A general guide to data preparation for computer analysis of farm survey data

Abstract

If a computer and an appropriate program for survey data analysis are available, the researcher has to decide whether or not these computer facilities should be used to analyse the survey data or whether an analysis by hand is sufficient. In almost all cases a preliminary analysis by hand is very useful to get a "feel" for the data. In our experience a complete manual analysis will be more efficient if the number of farmers in the sample is less than 5O. If the sample size is about 50-100 farmers, a computer analysis may be helpful when the study area is rather complex and farmers' practices and circumstances need to be analised through a relatively large number of cross tabulations. Surveys with more than 100 farmers can nearly always be analysed more efficiently by a computer if appropriate computational facilities are available. This note provides a short overview on presently used computers ("hardware") and computer programs ("softwere") assuming that the researcher has no previous experience with computers. It also presents the steps involved in preparing the data for computer analysis.