Search

How does one assess an individual player's contribution to his team's score? Baseball statisticians have attempted to answer this question countless times. It's not just the runs he scores himself [...]

Major League Baseball (MLB) is awash in advanced metrics that more reliably describe key aspects of players' offensive and defensive performance compared to 'traditional' statistics. PITCHf/x and Statcast greatly contributed [...]

The measurement of the magnitude of association between a nominal independent variable and an ordinal dependent variable is an important, but neglected, component in psychological research. Two measures of nominal--ordinal association are described and compared. Resampling permutation methods are utilized to compute probability values for both measures.

An analysis on estimation of the mean and variance of the future price, using cobweb model with respect to risk averse producers, is presented.

We introduce a novel statistical procedure for clustering categorical data based on Hamming distance (HD) vectors. The proposed method is conceptually simple and computationally straightforward, because it does not require any specific statistical models or any convergence criteria. Moreover, unlike most currently existing algorithms that compute the class membership or membership probability for every data point at each iteration, our algorithm sequentially extracts clusters from the given dataset. That is, at each iteration our algorithm strives to identify only one cluster, which will then be deleted from the dataset at the next iteration; this procedure repeats until there are no more significant clusters in the remaining data. Consequently, the number of clusters can be determined automatically by the algorithm. As for the identification and extraction of a cluster, we first locate the cluster center by using a Pearson chi-squared-type statistic on the basis of HD vectors. The partition of the dataset produced by our algorithm is unique and insensitive to the input order of data points. The performance of the proposed algorithm is examined using both simulated and real world datasets. Comparisons with two well-known clustering algorithms. K-modes and AutoClass, show that the proposed algorithm substantially outperforms these competitors, with the classification rate or the information gain typically improved by several orders of magnitude. Computational complexity and run time comparisons are also provided. KEY WORDS: AutoClass; Categorical data; Clustering; Computational complexity; Distance vector; Hamming distance; K-modes; Modified chi-squared statistic.

Calibration is commonly used in survey sampling to include auxiliary information at the estimation stage of a population parameter. Calibrating the observation weights on population means (totals) of a set of auxiliary variables implies building weights that when applied to the auxiliaries give exactly their population mean (total). Implicitly, calibration techniques rely on a linear relation between the survey variable and the auxiliary variables. However, when auxiliary information is available for all units in the population, more complex modeling can be handled by means of model calibration; auxiliary variables are used to obtain fitted values of the survey variable for all units in the population, and estimation weights are sought to satisfy calibration constraints on the fitted values population mean, rather than on the auxiliary variables one. In this work we extend model calibration considering more general superpopulation models and use nonparametric methods to obtain the fitted values on which to calibrate. More precisely, we adopt neural network learning and local polynomial smoothing to estimate the functional relationship between the survey variable and the auxiliary variables. Under suitable regularity conditions, the proposed estimators are proven to be design consistent. The moments of the asymptotic distribution are also derived, and a consistent estimator of the variance of each distribution is then proposed. The performance of the proposed estimators for finite-size samples is investigated by means of simulation studies. An application to the assessment of the ecological conditions of streams in the mid-Atlantic highlands in the United States is also carried out. KEY WORDS: Auxiliary information; Generalized regression estimator; Local polynomials; Model-assisted approach; Neural networks; Nonparametric regression.

Statistical rituals largely eliminate statistical thinking in the social sciences even as rituals are indispensable for identification with social groups. They should be the subject rather than the procedure of science. The typical 'null ritual' consists of the following steps: setting up a statistical null process, use the 5% significance level for rejecting the null and accepting your hypothesis, and always perform this procedure.

DNA microarrays are important devices for high throughput measurements of gene expression, but no rational foundation has been established for understanding the sources of within-chip statistical error. We designed a specialized chip and protocol to investigate the distribution and magnitude of within-chip errors and discovered that, as expected from theoretical expectations, measurement errors follow a Lorentzian-like distribution, which explains the widely observed but unexplained ill-reproducibility in microarray data. Using this specially designed chip, we examined a data set of repeated measurements to extract estimates of the distribution and magnitude of statistical errors in DNA microarray measurements. Using the common 'ratio of medians' method, we find that the measurements follow a Lorentzian-like distribution, which is problematic for subsequent analysis. We show that a method of analysis dubbed 'median of ratios' yields a more Gaussian-like distribution of errors. Finally, we show that the bootstrap algorithm can be used to extract the best estimates of the error in the measurement. Quantifying the statistical error in such measurements has important applications for estimating significance levels, clustering algorithms, and process optimization.

