Int. Journal of Business Science and Applied Management, Volume 4, Issue 1, 2009
Reverse logistics for recycling: The customer service
determinants
Patricia Oom do Valle
Faculty of Economics, University of Algarve
Edificio Ciencias I, Campus de Gambelas, 8005-139 Faro, Portugal
Tel: +351 289800915
Email: pvalle@ualg.pt
Joao Menezes
ISCTE - Business School, Lisbon University Institute
Avenida das Forcas Armadas, 1649-026 Lisbon, Portugal
Tel: +351 217935000
Email: joao.menezes@iscte.pt
Elizabeth Reis
ISCTE - Business School, Lisbon University Institute
Avenida das Forcas Armadas, 1649-026 Lisbon, Portugal
Tel: +351 217935000
Email: elizabeth.reis@iscte.pt
Efigenio Rebelo
Faculty of Economics, University of Algarve
Edificio Ciencias I, Campus de Gambelas, 8005-139 Faro, Portugal
Tel: +351 289800915
Email: elrebelo@ualg.pt
Abstract
Customer service is a central concern in the logistics practice and a study topic in the forward logistics
research. This article investigates the elements of customer service and their importance in reverse
logistics for recycling. Since consumer is the first intervenient in any reverse system that aims to
recycle household residues, the provision of an adequate customer service gains an increased
importance. Applying multivariate statistical methods (exploratory factor analysis, confirmatory factor
analysis and discriminant analysis) to the data from a sample of 267 Portuguese citizens, this study
identifies the levels of customer service in this reverse logistics chain and evaluates their relative
importance in achieving consumers’ participation. The study finds that, as in forward logistics, the
customer service in reverse channels for recycling also has a hard and a soft level, being the former
more important than the later. The results of this research suggest important guidelines to improve such
a complex logistics service.
Keywords: reverse logistics, customer service, multivariate statistics
Acknowledgements: The authors acknowledge Susy Rodrigues for providing proofreading support.
Int. Journal of Business Science and Applied Management / Business-and-Management.org
2
1 INTRODUCTION
Reverse logistics is the continuous logistic process through which shipped products move from the
consumer back to the producer for possible reuse, recycling, remanufacturing or disposal (Johnson,
1998). The European Working Group on Reverse Logistics (RevLog, 2002) describes reverse logistics
as “the process of planning, implementing and controlling the flows of raw materials, in process
inventory, and finished goods, from a manufacturing, distribution or usage point to a point of proper
disposal”. The purpose of a reverse logistics process is to regain the value of returned materials or
provide the means for proper disposal (Rogers and Tibben-Lembke, 1999, 2001). Forward logistics, in
contrast to reverse logistics, focuses on the flow of goods from the producer to the consumer.
As Maltz and Maltz (1998) propose, customer service in the forward logistics channels is a
multifaceted concept that can encompass either objective or perceptual elements. Objective elements
correspond to basic customer service (or hard service) such as inventory availability, on time delivery
and order cycle time reliability. Perceptual elements (or soft service) are those related to the suppliers’
ability to respond to specific customer requests such as after-sale service and effective handling of
information requests. Several authors recognize that customer service is an issue of central concern in
logistics research and practice (Byrne and Deeb, 1993; Emerson and Grimm, 1998; Fuller, 1978;
Giuntini and Andel, 1995; Kopicki et al., 1993; Maltz and Maltz, 1998; Marien, 1998; Murphy, 1986;
Stock, 1992; Zikmund and Stanton, 1971).
Reverse logistics systems for recycling begin with the consumer and finishes with the end market
(Jahre, 1995). These systems can be more or less complex depending on whether they possess
intermediate levels, such as, the collection level, the transfer level and the processing level. Consumers
have a particularly important role in this reverse logistics system since they are the first link in the
overall logistics chain. Without consumer participation (through the sorting and disposing of recyclable
materials), this system would not be possible. By providing a convenient system, customer service
becomes the touchstone in creating value for consumers as well as in securing their participation
(Turner et al., 1994).
As recently pointed out, most research in the reverse logistics field is essentially descriptive and
based on subjective evidence rather than on theoretical bases (Alvarez-Gil et al., 2007). In terms of the
reverse logistics systems for recycling, one gap that remains open is the comprehensive investigation of
the main elements of customer service that explain the consumer involvement in selective-collection
programs. This analysis would provide fundamental information about the most important customer
service elements and, thus, that require more attention and investment. The contribution of this study
lies in bridging this research gap.
Data for this research results from the outcome of a structured questionnaire collected from a
random sample of 267 Portuguese citizens. This study uses a three-step procedure to assess the
elements that comprise customer service. First, an exploratory factor analysis identifies the main levels
of customer service (both hard and soft) that shape consumer participation in the Portuguese recycling
program. Second, a confirmatory factor analysis validates the underlying levels and the corresponding
elements. Third, a discriminant analysis identifies the level of customer service that strongly predicts
consumer involvement, in this way offering future guidelines for the reverse logistics system at the
collection stage.
The structure of this paper is as follows. Section 2 summarizes the background literature in terms
of: (1) concept and origins of reverse logistics and (2) reverse logistics for recycling. Section 3
proposes a conceptual model that forms the basis for this research and puts forward a set of research
hypotheses. Section 4 describes the research methods used including data information and statistical
techniques. Section 5 presents the results and provides conclusions based on the research hypotheses.
Section 6 discusses the study’s theoretical and managerial implications, identifies its limitations and
proposes guidelines for further research.
2 BACKGROUND ON REVERSE LOGISTICS FOR RECYCLING
Recycling is a resources recovery option that enables the use of part or all materials from returned
goods, either by their original producer(s) or by other industries (RevLog, 2002). The recycling process
essentially encompasses two stages (Jahre, 1995). The first is the collection service stage and includes
all the necessary procedures that make recyclables possible for further reprocessing. The second is the
reprocessing stage from the collection of materials to the replacement of primary raw materials. Table 1
lists the studies that explore particular issues on reverse logistics for recycling.
Patricia Oom do Valle, Joao Menezes, Elizabeth Reis and Efigenio Rebelo
3
Table 1: Summary of articles on reverse logistics for recycling
Reference
Main topics
investigated
Material (in)
Material (out)
Driver(s)
Guiltinan and
Nwokoye (1975)
Reverse logistics
networks
Recyclables in
general
Materials
Social benefits
Economic benefits
Pohlen and Farris
(1992)
Reverse logistics
networks
Transportation
issues
Plastics
(*)
Environmental
concerns
Bronstad and
Evans-Correia
(1992)
Purchase of
recycled materials
Paper
Paper
(*)
Kopicki et al.
(1993)
Logistic
implications of
recycling (and
reuse) programs
Recyclables in
general
Materials
Social benefits
Economic benefits
Gupta and
Chakraborty (1994)
Planning and
control of recovery
activities
Glass scrap
Raw materials
Cost savings
Jahre (1995)
Reverse logistics
networks
Household waste
Substitutes for
primary
materials
Legislation
Faria de Almeida
and Robertson
(1995)
Incentives to
stimulate recovery
(timely and clear
information)
Batteries
Materials
(*)
Spengler et al.
