Distributions, selection random-effects analysis

In addition it is now time to think about the two assumption models, or types of analysis of variance. ANOVA type 1 assumes that all levels of the factors are included in the analysis and are fixed (fixed effect model). Then the analysis is essentially interested in comparing mean values, i.e. to test the significance of an effect. ANOVA type 2 assumes that the included levels of the factors are selected at random from the distribution of levels (random effect model). Here the final aim is to estimate the variance components, i.e. the variance fractions with respect to total variance caused by the samples taken or the measurements made. In that case one is well advised to ensure balanced designs, i.e. equally occupied cells in the above scheme, because only then is the estimation process straightforward. 

Grouping and Segregation Error this error arises from nonrandom distribution of particle sizes, usually a result of gravitational effects it can be minimized by selecting samples for analysis from many randomly selected primary samples or by careful homogenization and splitting of the sample. 

Selected entries from Methods in Enzymology [vol, page(s)] Analysis of GTP-binding/GTPase cycle of G protein, 237, 411-412 applications, 240, 216-217, 247 246, 301-302 [diffusion rates, 246, 303 distance of closest approach, 246, 303 DNA (Holliday junctions, 246, 325-326 hybridization, 246, 324 structure, 246, 322-324) dye development, 246, 303, 328 reaction kinetics, 246, 18, 302-303, 322] computer programs for testing, 240, 243-247 conformational distribution determination, 240, 247-253 decay evaluation [donor fluorescence decay, 240, 230-234, 249-250, 252 exponential approximation of exact theoretical decay, 240, 222-229 linked systems, 240, 234-237, 249-253 randomly distributed fluorophores, 240, 237-243] diffusion coefficient determination, 240, 248, 250-251 diffusion-enhanced FRET, 246, 326-328 distance measurement [accuracy, 246, 330 effect of dye orientation, 246, 305, 312-313 limitations, 246,... 

For the interpretive optimization of the primary (program) parameters in the programmed analysis of complex sample mixtures it may well be sufficient to optimize for the major sample components. This may be done if it is assumed that the primary parameters do not have a considerable effect on the selectivity, so that if the major sample components are well spread out over the chromatogram, the minor components in between these peaks will follow suit automatically, and if it is assumed that the minor peaks are randomly distributed over the chromatogram. The major chromatographic peaks can be separated to any desired degree if optimization criteria are selected which allow a transfer of the result to another column. 

There have been numerous studies examining the selection of data for an SSD. Forbes and Calow (2002) made the point that only a fraction of the species going into the SSD determines the effects threshold. With all species being weighted equally, the loss of any species is of equal importance to the system, while keystone or other important species are assumed to be randomly distributed in the SSD. For example, the ecologically realistic distribution of species by trophic level was 64% primary producers, 26% herbivores (invertebrates), and 10% carnivores (fish), compared to the mean ratio from SSDs for different chemicals of 27.5, 34.7, and 37.8%, respectively. Such variations were shown to alter the SSDs by as much as 10% (Duboudin et al. 2004). A sensitivity analysis performed on available data for chromium (VI) in marine waters (Table 4.8) shows how additional data points, or selective removal of data, have an impact on the derived 5th percentile (HC5). The effects are relatively small but can be higher for the 1st percentile data (HC1). Our view is that, provided the data set includes numbers of sensitive and insensitive species equal to or above the minimum data set, it is considered to be adequate. 

---