Download as pdf or txt
Download as pdf or txt
You are on page 1of 78

kosovo consumer buying behavior preferences and

demand for milk and dairy products

iliriana miftari
Master Thesis 30 credits 2009
Department of International Environment and Development Studies
Supervisor: Fred Håkon Johnsen
DECLARATION

I hereby declare that this thesis is accomplished with my own work and all sources of
literature that I have used are cited. I also assure that this work has not been presented to any
other university.

II
DEDICATION

This work is dedicated to my parents, and my two brothers, Artan and Arian, with much love
and thanks.

III
ACKNOWLEDGEMENT

My special gratitude goes to my supervisor Ass.Prof.Fred Håkon Johnsen for his wise advice
and great support in accomplishment of my master thesis. Great acknowledge goes to
Lånekassen, without its financial support my study program would be onerous manageable. I
wish to acknowledge Prof. Mujë Gjonbalaj for his continues support and encouragement
throughout my study. I also wish to thank my field assistant Rozafa Miftari for her sincerity
and tireless during field work. Great thanks go to those surveyed for their time and patience
during the interviews. My appreciation goes also to Ass.Prof.Mensur Vegara and Ass.Prof.
Hysen Bytyqi, who were the main initiator for my study program at University of Life
Sciences. I also wish to thank my colleagues at University of Prishtina, Prof. Halim Gjergjizi,
Prof. Mustafë Pllana, Prof. Shukri Fetahu, Prof.Skënder Muji, Jehona Shkodra. My special
gratitude go to my dear parents who continuously made sacrifice to keep me in school, thanks
a lot for all their love, support and encouragement that gave it to me. Special thanks go to my
best friend Panadda Larpkern, Besim Gojnovci, Ilbrahim Mehmeti, Arbina Kaja, and Paul
Okullo.

IV
Abstract
The aim of this study was to analyse consumer buying behaviour, preferences, attitudes,
needs and wants toward dairy products. The study was carried out in five Kosovo regions
(Prishtina, Prizren, Mitrovica, Peja, and Gjilan). The stratified random sampling has been
applied in selecting people who would be included in the sample. The data used in this study
were obtained through direct interviews with Kosovo consumers aged 18 years and older.
Two associated statistical techniques, multiple linear function and binary logistic function
have been used to explain the relationships between the quantity and expenditures on milk
and other dairy products with demographic and socioeconomic household characteristics. The
results show that milk, yoghurt, cream, curd, and cheese were the main dairy products
consumed by majority of Kosovo consumers. Supermarkets and grocery stores were the most
preferred market places by the consumers when buying milk and other dairy products. The
consumer’s preference towards market place was significantly dependent on demographic and
socioeconomic household characteristics Dairy products with shorter shelf life were bought
more frequently by the consumers compared to those with longer shelf life. Apart from
product life, other demographic and socioeconomic factors had significantly impact on the
frequencies of buying milk and other dairy products. The consumers’ demand for milk and
other dairy products was quite stable throughout the year. The consumers’ attitudes toward
product features such as nutritive content, taste, product safety, price, brand, wrapping,
package size, and the origin of product were significantly dependent on demographic and
socioeconomic factors. Kosovo consumers preferred bigger packages for curd and cheese.
Smaller packages were predominantly more preferred for fruit yoghurt and butter. Majority of
the Kosovo consumers had favourable bias towards domestic versus foreign dairy products.
Product attributes such as quality, safety, taste, and price were the main features motivating
Kosovo consumers to purchase domestic dairy products. Television and newspapers were the
media most often used by the consumers to get information about the dairy products. The
preference towards new dairy products and innovation was dependent on respondent’s
characteristics. The Household’s characteristics such as income, size, employment, the
number of children, respondent’s age, and education were significant in explaining variation
in quantity consumed and the expenditures on milk and other dairy products.

V
TABLE OF CONTENTS

DECLARATION………………………………………………………………...... II
DEDICATION…………………………………………………………………….. III
ACKNOWLEDGEMENT………………………………………………………… IV
ABSTRACT………………………………………………………………………. V
TABLE OF CONTENTS…………………………………………………………. VI
LIST OF TABLES………………………………………………………………... VIII
LIST OF FIGURES……………………………………………………………….. X
LIST OF ACRONYMS…………………………………………………………… XI
CHAPTER I……………………………………………………………………… 1
1. INTRODUCTION……………………………………………………………… 1
1.1 Background……………………………………………………………………. 1
1.2 Overview of the Kosovo Dairy Sector………………………………………... 2
1.3 Agriculture Sector Strategy and Policies……………………………………… 5
1.4 Problem Statement…………………………………………………………….. 8
1.5 Justification……………………………………………………………………. 9
1.6 Objectives and Research questions……………………………………………. 9
CHAPTER II…………………………………………………………………….. 11
2. THEORETICAL BACKGROUND……………………………………………. 11
2.1 Factor Influences Consumer Behaviour………………………………………. 11
2.2 The Stimulus model of the Consumer Behaviour…………………………….. 12
CHAPTER III……………………………………………………………………. 15
3. METHODOLOGY……………………………………………………………... 15
3.1 The Study Area………………………………………………………………... 15
3.2 Sample Selection……………………………………………………………… 16
3.3 Data Collection………………………………………………………………... 18
3.4 Model Specification…………………………………………………………… 19
3.4.1 Variable Description and Measurement…………………………………….. 21
3.4.2 Estimation procedure………………………………………………………... 23
CHAPTER IV……………………………………………………………………. 25
4. Results and Discussion…………………………………………………………. 25
4.1 Descriptive Statistics on the Household Characteristics……………………… 25

VI
4.2 Descriptive Statistics on Consumer buying Behaviour……………………….. 29
4.3 Model Estimation……………………………………………………………... 40
4.3.1 Binary Logistic linear Estimated Parameters……………………………….. 40
4.3.2 Multiple linear Estimated Parameters……………………………………….. 43
CHAPTER V……………………………………………………………………... 53
5. Conclusion……………………………………………………………………… 53
LIST OF REFERENCES………………………………………………………... 59
APPENDIX……………………………………………………………………….. 62
Appendix A1/Questionnaire………………………………………………………. 62
Appendix A2/Tables………………………………………………………………. 66

VII
LIST OF TABLES

Table 1. The estimated livestock situation in 2005……………………………... 3


Table 2. Dairy farm structure in Kosovo……………………………………….. 3
Table 3. Evolution of the local processors on daily milk processing…………… 4
Table 4. Dependent variables used in the estimation of the average quantity of 21
the product i purchased in litre or kilogram in the HHi…………………………
Table 5. Dependent variables used in the estimation of the average monthly 21
expenditures on the product i in euro per month in the HHi…………………….
Table 6. Independent variables used in the estimation of the quantity and the 22
expenditures on milk and six other dairy products……………………………...
Table 7. Dependent variables used in the estimation of the fitted model in the 22
equation 1………………………………………………………………………..
Table 8. Dependent variables used in the estimation of the fitted model in the 22
equation 1………………………………………………………………………..
Table 9. Independent variables (factors) used in the estimation of the fitted 23
model in the equation 1………………………………………………………….
Table 10. Recoded variables used in the test of independence…………………. 23
Table 11. Summary statistics of the household characteristics…………………. 25
Table 12. Pairwise comparison between females and males…………………… 27
Table 13. Pairwise comparison between rural and urban household…………… 27
Table 14. Pairwise comparison between regions……………………………….. 28
Table 15. Summary statistics of the average quantity of milk and other dairy 33
products purchased by the HHs………………………………………………….
Table 16 The annual average per capita consumption………………………….. 33
Table 17. Summary statistics of the average expenditures on milk and other 34
dairy products by the HHs………………………………………………………
Table 18. Relationship of whether the respondent i buys milk and the 40
predictors included in the equation 1……………………………………………
Table 19. Relationship of whether the respondent i buys yoghurt and the 40
predictors included in the equation 1……………………………………………
Table 20. Relationship of whether the respondent i buys fruit yoghurt and the 41
predictors included in the equation 1……………………………………………
Table 21. Relationship of whether the respondent i buys cream and the 41

VIII
predictors included in the equation 1……………………………………………
Table 22. Relationship of whether the respondent i buys curd and the 41
predictors included in the equation 1……………………………………………
Table 23. Relationship of whether the respondent i buys cheese and the 41
predictors included in the equation 1……………………………………………
Table 24. Relationship of whether the respondent i buys butter and the 42
predictors included in the equation 1……………………………………………
Table 25. Relationship of whether the respondent i fulfils the demand for milk 42
and dairy products and the predictors included in the equation 1………………
Table 26. Relationship of whether the respondent i prefers domestic dairy 42
products and the predictors included in the equation 1………………………….
Table 27. Relationship of whether the respondent i prefers foreign dairy 42
products and the predictors included in the equation 1………………………….
Table 28. Relationship of whether respondent i prefers new dairy products and 43
the predictors included in the equation 1………………………………………..
Table 29. Relationship of whether respondent i started buying new dairy 43
products last year and the predictors included in the equation 1………………
Table 30. Test of independence between the market places and demographic 56
and socioeconomic characteristics………………………………………………
Table 31. Test of independence between the frequencies of buying milk, dairy 57
products and demographic and socioeconomic characteristics………………….
Table 32. Test of independence between evaluation of product attributes and 58
demographic and socioeconomic characteristics………………………………..

IX
LIST OF FIGURES

Figure 1. Stimulus-response model of the consumer behaviour………………... 13


Figure 2. Stage model of the consumer buying process………………………... 14
Figure 3.Stratification of the sample size……………………………………….. 17
Figure 4. Respondents’ answers in terms of buying milk and six other dairy 31
products………………………………………………………………………….
Figure 5.Market places preferred by consumers when buying milk and other 32
dairy produces…………………………………………………………………...
Figure 6. Frequencies of buying milk and other dairy products………………... 32
Figure 7. Seasonal consumption patterns on milk……………………………… 35
Figure 8. Seasonal consumption patterns on yoghurt…………………………... 35
Figure 9. Seasonal consumption patterns on fruit yoghurt……………………... 36
Figure 10. Seasonal consumption patterns on cream…………………………… 36
Figure 11. Seasonal consumption patterns on curd……………………………... 36
Figure 12. Seasonal consumption patterns on cheese…………………………... 37
Figure 13. Seasonal consumption patterns on butter…………………………… 37
Figure 14. The evaluation of product features in order of importance…………. 38
Figure 15. The level of the mass media used by the respondents………………. 39

X
LIST OF ACRONYMS

ANOVA Analysis of Variance


ARDP Agricultural Rural Development Plan
CEFTA Central European Free Trade Agreement
EU European Union
GDP Gross Domestic Product
GMP Good Manufacturing Practices
HACCP Hazard Analysis Critical Control Points
HH Household
KARDP Kosovo Agricultural Rural Development Plan
KCBS Kosovo Cluster Business Support
KDPA Kosovo Dairy Processing Association
MAFRD Ministry of Agriculture Forestry and Rural Development
MPS Ministry of Public Services
SOK Statistical Office of Kosovo
SWOT Strengths Weaknesses Opportunities and Threats
UHT Ultra High Temperature
UN United Nations
UNMIK United Nations Mission in Kosovo
VAT Value Added Tax

XI
CHAPTER I

1. INTRODUCTION

1.1 Background

After the end of the last war in 1999, the United Nations (UN) Security Council
Resolution 1244 (UNSCR1244), placed Kosovo under temporary UN Administration
(UNMIK – United Nations Mission in Kosovo). Building self-governed democratic
institutions, economic re-activation, an effort to build the peace, inter-ethnic reconciliation
and building new bridges of cooperation with the countries in the region were the main
challenges accompanying Kosovo after the last conflict. The decline of the agriculture sector
which started during the 1990s resulted from the emigration of the Kosovo population (18-
20%) to the Western European Countries. While overall stagnation culminated in 1999, most
of the farms stopped production, having consequences to the food processing industry. The
agriculture sector was heavily harmed by the last conflict. The total amount of the destruction
and deprivation in rural areas was estimated to be 737 million dollars (MAFRD, 2003).
Despite the difficult circumstances after the last conflict, Kosovo has gone ahead in the
building of its democratic institutions and macroeconomic stabilization. ‘’It experienced a
strong recovery mostly based on an immense inflow of foreign assistance and external private
inflows estimated to equal up to 50 percent of GDP’’ (Fock, 2007, p.4). The estimated
average real GDP growth between 2002 and 2007 has been slow at less than one and half
percent. ‘’This slow expansion was mainly due to a combination of low investment and the
ongoing withdrawal of the international community in Kosovo’’ (World Bank, 2007, pp.1-2).
The poverty phenomenon has remained persistent and is widespread within the Kosovo
society. According to the World Bank (2007) estimation, about 45% of the Kosovo
population is considered to be poor, while a smaller fraction (18%) was in extreme poverty. In
2005 Kosovo had the highest poverty rate of the countries in the Western Balkans. This large
fraction of the Kosovo population just around the poverty line reflects the socio-economic
phenomenon where sources available to society are used to satisfy the wants of small fraction
while many have not even met their basic needs (Chambers, 1983).

-1-
Agriculture has historically been an important sector in Kosovo’s economy. It is
characterized by small subsistence farms, high input costs, low productivity, poor
infrastructure and poor advisory services (Ministry of Agriculture Forestry and Rural
Development, MAFRD, 2006). The sector’s contribution to GDP during the period 2000-
2005 was 25-27%. The high contribution of the agriculture sector to GDP was not due to high
productivity, but rather to the declining contribution of the other sectors in GDP. Given that
majority (55%) of the Kosovo population is concentrated in rural areas, agriculture has
remained an important sector in mitigating rural poverty. During the period 2000-2005, its
contribution to the employment rate was between 25-30% (informal self-employed agriculture
workers). The proportion of the Kosovo labour force engaged in agriculture sector/farming
was the highest in the Western Balkans countries and it is around five times higher than in
European Union (EU) countries. This high proportion of the labour force involved in
agriculture clearly indicates the low efficiency of this sector. Despite its low efficiency and
loss of traditional export markets, its contribution to the value of total exports in 2005 was
16%. Agriculture remains an important sector and an engine towards economic growth and
EU accession.

1.2 Overview of the Kosovo Dairy Sector

The Kosovo dairy sector is one of the most promising sectors and has consistently
performed well since the end of the last war. As with most other Kosovo sectors the last
conflict in 1999 caused considerable damage to the agricultural sector. Particular damages
affected the dairy sector, where approximately 50% of the livestock was killed and roughly
40% of the livestock infrastructure (stalls) was destroyed (MAFRD, 2003). Due to the
creation of this situation there has been a market shortage in animal and dairy commodities.
Consequently, there has been a sharp increase of the imported animal and dairy products for
market equilibrium to be established as well as to meet the market demand. Moreover, many
donors helped in the restocking of the cattle herd. Since the end of the war ‘’up to the first
quarter of 2003, 10,000 pregnant heifers were imported and distributed to farmers’’ (Rural
Development Plan 2007-2013, 2006, p.19). The livestock census conducted by Statistical
Office of Kosovo (SOK) in November 2003 was used as the determinative base of the animal
numbers in Kosovo. Even though efforts were made to improve the livestock situation

-2-
through import and donor support, livestock in Kosovo has still remained at a very low level.
Table represents the livestock census conducted by SOK in 2005.

