2. REVIEW METHODOLOGY

Location and staff

The call for studies by interdisciplinary groups is a frequent one, and this review is an example. The authors’ experience in Wageningen has taught them that an important prerequisite of interdisciplinary research is that those involved should spend a considerable amount of time and effort developing mutual understanding and insight into each others’ field of knowledge. The Wageningen review team has participated in a series of discussion workshops on food security over the last three years, and these exchanges have led to the development of a common base. These interdisciplinary activities have, moreover, taken place in the context of the ongoing debate on development within WUR, in which issues of food security and poverty are central.

An interdisciplinary review group was created at WUR. Five senior scientists planned the review, provided guidance when the individual country reports were being reviewed, and also analysed them. In cooperation with FIVIMS Secretariat staff at the Economics and Social Department (ES) of FAO, the WUR team carried out the review study of selected CCA reports and PRSPs. This involved reviewing 50 CCA reports and 25 PRSPs through a predefined framework for analysis. Four junior researchers carried out the analysis of the individual country studies, and they also participated in the drafting of this report.

The entire group met twice a week to discuss the review process and to define tasks for individual group members. Appendix II provides the names and disciplines of the entire review team.

Country selection

Selection procedure

The countries were selected for CCA report and PRSP review based on the status of completion of CCA reports and PRSPs as of 1 September 2002 (see Table 1).

Table 1. Status of completion of CCA reports and PRSPs as of 1 September 2002, disaggregated for five regions

Notes:

¹ List used for selecting the countries for this study.

² The status of interim PRSPs differs from country to country, depending on their purpose. Some are draft documents, while others resemble project activities that need to be carried out to develop a PRSP.

³ Total number of countries in CCA report and PRSP listing.

As of 1 September 2002, 106 countries had completed one or two CCA reports, with 12 in progress, and 17 countries had completed a PRSP and 30 countries had delivered an interim PRSP. For the present CCA/PRSP review, countries were selected on the basis of proportional sampling and stratified according to five regions (Africa; Asia and the Pacific; West Asia and North Africa; Latin America and the Caribbean; Europe and Commonwealth of Independent States - CIS). In total, 50 countries were selected for the CCA review, corresponding to 38 percent of all countries. For each region, 38 percent of the number of countries in that region was taken to determine the number of CCA reports to be reviewed. Likewise, 25 countries were selected for the PRSP review, corresponding to 19 percent of all countries, and, for each region, a sample of 19 percent of the countries in that region was taken to determine the number of PRSPs to be reviewed (see Table 2).

Table 2. Number of countries selected for the review, per region

Countries were randomly selected. In a first step, countries were selected for CCA review. If a selected country had also completed a PRSP (interim or complete), the PRSP of this country was selected for the review. Where the sample of countries selected for CCA review did not result in the required number of countries that had also completed a PRSP (interim or complete), random selection continued until the required number of countries for PRSP review was reached. Additional countries selected for review of their PRSP were not included in the CCA review. Where the selected countries for CCA review outnumbered the number required for PRSP review, the ‘first come, first served’ principle was used for the PRSPs.

Latin America and the Caribbean, Asia and the Pacific, and West Asia and North Africa did not have enough countries with completed or interim PRSPs. Therefore, two additional countries in Africa and one in Europe and CIS were randomly selected and included. Appendix III lists the countries selected for CCA and PRSP review.

Characteristics of countries selected for review

The selection comprised a large variety of countries. Of the CCA countries and PRSP countries, 36 percent and 44 percent, respectively, were classified as least developed countries (LDCs), and most of these were from the African region (see Appendix IV). The sample comprised both small and large countries: for example, the area of CCA countries varied from 1 800 km² (Mauritius) to 9 597 000 km² (People’s Republic of China), and for PRSP countries from 26 000 km² (Rwanda) to 2 345 000 km² (Democratic Republic of the Congo). Likewise, the population sizes for CCA countries varied from 400 000 (Suriname) to 1 253.8 million (People’s Republic of China), and for PRSP countries from 800 000 (Guyana) to 76.1 million (Viet Nam).

Also, the proportion of undernourished people varied enormously (see Appendix III). In CCA countries, the percentage of undernourished people varied from 1 percent in Bulgaria to 66 percent in Burundi, whereas for PRSP countries the percentage varied from 10 percent in Kyrgyzstan and Moldova to 64 percent in the Democratic Republic of Congo. However, note that for 16 percent of the CCA countries and 8 percent of the PRSP countries, the percentage of undernourished people was not known. Table 3 shows that in the CCA and PRSP samples, most of the countries where more than 35 percent of the population was undernourished were situated in Africa and, to a lesser extent, in Asia and the Pacific.

