Suchergebnisse

Regressions can be weighted by propensity scores in order to reduce bias. However, weighting is likely to increase random error in the estimates, and to bias the estimated standard errors downward, even when selection mechanisms are well understood. Moreover, in some cases, weighting will increase the bias in estimated causal parameters. If investigators have a good causal model, it seems better just to fit the model without weights. If the causal model is improperly specified, there can be significant problems in retrieving the situation by weighting, although weighting may help under some circumstances.

Schon im Frühstücksfernsehen reden sie auf uns ein, die Experten. Wir sehen, hören, lesen von Ernährungs-, Sport-, Politik- oder Rechtsexperten. Freedman blickt ihnen auf die Finger. Er gibt uns gar einen Leitfaden an die Hand, wie wir Fachleute einschätzen können und ihnen nicht so schnell auf den Leim gehen. Faktenreich, unterhaltsam und selbstironisch beschreibt der Wissenschaftsjournalist den Markt der Eitelkeiten. Er nennt Beispiele für Betrugsfälle, Forschungsirrtümer und fragwürdige Studien. Was soll der Normalbürger machen, wenn sich Ratschläge widersprechen? Ein paar stellt er exemplarisch gegenüber. Wie kommt es zur Expertenkonfusion, warum kommt es zur Rudelbildung unter den Fachleuten? Der Autor konstatiert, dass es keine Alternative zum Wissenschaftsbetrieb und Expertentum gibt. Noch nie hat es so viele Wissenschaftler auf der Welt gegeben wie heutzutage. Da erscheint dieses Buch gerade recht! (2)

We look at conventional methods for removing endogeneity bias in regression models, including the linear model and the probit model. It is known that the usual Heckman two-step procedure should not be used in the probit model: from a theoretical perspective, it is unsatisfactory, and likelihood methods are superior. However, serious numerical problems occur when standard software packages try to maximize the biprobit likelihood function, even if the number of covariates is small. We draw conclusions for statistical practice. Finally, we prove the conditions under which parameters in the model are identifiable. The conditions for identification are delicate; we believe these results are new.

The salt hypothesis is that higher levels of salt in the diet lead to higher levels of blood pressure, increasing the risk of cardiovascular disease. Intersalt, a cross-sectional study of salt levels and blood pressures in 52 populations, is often cited to support the salt hypothesis, but the data are somewhat contradictory. Four of the populations (Kenya, Papua, and 2 Indian tribes in Brazil) do have low levels of salt and blood pressure. Across the other 48 populations, however, blood pressures go down as salt levels go up, contradicting the hypothesis. Experimental evidence suggests that the effect of a large reduction in salt intake on blood pressure is modest, and health consequences remain to be determined. Funding agencies and medical journals have taken a stronger position favoring the salt hypothesis than is warranted, raising questions about the interaction between the policy process and science.

One form of error that can affect census adjustments is correlation bias, reflecting people who are doubly missing—from the census and from the adjusted counts as well. This article presents a method for estimating the total national number of doubly-missing people and their distribution by race and sex. Application to the 1990 U.S. census adjustment leads to an estimate of 3 million doubly-missing people. Correlation bias is likely to be a serious problem for census adjustment in 2000. The methods of this article are well suited for measuring its magnitude.

Considering the difficulties, the Census Bureau does a remarkably good job at counting people. There are two current techniques for evaluating or adjusting the census: (a) demographic analysis uses administrative records to make independent population estimates, which can be compared to census counts; (b) datafrom an independent sample survey can be used to estimate population coverage. If there is a large undercount, these techniques may be accurate enough for adjustment. With a small undercount, it is unlikely that current adjustment methodologies can improve on the census; instead, adjustment could easily degrade the accuracy of the data. This article reviews plans for Census 2000, including proposals for adjustment, in the light of past experience.