2.1. Conceptual framework
2.2. Sample design
2.3. Data collection techniques
2.4. Data analysis
Marketing studies are primarily geared toward identifying areas for improvement in marketing in order to foster economic development. A study of dairy marketing should be viewed against this basic philosophy. The objective of the Ethiopian dairy marketing study was to evaluate the relative efficiency of the alternative systems through which dairy products are marketed in Addis Ababa and its immediate environs (i.e. the catchment area for fresh milk within a 130 km radius of the City of Addis Ababa) and draw relevant policy implications. In this study, alternative marketing systems are defined in terms of the different marketing chains for a given type of dairy product. Such marketing chains are normally described as marketing channels, but the term "marketing chains" is preferred in this study as it is more appropriate in view of its implication that there are linkages within what is ordinarily described as a marketing channel.
Marketing is a productive activity whose output may be defined in terms of either the objectives set externally (e.g. by government) or the utilities (i.e. satisfaction) afforded to users of the particular marketing system. Hence the output of a marketing system can be evaluated in terms of the subjective satisfaction derived by external arbiters or users from the attainment of their objectives or utilities. Since a marketing system also requires some inputs, such as labour, transport and other types of facilities, a holistic concept of marketing efficiency ideally relates to the optimisation of the output in relation to the inputs. In practice, such holistic optimisation procedures in the evaluation of marketing efficiency are difficult to operationalize and partial efficiency criteria are often substituted. One such partial efficiency criterion is least cost in accomplishing a specified task, or the ability to offer services at the lowest price while recovering costs. According to Abbott and Makeham (1979), a marketing system will be judged to be efficient by this criterion if it operates at the lowest cost which is feasible with the available techniques, skills and knowledge and such efficiency will be reflected in both the consumer price and the quality of the goods and services provided. Other partial efficiency criteria relate to particular aspects of goods and services, such as provision of particular services that are desired by the consumers (e.g. home delivery services) and the attainment of particular standards of product quality and purity or hygiene. Theoretically, it is possible to construct an all-embracing index of efficiency, but such an index would be difficult to operationalize.
Ideally, a dairy marketing study would require a close scrutiny of one or more of the following aspects:
i) marketing options of all the different kinds of producers;ii) marketing activities of all the different kinds of middlemen, including the dairy parastatal organizations, such as the Dairy Development Enterprise (DDE) in the case of Addis Ababa;
iii) purchasing patterns of all the different kinds of consumers, including households arid institutions.
Given time and other resource constraints, it appeared plausible that the objective of this study could be achieved by examining the dairy marketing systems in Addis Ababa and its immediate environs from the retail end of the systems, and by giving an emphasis to the study of milk purchases patterns of the consuming households in Addis Ababa. However, the study also examined the purchases patterns for the other dairy products, especially butter. There is little doubt that a study of the consuming households can give answers to the major questions raised in relation to the relative efficiency of alternative marketing systems for dairy products in terms of (i) consumer price levels; (ii) the types, reliability, and stability or flexibility of the services offered to the consumers; (iii) the responsiveness of alternative marketing systems to the consumer demands; and (iv) the contribution of the alternative marketing systems to the maintenance or achievement of some specified goals of marketing policies, such as product quality and hygiene standards. The latter criterion would involve the assessment of the suitability of alternative marketing systems for different sets of circumstances.
In this study, the alternative marketing systems for dairy products through which consuming households in Addis Ababa procure their dairy products were identified on the basis of the types of the final links in the different marketing chains from which the consuming households purchased their products, i.e. the retail sources. These kinds of final links include individuals and private and government commercial concerns. The systems catering for the marketing of fresh liquid milk were treated as the main alternative dairy marketing systems available to Addis Ababa households. However, the systems catering for butter were also examined separately in some detail.
For convenience and clarity, the details of the approaches taken in the development of specific indicators of relative efficiency of alternative marketing systems are presented as part of the presentation and discussions of the results of the study in the subsequent sections of the report.
For the purpose of this study, Addis Ababa was mapped out and divided into three concentric zones, differentiated as the Inner City Zone (ICZ), the Middle City Zone (MCZ) and the Peripheral City Zone (PCZ) on the basis of the distance from the conventional location of the city centre. The expectation in this case was that the location of the consumers relative to the centre of a given consumer area (such as Addis Ababa) would influence the types of marketing chains through which a given product is sold.
