AP Statistics U4 Note

4.1 Population and Samples

Population: 总体
Sample: 样本
根据样本估计总体

样本调查 (Sample survey) 三个步骤：

1. What we want to describe
2. What we want to measure
3. How to choose a sample

怎样是不好的样本

1. Convenience Sample 为了方便获取最容易接触到的个体
   • Bias 偏见
   • Unrepresentative 不具有代表性
2. Voluntary Samples 自愿样本，持有强烈相同观点的人更容易做出回应

三种抽样调查 (How to sample well)

simple random sample (SRS) 纯随机抽

• Using technology choose
  1. label
  2. randomize – random generator

stratified random sample

1. classified the population – strata
2. choose a separate SRS from each group
3. combine thses SRSs

cluster sample 聚类样本

1. divided the population into smaller groups
2. each group could reflex the overall feature
3. choose SRSs from each group

Inference for sampling 从样本推断总体

• sampling error: 预期会有多少样本变异 –> control ramdom sample size
  • bad sample
  • convenience sample
  • voluntary
  • undercoverage
  • nonresponse

4.2 Experiments and Observational study

Observational study –> 得不到因果关系

observe and measure variables interested in but not influence the responses

观察个体互动，看特征和行为，但是不去干预，就看着

Experiemnt –> 能得到因果关系

impose some treatment to measure response

实施一些方法，测反应

the only source of fully convincing data

Language

treatment (干预) 手段，措施

experimental units 实验单位（被施加了treatment的最小个体集合），如果是人的话叫做 subject

描述解释变量的时候经常能见到 whethe，用于表示是否启用某个 treatment

有时候explanatory variable也叫做 factor，尤其是出现多种 explanatory variables 且有嵌套关系的时候

How experiment badly

• too simple design
• too many confounding variables

How experiment well

Confounding variables 混杂变量

与explanatory variable关联 & 影响response variable

当题目要找混杂变量的时候，需要解释所选的变量如何与explanatory variable关联且影响response variable