April 26, 2012 1 Comment
These are attempts at examples of uncertainty that may be familiar.
You go to the doctor, who tells you that you have a disease for which there are two treatments, a well-established one and a new one, which – based on one trial – seems slightly better. You are undecided between treatments.
You now realise that your identical twin had the same disease a while back, had the old treatment, and it worked. Is this a good reason to decide on the old treatment? Why, or why not?
Now, a variation. Suppose that the reported advantage of the new treatment is such that after realising that your twin had the same disease, you are undecided. You get home and hear on the news that, whereas you had thought of all instances of the disease as being similar, it has just been found that there are 10 distinct variants that may respond differently to particular treatments. Which treatment do you now prefer, and why?
This is – hopefully – a simpler example, raising a sub-set of the issues.
You need to go to Bigcity regularly. For the same cost, and at convenient times, you could go by X’s or Y’s train (by different routes). The rail companies publish standardised, audited, statistics concerning numbers of trains cancelled per week, average delays and number of trains that are more than 10 minutes late. X’s trains seems marginally better. But you have a colleague who uses the same train that you would, provided by Y, and has found it to be reliable and on time. Which train do you choose, and why?