(1997)
Reverse logistics
networks (private
networks)
Steel products
Reusable
products
Disposal cost saving
Public waste
management
Fuller and Allen
(1997)
Reverse logistics
networks
Post-consumer
recyclables
Substitutes for
primary
materials
Legislation
Yender (1998)
Incentives to
stimulate recovery
(Easy and simple
method of supply)
Batteries
Raw materials
Batteries
(*)
Barros, Dekker and
Scholten (1998)
Reverse logistics
networks (public
networks)
Construction waste
Sand
Waste disposal
Environmental
regulation
Nagel and Meyer
(1999)
Information and
communication for
reverse logistics
End-of-use
refrigerators
Plastics
Metals
No longer needed
Legislations
Costs savings
Lowers, Kip,
Peters, Souren and
Flapper (1999)
Reverse logistics
networks (private
networks)
Carpets
Fibres, etc.
Image
Expected legislation
Economic advantages
Realff, Ammons
and Newton (2000)
Reverse logistics
networks (private
networks)
Carpets
Fibres
(*)
Chang and Wei
(2000)
Reverse logistics
networks (public
networks)
Household waste
(*)
Waste disposal
Reducing costs
Environmental
concern
Note: (*) Not mentioned.
Although the concept of reverse logistics arises in the 1990s, the discussion on the structure of
logistics channels begins much earlier. Guiltinan and Nwokoye (1975) identify the main types of
logistics structures, functions and members that form part of the distribution channel. The study also
points out a number of key factors for the future development of recycling channels, such as “[the need
to expand] efforts in identifying potential markets and buyers of recycled materials; more extensive
contact with, and promotion to, final buyers; [in expanding] capacity for moving increased volumes of
material to achieve and maintain scale economies; and [in improving] flexibility in transportation”
(Guiltinan and Nwokoye, 1975: 35).
While the study of Guiltinan and Nwokoye (1975) does not focus specific recyclable materials,
Table 1 presents other contributions that address particular reverse logistic networks for recycling.
Int. Journal of Business Science and Applied Management / Business-and-Management.org
4
Pohlen and Farris (1992), for instance, analyze the set-up of recycling networks for plastics and
propose a more complex structure for the reverse channel when compared to the general Guiltinan and
Nwokoye (1975) approach. Pohlen and Farris (1992) discuss the main issues that affect reverse
logistics channels for recycling, namely, efficiency improving factors in terms of existing channels and
common forms of improving recyclables.
As Table 1 shows, some of the studies that address the organization of recycling networks focus
on public networks, while others describe private systems. In the first case, environmental concerns and
waste disposal legislation are the main motivations underlying reverse logistics. Contrary to this notion
are private reverse logistics networks that handle residues or end-of-life products in which recycling is
economically more attractive. Private processors finance the transportation of these materials as well as
the recycling process itself. For recycling to be economically viable, a significant amount of discarded
products (or parts) need to be processed.
The reverse logistics literature for recycling also explores the planning and control of recovery
activities (i.e., the decisions about what to collect, disassemble and process, and in what quantities,
how, when and where), the available information and communication systems (e.g., software, data
requirements), the logistical implications of recycling, and the implementation of programs to increase
the demand of recyclable materials. As Table 1 shows, the studies examining these issues explore only
one type of recyclable material.
Finally, the scope of this research explores the incentives that may stimulate a desired behavior in
specific members who form part of the reverse channels. These incentives look for encourage / impose
cooperation either in terms of reception or delivery of goods for recovery. In the first case, companies
may have some goods that they wish to dispose of, and, through incentives, influence others (e.g., the
goods’ providers) in accepting such requests, in order to avoid high disposable costs. The second case
includes situations in which the purpose is to encourage others (final consumers) to take part in efforts
which allow companies to manage goods (products, parts, packaging) for recovery. Table 1 identifies
three types of non-economic incentives: timely and clear information, general convenience, and an easy
and simple method of supply.
Table 1 evidences that only two studies on reverse logistics for recycling have specifically focused
on household residues and they address the topic of how to design and manage the logistics networks
(Jahre, 1995; Chang and Wei, 2000). In other words, despite the significant amount of research on
reverse logistics during the last years, no study has so far identified what elements of customer service
are important predictors of consumer involvement in the reverse logistics system for recycling. To
fulfil this research gab is the overall objective of the current study.
3 CONCEPTUAL FRAMEWORK AND RESEARCH HYPOTHESES
Incentives are of particular relevance in the context of encouraging consumer to separate and
properly dispose of household packaging residues for recycling. In this case, the consumer is the
starting point of any reverse logistics for recycling household waste and, therefore, his or her
participation is a needed condition for a recycling system to exist. Although this topic has not been
explored in the reverse logistics literature, in the field of environmental social-psychology, some
studies address the predictive effect of a convenient recycling program in articulation with other
potential determinants of environmentally friendly behavior. Essentially, the review of the literature
shows that an increase in consumer involvement can result from the following aspects: (1) closer
proximity of disposal recipients, (2) minimal complexity in storing and storage of recyclable materials,
(3) accessible information on what is recyclable and the location of collection points, and (4) reliable
frequency of collection. Table 2 summarizes the main characteristics and conclusions from these
studies.
Patricia Oom do Valle, Joao Menezes, Elizabeth Reis and Efigenio Rebelo
5
Table 2: Customer service elements as predictors of recycling behavior
Consumer-service
levels
Variable under
analysis
Participants
Finding (*)
Proximity of disposal
recipients
Self-reported
recycling
Residential
households
+
Participation in
recycling
Residential
households
+
Participation in
recycling
Residential
households
+
Self-reported
recycling
Residential
households
+
Willingness to
participate in a
recycling program
Residential
households
+
Self-reported
frequency of recycling
and observed amount
of collected materials
Residential
households
+
Participation in
recycling
College students
+
Self-reported
recycling
Residential
households
+
Minimal complexity in
separating and storing
Self-reported
recycling
Residential
households
+
Reliable frequency of
collection
Participation in
recycling
Residential
households
+
Amount of collected
material
Residential
households
+
Participation in
recycling
Residential
households
0
Amount of collected
material
Residential
households
+
Availability of
information about
Amount of collected
material
College students
+
what to recycle and
where
Participation in
recycling
Residential
households
+
Participation in
recycling
Residential
households
+
Self-reported
recycling
Residential
households
+
Self-reported
recycling
Households from
communities with
recycling education
programs
+
Observed participation
in recycling
College students
+
Self-reported
frequency of recycling
Residential
households
+
Participation in
recycling
Residential
households
0
Participation in
recycling
Residential
households
0
Amount of collected
material
College students
+
Self -reported and
observed recycling
Residential
households
+
Self-reported
frequency of recycling
Residential
households
+
Rate of waste
reduction
Residential
households
+
Note:(*) Legend +: significant positive relationship; -: significant negative relationship; 0: non-significant
relationship.
Int. Journal of Business Science and Applied Management / Business-and-Management.org
6
Based on the literature review, Figure 1 depicts the proposed model of consumer involvement in
the reverse logistics system for recycling. The model establishes a direct causal-effect relationship of
customer service in terms of consumer involvement. The following hypothesis establishes this
connection:
Figure 1: Conceptual model of customer service in the reverse logistics system for recycling
Hypothesis 1: Customer service explains consumer involvement in the reverse logistics system
for recycling household packaging.
