Estimation and testing with overdispersed proportions using the beta-logistic regression model of Heckman and Willis

Methods are presented for modeling dose-related effects in proportion data when extra-binomial variability is a concern. Motivation is taken from experiments in developmental toxicology, where similarity among conceptuses within a litter leads to intralitter correlations and to overdispersion in the observed proportions. Appeal is made to the well-known beta-binomial distribution to represent the overdispersion. From this, an exponential function of the linear predictor is used to model the dose-response relationship. The specification was introduced previously for econometric applications by Heckman and Willis; it induces a form of logistic regression for the mean response, together with a reciprocal biexponential model for the intralitter correlation. Large-sample, likelihood-based methods for estimating and testing the joint proportion-correlation response are studied. A developmental toxicity data set illustrates the methods.     

Birth month associations with height,head circumference,and limb lengths among peruvian children

Associations between season of birth and body size, morbidity, and mortality have been widely documented, but it is unclear whether different parts of the body are differentially sensitive, and if such effects persist through childhood. This may be relevant to understanding the relationship between early life environment and body size and proportions. We investigated associations between birth month and anthropometry among rural highland (n = 162) and urban lowland (n = 184) Peruvian children aged 6 months to 8 years. Stature; head‐trunk height; total limb, ulna, tibia, hand, and foot lengths; head circumference; and limb measurements relative to head‐trunk height were converted to internal age‐sex‐specific z scores. Lowland and highland datasets were then analyzed separately for birth month trends using cosinor analysis, as urban conditions likely provide a more consistent environment compared with anticipated seasonal variation in the rural highlands. Among highland children birth month associations were significant most strongly for tibia length, followed by total lower limb length and stature, with a peak among November births. Results were not significant for other measurements or among lowland children. The results suggest a prenatal or early postnatal environmental effect on growth that is more marked in limb lengths than trunk length or head size, and persists across the age range studied. We suggest that the results may reflect seasonal variation in maternal nutrition in the rural highlands, but other hypotheses such as variation in maternal vitamin D levels cannot be excluded. Am J Phys Anthropol 154:115–124, 2014. © 2014 The Authors. American Journal of Physical Anthropology Published by Wiley Periodicals, Inc.     

An exploration of adult body shape and limb proportions at Kellis 2, Dakhleh Oasis,Egypt

Several studies have shown that the human body generally conforms to the ecogeographical expectations of Bergmann's and Allen's rules; however, recent evidence suggests that these expectations may not hold completely for some populations. Egypt is located at the crossroads of Sub‐Saharan Africa, Southern Europe, and the Near East, and gene flow among groups in these regions may confound ecogeographical patterning. In this study, we test the fit of the adult physique of a large sample (N = 163) of females and males from the Kellis 2 cemetery (Dakhleh Oasis, Egypt) against ecogeographical predictions. Body shape (i.e., body mass relative to stature) was assessed by the femur head diameter to bicondylar femur length index (FHD/BFL), and brachial and crural indices were calculated to examine intralimb proportions. Body shape in the Kellis 2 sample is not significantly different from high‐latitude groups and a Lower Nubian sample, and intralimb proportions are not significantly different from mid‐latitude and other low‐latitude groups. This study demonstrates the potential uniqueness of body shape and intralimb proportions in an ancient Egyptian sample, and further highlights the complex relationship between ecogeographic patterning and adaptation. Am J Phys Anthropol 153:496–505, 2014. © 2013 Wiley Periodicals, Inc.     

Effects of plot vegetation diversity and spatial scale on Coccinella septempunctata movement in the absence of prey

The influence that vegetation diversity and the spatial scale of that diversity exert on insect behavior has increasingly been explored in the ecological literature, but relatively few experiments have explicitly incorporated both factors in experimental treatments. We conducted a field study designed to explore the effect of both of these factors on insect movement behavior in a broccoli agroecosystem. We caught and released seven‐spotted ladybird beetles (Coccinella septempunctata L.) in plots containing different degrees of vegetation diversity at two different spatial scales in which prey had been removed. Beetle movement was recorded at timed intervals, and move lengths and turning angles were used to generate discrete path maps for each beetle. Observed mean beetle net squared displacements were compared with predicted net squared displacements, and 95% confidence intervals were generated using a bootstrap method described by Turchin (1998 ) [Quantitative Analysis of Movement: Measuring and Modeling Population Redistribution in Animals and Plants. Sinauer Associates Inc., Sunderland, MA.]. Predicted net squared displacements underestimated beetle movement in smaller plots with both low and higher vegetation diversity for the first five move lengths, whereas no significant difference between observed and predicted net squared displacement for beetles in larger plots of either level of vegetation diversity were detected. These findings highlight the need for a better understanding of how natural enemies are influenced by vegetation diversity and the spatial scale of that vegetation in agroecosystems. The implications of these results for biological control are discussed.     