Table 1. The estimated livestock situation in 2005


Livestock Cattle Milk cows Buffaloes Sheep and goats
Number 351,827 186,707 622 151,880

Source: Agriculture Household Survey, 2006, SOK.

It was estimated that 67.7% of rural households have dairy cows, with an average 1.2 dairy
cows per household (HH) (Rural Development Plan 2007-2013, 2006). According to the
MAFRD classification, the most common type of the dairy farms in Kosovo is the
traditional/subsistence farm (94%), followed by the semi-commercial farm (5%) and the
commercial farm (1%).

Table 2. Dairy farm structure in Kosovo


Type of farm No. of farms Percentage No. of cows
Traditional/Subsistence 78,124 94% <5
Semi Commercial 4,378 5% 5–9
Commercial 787 1% > 10
Total 83,289 100%
Source: ASPAUK Project

Kosovo dairy sector is mainly dominated by milk, yoghurt and white cheese. Milk
production in Kosovo is derived only from the private sector, since the public sector collapsed
during the transition period between 1990 and 1999. According to the official statistical data,
milk production in Kosovo was estimated to be approximately 257,500 tons/year. Local milk
production is mainly intended for the domestic market. Of the total domestic milk produced,
45.6% is consumed on the farm, 41.4% is sold on the local green market1 mostly as raw milk
or white cheese, while only 13% is sold on the processing market. There are 19 dairy
processors where commercial and semi-commercial farms have the possibility to sell the
produced milk (Nushi and Selimi, 2009). The most important elements affecting marketability
of the domestic milk produced to the processing market were milk collection cost, low
technology and low milk quality. As is shown in table 2 most of the Kosovo dairy farms are
small traditional/subsistence farms dispersed across the countryside, with low capacity for
1
Green market - signifies local markets where un-pasteurised milk and cheese is sold directly to consumers by
farmers.

-3-
milk production and low milk quality2, which makes the milk collection procedure more
expensive.
‘’Considering that milk quality has a big impact on the final product and shelf-life’’, two
technical assistant projects worked in the improvement of the milk quality and ability of dairy
processors to test quality when buying milk (Oldham, Bajraktari and Wittkowsky, 2006).
However, Kosovo milk quality at the farm level as well as dairy products has still remained at
a poorer quality compared to the imported articles. Large market share and successful
penetration in the market of some imported dairy products above Kosovo prices proved that
consumers are willing to pay high prices for better quality. However, there is still not enough
good milk quality coming to the processing plants.
During the period of 2003 through 2007 there was a remarkable investment by local
investors in milk processing capacity. In 2003, the capacity of daily milk intake for processing
was 30,800 litres, which increased to 114,000 litres in 2007. Yearly milk processing capacity
was estimated to be 16,000 tons, but due to the evident impediments highlighted above, only
12,300 tons are processed in the dairy plants (Rural Development Plan 2007-2013, 2006).
Seasonal difference in the quantity of raw milk production is an additional obstacle for the
local dairy processors, which renders them unable to steadily and fully utilize their capacities.
Fresh milk, yoghurt, kos3, and white cheese are the main items produced by the local dairy
processors.

Table 3. Evolution of the local processors on daily milk processing


Year 2003 2004 2005 2006 2007
Milk in litres 30,800 71,000 74,800 92,200 114,000
Source: Kosovo Dairy Processing Association - KDPA

In Kosovo, dairy commodities are dominant in the daily diet and the main source of
protein. A study conducted by Kosovo Cluster Business Support (KCBS, 2008) on dairy
market assessment, proclaimed that among dairy commodities, ultra high temperature
processed (UHT) milk, white cheese, sour cream, fresh milk and yoghurt, were most
frequently dairy commodities purchased by Kosovo consumers. This study revealed that
89.7% of the Kosovo households occasionally buy milk. It was estimated that milk
consumption per capita was approximately 170 litres per year. There are no estimations
available with regard to the quantity consumed of other dairy commodities.

2
Low milk quality in terms of a high bacterial count.
3
Kos - milk drink between yoghurt and sour crème.

-4-
Kosovo is considered to be a huge importer of dairy products. Of the total domestic milk
demand, 76% was made up of Kosovo farms supply, while 24% was supplemented by import.
Among the imported dairy commodities, UHT milk accounts for the highest share of the total
value of imports (about 40%) followed by cream and fruit yoghurt. Production of milk and
other dairy products in Kosovo is relatively costly as most of the livestock is reared in a stall-
feed system rather than pasture fed. Moreover, quality at farm and factory level is still
considered to be at a lower level. This is why imported UHT milk and other dairy products
and compete successfully on the local market. Export of Kosovo dairy products is limited to a
few items, mostly UHT milk, cream and cheese. Export increases during the summer season,
while the Albanian market is the main absorber of Kosovo’s dairy commodities.

1.3 Agriculture Sector Strategy and Policies

Administratively, the agriculture sector and policy is set at national level, while the
execution of the policy is accomplished at national and municipality levels. The Ministry of
Agriculture, Forestry and Rural Development (MAFRD) in partnership with other
stakeholders such as the Farmers’ Associations, the Rural Advisory Service, and the
Veterinary and Food Agency play major roles in policy formulation and implementation. In
2006 Kosovo launched its Agriculture and Rural Development Plan 2007-2013 (ARDP)
(MAFRD, 2006) which outlines objectives and key measures for the agriculture sector. The
aim of ARDP was to provide a framework to guide future agriculture sector policies. Its
vision for Kosovo agricultural and rural development during 2007-2013 was to ‘’make a
balanced contribution to the economic, environmental, social and cultural well-being of rural
areas, and Kosovo as a whole, through effective and profitable partnerships between the
private sector, central/local government and local communities within the European context’’.
Based on the Kosovo’s agri-rural situation and after SWOT (strengths, weaknesses,
opportunities, and threats) analysis, general objectives for agricultural rural development in
Kosovo were set as below:

(i) Additional income for farmers and rural population on purpose to improve living
standards and working conditions in rural areas;
(ii) Improved competitiveness and efficiency of primary agricultural production in
order to achieve import substitution and take advantage of export markets;

-5-
(iii) Improved processing and marketing of agricultural produce, through increased
efficiency and competitiveness;
(iv) Improved on-farm / in-factory quality and hygiene standards;
(v) Sustainable rural development and improved quality of life (including
infrastructure) through promotion of farming and other environmentally
sustainable economic activities;
(vi) Creation of employment opportunities in rural areas, particularly through
diversification;
(vii) Alignment of Kosovo’s agriculture with that of the EU.

Two key approaches for achieving KARDP 2007-2013 objectives were set as below:
I. Undertake actions that overcome the bottlenecks that are holding back
sustainable rural development in Kosovo, and
II. Start aligning Kosovo’s rural sector with four axes of current EU rural
development strategy.

The achievement of the KARDP 2007-2003 was built on the following four axes with the
eight key measures:

Axis 1: Competitiveness

 Development of vocational training to meet rural needs (Measure 1)

 Restructuring physical potential in the agri-rural sector (Measure 2)

 Managing water resources for agriculture (Measure 3)

 Improving the processing and marketing of agricultural products (Measure 4)

Axis 2: Environment and improved land use

 Improving natural resource management (Measure 5)

Axis 3: Rural diversification and quality of rural life

 Farm diversification and alternative activities in rural areas (Measure 6)

 Improvement of rural infrastructure and maintenance of rural heritage (Measure 7)


Axis 4: Community-based local development strategies
 Support for local community development strategies (Measure 8)

-6-
During the period 1999-2004, fiscal rates were was 10% for the standard import duty and
15% Value Added Tax (VAT) for agriculture inputs. Presently, the agriculture sector in
Kosovo does not benefit from direct subsidies by the government. However, indirect subsidies
are given to the sector in terms of zero VAT and zero import duty on agricultural inputs,
reproductive material in animal production, mechanization and on most capital goods used in
agriculture or agro-processing. Kosovo applies 10 percent customs duty on imports, while
there is no encouraging instrument (export subsidy) that induces Kosovo’s export. In 2006,
Kosovo joined the Central European Free Trade Agreement (CEFTA). The market created by
CEFTA comprises nearly 30 million consumers and it was signed by ten countries4. This
trade agreement resulted in market access at zero import duty for all participating countries.
Even though this trade agreement offered Kosovo the opportunity to be part of regional
integration and EU association, there was also an increased pressure in terms of
competitiveness on the local market.
The expectations of ARDP for Kosovo’s dairy sector were to have a modern dairy
industry, by increasing the number of commercial farms with an average of 30 cows, an
achievement of 100% ‘’A’’ grade milk quality, increasing the average of milk productivity up
to 20/litres/cow/day, and increasing milk consumption from 170 to 190 litres/capita/year by
the end of 2013. Due to an increase of milk consumption per capita, demand for milk at the
end of 2013 was estimated at 28% higher than in 2005.
The following key actions were taken by government in order to achieve its objectives
towards dairy sector: ‘’a 10-yearly strategy for livestock development, approval of Law on
Livestock, U/A (Administrative Order) in proceeding, changes in fiscal policy, free trade
agreements, flattening the customs entering price for imported products, improving animal
nutrition, improving the breeding structure, increasing livestock production, lowering imports,
improving quality production and place exemption of customs tariffs and VAT for the
livestock inputs’’(Nushi and Selimi, 2009, p.12). Actually, although several actions were
taken by the government, Kosovo’s dairy sector is still facing significant difficulties with
regard to ‘’land availability, breed quality, high cost of milk production, transport and
distribution, lack of knowledge on new technology and marketing strategies, unfair
competition from import, lack of experience in processing and low marketing capacities due
to small dairies’’ (Nushi and Selimi, 2009, p.11).

4
Member of CEFTA 2006: Kosovo, Albania, Bosnia and Herzegovina, Croatia, Macedonia, Montenegro,
Moldova, Serbia. Except, Bulgaria and Romania (from January 1st 2007 are member of EU-27).

-7-
Increasing farm size through land consolidation, a developing land market, decreasing
interest rates for agricultural credits, implementing an integrated quality management system
through Good Manufacturing Practices (GMP) and Hazard Analysis Critical Control Points
(HACCP), were seen as key actions to be taken in order to improve the competitiveness of the
Kosovo dairy sector on the local and regional markets.

1.4 Problem Statement

As highlighted in the above section, the dairy sector in Kosovo has performed well and
seems to be one of the most promising agriculture sectors. However, the sector has passed
through a very onerous period during the last decade. Significant economic changes after the
last conflict had dramatically impacted consumers’ buying behaviour, their attitudes, needs
and demand for dairy products. Moreover, there was an enormous increase of foreign
competitors on the domestic dairy market. Therefore, new market approaches such as
improved product quality, wider assortment of the dairy produce, new management on the
sales system, an improved marketing information system, market segmentation, and price
differentiation, were required for the local dairy processors to become a competitive on local
and regional markets.
Previously, the market orientation of Kosovo’s dairy industry was to produce cheap bulk
dairy commodities with a limited assortment. Nowadays marketing has broadened its concept,
contemporary marketers include the study of transfer behaviour as well as transaction of
consumer behaviour (Kotler, 2002). Having information on consumers’ buying behaviour,
their preferences, attitudes, needs and demand for dairy products are considered to be key
determinant factors for the efficiency of dairy. However, the dairy industry in Kosovo lacks
information on consumers’ buying behaviour and demand for dairy products level. Thus,
among other significant problems that the dairy sector is facing, lack of information is an
additional disadvantage upon the sales opportunities and the efficiency of dairies in Kosovo.

-8-
1.5 Justification

The consumers represent the last component within the food chain supply but they are set
as a major player on the market, thus they deserve special attention (Kapsdorferová and
Nagyová, 2005). Significant economic changes and continued evolution of consumers’
preferences, needs and wants for the dairy produce, makes the dairy market become more and
more segmented as well as more complex. Therefore, a study of consumers’ buying
behaviour, their preferences, attitudes, needs and wants towards dairy products helps the dairy
industry (local processors) in their plan projections and development of their marketing
strategies. An understanding of the consumers’ buying behaviour and identifying the major
forces influencing structural changes in their consumption patterns, helps businesses of this
industry to improve and get the most rational way to meet the consumers’ needs.
Currently, there is no data available or comprehensive study devoted to the dairy
consumption patterns in Kosovo (Nushi and Selimi, 2009). There is no study on how
consumers’ buying behaviour reacts to demographic and socio-economic factors. Moreover,
no previous studies addressed the estimates of demand and household budget spent on dairy
products with respect to demographic and socio-economic factors. It is thus very important to
have good estimates of how the demand and household budget spent on dairy products reacts
to demographic and socio-economic changes. Furthermore, analyses of changes in
consumption patterns and consumption trends due to the changes of demographic and socio-
economic factors (particularly income changes) are very important and applicable for policy
modelling purposes. Having an accurate analysis and good estimates of demand for dairy
products helps projection of the future development of the dairy sector in Kosovo.

1.6 Objectives and Research questions

The overall objective of this study was to assess the evolution of Kosovo consumers with
regards to their new consumption patterns for dairy products. The study was carried out to
determine factors that influence consumers’ buying behaviour, their preferences, attitudes,
needs and wants towards dairy products. The following research questions were drawn in
pursuit of the outlined objective:
1. ‘’What, Who, Where, How and When” do consumers buy milk and other dairy
products such as yoghurt, fruit yoghurt, cream, curd, cheese and butter?

-9-
2. What are the consumers’ preferences for the package size, country of origin, new dairy
products and innovation?
3. What are the consumers’ attitudes toward product attributes such as brand, packing,
nutritive content, economic value, taste, safety, and country of origin?
4. Do the consumers meet their needs for milk and six other dairy products with their
current monthly income?
The aim of this study was also to estimate how the demand and household budget spent on
milk and six other dairy products react to demographic and socio-economic changes. The
following research questions were addressed in pursuit of this objective.
1. What is the average monthly quantity purchased of milk and other dairy products?
2. What is the average monthly expenditure for milk and other dairy products?

