ABSTRACT
Most of the methodology described in the preceding chapters has been developed within the classical (“frequentist”) framework of statistical inference. In this framework, populations are represented by probability models whose parameters are treated as fixed but unknown quantities about which inferences are to be made, such inferences are based exclusively on sample data, and the assessment of quantities such as bias, significance or confidence depends on the frequencies with which given sample values might arise if repeated samples were to be taken from the relevant populations. Thus, within this framework there is no scope for incorporating any extraneous information that the researcher might have gleaned about some of the population parameters (say from a series of previous experiments with similar materials), or for incorporating any subjective elements into the analysis.
