is required. The per capita waste genera- (EIO), is based on the principle that the sales of one
tion rate is needed to predict future waste generation sector are the inputs to another. Matrix analysis is
rates and for evaluating the waste generation trends in used to derive a technical coefficient matrix that
given communities. Once developed, such informa- reveals the impact of an incremental change in sales
tion could be extrapolated to other communities with of any sector on the demands for added purchases
similar socio-economic conditions. from all other sectors.
Waste generation rates and waste characteristics Although different methods of waste characteriza-
had been estimated using several methods. Purdy and tion are discussed in literature, the most common is
Sabugal (1999) conducted a field study to determine the classical method of direct waste analysis. A
the waste composition and waste generation for the comparison of this with two alternative methods,
city of Davao, Mindanao, Philippines. A sample waste product analysis (analysis of products from
waste load was collected from randomly selected waste processing) and market product analysis (ma-
collection trucks, and the sample waste load was terial balance of market products) was conducted by
roughly divided into four quarters and manually Brunnr and Ernst (1986). They concluded that direct
categorized into various waste types and each put waste analysis is an appropriate method to determine
into plastic sacks and weighed to estimate the MSW composition and determine the effect of spatial,
composition of waste. The total amount of waste temporal and socio-economical variations on MSW
Environ Monit Assess (2007) 135:31–39 33

Table 1 Sample distribution in wards and annual property assessment tax value ranges

Wards→Category 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Number of samples

<1,000 3 4 8 6 8 2 3 4 8 2 1 1 4 1 5 60
1,000–3,000 4 9 9 6 6 3 3 4 6 3 1 1 3 2 4 64
3,000–6,000 7 11 7 6 5 5 3 3 6 3 2 2 2 2 4 68
>6,000 11 15 21 12 11 9 8 5 14 5 2 4 4 2 7 130
25 39 45 30 30 19 17 16 34 13 6 8 13 7 20 0 322

composition. They also claimed that the method facturing and construction do not significantly impact
however has several limitations such as high labour waste generation rates.
cost and errors in sample collection and sample Although comprehensive studies that include direct
preparation. waste analysis and consideration of socio-economic
Although time consuming and labour intensive, parameters have been conducted in developed coun-
direct waste analysis will provide a significant amount tries, such studies are sparse in developing countries.
of information if combined with factors affecting Undoubtedly the data and the models generated in de-
waste generation. In the past, some researchers have veloped countries cannot be implemented in developing
attempted to correlate socio-economic factors to solid country situations without site specific data gathering
waste generation. Hockett et al (1995) conducted a and analysis. The paper presents a study conducted in a
study to identify and measure the variables which representative suburban municipal area in Sri Lanka to
influence per capita MSW generation in the south- determine the waste generation rates (per capita, as well
eastern USA using information from counties of as per household, waste generation) and waste compo-
North Carolina as a data set. They developed a sition. The generated data was used to identify the
predictive model of the demographic, economic and parameters that affect characteristics of household solid
structural determinants of per capita waste generation. waste. The results are applicable to suburban munici-
The amount of waste generated per capita per day was palities in other regions of Sri Lanka and in other
the dependent variable. The economic variables developing countries. The paper provides details of data
chosen for the study included per capita retail sales, collection, data analyses, and interpretation of results.
per capita value added by manufacturing and per The authors intend to use the data obtained to develop a
capita construction costs. The landfill tipping fee was predictive model for waste generation patterns in
chosen as the structural variable. The per capita suburban municipalities in Sri Lanka.
income and the urban population percentage were
taken as the demographic variables. Regression
analysis showed that only two variables, the per Study methodology
capita retail sales (P=0.0001) and tipping fees (P=
0.0104), are significant determinants of waste gener- Moratuwa, a suburban municipality of Sri Lanka with
ation. They found that income, urbanization, manu- an area of 21.6 km2 and a population of 189,150, was

Table 2 Descriptive statis-

tics of waste generation from Waste No. of samples Mean (kg) Std. deviation Skewness Kurtosis
selected households in the component
municipality of Moratuwa
Statistic

