A trainer wants to investigate the training regimes used by competitors wanting to improve their run times. The trainer decides to use simple random sampling to select 20 competitors at the winter triathlon to take part in their investigation. (a) Why is the sample not guaranteed to be a representative sample with respect to the classes of events? One variable about the running training that could usefully be collected is how easy the typical terrain used in training was, split into the following categories: 'very easy', 'easy', 'moderate', 'hard', 'very hard'. (b) Give a reason why this variable could be regarded as subjective data. (c) Give a reason why this variable could be regarded as ordinal data. Suppose that the trainer has come up with a new idea for a running training program and wishes to investigate whether this new training program is more successful than their standard training program. The trainer is most interested in whether the time taken to complete the run portion of the triathlon will be reduced if athletes use the new training program to prepare for their next event. (d) Explain why a double-blind trial would not be possible in this investigation. (e) Give a reason why a crossover design might not be suitable for this investigation. Suppose a matched-pairs design is used and 10 pairs of competitors are selected. One member of each pair is randomly selected to follow the new training program and the other member follows the standard training. (f) State the hypotheses that can be used for this matched pair t-test, using notation that you should define. (g) Give a reason why this hypothesis test might not turn out to be suitable. (h) Suppose that further analysis later suggested a Type 2 error had been made. What does this mean in terms of the evidence against the null hypothesis that was found in the initial analysis? What impact might this error have had on the choice of training program in the future