哈密瓜啊啊啊 发表于 2018-8-31 19:49
bs1是直接效应的结果,上下限不包括0则证明存在中介效应,同理bs2是直接效应的结果。
p表示置信区间
BC是 ...
您好,请教一下,为什么做出来的结果校正的置信区间上下限数值是相同的?存在什么问题?
Coef Bias Std Err [95% Conf Interval]
ind eff -0.018585 0.000161 0.007794 -0.03392 -0.005006 (P)
-0.038323 -0.038323 (BC)
dir eff 0.163847 0.000659 0.06719 0.03687 0.309134 (P)
-0.00582 -0.005824 (BC)
tot eff 0.145263 0.00082 0.066838 0.014513 0.290271 (P)
-0.02474 -0.024743 (BC)
(P)percentile confidence interval
(BC)bias-corrected confidence interval
还是要看下面的值?
command: sgmediation eff_w, mv(maer_w) iv(nonstate_w) cv(size_w lev_w growth_w tang_w ato_w indep_w dual
board_w concern1_w control)
_bs_1: r(ind_eff)
_bs_2: r(dir_eff)
------------------------------------------------------------------------------
| Observed Bootstrap
| Coef. Bias Std. Err. [95% Conf. Interval]
-------------+----------------------------------------------------------------
_bs_1 | -.01858472 -.0002732 .00932939 -.0378009 -.0003532 (P)
| -.0378945 -.000491 (BC)
_bs_2 | .16384746 .0027889 .07928715 .0011138 .320415 (P)
| -.0035124 .3104494 (BC)
------------------------------------------------------------------------------
(P) percentile confidence interval
(BC) bias-corrected confidence interval