ABSTRACT
Experimental science aims to answer research questions by investigating the relationships between variables using quantitative data obtained in an experiment and assessing the significance of the results statistically (Yeadon and Challis, 1994). In an ideal experiment the effects of changing just one variable are determined. While it may be possible to change just one variable in a carefully controlled laboratory experiment in the natural sciences, this is problematic in the sports sciences in general and in sports biomechanics in particular. If a typical performance in a sport such as high jumping is to be investigated then any intervention must be minimal lest it make the performance atypical. For example, if the intent is to investigate the effect of run-up speed on the height reached by the mass centre in flight, asking the jumper to use various lengths of run-up might be expected to influence jumping technique minimally if the athlete normally does this in training. In such circumstances the run-up speed may be expected to vary as intended but other aspects of technique may also change. Faster approaches may be associated with a greater stride length and a more horizontal planting of the take-off leg. As a consequence the effects of a faster approach may be confounded by the effects of larger plant angles and changes in other technique variables. To isolate the relationship between approach speed and jump height statistical methods of analysis that remove the effects of other variables are needed (e.g. Greig and Yeadon, 2000). For this to be successful there must be a sufficient quantity and range of data to cope with the effects of several variables.
