# Northern Ireland group meeting: 40 years a general practitioner in statistics

Written by Gilbert MacKenzie on . Posted in Sections and local group meeting reports

The Northern Ireland group held a meeting at 4pm on Thursday the 9th of May, 2013 in the David Bates Building in the Queen's University of Belfast. The speaker was Michael Stevenson of  Queen's University Belfast.

Michael Stevenson and his keen interest in trains (Stevenson, 2000, Steam's last fling, Colourpoint Books) is well known to the local group. Perhaps, not unsurprisingly, we learned that it was this early hobby which had inspired his interest in statistics.

The first part of his talk dealt with his early days of recording locomotive data and how this led to a set of initial data based on 50 records of the semi-fast 17.30 business express from York Road train station in Belfast to Larne (a port on the coast approximately 22 miles north-east of Belfast) recorded between March 1967 and August 1969. Michael was 17 years old in 1969. In all he accumulated data on 160+ starts from York road station heading to several different destinations. Unsure as how to proceed he had first read MJ Moroney's book Facts from Figures.

Michael then described the train terrain as it existed at that time. This was a walk down memory lane with many old photographs of York Street terminus and the trains of the time including steam driven 'Jeeps', but working up to engines of modern times. From the terminus to milepost 3, the level line follows the old shore of Belfast lough. Just before the former station of Whitehouse (Merville Garden Village/Gideon’s Green) the Larne line begins to climb at a fairly steep gradient of 1/89.

This led to an analysis of 122 starts from York street terminus between August 1966 and March 1970 involving Jeeps. No Jeep had ever reached 60mph on this stretch, but an interesting outlier was found at 69mph and the reading was found to be correct! A linear regression analysis was then undertaken in which the dependent variable was 'the maximum speed before Whitehouse' and several categorial covariates were studied, including: Load (in equivalent bogies), Engine, Driver class (elite, solid, dull, dreadful, Double-headed (Yes/No), Destination Main Line (Yes/No), and Train heating required (Yes/No). All of these variables reached statistical significance and Michael was able to identify the fastest engine.

Later he was to pursue a further regression analyses of more modern engines on the south-westerly runs to Lambeg, but by this time he had convinced the meeting of his love of Statistics and his great knowledge of the Northern Irish railway system and its engines.

His talk was received with acclaim and a lively discussion ensued.

Northern Ireland local group