作者evilove (願意感動就是一種享受)
看板Statistics
標題[問題] 請問怎麼解釋這個交互作用的結果
時間Sun May 5 10:13:31 2019
請問怎麼解釋交互作用變項不顯著 但是他們的marginal effects顯著? 謝謝!
Random-effects GLS regression Number of obs = 1,360
Group variable: id1 Number of groups = 172
R-sq: Obs per group:
within = 0.1152 min = 2
between = 0.1123 avg = 7.9
overall = 0.1308 max = 10
Wald chi2(12) = 76.82
corr(u_i, X) = 0 (assumed) Prob > chi2 = 0.0000
(Std. Err. adjusted for 172 clusters in
id1)
-------------------------------------------------------------------------------
| Robust
asp_pct | Coef. Std. Err. z P>|z| [95% Conf.
Interval]
--------------+----------------------------------------------------------------
mp | -1.332043 .5194244 -2.56 0.010 -2.350096
-.3139898
mp_sqr | .1547854 .043082 3.59 0.000 .0703462
.2392245
coh2 | 14.89091 5.379513 2.77 0.006 4.347255
25.43456
coh3 | 22.1159 5.544094 3.99 0.000 11.24967
32.98212
fqhc | 12.39198 4.62368 2.68 0.007 3.329734
21.45423
system | 20.40573 4.582832 4.45 0.000 11.42355
29.38792
interv | 4.570367 1.425087 3.21 0.001 1.777247
7.363487
maint | 1.091151 .7326216 1.49 0.136 -.3447612
2.527063
interv_fqhc | -5.622486 3.999668 -1.41 0.160 -13.46169
2.216719
maint_fqhc | -1.76751 1.225599 -1.44 0.149 -4.16964
.6346198
interv_system | -5.573831 1.594597 -3.50 0.000 -8.699185
-2.448477
maint_system | .9807776 .8697199 1.13 0.259 -.723842
2.685397
_cons | 37.42812 5.35765 6.99 0.000 26.92732
47.92892
--------------+----------------------------------------------------------------
sigma_u | 24.31059
sigma_e | 5.9862898
rho | .94283128 (fraction of variance due to u_i)
-------------------------------------------------------------------------------
. margins, at(maint = (0 1) fqhc = (0 1) system = (0 1))
Predictive margins Number of obs = 1,360
Model VCE : Robust
Expression : Linear prediction, predict()
1._at : fqhc = 0
system = 0
maint = 0
2._at : fqhc = 0
system = 0
maint = 1
3._at : fqhc = 0
system = 1
maint = 0
4._at : fqhc = 0
system = 1
maint = 1
5._at : fqhc = 1
system = 0
maint = 0
6._at : fqhc = 1
system = 0
maint = 1
7._at : fqhc = 1
system = 1
maint = 0
8._at : fqhc = 1
system = 1
maint = 1
------------------------------------------------------------------------------
| Delta-method
| Margin Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
_at |
1 | 49.37554 3.573883 13.82 0.000 42.37085 56.38022
2 | 50.46669 3.521219 14.33 0.000 43.56522 57.36815
3 | 69.78127 2.973859 23.46 0.000 63.95261 75.60993
4 | 70.87242 3.003034 23.60 0.000 64.98658 76.75826
5 | 61.76751 3.08916 19.99 0.000 55.71287 67.82216
6 | 62.85867 3.068848 20.48 0.000 56.84383 68.8735
7 | 82.17325 5.309122 15.48 0.000 71.76756 92.57894
8 | 83.2644 5.348805 15.57 0.000 72.78093 93.74786
------------------------------------------------------------------------------
fqhc, system and one more variable are coded as dummy variables.
--
※ 發信站: 批踢踢實業坊(ptt.cc), 來自: 67.255.36.188
※ 文章網址: https://www.ptt.cc/bbs/Statistics/M.1557022413.A.FB5.html
→ KirinGuess: 呃...花一點時間把報表整理好會不會比較有發問的誠意? 05/05 12:50
→ andrew43: 伸手要答案請直接花錢請顧問 05/05 18:47
→ evilove: stata裡的output看起來是ok的 但是貼過來就跑掉了 05/05 22:35
→ evilove: 至於花錢請顧問? 這裡不就是交流的地方嗎 需要這樣酸嗎 05/05 22:36
推 andrew43: 交流？我只看到你在要答案吧 05/06 00:00
推 andrew43: 何不看看別人在本版問問題有充分回答的是怎麼個交流法 05/06 00:06
→ DrSingularit: 你應該知道有修改文章的功能吧?不修一下,排板一下 05/06 10:48
→ DrSingularit: 嗎? 05/06 10:48
→ if266: 伸手文拿來的交流，多點圖片報表或文字敘述更好吧 05/21 19:32