Spark plug fouling is a common problem when vehicles are repeatedly operated for very short periods, particularly at low temperatures. This paper describes a test procedure which uses a series of short, high-load drive cycles to assess plug fouling under realistic conditions. The engine is force cooled between drive cycles in order to increase test throughput. Spark plug resistance is shown to be a poor indicator of the effect of fouling on engine performance and the rate of misfiring is given as an alternative measure. An automated technique to detect misfires from engine speed data is described. This has been used to investigate the effect of spark plug type, fuelling level and spark timing on fouling. Spark plugs which are designed to run hotter are found to be more resistant to plug fouling. Isolated adjustments to fuelling level and spark timing calibrations within the range providing acceptable performance have a weak effect on susceptibility to plug fouling. These effects are synergetic, and when fuelling and spark timing are both optimised this susceptibility is reduced substantially compared to cases when production calibration values have been used.