S.M. Das
Ministry of Agriculture, Directorate of Research and Development, Livestock Production
Research Institute (LPRI), P.O. Box 202, Mpwapwa, Tanzania
E-mail: drd@ud.co.tz
The Livestock Production Research Institute (LPRI) at Mpwapwa in Central Zone of Tanzania is involved mainly in livestock research on dairy and beef cattle, small ruminants (meat and milk), pastures and forages and in socio-economic studies. These programmes are integrated within disciplines such as animal breeding and genetics, animal nutrition, animal health and husbandry, range ecology and management and animal reproduction. Currently, the institute has also embarked on biodiversity and conservation projects of livestock germplasm. The institute collaborates with the livestock research departments and centres in the six other zonal centres in Tanzania and also with the Animal Science and Production Department at Sekoine University of Agriculture at Morogoro. Some of the multidisciplinary research projects conducted on farms are also associated with crop research scientists and socio-economists.
The paper is a result of discussions with some livestock research scientists at LPRI and other centres in Tanzania. A full detailed survey of biometrics needs was not done.
Most scientists are involved in both on-station and on-farm research. Insufficient knowledge of biometrics has usually been the main setback during research project design, data analysis and interpretation of results. Most research scientists have limited knowledge of basic statistics. This has led to various projects being either abandoned halfway due to bad experimental design or not being analysed correctly. Whereas a research scientist has been trained in basic statistics, and some have been introduced to statistics for on-farm research, the general capacity in the application of biometrics from design to presentation of results is poor. Furthermore, livestock researchers have not been able to avail themselves of new database management and statistical software due to lack of training. Assistance in increasing the capacity in applied biometry at the institute and livestock research centres is important, especially for researchers working on long-term research projects conducted both on station and on farm.
Expertise in biometrics is limited due to the small number of scientists with an advanced knowledge of biometrics. Table 1 summarises the level of biometric knowledge among the livestock scientists. Thus, data analysis of small trials has often relied on simple statistical techniques. Sometimes, scientists have depended on the Statistics Unit within the Rural Economics Department, Sekoine University of Agriculture for data involving both on-station and on-farm research. The assistance of the International Livestock Research Institute (ILRI) has occasionally been sought for the analysis of large data sets covering many years of data collection. Staff from universities abroad such as the University of Reading and the University of Agriculture, Norway have also sometimes assisted in data analysis and interpretation of results.
Table 1. Level of biometric knowledge among 41 livestock research scientists (11 with PhD, 26 with MSc and 4 with BSc qualifications).
Knowledge of biometrics and software |
Scientists |
Use of computer for data management and analysis |
<40 |
Basic statistics (mean, SD, SE) |
>60 |
Experimental design |
<20 |
One or two-way analysis of variance |
<40 |
Regression analysis, correlation |
<40 |
Covariance analysis |
<10 |
Categorical data analysis |
<5 |
Animal breeding and genetic data analysis |
<5 |
Designing and coding of survey questionnaires |
<5 |
Survey and on-farm data analysis |
<5 |
Principal component analysis |
0 |
Interpretation of analysed data |
<10 |
Use of Minitab, Systat statistical software |
<20 |
Use of SAS or other advanced statistical software |
<5 |
Computer hardware has previously been provided through a project funded by the World Bank. However, the number of computers available in working condition to research scientists is low (i.e. 1 computer for 10 scientists). The software available consists mainly of word-processing (Word or Word Perfect). Excel, Lotus or Dbase are occasionally available.
The use of computers for data management is patchy. Some data from small trials are entered in(to) Minitab or Excel. Scientists and technicians have also entered some animal breeding and survey data in Dbase. The institute has only two computers running on Windows 95. These are used for administration, accounting and sometimes research data management. The institute does not have a licensed statistical package and scientists rely on unsupported versions of Minitab and Systat. Several animal breeding and husbandry projects have not been analysed due to lack of professional biometric support and lack of appropriate statistical software.
There have been only two basic courses in biometrics for scientists and technicians during the past ten years. These were conducted at the Sekoine University and two scientist from LPRI attended. Two scientists also attended a course organised by ILRI on an introduction to Dbase and SAS. Thus, most scientists have not had any formal training in data management or use of statistical software.
The main areas of biometric assistance required by livestock scientists both at LPRI and elsewhere in Tanzania are as follows:
The present efforts to enhance the biometric capacity of scientists is timely due to an increased thrust in livestock research in Tanzania during the past five years.