Few concepts in the social sciences have wielded more discriminatory power over the status of knowledge claims than that of statistical significance. Currently operationalized as a = 0.05, statistical significance frequently separates publishable from nonpublishable research, renewable from nonrenewable grants, and, in the eyes of many, experimental success from failure. If literacy is envisioned as a sort of competence in a set of social and intellectual practices, then scientific literacy must encompass the realization that this cardinal arbiter of social scientific knowledge was not born out of an immanent logic of mathematics but socially constructed and reconstructed in response to sociohistoric conditions.

This article evaluates the variation in return volatility using a Fourier Flexible Form regression framework rather than the usual variance-ratio tests, which the authors find to be misleading. Results are illustrated through tests for changes in the intraday volatility pattern after removing trade restrictions in Tokyo, finding no changes outside the lunch period.

Research is presented concerning the placement of a monetary value on a statistical life. The use of empirical estimates for life expectancy where consumption is not constant throughout the life cycle is discussed.

THE APPLICATION OF SCIENTIFIC METHODS to military operations in WWII came to be known as operations research (O.R.). This was one among the many applications of science in what was [...]

A new study of Merriam's wild turkey hens with young in the Black Hills of South Dakota compares six statistical techniques for analyzing resource selection data. Habitat was the single variable used for the comparisons. The effects of varying the available habitats on the statistical methods was also examined. The study shows that method choice for evaluating resource selection affects inferences about wildlife habitat requirements.

An analysis of production and futures equilibria in the context of risk aversion reveals that results are affected by assumptions of speculator non-productivity and entry/exit cost imposition. The study's results are in contrast with previous studies on hedging as only risk aversion behavior is sufficient to derive results. The study also reveals that econometric tests may be developed and applied using approaches of Frank and Antonovitz, among others.

An evaluation of Bertrand and Cournot equilibrium in a homogenous product market reveals that the point of comparison of the two equilibrium lies in the sensitivity of Cournot statistics to market sharing rules. The study also shows that results using Bertrand equations may not hold when costs are asymmetric if the 'capacity sharing rule' does not apply to the homogenous market unlike in the diversified market.

There is an urgent need to reform the US statistics system. Failure to address the fundamental and structural problems of the system will result in an economic data crisis. Confidence in economic data, which make the markets work, will decline and eventually be lost if data continue to show quirks and inconsistencies and to involve contorted explanations. Such a crisis can be averted given the will to update economic measurements, an organization structure that will facilitate reform and adequate resources to undertake this task.

Four statistical methods are applied to a set of synthetic data to evaluate their effectiveness in validating estimates of racial disparity in the allocation of public contracts in the presence of relevant control variables. The ordinary least square (OLS) regression, tobit and logit (TAL) analysis and a multivariate procedure for comparing expected and observed outcomes are used to remove spurious disparities. The TAL model provide more accurate and sensitive estimates than OLS regression. Analysis of expected and observed outcomes produces more sensitive probability estimates than the TAL models.

Businesses should be concerned about the growing number of criticism against the accuracy of national statistics. Economic policy and business decisions are based on these numbers. Thus, inaccurate data leads to incorrect policies and decisions. Certain measures can be taken to control the decline of national statistics. These include the combination of agencies under one organization, contracting of research services and increasing the accountability for statistical information.

Sociologist Emile Durkheim opposed the popular assertion of many statisticians of his time that social problems can be explained through statistics and correlations. He believed that social phenomena revealed by statistics had deeper and more complex social explanations. Durkheim represented a sociological thought which was prevalent during the 19th Century among German sociologists. However, such a sociological philosophy eventually died out with the dominance of English sociological methods such as biometry.

A model based on data from the Challenger disaster of 1986, to find the relationship between launch temperature and damaged O-rings, helped students understand the role of statistics in decision making. The students were provided with the data the scientists used to decide when to launch the shuttle. They found that at a launch temperature of 31 degrees Fahrenheit, four or five O-rings could be damaged, and that the scientists overlooked the information from flights that had no O-ring problems.

