Showing papers in "Journal of The Royal Statistical Society Series A-statistics in Society in 2011"

The use of propensity scores to assess the generalizability of results from randomized trials

Abstract: Randomized trials remain the most accepted design for estimating the effects of interventions, but they do not necessarily answer a question of primary interest: Will the program be effective in a target population in which it may be implemented? In other words, are the results generalizable? There has been very little statistical research on how to assess the generalizability, or "external validity," of randomized trials We propose the use of propensity-score-based metrics to quantify the similarity of the participants in a randomized trial and a target population In this setting the propensity score model predicts participation in the randomized trial, given a set of covariates The resulting propensity scores are used first to quantify the difference between the trial participants and the target population, and then to match, subclassify, or weight the control group outcomes to the population, assessing how well the propensity score-adjusted outcomes track the outcomes actually observed in the population These metrics can serve as a first step in assessing the generalizability of results from randomized trials to target populations This paper lays out these ideas, discusses the assumptions underlying the approach, and illustrates the metrics using data on the evaluation of a schoolwide prevention program called Positive Behavioral Interventions and Supports

New Orleans business recovery in the aftermath of Hurricane Katrina

Abstract: Summary. We analyse decisions to reopen in the aftermath of Hurricane Katrina made by business establishments on major business thoroughfares in New Orleans by using a spatial probit methodology. Our approach allows for interdependence between decisions to reopen by one establishment and those of its neighbours. There is a large literature on the role that is played by spatial dependence in firm location decisions, and we find evidence of strong dependence in decisions by firms to reopen in the aftermath of a natural disaster such as Hurricane Katrina. This interdependence has important statistical implications for how we analyse business recovery after disasters, as well as government aid programmes.

Small area estimation using a reweighting algorithm

Abstract: Summary. The paper describes a method of small area estimation which uses a reweighting algorithm to reweight survey data to a number of known totals (benchmarks) for small areas. The method has so far been used to estimate small area poverty rates and housing stress. The method gives poverty rates for small areas that are similar to those available from the 2006 Australian census, when the same definition of poverty was used. Various methods of validating the poverty rates have been used, including aggregating the poverty rates to a larger area and comparing them with official Australian Bureau of Statistics estimates from a survey, and applying the spatial microsimulation to larger areas and comparing with official Australian Bureau of Statistics survey results. Both these tests show that the estimates are comparable and fairly robust for most states in Australia.

Validating the use of anchoring vignettes for the correction of response scale differences in subjective questions

Abstract: Comparing self-assessed indicators of subjective outcomes (such as health, or job satisfaction) across countries or socio-economic groups is often hampered by the fact that different groups use different response scales. Anchoring vignettes are used as a tool to identify and correct for such differences. The paper develops a model in which objective measurements are used to validate such vignette-based corrections. The model is illustrated with an application to objective and subjective self-assessments of drinking behaviour by students in the Republic of Ireland. Model comparisons using the Akaike information criterion favour a specification with response consistency and vignette-corrected response scales. Put differently, vignette-based corrections appear quite effective in bringing objective and subjective measures closer together.

Towards more accessible conceptions of statistical inference

Abstract: Summary. There is a compelling case, based on research in statistics education, for first courses in statistical inference to be underpinned by a staged development path. Preferably over a number of years, students should begin working with precursor forms of statistical inference, much earlier than they now do. A side benefit is giving younger students more straightforward and more satisfying ways of answering interesting real world questions. We discuss the issues that are involved in formulating precursor versions of inference and then present some specific and highly visual proposals. These build on novel ways of experiencing sampling variation and have intuitive connections to the standard formal methods of making inferences in first university courses in statistics. Our proposal uses visual comparisons to enable the inferential step to be made without taking the eyes off relevant graphs of the data. This allows the time and conceptual distances between questions, data and conclusions to be minimized, so that the most critical linkages can be made. Our approach was devised for use in high schools but is also relevant to adult education and some introductory tertiary courses.

A multilevel modelling approach to investigating the predictive validity of editorial decisions: do the editors of a high profile journal select manuscripts that are highly cited after publication?

Abstract: Summary. Scientific journals must deal with the following questions concerning the predictive validity of editorial decisions. Is the best scientific work selected from submitted manuscripts? Does selection of the best manuscripts also mean selecting papers that after publication show top citation performance within their fields? Taking the journal Angewandte Chemie International Edition as an example, this study proposes a new methodology for investigating whether manuscripts that are most worthy of publication are in fact selected validly. First, the influence on citation of the accepted and rejected but then published elsewhere manuscripts was appraised on the basis of percentile impact classes scaled in a subfield of chemistry and, second, the association between the decisions on selection and the influence on citation of the manuscripts was determined by using a multilevel logistic regression for ordinal categories. This approach has many advantages over methodologies that were used in previous research studies on the predictive validity of editorial selection decisions.

