Statistics Section 2-3 Day 1 - Sampling in the Real World
Sampling Errors 1. Natural Errors (Random Sampling Error) Caused by Random Sampling Error...Caused by taking a sample that cause sample results to be different from a census Example – Every person in class has a different list of names to go and collect data from 2. Undercoverage Groups of the population are left out of the process Often the list of the entire population is rarely available Example – homeless, groups of individuals, etc…
Sampling Errors If the Sampling frame (list of individuals from which the sample is chosen) leaves out certain classes of people, even an SRS will be biased Example – Sampling frame - Telephone directory for a telephone survey. Who is left out due to sampling error? Unlisted numbers and cell phone numbers
Non-Sampling Errors Processing Error ~Mistakes in handling the data ~Mistakes in mechanical tasks such as basic arithmetic or entering responses incorrectly ~Computers have made these mistakes less common, but they still exist.
Non-Sampling Errors Response Error ~Occurs when an individual(s) or subject(s) gives incorrect response through lies, lack of memory, or does not understand a question so they guess ~Often occurs with sensitive issues (age, income, drugs, etc...)
Non-Sampling Errors How can we reduce response error? ~Wording needs to be unbiased ~The administer of a survey needs to be unbiased Ex - Given the fact that those who understand Statistics are smarter and better looking than whose who don’t, don’t you think it is important to take a course in Statistics?
Non-Sampling Errors Non-Response Error ~Failure to obtain data from an individual selected for a sample ~Occurs because they cannot be contacted, refuse to cooperate, refuse to answer Causes bias because different groups of people have different rates of non-response (higher in large cities and with the elderly)
Non-Sampling Errors How can we reduce non-response? ~Calling back multiple times over longer periods of time ~Letters sent ahead of time informing that the survey will take place ***However - this slows down the surveys