Organics (biodegradable) 327 0.3744 0.2714 1.452 2.506

Paper 327 1.891E-02 3.411E-02 7.607 69.662
Plastic 327 1.407E-02 3.521E-02 12.711 196.578
Glass 327 6.689E-03 2.011E-02 4.221 20.114
Metal 327 2.967E-03 9.198E-03 6.063 45.757
34 Environ Monit Assess (2007) 135:31–39

Table 3 Average composition of household waste generation – Table 5 Regression analysis of per capita organic waste
municipality of Moratuwa generation for the lower-middle income group (LM)

Waste type Waste composition (as % of total) Variable Unstandardized Std. Standardized t Sig.
coefficients Error coefficients
Organic 90 B Beta
Paper 5
Plastic 3 Constant 0.559 0.090 6.179 0.000
Glass 2 Persons −4.221E-02 0.017 −0.277 −2.476 0.016
Metal 1 per
dwelling

Dependent variable: per capita organic waste

selected for the study. This municipality is located
Selecting only cases for which income level = 2_LM
18.5 km from the Colombo city centre. It consists of
16 wards and has families from diverse economic and
social backgrounds.
Although family income would be the most suitable dwelling was used to estimate the amount of waste
parameter to indicate the living standard which has generated per person or UWF.
been recorded to be proportional to the per capita unit At each household, the amount and the composi-
waste factor (UWF), it was difficult to obtain reliable tion of waste generated were measured over 1 week.
income data. Interviews conducted with the officials For this purpose, the households were given different
of various municipalities in Sri Lanka showed that the coloured garbage bags for the various waste compo-
most reliable data they have that can be related to nents in the waste stream: organics, paper, plastic,
living standards is an estimate of the annual property glass and metal. The organic fraction of the waste was
assessment tax value of a household, which is es- weighed every other day whereas the other materials
timated by the valuation department of the municipal- were weighed only at the end of the week. Hence the
ity for the purpose of levying property taxes. In Sri organic fraction was measured 3–4 times a week in a
Lanka, this tax is based on the size of the property, the given household. The entire exercise covering the
use of the property and material used in its construc- selected sample took 4 months to complete.
tion. The annual property assessment tax value, and To establish the socio-economic factors of the
therefore the living standard, was taken as an households, occupants of each of the households were
independent variable. The living standard was deter- surveyed using a questionnaire. The results of the
mined by vehicle ownership, which indirectly repre- waste composition and waste generation study and the
sents income or affordability, and by the number of data from questionnaires for 12 wards were used to
rooms in the house and the area of the house, which develop the empirical relationships of waste genera-
indicates the value of the property. Religious affilia- tion. The results of the remaining wards will be used
tion, education level and ethnicity were also taken as later for verification of the validity of the developed
independent variables. The number of persons per relationships.

Table 4 Regression analysis of per capita organic waste Table 6 Regression analysis of per capita organic waste
generation for the low income group (LO) generation for the upper-middle income group (UM)

Variable Unstandardized Std. Standardized t Sig. Variable Unstandardized Std. Standardized t Sig.
coefficients Error coefficients coefficients error coefficients
B Beta B Beta

Constant 0.597 0.082 7.236 0.000 Constant 0.628 0.088 7.178 0.000
Persons −4.672E-02 0.015 −0.377 −3.126 0.003 Persons −5.014E-02 0.016 −0.347 −3.047 0.003
per per
dwelling dwelling

Dependent Variable: Per capita organic waste Dependent variable: per capita organic waste
Selecting only cases for which income level = 1_LO Selecting only cases for which income level = 3_UM
Environ Monit Assess (2007) 135:31–39 35

Table 7 Regression analysis of

per capita organic waste genera- Variable Unstandardized Std. error Standardized t Sig.
tion for the upper income group coefficients coefficients
(UP) B Beta

Dependent variable: per capita Constant 0.662 0.063 10.544 0.000

organic waste Persons per dwelling −5.744E-02 0.013 −0.387 −4.518 0.000
Selecting only cases for which Number employed 6.952E-04 0.000 0.183 2.135 0.035
income level = 4_UPI