- 10 -
CHAPTER II

2. THEORETICAL BACKGROUND

2.1 Factors Influences Consumer Behaviour

Cultural, social and personal factors are always considered to be the major forces
influencing consumers’ buying behaviour (Kotler and Keller, 2006). An understanding of
such factors helps businesses at tailoring products that meet consumers’ needs and wants.
Among important influences on consumer buying behaviour (culture, subculture, and social
class) ‘’culture is the fundamental determinant of a person’s wants and behaviour’’ (Kotler
and Keller, 2006, p.174). Given that ‘’all human societies exhibit social stratification’’,
distinctive consumer buying behaviour and preferences exist as well among social classes
(Kotler and Keller, 2006, p.175). These distinctive buying patterns and preferences among
social classes are significantly determined by occupation, income, wealth, education etc.
Consumers can shift up and down the social strata but the extent of this mobility depends
much on how rigid the social stratification is in a given society (Kotler and Keller, 2006).
An individual’s values, perceptions, preferences and behaviours in a direct or indirect way
are significantly influenced by reference groups (family, friends, neighbours, and co-workers)
to which they belong. Such reference groups expose an individual to new behaviours and
lifestyles, and often have an effect on their attitudes, products or brand choices (Kotler and
Keller, 2006). Family is considered to be one of the most important consumer buying groups
in the society (Tour and Henthorne, 1995). It is thus very important for the marketers to
identify the roles and influence of the family members in the purchase of products. Marketers
may accurately address their marketing messages, only if they fully understand who of the
family members is a leader and has direct influence on the buying decisions. ‘’In countries
where parents live with grown up children, their influence can be substantial’’ (Kotler and
Keller, 2006, p.177). The buyer’s characteristics such as age, stage in the life cycle,
occupation, economic circumstances, personality, self-concept, lifestyle and values have a
significant impact on the consumer behaviours and the buying decisions. The consumption
patterns and taste in food are often shaped by the family life cycle and the number, age, and
gender of people in the household, and occupation (Kotler and Keller, 2006). The consumer’s
decisions in product and brand choices are greatly influenced by the income level, stability,

- 11 -
personality, self-concept, core values and life style. The consumers’ characteristics, buying
behaviour, and preferences vary over the consumers’ lifetime. This is why successful
marketers make an attempt to follow carefully trends and changes of the consumers’
characteristics.

2.2 The Stimulus model of the Consumer Behaviour

Many theorists have addressed the issue of understanding the consumer behaviour in the
decision buying processes and the purchase decisions. The marketing and environmental
stimuli are key enter points of understanding the consumer buying behaviours. The consumer
psychology combined with the consumer characteristics result in the buying decision process
and the purchase decision. The crucial point of consumer behaviour is to understand what
happens in the consumer’s consciousness between the arrival of outside stimuli (marketing
and environmental) and the ultimate purchase decision (Kotler and Keller, 2006). The
consumer responses to the outside stimuli are mainly influenced by the psychological
processes such as motivation, perception, learning, and memory, and the consumer
characteristics such as cultural, social, and personal.
An individual has different needs, biogenic needs that arise from physiological tension and
psychogenic needs that arise from psychological tension. ‘’A need becomes a motive when it
is aroused to a sufficient level of intensity’’ (Kotler and Keller, 2006, p.184).
The best-known theories of human motivation such as Sigmund Freud, Abraham Maslow, and
Frederic Hezberg, are often used by the marketers for consumer analysis, understanding their
behaviours and develop marketing strategies. Freud’s theory supposes that an individual may
not completely understand his/her motivations, as psychological processes shaping his/her
behaviour are mostly unconscious (Kotler and Keller, 2006). Maslow’s theory explains why
people are driven by particular needs at particular times (Maslow, 1954). Based on the
Maslow’s theory a person needs are in order of importance, starting from physiological,
safety, social, esteem, and self actualization needs. This theory was used by the marketers in
understanding of how products fit into the consumer’s plan, goal and life (Kotler and Keller,
2006). Herzberg’s theory distinguishes two-factor theory dissatisfiers and satisfiers factors
(Herzberg, 1966). According to Herzberg’s theory, marketer’s task is first to avoid
dissatisfiers factors and second to identify satisfiers or motivators of purchase in the market
and then supply them (Kotler and Keller, 2006).

- 12 -
Figure 1. Stimulus-response model of the consumer behaviour

Consumer
Psychology
Motivation
Perception
Learning Buying Decision Purchase
Memory
Marketing Other Process Decision
Stimuli Stimuli Problem recognition Product choice
Information search Brand choice
Products Economic
Evaluation of Dealer choice
Services Technological
alternatives Purchase amount
Price Political
Purchase decision Purchase timing
Distribution Cultural
Post-purchase Payment method
Communications
behaviour

Consumer
Characteristics
Cultural
Social
Personal

Source: Kotler and Keller (2006).

The information about the product is one of the prime determinants of consumer decision
making. Other psychological processes such as motivation, perception, learning, and memory
have a significant impact in the consumer buying decisions. Key consumer behaviour
questions such as “who, what, when, where, how, and why” are used by marketers to
understand reality and every facet of the consumer behaviour (Kotler and Keller, 2006). The
starting point of the buying process is problem recognition. The consumer starts buying, when
he or she recognizes a problem or need (Kotler and Keller, 2006). There are mainly five
stages that the consumer passes through during the buying decision process: problem
recognition, information search, evaluation of alternatives, purchase decision, and post
purchase behaviour. The consumer does not necessarily pass through all stages when he or
she buys the product. He or she may skip some stages if he or she already is familiar with the
product (Kotler and Keller, 2006). Figure 2 represents the five stages of the buying decision
process.

- 13 -
Figure 2. Stage model of the consumer buying process
Problem Information Evaluation of Purchase Postpurchase
recognition search alternatives decision behaviour

Source: Kotler and Keller (2006).

- 14 -
CHAPTER III

3. METHODOLOGY

3.1 The Study Area

Kosovo is situated in South-East Europe surrounded by Albania, Macedonia, Serbia and


Montenegro, with a total land area of 10.908 km2 and population density 193 people per km2.
Kosovo is divided into 5 regions and 30 municipalities with 1466 settlements (Law on
Territorial Structure, 2004). Taking into account that the system of civil registration was not
up to date and there has not been any census taken since 1981, it is difficult to provide reliable
figures for the population as a whole (Ministry of Public Services-MPS, 2007). Referring to
the latest registration conducted by Statistical Office of Kosovo (SOK) in 1981, the total
number of inhabitants was estimated to be 1,959,714. In terms of population size, Prishtina is
the biggest region with 25.5% of the total population, followed by Peja with 21%, Prizren
with 19.6%, Gjilan with 19.6% and Mitrovica with 14.6%. Traditionally more of the
population of Kosovo lives in the rural areas than in urban areas, with proportion 55:45%.
Kosovo is considered to have the youngest population in Europe, where 33% of the total
population is less than 14 years old, 61% is between 15-64 years old, while only 6% of the
population is over 65 years old. Ministry of Public Services (MPS, 2009a) states that for
every 100 female births there are around 109 male births.
Statistics of Living Standard in 2007 indicated that households living in rural areas as well
those less educated had lower general food consumption than household living in urban areas.
In 2008 food consumption per households living in urban areas was 2.144 euro or 35% of the
total consumption, compared 2384 euro or 44% for those living in rural areas (MPS, 2009b).
Statistical data on the level of education in 2007 has shown that more than 40% of the males
and about 60% of the females had not completed upper secondary education. The
unemployment rate in 2007 for the labour force in the 15-64 age groups was denoted 33.7%
for males and up to 49.2 % for the females.

- 15 -
3.2 Sample Selection

This section emphasizes the sampling principles and methods involved in selecting people
(respondents) who would be asked questions by questionnaire. Before getting to the
fieldwork, matters that concerned this survey were the definition of the population (N),
sample size (n), type of sampling and formulation of the research instrument that suits the
investigation of the topic. The intention of this study was to gain an understanding of
consumers’ behaviours, their attitudes, preferences and demands towards milk and dairy
products. The total number of inhabitants in Kosovo was the targeted population from which
we were interested to draw a sample.
One of the most relevant and basic considerations in conducting a survey research by
interview is the money and time consumed. It was impractical to examine each and every unit
of the population, thus, sampling was needed. Even though a sampling approach was less
costly in terms of money and time, other considerations came to the fore as well. Since the
aim of this survey was to interview a sample drawn from the national population which was
likely to be highly varied, there was a possibility of sampling error and bias. It was important
to ensure that our drawn sample and findings reflected the national population accurately. One
way in which it was possible to deal with the problem of sampling error and to minimize the
bias was to perform stratified random sampling. It is important to emphasize that this does not
mean that this type of sampling can eliminate bias and sampling error completely. According
to Bryman (2004) stratified probability sampling keeps the sampling error in check better than
non-probability sampling.
Another issue that is relevant to this survey relates to the choice of the sample size. A
crucial criterion for the decision about the sample size was the level of standard error that we
were prepared to tolerate in our findings. Considering this criterion we decided our sample
size would be n = 385 (interviewees) out of 1.959.714 which was the total number of
inhabitants in Kosovo. The calculation of the sample size was performed using a sample size
calculator with a 5% margin of error, 95 % level of confidence and 50% response distribution.
The stratified random sampling has been applied as a type of probability sample in
selecting people who would be included in the sample. One of the main reasons that we
performed this type of sampling was that we wanted our sample to exhibit a proportional
representation of the different strata of the population. Bryman (2004) states that by selecting
a stratified sample, the standard error of the mean will be smaller since the variation between

- 16 -
strata is eliminated and the population will be better represented in the sample in terms of the
stratification criteria used. Moreover, this type of sampling permits us to employ a test of
statistical significance and draw inferences about the whole population from which the sample
was drawn.
The stratification of the sample was done according to three criteria: region, age and
gender of the respondents. The sample size of n = 385 respondents was first stratified among
regions according to the number of inhabitants in each region. The allocated number in each
region was then stratified in terms of age and gender. The following figure illustrates the main
stages involved in stratification of the sample units.

Figure 3.Stratification of the sample size

n = 385

Prishtina Prizreni Mitrovica Peja Gjilani

Age 18-30 Age 18-30 Age 18-30 Age 18-30 Age 18-30

Female 16 Male 16 Female 13 Male 13 Female 9 Male 9 Female 13 Male 13 Female 12 Male 12

Age 31-40 Age 31-40 Age 31-40 Age 31-40 Age 31-40

Female 11 Male 11 Female 9 Male 9 Female 7 Male 7 Female 10 Male 10 Female 9 Male 9

Age 41-50 Age 41-50 Age 41-50 Age 41-50 Age 41-50

Female 10 Male 10 Female 7 Male 8 Female 6 Male 6 Female 8 Male 8 Female 7 Male 7

Age 51-60 Age 51-60 Age 51-60 Age 51-60 Age 51-60

Female 8 Male 8 Female 6 Male 6 Female 4 Male 4 Female 6 Male 7 Female 6 Male 6

Over 60 Over 60 Over 60 Over 60 Over 60

Female 3 Male 4 Female 3 Male 3 Female 2 Male 2 Female 3 Male 3 Female 3 Male 3

Total Total Total Total Total

Female 48 Male 49 Female 38 Male 39 Female 28 Male 28 Female 40 Male 41 Female 37 Male 37

In spite of its advantages, this method of sampling could be feasible and economically
reasonable only when it is relatively easy to identify and allocate units within strata. In our
case stratification of the sample size into the strata entailed a great deal of work since there
was no available list on distribution of the population by age and gender. Due to that, the
actual respondents deviated slighlty in terms of these two strata compared to the one that is
shown in Fig.3 (see Appendix A2 / Table 2).

- 17 -
3.3 Data Collection

In order to find out about the consumers’ behaviour, their attitudes, preferences and
demand for milk and dairy products, a structured questionnaire was developed and used for
the collection of data. The main reason for employing this method of interviewing was to
provide exactly the same context of questioning to all interviewees. Given that replies of the
interviewees are in the response to identical interview inducements, the interviewees’
responses can then be aggregated. The advantages of this type of interviewing go further than
this. One advantage that is particularly significant is the reduction of error due to variation in
the questions asked (Bryman, 2004). This method of interviewing also ensures greater
accuracy and an easier way of processing the respondents’ answers.
The data used in this study were obtained through direct interviews (face to face) with
Kosovo consumers aged 18 years and older. In order to avoid overestimation of the market
demand for milk and dairy products, even those households who happen to have their own
cows were included in the interviewing process. The survey was conducted during the period
of January-February in 2009. A pre-test was conducted in order to ensure that survey
questions and instrument as a whole functions well. Interviews with the respondents were
administered by researcher and trained field assistants. Statistical Program for Social Sciences
(SPSS) was employed for processing and analysing of data. The data is on a monthly basis
and provides information on consumers’ buying behaviour, preferences, their attitudes toward
product attributes, the quantity consumed and expenditures for milk and dairy products
(yoghurt, fruit yoghurt, cream, curd, cheese and butter). The survey also provides data on
respondents and households characteristics including sex, age, level of education, profession,
number of family members, number of children aged 14 years and younger, respondent’s
monthly income, HH income and number of employed family members.
Respondents were therefore asked to state whether they fulfil their needs for those
products by monthly income available. The questionnaire also included retrospective
questions on changes that had occurred over the previous year in consumers’ preferences
towards milk and dairy products. These questions give responses about whether the
consumers prefer new dairy products (innovation), what attracted them to start buying new
dairy products and channels they used to become aware of the new dairy products. The main

- 18 -
intention of asking these questions was to find out whether Kosovo consumers belong to
conservative or progressive consumers category with regard to dairy products.

3.4 Model Specification

Two associated statistical techniques, multiple linear function and binary logistic function
have been used to relate the dependent variable(s) to the independent variables. The model of
multiple linear functions was performed for those dependent variables measured on
continuous scales. There were, however, many dependent variables with binary responses that
had to be evaluated for the effects of multiple independent variables on a dichotomous
outcome. For the variables with dichotomous responses the model of binary logit-function
was categorized to take on two values, Y = 1 when the response outcome for each subject is a
“success” and Y = 0 otherwise.
The logit- function with all the main effects is given as:

logit [(Yi = 1)] = α + β1 Age + β2 Education + β3 HH size + β4 Children + β5 Employment + β6 HH income + β7p1
+ β7r1 + β8r2 + β9r3 + β10r4 + β11e1 (1)

Where, Yi indicates the exhibited preference of buying product i by jth respondent {1 =


Yes, 0 = No}; α is intercept; β1……….. β11 are estimators coefficients; {p1} is the indicator
variable for the first (of two) places; {r1, r2, r3, r4} are indicator variables for the first four (of
five) regions; {e1} stands for the first indicator variable (of two) employment R. Consumers’
statements on the question as to whether they buy milk and other dairy products was defined
as dependent variable Y (= 1 if the respondent buys milk and other dairy products and 0 if
not). The statement of the consumers is then estimated as a function of respondent’s age,
respondent’s education, household (HH) size, number of children below 14 years old, number
of employed members within the HH, HH income, place as a factor with two indicator
variables: Rural and urban, region as a factor with five indicator variables: Prishtina, Prizren,
Peja, Gjilan, Mitrovica, respondent’s employment as a factor with two indicator variables:
Yes and no. The model treats place, region, and respondent’s employment as nominal-scale
factors, whereas, respondent’s age, level of education, HH size, children below 14 years old,
employed members within the HH, and HH income are treated as continuous-scale factors.
Similar models have been applied for the statement of the consumers on buying yoghurt, fruit

- 19 -
yoghurt, cream, curd, cheese and butter. A full description of other variables with
dichotomous outcome which are used as dependent variables and estimated with the same
model and the same explanatory variables is presented in Table 8.
In the model of the multiple linear function the quantity of milk and other dairy products
purchased by a household within a month is a function of respondent’s age and education, HH
size, number of children aged 14 and younger, number of employed family members and HH
income. The same model has been used for the total expenditure on milk and other dairy
products.
Linear function equations are specified as follows:

Q i = b0 + b1Ag + b2De + b3Hh + b4Ch + b5Me + b6In + Ui (2)

E i = b0 + b1Ag + b2De + b3Hh + b4Ch + b5Me + b6In + Ui (3)

T = b0 + b1Ag + b2De + b3Hh + b4Ch + b5Me + b6In + Ui (4)

where, Qi is the quantity of product i purchased in litre/month in the HHi ; E i is the


expenditure on product i in euro/month in the HHi; T is total expenditure on milk and six
other dairy products in euro/month in the HHi; (i = 1………..385); b0 indicates the intercept;
Ag denotes age of the respondent i; De indicates education of the respondent i; Hh stands for
the HH size; Ch indicates number of children below 14 years old in the HHi; Me stands for
number of employed family members in the HHi; In indicates average monthly income in the
HHi; b1..........b6 are estimated coefficients of the variable (s); and Ui is a random error term.
Other statistical models were employed in order to analyse the set of data obtained by the
research study. Analysis of variance (ANOVA) has been used to find out whether there were
significant differences between means of the HH size, number of children below 14 years old,
number of employed family members, HH income, respondent’s income and education as a
factor of gender, place and region. Test of independence was performed to study whether
there was dependency between the market places that consumers preferred to buy dairy
products, the frequencies of buying dairy products, and the evaluation of product attributes
with demographic and socioeconomic characteristics.