In terms of forward distributor channels, Maltz and Maltz (1998) state that the concept of
customer service includes two common levels: (1) hard level that corresponds to objective or basic
customer services, and (2) soft level that refers to perceptual customer service elements, besides those
of basic service. The logistics literature widely accepts this classification of customer service levels
(Dadzie et al., 2005; Mentzer et al., 1989; Stock and Lambert, 2001).
Earlier research is clear about the meaning of basic customer service in forward reverse logistics.
As Dadzie et al. (2005) summarize, this construct includes in-stock availability and cycle time. The
literature on reverse logistics for recycling does not address this issue. This study proposes that the hard
elements of customer service lie in the accessibility to the selective collection recipients. By providing
this basic element to consumers, the reverse logistics system for recycling can operate. The remaining
customer service elements that Table 2 presents depend on this element, that is, they are important only
if communities have specific recipients for separated materials at their disposal. Consumers may
perceive the separation process as satisfactory, understand how and why to separate, believe that, in
general, the collection frequency is adequate, but without a collection service in place (well located and
not too distant) they will not participate in the recycling program. The second research hypothesis
represents these considerations:
Hypothesis 2: The hard level of customer service in the reverse logistics system for recycling
household packaging consists in the access to the selective collection facilities.
Identifying the soft elements of customer service in the forward reverse logistics systems is not as
clear as identifying hard elements. According to the definition proposed by Dadzie et al. (2005), soft
elements are those other than in-stock availability or time cycle, that is, those that are not basic service.
These elements are not objective but instead perceptual and result from suppliers’ ability to respond to
specific customer requests (Maltz and Maltz, 1998). In forwards reverse logistics systems, soft
customer service elements include error correction, follow up on customer complaints, after sales
services, and effectiveness in the handling of information (La Londe and Zinszer, 1976).
The current study on the elements of customer service in reverse logistics for recycling follows a
similar approach and considers that the soft elements go beyond just basic service, that is, beyond the
easy access to specific disposal recipients for recyclable materials. As Maltz and Maltz (1998) refer,
soft elements result from consumers’ perceptions about customer service rather than from the objective
characteristics of this service. Therefore, soft elements should necessarily include the perceived
Customer
service
Hard level
(Access to disposal
containers)
Consumer
commitment
Soft level
(Elements other
than accessibility to
disposal recipients)
H
1
H
2
H
3
H
4
H
4
Patricia Oom do Valle, Joao Menezes, Elizabeth Reis and Efigenio Rebelo
7
complexity in storing and separating the recyclable materials, the availability of information on how to
recycle and where to dispose of the recyclable materials and a reliable periodic collection. Table 2
summarizes the existing research on these elements. Besides these elements, this study explores
whether hygiene, design and security at the disposal locations, as well as available support and claim
services have an important soft elements of customer service. Based on this discussion, the study
establishes a third research hypothesis:
Hypothesis 3: The soft level of customer service in the reverse logistics system for recycling
household packaging includes all elements beyond accessibility to the selective collection facilities.
As Table 2 shows, research has focused mostly on the proximity element of disposal recipients
and, according to studies, has a consistent and positive effect on consumer participation in terms of
recycling. This Table also suggests that the effect of the remaining elements (soft elements, as this
study looks to show) on consumer participation is only moderately apparent. In fact, some studies show
that the more complex the sorting process is, together with limited recycling information, the lower the
recycling participation rates. However, a small number of studies find a non-significant relationship
between these variables. Given these considerations, and the fact that soft level of customer service can
only exist after the provision of the hard level, the final research hypothesis is as follows:
Hypothesis 4: The hard level of customer service is the main determinant for consumer to commit
to the reverse logistics systems for recycling, followed then by the soft level.
4 METHODS
4.1 Setting
The Green Dot Society (GDS) is a private company, created in 1997 with the purpose of managing
the Integrated Recovery System of Packaging Waste Management (IRSPWM). Currently, GDS is the
only company that develops this type of activity in Portugal. GDS is essentially a reverse logistics
aggregator with a shareholder structure composed of three holdings that represent almost 200
companies. The first holding represents the packagers/importers, the second represents the distribution
and retail trade, and the third represents the manufacturers and recyclers of packaging material. In
compliance with national legislation, GDS aims to recover 60% of the overall packaging weight and
recycle 55% of this material by the end of 2011. Recyclable materials include glass, paper/cardboard,
lightweight packaging (plastic, metal) and wood. With the exception of this last type of material, drop-
off systems, often referred to as eco-points, allows for the collection of packaging residues.
As in other European countries, the IRSPWM relies on the principle of shared environmental
responsibility. Packers and importers finance the system, based on the polluter-pays principle in which
the amount and weight of the corresponding packaging material, commonly known as the green spot
value, regulates the fee they must pay. In turn, packers and importers receive permission to mark their
packaging with the Green Spot symbol, which shows that these companies transfer their recovery
responsibility to GDS and the IRSPWM. The distribution role ensures that their commercial confines
only sell non-reusable packaging through the Integrated System. The GDS’s business structure does not
include municipalities though they are responsible through contract agreement for the multi-material
collection and sorting of household packaging residues.
Consumers should necessarily separate and dispose of their packaging waste at the eco-point. The
packaging manufacturers complete the cycle by securing the recycling of collected household
packaging. The GDS’s overall mission is to manage the reverse supply chain, finance and guarantee the
functioning of the entire system. This corporation invests a major part of its annual overall income to
compensate for the additional costs that municipalities incur with multi-material collection and sorting.
GDS also sub-contracts transportation services that handle packaging residues for recycling companies
and ensures that they receive, store and recycle recovered material.
4.2 Questionnaire
Data of this research result from personal interviews performed in April and May of 2006 based
on a structured questionnaire (appendix 1). The questionnaire design took into account an extensive
review of scientific and practitioner publications on recycling behavior, interviews on key elements of
GDS management and benchmark studies carried out in other European countries (Spain, Italy and
Belgium). The questionnaire encompasses three sections. Section 1 conducts an inquiry of the socio-
demographic characteristics: gender, age, educational qualification, marital status, occupation,
Int. Journal of Business Science and Applied Management / Business-and-Management.org
8
residence type, home ownership and family monthly income. Section 2 involves eleven elements
(included in the earlier study) characterizing customer service in the reverse logistics system for
recycling: (1) location of disposal recipients, (2) frequency of waste collection, (3) distance to the
disposal recipients, (4) number of disposal recipients, (5) cleaning and maintenance of disposal
recipients, (6) local safety, (7) emptying regularity, (8) available information, (9) support and claim
service, (10) system adequacy to lifestyle, and (11) number and type of suitable waste materials. A
Likert five-point scale assesses these elements, ranging from 1 – very unsatisfied to 5 – very satisfied.
Section 3 looks to measure consumer involvement in the recycling program and considers two
questions. The first measures the self-reported household recycling behavior (scale: 1 – separates and
selectively discards recyclable waste, 0 – does not separates and selectively discards recyclable waste).