A regression analysis was performed to delineate a the total numbers of households falling within the
relationship for the different income groups separately individual property assessment tax value ranges were
by taking the amounts of per capita organic waste, estimated from municipal data and the number of
generated per day in kilogram as the dependent households that should be surveyed from these
variable. The independent variables are; property different ranges was determined from the total sample
assessment tax value, total number of persons per number allocated per individual ward.
dwelling, number of families in a household, total This analysis showed that the number of house-
number of employed members in a household, and holds is larger for the lower property value ranges, and
the number of motor vehicles owned by a household. the lowest number of households is in the property
In analyzing the total waste generation data, the assessment tax value range of Rs.6,000 and above.
same independent variables were considered, with the However, the standard deviation is highest in the
household generation of organic waste per day in annual property assessment tax value range of
kilogram as the dependent variable. Rs.6,000 and above. It should be noted that the per
capita waste generation in households with relatively
low property assessment tax value is lower than that of
Selection of households higher property value households; and, the composi-
tion of waste produced by each group is also different.
In total, 322 households were selected for the study, Thus, if only the number of households in each group
which is about 1% of the number of households in the was considered, only a very small number of high
municipality of Moratuwa. Households were selected income property assessment tax value households
using the stratified random sampling technique. The would be sampled. A sampling strategy based solely
samples were stratified according to wards first, so that on the number of households in each group will not,
all areas of the municipality were represented in the therefore, be representative. On the other hand, a
study. The number of samples from each ward was sampling strategy based solely on standard deviation
selected in proportion to the number of households in will also be deficient, because the representation from
each ward. The households then were stratified according the low property value households will be very small.
to property assessment tax values to obtain a good
representation of all living standards. About 100 addi-
tional households were selected for verification studies.
Table 8 Regression analysis of the amount of organic waste
When stratifying according to annual property generated per household (OHH) for the LM (lower middle
assessment tax values, it was found that about 30% income) group
of households in the municipality have an annual
Variable Unstandardized Std. Standardized T Sig.
property value below Rs.1000 (1 US$=Rs.95), and
coefficients error coefficients
another 30% are in the range of Rs.1,000–3,000. The B Beta
rest of the households were equally distributed
between the ranges of Rs.3,000–6,000 and above Constant 1.228 0.234 5.244 0.000
Rs.6,000. Therefore, the households were stratified Number 0.244 0.116 0.238 2.107 0.039
according to four ranges of annual property assess- employed
ment tax values: Rs.1,000, Rs.1,000–3,000, Dependent variable: OHH
Rs.3,000–6,000 and above Rs.6,000. For each ward, Selecting only cases for which income level = 2_LM
36 Environ Monit Assess (2007) 135:31–39

Table 9 Regression analysis of the amount of organic waste generated per household OHH) for the upper-middle (UM) income group

Variable Unstandardized coefficients Std. error Standardized t Sig.

coefficients
B Beta

Constant 1.611 0.128 12.619 0.000

Number employed 8.091E-03 0.004 0.264 2.254 0.027

Dependent Variable: OHH

Selecting only cases for which income level = 3_UM
The organic waste amount generated per household in the UP group (Table 10) showed only a dependence on the total number of
people in the household.

To consider both the number of households in a refused to participate in the study, but others, though
particular property assessment tax value range and the willing, were not able to participate due to practical
standard deviation of the property assessment tax value reasons. In some households, all members were
within that group, sample sizes were determined in employed and there was no one at home during day
proportion to the number of households and the stan- time. A few residents who agreed to participate in the
dard deviation of property assessment tax values by beginning did not continue the process of sorting of
taking an average value from both. Table 1 summarizes waste throughout the study period. In such instances
the distribution of samples according to the wards and the waste had to be sorted and weighed at the time of
the different property assessment tax value ranges. measurement.
The Municipality of Moratuwa, similar to other Sri However, no difficulties were encountered during
Lankan municipalities, maintains a database of the the collection of information for the survey.
annual property assessment tax values of all house-
holds for each ward which is regularly updated for tax
purposes. This list was obtained, and the households Results and discussion
were first stratified according to the criteria explained
above. Households were then randomly picked for Average waste composition and per capita waste
sampling. generation rates

Assumptions in the approach and limitations The per capita waste generation rates and the average
composition of waste along with descriptive statistics
The primary assumption made in the study was that the
annual property assessment tax value is an indication
of the income and the living standard of the people.
Previous experience has shown that it is difficult to
obtain the correct income directly from people. Annual
property assessment tax value is levied based on the
value of the building occupied. This is assessed by the
local authorities in an objective and independent
manner and all local authorities maintain this data.
However, it is best to use income data where possible.

