Probit regression, reporting marginal effects Log pseudolikelihood-99.484992 ea0a28:989 Pseudo R Play dF/dx P>Izl x-bar【95c.1.1 Relitpe-o -.502169 392124 36527 :85903245069 .06653 .0102349 046470865q 30 8195681（atx-bar) ,分享的比例和进入的可能性成反比例，这主要是因为那些一开 怡分，较多的不情愿的分享者不进行游戏所导致的。分享比小进行游戏可以多得 享比例大 进行游戏没有不进行游戏的收益大) dprobit Play InitPercentShared Endow if Decision 2 ReluctantSharer =1,robust -93.836 Iteration1: Iteration 4: log paeudolikelihood -72.917475 Probit regression reporting marginal effect Log pseudolikel 。-72.91747 Prob>ch20.0000 ■0.222 dE/dx x-bar【958c.1.】 d. 在不情愿的分享者中，分享比例和进入可能性两者之间的负相关关系非常明显14 pred. P .8195681 (at x-bar) obs. P .7307692 Endow .06653 .0102349 6.21 0.000 13.5256 .04647 .08659 Reluct~r* -.349695 .0522464 -5.98 0.000 .602564 -.452096 -.247294 InitPe~d -.5021222 .182174 -2.77 0.006 .365278 -.859177 -.145068 Play dF/dx Std. Err. z P>|z| x-bar [ 95% C.I. ] Robust Log pseudolikelihood = -99.484992 Pseudo R2 = 0.2701 Prob > chi2 = 0.0000 Wald chi2(3) = 62.38 Probit regression, reporting marginal effects Number of obs = 234 对不情愿的分享者而言，分享的比例和进入的可能性成反比例，这主要是因为那些一开 始分享较多的不情愿的分享者不进行游戏所导致的。（分享比例小，进行游戏可以多得，分 享 比 例 大 ， 进 行 游 戏 没 有 不 进 行 游 戏 的 收 益 大 ） pred. P .6772361 (at x-bar) obs. P .6170213 Endow .0864788 .0144297 5.75 0.000 13.7766 .058197 .114761 InitPe~d -.8231926 .2648395 -3.13 0.002 .30922 -1.34227 -.304117 Play dF/dx Std. Err. z P>|z| x-bar [ 95% C.I. ] Robust Log pseudolikelihood = -72.917475 Pseudo R2 = 0.2229 Prob > chi2 = 0.0000 Wald chi2(2) = 36.92 Probit regression, reporting marginal effects Number of obs = 141 Iteration 4: log pseudolikelihood = -72.917475 Iteration 3: log pseudolikelihood = -72.917536 Iteration 2: log pseudolikelihood = -72.957184 Iteration 1: log pseudolikelihood = -74.277965 Iteration 0: log pseudolikelihood = -93.836 . dprobit Play InitPercentShared Endow if Decision > 2 & ReluctantSharer == 1 , robust 在不情愿的分享者中，分享比例和进入可能性两者之间的负相关关系非常明显