ON ESTIMATING PARAMETERS FOR BETA DISTRIBUTIONS

The purpose of this paper is to comment on and give historical perspective to two methdologies for estimating parameters of beta distributions. Fielitz and Myers [3] [4] developed and advocated a methodology using the method of moments, while Romesburg [20] advocated a methodology usingthe method of maximum likelihood. However, what Fielitz and Myers presented as new research and suggested as an area needing further study is ground already trampled. The authors have prepared a graph to underline the superiority of the maximum likelihood method in fitting beta distributions.     

Does product complexity matter for competition in experimental retail markets?

We describe a first experiment on whether product complexity affects competition and consumers in retail markets. We are unable to detect a significant effect of product complexity on prices, except insofar as the demand elasticity for complex products is higher. However, there is qualified evidence that complex products have the potential to induce consumers to buy more than they would otherwise. In this sense, consumer exploitability in quantities cannot be ruled out. We also find evidence for shaping effects: consumers’ preferences are shaped by past experience with prices, and firms may in principle exploit this to sell more.     

Transforming sub-Saharan Africa

This article provides an analysis of my personal experience and research on the need for sub-Saharan African economic transformation. It contains relevant references to econometric modeling which are consistent with this analysis. After presenting the causes of African failures, ranging from the alleged role of slavery and colonization, to trade composition and relative decline, to present day problems of governance, low level of foreign investment in Africa, and mass unemployment, the article concludes with an agenda for the transformation of sub-Saharan Africa.     

Identifying the most successful dose (MSD) in dose-finding studies in cancer

For a dose finding study in cancer, the most successful dose (MSD), among a group of available doses, is that dose at which the overall success rate is the highest. This rate is the product of the rate of seeing non-toxicities together with the rate of tumor response. A successful dose finding trial in this context is one where we manage to identify the MSD in an efficient manner. In practice we may also need to consider algorithms for identifying the MSD which can incorporate certain restrictions, the most common restriction maintaining the estimated toxicity rate alone below some maximum rate. In this case the MSD may correspond to a different level than that for the unconstrained MSD and, in providing a final recommendation, it is important to underline that it is subject to the given constraint. We work with the approach described in O'Quigley et al. [Biometrics 2001; 57(4):1018-1029]. The focus of that work was dose finding in HIV where both information on toxicity and efficacy were almost immediately available. Recent cancer studies are beginning to fall under this same heading where, as before, toxicity can be quickly evaluated and, in addition, we can rely on biological markers or other measures of tumor response. Mindful of the particular context of cancer, our purpose here is to consider the methodology developed by O'Quigley et al. and its practical implementation. We also carry out a study on the doubly under-parameterized model, developed by O'Quigley et al. but not     

A Multiple Imputation Approach to Linear Regression with Clustered Censored Data

Weextend Wei and Tanner's (1991) multiple imputation approach insemi-parametric linear regression for univariate censored datato clustered censored data. The main idea is to iterate the followingtwo steps: 1) using the data augmentation to impute for censoredfailure times; 2) fitting a linear model with imputed completedata, which takes into consideration of clustering among failuretimes. In particular, we propose using the generalized estimatingequations (GEE) or a linear mixed-effects model to implementthe second step. Through simulation studies our proposal comparesfavorably to the independence approach (Lee et al., 1993), whichignores the within-cluster correlation in estimating the regressioncoefficient. Our proposal is easy to implement by using existingsoftwares.     

Bounding convergence rates for Markov chains: An example of the use of computer algebra

Kolassa and Tanner (J. Am. Stat. Assoc. (1994) 89, 697–702) present the Gibbs-Skovgaard algorithm for approximate conditional inference. Kolassa (Ann Statist. (1999), 27, 129–142) gives conditions under which their Markov chain is known to converge. This paper calculates explicity bounds on convergence rates in terms calculable directly from chain transition operators. These results are useful in cases like those considered by Kolassa (1999).     

A modified likelihood ratio test for homogeneity in finite mixture models

Testing for homogeneity in finite mixture models has been investigated by many researchers. The asymptotic null distribution of the likelihood ratio test (LRT) is very complex and difficult to use in practice. We propose a modified LRT for homogeneity in finite mixture models with a general parametric kernel distribution family. The modified LRT has a χ-type of null limiting distribution and is asymptotically most powerful under local alternatives. Simulations show that it performs better than competing tests. They also reveal that the limiting distribution with some adjustment can satisfactorily approximate the quantiles of the test statistic, even for moderate sample sizes.     

Confidence intervals for secondary parameters following a sequential test

In sequential studies, formal interim analyses are usually restricted to a consideration of a single null hypothesis concerning a single parameter of interest. Valid frequentist methods of hypothesis testing and of point and interval estimation for the primary parameter have already been devised for use at the end of such a study. However, the completed data set may warrant a more detailed analysis, involving the estimation of parameters corresponding to effects that were not used to determine when to stop, and yet correlated with those that were. This paper describes methods for setting confidence intervals for secondary parameters in a way which provides the correct coverage probability in repeated frequentist realizations of the sequential design used. The method assumes that information accumulates on the primary and secondary parameters at proportional rates. This requirement will be valid in many potential applications, but only in limited situations in survival analysis.     

Measuring people's trust

Summary. We measure trust and trustworthiness in British society with a newly designed experiment using real monetary rewards and a sample of the British population. The study also asks the typical survey question that aims to measure trust, showing that it does not predict 'trust' as measured in the experiment. Overall, about 40% of people were willing to trust a stranger in our experiment, and their trust was rewarded half of the time. Analysis of variation in the trust behaviour in our survey suggests that trusting is more likely if people are older, their financial situation is either 'comfortable' or 'difficult' compared with 'doing alright' or 'just getting by', they are a homeowner or they are divorced, separated or never married compared with those who are married or cohabiting. Trustworthiness also is more likely among subjects who are divorced or separated relative to those who are married or cohabiting, and less likely among subjects who perceive their financial situation as 'just getting by' or 'difficult'. We also analyse the effect of attitudes towards risks on trust.     

Sparse additive models

Summary.  We present a new class of methods for high dimensional non-parametric regression and classification called sparse additive models. Our methods combine ideas from sparse linear modelling and additive non-parametric regression. We derive an algorithm for fitting the models that is practical and effective even when the number of covariates is larger than the sample size. Sparse additive models are essentially a functional version of the grouped lasso of Yuan and Lin. They are also closely related to the COSSO model of Lin and Zhang but decouple smoothing and sparsity, enabling the use of arbitrary non-parametric smoothers. We give an analysis of the theoretical properties of sparse additive models and present empirical results on synthetic and real data, showing that they can be effective in fitting sparse non-parametric models in high dimensional data.