Robert Crouchley

Stochastic frontier estimation of a CES cost function: the case of higher education in Britain

Abstract Many estimates exist in the literature of multi-product cost functions estimated across a sample of institutions of higher education. Typically these have not employed the appropriate frontier estimation techniques. The cost functions usually estimated — such as the constant elasticity of substitution (CES) function — are highly non-linear, and the standard software packages do not allow stochastic frontier estimation in such cases. We derive and maximise the likelihood function associated with this problem, and hence construct measures of economies of scale and scope which obtain in British higher education. We use the half-normal residuals generated by the stochastic frontier estimation to construct measures of technical efficiency for each university in our sample.

Testing for Quasi-Market Forces in Secondary Education

This paper investigates the effect of introducing quasi-market forces into secondary education on the allocation of pupils between schools and on the exam performance of pupils. A unique database is used which covers all publicly-funded secondary schools in England over the period 1992-98. We find several effects consistent with the operation of a quasi-market. Firstly, new admissions are found to be positively related to a schools own exam performance and negatively related to the exam performance of competing schools. Secondly, a schools growth in pupil numbers is positively related to its exam performance compared to its immediate competitors. Thirdly, there is strong evidence that schools experiencing an excess demand for places have responded by increasing their physical capacity. Fourthly, there is some evidence of an increase in the concentration of pupils from poor family backgrounds in those schools with the poorest exam performance of schools during 1992-98 can be attributed to the introduction of quasi-market forces.

A Random-Effects Model for Ordered Categorical Data

Abstract This article presents a random-effects model for multivariate and grouped univariate ordered categorical data. The assumed family of distributions for the random effects adopts a wide variety of forms and shapes. The models likelihood has a closed expression and can be computed without recourse to numerical integration or Gaussian quadrature. A bivariate and a grouped univariate example are used to illustrate the proposed model.

Social exclusion and labour market transitions: a multi-state multi-spell analysis using the BHPS

In this empirical paper, we assess how social exclusion arises in the context of labour market transition behaviour. We estimate a multi-state multi-spell competing risks model and identify five states: high skilled employment, intermediate skilled employment, low skilled employment, unemployment and out-of-the-labour market. Using data from the first seven waves of the British Household Panel Survey, we show that a substantial number of workers were trapped in a vicious circle of low-skilled employment, unemployment and inactivity in the 1990s. Workers who are part of the so-called flexible workforce are more likely to suffer social exclusion.

Integration of Existing Grid Tools in Sakai VRE

The integration of existing grid tools into the Sakai VRE (virtual research environment) is discussed in this paper. In particular we describe the integration of the business logic and JSR 168 compliant portlets through presentation-oriented Web services, via WSRP (Web services for remote portlets). A set of JSR 168 compliant portlets were developed for the UK NGS (National Grid Service) portal and have been published using WSRP4J and consumed within Sakai successfully, which proves re-use of portlets as Web components is a practical option. With the help of the WSRP consumer tool, the Sakai VRE has been successfully extended to support the JSR 168 specification

Using the Sakai collaborative toolkit in e-Research applications

The Sakai Project (http://www.sakaiproject.org) is developing a collaborative environment that provides capabilities that span teaching and learning as well as e‐Research applications. By exploiting the significant requirements overlap in the collaboration space between these areas, the Sakai community can harness significant resources to develop an increasingly rich set of collaborative tools. While collaboration is a significant element of many e‐Research projects, there are many other important elements including portals, data repositories, compute resources, special software, data sources, desktop applications, and content management/e‐Publication. The successful e‐Research projects will find ways to harness all of these elements to advance their science in the most effective manner. It is critical to realize that there is not a single software product that can meet the requirements for such a rich e‐Research effort. Realizing that multiple elements must be integrated together for best effect leads us to focus on understanding the nature of integration and working together to improve the cross‐application integration. This leads us not to drive towards a single toolkit (such as Sakai or Globus), but instead to a meta‐toolkit containing well‐integrated applications. When considering a technology for use, perhaps the most important aspect of that technology is how well it integrates with other technologies. Copyright

A comparison of GEE and random effects models for distinguishing heterogeneity, nonstationarity and state dependence in a collection of short binary event series

GEE transition models and Markov random effect models are applied to a simple panel data set on depression. In each case, the precise specifications adopted were derived from the authors’ interpretation of best practice in the literature. The two approaches result in quite different inference on the three process characteristics of interest: state dependence, heterogeneity and nonstationarity. The design of the analyses permits indirect goodness of fit measures to be derived for the GEE models and these indicate serious deficiencies in this approach. It is shown through simulation and further analyses of the depression data that these deficiencies may be corrected by including the initial observation properly in the analyses and by adopting an appropriate variance-covariance structure. The former problem is widely understood in random effects modelling and is relatively straightforward to address within GEE. The latter problem is more difficult because, without model selection or goodness of fit measures generally available for GEE models, it is not clear how one may select empirically between alternative variance-covariance structures. Inappropriate variance-covariance specifications prejudice consistent estimation of state dependence and nonstationarity.

The common structure of several models for non-ignorable dropout

This paper presents a multivariate generalization of the classical Heckman selection model and applies it to non-ignorable dropout in repeated continuous responses. Many of the recent models for dropout in repeated continuous responses can be written as special forms of this generalized Heckman model. To illustrate this, we present the parameterizations needed to obtain the form of dropout model that occurs when (1) the separate models for the response and dropout are linked by common random parameters, (2) the dropout model is an explicit function of the previous responses and the possibly unobserved current response, (3) the dropout model is both a function of the current response and a common random parameter, and (4) there is a covariance between the stochastic disturbances of the response and dropout processes. We present the joint likelihood of the generalized Heckman model and a residual for the responses. We contrast two of the dropout models in a simulation study. We compare the results obtained from several dropout models on the well known mastitis data.

Dynamic Models of Shopping Behaviour: Testing the Linear Learning Model and Some Alternatives

AbstractStochastic process models are identified as a suitable framework within which to examine dynamic consumer choice behaviour and to test the assumptions which underlie other simple behavioural models. A distinction is drawn between choice series and renewal series and the concepts of heterogeneity, non-stationarity and feedback are introduced. Some empirical problems in the testing of stochastic process models are discussed and illustrated by reference to a study by Burnett (1977) who applied a Linear Learning Model to data of store choice behaviour from Uppsala, Sweden. Careful calibration procedures are shown to be essential for useful model testing. Failure to consider more parsimonious models led to a failure to identify the superior performance of the simple Heterogeneous Bernoulli Model. Tests of non-stationarity and heterogeneity are presented. Finally, certain reservations are expressed about the sampling procedure adopted and of the interaction between consumer ‘brand’ choice and ‘store’ ch...

A specification test for univariate and multivariate proportional hazards models.

This paper applies Whites (1982, Econometrica 50, 1-25) information matrix (IM) test for correct model specification to proportional hazards models of univariate and multivariate censored survival data. Several alternative estimators of the test statistic are presented and their size performance examined. White also suggested an estimator of the parameter covariance matrix that was robust to certain forms of model misspecification. This has been subsequently proposed by others (e.g., Royall, 1986, International Statistical Review 54, 221-226) and applied by Huster, Brookmeyer, and Self (1989, Biometrics 45, 145-156) as part of an independence working model (IWM) approach to multivariate censored survival data. We illustrate how the IM test can be used for both univariate data and as part of the IWM approach to multivariate data.