The second evaluates the frequency of separation and disposal of recyclable materials at the eco-points
(scale: 1 – never, 2 – sometimes, 3 – always).
4.3 Sample and data
The study population encompassed the adult Portuguese citizens living in Faro city. Faro is the
capital of Algarve, located in the southern Portugal, comprehending six parishes. Faro has a total
population of 58 350 inhabitants and its most important economic activities are tourism and services.
From this population, the study selected a random sample of 267 citizens. The calculation of the sample
used the most conservative estimate for a single proportion (p = 0.5), a confidence level of 95% and a
maximum error of 6%. The study used stratified sampling and the distribution of the interviews
according to parishes was proportional to the resident population. In each parish the most important
shopping street was selected as the location to perform the interviews. College students administered
the questionnaires to respondents in those streets, with respondents chosen at random, according to a
systematic procedure. A questionnaire was delivered to the first person (older than 14) passing near the
interviewer at a defined hour. Then, a sampling interval of 5 people was established in order to select
the remaining respondents and, thus, to fill the sample stratum defined to each parish. Each respondent
received an explanation of the nature of the questionnaire.
Table 3 summarizes the socio-demographic characteristics of the respondents along with some
household features and participation patterns. Around 61% of respondents replied as being active
participants in the recycling program. The profile of the respondents corresponds roughly to that in the
previous national study (GDS, 2000). Most respondents were females (59.4%), between the ages of 26–
35 (27.2%) and 36–45 (18.2%), married (51.1%) and university degree holders (34.2%).
Table 3: Characteristics of the sample
Variables
Distribution of answers
Gender
59.4 % female
40.6% male
Age
14 – 25: 15.1%
26 – 35: 27.2%
36 – 45: 18.2%
46 – 55: 17.0%
56 – 65 :11.0%
66 – 94: 11.5%
Education level
4 years: 10.1%
6 years: 7.3%
9 years: 9.0%
12 years: 26.7%
technical/ professional: 12.7%
College or higher: 34.2%
Median education level: 12 years
Occupation
Farmer/fisher: 1%
Workman: 14.1%
Services worker: 22.4%
Public worker: 8.2%
Teacher: 7.3%
Liberal worker: 5.4%
Manager: 9.2%
Retired: 5.1%
Housewife: 4.3%
Student: 18.0%
Other: 5.1%
Marital status
Married: 51.1%
Single: 34.0%
Divorced: 9.8%
Widow:5.1%
Residence type
Apartment: 63%
House: 27%
Farm: 10%
Home ownership
Own/are buying: 75%
Renting: 18%
Familiar: 7%
Monthly family income
Less than 324 €: 2.1%
324 € – 499 €: 10.3%
500 € – 999 €: 39.3%
1000 € – 1999 €: 31.2%
2000 € – 2999 €: 11.7%
At least 3000 €: 5.4%
Self-reported recycling
behavior
Uses to separate and selectively
dispose of: 61%
Do not separate nor selectively
dispose of: 39%
Frequency of separation and
disposal of recyclable materials
Never: 22%
Sometimes: 27%
Always: 51%
Patricia Oom do Valle, Joao Menezes, Elizabeth Reis and Efigenio Rebelo
9
4.4 Data analysis methods
This study uses the following methods of multivariate data analysis: exploratory factor analysis
(EFA), confirmatory factor analysis (CFA) and discriminant analysis (DA). EFA and CFA enable the
assessment of hypotheses 2 and 3. DA permits to test hypotheses 1 and 4.
The application of EFA with varimax rotation on the set of eleven elements of customer service
allows for the reduction of the proposed instrument’s dimensionality (Hair et al., 1998; Reis, 1997).
The computation of the Kaiser-Meyer-Olkin (KMO) statistic and the results from the Bartlett test
establishes whether using EFA is possible in this research. The KMO statistic is a ratio that ranges from
0 to 1, and should be at least 0.7 for EFA be acceptable. The Bartlett test examines whether the
correlation matrix of the variables is significantly different from the identity matrix. The use of EFA is
adequate if this test rejects the null hypothesis that these matrices are equal. The Alfa Cronbach
coefficients evaluate the reliability of the final customer service factors that result from the EFA.
CFA evaluates the factors’ psychometric properties in terms of reliability and validity, based on
the AMOS 6 software package. The prior EFA defines the model that CFA requires to associate the
latent factors to the observed customer service elements (the observed variables). Given the absence of
normality in the data, this analysis uses the weighted least squares estimation method (Schumacker and
Lomax, 1996).
The analysis of the overall model fit relies on three types of measures: absolute fit, incremental fit
and parsimonious fit (Hair et al., 1998). The absolute fit evaluation adopts the Chi-square goodness-of-
fit test, the goodness of fit index (GFI) (Joreskog and Sorbom, 1986) and the root mean square residual
of approximation (RMSEA) (Steiger, 1990). The Chi-square test is a general indicator of how well the
estimated model fits with the data. The Chi-square value should be low and not statistically significant
to achieve goodness of fit. Similarly, the GFI should exceed 0.9 (GFI range from 0 to 1 with 1 meaning
perfect fit) and the RMSEA show be low (zero suggests a perfect fit). This study considers the
following incremental fit measures: the adjusted goodness of fit index (AGFI) (Joreskog and Sorbom
1986), the normed fit index (NFI) (Bentler and Bonnet, 1980), the Tucker and Lewis index (TLI)
(Tucker and Lewis, 1973), the incremental fit index (IFI) (Bollen, 1988) and the comparative fit index
(CFI) (Bentler, 1990). These measures range from 0 (no fit) to 1 (perfect fit). The parsimonious fit
index that this study observes is the normed Chi-square measure (Joreskog, 1969) that should range
from 1 to 5, ideally.
The interpretation of the results of CFA follows with a reliability assessment of the proposed
measurement model. Reliability analysis refers to whether the observed variables (i.e., the customer
service elements), chosen to indicate each latent variable (i.e., each customer service factor or level),
are actually measuring the same concept. This study considers two measures of reliability: composite
reliability and variance extracted. Composite reliability shows the degree to which the observed
variables adequately represent the corresponding latent variable. Variance extracted complements the
composite reliability and expresses the total variance of the observed variables that the latent variable
explains. The latent variables present appropriate levels of reliability if composite reliability and
variance extracted exceeds the acceptance level of 0.7 and 0.5, respectively (Scharma, 1996).
Next, the analysis focuses on the evaluation of the convergent and discriminant validity of each
latent variable (customer service level). Convergent validity evaluates whether the observed variables
really measure the corresponding latent construct. The significance and the size of the observed
variables’ weights permit evaluating of this type of validity (Bollen, 1989). Discriminant validity
focuses on whether strong correlations exist between the latent constructs (which indicates poor
discriminant validity) or weak correlations (which suggests strong discriminant validity). Within the
CFA method, the matrix of standardized correlations between the latent variables allows assessment of
this type of validity. This study expects a significant correlation between the latent constructs since
they represent the dimensions of a general construct, the customer service. Another expected result is
that these correlations, although statistically significant, are not very high, because the latent constructs
should be measuring different levels of customer service (hard and soft). Hair et al. (1998) suggest that,
for purposes of discriminant validity, the correlations between latent variables should not exceed 0.7.