- 20 -
3.4.1 Variable Description and Measurement

Tables 4, 5, and 6 present a full description of all input variables that were accommodated
into fitted models in the equation 2, 3, and 4, while Tables 7, 8, and 9 describe the variables
used in equation 1.

Table 4. Dependent variables used in the estimation of the average quantity of the product i
purchased in litre or kilogram in the HHi
Dependent Variable Description of Variable
Q Milk (l/month) Quantity of milk purchased in the HHi
Q Yoghurt (l/month) Quantity of yoghurt purchased in the HHi
Q Fruit Yoghurt (kg/month) Quantity of fruit yoghurt purchased in the HHi
Q Cream (kg/month) Quantity of cream purchased in the HHi
Q Curd (kg/month) Quantity of curd purchased in the HHi
Q Cheese (kg/month) Quantity of cheese purchased in the HHi
Q Butter (kg/month) Quantity of butter purchased in the HHi
Note: Q indicates quantity of the product i.

Table 5. Dependent variables used in the estimation of the average monthly expenditures on
the product i in euro per month in the HHi
Dependent Variable Description of variable
E Milk (euro/month) Expenditures on milk in the HHi
E Yoghurt (euro/month) Expenditures on yoghurt in the HHi
E Fruit yoghurt (euro/month) Expenditures on fruit yoghurt in the HHi
E Cream (euro/month) Expenditures on cream in the HHi
E Curd (euro/month) Expenditures on curd in the HHi
E Cheese (euro/month) Expenditures on cheese in the HHi
E Butter (euro/month) Expenditures on butter in the HHi
E Total (euro/month) Expenditures on milk and dairy products in the HHi
Note: E indicates expenditures on the product i.

- 21 -
Table 6. Independent variables used in the estimation of the quantity and the expenditures on
milk and six other dairy products
Independent Variable Description of variable
Age Respondent’s age given in years
Education Respondent’s education given in years
HH size Number of members in the HHi
Children Number of children in the HHi aged 14 and younger
Employment Number of employed members in the HHi
HH income Average monthly income in the HHi
Note: HH monthly income is given in euro

Table 7. Dependent variables used in the estimation of the fitted model in the equation 1
Dependent Variable Description of Variable
Product Exhibited statement on buying product i by jth respondent
Milk Y = 1 if the respondent buys milk and Y = 0 otherwise
Yoghurt Y = 1 if the respondent buys yoghurt and Y = 0 otherwise
Fruit yoghurt Y = 1 if the respondent buys fruit yoghurt and Y = 0 otherwise
Cream Y = 1 if the respondent buys cream and Y = 0 otherwise
Curd Y = 1 if the respondent buys curd and Y = 0 otherwise
Cheese Y = 1 if the respondent buys cheese and Y = 0 otherwise
Butter Y = 1 if the respondent buys butter and Y = 0 otherwise

Table 8. Dependent variables used in the estimation of the fitted model in the equation 1
Dependent Variable Description of Variable
Fulfilment of the Y = 1 if the respondent fulfilled needs for the dairy products
respondent’s needs with his/her current monthly income and Y = 0 otherwise
Preferring domestic Y = 1 if the respondent preferred domestic dairy products
dairy products and Y = 0 otherwise
Preferring foreign Y = 1 if the respondent preferred foreign dairy products and
dairy products Y = 0 otherwise
Preferring new dairy Y = 1 if the respondent preferred new dairy products and
products innovation and Y = 0 otherwise
Buying new dairy Y = 1 if the respondent during the last year started buying
products new dairy products and Y = 0 otherwise

- 22 -
Table 9. Independent variables (factors) used in the estimation of the fitted model in the
equation 1
Factor Indicator Variable
Place 1 = Rural; 0 = Urban
Region 1 = Prishtina; 2 = Prizren; 3 = Peja; 4 = Gjilan; 5 = Mitrovica
EmploymentR Respondent’s employment 1 = Yes; 0 = No

Table 10. Recoded variables used in the test of independence


Variable Recoded Variable
Age 1= 18-30; 2 = 31-40; 3 = 41-50; 4 = 51-60; 5 = older than 60
Education 1 = Primary school; 2 = Secondary school;3 = University
IncomeR 1= up to 200; 2 = 201-400; 3 = 401-600; 4 = more than 600
HH size 1 = up to 2 members; 2 = 3-4; 3 = 5-6; 4 = more than 6
Children 1 = 1 child in the HH; 2 = 2-3; 3 = 4 and more
Employment 1 = 1 employed family member; 2 = 2-3; 3 = 4 and more
HH income 1 = up to 200; 2 = 201-400; 3 = 401-600; 4 = 601-800; 5 = more than 800
Note: Variable IncomeR indicates respondent’s income; HH income and respondent’s income are given in euro

3.4.2 Estimation Procedure

In view of the fact that this study was designed to answer certain questions about
consumers’ behaviour, preferences and demands for milk and other dairy products, several
predictor variables have been included in the models (equation 1, 2, 3 and 4). Models with
several explanatory variables often suffer from multicollinearity, making it seem that no one
variable is significant when all the others are in the model (Agresti, 2007). The selection
process of the fitted model becomes more complex as the number of explanatory variables
increases. The estimation procedure of the stated preference by consumers for buying milk
and other dairy products began with a preliminary model presented in equation (1). Backward
elimination procedure was used to select an optimal model that tends to have its fitted values
closest to the true expected values. With this algorithm, the fitted model began with all
explanatory variables presented in equation (1), and then the variable that produced the
smallest decrease in residual variance was removed from the model. This method takes out
terms successively, until there are no variables remaining in the equation that could be
removed without significantly increasing the residual variance. Backward elimination uses the

- 23 -
incremental sum of squares as a measure of the increase in the residual sum of squares
associated with removing the variable (Glantz and Slinker, 2001).
The next stage of the model selection checked the model fit. The validity of the fitted
model in the equation 1 was checked by using log-likelihood statistic, while Wald statistic
was used in assessing the contribution of predictors. Multiple linear functions in the equation
2, 3, and 4 were judged by using the goodness of fit (R2) statistical tests of significance (F
statistics for testing whether the fitted model reduced significantly residual variances, t-
statistics for testing individual regression coefficients). χ 2 Statistics has been carried out for
testing the independence between variables. The coefficient of Cramer’s V was used for
measuring the strength of association between two categorical variables.

- 24 -
CHAPTER IV

4. Results and Discussion

This chapter presents the results and analyses the set of data obtained from 385 surveyed
respondents. The chapter is structured as follows. Basic statistics description is given to
demonstrate the overall household characteristics. This section provides results of the
pairwise comparisons among levels of gender, place, and region with regard to the household
characteristics. Basic description is presented for some of the key consumer behaviour
questions such as “what, who, where, how and when” they buy milk and six other dairy
produces. It presents the estimated parameters obtained from the two types of functions:
binary logistic and multiple linear functions. To give a clear view and ease interpretation of
the estimated coefficients, estimation models are given successively for each dairy product.
The last section of this chapter presents tests of independence between the market places
preferred by consumers, the frequencies, and the evaluation of product features with
demographic and socioeconomic characteristics.

4.1. Descriptive Statistics on the Household Characteristics

Table 11 presents a summary statistics on the household characteristics according to


demographic and socioeconomic indicators. The average sample scores given below were
summarized for the whole country.

Table 11. Summary statistics of the household characteristics


Household characteristics Minimum Mean Maximum Std. Error Std. Deviation
HH size 1 6.17 27 ± 0.17 3.39
Children 1 2.21 13 ± 0.10 1.53
Employment 1 2.12 8 ± 0.026 1.16
HH income 60 585.79 3500 ± 26.53 52.51
IncomeR 40 281.74 1500 ± 13.29 219.5
Education 3 12.5 20 ± 0.15 3.04
Note: HH income and respondent’s income are given in euro per month

- 25 -
The overall household characteristics were further examined, in order to find out whether
there were significant variations of the average sample scores among the levels of gender,
place and region. The result in table 12 shows that there was significant variation between
females and males in terms of the average monthly income (PANOVA = 0.016). Males had a
tendency to generate higher income compared to females. Publication given by (MPS, 2009a)
states that unemployment rate is much higher for women (55 %) than for men (39 %). The
majority of employed women were engaged in education, health, trade, and the agricultural
sector (MPS, 2009). Disparity on the unemployment rate, as well as engagement of women in
the sectors which yield lower income caused the significant variation on the income levels
between males and females. Highly significant variation was also found in the level of
education (PANOVA = 0.003). This result clearly shows that women in Kosovo were less
educated compared to men. It also corresponds with the statistical data on the level of
education given by MPS (2007), where the proportion of females and males who had not
completed upper secondary education was 40:60 %.
Study results showed that there were discrepancies between rural and urban households, in
terms of the family members. Rural households had significantly more family members than
the urban households (PANOVA = 0.000). Moreover, the average number of children in the
household below 14 years old was distinctly higher in rural households than in the urban
(PANOVA = 0.003). This result was expected, considering that the rural households had
significantly more family members than the urban. It was proved that these two variables, the
household size and the number of children below 14 years old, were highly related with a
Pearson’s correlation coefficient r = 0.750**. However, insignificant variations were found
between rural and urban households, with regard to the number of employed family members
(PANOVA = 0.297). Concerning the average monthly income, rural households had
significantly higher income than urban households (PANOVA = 0.032). But income per capita
was still higher in urban areas (98.41 euro), than in rural areas (88.75). During the period of
1990-99, approximately 18 to 20% of the Kosovo population emigrated at Western European
countries. Majority of them were young people coming from rural areas. Statistics of Living
Standard in 2007 indicated that in Kosovo every tenth household lives by money sent abroad
the country (remittances). Hence, remittances by emigrants can be considered as an influential
factor in the variations of the average household incomes between rural and urban areas. Even
though, there were significant differences in the average household incomes, insignificant
variations were found in terms of the respondent’s average monthly income (PANOVA = 0.088).
Highly significant variations were marked also in the level of respondents education (PANOVA

- 26 -
= 0.005). Respondents coming from rural areas had predispositions to be less educated than
the respondents from urban areas.
The sample data showed highly significant variations on the average household size
between the regions (PANOVA = 0.000). The region of Peja differed significantly from
Prishtina, Prizren and Mitrovica. The households from Mitrovica had significantly more
family members than the households from Prishtina and Gjilan. A significant variation is
found also between Prizren and Gjilan. Significant variations of the household size appeared
to be important in explaining the variations of the number of children below 14 years old
(PANOVA = 0.019). Regions that had significantly more family members were predisposed to
have more children below 14 years. Significant differences were observed between the
Mitrovica and Prishtina, the first one had predominantly more children below 14 years old.
The households from Peja region appeared to have fewer children than any other region.
There were no significant variations between the levels of region, with regard to the
number of employed family members (PANOVA = 0.161). The insignificant variation between
regions was obtained also in terms of the respondent’s monthly income (PANOVA = 0.229) and
the respondent’s education (PANOVA = 0.057). However, highly significant variations were
observed in the level of income (PANOVA = 0.005). The households from Prishtina and Prizren
had significantly higher monthly income than the households from Peja and Gjilan.

Table 12. Pairwise comparison between females and males


Gender Female Male Difference
Sample Mean / Std. Deviation X SD X SD (F X - M X )
IncomeR 243.4 199.3 308.4 229.4 -65*
Education 11.69 3.272 12.620 2.709 -0.931**
Note: Significance of variations is denoted as follows: *P < 0.05; **P < 0.01; ***P < 0.001.

Table 13. Pairwise comparison between rural and urban household


Place Rural Urban Difference
Sample Mean / Std. Deviation X SD X SD (R X - U X )
HH size 7.593 4.453 5.581 2.630 2.012***
Children 2.639 2.009 2.00 1.178 0.639**
Employment 2.218 1.474 2.080 0.999 0.138
HH income 673.9 649.3 549.2 452.9 124.7*
IncomeR 247.2 188.6 296.6 230.4 -49.4
Education 11.487 3.071 12.430 2.984 -0.943**

- 27 -
Table 14. Pairwise comparison between regions
Region r1 r2 r3 r4 r5
Sample mean / SD SD SD SD SD
X X X X X
Std. Deviation
Reference r1 r2 r3 r4
Mean difference (r2-r1) (r3-r1) (r4-r1) (r5-r1) (r3-r2) (r4-r2) (r5-r2) (r4-r3) (r5-r3) (r5-r4)
*** ** **
HH size 6.10 3.42 6.96 3.9 5.02 1.9 5.9 3.1 7.25 4.1 0.86 -1.1 -0.24 1.15 -1.94 -1.097 0.288 0.84 2.23 1.385
**
Children 1.94 1.27 2.48 2.0 1.74 0.8 2.4 1.5 2.74 1.7 0.54 -0.20 0.41 0.80 -0.74 -0.13 0.26 0.61 1.00 0.39
Employment 2.23 1.24 2.31 1.2 1.88 0.8 2.0 1.2 2.11 1.3 0.08 -0.35 -0.20 -0.12 -0.42 -0.28 -0.19 0.15 0.23 0.083
*
HH income 710 587 673 625 494 427 460 232 548 518 -37.3 -215.5 -250 -162.3 -178.2 -212.8 -125 -34.55 53.16 87.71
IncomeR 315 230 264 176 315 270 263 213 232 183 -51.4 0.27 -51.68 -82.84 51.10 -0.31 -31.47 -51.41 -82.6 -31.16
Education 12.66 3.08 11.5 3.4 12.35 2.67 12 3.12 12.5 2.6 -1.13 -0.31 -0.99 -0.12 0.82 0.14 1.01 -0.68 0.19 0.87
* ** ***
Note: Given characters denote: r-region; r1-Prishtina; r2-Prizren; r3-Peja; r4-Gjilan; r5-Mitrovica; Significance of variations is denoted as follows: P < 0.05; P < 0.01; P<
0.001.

