Imagine being a painter with no brushes, or a builder with no trowel, or a chef with no wooden spoon… It would be very difficult to do your job; there are ways around it – but it might make your job much harder and take twice as long. All of us need the right tools to make our jobs easier and make the results better.Now let’s imagine being a researcher or scientist, but not be able to access professional software to help with your analysis. Well, strictly speaking you can access the software, but the cost is too great for you or your organisation. This unfortunately is the reality for many universities and research stations in the Developing World; the repeated cry is inadequate infrastructure, lack of funding and little access to literature.
“This problem has now been solved with the release of GenStat Discovery Edition 3, professional statistical software that is free for non-commercial users such as government researchers, lecturers and students in developing countries” says Ric Coe, who heads the Research Methods Group in Nairobi, Kenya. This is a joint unit across two CGIAR centres that trains and gives advice to researchers on how to improve their research.
I spoke to Stewart Andrews (CEO of VSNi) to find out more:
Stewart: GenStat Discovery is the full version of GenStat but available only to non-commercial organisations in the Developing World; it’s as simple as that. We know that GenStat is a very powerful and easy to use statistical software system; and that it has features that are useful when doing research on agriculture, biosciences and related disciplines, including climate change. Africa is a continent where the effects of climate change are noticeable and can be dramatic i.e. even a small reduction in rainfall can affect the crop yields dramatically. We also know that many African institutes cannot afford to buy professional statistical software; so in 2003 we launched the first GenStat Discovery Edition – a free version of GenStat, initially only in Africa. The response was so positive that we made it available to all developing countries the following year, and we have continued to support this since.
Stewart: We believe that we have a social and moral responsibility to assist where we can. A single charitable donation would provide financial aid at that time, but we believe that it is better to give people the ability to solve their own problems in the long term; and therefore be able to plan for the future better…for example one of my great frustrations is when my computer stops working and I ask for help; what I actually want is the ability to solve the problem in the future, not a one off fix. By providing researchers and scientists in the Developing World with a tool for analysis and prediction and the support that goes with that I believe we are giving everyone a fair and equal chance to access the best in data analysis systems so no matter where you are in the world you need not necessarily be hampered by lack of resource.
Stewart: The supported users from the developed world essentially pay for it – as without them and their business we could not provide GenStat Discovery for free. We are a commercial organisation, not a charity; therefore we have commitments to our employees and our board. However, the more business we do – the more we can help organisations from developing countries, who do not have the financial capacity to purchase GenStat. Having said that, our prices are not inflated to cover this project; GenStat is priced fairly throughout the world – we see this as putting our ethics into practice.
Research is about giving information to people so that they can make decisions based on hard evidence and at the same time have an idea about risks and uncertainties. If people have access to information, they can start to make informed decisions; information can open minds to ideas and doors to the future.
GenStat Discovery 3 was launched in November 2007 in Nairobi, Kenya and Kampala, Uganda. This picture is of Wim Buysse (Flemish Association for Development Cooperation and Technical Assistance), Professor Adipala (RUForum) and Ric Coe (ILRI). GenStat Discovery 3 is available direct from VSNi, simply follow the link below to download your copy. In Africa, a free CD-Rom is available from RUForum, the Regional Universities Forum for Capacity Building in Agriculture. The license key for GDE3 is automatically installed from the download version and CD-Rom, but expires at the end of 2008. To renew your license key and receive information on Discovery and related projects, register on the VSNi website – see below.
RUForum is an initiative by a consortium of 12 universities in east and southern Africa, to develop and strengthen human resource capacity for inter-disciplinary problem-solving. For more information go to or email.
For more information on GenStat Discovery 3 go here
GenStat provides control over the appearance of graphs so that they can be customized to suit your needs. Amongst other things, you can choose tick mark positions, axes label directions, line styles, colours and symbols of points, and switch grid lines on or off. Together these settings define a graphics environment. You can modify the graphics environment by using the menu provided (Tools | Graphics Environments) or by using the GenStat commands FRAME, XAXIS, YAXIS, ZAXIS, PEN and COLOUR. Graphics environments can be saved to files that can be loaded back into GenStat using the DSAVE and DLOAD commands. So, you can create a graphics environment to suit a specific project or publication and save it to a file so it can be shared with colleagues and used again when similar plots are required.
