M. Nabasirye
Department of Crop Science, Faculty of Agriculture, Makerere University
P.O. Box 7062, Kampala, Uganda
E-mail: mnabas@math.mu.ac.ug
The mandate of the National Agricultural Research Organisation (NARO) in Uganda is to undertake, promote and co-ordinate research in all aspects of crops, fisheries, forestry and livestock and to ensure dissemination and application of research results. Research is conducted at the centre and eight research institutes. There are 195 research scientists with varied disciplinary backgrounds and experience throughout the organisation. These are supported by 168 technical support staff. Having myself recently moved to Makerere University there are currently no biometricians on the organisation's staff.
A recent survey conducted throughout NARO showed that research was carried out either on station (15%), on farm (15%) or both (70%). More than 200 on-farm trials were conducted during 1998 and 1999 with an increased emphasis on studies involving farmer participation and impact assessment.
Prior to 1980 there was good access to biometric expertise with each research institute having its own biometrics unit, which offered advisory and analytical services and training to newly recruited scientists. During the 1980s, however, most biometricians left the research system and units became virtually non-functional. NARO was established in 1992. However, the role of biometrics was not explicitly defined with the consequence that there has been a major decline in the quality of biometric input in research compared with earlier years. Results of a survey on the state of biometrics in NARO carried out in 1997/98 justified institutionalisation of the discipline. The lone biometrician within NARO was given responsibility for biometrics in 1998. But, given the minimal ratio of biometrician to scientist (1:194) with no supporting staff, this initiative has obviously had very little impact.
A survey of current computer use shows that the majority of scientists are using spreadsheet software for data entry and management prior to statistical analysis (Table 1). Many also rely on Mstat. The use of statistical software is illustrated in Table 2. The table illustrates the variety of packages used. However, scientists have very limited access to manuals. The use of major statistical packages, such as SAS and Genstat, is limited by the inability to purchase software and inadequate skills in their use.
Table 1. Percentages of scientists using different software packages for data entry and management.
Software package |
Use by scientists |
Lotus |
42 |
Excel |
38 |
Quatro-pro |
26 |
Access |
2 |
Mstat |
32 |
Although there are currently 156 computers with 486 and above processors in NARO, not all are accessible by research scientists. Only a limited number of computers have appropriate statistical software installed, and scientists are required to share these computers. Scientists have varying levels of statistical expertise gained through training courses, some of which date back to the 1970s (7%) and 1980s (50% of scientists). Thus, there is a major need for training courses in biometrics, both at the introductory and advanced level.
Table 2. Percentages of scientists that have been trained in the use of statistical packages, use a statistical package and have access to a manual.
Software package |
Have received training |
Used by |
Access to a |
Mstat |
32 |
72 |
24 |
SAS |
25 |
50 |
13 |
Genstat |
14 |
27 |
7 |
SPSS |
8 |
20 |
7 |
Statgraphics |
6 |
9 |
6 |
Minitab |
11 |
9 |
1 |
Other |
14 |
19 |
12 |
There are important requirements for improving levels of expertise in biometrics.
Through such training activities there may be opportunities to promote and strengthen collaboration among biometricians both within and outside the East and southern Africa region. This should be a major consideration by participants at this workshop.