28
4.2 Descriptive Statistics on Consumer buying Behaviour
Based on the information gathered from interviewed respondents, it was clear that milk
and other dairy produces stand as regular staple food to most of the Kosovo households. Out
of 385 surveyed respondents 87% proclaimed that they buy milk on a regular basis, (see
Figure 4) only 13% of the respondents did not buy milk at all. An attempt was made to come
across the reasons for the proportion of the respondents who did not use to buy milk. Hence,
the remaining percentage 13%, was scrutinized further and it was remarked that 90% of the
respondents were from rural areas, who predominantly own cows and this was the solely
reason given by them for not buying milk. Respondents were questioned whether they buy
other dairy produces such as yoghurt, fruit yoghurt, cream, curd, cheese and butter. With
regard to the buying habits towards yoghurt and fruit yoghurt, about 79 % of the respondents
stated that they regularly buy yoghurt. The proportion of the respondents who were positively
responding to the question on buying fruit yoghurt was 50:50%.
Consumer’s behaviour to product choice is greatly affected by economic circumstances and
lifestyle pattern which is shaped by whether consumers are money-constrained or time-
constrained (Kotler and Keller, 2006 pp182-183). The sample included respondents who
happened to own cows and it was the only reason stated in terms of the response of not buy
milk. The same matter has accompanied other dairy produces, where out of 21%, who did not
use to buy yoghurt 12% stressed the similar reason as those who did not use to buy milk.
Apart from this reason, there were however other factors influencing consumers lifestyle
and their behaviour to dairy products. Income was an additional influential factor affecting
consumers buying behaviour. It was realized that lower income households could not afford
buying yoghurt direct from the market. A cheaper alternative for this consumers group was
buying fluid milk from farmers and process it into value-added products such as yoghurt.
Out of 50% of the respondents who did not use to buy fruit yoghurt, more than 25% do not
prefer and lack the habit of buying it. This attitude was notably noticed among the old age
group. On the other hand, no presence of children in the household was another indirect factor
influencing the product choice, as most of them consider it as a children product. In fact the
responses by this consumers group in terms of not buying fruit yoghurt were in some way
intertwined by lack of preferences, habits and no presence of children in the household. For
the remaining proportion, 23% of the respondents that did not use buying fruit yoghurt,
economic issue stand as the main influential factor on their attitudes towards this product. As

29
a result of budgetary constraints on this consumers group, fruit yoghurt was considered as
luxury product and predominantly unachievable given their current economic circumstances.
Nearly 70% of the respondents claimed that they commonly buy cream, 66% was the
proportion of the respondents who buys curd. The proportion of the respondents who buys
cheese was similar to the one for milk (86%). The proportion of the respondents that use to
buy butter was roughly as one for the fruit yoghurt, which accounted for 52%. As it is
previously cited, an effort was also made to acquire the reasons given by the proportion of the
respondents that did not use to buy cream, curd, cheese and butter. Before getting to the
depiction of the given reasons, it is worth noting that an open ended question was asked and
we ended up with diverse responses as the respondents answered in their own terms.
Consequently, the given responses were interpreted and coded in order to be analyzed
quantitatively.
The study results have shown that the given main reasons for not buying above mentioned
products were related to personal factors, including respondent’s lifestyle, economic
conditions and age. For the respondents who owned cows it still remained the sole reason for
not buying milk, cream, curd, cheese and butter which accounted roughly the same proportion
for all of these produces, 12%. About 11% could not afford buying cream, while 7% was the
proportion of the respondents that did not prefer buying this product. The proportion of the
respondents that did not use to buy curd was 34%. Lack of preference towards this product
was the main reason stated by the respondents, which counted roughly 19%. There were few
respondents that could not afford buying it. The results of this study indicated that 99% of the
respondents were consumers of cheese. The proportion of the respondents that did not
favoured buying butter was almost 32%. There was small a percentage (5%) that could not
afford buying butter. It was particularly marked that buying habits towards butter and cream
differed by personal characteristics as age and geographic niches. Respondents among urban
households and young adults aged 18 to 40 appeared to buy less dairy products with lower
content of fat.
The family is incessantly considered as the most important consumer buying organization
in society, and family members as the most influential primary reference group (Tour and
Henthorne, 1995). Nowadays, most of the Kosovo households consist of husband, wife,
children and often grandparents. To examine the role and influence of family members in the
purchase of milk and other dairy products, key consumer behaviour questions were asked,
like ‘who buys’ and ‘who makes buying decisions’ on the dairy products choice. Involvement
of husband and wife in the purchase of products varies by product category. ‘’Wife has

30
usually acted as the family’s main purchasing agent for food items’’ (Kotler and Keller, 2006,
p.179). However, respondents’ answers showed highest proportion 34%, for joint purchase
(husband and wife) of milk and dairy products. Men were disposed to be more active in the
purchase of milk and dairy products compared to women. The percentage of men accounted
for 30%, which was higher than the proportion of women (21%). The remaining percentage
indicates involvement of children in the purchase of milk and dairy products. Even though
traditional purchasing roles are changing, study result proved that men were buying
substantially more than women. In fact this result was not surprising, given that
unemployment rate was statistically significant higher for the women than for men. Moreover,
income generating ability for the women was considerably less than for men.

Figure 4. Respondents’ answers in terms of buying milk and six other dairy products

100
90
80
70
60
50
40
30
20
10
0
Milk Yoghurt Fruit Cream Curd Cheese Butter
Yoghurt

Yes No

To understand how the consumers make their buying decisions in connection with milk
and other dairy products, it was particularly important to identify who among the family
members contributes and makes the buying decision. In the purchase decision task, family
members can be initiators, influencers, deciders, buyers, or users (Kotler and Keller, 2006 p
203). Considering different purchasing roles by the family members in the buying decisions,
men had much more buying power and were mainly engaged as buyers of milk and other
dairy products. Women, on the other hand, had a tendency to be much more influencers and
deciders in the purchasing decisions. Nearly 50% of the respondents stated that the wife is the
one of the family members who usually decides what kind of milk and other dairy products
should be purchased. About 20% of the respondents stressed that decisions were usually made

31
jointly by wife and husband. The proportion of the respondents stating that husband makes
buying decisions was smaller, about 15%. Only 8% of the respondents stated that purchasing
decisions towards kind of milk and other dairy products were taken by children.
The identification of the most frequented market place by consumers, when buying milk
and other dairy products was captured by asking ‘where do you usually go when you buy milk
and other dairy products’? As shown in Figure 5, it was revealed that the most frequented
market place by consumers for milk and other dairy products was supermarket.

Figure 5.Market places preferred by consumers when buying milk and other dairy produces

Butter

Cheese

Curd

Cream

Fruit Yoghurt

Yoghurt

Milk

0% 20% 40% 60% 80% 100%

By farmer Street vendor Green market Grocery store Supermarket

Figure 6. Frequencies of buying milk and other dairy products

Butter

Cheese

Curd

Cream

Fruit Yoghurt

Yoghurt

Milk

0% 20% 40% 60% 80% 100%

Every day Twice a week Once a week Twice a m onth Once a m onth

32
Figure 6 shows that milk and yoghurt, were the most frequent dairy products bought by the
consumers, followed by fruit yoghurt and cream. The other dairy products such as curd,
cheese, and butter were less frequently bought by the consumers. The obtained results
indicated that the frequency of buying milk and other dairy products was related to freshness
and durability of the products. The above figure illustrates that majority of the Kosovo
consumers prefer more fresh dairy products, particularly milk and yoghurt.
Table 15 provides the averages of monthly quantity of milk and six other dairy products
purchased by the HHs. The research study did not provide directly per capita consumption of
the various dairy products. However, given that the averages of purchased quantity by the
HHs was already known, as well as the HH size, per capita consumption for the various dairy
products could be estimated, and the estimates are presented in Table 16.

Table 15. Summary statistics of the average quantity of milk and other dairy products
purchased by the HHs
Quantity of product i Minimum Mean Maximum Std. Error Std. Deviation
Q Milk (l/month) 3 26.30 90 ± 0.79 14.53
Q Yoghurt (l/month) 2 10.51 40 ± 0.43 7.53
Q Fruit Yoghurt (kg/month) 1 2.51 30 ± 0.22 3.08
Q Cream (kg/month) 1 2.34 35 ± 0.17 2.86
Q Curd (kg/month) 1 3.93 20 ± 0.18 2.88
Q Cheese (kg/month) 1 4.67 20 ± 0.14 2.53
Q Butter (kg/month) 1 1.39 10 ± 0.0887 1.27

Table 16 The annual average per capita consumption


Product Milk Yoghurt Fruit yoghurt Cream Curd Cheese Butter
Quantity l/year l/year Kg/year Kg/year Kg/year Kg/year Kg/year
51.15 20.44 4.88 4.55 7.64 9.08 2.70

The averages of monthly expenditures by the HHs on milk and other dairy products are
given in table 17. The monthly average expenditure was higher for milk, cheese, and yoghurt
than for other products. The total average monthly expenditure by the HHs on milk and other
dairy products was estimated to be 56.17 euro/month (Std. Error ± 1.55; Std. Deviation
29.25). The total average monthly expenditure on milk and other dairy products counted at

33
9.6%, of the total average household monthly income. Regarding the fulfilments of the
consumers’ needs for milk and other dairy products, roughly 57% of the respondents stressed
that they manage to fulfil demand for milk and dairy products with their current monthly
income. About 43% of the interviewed respondents stated that they did not fulfil their needs
with the current monthly income. Majority (72%) of the respondents who did not fulfil their
demand for such products, were from urban areas. This result was not surprising as the
households in the urban areas had significantly lower income (PANOVA = 0.032) compared to
the rural households. Within the proportion of the respondents who did not fulfil the needs
(73%) were households who had lower monthly income (less than 400 euro). It has been
realized that the households who had more than five family members, had less tendency to
fulfil the demand for milk and other dairy products.

Table 17. Summary statistics of the average expenditures on milk and other dairy products by
the HHs
Expenditures on product i Minimum Mean Maximum Std. Error Std. Deviation
E Milk (euro/month) 2 19.37 60 ± 0.59 10.75
E Yoghurt (euro/month) 2 10.02 40 ± 0.40 7.11
E Fruit yoghurt (euro/month) 1 5.85 36 ± 0.35 4.88
E Cream (euro/month) 1 6.11 70 ± 0.39 6.44
E Curd (euro/month) 1 6.69 30 ± 0.31 4.93
E Cheese (euro/month) 2 14.69 50 ± 0.45 8.16
E Butter (euro/month) 1 5.08 40 ± 0.30 4.31

Despite the seasonal variation on milk production, which is characterized by over-


production during the summer season, the demand for milk and other dairy products was quite
stable throughout the year. Figure 7 shows that majority of the Kosovo consumers’ (64 %),
consumed milk on a regular basis throughout the spring, summer, and autumn months. The
consumption of milk appeared to be higher during the winter months. Concerning the
seasonal consumption patterns on yoghurt, Figure 8 shows that the consumers tend to behave
differently compared with the seasonal consumption patterns on milk. The consumption of
yoghurt seemed to be quite stable during the spring, autumn, and winter months, while its
consumption was higher throughout the summer months.
Figure 9, 10, and 13 indicate that Kosovo consumers’ tend to behave in a seasonal manner
with regard to the fruit yoghurt, cream, and butter. The consumption of fruit yoghurt was less

34
throughout the year, with small increases during the spring and the summer months. The
products with the higher content of fat such as cream and butter were prone to be consumed
less throughout the year except of butter, its consumption tends to be higher during the winter
months. Regarding the seasonal consumption patterns on curd and cheese, Figure 11 and 12
show that consumption of these dairy products was quite stable throughout the year.

Figure 7. Seasonal consumption patterns on milk


70 65.2 64.7
61.6 61
60

50

40 35.3
31.4
30 22.6
19.7
20 14.8
11.4
10 6.2
2.9
0
Spring Summer Autumn Winter

Milk
Less Constant More

Note: The variable was measured on scale of 1 to 3, where 1 is “when the respondent consume less of the
product i for a given season”, 2 is “when the respondent consumes product i on a regular basis”, and 3 is “when
the respondent consumes more the product i for given season”.

Figure 8. Seasonal consumption patterns on yoghurt


80
67
70 62.6 62.9
60
60

50

40

30 26.5
22.6 23.4 22.9
20 14.3
10.9 11.2
10 3.4
0
Spring Summer Autumn Winter

Yoghurt

Less Constant More

35
Figure 9. Seasonal consumption patterns on fruit yoghurt

40 35.8 36.4
35 30.6
30
24.9
25
20 15.6
15 12.5 12.2 11.7
10.4
10 7.8

5 2.6 2.6

0
Spring Summer Autumn Winter

Fruit Yoghurt

Less Constant More

Figure 10. Seasonal consumption patterns on cream

60
48.6
50 42.6 44.4 44.4

40

30 23.6 24.9 23.4


22.1
20 15.8 15.6 14.3
8.6
10

0
Spring Summer Autumn Winter

Cream

Less Constant More

Figure 11. Seasonal consumption patterns on curd


50 44.9
42.9
45 39.7
38.4
40
35
30
25 21.6 21 21.6
20 19.2 18.7
20 16.4 15.6
15
10
5
0
Spring Summer Autumn Winter

Curd
Less Constant More

36
Figure 12. Seasonal consumption patterns on cheese
70 65.7 64.4 62.1 59.7
60
50
40 35.3
32.3
28.8 28.8
30
20
10 4.2 5.5 4.4 3.6
0
Spring Summer Autumn Winter

Cheese

Less Constant More

Figure 13. Seasonal consumption patterns on butter


70
56.6 57.4
60
51.9
50
39
40

30
19.5
20

10 5.5 4.9 4.4 6.2 4.2


0.5 0.3
0
Spring Summer Autumn Winter

Less Constant
Butter More

Butter

Concerning the consumers’ evaluation on product features, Figure 14 displays that


nutritive content, taste, product safety, price and origin of the product were highly ranked by
the consumers in the order of importance. Package size, wrapping, and brand appeared to be
not very important.
With respect to the consumers’ preference for the package size, the results have shown that
majority (84%) of the respondents preferred a milk package size at 1 litre, a small fraction
(about 8%), preferred a package size at 2 litres. A 1 litre package was the most preferred one
for the yoghurt (60%), followed by 500 ml which counted at 26%, while the responses for the
remaining proportion were quite diverse. The packages from 250, 200, and 180 gram were the
most preferred for the fruit yoghurt. The consumers’ preference towards package size for
cream were 1000, 500, and 200 gram. Bigger packages were preferred for curd and cheese (3,

37
2, and 1 kg), whereas butter packages from 500, 250, and 200 gram were the most preferred
by interviewed respondents. It is important to note that preferences for the package size of
curd, cheese, and butter were positively related with the HH size. This would mean that the
households with more family members, had tendency to prefer bigger packages for curd,
cheese, and butter.