Difficulties faced during the study

It was not possible to achieve a 100% participation

Fig. 1 Amount of organic waste per household (ohh) in
rate. Some of the households selected from the list of
kilogram as a function of property value (or income level).
assessments available with the municipality did not LO – Low income group, LM – Lower-middle income group,
participate due to many reasons. Very few residents UM – Upper-middle income group, UP – Upper income group
Environ Monit Assess (2007) 135:31–39 37

Table 10 Regression anal-

ysis of the amount of or- Variable Unstandardized Std. Standardized t Sig.
ganic waste generated per coefficients error coefficients
household (OHH) for the B Beta
upper income (UP) group
Constant 1.087 0.271 4.016 0.000
Persons 0.137 0.054 0.227 2.508 0.014
Dependent Variable: OHH per
Selecting only cases for household
which income level = 4_UP

obtained from the statistical analysis are presented in The results of a similar regression analysis under-
Table 2. taken for the lower-middle income group, LM (prop-
From the results in Table 2, the mean composition erty assessment tax value between Rs.1,000 and
of household waste generation in the municipality of Rs.3,000), are presented in Table 5. The same trend
Moratuwa has been determined and is presented in is observed here; however, the constant value was
Table 3 (as a percentage of total waste). slightly lower in this case than for the LO group.
The results of the regression analysis undertaken for
the upper-middle income group, UM (property assess-
Regression analysis of per capita waste ment tax value between Rs.3,000 and Rs.6,000) are
generation data presented in Table 6. Although the same trend is
observed, there is a significant increase in the amount
Regression analyses were conducted on the individual of organic waste generated by this group, indicating
income groups with the amount of waste generated (by consumption trends in Sri Lankan communities. It
different waste type). Although regression analyses appears that higher income groups tend to consume
were conducted for each waste type (organics, paper, higher quantities of food products with a corres-
plastics, etc.) and several independent variables, only ponding increase in organic waste production. This
the results related to the organic waste fraction are indicates a higher propensity to prepare meals at home
presented in this paper. The most important fraction of and a correspondingly higher wastage rate.
household waste is organic waste, in terms of potential The relationship for the upper income group, UP
to cause environmental pollution and resource recov- (property value above Rs.6,000), is shown in Table 7.
ery. The organic waste fraction is responsible for the Again, an increase in the amount of organic waste
quality of leachate created (the higher the organic waste generated is shown, in comparison to the other three
fraction, the higher the biochemical oxygen demand or groups, and there is the same negative correlation
BOD of leachate) and the amount of gas produced in a
sanitary landfill. However, a waste stream with a high
organic content can be processed to produce high
quality compost and thus is advantageous.
A regression analysis performed to delineate a
relationship for the low income group, LO (property
assessment tax value below Rs.1,000), with socio-
economic factors as independent variables shows that
the only statistically significant variable was the total
number of persons per dwelling. The results in Table 4
show a negative correlation between the per capita
organic waste generation and the total number of
persons per dwelling. The results also indicate that
that there is a minimum non-zero amount of organic
Fig. 2 Amount of paper per household (pahh) in kilogram as a
wastes that is generated from a household, irrespec-
function of property assessment tax value. LO – Low income
tive of the number of individuals living in the group, LM – Lower-middle income group, UM – Upper-middle
household. income group, UP – Upper income group
38 Environ Monit Assess (2007) 135:31–39

