Choosing a Cluster Sampling Design for Lot Quality Assurance Sampling Surveys

Lot quality assurance sampling (LQAS) surveys are commonly used for monitoring and evaluation in resource-limited settings. Recently several methods have been proposed to combine LQAS with cluster sampling for more timely and cost-effective data collection. For some of these methods, the standard binomial model can be used for constructing decision rules as the clustering can be ignored. For other designs, considered here, clustering is accommodated in the design phase. In this paper, we compare these latter cluster LQAS methodologies and provide recommendations for choosing a cluster LQAS design. We compare technical differences in the three methods and determine situations in which the choice of method results in a substantively different design. We consider two different aspects of the methods: the distributional assumptions and the clustering parameterization. Further, we provide software tools for implementing each method and clarify misconceptions about these designs in the literature. We illustrate the differences in these methods using vaccination and nutrition cluster LQAS surveys as example designs. The cluster methods are not sensitive to the distributional assumptions but can result in substantially different designs (sample sizes) depending on the clustering parameterization. However, none of the clustering parameterizations used in the existing methods appears to be consistent with the observed data, and, consequently, choice between the cluster LQAS methods is not straightforward. Further research should attempt to characterize clustering patterns in specific applications and provide suggestions for best-practice cluster LQAS designs on a setting-specific basis.     

Counting Alleles with Rarefaction: Private Alleles and Hierarchical Sampling Designs

The number of alleles (allelic richness) in a population is a fundamental measure of genetic variation, and a useful statistic for identifying populations for conservation. Estimating allelic richness is complicated by the effects of sample size: large samples are expected to have more alleles. Rarefaction solves this problem. This communication extends the rarefaction procedure to count private alleles and to accommodate hierarchical sampling designs.     

Within-Otolith Variability in Chemical Fingerprints: Implications for Sampling Designs and Possible Environmental Interpretation

Largely used as a natural biological tag in studies of dispersal/connectivity of fish, otolith elemental fingerprinting is usually analyzed by laser ablation-inductively coupled plasma-mass spectrometry (LA-ICP-MS). LA-ICP-MS produces an elemental fingerprint at a discrete time-point in the life of a fish and can generate data on within-otolith variability of that fingerprint. The presence of within-otolith variability has been previously acknowledged but not incorporated into experimental designs on the presumed, but untested, grounds of both its negligibility compared to among-otolith variability and of spatial autocorrelation among multiple ablations within an otolith. Here, using a hierarchical sampling design of spatial variation at multiple scales in otolith chemical fingerprints for two Mediterranean coastal fishes, we explore: 1) whether multiple ablations within an otolith can be used as independent replicates for significance tests among otoliths, and 2) the implications of incorporating within-otolith variability when assessing spatial variability in otolith chemistry at a hierarchy of spatial scales (different fish, from different sites, at different locations on the Apulian Adriatic coast). We find that multiple ablations along the same daily rings do not necessarily exhibit spatial dependency within the otolith and can be used to estimate residual variability in a hierarchical sampling design. Inclusion of within-otolith measurements reveals that individuals at the same site can show significant variability in elemental uptake. Within-otolith variability examined across the spatial hierarchy identifies differences between the two fish species investigated, and this finding leads to discussion of the potential for within-otolith variability to be used as a marker for fish exposure to stressful conditions. We also demonstrate that a ‘cost’-optimal allocation of sampling effort should typically include some level of within-otolith replication in the experimental design. Our findings provide novel evidence to aid the design of future sampling programs and improve our general understanding of the mechanisms regulating elemental fingerprints.     

Spatial Variation in Soil Properties among North American Ecosystems and Guidelines for Sampling Designs

Soils are highly variable at many spatial scales, which makes designing studies to accurately estimate the mean value of soil properties across space challenging. The spatial correlation structure is critical to develop robust sampling strategies (e.g., sample size and sample spacing). Current guidelines for designing studies recommend conducting preliminary investigation(s) to characterize this structure, but are rarely followed and sampling designs are often defined by logistics rather than quantitative considerations. The spatial variability of soils was assessed across ∼1 ha at 60 sites. Sites were chosen to represent key US ecosystems as part of a scaling strategy deployed by the National Ecological Observatory Network. We measured soil temperature (T_s) and water content (SWC) because these properties mediate biological/biogeochemical processes below- and above-ground, and quantified spatial variability using semivariograms to estimate spatial correlation. We developed quantitative guidelines to inform sample size and sample spacing for future soil studies, e.g., 20 samples were sufficient to measure T_s to within 10% of the mean with 90% confidence at every temperate and sub-tropical site during the growing season, whereas an order of magnitude more samples were needed to meet this accuracy at some high-latitude sites. SWC was significantly more variable than T_s at most sites, resulting in at least 10× more SWC samples needed to meet the same accuracy requirement. Previous studies investigated the relationship between the mean and variability (i.e., sill) of SWC across space at individual sites across time and have often (but not always) observed the variance or standard deviation peaking at intermediate values of SWC and decreasing at low and high SWC. Finally, we quantified how far apart samples must be spaced to be statistically independent. Semivariance structures from 10 of the 12-dominant soil orders across the US were estimated, advancing our continental-scale understanding of soil behavior.     

