June 2009
Front Matter
Training
46
Critical Issues for Interpretation
•Issues to consider when planning and performing data analysis:
–Data quality
–Data availability
–Sampling duration
–Sampling frequency
–Complementary data
Data quality.  Information from collection and chemical analysis such as standard operating procedures, audits, accuracy and precision, and data validation provide insight into sample and collection biases and errors.  This information is necessary for data validation.  Metadata such as precision and accuracy are required for other analyses (e.g., receptor modeling).
Data availability.  The number of species and amount of data above detection give insight into what analyses can be performed and provide a starting point for planning data analysis.
Sampling duration.  Duration provides information about analysis possibilities, for example, 24-hr data cannot be used to investigate diurnal patterns.  This information may also be necessary for calculating completeness criteria when aggregating data.
Sampling frequency.  Frequency information provides further insight into what analyses will be possible; for example, one year of 1-in-6-day data may not be sufficient to investigate day-of-week tendencies.  Sample frequency will also be necessary to calculate data completeness and to aggregate data.
Complementary data.  Additional data for criteria pollutants, speciated PM, and non-toxic hydrocarbons and meteorological data can be useful in a variety of analyses such as data validation, understanding transport, and source identification.