DA uses the final customer service levels that EFA and CFA suggest as discriminant variables.
This study carries out two DAs. In the first one, the dependent variable represents self-reported
recycling behavior, with two categories (1 – separates and selectively discards reusable waste, 0 – does
not separate nor selectively discards reusable waste), giving rise to one discriminant function. For the
second DA, the dependent variable is the frequency of separation and disposal of recyclable materials
at the eco-points, which is tri-categorical (1 – never, 2 – sometimes, 3 – always), producing two
discriminant functions. In each case, the Box M’s test assesses the underlying assumption of the DA,
that is, that the matrices of variances and covariances for all groups that the dependent variable defines
are equal.
Int. Journal of Business Science and Applied Management / Business-and-Management.org
10
For each DA, the estimation of the discriminant functions uses a random sample of half of the
cases (the analysis sample). The other half (holdout sample) allows validation of DA (Fernández and
Martinez, 2000). For each analysis, the Wilks’s lambda statistics test whether the discriminating
function(s) significantly differentiates the groups defined by the dependent variable. In this test, the
null hypothesis is that the groups defined by the dependent variable have the same mean in the
discriminating function(s).
To assess the predictive ability of the discriminant function(s), this study analyses the
classification matrix that results from each DA. Inside this matrix, the number of cases that the analysis
correctly classifies within each group appears on the principal diagonal. The overall hit ratio is the
global percentage of correct classifications in all the groups.
Two procedures evaluate the classification accuracy of the discriminant functions (Huberty, 1994;
Klecka, 1980). The first procedure compares the overall hit ratio for both samples (the analysis sample
and the validation sample) either with the maximum chance criterion or with the proportional chance
criterion. The maximum chance criterion is the percentage of cases in the larger group. Alternatively,
the proportional chance criterion takes into account the proportion of cases in all groups. In particular,
for a dependent variable with k categories (i.e., k groups), the expression is as follows:
k
2
pro i
i1
Cp
where p
i
is the percentage of cases in group i (i=1,2,…k). The second procedure is to establish and
evaluate the Press' Q statistic. This statistic tests the null hypothesis which states that the discriminating
ability of the classification function(s) is not significantly different than the classification by chance.
The null hypothesis in this test is that the number of cases correctly classified resulting from the
discriminant analysis does not exceed the number of cases correctly classified by chance. The Press' Q
statistic should be high and statistically significant for the purpose of DA’s validation.
5 RESULTS
5.1 Exploratory factor analysis
EFA allows the reduction of the original eleven elements into two factors, in which both account
for 71.3% of the total variance (KMO = 0.87; Bartlett test p = 0). Table 4 summarizes the main results
of this analysis. The observation of the elements with higher loadings for each factor justifies the
corresponding chosen label. Factor 1, hard level, gathers the customer services elements (CSE) that
reflect the availability and accessibility to selective collection recipients: the distance to the disposal
recipients, their location and the available quantity. Factor 2, soft level, includes the remaining elements
such as aspects relating to disposal conditions (indicated by local safety, cleaning and maintenance and
frequency of waste collection), available information on recycling (indicated by information
availability, support and claim services) and system adequacy (indicated by system adequacy to
lifestyle and number and type of accepted materials). The high Cronbach’s alpha coefficients for each
factor suggest that they have a very good degree of internal consistency. The factors take into account
the hypothesized customer service elements, a sign in support of hypotheses 2 and 3.
Table 4: Customer service elements (CSE) and factors (After varimax rotation)
Factors
Loadings
% of Explained Variance
and Cronbach
Factor 1 – Hard level
CSE3 – Distance to the disposal recipients
0.86
42.1%
CSE1 – Location of disposal recipients
0.78
Cronbach =0.92
CSE4 – Number of disposal recipients
0.56
Factor 2 – Soft level
CSE2 – Frequency of waste collection
0.84
29.2%
CSE6 – Local safety
0.81
Cronbach =0.87
CSE11 – Number and type of accepted waste materials
0.73
CSE7 – Emptying regularity
0.71
CSE5– Cleaning and maintenance of disposal recipients
0.69
CSE8 – Available information
0.64
CSE9 – Support and claim service
0.62
CSE10 – System adequacy to lifestyle
0.59
Patricia Oom do Valle, Joao Menezes, Elizabeth Reis and Efigenio Rebelo
11
5.2 Confirmatory factor analysis
Figure 2 shows the standardized estimates on the CFA model. Regarding the absolute fit
evaluation, the Chi-square statistic reports a low and not statistically significant value (p > 0.05),
suggesting that the model adequately describes the data. The remaining measures of overall fit also
present favorable results, indicating an adequate incremental and parsimonious fit: the GFI exceeds 0.9
and the RMSEA are close to 0; the AGFI, the NFI, the TLI, the IFI and the CLI exceed 0.9; the normed
Chi-square lies in the 1 to 5 interval.
For the latent variable hard level, the composite reliability coefficient is 0.76 and the variance
extracted is 0.51. For the soft level, the reliability measures are 0.89 and 0.50, correspondingly. For
both latent variables, these values exceed the minimum threshold of 0.7, in terms of composite
reliability, and are at least 0.5, in terms of the variance extracted. These results support the EFA
findings concerning the factors’ reliability, according to the Cronbach’s alpha coefficients. In terms of
the convergent validity analysis, Figure 2 shows that all observed variables have positive weights that
exceed the acceptable level of 0.4 (Hair et al., 1998). All weights are statistically significant (t tests: p =
0.00). In terms of discriminant validity, Figure 2 reveals that the correlation level between the latent
variables is 0.58. This correlation is moderately high, indicating a correlation between the two latent
variables (i.e., both represent levels of the same construct: customer service) which is not too strong
(i.e., each level captures a somewhat different perspective of the customer service construct).
Thus, EFA and CFA specify two final customer services levels whose contents provide support to
research hypotheses 2 and 3.
Figure 2: Confirmatory factor analysis of customer service elements. Standardized estimates
Note: Chi-square = 59.6 (p = 0.07), GFI = 0.94, RMSEA = 0.03, AGFI = 0.91, NFI = 0.89, TLI = 0.97, IFI =
0.97, CFI = 0.98, Normed Chi-square = 1.32.
5.3 Discriminant analysis
The first DA uses the two factors produced by the EFA as independent or discriminating variables
and the self-reported recycling behavior as the dependent variable (Box’s M test p = 0.17). For the
second DA, the dependent variable indicates participation frequency (Box’s M test p = 0.08). In both
cases, the Wilks’s lambda tests reveal that the groups, defined by the dependent variables, are
statistically different in terms of the customer service satisfaction level (Wilks’s lambda tests: p = 0).
The two discriminant functions arising from the second DA, are both significant in separating the
groups (canonical correlation for the first discriminant function = 0.93; canonical correlation for the
second discriminant function = 0.86; explained variance for the first discriminant function = 68.8%;
explained variance for the first discriminant function = 31.2%). Thus, these findings support the
research hypothesis that variables related to customer service elements determine consumer
involvement in the reverse logistics system for recycling (Hypothesis 1).
