Methodologies for analysis of household survey data
MetadataShow full item record
Permanent link to this item: https://cgspace.cgiar.org/handle/10568/2874
Internet URL: http://mahider.ilri.org/handle/10568/3140
Characterisation of a benchmark site is a research activity, therefore the scientific method has to be applied. Within this context, Rey et al (1993) suggested seven general steps for the characterisation of market-oriented dairy systems in sub-Saharan Africa which are also applicable to the crop-animal systems practised in Southeast Asia. Most of these steps are discussed in detail in this paper, but the emphasis of the paper is on the methodologies used to analyse household survey data. Some examples of hypotheses that can be formulated when studying crop-animal systems are outlined. These include animal component; crop-animal interactions; supply, processing and marketing; socioeconomic aspects; gender issues; test of hypotheses with categorical data; test of hypotheses with continuous data; multiple and partial regression and correlation analysis; multivariate analysis; cluster analysis; and descriminant analysis.