For the purposes of official statistical data collection, Addis Ababa is mapped into several Enumeration Areas (EA's) by the Central Statistical Office of the Government of Ethiopia. Each EA covers between 150 and 200 households on average. An EA was adopted as the basic sampling unit in this study, and 25 EA's were randomly selected for the purposes of the survey. The 25 EA's were allocated to the three study zones (i.e. ICZ, MCZ and PCZ) with probability proportional to the distribution of the EA's in the three zones. After listing the monthly cash income of each household in the 25 EA's which were included in the sample. Three broad sets of data were collected:
i) exploratory survey data set;
ii) initial survey data set;
iii) subsequent survey data set.
The exploratory survey of the consuming households involved a one-time interview of each household in the random sample of 25 EA's which were included in the study. This survey was undertaken primarily as the basis for generating the data to be used in further sample selection procedures. Under this survey, a sample of 3,888 households in Addis Ababa (which represent close to 2% of all the households in the city) were interviewed in order to obtain information on their sources and levels of monthly cash income and also to determine the origin of the milk for those households who had acquired some milk any day during the course of the week prior to the interview day. The origin of the milk was defined in terms of whether the milk was produced by the households, purchased or obtained as gift. Purchases of the other dairy products, if any, were also recorded.
The initial survey of the consuming households involved a one-time interview of each household and was conducted on a purposively-selected stratified random sample of 482 households. A random sample of 20 households was purposively selected from each EA for the study using the following criteria as the basis for sampling as far as possible:
i) 16 households should belong to those household income groups whose average cash income was at least 250 Ethiopia birr1 per month;ii) 4 households should belong to those household income groups whose average cash income was less that 250 Ethiopian birr per month, but who had purchased some milk at least once during the week prior to the date when the households' monthly cash incomes were listed.
1 One Ethiopian birr was approximately equivalent to 0.5 United States dollar at the time the survey was conducted in 1984-85.
This distribution was based on findings of the exploratory survey. In case there were difficulties in obtaining and selecting the required number of households within an EA according to the above criteria, compensation was made by allocating proportionally the number of households that a given EA fell short of its quota among the other EA's within the same zone. The above sampling criteria were justified on the grounds that only those households who could provide information that could be used to ascertain the marketing chains for dairy products in Addis Ababa were relevant to the study.
The subsequent survey of the selected consuming households involved only those households covered in the initial survey who consume milk. Each of such milk-consuming households was interviewed consecutively for seven days in order to determine the major factors that influence the households to deviate from their long-term purchases and consumption patterns for milk and other dairy products. Such deviations were defined primarily in terms of changes in the sources and in the geographical locations of the places from which the households purchase and collect their dairy products, following a prior determination during the initial survey of the types of sources and geographical locations that are most frequently used by a particular household. The post-enumeration interview of each household included in the subsequent survey sample thus constituted the eighth interview of the household. This post-enumeration interview was meant to be a check on the consistency of the data collected by the enumerators and was thus not used as a source of data in subsequent analyses.
Background information on the dairy industry in Ethiopia and the marketing of milk, especially in Addis Ababa and its immediate environs, was obtained from secondary data sources. Primary data were collected through personal interviews using structured questionnaires. Errors in data collection were minimized through the use of carefully trained enumerators and supervisors, and through the retention of their services throughout the course of the field data collection. Five supervisors and 16 enumerators were recruited and trained in interviewing techniques, with emphasis on data collection and handling techniques. The first 4 supervisors were each assigned to supervise 4 enumerators. The 5th supervisor was designated as post-enumeration supervisor, and his role was to conduct a post-enumeration interview on each household that had been interviewed by an enumerator during the course of the "subsequent survey'' (to be explained later) in order to check and correct for any inconsistencies in the data collected by the enumerators. Variance in observations was minimized through the collection of data from stratified or purposively selected random samples, for instance through the sampling of households by city zones and income levels, or through the sampling of households from a class that was most likely to purchase milk as described above.
Data were entered into the HP 3000 computer system and verified using a computer package called DEVAPS. The DEVAPS package has entry, verify, transact and extract programmes. The data entry and verification format was built using the transact programme. The DEVAPS package also has the capacity to automatically duplicate the data which has previously been entered without requiring it to be edited. The package, therefore, is advantageous when entering data where some of the information occurs more than once and when information does not vary by interview days. Data editing through the verification programme facilitated the detection of any errors in data entry.
The data were analysed primarily by using the SPSS (Statistical Package for Social Sciences) programme. The actual details of the types of analysis undertaken at each stage in the evaluation of the merits and, hence, the relative efficiency of the alternative marketing systems for milk and milk products are presented in the relevant sections of the report under the sections dealing with the general presentation and discussion of the results.