The overall hit ratio for both the analysis and holdout samples evaluates the predictive accuracy of
the discriminant functions (one in the first DA, two in the second DA). These ratios (72% and 68.2%,
respectively, in the first DA and 69% and 65.4%, in the second DA) significantly exceed the levels of
the proportional chance criterion and the maximum chance criterion (in both analyses, and in terms of
both criteria, Press Q statistic > χ
2
(1)
for a significance level of 0.05).
Table 5 reports the structural coefficients, representing the correlations between the discriminating
variables and the discriminant functions. Both levels are statistically significant in discriminating the
Hard level
CSE1
Soft level
CSE3
CSE2
CSE6
CSE11
CSE7
CSE5
CSE4
CSE8
CSE9
CSE10
0.89
0.87
0.60
0.81
0.77
0.87
0.86
0.78
0.52
0.56
0.63
0.58
e3
e1
e4
e2
e6
e11
e7
e5
e8
e9
e10
0.48
0..61
0.71
0.52
0.48
0.42
0.59
0.61
0.65
0.54
0.50
Int. Journal of Business Science and Applied Management / Business-and-Management.org
12
groups because in the both DAs and in all the discriminant functions, the structural coefficients exceed
the minimum absolute value of 0.3 (Hair et al., 1998). However, as this table shows, of the two
variables in the first DA, the hard level has the highest structural coefficient and, thus, this factor
discriminates the most; on the contrary, the soft level reports the lowest structural coefficient and,
therefore, this factor discriminates the least. The same finding occurs in the second DA in terms of the
first discriminant function, this of which is the most important. In other words, the fourth research
hypothesis (Hypothesis 4) should not be rejected.
Table 5: Structure matrix of discriminant analyses
Discriminating Variables
Structural Coefficients
DA1
DA2
Discriminant function 1
Discriminant function 2
Hard level
0.75
0.71
-0.33
Soft level
0.62
0.54
0.68
6 DISCUSSION AND CONCLUSION
Consumers are the foremost and decisive link in a reverse logistics chain that aims to recycle
household packaging residues. In fact, without consumers’ involvement and continuous collaboration,
this system cannot exist. This article explores the importance of consumer motivation to participate in
the IRSPWM by ensuring that the recyclable materials are available for the recycling industries. By
using a combination of multivariate statistical methods, this study shows the importance of providing
consumer convenience in order to gain greater involvement in reverse logistics systems for recycling.
On the whole, the research supports the conceptual model of customer service (Figure 1). As in the
traditional forward logistics systems, customer service in the reverse logistics system for recycling
comprises hard and soft levels (hypotheses 2 and 3). As hypothesized, both levels explain consumer
involvement (hypothesis 1), with the hard level representing the most important predictor (hypothesis
4).
This study has theoretical contributions and managerial implications.
Previous studies in forward logistics identify the elements that form the customer service construct
and explore the importance of providing an adequate level of customer service. The first theoretical
contribution of this article lies in exploring this issue in the specific context of reverse logistics for
recycling. The EFA and CFA that this study employs show that the hard level of customer service in
the reverse logistics chain also corresponds to the basic service, which, in this case, means easy access
to specific disposal recipients for recyclable materials. These analyses also show that other customer
service elements form the soft level. These findings support those found in the forward logistics
research. The second theoretical contribution of this study concerns a comprehensive analysis of a
service based on convenience in order to enhance recycling behavior. In fact, existing literature in the
field of environmental social psychology addresses some customer service elements individually
(Table 2). This study focuses on the various elements as a whole, combining them into a single
analysis, and adding new elements (hygiene, security at the disposal sites and the existence of support
and claim services), revealing to be significant factors.
Overall, this study substantiates that consumers are sensitive to several customer service elements
and that their evaluation of this service determines the current self-reported recycling behavior and the
frequency of involvement. These results have managerial implications. The first one is that meeting
customer service demands in terms of customer service requirements must be a priority in planning this
type of reverse logistics networks. This is not difficult to carry out because the customer service
elements of the reverse logistics system for household packaging are manageable variables. Therefore,
their evaluation reveals opportunities and insight to improve the effectiveness of the customer service
concept and, as a consequence, increase consumer involvement in the system.
The study also observes the customer service levels by taking into account their relative
importance in fostering consumer involvement: the hard level is relatively more important than the soft
level. This finding also has managerial implications since it helps prioritize the overall logistics needs
for a more effective selective-collection system. Although the overall organization and performance of
the Portuguese reverse logistics system for recycling requires global improvement, an important
priority us defining the location of the eco-points in terms of easier and more convenient population
access. In establishing this, focus should turn to aspects such as available support and claim service,
more recycling awareness campaigns, and general disposal conditions (cleaning, maintenance, safety,
etc).
Patricia Oom do Valle, Joao Menezes, Elizabeth Reis and Efigenio Rebelo
13
Reverse logistics systems with centralized the disposal facilities (as is the case with the eco-points
in Portugal) are more inconvenient because consumers must transport and deposit recyclable materials
at drop-off points. However, these systems are also less expensive than the curbside alternative. In
curbside schemes, collection is door-to-door, which increases convenience but also collection costs and
ultimately the overall cost of the system. A less expensive collection option is to maintain eco-points
and invest in more convenient locations. However, considering the relative importance of the hard
customer service level as the main determinant of consumer involvement, the possibility of providing
curbside collection, at least temporarily and in a few municipalities, should be considered. As this study
shows, shortening the distance that consumers have to travel to reach the collection points is the best
way to obtain greater involvement. The improvement of the quality and quantity of the collected
materials may compensate the additional collection costs of curbside collection. This is as aspect that
clearly deserves further investigation.
As this study demonstrates, the soft level also explains consumer involvement in the reverse
logistics system. Systems based on eco-points reduce the monetary separation costs because consumers
do not receive financial compensation for their sorting and discarding activities. Given that the current
system demands significant efforts on the part of consumers, this may reduce their willingness to
recycle. A more convenient alternative would be not to expect consumers to separate recyclables, that
is, to allow them to discard all the recyclable materials into a single recipient and assign the
responsibility of the separation process to Material Recovery Facilities. Although such a strategy can
reduce the system’s complexity from the consumer perspective this substantially increases separation
costs and, thus, the total cost of the system. A similar problem would incur if the improvement of
customer service implies a change to a commingled collection system (in contrast to multi-material
collection) or the implementation of a more frequent collection system. These strategies are likely to
improve the soft level of customer service in terms of reducing the perceived sorting complexity and
solving the (lack of) maintenance and design appeal of disposal recipients. The potential of these
strategies and their effects on separation and transportation costs is also a future research avenue.
An additional alternative and intermediate strategy that can improve the soft level of customer
service is to maintain the current multi-material collection based at the eco-points and promote
marketing campaigns to increase consumer awareness for greater involvement. Campaigns can also
demystify exaggerated negative expectations about the recycling system. Implementing curbside
collection, on a larger scale, is also a solution for reducing the sorting complexity, since information
exchange is possible on a one-to-one basis. However, and as referred, the implementation of such a
collection system requires a detailed cost-benefits analysis.
