Читать книгу Demystifying Research for Medical and Healthcare Students - John L. Anderson - Страница 41

Randomisation in experimental designs refers to the process by which your research participants are allocated to one group or another. (It does not, as some people think, refer to the recruitment of the sample.) The idea is that, by assigning participants to groups entirely by chance, this will avoid any bias or preference by the researchers. It is a means of ensuring equity, or fairness, in the distribution of participants to groups. Successful randomisation will ensure that each group of participants is roughly equal in characteristics, and should be roughly equal in their chances of performing in any tasks or tests they are given. When I started out as a researcher, we used to have ‘tables of random numbers’ which we had to refer to for allocations. Nowadays it is all done by computer, and quite rightly so. Never trust any method of randomisation which a human being can control or influence!