This paper discusses five very basic considerations in the statistical analysis of field accident data. Although these points are not new, the paper brings them together to form a checklist which can be used to evaluate the credibility of statistical studies. The five points considered are:​

1. Sampling Technique
2. Random Variations
3. Definitions, Missing Data, Data Accuracy
4. Performance Measuring Techniques
5. Population Differences

These points are demonstrated by using a hypothetical restraint study as an example.