Table 3. Number of countries in prevalence categories of undernourished people in 1998-2000 per region

Notes:

¹ Data missing for 1 CCA country.
² Data missing for 2 CCA countries.
³ Data missing for 2 CCA countries and 1 PRSP country.
⁴ Data missing for 3 CCA countries and 2 PRSP countries.

The percentage of moderately and severely underweight^[2] children (0-59 months old) in CCA countries varied from 0.8 percent in Chile to 47 percent in India, and in PRSP countries from 3 percent in Georgia to 46 percent in Yemen (although for 18 percent of the CCA countries and 3 percent of the PRSP countries, the percentage of moderately and severe underweight children was not known). Child anthropometric indicators do show empirically a broad relation to calorie availability, but the relation is probably not linear (Mason, 2002). When child growth is adequate, then the calorie availability is likely to be adequate as well, but the opposite may not apply - poor growth can be due to other factors (Shetty, 2002).

Given the large variation of countries in the sample, it can be stated that the results of the review of the sampled countries provide a good impression of the status of both the CCA reports and the (interim) PRSP reports in all regions in terms of the attention given to food insecurity and vulnerability.

Approach to data collection

The TOR checklist of 20 items (Appendix I) formed the starting point for the development of a predefined framework for analysis of the selected country reports. The topics were disaggregated into several questions or groups of questions. These could be answered either by yes/no, a numerical code representing an answer from a list of possible answers, or by a textual explanation in case no predetermined answer was considered possible. Because the review also included PRSPs that do not directly emphasize food security, a number of questions regarding poverty were also included. The resulting 255 questions that formed the predefined framework for analysis were grouped in seven domains.

1. General (regarding the analysis itself and country details).

2. Report preparation details (questions regarding the process of preparing the reports).

3. Definitions (questions probing the definitions used for food security and for poverty).

4. Policy statements (questions on the position of food insecurity and vulnerability within the quoted country policies).

5. Data collection (an elaborate set of questions regarding indicators, sources, visualization (maps), disaggregation, representativeness, and temporal dimensions of the data stated).

6. Analysis (an elaborate set of questions regarding approaches and methodologies applied, and causes and trends identified in the reports).

7. Policies, strategies and interventions (questions regarding the transformation of the analysis into policies, strategies and interventions).

The first two domains refer to general data on the country and the report preparation process. The subsequent five domains reflect the cyclical process of policy analysis and policy implementation at country level (Ellis, 1992) (see Figure 1).

Figure 1 The cyclical process of policy analysis and policy implementation concerning food insecurity and vulnerability

The first step in this process is the identification (or not) of food security issues as a national problem. If food security is a national problem, then it needs to be ranked among other issues in development priorities. Thereafter follows a process of data collection and analysis that finally results in specific policies with strategies and interventions to tackle the food security problem. The process is cyclical and the results of interventions, and the development process in general, will determine the priority setting and the priority specification that will guide the analytical framework and action in the next cycle.

The development of the predefined framework for analysis took place in stages: preparation of sets of questions; testing on a number of country studies; discussion of the results and the revisions required; new tests on other country studies, followed by new discussions. This took place in the first two months of the operational period, and towards the end of this period amendments to the framework became less and less substantial. In this process, the “conceptual framework for understanding possible causes of low food consumption and poor nutritional status” (see Appendix V), as adopted by the Committee on World Food Security, was used for refining the framework regarding causes, strategies and interventions. FAO staff also made amendments. The final form of the predefined framework for analysis can be found in Appendix VI.

The framework was developed as a spreadsheet (Excel), allowing answers to be entered into one column, with easy transfer of the data set of a country report to one central database. The numerical answers facilitated data processing and subsequent statistical analysis through specialized programmes. The relationship between the individual questions and the checklist of 20 items in the TOR was maintained through direct references in the predefined framework for analysis.

Analysis

The analysis was performed in two stages. First, an analysis was made of the major issues as identified in the TOR, based on the individual questions of the predefined framework for analysis. The resulting comparison of the data from the 75 country studies enabled a detailed discussion of the breadth and depth of food insecurity and vulnerability as dealt with in the country reports. Most of this information is given in Chapter 3.

Second, an analysis of selected domains per country report was performed. Three domains were selected for this analysis: (i) from the report preparation process to the definition of food insecurity and vulnerability, and poverty; (ii) aspects of the analysis of food insecurity and vulnerability, and poverty; and (iii) the link between the analysis and interventions. This analysis sought to answer questions about the consistency with which the domains on food (in)security, vulnerability and poverty were dealt with within the country reports. The results of this analysis are discussed in Chapter 4.