ABSTRACT
In order to draw valid conclusions, it is important to collect ‘good’ data. It is usually helpful to distinguish Between data collected in designed experiments (including aliniaal trials), sample surveys, pure observational studies and more specialist investigations such as controlled prospeativr and retrospective studies. T he general principles of experimental design and survey sampling are summarized in sections A.10-A.12 of Appendix A but will probably be familiar to many readers. Here wc make some more general comments.