Figure 14. The evaluation of product features in order of importance


Product origin

Package size

Wrapping

Brand

Price

Product safety

Taste

Nutritive Content

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%

Unimportant Not very important Important Very important

Regarding the preferences of domestic versus foreign dairy products, consumers had a
favourable bias towards domestic dairy products (78%). Quality, safety, taste, and price were
the main product features related to the consumers’ preference for purchase of the domestic
dairy products. Apart from the product features above cited, national patriotism seemed to be
another factor influencing consumers’ preference towards domestic dairy products. Based on
the consumers’ view, the purchase of domestic dairy products helps development of dairy
industry and domestic economy in general. The package, durability, assortment, and visual
aspect, were the main product features motivating consumers to purchase foreign dairy
products. The consumer’s preference towards foreign dairy products was significantly
dependent on the respondent’s location. Majority of the respondents who preferred foreign
dairy products (82%), were from urban areas.
With respect to mass media preferences, Figure 15 shows that television, and newspapers,
were the media most often used by the respondents to get information about the dairy
products. The preference towards media was significantly associated with the consumer’s
education. The preference of the newspaper was positively related with the respondent’s

38
education, which means that newspapers tend to be preferred more by the respondents that
had higher education. The preferences toward television and radio were negatively associated
with the respondent’s education. It means that a media category such as television and radio
was most often used by the respondents who had less education.

Figure 15. The level of the mass media used by the respondents

70

60

50

40

30

20

10

0
Media most rarely Media rarely used Media often used Media most often
used used

New spapers Flyers Television Radio

Note: The variable was measured on scale of 1 to 4, where 1 is “for the media most rarely used”, 2 is “for the
media rarely used”, 3 is “for the media often used”, and 4 is “for the media most often used”.

The proportion of the respondents that did not prefer new dairy products and innovation
was 59:41%. It has been realized that the respondents aged 18 to 40 had a tendency to prefer
more new dairy products and innovation. The respondents aged 60 and older were more
conservative, they prefer dairy products that they were used to consume regularly. Majority of
the respondents (85%) did not start buying any new dairy product since the last year. It was a
small fraction (15%) of the respondents who started buying new dairy products. The
consumers’ habit for the new dairy products was significantly dependent on the HH monthly
income. Milk, yoghurt, and fruit yoghurt were the main dairy products that consumers started
buying since the last year. They were strongly motivated by curiosity, quality, taste, and price
of the new dairy products. It is important to note that milk and yoghurt are bought previously
by the consumers. It was mostly a shift from one brand to another. The supermarkets and
television were the main sources of the information for the new dairy products.

39
4.3 Model Estimation

This section provides the parameters estimated by the two statistical techniques, multiple
linear function and binary logistic function. The following subsection describes the
parameters estimated by the logistic function for the outcome variable that is categorical with
the dichotomy response. The next subsection (4.3.2) presents the parameters obtained by the
multiple linear functions.

4.3.1 Binary Logistic linear Estimated Parameters

These parameters are estimated by fitting models, based on the predictors in the equation
1. The values of the parameters are estimated using maximum –likelihood estimation “which
selects coefficients that make the observed values most likely to have occurred” (Filed, 2005).

Table 18. Relationship of whether the respondent i buys milk and the predictors included in
the equation 1
Variable B SE Wald Exp (B)
Constant 3.920*** 0.714 30.187 50.425
Place (Rural) -3.731*** 0.765 23.785 0.024
Note: B = Logistic regression coefficient; SE = Standard error; Wald = Wald statistic (which has a special
distribution known as the χ distribution); Exp (B) = Indicates the change in odds resulting from a unit change
2

in the predictor. Significant contribution of each independent variable to the model is denoted as follows: *P <
0.05; **P < 0.01; ***P < 0.001.

Table 19. Relationship of whether the respondent i buys yoghurt and the predictors included
in the equation 1
Variable B SE Wald Exp (B)
Constant 1.233* 0.582 4.489 3.431
Place (Rural) -1.640*** 0.413 15.811 0.194
Region (Prishtina) 0.438 0.667 0.431 1.550
Region (Prizren) -0.367 0.631 0.338 0.693
Region (Peja) 1.208 0.800 2.278 3.346
Region (Gjilan) -0.747 0.631 1.400 0.474
HH income 0.001 0.000 2.519 1.001

40
Table 20. Relationship of whether the respondent i buys fruit yoghurt and the predictors
included in the equation 1
Variable B SE Wald Exp (B)
Constant 0.408 0.389 1.099 1.504
HH size -0.170** 0.059 8.210 0.844
IncomeR 0.547** 0.191 8.204 1.729

Table 21. Relationship of whether the respondent i buys cream and the predictors included in
the equation 1
Variable B SE Wald Exp (B)
Constant -1.024 1.084 0.893 0.359
Place (Rural) -1.700*** 0.445 14.582 0.183
Region (Prishtina) -0.725 0.714 1.032 0.484
Region (Prizren) -1.085 0.703 2.382 0.338
Region (Peja) 0.693 0.829 0.698 1.999
Region (Gjilan) -1.031 0.704 2.148 0.357
Age 0.033 0.018 3.205 1.034
IncomeR 0.003* 0.001 4.857 1.003
HH size -0.134 0.080 2.853 0.874
Employment 0.897** 0.270 11.006 2.453

Table 22. Relationship of whether the respondent i buys curd and the predictors included in
the equation 1
Variable B SE Wald Exp (B)
Constant 1.057** 0.324 10.663 2.878
Place (Rural) -2.235*** 0.402 30.994 0.107
HH income 0.001* 0.000 4.099 1.001

Table 23. Relationship of whether the respondent i buys cheese and the predictors included in
the equation 1
Variable B SE Wald Exp (B)
Constant 3.506*** 0.586 35.815 33.329
Place (Rural) -3.393*** 0.647 27.477 0.034

41
Table 24. Relationship of whether the respondent i buys butter and the predictors included in
the equation 1
Variable B SE Wald Exp (B)
Constant -0.500** 0.172 8.390 0.607
Place (Rural) 1.214*** 0.307 15.599 3.367

Table 25. Relationship of whether the respondent i fulfils the demand for milk and dairy
products and the predictors included in the equation 1
Variable B SE Wald Exp (B)
Constant -0.854 0.589 2.098 0.426
Place (Rural) 0.748 0.436 2.939 2.112
IncomeR 0.006*** 0.002 13.869 1.006
HH size -0.261** 0.075 12.107 0.770
Employment 0.663** 0.230 8.331 1.940

Table 26. Relationship of whether the respondent i prefers domestic dairy products and the
predictors included in the equation 1
Variable B SE Wald Exp (B)
Constant -0.734 0.817 0.808 0.480
Place (Rural) 1.446** 0.544 7.074 4.247
Age 0.055** 0.020 7.758 1.057
HH size -0.123 0.069 3.202 0.884
Employment 0.416 0.245 2.872 1.515
*
HH income -0.001 0.000 5.417 0.999

Table 27. Relationship of whether the respondent i prefers foreign dairy products and the
predictors included in the equation 1
Variable B SE Wald Exp (B)
Constant 0.734 0.817 0.808 2.084
Place (Rural) -1.446** 0.544 7.074 0.235
Age -0.055** 0.020 7.758 0.946
HH size 0.123 0.069 3.202 1.131
Employment -0.416 0.245 2.872 0.660
HH income 0.001* 0.000 5.417 1.001

42
Table 28. Relationship of whether respondent i prefers new dairy products and the predictors
included in the equation 1
Variable B SE Wald Exp (B)
Constant -2.329** 0.897 6.732 0.097
Education 0.149* 0.068 4.832 1.161

Table 29. Relationship of whether respondent i started buying new dairy products last year
and the predictors included in the equation 1
Variable B SE Wald Exp (B)
Constant -4.044** 1.340 9.111 0.018
Education 0.261* 0.105 6.133 1.298
*
IncomeR -0.004 0.002 4.789 0.996

4.3.2 Multiple linear Estimated Parameters

Fluid milk

The estimated quantity of milk consumed in litre/month in the HHi was:

Q̂ Milk (l/month) = 14.352 + (0.142 Age) + (1.789 Children) + (0.009543 HH income)


3.742*** 1.665 2.535* 4.482***
(R2 = 0.15) (5)

Note: values in italics are t-statistics; significant variable in the model is denoted as follows:
*
P < 0.05; **P < 0.01; ***P < 0.001.

The estimated coefficients in the equation (5) indicated that there was positive relationship
between the quantity of milk consumed in the HHi and the three predictors included in the
model. The results suggest that given a year increase in the respondent’s age, fluid milk
consumption increases by 0.142litre/month. This interpretation stands only if the number of
children in the HHi and the average monthly income in the HHi are held constant. One child
more in the HHi is expected to cause an increase in the fluid milk consumption by
1.789litre/month. This interpretation is consistent only if the respondent’s age and the average
monthly income in the HHi are held constant. The fitted model predicts 0.009543litre/month

43
increases in the fluid milk consumption, for a one euro increase in the HH income. This
interpretation is true only if the respondent’s age and the number of children in the HHi are
held constant. The magnitude of the t-statistics showed that the average monthly income in
the HHi had greater contribution to the model, compared to the respondent’s age and the
number of children in the HHi. The R2 is only 15%, though F ratio is 10.587 and P < 0.001,
indicating that the probability of obtaining the value of F by chance was very small.

Yoghurt

The estimated quantity of yoghurt consumed in litre/month in the HHi was:

Q̂ Yoghurt (l/month) = 4.547 + (0.316 HH size) + (0.006767 HH income)


3.767*** 2.154* 5.958***
(R2 = 0.24) (6)

Positive values of the estimated coefficients indicated that there was positive relationship
between the quantity of yoghurt consumed in the HHi and two predictors HH size and HH
income. The fitted model in the equation (6) predicts 0.316litre/month increase in the
consumption of yoghurt, if the HH size increases by one member. This interpretation is true
only if the HH income is held constant. The magnitude of the t-statistics indicated that the HH
income had greater contribution to the model, meaning that explained much more of the
variability in the outcome than the HH size. The consumption of yoghurt increases by
0.006767litre/month, for every additional euro in the HH monthly income. This interpretation
is consistent only if the HH size is held constant. Even though the fitted model produces a
significant reduction in the residual variance, F = 24.844; P < 0.001, the coefficient of
determination (R2) is only 24%.

44
Fruit Yoghurt

The estimated equation for fruit yoghurt was:


Q̂ Fruit yoghurt (kg/month) = - 2.910 + (0.163 Education) + (0.167 HH size) + (0.709 Children)
- 2.113* 1.706 1.834 2.662**
+ (0.001119 HH income)
2.441*
(R2 = 0.30) (7)
The best fitted model in the equation 7 predicted 0.163kg/moth increase in the quantity of
fruit yoghurt consumption, if the respondent’s education increases by one year. This
interpretation is consistent only if other predictors in the equation are held constant. The
increase in the HH size by one member induces fruit yoghurt consumption by 0.167kg/month.
This interpretation is true only if other explanatory variables in the fitted model are held
constant. As it was expected, the number of children below 14 years old in the HHi had much
more impact than the other predictor variables included in the equation (7). One child more in
the HHi increases fruit yoghurt consumption by 0.709kg/month. This interpretation is
consistent only if other predictors in the equation are held constant. The model predicts
0.001119kg/month increases in the fruit yoghurt consumption, for a one euro increase in the
HH monthly income. This interpretation stands only if other predictors in the model are held
invariable. Although, the fitted model did not explained much of the variability in the
outcome by the predictors, R2 = 0.30. There was a significant reduction in the residual
variance, with F = 12.560; P < 0.001.

Cream

The estimated equation for cream was:

Q̂ Cream (kg/month) = - 3.363 + (0.134 Education) + (0.458 HH size) +


- 3.010** 1.722 6.255***
(0.438 Employment)
1.973
(R2 = 0.40) (8)

45
The estimated parameters in the equation 8 showed that there was positive relationship
between the quantity of cream consumed in the HHi and three explanatory variables. The rate
of cream consumption increases by 0.134kg/month, if the respondent’s education increases by
one year. This interpretation stands only if HH size and number of employed members in the
HHi are held constant. One more member in the HHi induces the cream consumption by
0.458kg/month. This estimation is consistent only if two predictors’ education and
employment are held constant. The rate of cream consumption rises by 0.438kg/month, given
a one more member employed in the HHi. This interpretation is true only if other two
predictors are held constant. The quality of the fit is considered as moderate, with R2 = 0.40
and significant reduction in the residual variance F = 31.237, P < 0.001.

Curd

The estimated quantity of curd consumed in kilogram/month in the HHi was:

Q̂ Curd (kg/month) = 0.164 + (0.0466 Age) + (0.169 HH size) + (0.477 Employment)


0.197 2.682** 2.357* 2.191*
(R2 = 0.19) (9)
The estimated parameters in the equation (9) indicated that there was positive relationship
between the quantity of curd consumed in the HHi and three explanatory variables included in
the model. Given a year increase in the respondent’s age, curd consumption increases by
0.0466kg/month. This interpretation is true only if HH size and number of employed
members in the HHi are held constant. The fitted model predicts 0.169kg/month increases in
the curd consumption, if the HH size increases by on member. This estimation is consistent
only if the respondent’s age and number of employed family members are held constant. One
more member employed in the HHi, curd consumption increases by 0.477kg/month. This
interpretation is true only if the respondent’s age and the HH size are held invariable.
Although, the coefficient of determination (R2) was only 19%, the fit of the model produces a
significant reduction in the residual variance, with F = 10.520; P < 0.001.

46
Cheese

The estimated equation for cheese was:

Q̂ Cheese (kg/month) = 1.153 + (0.03274 Age) + (0.221 HH size) + (0.467 Employment)


1.855 2.545* 4.205*** 2.993**
(R2 = 0.26) (10)
Positive values of the estimated coefficients in the equation (10) showed that there was
positive relationship between the predictors and the outcome. Given a year increase in the
respondent’s age, cheese consumption rises by 0.03274 kg/month. This estimation is
consistent only if the other two predictors HH size and employment are held invariable. From
the magnitude of the t-statistics, it was clear that the HH size had greater impact and
explained more the variability on the cheese consumption. One more member in the HHi
induces cheese consumption by 0.221 kg/month. This estimation stands only if the
respondent’s age and number of employed members in the HHi are held constant. The model
predicts an increase in the cheese consumption by 0.467 kg/month, if the number of employed
members in the HHi increases by one. This interpretation is true only if other two predictors in
the equation (10) are held invariable. Although, the fitted model explained only 26% of the
variability in the cheese consumption, there was a significant reduction in the residual
variance, with F = 20.286; P < 0.001.