On the whole, this study clarifies the need to address all customer service elements. An important
limitation, however, is the fact that the sample is small and drawn from a single city and, as a
consequence, the generalization of the conclusions needs additional research. Furthermore, the
improvement of customer service brings challenges whose overcoming requires additional research.
Entities that manage the system must weight the need of increasing consumers’ involvement without
compromising the system’s economic viability. Therefore, an important challenge that such a reverse
logistics system needs to overcome is to find the best way to minimize the strategic costs of the
collection system without affecting consumer service performance. Another challenge that arises in the
reverse logistics systems for recycling is to guarantee that the market absorbs the recycled materials.
The quality of collected packaging residues affects the performance of the system because most soiled
materials either cannot be further recycled or may lead to increased reprocessing costs. In this sense,
additional support should be given to research that promotes the development of new products using
recycled materials and also marketing campaigns that aim to increase consumer awareness in using
such materials.
Int. Journal of Business Science and Applied Management / Business-and-Management.org
14
APPENDIX: SURVEY ON CONSUMERS’ MOTIVATIONS TOWARDS HOUSEHOLD
PACKAGING RECYCLING
I – CHARACTERIZATION OF THE RESPONDENT
Gender: Female Male
Age: …..… years old
Education level:
4 years 6 years 9 years 12 years
Technical –professional College or higher
Marital status:
Married Single Divorced Widow
Occupation:
Farmer/fisher Workman Services worker Public worker Teacher
Liberal worker Manager Housewife Student Retired Other
Residence type: House Farm Apartment
Home ownership: Own/Are buying Renting Familiar
Monthly family income:
Less than 324 € 324 a 499 € 500 a 999 € 1000 a 1999 €
2000 a 2999 € At least 3000 €
II – SATISFACTION WITH THE RECYCLING SYSTEM OF HOUSEHOLD WASTE
Indicate your satisfaction level with the following aspects concerning the selective-collection
system of recyclable materials. Use the following answer scale:
Very unsatisfied = 1 … Very satisfied = 5
Location of disposal containers
Frequency of waste collection
Distance to the disposal recipients
Cleaning and maintenance of disposal recipients
Local safety
Emptying regularity
Available information
Support and claim service
System adequacy to lifestyle
Number and type of suitable waste materials
III – INVOLVEMENT IN THE RECYCLING PROGRAM
Do you usually separate and selectively discard recyclable waste?
Yes No
With what frequency do you separate and selectively discard recyclable waste?
Never Sometimes Always
Thank you very much for your participation
Patricia Oom do Valle, Joao Menezes, Elizabeth Reis and Efigenio Rebelo
15
REFERENCES
Alvarez-Gil, M. A., Berrone, P., Husillos, F. J. Lado, N. (2007). Reverse logistics, stakeholders’
influence, organization slack, and managers’ posture. Journal of Business Research, 60, 463-473.
Austin J., Hatfield, D., Grindle, A., Bailey, J. (1993). Increasing recycling in office environments: the
effects of specific informative Cues. Journal of Environmental Systems, 6, 355-368.
Barros, A., Dekker, R., Scholten, V. (1998). A two-level network for recycling sand: a case study.
European Journal of Operational Research, 110(2), 199-215.
Bentler, P., Bonnet, D. (1980). Significance tests and goodness of fit in the analysis of covariances
structures. Psychological Bulletin, 88, 588-606.
Bentler, P. (1990). Comparative fit indexes in structural models. Psychological Bulletin, 107, 238-246.
Bollen, K. (1988). A new incremental fit index for general structural equation models. Paper Presented
at 1988 Southern Sociological Society Meetings. Nashville: Tennessee.
Bollen, K. (1989). Structural Equations with Latent Variables. John Wiley and Sons, Inc., New York.
Bronstad, G., Evans-Correia, K. (1992). Green purchasing: the purchasing agent’s role in corporate
recycling. Conference Proceedings of the National Association of Purchasing Management, 117-
121.
Burn, S., Oskamp, S. (1986). Increasing community recycling with persuasive communication and
public commitment. Journal of Applied Social Psychology, 16, 29-41.
Byrne, P, Deeb, A. (1993). Logistics must meet the “green” challenge. Transportation and Distribution,
34(2), 33.
Chang, N, Wei, Y. (2000). Sitting recycling drop-off stations in urban areas by algorithm-based fuzzy
multiobjective nonlinear integer programming modeling. Fuzzy Sets and Systems, 114, 133-149.
Dadzie, K., Chelariu, C., Winston, E. (2005). Customer service in the internet-enabled logistics supply
chain: website design antecedents and loyalty effects. Journal of Business Logistics, 26(1), 153-78
De Young, R. (1989). Exploring the differences between recyclers and non-Recyclers: the role of
information. Journal of Environmental Systems, 18, 341-351.
De Young, R. (1990). Recycling as appropriate behavior: a review of survey data from selected
recycling education programs in Michigan. Resources, Conservation and Recycling, 3, 253-266.
Diamond, W., Loewy, B. (1991). Effects of probabilistic rewards on attitudes and behaviors. Journal of
Applied Social Psychology, 21, 1590-1607.
Emerson, C., Grim, C. (1998). The relative importance of logistics and marketing customer service: a
strategic perspective. Journal of Business Logistics, 19(1), 17-32.
Faria de Almeida, A., Robertson, A. (1995). Domestic batteries and the environment: a life-cycle
approach to consumer electronic products. Proceedings International Conference Clean
Electronics Products and Technology, Edinburgh, Scotland, 162-167.
Fernández, J., Martinez, T. (2000). Análisis discriminante, in: Martinez, T. (Ed.), Técnicas de Análisis
de Datos en Investigación de Mercados. Ediciones Pirámide, Madrid, pp. 383-430.
Folz, D., Hazlett, J. (1991). Public participation and recycling performance: explaining program
success. Public Administration Review, 51(6), 526-231.
Folz, D. (1991). Recycling program design, management, and participation: a national survey of
municipal experience. Public Administration Review, 51(3), 222-231.
Foshay, M., Aitchison, A. (1991). Factors affecting yield and participation in curbside recycling
collection programs. Recourse Recycling March,118-129.
Fuller, D., Allen, J. (1997). A typology of reverse channels systems for post-consumer recyclables, in:
Polonsky J., Mintu-Winsatt A. (Eds.), Environmental Marketing: Strategies, Practice, Theory and
Research. Haworth Press, New York.
Fuller, D. (1978). Recycling consumer solid waste: a commentary on selected channels alternatives.
Journal of Business Research, 6(January), 17-31.
Gamba, R., Oskamp, S. (1994). Factors influencing community residents’ participation in commingled
curbside recycling programs. Environment and Behavior, 26, 587-612.
Giuntini, R., Andel, T. (1995). Advance with reverse logistics – Part 1. Transportation and Distribution,
36(2), 73-77.
Int. Journal of Business Science and Applied Management / Business-and-Management.org
16
Guiltinan, P., Nwokoye, N. (1975). Developing distribution channels and systems in the emerging
recycling industries. International Journal of Physical Distribution, 6(1), 28-38.
