Basic Methods of Exploratory Analysis

Basic methods of exploratory analysis depend on the type of data being examined. For the most part, performance auditors will be dealing with combinations of categorical and continuous data. Other forms of data, such as text and geospatial data, do offer rich opportunities for analysis, but rarely form the basis of sampling frames in an audit environment. Exploratory analysis for most performance audits can be done with a limited number of basic tools, such as:

• descriptive statistics such as mean, median, variance, skewness (continuous variables)
• bar charts (categorical variables) and histograms (continuous variables)
• frequency distributions (categorical variables)
• cross-tabulations (relationships between pairs of categorical variables)
• scatter diagrams and bivariate correlations (combinations of continuous variables)
• box plots (combination of categorical and continuous variables)

As with any performance audit, the results of these tests can be combined with other audit procedures (such as inspection of procedural documents, benchmarking, or interviews of auditees) to understand program operations and the factors affecting them.

• PreviousQuality of Available Data
• NextOptimizing for Homogeneity