Testing equality of variances in the analysis of repeated measurements

The problem of comparing the precisions of two instruments using repeated measurements can be cast as an extension of the Pitman-Morgan problem of testing equality of variances of a bivariate normal distribution. Hawkins (1981) decomposes the hypothesis of equal variances in this model into two subhypotheses for which simple tests exist. For the overall hypothesis he proposes to combine the tests of the subhypotheses using Fisher's method and empirically compares the component tests and their combination with the likelihood ratio test. In this paper an attempt is made to resolve some discrepancies and puzzling conclusions in Hawkins's study and to propose simple modifications.
The new tests are compared to the tests discussed by Hawkins and to each other both in terms of the finite sample power (estimated by Monte Carlo simulation) and theoretically in terms of asymptotic relative efficiencies.     

Ratings and confirmation

We present a formalism which makes explicit and precise the confirming effect of multiple observers and repeated trials on composite ratings, considering the relevant subjective inputs.     

Latent Demand for Urban Housing in the People's Republic of China

A new micro-data set demonstrates that, in urban China, own-built dwellings are 75% larger than state-owned dwellings. A switching regression with endogenous switching demonstrates that families with less privileged positions in the Chinese economy are more likely to inhabit own-built housing. The relationships between family characteristics and dwelling sizes in the own-built and state-owned sectors differ dramatically. However, unobserved characteristics, probably includingguanxi, play crucial roles in both. Depending on these characteristics, families in state-owned dwellings might have two-thirds more living space were they to choose own-built dwellings.     

Analysis of arrest rates for trends in criminality

During the 1960s, the United States experienced a substantial increase in crime with the rates of reported crime more than doubling. It has been postulated that the increase in the proportion of young adults in the population which also occurred during the Sixties is a primary factor in the increase in crime. In this paper, we attempt to explore this question by partitioning the recent increases in rates between an increasing crime-committing propensity for various demographic groups on one hand, and a changing demographic mix to the other. We do this by analyzing arrest rates by demographic group over time. However, since arrest rates are a complex function of both criminality and police activity, a model is developed which attempts to distinguish between these phenomena. The model is estimated using data on age, race, and sex-specific arrest rates for Pittsburgh residents over the period 1967–1972. The results of the analysis suggest that there has been no significant increase in criminality for men in the Pittsburgh population, while there is some evidence of increased criminality for women. However, this increase may reflect an increased willingness of the police to arrest women. We hesitate to generalize our conclusions too widely from the experience of a single urban center. Additional analysis, using the model developed here, of arrest rates from other cities or regions would do much to illuminate this very significant issue.     

Municipal welfare functions

In this paper the concept of a municipal welfare function is defined. It reflects the evaluation by local authorities of several levels of local expenditures. On the basis of an extensive survey among all Dutch municipal authorities these functions are estimated for about 550 Dutch municipalities with respect to total expenditures and differentiated with respect to several portfolios, like public works, education, etc. The variation of the estimated municipal welfare parameters is explained by objectively measurable municipal characteristics like the number of the inhabitants, age distribution of inhabitants and houses, number of unemployed, regional situation.     

Export Composition and Manufacturing Employment in the US during the Economic Downturn of 1991-92

This paper examines the effect that export composition had upon manufacturing employment in the US during the 1991 recession. Although it takes, on average, approximately $66 000 in exports to create one job, the exact gains in terms of total employment depend upon the labour-intensity of the products being exported. Foreign sales by the chemical and textile industries result in a far greater increase in employment than exports by the petroleum refining or steel industries. This analysis estimates the employment effects of manufacturing exports over the 1989-95 period, utilizing an input-output model to capture both direct and indirect effects. The results demonstrate that export composition has, at times, both strengthened and reduced demand for labour. Consequently, if job-creation is a national goal, it may be in the interests of the US to promote exports from sectors that are labour-using.