Gupta, T., Chakraborty, S. (1994). Looping in a multistage production system. International Journal of
Production Research, 22, 299-311.
Hair, J., Anderson, R., Tathan, R., Black, W. (1998). Multivariate Data Analysis with Readings. fifth
ed. Prentice-Hall, Upper Saddle River.
Hopper, J., Nielsen, J. (1991). Recycling as altruistic behavior: normative and behavioral strategies to
expand participation in a community recycling program. Environment and Behavior, 23(2), 195-
220.
Huberty, C. (1994). Applied Discriminant Analysis. John Wiley and Sons, Inc., New York.
Jacobs, H., Balley, J, Crews, J. (1984). Development and analysis of a community-based recourse
recovery program. Journal of Applied Behavior Analysis, 17, 127-145.
Jahre, M. (1995). Household waste collection as a reverse channel: a theoretical perspective.
International Journal of Physical Distribution and Logistics Management, 25(2), 39-55.
Johnson, P. (1998). Managing value in reverse logistics systems. Transportation Research, 34(3), 217-
227.
Joreskog, K., Sorbom, D. (1986). LISREL VII: Analysis of Linear Structural Relationship by
Maximum Likelihood and Least Square Method. Mooresville, Scientific Software.
Joreskog, K. (1969). A general approach to confirmatory maximum likelihood factor analysis.
Psychometrika, 34, 183-202.
Klecka ,W. (1980). Discriminant Analysis. Sage Publications, Inc., Newbury Park.
Kopicki, R., Berg, M., Legg, L., Dasappa, V., Maggioni, C. (1993). Reuse and Recycling – Reverse
Logistics Opportunities. Council of Logistics Management. Oak Brook, Illinois.
La Londe, Bernard, J., Zinszer, P. (1976). Customer Service: Meaning and Measurement. Oak Brook,
Council of Logistics Management.
Leroux, K. (2000). Tips for winning the recycling race. American City and County, May, 44-56.
Louwers, D., Kip, B., Peters, E., Souren, F., Flapper, S. (1999). A facility location allocation model for
reusing carpet materials. Computers and Industrial Engineering, 36, 855-869.
Ludwig, T. D., Gray, T. W., Rowell, A. (1997). Increasing recycling in academic buildings: a
systematic replication. Journal of Applied Behavior Analysis, 31(4), 683-686.
Luyben, P., Cummings, S. (1981-82). Motivating beverage container recycling in a college campus.
Journal of Environment Systems, 11, 235-245.
Maltz, A., Maltz, E. (1998). Customer service in the distributor channel: empirical findings. Journal of
Business Logistics, 19(2), 103-130.
Margai, F. (1997). Analyzing changes in waste reduction behavior in a low-income urban community
following a public outreach program. Environment and Behavior, 29(6), 769-792.
Marien, E. (1998). Reverse logistics as competitive strategy. Supply Chain Management Review,
14(1), 43-52.
Mentzer, J., Gomes, R., Krapfel, R. (1989). Physical distribution service: a fundamental marketing
concept! Journal of the Academy of Marketing Science, 17(1), 53-42.
Murphy, P. (1986). A preliminary study of transportation and warehousing aspects of reverse
distribution. Transportation Journal, 25(Summer), 12-21.
Nagel, C., Meyer, P. (1999). Caught between ecology and economic: end-of-life aspects of
environmental conscious manufacturing. Computers and Industrial Engineering, 36(4), 781-792.
O’Connor, J. (1993). Adapting recycling programs to meet new needs. Solid Waste and Power, 22-31.
Pelton, L., Strutton, D., Barnes, J., True, S. (1993). The relationship among referents, opportunity,
rewards and punishments in consumer attitudes towards recycling: a structural equations
approach. Journal of Macromarketing, 13(1), 60-77.
Pohlen, L., Farris, M. (1992). Reverse logistics in plastics recycling. International Journal of Physical
Distribution and Logistics Management, 22(7), 35-47.
Realff, M., Ammons, J., Newton, D. (2000). Strategic design of reverse production systems. Computers
and Chemical Engineering, 24, 991-996.
Reis, E. (1997). Estatística Multivariada Aplicada. Edições Sílabo, Lisbon.
Patricia Oom do Valle, Joao Menezes, Elizabeth Reis and Efigenio Rebelo
17
RevLog., 2002. [http://www.fbk.eur.nl/OZ/REVLOG/PROJECTS/TERMINOLOGY/Definitions.html].
Retrieved: January 2007.
Rogers, D., Tibben-Lembke, R. (2001). An examination of reverse logistics practices. Journal of
Business Logistics, 22(2), 129-148.
Rogers, D., Tibben-Lembke, R. (1999). Going Backwards: Reverse Logistics Trends and Practices.
Reverse Logistics Executive Council Press, Pittsburgh.
Saphores, J., Nixon, H., Oladele, A., Ogunseitan, A. (2006). Shapiro household willingness to recycle
electronic waste: an application to California. Environment and Behavior, 38, 183-208.
Scharma S. (1996). Applied Multivariate Analysis. New York, John Wiley and Sons, Inc.
Schumacker, R. E., Lomax, R. G. (1996). A Beginner’s Guide to Structural Equation Modeling.
Lawrence Erlbaum Associates, Inc., New Jersey.
Spengler, T,, Puchert, H,, Penkuhn, T., Rentz, O. (1997). Environmental integrated introduction and
recycling management. European Journal of Operational Research, 97(2), 308-326.
Steiger J. (1990). Structural model evaluation and modification: an interval estimation approach.
Multivariate Behaviour Research, 25, 173-180.
Stock, J., Lambert, D. (2001). Strategic Logistics Management. McGraw-Hill, New York.
Stock, R. (1992). Reverse Logistics. Council of Logistics Management. Oak Brook, Illinois.
Thogersen J. (1994). A model of recycling behavior, with evidence from Danish source separation
Programs. International Journal of Research in Marketing, 11, 145-163.
Tucker, L., Lewis, C. (1973). A reliability coefficient for maximum likelihood factor analysis.
Psychometrika. 38, 1-10.
Turner, G., Lemay, S., Mitchell, M. (1994). Solving the reverse logistics problem: applying the
symbiotic logistics concept. Journal of Marketing Theory and Practice, 2(2),15-27.
Valle, P., Rebelo, R., Reis, E., Menezes, J. (2005). Combining behavioral theories to predict recycling
involvement. Environment and Behavior, 37(3), 364-396.
Valle, P, Reis, E., Menezes, J., Rebelo, R. (2004). Behavioral determinants of household recycling
participation: the Portuguese case. Environment and Behavior, 36(4), 505-540.
Vining, J., Ebreo, A. (1992). Predicting recycling behavior from global and specific environmental
attitudes and changes in recycling opportunities. Journal of Applied Social Psychology, 22(20),
1580-1607.
Yender, G. (1998). Battery recycling technology and collection processes. Proceedings IEEE
International Symposium on Electronics and the Environment, May, 30-35.
Zikmund, W., Stanton ,W. (1971). Recycling solid wastes: a channels-of-distribution problem. Journal
of Marketing, 35(July), 34-39.