It involves the collection of quantitative data on at least two variables at one point in time (or over a short period) and from a number of cases. 

The data is used to look for patterns of association or relationships either in the group as a whole (all cases) or in subgroups sharing characteristics or attributes (females or males for example). 

1. Relatively inexpensive and takes up little time to conduct;

2. Can estimate prevalence of outcome of interest as sample is usually taken from the whole population;

3. Many outcomes and risk factors can be assessed;

4. Useful for public health planning, understanding disease aetiology and for the generation of hypotheses;

5. There is no loss to follow-up.

1. Difficult to make causal inference in simple statistical tests of relationships but causal inferences can be made using sophisticated techniques such as regression analyses.

2. Only a snapshot: situation may provide differing results if another time-frame had been chosen;

3. Prevalence-incidence bias (Neyman bias): a risk factor causing death will be under-represented among those with the disease.

* * *