Butter

The estimated quantity of butter consumed in kilogram/month in the HHi was:

Q̂ Butter (kg/month) = - 0.655 + (0.03069 Age) + (0.432 Employment)


- 1.465 3.084** 4.964***
(R2 = 0.22) (11)

The fitted model in the equation (11) predicts 0.03069 kg/month increase in the butter
consumption, for every year increase in the respondent’s age. This estimation stands only if
the number of employed members in the HHi is held constant. Based on the magnitude of the
t-statistics, the number of employed members in the HHi had greater contribution to the model

47
compared to the respondent’s age. One more member employed in the HHi, increases butter
consumption by 0.432 kg/month. This interpretation is consistent only if the respondent’s age
is held constant. The variability in the butter consumption explained by the predictors was
only 22%. However, the model produces a significant reduction in the residual variance, with
F = 15.911; P < 0.001.

Expenditures on milk

The estimated expenditure on milk in euro/month in the HHi was:

Ê Milk (euro/month) = 16.513 + (0.007594 HH income)


13.535*** 4.670***
(R2 = 0.11) (12)

Positive value of the estimated coefficient in the equation (12) indicates that there was
positive relationship between the average monthly income in the HHi and monthly
expenditures on milk. The model predicts 0.007594 euro/month increases on the milk
expenditures, if the average monthly income in the HHi rises by one euro. Even though the
quality of the fit is not very good, with R2 = 0.11, there was a significant reduction in the
residual variance, F = 21.810; P < 0.001

Expenditures on yoghurt

The estimated expenditure on yoghurt in euro/month in the HHi was:

Ê Yoghurt (euro/month) = 1.052 + (0.327 Education) + (0.298 HH size) + (0.005043 HH income)


0.407 1.789 2.108* 4.584***
(R2 = 0.19) (13)

Based on the fitted model in the equation (13) monthly expenditures on yoghurt in the
HHi, increase by 0.327 euro/month given one year increase in the respondent’s education.
This estimation is consistent only if the HH size and average monthly income in the HHi are
held constant. One more member in the HHi increases the expenditures on yoghurt by 0.298

48
euro/month. This interpretation is true only if the respondent’s education and the HH income
are held constant. The magnitude of the t-statistics showed that the average monthly income
in the HHi had greater impact to the model compared to other predictors in the equation (13).
The expenditures on yoghurt increases by 0.005043 euro/month given one additional euro in
the HH income. This estimation is true only if the respondent’s education and the HH size are
held constant. The fitted model produces a significant reduction in the residual variance, with
F = 12.616; P < 0.001.

Expenditures on fruit yoghurt

The estimated expenditure on fruit yoghurt in euro/month in the HHi was:

Ê Fruit yoghurt (euro/month) = - 3.167 + (0.303 Education) + (1.472 Children)


- 1.563 2.138* 4.996***
+ (0.003348 HH income)
5.066***
(R2 = 0.36) (14)

The fitted model in the equation (14) predicted 0.303 euro/month expenditure increase on
fruit yoghurt in the HHi, given a one year increase in the education level of respondent i,
holding number of children in the HHi and HH income constant. As it was expected the
number of children in the HHi had significant contribution to the fitted model. One child more
in the HHi increases expenditure on fruit yoghurt by 1.472 euro/month, holding respondent’s
education and HH income constant. The magnitude of the t-statistics indicates that the HH
income had the highest contribution to the fitted model. Monthly expenditure in the HHi on
fruit yoghurt is expected to increase by 0.003348, per each euro increase in the HH income,
holding number of children in the HHi and respondent’s education constant. Even though, the
variability on monthly fruit yoghurt expenditures explained by the predictors was 36%.
However, the fitted model produced a significant reduction in the residual variance, with F =
22.740; P < 0.001.

49
Expenditures on cream

The estimated expenditure on cream in euro/month in the HHi was:

Ê Cream (euro/month) = - 3.287 + (0.950 HH size) + (1.349 Employment)


- 2.957** 5.898*** 2.745**
(R2 = 0.41) (15)
The estimation results indicate that there was a positive relationship of HH monthly
expenditures on cream consumption with the HH size and number of employed family
members. An increase of the HH size by one member increases HH monthly expenditures on
cream consumption by 0.950 euro/month, holding the number of employed family members
constant. One more member employed in the HHi increases monthly expenditures on cream
consumption by 1.349 euro/month. This interpretation is true only if HH size is held constant.
The fitted model produces a significant reduction in the residual variance, with F = 48.979; P
< 0.001.

Expenditures on curd

The estimated expenditure on curd in euro/month in the HHi was:

Ê Curd (euro/month) = 1.065 + (0.05923 Age) + (0.290 HH size) + (0.797 Employment)


0.696 1.852 2.193* 1.990*
(R2 = 0.15) (16)

The fitted model in the equation (16) predicted 0.05923 euro/month increase on curd
expenditure as the respondent’s age increases by one year. This interpretation stands only if
the HH size and the number of employed family members are held constant. The expenditure
on curd increases by 0.290 euro/month if the HH size increases by one member, holding other
predictors in the equation (16) constant. The model predicts 0.797 euro/month increase on the
curd expenditures if the number of employed family members increases by one. This
interpretation is true only if the respondent’s age and the HH size are held constant. The fitted
model produces a significant reduction in the residual variance, with F =8.077; P < 0.001.

50
Expenditures on cheese

The estimated expenditure on cheese in euro/month in the HHi was:

Ê Cheese (euro/month) = 7.773 + (0.920 HH size) – (0.941 Children) + (0.005169 HH income)


6.273*** 4.175*** - 1.846 4.685***
(R2 = 0.24) (17)

The fitted model in the equation (17) predicted 0.920 euro/month increase on the cheese
expenditure if the HH size increases by one member, holding number of children in the HHi
and HH income constant. The expenditures on cheese decrease by 0.941 euro/month if the
number of children in the HHi increases by one. This interpretation is true only if the HH size
and the HH income are held constant. The model predicts 0.005169 euro/month increase on
the cheese expenditure if the HH income increases by one euro, holding HH size and the
number of children in the HHi constant. The fitted model produces a significant reduction in
the residual variance, with F = 18.792; P < 0.001.

Expenditures on butter

The estimated expenditure on butter in euro/month in the HHi was:

Ê Butter (euro/month) = - 2.600 + (0.09451 Age) + (1.973 Employment)


- 1.599 2.611* 6.227***
(R2 = 0.28) (18)

The model in the equation (18) predicts 0.09451euro/month increase on the butter
expenditure if the respondent’s age increases by one year, holding number of employed
family members in the HHi constant. The magnitude of the t-statistics indicates that the
number of employed family members had highly significant contribution to the fitted model.
The expenditure on butter increases by 1.973 euro/month if the number of employed family
members increases by one. This interpretation stands only if respondent’s age is held
constant. The fitted model produces a significant reduction in the residual variance, with F =
21.558; P < 0.001.

51
Total monthly expenditures on milk and other dairy products

The estimated total monthly expenditure on milk and other dairy products in euro/month in
the HHi was:

Tˆ = 12.220 + (0.291 Age) + (1.691 HH size) + (6.091 Employment) + (0.02090 HH income)


1.823 2.051* 2.944** 2.883** 4.757***
(R2 = 0.41) (19)

The fitted model in the equation (19) predicted 0.291 euro/month increase in total monthly
expenditures on milk and dairy products if the respondent’s age increases by one year,
holding HH size, number of employed family members and the HH income constant. The
total monthly expenditures on milk and other dairy products increase by 1.691 euro/month if
the HH size increases by one member. This interpretation stands only if other predictors in the
equation (19) are held constant. The increase number of employed family members increases
total monthly expenditures on milk and other dairy products by 6.091 euro/month, holding
other predictors in the equation (19) constant. The magnitude of the t-statistics indicates that
the HH income had the highest contribution to the fitted model. The model predicts 0.02090
euro/month increase in the total monthly expenditures on milk and other dairy products if the
HH income increases by one euro, holding other predictors in the equation (19) constant. The
fitted model produces a significant reduction in the residual variance, with F = 32.787; P <
0.001.

52
CHAPTER V

5. Conclusions

The results of this study indicate that milk and other dairy products stand as regular staple
food to most of the Kosovo households. Milk, yoghurt, cream, curd, and cheese were the main
dairy products consumed by majority of the Kosovo consumers. Men had much more buying
power compared to women. Men’s role in the decisions process for milk and other dairy
products was mainly as buyers. Women’s role in the decisions process was mainly as
influencers and deciders of the product choice. Supermarkets and grocery stores were the
most preferred market places by the consumers when buying milk and other dairy products.
The consumer’s preference towards market place was significantly dependent on
demographic and socioeconomic household characteristics (see Table 30). The frequency of
buying dairy products was associated with the durability of the product. Dairy products with
the shorter shelf life such as milk, yoghurt, and fruit yoghurt were bought more frequently by
the consumers compared to those with longer shelf life. Apart from product life, other
demographic and socioeconomic factors had significantly impact on the frequencies of buying
milk and other dairy products (see Table 31). The consumers’ demand for milk and other
dairy products was quite stable throughout the year. The consumption of milk and butter was
slightly higher during the winter months, while yoghurt and fruit yoghurt were more preferred
during the summer season.
The consumer’s preference towards the package size was positively associated with the
HH size. Kosovo consumers preferred bigger packages for curd and cheese. Smaller packages
were predominantly more preferred for fruit yoghurt and butter. The country of origin
influenced the consumers’ preferences towards locally produced dairy products. Majority of
the Kosovo consumers had favourable bias toward dairy products. Product attributes such as
quality, safety, taste, and price were the main features motivating consumers to purchase
domestic dairy products. According to the consumers’ view, other product features such as
package, durability, assortment, and the visual aspect were seen as disadvantage of the
domestic compared to the foreign dairy products. Respondent’s location had significant
impact on the preferences of domestic versus foreign dairy products. The consumer’s
preference towards new dairy products and innovation was dependent on his/her

53
characteristics such as age, and monthly income. Respondents with higher monthly income
were more willing to try new dairy products that come to the market.
The consumers’ attitudes toward product features such as nutritive content, taste, product
safety, price, brand, wrapping, package size, and the origin of product were significantly
dependent on demographic and socioeconomic factors (see Table 32). Among the product
attributes nutritive content, taste, product safety, price and the origin of dairy products were
most highly ranked by the Kosovo consumers. The study has proved that households with
lower monthly income and more than five family members were less able to meet their needs
for milk and other dairy products.
The location of the respondents (rural or urban) was an important variable of whether the
respondents buy milk and other dairy products. It was proved that rural households buy
significantly less milk, yoghurt, cream, curd, and cheese compared to the urban households.
Income was a significant variable of whether respondents buy fruit yoghurt, cream, and curd.
The odds ratio of buying yoghurt, cream, and curd increases as the respondent’s and
household’s income increase. Household size was a significant variable of whether
respondents buy fruit yoghurt and cream. The probability of buying these dairy products
decreases as the number of family members increases. It was also proved that number of
family members employed had a significant positive effect of whether respondents buy
cream.
Respondent’s monthly income and number of employed family members were significant
variables of whether respondents fulfil their demand for milk and other dairy products. The
odds ratio for fulfilling the demand for milk and other dairy products decreases as the HH size
increases. Respondent’s location and age were important variables of whether respondents
prefer domestic dairy products. Respondents from rural areas significantly preferred more
domestic dairy products. The probability of preferring domestic dairy products increases as
the respondent’s age increases. The odds ratio of preferring domestic dairy products decreases
as the HH size and HH income increases. Education was a positive determinant factor of
whether respondents preferred new dairy products and innovation.
The household income was a highly significant variable in explaining the variation on
milk, yoghurt, and fruit yoghurt consumption, suggesting a trend toward increased
consumption of such products as the household income rises. The household size was an
important variable almost for all dairy products, and quite significant for cream and cheese
consumption. The effect of number of children in the household was highly significant in
explaining the variation in milk and fruit yoghurt consumption. It suggests that an increased

54
number of children in the household will be a determining factor in the increasing demand for
milk and fruit yoghurt. The number of family members employed was a significant variable in
explaining the variation on curd, cheese, and butter consumption. The estimation predicts an
increase on the curd, cheese, and butter consumption as the employment rate increases.
Respondent’s age was significant for curd, cheese, and butter. Changes in the respondent’s
age significantly affect curd, cheese, and butter consumption. Income was a highly significant
variable in determining the household expenditures on milk, yoghurt, fruit yoghurt and
cheese. It suggests that increase of income will significantly increase expenditures on milk,
yoghurt, and fruit yoghurt. An increase in the household size significantly increases
household expenditures on yoghurt, cream, curd, and cheese.
The average quantity of milk consumed by the HHs was estimated to be 26.30 l/month.
The average quantity consumed by the HHs for the other dairy products was estimated as
follows: yoghurt 10.5 l/month, fruit yoghurt 2.51 kg/month, cream 2.34 kg/month, curd 3.93
kg/month, cheese 4.67 kg/month, and the butter 1.39 kg/month. The average of monthly
expenditures by the HHs on milk and other dairy products were estimated as follows:
expenditures on milk 19.37 euro/month, yoghurt 10.02 euro/month, fruit yoghurt 5.85
euro/month, cream 6.11 euro/month, curd 6.69 euro/month, cheese 14.69 euro/month, and the
butter 5.08 euro/month. The total average monthly expenditure by the HHs on milk and other
dairy products was estimated to be 56.17 euro/month or 9.6% of the total average household
monthly income.

55
Table 30. Test of independence between the market places and demographic and socioeconomic characteristics
Product Milk Yoghurt Fruit Cream Curd Cheese Butter
Yoghurt
Variable χ2 Cramer’s
χ2 Cramer’s
χ2 Cramer’s
χ2 Cramer’s
χ2 Cramer’s
χ2 Cramer’s
χ2 Cramer’s
V V V V V V V

Place 1.423 0.065 0.632 0.046 0.001 0.003 1.795 0.082 8.901 0.188 3.331 0.10 1.156 0.075
**
Region 17.720 0.115 19.479 0.126 8.082 0.205 20.139 0.159 37.1 0.191 23.663 0.134 25.49 0.177
Age 10.741 0.090 11.545 0.097 1.861 0.098 3.089 0.062 10.73 0.103 17.341 0.115 13.50 0.129
Education 17.805* 0.163 17.687* 0.17 7.902* 0.202 11.757 0.148 13.76 0.165 19.42* 0.171 15.77* 0.197
EmploymentR 0.806 0.049 1.905 0.079 1.289 0.082 2.445 0.096 5.718 0.150 6.350 0.139 1.081 0.073
IncomeR 12.76 0.112 16.992 0.136 5.389 0.167 20.82* 0.161 14.50 0.138 13.612 0.117 12.55 0.144
HH size 39.7*** 0.199 21.612* 0.154 1.573 0.090 8.307 0.102 19.53 0.160 13.282 0.116 30.3** 0.223
* *
Children 5.981 0.094 19.749 0.18 0.536 0.053 4.134 0.088 15.7 0.176 7.747 0.108 5.119 0.112
Employment 24.23** 0.190 21.44** 0.187 8.586* 0.211 14.18* 0.163 24.3** 0.219 28.7*** 0.208 15.62* 0.196
HH income 24.437 0.135 27.041* 0.149 18.42** 0.309 19.002 0.154 33.2** 0.181 41.8*** 0.178 30.30* 0.193
Note: Market place where consumer use to buy milk and other dairy products: by farmer; street vendor; green market; grocery store; supermarket; Cramer’s V measure the
strength of association between two categorical variables; Level of significance is denoted as follows: *P < 0.05; **P < 0.01; ***P < 0.001.

