Like most guidebook ventures it would be easy to say "if I knew what I know now I would have run a mile." Luckily I know who to blame for this. As a postgraduate at Royal Holloway University of London I was fortunate to come under the influence of John Brew. John had recently joined the academic staff, trained as an engineer, he came from a position where he had been undertaking operational research using clinical trial data for a local health authority. An unusual background perhaps but this serves to demonstrate the vision of Prof. David Bowen (head of the Quaternary Research Centre) who realised the need to bring on board specialists who could utilise numerical and statistical techniques. I had always been interested in the application of statistical techniques from the days I had written my first BASIC program to allow me to complete my undergraduate regression practicals. In some ways I had been fortunate to catch the end of the era when data analysis formed a prominent role in undergraduate Physical Geography (what chance today's undergraduates?). However the quantitative revolution in geography had stalled, in fact it had stagnated, with only a few individuals continuing to take an interest in developments in main stream statistics. John introduced me to a new world which offered great possibilities for the analysis of qualitative data, in particular there was this rather intriguing family of statistical models grouped together under the rather uninformative title of Generalised Linear Models.
Enlightened by his experience I began a mini-crusade to promote utilisation of these methods. As my own research expanded into the area of palaeoecology I delved deeper into the statistical literature and discovered a wealth of different techniques including correspondence analysis (particularly the wonderful work of Cajo ter Braak,) and the exciting world of transfer function work. Although a battery of techniques existed their use was restricted to a select band of dedicated individuals with the wider Quaternary community seemingly largely ignorant of its potential use in a much wider context.
Mainstream statistics packages such as SPSS and Statgraphics had begun to include some of these techniques and desktop PCS were now cheap and powerful enough to make them widely available and accessible for the first time. Hence an idea was born what was needed was a cheap technical guide short and sweet to attract the interest of the wider community. From here we turned to the Quaternary Research Association, whose excellent series of technical guides were cheap and well respected exactly what we were looking for. Could a commercial publisher have produced a guide for £10?
Our vision was perhaps optimistic. How do you cover relatively advanced statistics when the general readership may have forgotten about regression or worse yet never undertaken any statistical analysis (yes they do exist!). We hoped to make it a practical manual, a kind of cookbook, but we had to modify this both in the light of the manuscripts and because it ultimately proved too ambitious. We decided instead to try and cater for a range of abilities with chapters which would act in some cases as memory joggers whilst encouraging further development. Other chapters would introduce new concepts.
Thanks to the help of Lou Maher who allowed us to call for contributors in this Newsletter, our quest was greatly aided by those who volunteered to write the various sections. An international list of contributors including John Birks (Norway) and David Green (Australia) has its own problems especially when trying to meet deadlines. However we decided to take a pragmatic view and allow the authors time to write contributions they could be proud of. We provided only a brief list of guidelines which were observed religiously by often surprisingly co-operative contributors. Although our production time-table slipped progressively, the QRA remained supportive.
As with all such ventures the contributors produced their manuscripts at widely differing rates. The chapters on Factor Analysis (John Waldon and John Smith) and methods appropriate to tree-ring sequence matching (Martin Bridge) were finished quickly as they present short case studies of specific techniques.
Warren Kovach had the unenviable task of summarising Multivariate Data Analysis and David Green the equally impossible task of summarising Time Series Methods - both in single chapters. Given the space limitations both have completed their contributions with considerable merit and should provide the reader with sufficient knowledge to approach these vast subject areas with confidence. Our own chapter on Generalised linear models went through several modifications with much soul searching over the precise balance between theory and practise I am not certain we achieved the balance we would have wished for. Perhaps the most amazing contribution is that of John Birks. With typically thoroughness his review of Quantitative palaeoenvironmental reconstruction is, in my view, the most impressive chapter ever written on the subject -- miss it at your peril!
During the preparation of the guide we suffered a number of further difficulties not the least of which was the departure of John from Royal Holloway. Unfortunately some of his colleagues did not share the early vision of David Bowen and it was time for him to move on. Utilising e-mail from home the work continued however. Final submission of the document produced a number of new headaches. Due to financial constraints the printing and binding of the Guide was undertaken in the Geography Department at Durham University. Early attempts suffered from inadequate binding (pages often falling out) and poor cover quality. The combined efforts of Dr David Bridgland and Dr Wishart Mitchell (both acting on behalf of the QRA) resolved these problems despite my persistent moaning.
Did we succeed? Well we certainly completed the guide! Perhaps the whole is not as coherent as we would have liked and perhaps not as practical as we first envisaged. There are many things that can be criticized (and I am sure they will be!) and indeed there are a number of things we would do differently if we started all over again but generally we are satisfied with the outcome.
Sales are evidently going well but we would be happy to know what you think. If you would like a copy of the guide, the full reference is listed at the beginning of its review by Andre Lotter.