Class Notes (834,756)
Economics (1,614)
ECON 2B03 (45)
Lecture

# Lecture #7: Types of Bias, Association vs. Causation.doc

4 Pages
109 Views

School
Department
Economics
Course
ECON 2B03
Professor
Jeff Racine
Semester
Fall

Description
Lecture #7- Chapter 3 continued (Part 2) • Probability Sample: Occurs when a sample is taken with the help of a randomizing device that assures each member a known (positive, not necessarily equal) probability of selection • (in other words; a subset of a population that has been taken from a randomizing device) o Simple random sample: obtained if every member of a population has an equal chance of being in the sample (main type) o Systematic random sample: randomly select 1 element, then include every kth element thereafter till sample complete o Stratitified random sample: take random samples from every stratum (clearly distinguishable subgroups) in a population. o Clustered random sample: population naturally subdivided into geographically distinct units in which you will randomly select clusters, and then take a sample form each cluster. Errors in Survey Data • We expect that there will exist random error when we conduct a survey (also known as chance error or sampling error) –byproduct of taking a subset of a population • Arises only in sample surveys (ie: this will be 0 in a census) • Sample error equals difference between: o Value obtained from single random sample o Value obtained by taking a census (true value)what we would get if we actually took a census • Size of error (and its frequency) often estimated and reported with observed data (eg: “A census would reveal a number within 2% of the value”) Bias in Survey Data • Random errors are the least of our problems, and if even if we took a census, we might encounter serious problems • Systematic error (Non-sampling error): equals the difference between o Value obtained by taking a census o The true (but unknown value) • Problem with systematic error is that it is hard to detect • Bias can creep into a survey in a number of ways: o Can be built into the survey’s design o Can occurs during the survey’s execution (during survey) o Can transpire during the final processing (post survey) Selection Bias • Selection Bias: A systematic tendency to choose certain elements from a population and ignore other elements from a population. Ie: if you call someone in 3 in afternoon you are more likely to get someone not in the labour force; than in the labour force • Sources of Selection Bias o Nonrandom samples (eg: surveying all employees who leave work at 5 pm) o Faulty design of random samples th (eg: selecting every 12 month to survey monthly stock prices) o Faulty executions of perfect sampling plans (eg: interviewers substituting other persons for those randomly selected) Nonresponse Bias • Nonresponse Bias: Systematic tendency for some elements of a population in a survey to choose not to respond (not to contribute data). • Sources of Nonresonse Bias o Contact problems (eg: potential respondent tears up
More Less

Related notes for ECON 2B03
Me

OR

Join OneClass

Access over 10 million pages of study
documents for 1.3 million courses.

Join to view

OR

By registering, I agree to the Terms and Privacy Policies
Just a few more details

So we can recommend you notes for your school.