POL SCI 3 Lecture Notes - Lecture 23: Sampling Bias, Sampling Error, Missing Data
22 Nov 2016
School
Department
Course
Professor
Document Summary
Equal ps: every unit has a known probability of selection. Statistic: summary feature in sample / parameter: summary feature in population. Affected by sample size (major factor): larger size, lower error: nonlinear, decline and flatten, not affected by population size, error in small in sample size of 5000. Sampling error extent to which the statistic departs by the parameter by chance. Multistage area probability sampling: exclude homeless, people living on ocean, people. Bias in the frame is much more common frame on military bases: ex. Rdd: exclude people who do not have a phone: ex. Ra who is sampling state senators and excludes nebraska because she is anti-nebraska. Problem in the sampling procedure selected: ex. Ra forgot to include last 100 people on the list: rare in real life, ex. Balls in an urn very difficult to mix properly: whatever goes in last is more likely to be: ex.