Objectives. This study used a recent national population survey on childhood and adolescent nonfatal injuries to investigate the effects of recall bias on estimating annual injury rates. Strategies to adjust for recall bias are recommended. Methods. The 1988 Child Health Supplement to the National Health Interview Survey collected 12-month recall information on injuries that occurred to a national sample of 17 110 children aged 0 through 17 years. Using information on timing of interviews and reported injuries, estimated annual injury rates were calculated for 12 accumulative recall periods (from 1 to 12 months). Results. The data show significantly declining rates, from 24.4 per 100 for a 1-month recall period to 14.7 per 100 for a 12-month recall period. The largest declines were found for the 0- through 4-year-old age group and for minor injuries. Rates of injuries that caused a school loss day, a bed day, surgery, or hospitalization showed higher stability throughout recall periods. Conclusions. Varying recall periods have profound effects on the patterns of childhood injury epidemiology that emerge from the data. Recall periods of between 1 and 3 months are recommended for use in similar survey settings., Parents' recall of children's injuries may vary significantly by the length of time between the injury and the interview. Such variability has impact on public health programs designed to prevent and treat injuries. Bias in responding to survey questions about children's accidents and injuries can be introduced from two sources. One is loss of information because of failure to recall older events. The second is what is known as telescoping. This is the tendency to recall events as happening more recently than they actually did. An analysis of recall bias shows that both these tendencies affect parents' responses to questions in the National Health Interview Survey. The injury rates vary by the length of time the recall must cover. Thus the estimate of an overall rate for children in the US can be as high as one for every three children per year if the recall period is every two weeks, to as low as one in every seven children if the recall is done for the entire year. The effect is less for serious injuries and higher for minor injuries.

This paper proposes a new procedure for computing the Cox statistic for tests of nonnested hypotheses using the method of stochastic simulation. The procedure is applicable to a wide class of probability distributions, is relatively simple to implement, and does not require an analytic derivation of the pseudo-true estimators that enter the Cox test statistic. The paper also contains an application of the proposed method to the test of probit versus logistic formulations of the univariate binary choice models. The empirical results show that the simulation method works reasonably well even for a moderate number of replications.

Policymakers and researchers interested in prevalence estimation often seek practical guidance on how to obtain and interpret quality estimates. Although the preceding chapters have addressed aspects of these issues, it is helpful to summarize some of the common problems and to provide practical solutions. This chapter assembles a set of questions and abbreviated answers in areas of policy needs, prevalence definitions, data characteristics and availability, and estimation techniques. Readers are directed to extensive discussion in other relevant chapters.

The author describes several of the most important quantitative procedures for estimating the size of an unobserved or partially observed population, with specific application to the estimation of the prevalence of drug use. The methods discussed include synthetic estimation, truncated Poisson estimates, multiple-capture surveys in both closed populations (the capture-recapture model and log-linear models) and open populations (the Jolly-Seber model and Markov models,) and, more briefly, system dynamics models.

Issues about data analysis were examined in a survey conducted on 551 psychological researchers. A questionaire served to assess the subjects' views about the reliability of Type1 and 2 errors, analysis of interactions, contrast analysis and power and effect siza in replication. Results implied that active researchers did not have sufficient comprehension of statistical issues as shown by low accuracy scores. The findings did not adequately assess statistical knowledge but it initiated issues that are important in statistical applications in psychological research.

This article outlines a general, argument-based approach to validation, develops an interpretive argument for a placement test as an example, and examines some key properties of interpretive arguments. Validity is associated with the interpretation assigned to test scores rather than with the test scores or the test. The interpretation involves an argument leading from the scores to score-based statements or decisions, and the validity of the interpretation depends on the plausibility of this interpretive argument. The interpretive arguments associated with most test-score interpretations involve multiple inferences and assumptions. An explicit recognition of the inferences and assumptions in the interpretive argument makes it possible to identify the kinds of evidence needed to evaluate the argument. Evidence for the inferences and assumptions in the argument supports the interpretation, and evidence against any part of the argument casts doubt on the interpretation.

Risk Within Reason SOCIETY'S SYSTEM FOR MANAGING RISKS TO LIFE AND LIMB IS deeply flawed. We overreact to some risks and virtually ignore others. Often too much weight is placed [...]