Good Practices in Extension Research and Evaluation

  • 18th August 2017
  • by secretary


Agricultural Extension in South Asia Network (AESA)
278 pages

This manual was developed as a hands-on reference tool to help young researchers, research students, and field extension workers in choosing the right research methods for conducting quality research and evaluation in extension.
Extension research is a unique social science inquiry where research ideas are gathered from the field problems and put through a systematic cycle of objective investigations that result in significant solutions. Apart from developing theories and models that advance scientific knowledge, extension research should also provide new insights for improving extension policy and practice.
A Workshop was organised on ‘Good Practices in Extension Research and Evaluation’ at the ICARNational Academy of Agricultural Research Management (NAARM), during 29 November-2 December 2016, for young extension researchers and PhD students in extension and this Manual is the outcome of this workshop. 
Re-orienting extension research is urgent, and calls for a coordinated approach by integrating state-of-the-art methods from other sciences in order to improve the utility and visibility of the extension research outcomes. Adopting several good practices, such as the following, can enhance the quality of extension research: 
  • Creative generation of relevant research ideas using an intuitive/common sense approach; • Selection of a rigorous and robust research design; 
  • Choice of right variables following alternate criterion-referenced validity assessment procedures; • Selection of appropriate sample sizes to maximise generalisability; 
  • Estimation of reliability and validity through robust modelling procedures, such as Structural Equation Modelling; 
  • Deployment of resource and time saving but accurate tools, such as shortened paper surveys and e-surveys; 
  • Compensation of respondents so as to maximise the accuracy of responses; 
  • Data cleaning by employing missing value estimation and assumption testing tools, and multivariate data modelling