PsychStatsWiki

本页面只读。您可以查看源文件，但不能更改它。如果您觉得这是系统错误，请联系管理员。
====== 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  |