Limb-size proportions in Australopithecus afarensis and Australopithecus africanus

Previous analyses have suggested that Australopithecus africanus possessed more apelike limb proportions than Australopithecus afarensis. However, due to the errors involved in estimating limb length and body size, support for this conclusion has been limited. In this study, we use a new Monte Carlo method to (1) test the hypothesis that A. africanus had greater upper:lower limb-size proportions than A. afarensis and (2) assess the statistical significance of interspecific differences among these taxa, extant apes, and humans. Our Monte Carlo method imposes sampling constraints that reduce extant ape and human postcranial measurements to sample sizes comparable to the fossil samples. Next, composite ratios of fore- and hindlimb geometric means are calculated for resampled measurements from the fossils and comparative taxa. Mean composite ratios are statistically indistinguishable (alpha=0.05) from the actual ratios of extant individuals, indicating that this method conserves each sample's central tendency. When applied to the fossil samples, upper:lower limb-size proportions in A. afarensis are similar to those of humans (p=0.878) and are significantly different from all great ape proportions (p< or =0.034), while Australopithecus africanus is more similar to the apes (p> or =0.180) and significantly different from humans and A. afarensis (p< or =0.031). These results strongly support the hypothesis that A. africanus possessed more apelike limb-size proportions than A. afarensis, suggesting that A. africanus either evolved from a more postcranially primitive ancestor than A. afarensis or that the more apelike limb-size proportions of A. africanus were secondarily derived from an A. afarensis-like ancestor. Among the extant taxa, limb-size proportions correspond with observed levels of forelimb- and hindlimb-dominated positional behaviors. In conjunction with detailed anatomical features linked to arboreality, these results suggest that arboreal posture and locomotion may have been more important components of the A. africanus behavioral repertoire relative to that of A. afarensis.     

Evolutionary reversals of limb proportions in early hominids? Evidence from KNM-ER 3735 (Homo habilis)

Upper-to-lower limb proportions of Homo habilis are often said to be more ape-like than those of its reputed ancestor, Australopithecus afarensis. Such proportions would either imply multiple evolutionary reversals or parallel development of a relatively short upper limb in A. afarensis and later Homo. However, assessments of limb proportions are complicated by the fragmentary nature of the two known H. habilis skeletons, OH 62 and KNM-ER 3735. Initially, KNM-ER 3735 was compared to A.L. 288-1 (A. afarensis) using a single modern human and chimpanzee as reference. Here, based on a larger comparative sample, we find that the relative size of the distal humerus, radial head, and shaft of both KNM-ER 3735 and A.L. 288-1 lie within the range of variation of modern humans, whereas their sacra are small as is the case for all early hominids. In addition, their manual phalanges are similar in having a gracile base but robust midshaft. Contrary to earlier studies, the fossils are not differentiable from each other statistically with respect to all features listed above. On the other hand, they differ in robusticity of the scapular spine and relative length of the radial neck. An exact randomization test suggests only a very low probability of finding a similar degree of difference within a single species of extant hominoids. In contrast to the consensus view, we conclude that A.L. 288-1 had a short, human-like forearm, whereas KNM-ER 3735 possessed a distinctly longer forearm and more powerful shoulder girdle. This interpretation fits with earlier conclusions that suggested human-like humerofemoral proportions but chimpanzee-like brachial proportions for Homo habilis. Thus, the scenario of a unidirectional, progressive change in limb proportions within the hominid lineage is not supported by our work.     

FDR control by the BH procedure for two-sided correlated tests with implications to gene expression data analysis

The multiple testing problem attributed to gene expression analysis is challenging not only by its size, but also by possible dependence between the expression levels of different genes resulting from coregulations of the genes. Furthermore, the measurement errors of these expression levels may be dependent as well since they are subjected to several technical factors. Multiple testing of such data faces the challenge of correlated test statistics. In such a case, the control of the False Discovery Rate (FDR) is not straightforward, and thus demands new approaches and solutions that will address multiplicity while accounting for this dependency. This paper investigates the effects of dependency between bormal test statistics on FDR control in two-sided testing, using the linear step-up procedure (BH) of Benjamini and Hochberg (1995). The case of two multiple hypotheses is examined first. A simulation study offers primary insight into the behavior of the FDR subjected to different levels of correlation and distance between null and alternative means. A theoretical analysis follows in order to obtain explicit upper bounds to the FDR. These results are then extended to more than two multiple tests, thereby offering a better perspective on the effect of the proportion of false null hypotheses, as well as the structure of the test statistics correlation matrix. An example from gene expression data analysis is presented.     

Behavioral testing of antidepressant compounds: an analysis of crossover design for correlated binary data

The differential reinforcement of low-rate 72 seconds schedule (DRL-72) is a standard behavioral test procedure for screening potential antidepressant compounds. The protocol for the DRL-72 experiment, proposed by Evenden et al. (1993), consists of using a crossover design for the experiment and one-way ANOVA for the statistical analysis. In this paper we discuss the choice of several crossover designs for the DRL-72 experiment and propose to estimate the treatment effects using either generalized linear mixed models (GLMM) or generalized estimating equation (GEE) models for clustered binary data.