✔ 最佳答案
Firstly, z-test statistics is used when
1) n(sample size) is equal or bigger than 30, or
2) when the population is "Normally Distributed" and
3) the 'population standard deviation' is known.
However, the t-test statistics is used when
1) n<30
2) the population is normally or approximately normally distributed
3) population standard deviation is unknown. (we use sample standard deviation instead.)
讀統計既人都會知道, 搵t-value既時候都要用到 df (degree of freedom),即係用 n-1 係個statistics table黎溫,而唔係就咁用n, 原因係個samlpe size太細會影響個準確性 (呢d要繼續讀統計, 讀上去先至會prove到)
而溫z-value既時候, 由於個sample size夠大, 所以我地會就咁對番個 level of significance 就ok
當你睇番個statistics table, 你會發覺:
z table (normal distribution) 果版係唔需要用任何sample size
t-distribution 果版就需要用 df 同signifance level 兩樣野
當然, 計z 同 t 既 formula 係好相似, 當個 sample size 增加既時候, t-distribution 就會愈來愈接近 z-distribution .係統計學上,當個n係30或者大過30, 佢地個差距其實已經唔大, 至於點解係30 呢個數, 係一路讀落去就可以學到點prove.
希望幫到你~!
2007-03-28 06:30:16 補充:
z-test 係用population standard deviation 黎計而 t-test 係用sample standard deviation(因為我地唔知個population standard dev. ,所以採用sample standard dev.)佢地既 formula 係有少少分別架 :p