Describing Data Distributions - Remember the objectives. Statistics are designed to suppress detail and reveal important findings.
- Always identify the scale types. The nature of the scale determines, in part, the
proper descriptive measures.
- Treat variables both ways. Continuous items with only a few scale values can
be treated as categorical.
- Portray information when feasible. Showing the entire distribution is more
effective than describing it.
- Describe data with care. The shape of the distribution affects the choice and
interpretation of descriptions.
- Begin with data description. Organize and interpret the descriptions of individual
variables before turning to measures of association.
- Choose the best average. Consider both the type of scale and the shape of the
distribution when selecting the best average.
- Assess confidence carefully. Choose an appropriate level of probability for
determining confidence intervals.
|