Confounding data

Confounded reactors are likely to stay confounded. Data correlations can produce excellent fits and can be useful for predicting the response of the particular system on which the measurements were made to modest changes in operating conditions. They are unlikely to produce any fundamental information regarding the reaction rate, and have very limited utility in scaleup calculations. 

Other techniques to improve throughput are instrumentation based and may involve multiple HPLC systems. The simplest method involves the automated use of solid phase extraction cartridges for sample cleanup followed by direct injection into the mass spectrometer [114], Coupling of multiple HPLC systems to one mass spectrometer allows one column to equilibrate and separate while another column to flow into the mass spectrometer. Multiple HPLC systems may be configured such that the mass spectrometer is only exposed to each serial HPLC eluent as the analyte of interest is eluted [115,116]. Although multiple H P LC-based methods may increase throughput, they also typically decrease sensitivity and may confound data workup and interpretation. 

Excess mortality may confound data interpretation Increased sensitivity may not be relevant... 

The most difficult part of troubleshooting is often identifying the true cause of a problem. There are four main barriers incomplete data, conflicting data, confounding data, and opinion. The best way to identify the physical cause of a problem may be to view the operations firsthand. Once the true cause of a problem is known, designing experiments to resolve the difficulty is often straightforward. 

A relatively simple example of a confounded reactor is a nonisothermal batch reactor where the assumption of perfect mixing is reasonable but the temperature varies with time or axial position. The experimental data are fit to a model using Equation (7.8), but the model now requires a heat balance to be solved simultaneously with the component balances. For a batch reactor. 

To identify ancillary data needs and potential confounding factors that should be considered or documented to ensure the defensible interpretation of data on monitored biological indicators... 

Criteria 1) Relevance to human health endpoints. 2) Sensitivity to change in loadings. 3) Overall historical data quality. 4) Data collection infrastructure. 5) Feasibility of data collection and analysis. 6) Ability to adjust for confounding factors. 7) Understanding of linkages with rest of ecosystem. 8) Broad geographic distribution. 9) Well-known life history (for fauna). 10) Nonintrusive sampling. 

---