Reproductive Trade-Offs and Direct Costs for Males in Arthropods

Until 30 years ago, the emphasis on reproductive costs for males was mainly on costs related to mate searching, courtship and fighting with rival males. However, costs for males are substantial and varied and often resemble the more thoroughly studied female reproductive costs. Costs can be referred to as trade-off costs, where investment in reproductive activity comes at the expense of another important activity or fitness component. Investment in reproduction at the expense of longevity and future reproduction is the ultimate cost, because it affects fitness directly. In contrast, flawed performance (e.g., of the immune system) is perceived as a mechanistic trade-off, because it affects fitness indirectly through a mediator (i.e., parasites). Finally, direct costs refer to direct measurements of the energy expenditure during involvement in reproduction-related activities. Both direct and mechanistic trade-off costs often result in decreased longevity compared to unmated males (an ultimate cost). Males incur costs during different reproductive phases: before copulation, when producing sperm, while searching for, courting and copulating with females, and subsequently when guarding females or taking care of offspring. This synthesis follows previous pioneering reviews addressing specific aspects of male costs, but strives to summarize all known male reproductive cost types more comprehensively, including their classification. We suggest several directions for targeted future research. While costs for males have been fairly well described, it is now necessary to uncover the ecological and evolutionary factors responsible for differences between closely related species and systems and to better link between directly-measured costs, mechanistic trade-off costs and ultimate trade-off costs.     

Nested ANOVA for Repeated Surveys with Autocorrelated Errors

Diet surveys in the science of human nutrition involve longitudinal observations on the same individuals repeated at distinct time points. Consequently a common feature of such surveys is inherent correlation between successive observations. In this paper, an autoregressive model of order 1 is assumed for the successive observations on the same individual in a survey. Further, observations on the same individual in two different surveys are assumed to be correlated. The technique of analysis of variance is modified suitably for this set up and applied to two nutritional surveys carried on army recruits.     

Designs Suited for Varietal Trials with Multiple Basals

DAS (1960) gave a method of construction of confounded balanced asymmetrical factorial designs of the type v × 2² by using BIB designs. In the present paper a method has been given for construction of balanced asymmetrical factorial designs of the type (v–t) × 2² by using truncated balanced incomplete block designs obtainable by omitting t treatments. Likewise, partially balanced asymmetrical factorial designs can also be obtained by omitting any particular treatment alongwith its first or second associate treatments from the v treatments of a PBIB design. We can get a large number of new designs not available in literature through this technique. These designs are well suited for varietal trials with multiple basals.     

Destructive thyrotoxicosis in a patient with anaplastic thyroid cancer

A 55-year-old man was admitted to our hospital with an anterior neck tumor, hoarseness, and dysphagia that had continued for a few weeks. He was diagnosed as anaplastic thyroid cancer by fine-needle aspiration cytology. He was treated by external radiation and chemotherapy, but left hemothorax developed and he died of respiratory failure on the 76th day in hospital. On admission, the levels of serum free triiodothyronine (FT3), free thyroxine (FT4), and TSH were 12.8 pg/ml, 4.2 ng/dl, and 0 microU/ml, respectively. The simultaneous thyroidal I-131 uptake rate was 1.2% at 24 hours. The levels of free thyroid hormones fell gradually without antithyroid drugs to result in hypothyroidism (FT3 0.8 pg/ml, FT4 0 ng/dl, and TSH 36 microU/ml). The rapid growth of anaplastic thyroid cancer seemed to be responsible for destructive thyrotoxicosis followed by hypothyroidism in this patient.     

Dealing with Uncertainty in Risk Assessment

This paper is a commentary on Hattis’ three laws of risk assessment. The first law, that “application of standard statistical techniques to a single data set will nearly always reveal only a trivial proportion of the overall uncertainty in the parameter value” is illustrated both by examining the relevance of animal models to man and by a retrospective view of exposure conditions whose importance has only recently been recognized to be important. The second law, that “any estimate of the uncertainty of a parameter value will always itself be more uncertain than the estimate of the parameter value,” is examined in terms of a model addressing multiple levels of uncertainty, e.g., the “uncertainty in the uncertainty”. A argument is made that the number of terms needed for convergence of this uncertainty hierarchy depends on how far from the central tendency of the risk distribution one goes. The further out the “tail” of the distribution, the more terms in the uncertainty hierarchy are needed for convergence. The third law, that “nearly all parameter distribu tions look lognormal, as long as you don't look too closely,” is illustrated with a number of examples. Several reasons are put forward as to why risk variables appear so frequently to be lognormal. Recognition of the lognormal character of variable distributions can provide insight into the proper form for the associated uncertainty distributions.     

