簡單隨機抽樣 (Simple Random Sampling)
分層隨機抽樣 (Stratified Random Sampling)
把母體先根據某種特性 (例如性別、年齡層、地區等) 去分群，然後再於每個群裡隨機去抽出受訪者。優點：可以確保這種特性中的每一種狀況 (如男性與女性) 都會有受訪者。缺點：用來分層的特性如果與要調查或測量的變項之間是沒有什麼關係時，可能調查結果會有較大的誤差。(也就是說，分群之後的組內變異過大，會造成抽樣結果的誤差變大)
集群抽樣 (Cluster Sampling)
把母體根據某種特徵 (例如班級、聚落等) 加以分群，然後抽出某一群或某幾群，對這幾群裡面的所有觀察體進行訪問或測量。優點：降低執行成本 (交通成本)，一般來說也會比較快速。缺點：容易出現集群效果，如果同一群裡的觀察體太類似了，調查的結果會有較大的誤差。 (也就是說，分群後的組內變異不大，反而是各群間的組間變異比較大的話，調查或測量結果就比較容易有較大的誤差)
- Anonymous1 decade agoFavorite Answer
[Sample sampling is defined ]
Utilize the appropriate method to take out limitedly from mother's colony (partially) Individual, is it for observe or research object that measure actually, come inference estimate characteristic, mother of colony by this to do. Purpose its in ' obtain proper size, and representative sample '.
Simple random sampling (Simple Random Sampling)
You have lists of a piece of parent, then pluck one by one from this list directly in way at random. The advantage is: Easily understood. The shortcoming is: Generally speaking, you can not have such a list, so can not carry out the simple random sampling at all.
Divide one layer of random samplings (Stratified Random Sampling)
A pair of parent, according to a certain characteristic first (such as sex, age level, area,etc.) Go to hive off, then take interviewees out at random in each one. Advantage: Can guarantee every kind of state in this kind of characteristic (such as man and women) There are interviewees. Shortcoming: If the characteristic used for dividing layer becomes with what has been investigated or measured between one while having no relation, perhaps the investigation result will have greater errors. (That is to say, hive off group after make a variation too big, will is it sample error of the result heavy to cause)
2007-12-16 12:05:26 補充：
Sample in cluster (Cluster Sampling)
A pair of parent, according to a certain characteristic (for example class, is it set to gather,etc.) Hive off, then take some one group or some several groups out,
2007-12-16 12:05:37 補充：
visit or measure to the body of these groups of observations inside. Advantage: Reduce and carry out the cost (traffic cost) ,Generally speaking will be faster too.
2007-12-16 12:05:55 補充：
Shortcoming: Apt to present the cluster result, if observation body in the same group is too similar, there are greater errors in the result investigated.
2007-12-16 12:06:01 補充：
(That is to say, hive off group after make a variation big, between-group variation heavy what of every, investigate or it is apter to have greater errors to measure the result)Source(s): 自己