COMM 101 Lecture Notes - Lecture 10: Central Limit Theorem, Sampling Frame, Sampling Error

• School: U of M
• Department: Communication Studies
• Course: COMM 101
• Professor: Julia Lippman

Document Summary

The goal: a sample that allows us to draw conclusions about a population. Necessary condition: a representative sample: easiest way to obtain this: random sampling, you can only do truly random sampling if you have a list of everybody to sample from. Scienti c random: you have a group of possibilities and you don"t know which to choose, so you roll the die on what will get chosen. The law of large numbers helps to wash out random error. If we choose people from a population at random, we know exactly how much random error we are likely to have. With a random sample, all of our error is random, and none is systematic (bias) There is an actual value in the population but you don"t ever really know what it is. With sampling, you can be pretty sure the value you generate is close. Sampling error: the difference between your number and the actual (but unknown) value.