Dealing with incongruence in phylogenomic analyses

Incongruence between gene trees is the main challenge faced by phylogeneticists in the genomic era. Incongruence can occur for artefactual reasons, when we fail to recover the correct gene trees, or for biological reasons, when true gene trees are actually distinct from each other, and from the species tree. Horizontal gene transfers (HGTs) between genomes are an important process of bacterial evolution resulting in a substantial amount of phylogenetic conflicts between gene trees. We argue that the (bacterial) species tree is still a meaningful scientific concept even in the case of HGTs, and that reconstructing it is still a valid goal. We tentatively assess the amount of phylogenetic incongruence caused by HGTs in bacteria by comparing bacterial datasets to a metazoan dataset in which transfers are presumably very scarce or absent.We review existing phylogenomic methods and their ability to return to the user, both the vertical (speciation/extinction history) and horizontal (gene transfers) phylogenetic signals.     

Cardiovascular Surveys in Areas with Different Water Supplies

Cardiovascular surveys were conducted in areas with different water supplies. In two areas with hard water 243 men were seen and in three areas with soft water 357 men were seen. There was no evidence of an important difference between the two areas in blood pressure, serum cholesterol, blood sugar, or packed cell volume. A variety of measurements were made on an E.C.G. The P-R interval was significantly longer in the men in the hard water area, and a 40-complex trace showed ventricular extrasystoles in a significantly higher proportion of these men than in those in the soft water area.     

How Enrichment Affects Exploration Trade-Offs in Rats: Implications for Welfare and Well-Being

We propose that a comparative approach to well-being could be the key to understanding ‘the good life.’ Inspired by current theories of human well-being and animal welfare, we designed a novel test of exploration behavior. Environmentally and socially enriched Long-Evans female rats (N = 60) were trained in four simultaneously presented arms of an eight-arm radial-maze. They learned to expect successes in two arms and failures in the other two. After training, 20 animals remained in enriched housing (enrichment-maintenance) while 40 animals were re-housed in standard, isolated conditions (enrichment-removal). Two weeks later, all animals were re-tested in the maze, initially with access to the four familiar arms only. In the final minute, they also had access to the unfamiliar ambiguous-arms. Though both groups showed significant interest in the ambiguous-arms (P<.0001), the enrichment-maintenance group showed a significantly greater exploratory tendency (P<.01) despite having equivalent levels of activity (P>.3). Thus, we show not only that rats will abandon known rewards and incur risk in order to explore, indicating that exploration is valuable in its own right, but also that individuals with (vs. without) enriched housing conditions are more likely to engage in such exploratory behavior. This novel test contributes to the body of knowledge examining the importance of exploration in humans and other animals; implications for animal welfare and human well-being are discussed.     

Trade-Offs Between Efficiency and Robustness in Bacterial Metabolic Networks Are Associated with Niche Breadth

The relation between structure and function in biologic networks is a central point of systems biology research. Key functional features—notably, efficiency and robustness—are linked to the topologic structure of a network, and there appears to be a degree of trade-off between these features, i.e., simulation studies indicate that more efficient networks tend to be less robust. Here, we investigate this issue in metabolic networks from 105 lineages of bacteria having a wide range of ecologies. We take quantitative measurements on each network and integrate this network data with ecologic data using a phylogenetic comparative model. In this setting, we find that biologic conclusions obtained with classical phylogenetic comparative methods are sensitive to correlations between model covariates and phylogenetic branch length. To avoid this problem, we propose a revised statistical framework—hierarchical mixed-effect regression—to accommodate phylogenetic nonindependence. Using this approach, we show that the cartography of metabolic networks does indeed reflect a trade-off between efficiency and robustness. Furthermore, ecologic characteristics related to niche breadth are strong predictors of network shape. Given the broad variation in niche breadth seen among species, we predict that there is no universally optimal balance between efficiency and robustness in bacterial metabolic networks and, thus, no universally optimal network structure. These results highlight the biologic relevance of variation in network structure and the potential role of niche breadth in shaping metabolic strategies of efficiency and robustness. Electronic supplementary material  The online version of this article (doi:) contains supplementary material, which is available to authorized users. Ransom A. Myers Died March 27th, 2007. He will be missed.