Suchergebnisse

In a recent issue of this journal, Larocca (2005) makes two notable claims about the best linear unbiasedness of ordinary least squares (OLS) estimation of the linear regression model. The first, drawn from McElroy (1967), is that OLS remains best linear unbiased in the face of a particular kind of autocorrelation (constant for all pairs of observations). The second, much larger and more heterodox, is that the disturbance need not be assumed uncorrelated with the regressors for OLS to be best linear unbiased. The assumption is unnecessary, Larocca says, because "orthogonality [of disturbance and regressors] is a property of all OLS estimates" (p. 192). Of course OLS's being best linear unbiased still requires that the disturbance be homoskedastic and (McElroy's loophole aside) nonautocorrelated, but Larocca also adds that the same automatic orthogonality obtains for generalized least squares (GLS), which is also therefore best linear unbiased, when the disturbance is heteroskedastic or autocorrelated.

The nature of political sophistication is examined through development of a nonlinear simultaneous equation model that combines three general explanations of political behavior: exposure to political information & ability & motivation to assimilate & organize it. Model estimates are derived from a national probability sample from the 1976 American National Election Study. They suggest that interest & intelligence, representing motivation & ability, have major effects on political sophistication, but education & media exposure do not. Implications for the sophistication of mass publics & for democratic theory are discussed. 5 Tables, 1 Figure, 101 References. Adapted from the source document.

Explored are the estimation & meaning, in both recursive & (especially) nonrecursive causal models, of two kinds of parameters that describe the disturbances: correlations between the disturbances of different structural equations & residual path coefficients. Presented is a straightforward & statistically consistent method of calculating estimates of both types of parameters -- a method applicable to any identified model in which these parameters appear, whether nonrecursive or recursive. The interpretation of both kinds of parameters is discussed. In the case of correlations between disturbances, the caveat is that a high value can occur even when the factors responsible for the correlation are not truly worthy of note & a low value can occur even when they are. For this reason, estimated correlations between disturbances should always be juxtaposed with estimates of the strength of the disturbances involved. For residual path coefficients, the caveat is that in nonrecursive structures, unlike in recursive structures, their estimators do not represent the proportion of variance in a dependent variable that the model leaves unexplained. Appendixes. Modified HA.

Several recent studies have claimed to explain the post-1960 decline in U.S. presidential and congressional election turnout in terms of just a few variables. Abramson and Aldrich (1982) attribute the great bulk of the decline to aggregate declines in partisanship and political efficacy. Shaffer (1981) attributes still more of the decline to aggregate decreases in these same two variables plus age and following the campaign in the newspapers. And Kleppner (1982) attributes "virtually all" of the decline to aggregate decreases in the first three of these variables. We show how these studies measure explanation too generously and how the very short-handedness of their models makes them seem more successful than they are. These few variables, we conclude, leave most of the decline unexplained.