Effective policymaking require rigorous evidence to be generated to guide the design and implementation of interventions that support social and economic wellbeing of citizens. However, challenges of undertaking controlled experiments of the type conducted by natural scientists possess a major hurdle for those concerned with generating social or economic evidence. Nonetheless, the field of econometrics has provided a useful approach for drawing inferences through analysis of non-experimental and non-biophysical experimental data. Econometrics translates data into models that assist to make forecasts to support decision making in a wide array of social and economic fields of agriculture, natural resource, social behavior, medicine, finance, marketing, among others. Useful as econometrics may be, it does not come handy for interdisciplinary non-economist researchers. This was also reflected in findings of the capacity needs assessment survey conducted for doctoral fellows supported by RUFORUM under the Social and Environmental Trade-offs in African Agriculture (Sentinel) Project in March 2021. A big proportion of respondents reported skills gaps in practical econometric data analysis techniques.

Responding to the capacity needs of the students, RUFORUM organised a basic econometrics and data analysis training using STATA on 28th June-2nd July 2021. Mr. Robert Asiimwe, a Senior Research Associate working with the International Center for Tropical Agriculture (CIAT), facilitated the training.  In order to widen the beneficiary audience for the training, all masters and doctoral students supported under the various RUFORUM programmes were invited to participate.  The training attracted 282 applicants of which 149 participants successfully completed the training. The participants who completed the training were drawn from 18 countries across Africa. Uganda had the highest number of participants (50) participants followed by Burundi (23) and Kenya (18). Students and early career researchers from 33 institutions participated. Gulu University in Uganda had the highest number with 29 followed by Makerere University in Uganda (25), University of Burundi (25) and Egerton University in Kenya (18). Twenty six (26%) percent of the participants were female. The main objective of the training was to build practical skills of graduate students in econometrics necessary to accomplish their research studies. The expected outcomes of the study were; students gain basic knowledge in econometrics concepts and models; and students gain practical skills in econometric modelling using STATA statistical software. The students were supported to install the STATA software, provided with training resource manuals before the actual training and received video recordings after each training session for future reference.

The training covered topics: general introduction to econometrics; introduction to Stata as an econometric software; understanding data and data management before and during the analysis process; regression analysis;  limited dependent variable models;  binary choice (probability) models; discrete (categorical) choice models (ordered and nominal dependent variables;  censored and  count data models.  The training was interactive involving presentations, Q&A analysis and practical data analysis sessions using STATA. The participants found the training very useful in broadening their knowledge base in preparation for their thesis writing and research careers for example one participants wrote as follows “The training was very helpful because it prepares a researcher on data collection and analysis which is something very critical in research”. Another participant remarked, “The training has been a success and I have personally learnt a lot and continue learning and referring to the materials that I have received”.

Looking forward, this econometrics training has opened a window of hope to several young researchers in the RUFORUM member universities that are struggling with technical capacities. It has created demand for more training. One of the participants wrote as follows, “I would like RUFORUM to organise another course on panel data analysis using pooled method, fixed model, random effect model or dynamic model”. This point to the importance of sharing technical expertise across the region to address these kinds of gaps. The training evaluation provided additional insights for improving online training programmes for RUFORUM. For instance, the participants noted the need to incorporate more practical sessions in trainings, allocate more practice time and arrange additional trainings in other relevant software. Through these trainings, the RUFORUM Secretariat has continued to engage with its network during the current COVID-19 disruptions on learning as well as provide learning and skills enhancement opportunities to students in the network beyond those directly receiving scholarships and/or research grants from the RUFORUM Secretariat.

This blog was originally posted on the RUFORUM website.