A 2023 paper about these values by economist Alberto Abadie et al. proposes “model-based” and “design-based” frameworks for choosing which type to use. In spatial contexts, the Conley adjustment can be applied to these values, which Morgan Kelly found to be unrealistically small in 27 past studies of historical persistence. James MacKinnon and Halbert White devised four types of these values called HC0 (“H-C-zero”) through HC3. In difference-in-differences studies, serial correlation requires estimating “clustered” versions of these values. When variances of residuals are unlikely to be constant, researchers usually use “robust” versions of these values. This value is the denominator of a one-sample t-statistic. For 10 points, name this value often defined as standard deviation over square root of sample size, a measure of how accurate an estimate is. ■END■
ANSWER: standard errors [or clustered standard errors; or heteroskedasticity-consistent standard errors; or robust standard errors; accept standard errors of the mean; prompt on errors; prompt on s over square root of n; prompt on sigma over square root of n; prompt on SE; reject “standard deviation”]
<Social Science>
= Average correct buzz position