Statistical test planning

Statistical test planning is an effective aid to recognizing significant quantities that influence chemical reactions. A systematic process for searching for suitable catalysts and optimizing diem is especially helpful in the case of catalytic reactions, with their numerous test parameters. 

In this chapter we shall largely dispense with the mathematical basis of statistical test planning and we will illustrate the method with the aid of some simple practical examples. 

Selection and Testing [Pg.379]

If the experiment is conducted in stages, precautions must be taken to ensure that possible differences between the stages do not invaUdate the results. Appropriate procedures to compare the stages must be included, both in the test plan and in the statistical analysis. For example, some standard test conditions, known as controls, may be included in each stage of the experiment. 

If there are separate analysis plans for the clinical and economic evaluations, efforts should be made to make them as consistent as possible (e.g., shared use of an intention-to-treat analysis, shared use of statistical tests for variables used commonly by both analyses, etc.). At the same time, the outcomes of the clinical and economic studies can differ (e.g., the primary outcome of the clinical evaluation might focus on event-free survival, while the primary outcome of the economic evaluation might focus on quality-adjusted survival). Thus, the two plans need not be identical. 

W.J. Youden and E.H. Steiner, Statistical Manual of the AOAC. Statistical Techniques for Collaborative Tests. Planning and Analysis of Results of Collaborative tests, AOAC, Washington,DC, (1975) 33-121. 

Step 6 allows us to create a statistical approach for the evaluation of the collected data. Using a statistical test and the statistical parameters selected in Step 6, we will be able to control decision error and make decisions with a certain level of confidence. Decision error, like total error, can only be minimized, but never entirely eliminated. However, we can control this error by setting a tolerable level of risk of an incorrect decision. Conducting Step 6 enables the planning team to specify acceptable probabilities of making an error (the tolerable limits on decision errors). At this step of the DQO process, the project team will address the following issues ... 

The planning team will use similar rationale for assigning tolerable decision error rates in an opposite scenario illustrated in Figure 2.4. In the planning phase, these probabilities will be factored into statistical tests for the calculation of the proper number of samples to be collected. In the assessment phase, they will be used in a statistical verification of the attainment of the action level. 

Define in the planning documents the types of statistical tests that will be used for data evaluation. 

A kinetic expression that is selected should satisfy certain statistical tests and physicochemical constraints, before it can be indicated as an adequate model. These tests often cannot be applied together, due to lack of information, but should be considered as far as possible in the evaluation and selection of the best rate expression. When still a set of competitive expressions are left some model discrimination techniques may be applied or additional experiments should be conducted, based on a careful, efficient planning. 

W. Nelson. In Accelerated Testing Statistical Models, Test Plans, and Data Analyses. 1990, John Wiley, New York. p. 115. 

Confusion of compendial standards with release tests and with statistical sampling plans occasionally occurs. Interpretation of results from official tests and assays requires an understanding of the nature and style of compendial standards. Tests and assays given in the USP prescribe operation on a single specimen that is, the singlet determination. This is the minimum... 

Power. The power is the probability of concluding that the alternative hypothesis is true given that it is in fact true. It depends on the statistical test being employed, the size of that test, the nature and variability of the observations obtained and the size of the trial. It also depends on the alternative hypothesis. In practice there is no single alternative h3fpothesis, so a reference alternative based on a clinically relevant difference is usually employed. The power of a trial is a useful concept when planning the trial but has little relevance to the interpretation of its results. (Caution not all statisticians agree with this last statement.)... 

The handling and analysis of pharmacoeco-nomic data should be along the lines familiar to those observing good clinical practices (GCP) for other purposes. Data collection instruments need to be selected, or created and incorporated into case report forms, just as for any other end-point. Data analysis plans should be created prospectively. The statistical analysis plan should be prospective, and should help put the pharmacoe-conomic measures in the context of other properties of the test medication (Table 19.3). Are they... 

The Plackett-Burmaa plan, which is based on statistics and combinatorial analysis, allows N— 1 effects of variables to be determined simultaneously in N tests. In this highly simplified test plan, only the main effects can be determined numerically at the same time, error estimation is performed by means of a blank variable. Interactions between the variables can not be determined. A test matrix for seven variables is shown in Table 13-11. 

Statistical test plarmiag can be carried out advantageously by expert systems which design the test plan, evaluate the results, and optimize the process in a single logically constructed sequence. An example is the program APO (Analysis Process Optimization) [30]. 

This chapter gives an overview of relevant statistical methods for the analysis of bioimpedance measurements, with an aim to answer questions such as How do I begin with planning an experiment How many measurements do I need to take How do I deal with large amounts of frequency sweep data Which statistical test should I use How do I validate my results ... 

Wayne, N. (1990). Accelerated testing statistical models, test plans and data analyses. Wiley. 

Statistical testing expertise is available at multiple locations outside of government. The committee observed evidence that there could be greater use of external expertise to review testing plans that are currently created internally and apparently not subject to independent external review (See also Test and Evaluation),... 

W.B. Nelson, Accelerated Testing - Statistical Models, Test Plans and Data Analysis, Wiley, New York, NY, USA, 1990. 

---