Selection of working correlation structure and best model in GEE analyses of longitudinal data

The Generalized Estimating Equations (GEE) method is one of the most commonly used statistical methods for the analysis of longitudinal data in epidemiological studies. A working correlation structure for the repeated measures of the outcome variable of a subject needs to be specified by this method. However, statistical criteria for selecting the best correlation structure and the best subset of explanatory variables in GEE are only available recently because the GEE method is developed on the basis of quasi-likelihood theory. Maximum likelihood based model selection methods, such as the widely used Akaike Information Criterion (AIC), are not applicable to GEE directly. Pan (2001) proposed a selection method called QIC which can be used to select the best correlation structure and the best subset of explanatory variables. Based on the QIC method, we developed a computing program to calculate the QIC value for a range of different distributions, link functions and correlation structures. This program was written in Stata software. In this article, we introduce this program and demonstrate how to use it to select the most parsimonious model in GEE analyses of longitudinal data through several representative examples.

Investigative interviewing of children : barriers to adopting 'best-practice' guidelines

This research examined the global difficulties of investigative interviewers in adhering to best-practice guidelines (i.e. open-ended questions) when interviewing children about abuse. It demonstrated that the importance of, and rationale for, using open-ended questions is not well understood by investigative interviewers, nor are they adequately reinforced with police organisations.

An innovation framework based on best practice exemplars from the Australian School Innovation in Science, Technology and Mathematics (ASISTM) project

This report describes research into 16 ASISTM projects selected to be broadly representative of exemplars in innovation. Case studies of each project were constructed from interviews with a range of key participants, and used to develop and refine an innovation framework that is used to make sense of and describe the key features of each project. The major issue binding these projects was found to be that of student interest and engagement, and this was pursued through involving students in contemporary science, technology and mathematics (STM) practices in authentic settings. The findings point to an enriched set of purposes of STM education implicit in these projects, a set of pedagogical practices that are varied and consistent with contemporary educational thinking, and a varied array of 'actors' recruited to these projects.

An empirical and logical exploration of the strategic compatibility of best practice and product innovation : preliminary insights from Australian and New Zealand manufacturing industry

In an attempt to enhance debate focused on an established academic controversy, this study re-investigated selected data from the 1994 AMC survey of Australian and New Zealand manufacturing practices to test the hypothesis that best practice and product innovation may be incompatible generic business strategies. A modification of Robert G. Cooper’s Stage-Gate product development model was used as a theoretical framework to create a measurable construct of ‘product innovation’ as a strategy and compare two groups: firms committed to a best practice strategy (BPs) and firms not utilising best practice (Non-BPs). Eight variables were scrutinised. After logical critique was added to statistical data analysis, four major insights emerged.

(1) Tests yielded several statistically significant but substantively inconclusive results because both studied groups had nearly identical profiles in rating innovation as the factor of lowest importance to commercial success and because the definitional framework which guided construction of the survey instrument treated innovation as a second-order issue. (2) Currently, best practice and product innovation are logically incompatible by definition. (3) Even if the definition of best practice were changed, it is likely that the additional key process of innovation would remain incompatible with the existing key process of benchmarking. (4) However, until the definition of best practice does make an attempt to include innovation as a key process rather than an outcome, testing any hypothesis of strategic compatibility between a best practice focus and an innovation focus will be both empirically difficult and logically unnecessary.

Best practice angels: early-stage funding at Pasadena Angels

This 8-minute DVD (PAL standard in DVD-R format) looks into the workings of one of the oldest angel investing organizations. Pasadena Angels is a group of leading private investors that provide long-term human and financial capital to help build successful companies. Two start-up companies are profiled as they apply for funds. Directors and angel investors describe their funding criteria and discuss how a “best practice” angel operates. This video is appropriate for new angel groups, upper-level university courses, and general venture capital interest.