56
Table 31. Test of independence between the frequencies of buying milk, dairy products and demographic and socioeconomic characteristics
Product Milk Yoghurt Fruit Cream Curd Cheese Butter
Yoghurt
Variable χ2 Cramer’s
χ2 Cramer’s
χ2 Cramer’s
χ2 Cramer’s
χ2 Cramer’
χ2 Cramer’s
χ2 Cramer’s
V V V V sV V V
* *
Place 6.755 0.142 7.821 0.160 4.015 0.144 10.729 0.20 8.36 0.181 4.825 0.121 0.978 0.069
** * *** *** ***
Region 32.4 0.155 25.529 0.144 29.672 0.196 23.208 0.147 42 0.234 68.3 0.227 42 0.227
Age 13.506 0.100 8.196 0.082 14.068 0.135 13.593 0.113 15.55 0.143 23.511 0.133 27.7* 0.185
Education 12.163 0.135 14.523 0.154 8.571 0.149 4.422 0.091 11.22 0.148 13.542 0.143 7.020 0.131
EmploymentR 1.917 0.076 17.58** 0.240 3.088 0.126 2.482 0.096 7.268 0.169 6.881 0.144 0.839 0.064
IncomeR 15.067 0.122 21.244* 0.152 11.206 0.139 8.862 0.105 8.872 0.108 8.641 0.093 9.719 0.126
HH size 13.998 0.118 13.029 0.119 9.110 0.125 14.917 0.136 7.810 0.101 13.541 0.117 8.601 0.119
Children 6.189 0.096 6.786 0.105 4.356 0.106 4.026 0.087 5.922 0.108 15.207 0.152 13.60 0.183
Employment 15.030 0.150 3.875 0.080 11.697 0.174 4.691 0.094 6.822 0.116 13.683 0.144 4.785 0.108
HH income 32.8** 0.156 20.678 0.130 22.833 0.172 9.839 0.096 9.643 0.112 23.792 0.134 12.74 0.125
Note: Frequencies were given as follows: every day; twice a week; once a week; twice a month; once a month. Level of significance is denoted as follows: *P < 0.05; **P <
0.01; ***P < 0.001.

57
Table 32. Test of independence between evaluation of product attributes and demographic and socioeconomic characteristics
Product 1 2 3 4 5 6 7 8
attribute
Variable χ2 Cramer’s
χ2 Cramer
χ2 Cramer’s
χ2 Cramer’
χ2 Cramer’s
χ2 Cramer’s
χ2 Cramer’s V
χ2 Cramer’s
V ’s V V sV V V V
**
Place 5.06 0.115 0.048 0.011 0.46 0.035 4.325 0.106 14 0.188 7.083 0.136 4.451 0.108 6.19 0.127
Region 14.0 0.110 4.355 0.137 7.88 0.101 29.8** 0.161 35*** 0.174 30.7** 0.163 41.5*** 0.190 10.0 0.093
Age 17.9 0.125 12.29 0.127 6.31 0.091 8.382 0.085 17.3 0.123 7.457 0.080 12.309 0.103 16.5 0.12
Education 13.6* 0.133 4.677 0.078 9.23 0.109 13.0* 0.130 12.8* 0.129 10.80 0.118 16.5* 0.146 23** 0.174
EmploymentR 2.26 0.077 7.155* 0.136 3.40 0.094 12.3** 0.179 13** 0.184 4.115 0.103 0.715 0.043 2.83 0.086
IncomeR 5.010 0.066 5.435 0.084 2.12 0.052 61*** 0.231 44*** 0.197 30.8*** 0.163 14.572 0.112 13.4 0.108
HH size 7.684 0.082 3.432 0.067 8.33 0.104 0.240 0.094 8.82 0.087 3.586 0.056 4.448 0.062 16.2 0.12
Children 5.211 0.082 5.829 0.087 5.51 0.085 2.558 0.058 1.22 0.040 7.456 0.098 3.864 0.071 8.92 0.108
Employment 5.198 0.082 6.840 0.094 1.35 0.042 19.1** 0.161 8.14 0.103 11.345 0.121 5.147 0.082 4.91 0.80
HH income 12.74 0.105 7.991 0.102 14.2 0.136 88*** 0.276 36*** 0.177 36.0*** 0.177 13.468 0.108 24.4* 0.145
Note: The variable was measured in order of importance, on scale of 1 to 4, where 1 is “unimportant”, 2 is “not very important”, 3 is “important”, 4 is “ very important”. Numbers stand
for 1= Nutritive content; 2 = Taste; 3 = Product safety; 4 = Prices; 5 = Brand; 6 = Wrapping; 7 = Package size; 8 = Product origin. Level of significance is denoted as follows:
*
P< 0.05; **P< 0.01; ***P < 0.001.

58
LIST OF REFERENCES

Agresti, A. (2007) An Introduction to Categorical Data Analysis. 2nd ed. New Jersey, John
Wiley & Sons, Inc.

Andy, F. ( 2005) Discovering Statistics Using SPSS. 2nd ed. London, Thousand Oaks, New
Delhi, University of Sussex, SAGE Publications Ltd.

Bryman, A. (2004) Social Research Methods. 2nd ed. New York, Oxford University Press.

Chambers, R. (1983) Putting the Last First. United States of America, Longman Inc.

Fock, K. (2007) Agriculture Public Expenditure Review. Draft 2. Kosovo.

Glantz, S. and Slinker, B. (2001) Applied Regression & Analysis of Variance. 2nd ed. United
States of America, McGraw-Hill, Inc.

Herzberg, F. (1966) Work and the Nature of Man (Cleveland: William Collins). Thierry and

Koopman-Iwerna, Motivation and Satisfaction, pp. 141-142. ‘Quoted in:’ Kotler, Ph. and

Keller, K. (2006) Marketing Management. 12th ed. Upper Saddle River, New Jersey, 07458,

Pearson Education, Inc.

Kapsdorferová, Z. and Nagyová, Ľ. (2005) Consumer behaviour at the Slovak dairy market.

Nitra, Slovak Republic, Slovak University of Agriculture. AGRIC. ECON. – CZECH, 51,

2005 (8): 362-368.

KCBS (2008) Dairy Market Assessment Study. Contract No. AFP-I-00-03-00030-00, TO


#800. Kosovo Cluster and Business Support Project Kosovo, USAID.

59
Kotler, Ph. (2002) Marketing Management, and Millennium Edition. United States of

America, University of Phoenix.

Kotler, Ph. and Keller, K. (2006) Marketing Management. 12th ed. Upper Saddle River, New

Jersey, 07458, Pearson Education, Inc.

MAFRD, (2003) Kosovo Green Book- Agriculture and Rural Sustainable Development
Strategy in Kosovo (in Albanian). Kosovo, Ministry of Agriculture Forestry and Rural
Development.

Maslow, A. (1954) Motivation and Personality. New York, Harper and Row, pp. 80-106.

‘Quoted in:’ Kotler, Ph. and Keller, K. (2006) Marketing Management. 12th ed. Upper Saddle

River, New Jersey, 07458, Pearson Education, Inc.

Ministry of Agriculture (2006) Agriculture and Rural Development Plan 2007-2013. Chapter
1, 2. The Rural Development Context of Kosovo. Kosovo.

Ministry of Public Services. (2006) Agriculture Household Survey 2005. Series 2. Agriculture
and Environment Statistics. Kosovo, SOK.

Ministry of Public Services. (2007) Living Standard Statistics. Kosovo, SOK.

Ministry of Public Services. (2009a) Women and Men in Kosovo. Kosovo, SOK.

Ministry of Public Services. (2009b) Kosovo in Figures 2008. Kosovo, SOK.

Nushi, M. and Selimi, F. (2009) An Assessment of the Competitiveness of the Dairy Food
Chain in Kosovo. Study 1. AgriPolicy Enlargement Network for Agripolicy Analysis.
Kosovo.

60
Oldham, P. Bajraktari, E. and Wittkowsky, A. (2006) Study of Competitiveness with Imports.
The Kosovo Dairy Sector. Economic Policy Office EU Pillar – UNMIK Kosovo.

Tour, L. and Henthorne, T. (1995) Perception of Marital Roles in Purchase-Decision

Processes: A Cross-Cultural Study. Journal of the Academy of Marketing Science (Spring

1995): 120-131. ‘Quoted in:’Kotler, Ph. and Keller, K. (2006) Marketing Management. 12th

ed. Upper Saddle River, New Jersey, 07458, Pearson Education, Inc.

World Bank. (2007) Kosovo Poverty Assessment. Volume 1. Accelerating Inclusive Growth
to Reduce Widespread Poverty. Poverty Reduction and Economic Management Unit, Europe
and Central Asia Region. Report No. 3973-XK.

61
APPENDIX

Appendix A1 / Questionnaire

No. of questionnaire ______ Date ______________ Phone no.______________


Place: _______________ Rural ____ Urban ____ Region_________________

Basic information about respondent

Gender F_____ M______


Age ______
Formal education (in years) ______ Profession ________________
Employed Yes______ No_____
Average monthly income (per respondent) _________€
Number of family members_______
Number of children aged 14 and younger ________
Employed number of family members_______
Average family income per month ________ €

1. Do you or your family members buy milk or other dairy products?


Response/Product Fluid Yoghurt Fruit Cream Curd Cheese Butter
milk yoghurt
Yes
No

If no, due to ________________________________________

If yes,
2. Who usually buys milk and other dairy products in your family?
Man_____ Woman_____ Man & Woman_____ Children_____

3. Who decides what kind of dairy products to buy?


Man____ Woman ______ Man & Woman ______ Children _____

4. Where do you usually buy milk and dairy products given in table below?
Location/Product Fluid Yoghurt Fruit Cream Curd Cheese Butter
milk yoghurt
By farmer
Street vendor
Green market
Grocery store
Supermarket
Other

62
5. How often do you usually buy the following products?
Frequencies/Product Fluid Yoghurt Fruit Cream Curd Cheese Butter
milk yoghurt
Every day
Twice a week
Once a week
Twice a month
Once a month

6. What is the average of quantity purchased and expenditures within a month for the
following products?

Quantity in Kg, l, Fluid Yoghurt Fruit Cream Curd Cheese Butter


and expenditures in € / milk yoghurt
Product
Monthly quantity purchased
per product
Monthly expenditures per
product

7. Total monthly expenditure on milk and dairy products (within family) _______ €

8. Are you fulfilling demand for milk and other dairy products with monthly income
available?

Yes ______ No_______

9. Do you personally consume the following products?

Response/product Fluid Yoghurt Fruit Cream Curd Cheese Butter


milk yoghurt
Yes
No

10. If no, due to:


Motive/product Fluid Yoghurt Fruit Cream Curd Cheese Butter
milk yoghurt
Do not like the taste
Can not afford
Due to allergy
Other

63
11. Regarding seasonality, when do you consume less/ on a regular basis/ or more the
following products?
Rank from 1 to 3: by 1 when you consume less, by 2 when you consume on a regular basis
and by 3 when you consume more.
Season/Product Fluid Yoghurt Fruit Cream Curd Cheese Butter
milk yoghurt
Spring
Summer
Autumn
Winter

12. How do you appraise attributes given below to milk and other dairy products?
Attribute/Appraisal Very important Important Not very important Unimportant
Nutritive content
Taste
Product safety
Price
Brand
Wrapping
Package size
Product origin
Other

13. What is your preference for the package size for the following products?

Package size / Fluid Yoghurt Fruit Cream Curd Cheese Butter


Product milk yoghurt
Package size in l/gr

14. When you buy milk and other dairy products, how do you give attention on: (rank from 1
to 5, with 1 when you do not give attention to 5 when you give the highest attention)?

a) Expired date _____


b) Product content (in package) _____
c) Producer name_____
d) Origin of the product______
e) Other______

15. As a consumer of the dairy products, do you prefer and buy more?
Domestic dairy products_____
Foreign dairy products_____

64
16. What are the main reasons preferring/not preferring domestic/imported products?
Origin of the product Domestic Imported
Preferences Yes / No Yes / No
(< Quality) or (> Quality)
(<Safety) or (> Safety)
(< Price) or (> Price)
(< Packing) or (> Packing)
(< Taste) or (> Taste)
(< Durability) or (> Durability)
(< Assortment) or (> Assortment)
(<Visual aspect) or ( > Visual aspect)
(<Advertisement) or( >Advertisement)
Other

17. How do you usually get information about dairy products (rank by 1 to 4, with 1 for
source most rarely used, 2 for source rarely used, 3 source often used and with 4 for source
most often used?
Newspapers____
Flyers_____
TV_______
Radio_____
Other_____

18. Do you prefer new dairy products and innovation?

Yes______ No_____

19. During the last year, did you start buying any new dairy product which you did not buy it
before?

Yes_____ No_____

If yes,

20. Which of the dairy products: __________________________________________

21. What did you find attractive about this product?


____________________________________________________________________

22. How did you get information about this new product?
_____________________________________________________________________

65
Appendix A2 / Tables

Table 1. The targeted stratification of the sample size


Region / Age 18 – 30 31 – 40 41 - 50 51 - 60 Over 60 Total
Prishtina 16F, 16M 11F, 11M 10F, 10M 8F, 8M 3F, 4M 48F, 49M
Prizreni 13F, 13M 9F, 9M 7F, 8M 6F, 6M 3F, 3M 38F, 39M
Mitrovica 9F, 9M 7F, 7M 6F, 6M 4F, 4M 2F, 2M 28F, 28M
Peja 13F, 13M 10F, 10M 8F, 8M 6F, 7M 3F, 3M 40F, 41M
Gjilani 12F, 12M 9F, 9M 7F, 7M 6F, 6M 3F, 3M 37F, 37M
Total 63F, 63M 46F, 46M 38F, 39M 30F,31M 14F,15M n=385
Note: Given characters F and M denote respondent’s gender: F for female and M for male

Table 2. The feasibility stratification of the sample size


Region / Age 18 – 30 31 – 40 41 - 50 51 - 60 Over 60 Total
Prishtina 16F,16M 11F,11M 10F,9M 8F,6M 3F,6M 48F,48M
Prizreni 12F,15M 10F,13M 6F,8M 3F,4M 3F,3M 34F,43M
Mitrovica 9F, 9M 6F, 7M 8F,6M 4F,3M 3F,2M 30F,27M
Peja 14F,13M 11F, 9M 8F,8M 5F,7M 4F,2M 42F,39M
Gjilani 14F,11M 8F, 10M 8F,6M 5F,6M 3F,3M 38F,36M
Total 65F,64M 46F,50M 40F,37M 25F,26M 16F,16M n=385

66

You might also like