Description: This study did an evaluation of the socio-economic spillovers of the introduction of a fruit processing factory in Tzaneen. The data file contains responses from the suppliers to the NDC fruit processors factory. The data contains information regarding the supplier's demographics, business activity, job creation, source of merchandise and the market environment.

The data set contains 62 records and 66 variables. Abstract: This study did an evaluation of the socio-economic spillovers of the introduction of a fruit processing factory in Tzaneen. The factory was setup to provide alternative income streams to local entrepreneurs and the community in order to reduce poverty and increase overall well-being of the local population.

The evaluation produced two data sets, one based on suppliers' experiences and the other based on households' experiences. The data in this data set is the suppliers' data only. It should be used together with the household data which will be made available in due course in order to get a clearer picture of the total spillover effects that could have materialised as a result of the operation of the factory.

In depth interviews were conducted with local suppliers of fruit to the factory. The major questions asked in the study were:

To what extent and in what ways did local entrepreneurs benefit from the arrival of the agro-processing project?

What were the employment effects in the community?

How can long run income diversification be achieved in the community?

The major variables in the data set are: Changes in income streams, sources of merchandise, payment procedures, changes in personal circumstances and sources of income.

Face-to-face interview

The universe is all the suppliers to the NDC factory since it started operating in 2011. The sampling frame was the suppliers to the factory who were active between January 2013 and December 2015.

The strategy was to interview every supplier who could be reached on the listed contact numbers in the register. The retention pattern of telephone contacts in a case like this, over a two year period, is unpredictable and totally random. So the realised sample can be deemed to be totally random, although it may not necessarily be representative of the whole sample.