Why sample? - Because we can't ask anyone unless we are doing a census (only every 10 years) - Small group that approximates the population Probability Sampling: - Everyone has a known non-zero chance of being selected - For me to know the probability of being selected I must know certain things about the population - Must know total population before I can know your chance of being selected - Poll the class- I know the probability because there are 90 students - 15/90 - Major limitation: we don't always know the population (ex: how many sex workers in Vancouver- therefore cannot do probability sampling) - Probability must be non-zero- everyone must have equal chance of being selected - Why do probability? - representativeness - We can draw a sample that looks like the population - Common in quantitative research- numbers Non-Probability Sampling - Describe, not looking for representativeness - Qualitative studies - Reduce bias- larger sample, representativeness Two big errors: - Sampling error: - We can't eliminate because every time we draw a sample, it is never going to match population perfectly - Want to reduce sampling error - How do we reduce sampling errors? - Reduce sample size- large the sample size, the smaller the sampling error - A homogeneous population- ex: in class most people are under 25, middle class, perspectives are likely to be similar - Coverage error: - Make sure there are not pockets of people who are excluded - Ex: survey on students perception of Oral, if he comes to class and gives out surveys, he has coverage error, because some student might not be in class. If he put it online there's a greater chance he'll cover everyone - Ex: phone polls- someone doesn't have a phone, using phonebook- someone isn't listed Example for drawing a probability Sample - Ex: Want to measure if our class supports the USMCA - Can do probability sampling because: we know the exact population of the class Basic probability Sampling Jargon
Representativeness: - Same characteristics are common across the population Elements: - Who are we making our inferences about Target population: - Who we want to study Sample frame: - List of all the members of our population- ex: class list - Polls may use electoral registry or phonebook - Want to make sure it's up to date (phonebooks not representative) Sampling unit - Same as elements, the people we are going to be researching Study population - Same as target pop except it's the list of people we are going to be studying Distinguishing between parameter and statistic: * MIDTERM* - S=S P=P - Statistic is always in reference to a sample- never the population - Parameter is always in reference to a population A researcher wants to know if residents in BC support or reject the new USMCA agreement 1. What would you suggest she uses as her: a. Target population A. Adult residents of BC b. Sampling frame A. Electoral registry, CRA, phonebook, MSP