The End of the Affair?

Abstract

An overriding theme running through this book is one of British pragmatism, in that time and time again method and theory have been developed from attempts to solve practical problems.^1,2 Yule’s interest in analysing time series stemmed, at least in part, from his desire to understand the nature of the nonsense correlation problem which bedevilled early empirical work, while Kendall’s research was prompted by his work at the Ministry of Agriculture on analysing detrended and oscillatory agricultural time series and later on practical forecasting issues emanating from his consultancy contracts. Durbin had long associations with governmental statistical agencies from which his research on seasonal adjustment and regression stability over time was a natural development. Jenkins first encountered spectral analysis whilst tackling aircraft design problems at the Royal Aircraft Establishment and later developed his own consultancy company specializing in complete forecasting and decision systems for industry and government. Box cut his statistical teeth on wartime chemical experiments and later worked on control problems for ICI, maintaining his interest in this area throughout his career and setting up the Centre for Quality and Productivity Improvement at Madison. Box and Jenkins’ joint research on forecasting and control was aimed at providing a practical method for understanding and solving such problems with observed time series. Granger’s research, across numerous areas, was often sparked by the desire to understand and solve real, practical problems in time series.