Data analysis test approach

The above two objectives, data examination and preparation, are the primary focus of this section. For data examination, two major techniques are presented the scattergram and Bartlett s test. Likewise, for data preparation (with the issues of rounding and outliers having been addressed in a previous chapter) two techniques are presented randomization (including a test for randomness in a sample of data) and transformation. Exploratory data analysis (EDA) is presented and briefly reviewed later. This is a broad collection of techniques and approaches to probe data, that is, to both examine and to perform some initial, flexible analysis of the data. 

To be effective the investigation must apply an approach which is based on basic incident causation theories and use tested data analysis techniques. Investigating incidents to determine root causes and make recommendations can be as much an art as a science. Within the industry, best practices in incident investigation have evolved substantially in the last 20 years. This chapter provides a brief overview of some of the more relevant causation theories. 

Progressive companies use a more structured and comprehensive team approach to identify root causes. Scientific principles and concepts are applied to determine root causes and make recommendations to prevent recurrence. Effective investigations should use tested data analysis tools and methodologies to seek the identification of multiple causes. To be repeatable, the investigation should use a systematic approach, which may also be prescriptive. As a rule, the benefits of this systematic approach result from two actions ... 

Exploratory data analysis, EDA, is an essential prerequisite of the examination of data by confirmatory methods. Time spent here can lead to a much greater appreciation of its structure and the selection of the most appropriate confirmatory technique. This has parallels in the analytical world. The story of the student s reply to the question Ts the organic material a carboxylic acid which was I don t know because the IR scan isn t back yet poses questions about the approaches to preliminary testing ... 

The first, called the integral method of data analysis, consists of hypothesizing rate expressions and then testing the data to see if the hypothesized rate expression fits the experimental data. These types of graphing approaches are well covered in most textbooks on kinetics or reactor design. 

The traditional way is to measure the impedance curve, Z(co), point-after-point, i.e., by measuring the response to each individual sinusoidal perturbation with a frequency, to. Recently, nonconventional approaches to measure the impedance function, Z(a>), have been developed based on the simultaneous imposition of a set of various sinusoidal harmonics, or noise, or a small-amplitude potential step etc, with subsequent Fourier- and Laplace transform data analysis. The self-consistency of the measured spectra is tested with the use of the Kramers-Kronig transformations [iii, iv] whose violation testifies in favor of a non-steady state character of the studied system (e.g., in corrosion). An alternative development is in the area of impedance spectroscopy for nonstationary systems in which the properties of the system change with time. 

All these people have some interest in data analysis or chemometrics, but approach the subject in radically different ways. Writing a text that is supposed to appeal to a broad church of scientists must take this into account. The average statistician likes to build on concepts such as significance tests, matrix least squares and so on. A... 

When compounds are selected according to SMD, this necessitates the adequate description of their structures by means of quantitative variables, "structure descriptors". This description can then be used after the compound selection, synthesis, and biological testing to formulate quantitative models between structural variation and activity variation, so called Quantitative Structure Activity Relationships (QSARs). For extensive reviews, see references 3 and 4. With multiple structure descriptors and multiple biological activity variables (responses), these models are necessarily multivariate (M-QSAR) in their nature, making the Partial Least Squares Projections to Latent Structures (PLS) approach suitable for the data analysis. PLS is a statistical method, which relates a multivariate descriptor data set (X) to a multivariate response data set Y. PLS is well described elsewhere and will not be described any further here [42, 43]. 

These properties are especially important in the design, data analysis, and interpretation of multispecies toxicity tests, field studies, and environmental risk assessment and will be discussed in the appropriate sections. This alternate approach rejects the smooth transition of effects and recognizes that ecosystems have fundamentally different properties and are expected to react unexpectedly to contaminants. 

---