Timeline for 95% CI vs. +/- 2 standard error bands for impulse response function
Current License: CC BY-SA 3.0
5 events
when toggle format | what | by | license | comment | |
---|---|---|---|---|---|
Apr 23, 2017 at 13:25 | vote | accept | jeffrey | ||
Apr 19, 2017 at 18:13 | answer | added | shadowtalker | timeline score: 2 | |
Apr 19, 2017 at 17:54 | comment | added | Michael R. Chernick | It is a property of the Normal distribution that the interval from -1.96 SE + mean to 1.96 SE + mean covers 95% of the distribution. People often round this off to 2. The confidence intervals for the mean are constructed using the sample estimates in place of the population parameters. But this does not apply in general to other distributions. | |
Apr 19, 2017 at 17:54 | comment | added | Carol Eisen | Others may contradict me, but I prefer CIs over SE, because 2*SE are close to CIs but not quite the same (e.g. z*=1.96). The CI is more intuitive to interpret for most people. | |
Apr 19, 2017 at 17:41 | history | asked | jeffrey | CC BY-SA 3.0 |