三种差异量数的比较 (The differences of three measures of variability)

1. 极端分数:全距受影响最大,四分位距受影响最小。
There are extreme scores in the data set: the full range is the most affected, and the quartile is the least affected.

2. 样本大小:全距可能随n增加而增加,四分位距和标准差不会。
Change the sample size: the full distance may increase with the increase of n, but the quartile distance and standard deviation will not.

3. 样本选取:同一总体多次选取不同样本,全距没有稳定的值,但四分位距和标准差是稳定的。
Select different samples: the same population selects different samples many times, the full distance has no stable value, but the quartile distance and standard deviation are stable.

4. 当存在不确定值的分布时,全距和标准差无法求得,四分位距可求。
When there is a distribution of uncertain values, the full distance and standard deviation can not be obtained, and the quartile distance can be obtained.

差异量数的优缺点 (Pros and cons of measures of variability)

优点 (Pros) 缺点 (Cons)
全距 (Range) 1. 计算便捷 1. 样本稳定性差
2. 受极端数值的影响
3. 可能与样本量有关
四分差 (Interquartile range) 1. 不易受极端分数的影响
2. 适用于有不确定值的数据
1. 在一定程度上样本稳定性差
标准差 (Standard deviation) 1. 样本稳定性好
2.包含最多的信息
1.受极端数值的影响
Pros Cons
Range 1. Easy to Calculate 1. Poor sample stability
2. Influenced by extreme values
3. Possibly related to sample size
Interquartile range 1. Less susceptible to extreme values
2. Suitable for data with uncertain values
1. Poor sample stability to some extent
Standard deviation 1. Good sample stability
2. Contains the most information
1. Influenced by extreme values