EUROMIND: a monthly indicator of the euro area economic conditions

Abstract: Summary. Continuous monitoring of the evolution of the economy is fundamental for the decisions of public and private decision makers. The paper proposes EUROMIND, which is a new monthly indicator of the euro area economic conditions, based on tracking real gross domestic product monthly, relying on information provided in the Eurostat Euro-IND database. EUROMIND has several original economic and statistical features. First, it considers both the output and the expenditure sides of the economy, as it provides a monthly estimate of the value added of the six branches of economic activity and of the main gross domestic product components by type of expenditure (final consumption, gross capital formation and net exports), and combines the estimates with optimal weights reflecting their relative precision. Second, the indicator is based on information at both the monthly and the quarterly level, modelled with a dynamic factor specification cast in state space form. Third, since estimation of the multivariate dynamic factor model with mixed frequency data can be numerically complex, computational efficiency is achieved by implementing univariate filtering and smoothing procedures. Finally, special attention is paid to chain linking and its implications, via a multistep procedure that exploits the additivity of the volume measures expressed at the prices of the previous year.

Biases in the healthcare luxury good hypothesis?: a meta-regression analysis

Abstract: Summary. Although a growing literature examining the relationship between income and health expenditures suggests that healthcare is a luxury good, this conclusion is debatable owing to heterogeneity of the existing results. The paper tests the luxury good hypothesis (namely that income elasticity exceeds 1) by using meta-regression analysis, taking into consideration publication selection and aggregation bias. The findings suggest that publication bias exists, which is a result that is robust to the meta-regression model employed. Publication selection and aggregation bias also appear to play a role in the generation of estimates. The corrected estimates of income elasticity range from 0.4 to 0.8, which cast serious doubt on the validity of the luxury good hypothesis.

Measuring inequality using censored data: a multiple-imputation approach to estimation and inference

Abstract: Summary. To measure income inequality with right-censored (top-coded) data, we propose multiple-imputation methods for estimation and inference. Censored observations are multiply imputed using draws from a flexible parametric model fitted to the censored distribution, yielding a partially synthetic data set from which point and variance estimates can be derived using complete-data methods and appropriate combination formulae. The methods are illustrated using US Current Population Survey data and the generalized beta of the second kind distribution as the imputation model. With Current Population Survey internal data, we find few statistically significant differences in income inequality for pairs of years between 1995 and 2004. We also show that using Current Population Survey public use data with cell mean imputations may lead to incorrect inferences. Multiply-imputed public use data provide an intermediate solution.

Is there a "heat or eat" trade-off in the UK?

Abstract: Do households cut back on food spending to finance the additional cost of keeping warm during spells of unseasonably cold weather? For households which cannot smooth consumption over time, we describe how cold weather shocks are equivalent to income shocks. We merge detailed household level expenditure data from older households with historical regional weather information. We find evidence that the poorest of older households cannot smooth fuel spending over the worst temperature shocks. Statistically significant reductions in food spending occur in response to winter temperatures 2 or more standard deviations colder than expected, which occur about 1 winter month in 40; reductions in food expenditure are considerably larger in poorer households.

Education-related inequity in healthcare with heterogeneous reporting of health.

Abstract: Reliance on self-rated health to proxy medical need can bias estimation of education-related inequity in healthcare utilization. We correct this bias both by instrumenting self-rated health with objective health indicators and by purging self-rated health of reporting heterogeneity that is identified from health vignettes. Using data on elderly Europeans, we find that instrumenting self-rated health shifts the distribution of visits to a doctor in the direction of inequality favouring the better educated. There is a further, and typically larger, shift in the same direction when correction is made for the tendency of the better educated to rate their health more negatively.

Using field process data to predict best times of contact conditioning on household and interviewer influences

Abstract: Establishing contact is an important part of the response process and effective interviewer calling behaviours are critical in achieving contact and subsequent co-operation. The paper investigates best times of contact for different types of households and the influence of the interviewer on establishing contact. Recent developments in the survey data collection process have led to the collection of so-called field process data or paradata, which greatly extend the basic information on interviewer calls. The paper develops a multilevel discrete time event history model based on interviewer call record data to predict the likelihood of contact at each call. The results have implications for survey practice and inform the design of effective interviewer calling times, including responsive survey designs.

Qualitative business surveys: signal or noise?

Abstract: The paper identifies the information content at the firm level of qualitative business survey data by examining the consistency between these data and the quantitative data that are provided by the same respondents to the UK's Office for National Statistics in official surveys. Since the qualitative data are published ahead of the quantitative data the paper then assesses the ability of the qualitative data to predict the firm level quantitative data.

