Book Review

DNA Microarrays and Gene Expression: From Experiments to Data Analysis and Modeling, by Pierre Baldi and G. Wesley Hatfield, 2002, Cambridge University Press, Cambridge, United Kingdom, 213?+?xvi pp., [ISBN: 0-521-80022-6]     

Reproducibility of clinical data I: continuous outcomes

Across a variety of clinical settings, repeated measurements on an individual, obtained under identical circumstances, often differ from one another. This implies the measurements lack perfect reproducibility. Topics related to reproducibility of clinical measurements are introduced in this paper. In this first of two parts, continuous outcomes are addressed. The intraclass correlation coefficient, ρ, has been the traditional coefficient of reproducibility for continuous outcomes. The importance of ρ regarding observations on an individual, and observations among populations, is outlined. Estimation and inferential procedures for ρ are reviewed and worked examples are provided. Copyright © 2004 John Wiley & Sons, Ltd.     

Measurement of bivariate attributes using a novel statistical model

Reducing process variability is essential to many organisations. According to the pertinent literature, a quality system that utilizes quality techniques to reduce process variability is necessary. Quality programs that respond to measurement precision are central to quality systems, and the most common method of assessing the precision of a measurement system is repeatability and reproducibility (R&R). Few studies have investigated R&R using attribute data.

In modern manufacturing environments, automated manufacturing is becoming increasingly common; however, a measurement resolution problem exists in automatic inspection equipment, resulting in clusters and product defects. It is vital to monitor effectively these bivariate quality characteristics. This study presents a novel model for calculating R&R for bivariate attribute data. An alloy manufacturing case is utilized to illustrate the process and potential of the proposed model. Findings can be employed to evaluate and improve measurement systems with bivariate attribute data.     

The poor availability of syntaxes of structural equation modeling

The syntax or codes used to fit Structural Equation Models (SEMs) convey valuable information on model specifications and the manner in which SEMs are estimated. We requested SEM syntaxes from a random sample of 229 articles (published in 1998–2013) that ran SEMs using LISREL, AMOS, or Mplus. After exchanging over 500 emails, we ended up obtaining a meagre 57 syntaxes used in these articles (24.9% of syntaxes we requested). Results considering the 129 (corresponding) authors who replied to our request showed that the odds of the syntax being lost increased by 21% per year passed since publication of the article, while the odds of actually obtaining a syntax dropped by 13% per year. So SEM syntaxes that are crucial for reproducibility and for correcting errors in the running and reporting of SEMs are often unavailable and get lost rapidly. The preferred solution is mandatory sharing of SEM syntaxes alongside articles or in data repositories.     

Demographic Changes in East Malaysia and their Relationship with those in the Peninsula 1960–80

Hardly anything is known about the demography of East Malaysia, that is of the lesser developed regions of Sabah and Sarawak, because of the wide range of deficiencies in the available data. This paper uses a range of indirect methods to construct a demographic profile of Sabah and Sarawak over the period 1960 to 1980. Comparisons are made between these areas and Peninsular Malaysia and some overall Malaysian estimates are given. The study suggests that between 1965 and 1980 there were large gains in life expectancy among the indigenous communities in Sabah and Sarawak, although their life expectancies are still shorter than those of the Chinese. Differentials in life chances between these regions and Peninsular Malaysia narrowed appreciably during the 1970s. Over the same period there were falls in levels of fertility. These were most spectacular among the Chinese communities and the groups living in and around the urban areas in Sarawak. There has been some decline in fertility among the indigenous groups in Sabah but current levels remain higher than among the indigenous groups in Sarawak and the Malays in the Peninsula.     

Reproducibility of clinical data II: categorical outcomes

In this second part of this paper, reproducibility of discrete ordinal and nominal outcomes is addressed. The first part deals with continuous outcomes, concentrating on intraclass correlation (ρ) in the context of one‐way analysis of variance. For categorical data, the focus has generally not been on a meaningful population parameter such as ρ. However, intraclass correlation has been defined for discrete ordinal data, ρ_c, and for nominal data, κ_I. Therefore, a unified approach to reproducibility is proposed. The relevance of these parameters is outlined. Estimation and inferential procedures for ρ_c and κ_I are reviewed, together with worked examples. Topics related to reproducibility that are not addressed in either this or the previous paper are highlighted. Considerations for designing reproducibility studies and for interpreting their results are provided. Copyright © 2004 John Wiley & Sons, Ltd.     

Managing the science in the presence of financial conflict of interest

ABSTRACT

We consider the question of why (primarily) and how (secondarily) to perform scientific oversight of research performed by investigators with a financial conflict of interest (COI). One way to frame the trade-off of having investigators with financial COI participate in research is through a decision rule: “Our institution is willing for financially conflicted investigators to participate in research around their intellectual property if (a) the science is likely to be significantly better with their participation (or if other significant benefits accrue); and (b) the COI can be adequately managed”.

A key component of COI management is the demonstration that the underlying science is sufficiently rigorous and transparent, and in turn, a scientific oversight plan is a key element of that demonstration. Scientific oversight plans should be proactively generated, by individuals (i.e., scientific reviewers) who are independent and expert, and they should assess the rigor and transparency of the research, in a fashion which is fair and efficient.

Abbreviations: COI: conflict of interest; SOP: scientific oversight plan     

Confidence Intervals for Variance Components in Measurement System Capability Studies

In Measurement System Analysis a relevant issue is how to find confidence intervals for the parameters used to evaluate the capability of a gauge. In literature approximate solutions are available but they produce so wide intervals that they are often not effective in the decision process. In this article we introduce a new approach and, with particular reference to the parameter γ_R, i.e., the ratio of the variance due to the process and the variance due to the instrument, we show that, under quite realistic assumptions, we obtain confidence intervals narrower than other methods. An application to a real microelectronic case study is reported.     

Effect of impact factor and discipline on journal data sharing policies

Data sharing is crucial to the advancement of science because it facilitates collaboration, transparency, reproducibility, criticism, and re-analysis. Publishers are well-positioned to promote sharing of research data by implementing data sharing policies. While there is an increasing trend toward requiring data sharing, not all journals mandate that data be shared at the time of publication. In this study, we extended previous work to analyze the data sharing policies of 447 journals across several scientific disciplines, including biology, clinical sciences, mathematics, physics, and social sciences. Our results showed that only a small percentage of journals require data sharing as a condition of publication, and that this varies across disciplines and Impact Factors. Both Impact Factors and discipline are associated with the presence of a data sharing policy. Our results suggest that journals with higher Impact Factors are more likely to have data sharing policies; use shared data in peer review; require deposit of specific data types into publicly available data banks; and refer to reproducibility as a rationale for sharing data. Biological science journals are more likely than social science and mathematics journals to require data sharing.