=>characteristic of population
用sample 的特性去推population
@CI=confidence Interval 信賴區間
99%CI(α=0.05)={mean-2.57SD,mean+2.57SD}99%左右的樣本的平均值會落入U±2.58SE
95%CI(α=0.05)={mean-1.96SD,mean+1.96SD}95%左右的樣本的平均值會落入U±1.96SE
90%CI(α=0.05)={mean-1.645SD,mean+1.645SD} 90%左右的樣本的平均值會落入U±1.64 SE

當數值超過CI範圍==>表示不屬於此sample
Hypothesis testing: intervention 之後,看進步是因為by chance or sampling error
費雪爾(R.A.Fisher),把假設分為虛無假設(null hypothesis)與對立假設(alternative hypothesis)
兩種,且分別以符號Ho及H1表示之。
費雪爾建議:將實驗者心目中盼望得到的研究結果當作是對立假設H1 ;而將與對立假設完全相反的結果當作是虛無假設Ho。在兩種假設當中,只有虛無假設是直接受到統計檢定。費雪爾希望藉由統計測驗推翻虛無假設,從而間接的為對立假設的可信性提供支持。
1-4steps:
1.state null hypothesis: intervention 無效
Ho=Ua=Ub=Ua-Ub=0
#拒絕 虛無假設
#接受 虛無假設(介入無效)
evidence is too weak to support the effect (intervention 不一定無效)
2.state the alternative hypothesis:intervention 有效
H1=Ua≠Ub or Ua-Ub≠0 (nondirectional hypothesis)
H1=Ua>Ub(directional hypothesis)較常用
3.select level of significance (α)
error:decision to reject
Type 1 error(α)α=0.05(p<0.05)
type 2 error(β)β=0.2(power=0.8)
p=0.05 有5%的機率效果是by chance,有5%是Type 1error
p values: p<0.001 or p=0.001 在統計的意義=>都小於0.05,reject Ho
可以看出有進步或改變的差異
要看效果則要看effect size(ES)
null hypothesis(虛無假設)
「虛無假設」是母群體的變數。假設驗證是要根據實驗資料測試虛無假設的可行性,它可能被推翻或不被推翻。The null hypothesis is an hypothesis about a population parameter. The purpose of hypothesis testing is to test the viability of the null hypothesis in the light of experimental data. Depending on the data, the null hypothesis either will or will not be rejected as a viable possibility.
研究者想研究酒精和人對音調反應的影響。當µ1 表示喝酒者的平均反應時間,而µ2表未喝酒者的平均反應時間,則此虛無假設為µ1 - µ2 = 0。
Consider a researcher interested in whether the time to respond to a tone is affected by the consumption of alcohol. The null hypothesis is that µ1 - µ2 = 0 where µ1 is the mean time to respond after consuming alcohol and µ2 is the mean time to respond otherwise.
虛無假設通常和實驗者的預測結果相反,為的是要讓實驗資料將之推翻。此實驗中,實驗者認為酒精對反應力有不良的影響,若實驗資料亦顯示如此,則可以將虛無假設推翻。
The null hypothesis is often the reverse of what the experimenter actually believes; it is put forward to allow the data to contradict it. In the experiment on the effect of alcohol, the experimenter probably expects alcohol to have a harmful effect. If the experimental data show a sufficiently large effect of alcohol, then the null hypothesis that alcohol has no effect can be rejected.
http://www.ruf.rice.edu/~lane/hyperstat/A29337.html
null hypothesis:先假設和原本假設相反的假設,也就是「虛無假設」,再推翻它,進而證明原假設成立。(這麼做的原因是因為要直接證明一個假設成立很難,因為它可能在各種狀況下都成立,而只要所有情況中有任一個可以證明假設不成立,即可推翻此假設,但我們又不能保證已經找出所有的情況故最好的方法就是先假設和預期相反的虛無假設,再進一步推翻它,此方法較容易,且可間接證明原先預期成立。
沒有留言:
張貼留言