between per capita waste generation and the number Conclusions

of inhabitants in a household. In this case, however,
the total number of employed individuals in a The study showed that the average amounts of waste
household was also taken as a significant independent generated per capita per day in the municipality of
variable in the second run of the regression analysis. Moratuwa were: 374 g of organic waste, 18.9 g of paper,
The analysis showed that, in the upper income group, 14.1 g of plastic, 66.9 g of glass and 29.7 g of metal. The
as the number of employed individuals in a household biodegradable organic constituents form the largest
increases so does the amount of organic waste category (about 90%) in the household waste stream.
generated, perhaps due to the high level consumerism Thus, a waste management strategy for the region
within this income group. should focus on managing biodegradable organic waste,
considering this fraction has the potential to impact the
quality of leachate and gas generation if deposited in a
Regression analysis of household total waste sanitary landfill.
generation data The data analyses showed that per household
generation of organic and paper waste increase with
The analysis of organic waste generated per house- increase in property assessment tax value or income
hold (OHH) in the LO group showed no relationship level (Fig. 2). This is explained by the relatively high
between OHH and independent variables. However, food consumption trends of higher income groups and
in LM group, the total number of employed individ- the increased purchases of packaged products and
uals in a household showed a statistically significant reading material.
correlation with OHH (see Table 8). Therefore, this The study also showed that as the number of people
variable was chosen for the analysis. A positive in a household increases, there is a reduction in the per
relationship was shown indicating that as the number capita waste generation rate, thereby establishing the
of employed individuals increased so did the amount fact that when waste generation parameters are consid-
of organic waste generated. This is expected because ered, per household waste generation is as important as
as the number of employed individuals increases so the per capita waste generation rate. The number of
does the household income; and previous analyses employed people in a household also appears to be a
showed a direct relationship between the income and contributing factor to waste generation. The average
organic waste generation. The UM group also showed amounts of waste generated per households of different
a similar relationship (see Table 9). income levels can be used to predict the total amount of
waste generated within a municipality. In Sri Lanka,
since municipalities in the country maintain a database
Relationship between organic waste quantities of property assessment tax values, the study findings
and living standards can be extrapolated to other municipalities to estimate
waste generation and composition.
The Fig. 1 shows the organic waste generation per
household as a function of the property assessment
tax value groups, which is the surrogate for living References
standards of occupants of the household. Although a
statistically significant relationship could not be Al-Momani, A. H. (1994). Solid waste management: Sampling,
obtained through regression analysis, the bar chart analysis and assessment of household waste in the city of
shows a relationship between the two (Table 10). Amman. International Journal of Environmental Health
Research, 4, 208–222.
Except for the lower-middle income category, there is
Brunnr, P. H., & Ernst, W. R. (1986). Alternative methods for
a clear increase in organic waste quantities as the the analysis of municipal solid waste. Waste Management
property assessment tax value increases.Bar charts & Research, 4, 147–160.
between property assessment tax values and glass, Daskapoulos, E., Badr, S. D., & Probert, S. D. (1998).
Municipal solid waste: A prediction methodology for the
plastic and metal waste have shown that there is no
generation rate and composition in the European Union
relationship between income level and household countries and the United States of America. Resources,
generation of glass, plastic and metal. Conservation and Recycling, 24, 155–166, (1998).
Environ Monit Assess (2007) 135:31–39 39

Gay, A. E., Beam, T. G., & Mar, B. W. (1993). Cost-effective Medina, M. (1997). The effect of income on municipal
solid-waste characterization methodology. Journal of solid waste generation rates for countries of varying
Environmental Engineering, 119(4), 631–644. levels of economic development: A model. Journal of
Grossmann, D., Hudson, J. F., & Marks, D. H. (1974). Waste Resource Management and Technology, 24(3), 149–155,
generation models for solid waste collection. Journal of 1997.
the Environmental Engineering Division, 100, 1219–1230. Purdy, S., & Sabugal, F. (1999). Waste composition/genera-
Hockett, D., Lober, D. J., & Pilgrim, K. (1995). Determinants tion study for the city of Davao, Mindanao, Philippines.
of per capita municipal solid waste generation in the sub- Proceedings of Sardinia 99, seventh international waste
eastern United States. Journal of Environmental Manage- management and landfill symposium, Cagliari, Italy,
ment, 45, 205–217. 4–8 October, 1999.
King, B. J., & Murphy R. C. (1996). Survey to estimate Wertz, K. L. (1976). Economic factors influencing house-
residential solid waste generation. Journal of Environmental holds’ reproduction of refuse. Journal of Environmental
Engineering, 122(10)897–901. Economics and Management, 2, 263–272.

View publication stats