Forecasting mortality in the event of a structural change

Abstract: In recent decades, life expectancy in developed countries has risen to historically unprecedented levels driven by unforeseen declines in rates of mortality. The prospects of further reductions are of fundamental importance in various areas. In this context, we consider the problem of forecasting future mortality and life expectancy in the event of a structural change. We show how recent advances in statistical testing for structural changes can be used to arrive at properly specified models for the general level of mortality in the context of the Lee-Carter model. Specifically, the results of tests for a change in the trend of the index of mortality and for the presence of a unit root are used to identify appropriate forecasting models. The methodology proposed is applied to post-1950 mortality data for a set of developed countries to test the usual assumption that is made in the literature of a sustained decline in mortality at roughly constant rates in this period. Structural changes in the rate of decline in overall mortality are found for almost every country, especially in male populations. We also illustrate how accounting for such a change can lead to a major effect in mortality and life expectancy forecasts over the next decades.

Multiple Testing Problems in Pharmaceutical Statistics

Abstract: Multiplicity Problems in Clinical Trials: A Regulatory Perspective, Mohammad Huque and Joachim Rohmel Introduction Common multiplicity problems in clinical trials Reducing multiplicity in clinical trials Multiplicity concerns in special situations Multiplicity in the analysis of safety endpoints Concluding remarks Multiple Testing Methodology, Alex Dmitrienko, Frank Bretz, Peter H. Westfall, James Troendle, Brian L. Wiens, Ajit C. Tamhane, and Jason C. Hsu Introduction Error rate definitions Multiple testing principles Adjusted significance levels, p-values, and confidence intervals Common multiple testing procedures Multiple testing procedures based on univariate p-values Parametric multiple testing procedures Resampling-based multiple testing procedures Software implementation Multiple Testing in Dose Response Problems, Frank Bretz, Ajit C. Tamhane, and Jose Pinheiro Introduction Dose response trend tests Target dose estimation using multiple hypothesis testing Power and sample size calculation for target dose estimation Hybrid approaches combining multiple testing and modeling Analysis of Multiple Endpoints in Clinical Trials, Ajit C. Tamhane and Alex Dmitrienko Introduction Inferential goals At-least-one procedures Global testing procedures All-or-none procedures Superiority-noninferiority procedures Software implementation Gatekeeping Procedures in Clinical Trials, Alex Dmitrienko and Ajit C. Tamhane Introduction Motivating examples Serial gatekeeping procedures Parallel gatekeeping procedures Tree gatekeeping procedures Software implementation Adaptive Designs and Confirmatory Hypothesis Testing, Willi Maurer, Michael Branson, and Martin Posch Introduction Basic principles and methods of error rate control Principles of adaptive testing procedures Adaptive multiple testing procedures Case studies Discussion Design and Analysis of Microarray Experiments for Pharmacogenomics, Jason C. Hsu, Youlan Rao, Yoonkyung Lee, Jane Chang, Kristin Bergsteinsdottir, Magnus Karl Magnusson, Tao Wang, and Eirikur Steingrimsson Potential uses of biomarkers Clinical uses of genetic profiling Two stages of pharmacogenomic development Multiplicity in pharmacogenomics Designing pharmacogenomic studies Analyzing microarray data by modeling A proof of concept experiment Software implementation Bibliography

On the distribution of runs scored and batting strategy in test cricket

Abstract: Negative binomial distributions are fitted to partnership scores and innings scores in test cricket. For partnership scores, we use a parametric model that allows us to consider run rate as a covariate in the distribution of runs scored and hence to use run rate as a surrogate for batting strategy. Then we describe the implied influence of run rate on match outcome probabilities given the state of the match at some point during the third innings; we refer to such a point in the match as the current position. Match outcome probabilities are calculated by using a model for the outcome given the end of the third-innings position, and a model for transitions from the current position to the end of the third-innings position, with transition probabilities considered as a function of run rate. Although the run rate is not wholly in the control of the batting side, our approach at least allows a captain or team analyst to consider the match outcome probability if the team can bat towards a target at a particular run rate. This will then at least indicate whether an aggressive or defensive batting strategy is desirable.

Parental unemployment and young people's extreme right-wing party affinity: evidence from panel data

Abstract: Summary The paper investigates the extent to which parental unemployment affects young people's far right-wing party affinity Cross-sectional estimates from the German Socio-Economic Panel show a positive relationship between growing up with unemployed parents and support for the extreme right The paper uses differences in parental unemployment experience during childhood across siblings to investigate a causal relationship Sibling differences estimates suggest that young people who experience parental unemployment have a significantly higher chance of supporting extreme right-wing parties in Germany The results show that the effect is particularly strong among East Germans, and stronger among sons than daughters Moreover, the estimates point to a strong and positive effect of growing up in a single-parent family on young people's far right-wing party affinity, whereas household income appears to be an insignificant predictor