How Fragile are Fragile Inferences? A Re-Evaluation of the Deterrent Effect of Capital Punishment
用极端边界分析和自助法重新评估死刑的威慑效应，发现统计推断的脆弱性需谨慎解读。
Extreme bounds analysis attempts to measure the effects of the uncertainty in the specification of the explanatory variables in a regression model on the estimated coefficients of interest. Standard errors for the stochastic extreme bounds are computed using the bootstrap technique. State-by-state cross section data are used to study the deterrent effect of capital punishment in the United States in 1950. The bootstrap standard errors are sufficiently large for some bounds to suggest caution in the interpretation of the empirical results regarding the fragility of inferences for the deterrent effect of capital punishment. Copyright 1989 by MIT Press.