Here is an example where the same data is plotted using the default graphics environment and again using a customized one. This customized graphics environment uses larger sizes for symbols, axes labels and the plot title as well as adding gridlines, sub tick marks and a box around the plot and placing the tick marks on the inside.
We’re pleased to report back on Professor Roger Payne’s (VSNi’s CSTO) invited talk at the International Annual Meeting of the American Society of Agronomy, held in New Orleans. The ASA had invited Roger to give a special lecture entitled “Fisher, Yates, Nelder and Thompson: The Development of Statistical Design and Analysis Concepts at Rothamsted”. The lecture was the first of three in a symposium called “The Role of Statistics in Agricultural Research: Past, Present and Future”. Rothamsted Research, probably the oldest research station in the world, saw the beginnings of statistics. This led to the development of GenStat, now of course developed by us at VSNi. Go here to see Roger’s slides. While at the conference, Roger also met with Dieter Mulitze of Agronomix, developers of the Gen II database, and demonstrated the links between Gen II and GenStat to visitors at the Agronomix stand.
Biometricians the world over would no doubt have wanted to join the IBS Australian Region Conference – Biometrics by the Beach, where an exciting scientific programme included Professor Ari Verbyla talking on Whole genome analysis of QTL, including analyses of various experiments presented using an implementation of the approach in the ASReml software. This picture shows Ari with VSNi technical developer David Baird and our good friend Maryann Ebsworth.Don’t forget to register for the European GenStat Applied Statistics Conference at AFBI, in Belfast, Northern Ireland, from 21st- 22nd July 2008, for more details see the event page.
To see where VSNi will be in the future look here.
If you are involved in organising an event which may be of interest to VSNi and our users please let us know by emailing us.
Don’t forget to check out the latest VSNi training course schedule to find out when your nearest course will be held, or email us to request a specific event.Another new course has been developed and is being held in February 2008 in Hemel Hempstead on using Regression, non-linear and generalized linear models in GenStat. The course shows how GenStat’s menus guide you from simple to very complicated analyses. It aims to familiarize you with these techniques, and give you the underlying knowledge and confidence to use them correctly and effectively.
Some familiarity of GenStat is necessary and a practical experience of the standard probability distributions (Normal, binomial and Poisson), but there will be no need for any complicated maths. On each day the lectures will be interspersed with practicals to introduce you to real-life data sets and illustrate the methods. The practicals also give you the opportunity to discuss your own problems and investigations with the presenter.
For more details on this Regression, nonlinear and generalized linear models in GenStat course, or to book a place and for more information view the current course schedule, or contact the VSNi training team.
For those who know and love the Christopher Robin stories by A. A. Milne, you’ll know that Piglet’s favourite food is acorns. Given that Piglet, like all of Christopher Robins’ friends, lives in The Hundred Acre Wood, you would think that Piglet’s food should be easy to find…after all The Hundred Acre Wood is in England, acorns come from oak trees, and oak trees are common enough in England…But what if it wasn’t that simple? A typical English woodland usually comprises a variety of different deciduous and evergreen trees…all of whom are fighting for survival against each other, against groundcover/vegetation and damage from squirrels, deer etc …so how can Piglet be sure that the mighty oak will survive and regenerate to produce more of his beloved acorns?
Well, perhaps GenStat holds the key…a recent study undertaken by Forest Research on oak regeneration indicates that a number of factors can affect the regeneration of oak trees – such as the amount and type of groundcover that is best suited for oak seedlings to survive, how far the seedlings are from the parent tree, and therefore how much the canopy might need cutting back…and so the list goes on. But with the help of GenStat’s Generalised Linear Models, Piglet could investigate the relationships between location and vegetation variables, and seedling data; linear regression could be used to analyse the height and diameter of seedlings and vegetation cover. Piglet would need to analyse this experiment over many years and take many measurements, therefore the mixed model facilities in GenStat would help to manage the issues of unbalanced data…
So with a little statistical analysis, time and a great software package like GenStat, Piglet could learn to manage the woodland for optimum acorn production…well, providing Christopher Robin lets him borrow his computer…
 Development of Quercus robur advance regeneration following canopy reduction in an oak woodland, Ralph Harmer & Geoff Morgan, Forestry 2007 80(2): 137-149
The original article featured in Forestry and Timber News (Dec 07): Statistical software in forestry research, can be found here (as a pdf):
Finally, we wish you all a very Merry Christmas and Happy and prosperous New Year from all at VSNi